Rural Settlentents • ID Medieval Europe Papers of the 'Medieval Europe Brugge 1997' Conference Volume 6 edited by Guy De Boe & Frans Verhaeghe LA.P. Rapporten 6 Zellik 1997 I.A.P. Rapporten uitgegeven door I edited by Prof Dr. Guy De Boe Ru al Settlements .. Ill Medieval Europe Papers of the 'Medieval Europe Brugge 1997' Conference Volume 6 edited by Guy De Boe & Frans Verhaeghe I.A.P. Rapporten 6 Zellik 1997 Published by the Een uitgave van het Instituut voor het Archeologisch Patrimonium Institute for the Archaeological Heritage Wetenschappelijke instelling van de Vlaamse Gemeenschap Departement Leefmilieu en Infrastructuur Administratie Ruimtelijke Ordening, Huisvesting en Monumenten en Landschappen Doomveld Industrie Asse 3 nr. 11, Bus 30 B -1731 Zellik - Asse Tel: (02) 463.13.33 (+ 32 2 463 13 33) Fax: (02) 463.19.51 (+ 32 2 463 19 51) DTP: Arpuco. Seer.: M. Lauwaert & S. Van de Voorde. ISSN 13 72-0007 ISBN 90-75230-07-9 D/1997 /6024/6 Scientific Institution of the Flemish Community Department of the Environment and Infrastructure Administration of Town Planning, Housing and Monuments and Landscapes 06 RURAL SETTLEMENTS - DE LANDELIJKE WERELD - LE MONDE RURAL LANDLICHE SIEDLUNGEN was organized by werd georganiseerd door fut organisee par wurde veranstaltet von Elisabeth Zadora-Rio Marnix Pieters PREFACE It can reasonably argued that fieldwork and research concerning the rural world is one of the oldest branches of modem medieval and -be it to a somewhat more limited extent -later archaeology. In view of the importance ofthe rural component in medieval and later society, this is hardly surprising. It can also be argued that the emergence and development of this particular field of archaeological research is strongly indebted to questions raised by historians and related to a particular problem, to wit the- mainly late medieval- desertion of villages. The names of scholars such as Maurice Beresford, John G. Hurst, Walter Janssen, Stanislaw Tabaczynski, Jean-Marie Pesez, Gabrielle Demians d' Archimbaud and many others speak for themselves, as do those of sites such as Wharram Percy in Yorkshire. The early archaeological interest in medieval deserted villages also led to specific research groups, amongst which the British D. M. VR. G. or Deserted Medieval Village Research Group immediately comes to mind. The D.M. VR. G. was of course mainly concerned with the British Isles but its evolution over Medieval Village Research Group to - after joining with the Moated Sites Research Group, which focused on yet another type of mainly if not exclusively rural settlement- Medieval Settlement Research Group does provide a nice illustration ofhow the field developed, both in the United Kingdom and on the European Continent. Indeed, since the 50s and 60s, the archaeological study ofthe medieval and later rural world has moved on and this in more ways than one. The work no longer focuses almost exclusively on deserted medieval villages and the field has expanded in many directions to encompass all kinds of new issues. These include not only the many different forms of rural settlement isolated, nucleated or agglomerated- their variations, their origins and their development through time but also their setting, the landscapes in which the settlements developed and how they changed those landscapes, and the intricate relations between rural settlement and other material components of medieval and later society such as towns, castles, monasteries or other kinds of ecclesiastical, military and/or residential units. The archaeological study of individual buildings and hamlets or villages continues to form the backbone of the work and case-studies continue to abound. But since the 50s and 60s the field has also changed in other ways. Over the past three decades, the attention paid to issues related to the different kinds and levels of spatial organisation has grown continuously and is arguably the main or at least the most seminal development to be noted. Specifically over the past 15 years, intra-site approaches have increasingly been complemented with inter-site, local, micro-regional and regional ones; simultaneously, more and more attention now also goes to site hierarchies and spatial organisation on a much larger scale as well as to the spatial organisation of the rural landscape in terms of estates, holdings, plots, and what can reasonably be termed catchment areas. In addition, the growing impact of the natural sciences and more particularly of the environmental studies is also making itself felt - though still far less obviously than in the case of other categories of settlements. At the same time medieval and later rural archaeology seem to have moved away from the older deterministic approaches and interpretations. Rather, the development of the many forms of rural settlement and the ways in which they were influenced by and changed the local and regional environmental, social and economic setting is more and more looked in terms ofhuman behaviour and interactions. Again- and as with the other themes discussed at the Brugge conference- the subject of rural settlement cannot and should not be divorced from the many other concerns of medieval and later archaeology. The sheer complexity of the subject entails inevitable links with many other topics presented and discussed within the context of the other sections of the MEDlEY AL EUROPE BRUGGE 1997 conference, among them material culture (section 07), trade and exchange (section 03), environment and subsistence (section 09). But the rural world also includes and interacts with specific types of sites and settlements such as towns (section 01), castles and defense-works (section 11) and religious buildings (section 04) and cemeteries (section 02). In addition, art and architecture, including the meanings and the perception thereof are not absent from the rural world (section 05), while the environment and the subsistence problems (section 09) equally constitute major concerns. Some aspects of transportation and travel technology and organization (section 08) should not be forgotten either, and, finally, spatial analysis- the importance of which has already been emphasized earlier- raises many theoretical and methodological problems relevant to the study of the rural world and some of these have been discussed within the context of section 10 (Method and Theory in Historical Archaeology) of the Brugge conference. All this sufficiently illustrates the need for bringing together the information related to the many aspects of the medieval and later worlds and societies, which was one of the basic philosophies of the York and Brugge MEDIEVAL EUROPE conferences. For sheer practical reasons, however, a number of contributions presented on the occasion of the international conference on medieval and later archaeology MEDIEVAL EUROPE BRUGGE 1997 which took place in Brugge, Belgium, on 1 through 4 October 1997, have been grouped in section 06 under the general heading Rural settlements -Le monde rural- Landliche Siedlungen- De landelijke were/d. The section was organized by Elizabeth Zadora-Rio (C.N.R.S., France) and Marnix Pieters (Institute for the Archaeological Heritage, Flanders, Belgium). The present volume offers a collection of pre-printed papers, a number of which were presented orally and debated during the sessions of section 06. Unfortunately, a number of contributors to this section did not submit a text in time for inclusion in the present volume while other colleagues could not attend and present their contribution. In a few cases, texts were graciously made available for inclusion in the present volume but due to the large offer of papers could not be presented orally. All this explains why the general structure and the contents of the present volume do not conform in all details to the programme of the conference. Nevertheless, the volume has been organized keeping in mind both the complexity of the subject and the general lines of the structure of section 06 of the conference as originally proposed by the organizers. The texts available for the present volume have therefore been grouped in five sections, which at the same time largely reflect the current concerns and state of research in the archaeological study of the medieval and later rural world: - A first subsection, sub-titled Regional settlement studies, groups eight contribution which present regional surveys and/or an assessment of the current state of the work and future issues related to the rural settlement of larger areas. - The second subsection groups twelve studies concerning specific sites and has been sub-titled Case studies. They reflect the continuing importance of this type of work which offers ever more building blocks for new questions and approaches, including regional and even supra-regional ones. While focusing of specific sites, these studies often also address broader issues relevant to the individual site in question and should therefore not be considered simple site-reports. - With four contributions, the subsection Territorial organisation and central places looks at relations between different sites and types of sites as well as at relations between different components of a territorial unit. - Three papers, all of them concerned with the Iberian peninsula and with the specific subjects of silos, storage and water management as well as with their potential for understanding the often regional organization and development of rural settlement, have been brought together in a special subsection Settlement and rural infrastructure which follows the previous subsection because the notion of 'organization' is very much present in the comments. - Finally, five contributions brought together in the subsection Settlements and landscapes return to the theme of the broader context and setting of rural settlements, in terms of either chronological development or in terms oflandscape development and organization. Of necessity, the papers are rather short and the volume of course does not do total justice to the many excavations and the wealth of other types of research work concerning rural settlement, the rural landscape or some specific types of rural sites. Thus, for instance, mills and what has been termed 'moated sites' are largely if not totally absent from the picture. Nor does the volume provide a complete overview of the results attained and knowledge acquired. Nevertheless, the 33 papers included in the present volume emphasize the continuing importance and vitality of the study of the medieval and later rural world while at the same time providing a good idea of the potential and of the present state of the work in this particular field of research. They also demonstrate both the complexity of the subject and the recent developments in terms of spatial approaches. This is even more true when the volume is considered within the context of the other volumes in the present series of volumes linked to the Medieval Europe conference and when the reader takes into account that the rural world is also very much present - directly or indirectly - in these other volumes. Frans Verhaeghe & Guy De Boe 6 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Bruggc 1997' Conference- Volume 6 Matej Ruttkay Early Medieval Settlement Structure in Slovakia The tradition of archaeological research on medieval Slovakia is more than 50 years old. During this period, scores of studies have been published and excavations carried out, which resulted in a vast source base. The macro-region under discussion is oflowland character with elevations ranging from 94 to 300 m a.s.l. Approximately one sixth of the region is situated at an elevation of over 300 m a.s.l. The existence of man in this region in the past was determined by several factors. The density of settlement was to a significant degree determined by the quality of local sources of energies, water sources being among the most important ones. Other important factors are the quality of soil, thermal energy, slope tilt, etc. Undoubtedly, a very important role was played by social and economic relations in the individual societies and the global geopolitical situation. Moreover, the development of the settlement process was determined by the occurrence of mineral resources (iron ore, stone, ores of non-ferrous metals, etc.). At the present time we have at our disposal information from more than 2100 sites from within the macro-region. More than 1800 of them have been documented archaeologically (66% of these are settlements, 26% burial places, 5.6% sacral buildings and 2.4 % are not documented in a detailed way). The cartographic evaluation of the sites discovered in individual decades reveals a certain lack of balance of research in certain regions. The rise in the number of archaeological sites from 1950 to 1995 is illustrated in Fig. 1. While about 320 sites were known in 1950, by the year 1996 this number increased to 1995 sites. As far as the way of acquisition of information is concerned, the absolute majority of sites were identified through surface surveys (998, i.e. 45 %) and in the surveys carried out during construction activities (376, i.e. 17 %). Together, these two ways represent 62 %. The second largest group consists of localities where minor conservation actions were carried out (601, i.e. 28 %). In 105 (5 %) sites major conservation actions or systematic investigations took place. The rest of the finds represents isolated finds or objects from different collections which cannot be localized in a more exact way. As far as dating is concerned, the localities can be divided into three groups: 1) relatiyely reliably dated (larger collections obtained mostly by systematic research) - 31 %, 2) less reliably dated (e.g. larger collections of sherds from investigations, etc.)- 63 %, 3) approximate dating (isolated, less pronounced finds)- 6%. The majority of finds (84 %) are reliably localized, 11 %are approximately localized and the localization of 5 % of the finds is impossible. The settlements from the 6th to 12th centuries in southwestern Slovakia include a broad typological range of objects Taken together, 2978 objects have been documented. Unfortunately, only limited or no information is available for a large number of them. The basic classification is purely typological with respect to the appearance of objects and not to their functions. The database which has been worked out gives different numbers of individual objects: I. buildings- houses (647 objects), n. isolated heating devices (376 objects), III.A storage pits (559 objects), III.B cooking pits (127 objects), III.C other pits (282 objects), 1600 1400 1200 1000 800 1-~------- 600 400 200 0 0 LO 0> 0 (!) Ol 0 0 Ol 0> t'- CO 0 0> 0> .....- LO 0> 0> Fig. 1. - Western Slovakia. Gradual growth of known archaeological sites from 1950 to 1995. 7 M. Ruttkay :'·,· I ·' !- .:~ [ Y:./ 0 nad 300m 25km L _ _ _ _ _....c Dunaj Fig. 2.- Settlement of western Slovakia in 6th century (1. Slavs). IV. troughs (50 objects), V. wells (15 objects), VI. manufacturing objects (101 objects), VII. other objects, not specified (334 objects) For almost 500 other objects (sites), no information is available and these were not included in the reported work at all. During the whole studied period a relatively high proportion of finds belong to houses. From the 6th to the 12th century, the proportion of storage pits and isolated clay ovens gradually rises, while the number of cooking pits decreases. The representation of the remaining types of objects shows no significant fluctuations. The principal type of house in the 6th-12th centuries is represented by slightly sunken or aboveground objects of a quadrangle ground plan (surface: 6-20m2 ; most frequent depth: 30-60 cm). When compared to surrounding regions of Central Europe, the presence of a considerable variety of these houses is a specific feature of southwestern Slovakia. it is a feature which is very pronounced, particularly in the 6th-1Oth centuries. During the period under discussion, the heating device used most frequently in the house was the stone 8 oven. Open fireplaces were used somewhat less frequently, while clay ovens occurred the least. From the chronological point of view, the representation of the fireplace is the most stable one. Stone ovens occur most commonly in Horizon I and II (6th-10th centuries), while in the next period their occurrence drops significantly. Clay ovens reveal a reverse development and their occurrence rises especially from the 11th century onwards. The utilization of a specific type of oven depended first and foremost on its function and local sources of building materials were of minor importance. The analysis proved that the theory of ethnic determination of clay oven occurrence is unjustified (these ovens were found even in contexts with the Prague Type ceramics). An exception may be the clay ovens of the 11th-13th centuries, which markedly 'penetrated' the region of the middle course of the river V ah and the lower part of the Dudvah basin, i.e. the regions where they had not been found before. This fact may be possibly connected with the migration of populations from the middle course of the river Nitra and the lower course of the river Hron to this territory (the retreat of inhabitants before the expansion of Old Magyars?). Early Medieval Settlement Structure in Slovakia ! · >;:::j 0 \___ nad 300m ___ 25km Fig. 3.- Settlement of western Slovakia in the middle of 7th century (1. Slavs, 2. Avars). The cartographic demarcation of the sites dating from the 6th century does not give a picture of the main spheres of interest of the oldest Slavonic groups (Fig. 2). There was a moderate concentration found in the region of the middle part of the Nitra basin, in the basin of the Vah and in Zahorie (western Slovakia). In the first half of the 7th century, the area of the middle course of the river Nitra seems to have been again the most densely settled one. In the remaining territory of the region, only sparse islets were settled. A relatively thin network of settlements represented by the Prague Type ceramics is supplemented, like in the previous phase, by a system of cremation burial places. In the close proximity to the Danube a new phenomenon occurs: the first burial sites with skeletons (Fig. 3). In the second half of the 7th century, the number of settlements increases markedly (Fig. 3). The boundary of localities represented by the Prague Type ceramics (Slavonic settlement) moves southward to the line Bratislava- Galanta- Bajc - Chl 'aba. At the same time, settlements characterized by a different material culture (Avar Kaganate) were being founded on the left bank of the Danube - Sturovo - Obid (Zabojnik 1988) and Komarno (Trugly 1993). Between these groups an uninhabited (buffer) strip of land some 20-40 kilometres wide remained open. This is an important discovery proving that between the Slavonic world north of the Danube (western Slovakia) and the Avar Kaganate, there were no closer links, either peaceful or warlike. In this connection the question of the localization of the Empire of Samo appears in a new light. It is very likely that the empire of the Slavonic tribes, led by the Frankish merchant Samo, was established in the region of immediate contacts between the Slavs and the Avars. As this relation is absent in the studied region, it is likely that the crystallizing core of Samo 's Empire cannot have been the region of western Slovakia (with the exception of its westernmost edge). The situation changed more markedly at the end of the 7th century and in the 8th century. After the relatively thin and sparse settlement in the 6th-7th centuries, a pronounced rise in the total number of settlement units was recorded (Fig. 4) and the density of population increased. Besides 82 localities dated back to the 7th-8th centuries or to the 7th-9th centuries respectively, 112 localities dated back to the 8th 9 M. Ruttkay ·· i nad 300m ~~~· 0 25km Fig. 4.- Settlement of western Slovakia in 8th century (1. Slavs, 2. Avars and Slavs, 3. Avars). or to the 8th-9th centuries belong to this group. The most densely populated areas - especially in the last third of the 8th century- belong to the region of the middle course of the river Nitra, the area demarcated by the triangle Sered' - Senec- Galanta, the left bank of the Morava river between Bratislava and Skalica and the middle course of the river V ah. A relatively dense network of settlements is documented in the basin of the Zitava river, on the upper course of the river Nitra and on the middle course of the river Ipel' (near Sahy). As far as settlements are concerned and on the basis of the overall habitus and material culture, nearly the whole region studied seems to be a uniform unit linked up to an older settlement. An exception here are the southernmost areas where besides the typical ceramics of the Danubian type, the material culture includes a type of coarser ceramics (so-called Avar ceramics) together with a high quality type of grey ceramics made on the potter's wheel (Srurovo, Obid, Komarno). As follows from the above-mentioned facts, in the 8th century, the settlement network gradually becomes denser and regions with a relatively compact settlement pattern were being established. At the end of the 10 8th century, the nuclei of economic and social centres - hillforts - were gradually established (Pobedim, Majcichov, Nitra, Bratislava). They are generally preceded by a gradual accumulation of settlements in a relatively small region. In the 9th century, the Slavonic substratum became predominant. As it follows from the total number of known localities, the increase in comparison with the previous horizon represents a significant change per century. The overall character of the settlements and ofthe material culture discovered is of the same type throughout the whole region. Due to the lack of significant chronological elements it is very difficult, for the time being, to distinguish the objects and localities from the 9th and 1Oth centuries reliably, which would be very important from a historical point of view. On the other hand, the pronounced similarity of the material culture from the 9th and 1Oth centuries clearly negates the theories about the destructions by or at least the destructive arrival of Old Magyar military companies. On one side there was a clear tendency towards the extension of settled areas, on the other side the gradual concentration of the settlement network in Early Medieval Settlement Structure in Slovakia l . /\ nad 300m o 25km L___ _ __ [l]]]J 2 nai Fig. 5.- Settlement of western Slovakia in 1Oth century (1. Slavs, 2. majority of Old Hu~garians, 3. majority of Slavs). the regions which had already been settled in the previous horizon took place. At the same time, some territories became almost devoid of settlements or were settled sparsely, as with the larger part of Zitny ostrov on the south and a relatively large area between the rivers Zitava and Hron reaching almost up to the Pohronsk)' Inovec Mountains in the east. In the lOth century, it is possible to document the boundary of the Old Magyar interference (the river Dudvah in the west, the line Hlohovec - Nitra - Levice, Fig. 5, in the north) also in material culture (graves with armaments and a horse harness - sabre, sword, bow, bit, stirrup). Their establishment is a result of vast social and economic changes in the Middle Ages. On one side there were huge organizational and administrative units built by a higher feudal power (residential, organizing, military and administrative functions), on the other side there were smaller units often performing a wide range of functions (defensive, economic, administrative etc.). The majority of central settlements or settlement areas respectively kept their role in the following centuries. Archaeological and also historical sources document a marked influence of Christian faith on the forming of contemporary society. This is linked with the origins of church building in the territory of present-day western Slovakia. The occurrence of stone sacral objects is sporadic and it is linked with the presence of the highest social strata. So far four (Bratislava - Castle, Bratislava - Devin, Ducove, Nitra- Martinsk)' vrch) or six respectively (Nitra- Na vrsku, Nitra - Castle) sacral buildings have been convincingly dated to Horizon II. All of these sacral buildings are directly linked with hillforts. As follows from the above, the principal tendency towards the densifying and expansion of the settlement network in the 9th-lOth centuries took place in the northern half of the region studied and was directed to the north. It means that for some reason, the contemporary inhabitants preferred localities situated relatively higher to apparently more favourable places (from our point of view) in the southern part of the Danubian basin. The expansion of the Old Magyars was directed at a relatively sparsely settled area and did not aim at having a direct destructive impact on the settlement structure. 11 M. Ruttkay H:::::::::j:)jl ~km 0 I nad 300rn 1 Dunaj Fig. 6.- Western Slovakia. Distribution map offortifications within the macro-region (1. castle, 2. moated site). After the break-up of the Great Moravian Empire, the settlement declined in the 11th and in the first half of the 12th century when the territory became part of the Early Hungarian feudal state. At the beginning of this period, most of the hillforts ceased to exist and only those of primordial importance survived and later became centres of power and administration (Nitra, Bratislava, Stary Tekov). This situation may be explained by the inability of the Slavonic economic and administrative structures of the 9th and 1Oth centuries to accommodate to the newly formed Early Hungarian establishment. When compared with the previous horizon, the total area settled does not change significantly. The whole region of the basin of the river Nitra remains the central settlement area. A characteristic phenomenon is a drop in the density of the settlement network in Zahorie and a mild concentration in the southern parts of the region (eastern part of Zahoria, lower part of the Hron basin, the basin of the river Zitava), in the central part of the Nitra basin and in the basin of the river Vah. At the same time a noticeable decrease in the settlement density in the upper part of the Zitava basin and in the Nitra basin in the 12 area between Nitra and Nove Zamky can be discerned. A conspicuous decrease in the number of settlements - the destruction of the settlement network can be discerned only in the western part of the region. This development was caused by several factors- adverse climatic conditions (rise in the levels of water courses), decreasing soil fertility, permanent conflicts with western neighbours. In the 11th century, the border of the Hungarian Kingdom was situated on the river Vah (the Dudvah respectively). The western part of the region probably belonged in the political sphere of the Bohemian Kingdom. In the 11th century, monasteries were established on the northern edge of the Old Magyar dominion (Nitra, Skalka Hronsky Benadik) which played an important role in the conquest of the northern part of present Slovakia. The 11th century is the time when hillforts could no longer find their place in the qualitatively new settlement structure. The only exceptions are the central hillforts which were gradually transformed to comitate castles (Nitra, Bratislava, Stary Tekov). These can already be considered to be real castles. Early Medieval Settlement Structure in Slovakia nad 300m 25km 0 @1 ti.§l2 --3 Fig. 7.- Western Slovakia. Towns with privileges and their economic background (1. town, 2. market). During the 11th century, the whole region became unified from the cultural point of view- the material inventory obtained a supra-ethnic character and it is almost impossible to trace any ethnic differences. The fact that beside the Slavs a considerably high number of other ethnic groups were found in this region is documented by written sources. The multiethnic character of southwestern Slovakia is documented by the investigation of the meso-region (the Slavs, the Hungarians, the Kalisz, the Cuman, the Pechenegs). By the end of the first half of the 12th century, the situation at was already consolidated. Together with economic development, settlement started to increase again. This process is directly connected with the gradual rise of towns. Increasing social differentiation within the early feudal society activated the building of castles (most often they were small fortified areas with a wooden or stone tower-shaped construction). The castles served military purposes as well as being symbols of the rising nobility. The settlement complexes show considerable stability and generally they contain several 'central points', a church or from the 13th century onwards also residences of local nobles. In southwestern Slovakia 43 fortifications of the lower nobility have been documented so far and another 12 are uncertain (Fig. 6). The dating of their construction back to the 11th century is more or less hypothetical, without any convincing evidence. Eleven registered fortifications are supposed to have been built at the end of the 12th and in the first half of the 13th century. The locality Partizanske Simonovany and the wood-and-earth fortification in Topol'cianky are the most pronounced examples. In two cases, stone constructions are documented; in the remaining localities wooden structures were probably erected. The boom in the construction of small fortifications started as late as the half of the 13th century, i.e. after the Tartar invasion, when the need for a more thorough defense became obvious. One of the results of the structural change of original dominion royal property in Hungary is the formation of towns (Fig. 7). The process of medieval town formation can be summarized as follows (Kucera 1978): 1) up to the end of the lOth century: a period of large popular and densely populated towns of the 13 M. Ruttkay H:::::~:::H o nad 300m 25km Dunaj Fig. 8.- Western Slovakia. Early Medieval settlement (6th-13th century), general distribution map. castle town type with a pronounced political and military function and with an economy which was strongly subordinated the sovereign's power; 2) from the 11th century up to the year 1242: a period of the formation of large towns of different types on the basis of older settlement and town cultures and centres; political, administrative and military functions gradually recede to the background and the functions of commerce and merchants' guilds together with the formation offree town artisans developed in a pronounced way; 3) from 1242 to 1306-1309: a period of the most productive growth of medieval towns. The social and economic function of these towns is a characteristic feature; the towns were now acting as the implementors of merchandising and monetary relations; the free mercantile guild or entrepreneurial activity respectively (commonly characterized as 'ius fori') came to the fore. The whole of economic and social life was subordinated to the town community and the town also acted as a social and judicial institution to the outside. The town was often built and developed according to the proved organizational forms known abroad. 14 Town privileges started to be given to the settlements of an urban nature as late as the second third of the 13th century. Five (or six respectively) towns are documented in the studied macro-region: Tmava - 1238 Zumbothel, Statj Tekov - 1240 Suburbium castri de Bars, Nitra- 1248 Nitra, Komamo - 1265 Camarum, Bratislava- 1291 Posonium, Nove Mesto nad Vahom 1253 (? ). The comparison of archaeological sources and excavations with historical sources shows that archaeological excavations and research are the main information sources up to the end of the 12th century. Starting from the 13th century, both types of archaeological information are balanced and later historical sources gradually prevail over archaeological ones (Fig. 8). The cartographic evaluation of finding localities also allowed us to identify the most important crossings - fords on the river Danube and so-called 'main roads-. The most important fords were located near Srurovo, Patince, Komamo and Bratislava. From the above-mentioned facts, it appears that in the 6th to 13th centuries the lowland regions were most densely settled areas. The mapping of find- Early Medieval Settlement Structure in Slovakia spots and localities indicates that the southern part of the region was relatively less densely settled (Zitny ostrov, the areas between the rivers Zitava and Hron and between the Nitra and the Vah). In the areas which are the most important from the point of view of settlement development (the basins of the Zitava and the Nitra), settlement density gradually increased during the whole of the period studied. Literature M. 1978: Geneza miest na Slovensku, Archaeol. hist. 3, Bmo, 147-164. TRUGL Y S. 1993: Griffek oroszlanok nepe, Pozsony. ZABoJNiK J. 1988: On the problems of settlements of the A var Khaganate Period in Slovakia, Arch. Rozhl. 40,401-437,480. KUEERA es Dir. Matej Ruttkay Archeologicky Ustav Sav Akademicka 2 94921 Nitra Slovensko 15 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 M aria Beres Regionale Beziehungen der Dorfsiedlungen Siidostungarns vom 10. bis 13. Jahrhundert In diesem Referat wird eines der Elemente des si eh zwischen dem 10. und 13. Jahrhundert in der AlfoldRegion des Karpatenbeckens entfaltenden Siedlungsnetzes untersucht, namlich die Dorfer in einer engeren Region zwischen der Koros- und Maros-Mi.indung beiderseits der TheiJ3. Im Rahmen der ungarischen dorflichen Ausgrabungen, die in unserem Komitat nicht planmal3ig und systematisch durchgefiihrt wurden, wurden die innere Struktur und die Objekte dieser in erster Linie untersucht. Dementsprechend, also wegen der Zufalligkeit des archaologischen Quellenmaterials konnte ich die Aufmerksamkeit der Analyse des sich entwickelnden Dorfnetzes nur in bescheidenem Ma13e zuwenden (Abb. 1). In der ersten Halfte unseres Jahrhundertes entfaltete Gabor Csallany eine regelmal3ige Tatigkeit m it der Freilegung von Graberfeldem aus dem 10. bis 12. Jahrhundert, was das Sammeln des mittelalterlichen archaologischen Denkmaterials anbelangt. Als Marta Szell die mittelalterliche Siedlungskarte einiger Gebiete dieses Areals skizzierte, wendete sie seine Ausgrabungsergebnisse, femer die Dokumentation einiger Kirchenfreilegungen und die urkundlichen Angaben an. In dieser Region finden archaologische topographische Arbeiten mit kleineren und gro13eren Unterbrechungen ab 1977 statt, femer aul3er den Details mehrerer Dorfsiedlungen wurde eines der Verwaltungs- und Kirchenzentren aus der Staatsgri.indungszeit: Szer erschlossen. Abb. 1. - Die Lage das Gebiet Komitat Csongrtid. 17 M. Beres -- Abb. 2.- Funde des Dmfes Szer aus dem 9. Jh . (nach Katalin Vcily1). .................... 0 9 ······ ······ 11 6 Zur Bestimmung des arpadenzeitlichen Siedlungsbestandes bieten die regelmal3igen archaologischen Gelandebegehungen aul3er den Urkunden die meisten Angaben, Ober die innere Struktur und die Bauten dieser Siedlungen erwarben wir Kenntnisse durch Ausgrabungen und Analyse van Luftaufnahmen. In meinem Referat werden drei Siedlungstypen ausftihrlich analysiert. 1. Aufgrund von Urkunden datierbare, zur Staatsgri.indungszeit sich auf dem koniglichen Burggut entwickelnde grol3e Dorfer mit einer Kirche. Die Bewohner beschaftigten si eh mit Ackerbau und Tierzucht. Solche Dorfer existierten vom 10. his 13. Jahrhundert kontinuierlich. Fallweise haben diese Siedlungen- in unserem Area!- (bulgarische, slawische) VorHiufer aus dem 9. Jahrhundert, und die Bevolkerung dieser tilgte die Steuer durch spezielle Dienstleistungen (Csomorkany, Sap, Szer, Abb. 2-3). 2. Kleinere Siedlungen mit Kirche, Ackerbau und Tierzucht treibende Bevolkerung. Ober diese stehen uns keine Urkunden zur Verfi.igung. In der untersuchten Periode bestanden sie kontinuierlich. 18 3. Kleinere Siedlungen, Dorfkeime. Ihr Bestehen ist nicht kontinuierlich: Zeitweise entvolkerten sie sich, dann wurden sie emeut besiedelt. Keine Urkunden stehen zur Verftigung. Aufgrund der archaologischen Funde und Objekte konnen sie von der zweiten Halfte des 10. Jahrhunderts an datiertwerden (Tape, Abb. 4). Im von uns untersuchten Gebiet (Ungam, nach der heutigen Verwaltungseinteilung: Komitat Csongrad) sind die untersuchten Dorfer aus dem 10. his 13. Jahrhundert von ungi.instiger Lage, was die Moglichkeit der archaologischen Beobachtungen betrifft: Sie befinden sich namlich im Weichbild oder in der, unter landwirtschaftlicher Bestellung stehenden Gemarkung von heutigen Siedlungen. Vollkommen konnte die Struktur keines Dorfes- unabhangig von der ZugehOrigkeit zu einem der angegebenen Typen - geklart werden, da es keine, die ganze Siedlung untersuchende, komplette Freilegung gab. Darum ist unser Vergleichsmaterial nicht komplett, weder was den Vergleich des Siedlungsdetails ahnlichen Alters in anderen Regionen noch was die Beobachtung der strukturellen Veranderungen innerhalb der einzelnen Regionale Beziehungen der Dorfsiedlungen Slidostungarns vom 10. bis 13. Jahrhundert Abb. 3. - Funde des Dorfes Sap aus den 9-12. Jh. (nach Maria Beres). . . -- D. .. '. _-~: ., . .. . . Siedlungen betrifft. Es scheint doch, da/3 die Dorfer des Typs 1 durch ein inneres Grabensystem gegliedert wurden. Die si eh danach gerichteten Ha user und die umliegenden Speichergruben wurden in Reihen angeordnet. Fur die innere Struktur der Dorfer, gehOren sie zu irgendwelchem genannten Typ, ist die innere Bewegung am kennzeichnendsten (Abb. 5). Unter den Ansiedlungsfaktoren der untersuchten Dorfer spielten die Besitzverhaltnisse au/3er den naturgeographischen Gegebenheiten (Su/3wasser, ErhOhung, Weg, Flu/3ubergang, Rohstoff usw.) eine bestimmende Rolle. Die Umstande und Ursachen ihrer Zerstorung oder ihres Fortbestehens mussen aber sorgfaltig erwogen werden. Die Zahl der Dorfer des Siedlungsnetzes im 13. Jahrhundert nahm nach dem Ende der behandelten Epoche bedeutend ab. Das Verschwinden der Dorfer kann nicht ausschlie/3lich an den Tatarensturm 1241-1242 gekntipft werden, sondern die, das Siedlungsnetz verandernde Wirkung des Wirtschaftssystems kann in den meisten Fallen als bewiesen betrachtet werden. ., Anmerkungen und Literatur BALINT Cs., Siidungarn im 10. Jahrhundert, Studia Archaeologica XI, Budapest, 1991. BERES M., Regeszeti adatok Csongrad megye Arpadkori telepiilestortenetehez. Archaologische Angaben zur arpadenzeitlichen (I 0-11. Jh.) Siedlungsgeschichte der Umgebung Komitat Csongrad, Kezirat, Szeged, 1986, Manuskript. BLAZOVICH L. (Ed), A Koros- Tisza-Maros-koz telepiilesei a kozepkorban. Mittelalterliche Siedlungen der Umgebung der Koros-Teij3-Maros Fliissen, Szeged, 1996. CSALLANY G., Avarkori es XI-XIII. szazadi leletek a szentesi muzeumban. Awarische und ungarische Funde zu Szentes, Dolgozatok. Arbeiten des Archaologischen Institut der Kon. Ung. Franz Josef Universitiit in Szeged IX-X, Szeged, 19331934, 221-280. FEHER-ERY-KRALOV AN SKY, A Kozep-Duna medence magyar honfoglalas- es kora Arpad-kori sirleletei, 19 M. Beres Regeszeti tanulmanyok II, Budapest, 1962. KovALOVSZKI J., E!Ozetes jelentes a dobozi Arpadkori falmisatasr61 1962-1974. Vorbericht Uber die Ausgrabung des arpadenzeitlichen Dorfes Doboz 1962-197 4, Archaeologiai Ertesito 102, Budapest, 1975, 204-223. KURTI B., Regeszeti adatok a Maros-torok videkenek 10-11. szazadi tOrtenetehez. Archaologischen Angaben zur Geschichte der Umgebung der Marosmi.indung in den 10-11-ten Jahrhunderten, in: A kokort6l a kozepkorig. Van der Steinzeit bis zum Mittelalter, Szeged, 1994, 369-386. SZELL M., Elpusztult falvak, X-XVII. szazadi regeszeti leletek Csongrad varmegye teri.ileten. Vemichtete Dorfer und Funde aus den X.-XVII. Jahrhunderten in Kom. Csongrad, Dolgozatok. Arbeiten des Archaologischen Institut der Kon. Ung. Franz Josef Universitat in Szeged XIX, Szeged, 1943, 176-182. V AL YI K., Szer kozepkori teleptilestOrtenete a regeszeti leletek ti.ikreben. Die mittelalterliche Siedlungsgeschichte von Szer im Spiegel der archaologischen Funden, in: Novak-Selmeczi, Falvak, mezovarosok az Alfoldon, Nagykoros, 1986, 117-129. Ab b. 4. - Funde des Dorfes Tape a us den 10-11. Jh. (nach Maria Beres). Maria Beres Mora Serene Museum Roosevelt per 1-3 H- 6720 Szeged Ungarn Abb.5.- Die Lage des Dorfes Typ 1 (nach J. Kovalovszki). ·-i . 1 o~e"! .,..,.:.;:)..:.•~•ls r.:.n:iud.i ll.li•\1'\ .i:~i:·~:lo 20 i ' Rural Settlement in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 J.L. Bintliff The Archaeological Investigation of Deserted Medieval and Post-Medieval Villages in Greece Abstract Since 1978 the Boeotia Archaeological Survey has discovered some 300 rural settlements and analysed three ancient city sites through surface study. The Medieval Village Project is a subproject which is focused on deserted Medieval and Postmedieval village sites, combining surface mapping and artefact collection with information from Byzantine, Frankish and Ottoman chronicles and tax-cadasters. This paper will present the various approaches that the Boeotia Project is currently taking to study rural communities in Medieval and Post-Medieval Boeotia. The accompanying ceramic analysis will be presented by Miss Joanita Vroom in her separate contribution to the Bruges Conference. The traditional and primary source for studying the archaeology of Medieval and Post-Medieval Boeotia, a large agricultural province (Fig. 1) in Central Greece, has been the surviving monuments across the countryside, such as the Middle Byzantine church at Slaipou (modem Orchomenos). A useful map and gazetteer of such monuments and of excavated medieval sites in the region has been published by Koder and Hild (1976) as Volume 1 Hellas und Thessalia ofthe Austrian Byzantine Atlas Project. In an early phase of our Medieval Boeotia Project our Frankish specialist Peter Lock published a study (1986) of the bestpreserved Frankish towers ofBoeotia; recently he has published the first modern textbook on Frankish Greece ( 1995). Figure 2 shows the location of the major pre-Ottoman medieval monuments in Boeotia together with medieval settlements (open circles) studied by the Boeotia Project or recorded in medieval archives. Monuments that still lack both study and conservation include Post-Medieval water-mills, of which a group of three datable to the 16th century from documentary sources and surface finds has been recorded during our intensive field survey of the Valley of the Muses (Bintliff 1996b). Our project has also made a beginning with localised study of traditional domes- tic housing, such as the ubiquitous early Modem 'longhouse' (makrinari) which our analysis of deserted villages at Rhadon (between Pyrgos and Pavlo, North Boeotia) and site VM4 (near Palaiopanagia, South-West Boeotia), together with oral history reports, suggests formed the typical rural dwelling from the 16th century at least in the region. Some of our earlier work (Fig. 3) on village houses has recently been published by Nancy Stedman in an edited volume on the archaeology of Medieval Greece (Lock & Sandars 1996), but Boeotia desperately needs a locally-organized project, perhaps through schools, to record traditional houses by photography, drawings and interviews before they are all demolished to make way for more spacious modem houses. Secondly, we have a range of graphic sources, beginning with artistic representations such as an icon representing the townscape of Thebes (one of the two major towns of the region) at the turn of the 16th-17th centuries, and including photographs from the last century oflocal villages and their inhabitants - but also other regional personalities such as the notorious Dilessi bandits whose like made much of the rural areas ofBoeotia insecure in the third quarter of the 19th century. Of especial importance for Boeotian village history are the numerous maps, often linked to the descriptions of both Western and Eastem Travellers, and commencing in the 17th century. One of the most helpful of these for tracing deserted villages is the Atlas de la Grece of 1852, drawn up by the French Army, which marks both contemporary villages and the location of many abandoned villages. A third source is that of official archives. Apart from state censuses of the Greek state from the late 19th century onwards, and the limited records of Byzantine and Frankish times for Boeotia, the most important archives are those of the Ottoman Empire. Our project Ottoman specialist Machiel Kiel (Kiel, in press) has provided us with the tax reports for some 200 villages in Boeotia, as well as for its towns of Thebes and Livadhia, from 1466-1687, together with fragments from the 18th century. I have been able to localize some 160 of these villages (Fig. 4 ), shown 21 '--< N ~ N to EUBOEA [ 6i ~ " 0 4 6 12 16Km ATTICA Fig. 1. - Location map of the modern province of Boeotia, Central Greece, with temporaty towns and villages. Modern villages in So ut- West and North- West Boeotia mentioned in the text are underlined. The Archaeological Investigation of Deserted Medieval and Post-Medieval Villages in Greece 0 BYZANTINE-FRANKIS,..H BO_ OTLA ~ -====- / ., I . '§¥Et -==== -=- ........... Castle ~Tower Gjln Vundre: ...., 0 =I• ~~~~!~:'· ---=- \ e Church 0 Sotllomont B Bishop 0 ::::~; ' \" Salgoneua~ EU:OIAN ~"" ~·'s_•~-·--~~~~~-~v RutamUea • Harm en a 0 ~•Dramlel o SJ!tal c) .o o• M.Harmena. .t..oe B . . . • Panayia Neochorl 0• -Tallzeo£. 0 o Kakoal 0• o Andrltoa THESES Muataphe.doa Erlmokaatro o0 Archondlkl oA• ~ • Bra tal Bubokao~- I Stanlatee • Klobucar • Kleldl .87 I lno Plgadl 8 0 I I I I _,_ OoKokla o..___ _~_ ___j10 km GULF OF CORINTH Fig. 2. - Distribution of the major pre-Ottoman medieval monuments in Boeotia, together with medieval settlements (circles) studied by the Boeotia Project or recorded in medieval archives. Fig. 3. - The traditional peasant house in post-medieval Boeotia: the Makrinari or long-house. From Stedman 1996. 23 J.L. Bintliff ~- - 1Bokrakao/ 1 Lakroktu • Buboka _ aleei Buzurg e 0 Kruyekuki _ Skourte. A.Trladh•aKokla S. 1 ..... - - 1 • Kokla K. 8 GULF OF CORINTH Fig. 4. -Hitherto-located villages listed in the Ottoman Imperial archive census lists for Boeotia, 1466-1687. VILLAGES EXISTING • Ill .A GULF OF CORINTH DESERTED 0 0 6 Hoykal/ ---Koata Rabduea - - Fig. 5.- Locatable villages listed in the Ottoman census archives, showing current status (deserted, sit!! occupied), and ethnicity in Early Ottoman records (Albanian, Greek, unknown). Urban Muslim population not shown. 24 The Archaeological Investigation of Deserted Medieval and Post-Medieval Villages in Greece U! E > (j) (\J 0! (\J c > E > Fig. 6.- The Valley of the Muses (west of modern Palaiopanagia village, S. W. Boeotia), archaeological sites discovered through field survey identified by sample grids. The largest in north-centre of the Valley is the medieval village of Panagia/site VM4 (1 1 ha). 25 J.L. Bintliff Valley of the Muses Overall Pottery Density and Site Sampling Grids Fig. 7. - Overall potte1y density by fie!dwalking transect and identified sites (open or black grids) in the Valley of the Muses. N A The shading scheme used to represent the overall pottery density ranges from a minimum white to a maximum grey 0 1500 3000 Meters ~~~~~------~ here; only half of these are still inhabited (Figure 5 shows the ethnicity in the Ottoman records, where known, and the villages still occupied today). Those that are deserted offer excellent opportunities for the study of the development of houses and of pottery assemblages, especially as often their date of foundation and desertion can be estimated approximately. The fourth major approach is through field survey archaeology; the Boeotia Project has been carrying out surface survey since 1978 both in south-west Boeotia, from a base in the village of Mavrommati, and in the far north around the villages ofPavlo and Loutsi (Fig. 1) (Bintliff & Snodgrass 1988a & b; Bintliff 1991, 1995, 1996a, b, c). Such work involves teams walking every field at close intervals, counting and collecting pottery fragments found on the surface. Where concentrations of potsherds or freshlyploughed up scatters are seen, a grid is placed across the area so that the site- whether farm, village, cemetery or sanctuary, can be measured for its size and date. After many square kilometres of countryside have been covered in this fashion as completely as possible, we can produce maps of past settlement and other kinds of site. In figure 6 we see the Valley of the Muses (west ofPalaiopanagia, South-West Boeotia), with the archaeological sites identified through 26 their sample grids; one of the two largest is the medieval village of Panagia/ Site VM4 in the right upper centre of the picture (11 ha). The density of pottery across the entire surveyed landscape is shown by grey-scale shades in Figure 7 and mainly reflects agricultural manuring in Greco-Roman times. Detailed study of particular periods such as the Early to Middle Byzantine era will begin by using maps of sites datable to this phase (cf. Figures 8 & 9, for the South-West survey block which includes the Valley of the Muses), which give an overview of the density and distribution of population. More information comes through intensive study of particular village or farm sites. Some of the key villages with plentiful surface ceramics and archival references are shown on Figure 2: Neochori, Harmena and Archontiki. Some medieval and post-medieval settlements provide little for surface survey; the deserted village of Palaeomazi, for example (in the mountains between Palaiopanagia and Evangelistria in South-West Boeotia), is under pasture and shows almost no surface pottery. Only a road cutting allowed us to observe medieval occupation material. In contrast the dramatic drying of Lake Ilike in the centre of the province ofBoeotia during the late 1980's and the The Archaeological Investigation of Deserted Medieval and Post-Medieval Villages in Greece • 0 (i) Definite Probable Possible occupation occupation occupation •? I MIDDLE BYZANTINE 41. MIDDLE/LATE BYZANTINE-FRANKISH ? "BYZANTINE" 1km I I I .,.,. .... L::. • ,./ \_, ,. .... -,......... ,. ... ..... _,_ ( \ D I \ \ 1111 0 i"' I e EARLY BYZANTINE \ \ \- \ \ .-""""' / \ ........ ..... -"" ..... ..... ..... .... ..... ' \\ ? • ASKRA . 111141. CD • I &. \.---, 1111 VM21 •III 1111 \ I \ ? • , b.• . 0 •? . L::. PALAEO- p NEOCHORI • III -~ \'- .... _.... { .!""' ........ PALAEO-THESPIAE;\ EAIMOKASTRf> .PP16 I 111141. 1111 ~--, -~ I ,' ~ I \...._ CD • _.,. /'/ I -- Fig. 8.- Provisional distribution of Early to Middle Byzantine, and Transitional Middle-Late Byzantine (Frankish) sites in the South-West Boeotia survey region. early 1990's exposed a lost Frankish tower and surrounding settlement at Klimmataria (Fig. 1 & 2), where our project obtained a permit to plan and collect ceramics from a site as clean as an excavation. Usually however, medieval and later settlements show no surface traces beyond the occasional foundations of abandoned longhouses and the generally ubiquitous and abundant surface pottery. The potential and significance ofhighly-intensive surface survey can be illustrated through our discovery some 500 m north of the ancient city ofHyettos, at the locality ofGjin Vendre (Fig. 2) near Pavlo, of five discrete medieval and post-medieval settlement sites, each one showing a slightly different range of pottery; it is possible that the entire sequence here may cover most of the period from Early Byzantine to the late 19th century. It is the pottery of such deserted settlement sites that is the richest and least exploited area ofBoeotian medieval history. Finally I shall mention some of the major conclusions of the medieval to post-medieval project in Boeotia so far: 27 J.L. Bintliff 0~ "' \ \ e L BYZ I FRANKISH A FR./EARLY TURKISH • 0 \ \ \ ,. .,."" (!) Definite Probable Possible occupation occupation occupation . '' A ( A 6... ..... , ...., • ... &.• .. \ .... .... -' ...... \ (!) ~.. .. <D PALAEO-NEOCHORI •., MUSES SANCTUARY • PP16. (1)" ~ r-~~,_ ~ 0 --------- __, .r' JiJ. \ ' • { PALAEO THESPIAE ERIMO~A~T~ \ .....--, \ A • / I , __ _ / / 'J Fig. 9.- Provisional distribution of Late Byzantine (Frankish) and Early Turkish sites in the South- West Boeotia survey region. 1. The map of major Byzantine to Frankish sites for Boeotia as a whole (Fig. 2), and our detailed examination of the sequence in South-West Boeotia (Fig. 8) and at Gjin Vendre in the North, suggest that there was considerable continuity between Greco-Roman settlement patterns (Fig. 10: towns {triangles} and villages {circles} of ancient Boeotia) and those of the early to high middle ages. The colonisation of this landscape by Slavs in the 6th-7th centuries AD probably involved a merging with local populations, often on pre-existing settlement sites. Subsequent population growth through the Middle Byzantine and 28 Frankish eras, to the 13th century, remained largely within the ancient settlement network. 2. The 14th and early 15th centuries AD appear to have been a disastrous time for Boeotian settlements. Incessant warfare between the Franks, the Byzantines and the encroaching Ottoman state, together with the impact of the Black Death and climatic deterioration, led to the large scale abandonment of most of the countryside, and a nucleation of population into the two regional towns of Thebes and Livadhia and a limited number of large villages. This can be shown The Archaeological Investigation of Deserted Medieval and Post-Medieval Villages in Greece , ........... / / / ............. / ...... __ -- HYETTOS.A --..- ._, • ?Oimones / / / / TegyraEII / I CHAERONEA.A ORCHOMENOS.Ii( I / / I I \ ALEVADEIA ''\ Alalkomenai El \ DelionE~ KORONEIAA \ El ?Okalea THEBESA \ eAskra \ '' / El Hippotai \ / / / CHORSIAEA ATHESPIAE ?Donakone ATHISBE I eStephon • ?Leontarne • Eutresis ?EIIopia "' ?SkolosEII ?Eteonos/ Skarphe PLATAIAA et I ATANAGRA ?Eilesione I ?Pharai ?Skhedia Ell I I e Erythrai - e Hysiai ---- I I I I ..---....... I ............... .._..._--_..I Fig. 10. - The distribution of towns (triangles) and villages (circles) in Classical Greek Boeotia. archaeologically through surface survey of several deserted villages but even more clearly from the first Ottoman census map preserved, that of 1466 (Fig. 11): the Greek villages are notable in their size and concentration. During the final half century of Frankish rule the Dukes of Athens attempted to recolonize the landscape through encouraging settlement by Albanian clans (Jochalas 1971 ); this was continued under the first Ottoman authorities; the tiny new foundations of Albanians seem generally to have been settled close to abandoned Byzantine-Frankish villages. 3. Under the Pax Ottomanica Boeotian populations and economy flourished. This can be shown not only in the census statistics (Figure 12 compared to Figure 11 evidences a generalized population boom up to 1570), but also in the surface archaeology of villages studied by the Project. Thus at the village of VM4/ Panayia in the Valley of the Muses, the dramatic expansion of the community between Frankish (Fig. 13) and Early Turkish (Fig. 14) times is clearly recorded from the spread of diagnostic surface ceramics, mirroring the Ottoman census statistics where well over 1000 people are recorded for the village by the later 16th century. 4. The severe decline in Boeotia's fortunes during the troubled 17th century is likewise documented both at the province level through census records (Fig. 15, for 1687 /8), and through the study of deserted village sites using surface ceramic distributions. Figure 16 shows the contraction, followed by abandonment, of the village ofVM4/ Panayia which took place during the late 17th century. 5. The richness of archive materials and the extraordinary abundance of surface ceramics of postRoman date in Boeotia, together with the growing recognition of discrete assemblages of ceramics for each major phase of medieval and post-medieval times (the work of Professor John Hayes and Joanita Vroom), offer an unusual opportunity to bring together historic sources and field archaeology, so that the development ofboth landscapes and townscapes in post-Roman Greece are becoming increasingly better understood. Figure 17 provides a comparison between demographic change for 16 Boeotian vil- 29 J.L. Bintliff BOEOTIA IN 1466 EUBOIAN GULF Spata/Sipahll&rO LalokaD 0 Mavrommatl Yorgl Panaya.KunlaO 0 ~la - 30 "Greek" village FAMILIES Unknown ethnic I l 31 -50 51 - 134 I Slamadl OomenlkoQ? I KelmendiQ 0 luta Kaparell\0 0 a KlobuCIIf 0 Andre lvlu I I Kino• SalaD 0 Parapungl 0 SkymalariD ?Q Golundlye Orallu 'Albanian• village Andrlha Boreure THESES Q?Poravanka Zogra Kobill Atehondilea Bala 0 OTTOMAN ARCHIVE 0 QUia Aabduu r-, 1 0 1eakrakui/Lakroku ~ L--' 0 Kruyekukl Saln~zuroO KeblrO 135- 199 200+ LEVADEIA Muslim 0 Greek THEBES 51 Greek 164 467 • A Fig. 11. - Population and ethnicity in Boeotia in 1466, after Ottoman census records. Fig. 12.- Population and ethnicity in Boeotia in 1570, after Ottoman census records. EUBOIAN GULF L. Copals Platanakl~ 0 MurlkiG, - ~----~-------'0 km OTTOMAN ARCHIVE 'Albanian' village 'Greek' village Unknown ethnic 30 _ ====- Lavdlye/Lavda Klrnlal __ 1 1 Kapandi'HI 0 '. ~~ 0 · ~GHCI/ Ver&dil Oramlal GULF OF CORINTH I - 30 ""'"" !: lA 31 - 00 0 51 - 134 0 135 - 190 200+ LEVADEIA THEBES Muslim 210 Greek 542 Greek 1497 The Archaeological Investigation of Deserted Medieval and Post-Medieval Villages in Greece Site Vm4 (f) Fig. 13.- The occupation swface of the deserted medieval village of VM4 in Frankish times, based on surface sherds. !Grab Sample= 8 Sherds N 0 100 200 300 400 500 A 600 Meters Site Vm4 (If-et) Fig. 14.- The occupation suiface ofthe deserted medieval village of VM4 in Early Turkish times, based on surface sherds. I Grab Sample = 34 Sherds N 0 100 200 300 400 500 600 Meters A 31 J.L. Bintliff BOEOTIA IN 1687/8 Ayo. Nikola/Kamba~ I -- Kapreno. Peraci/Yeradji __ ,Kirnlai 0 Lavda/Lavdlyo 1 A I ~-J Spat a/ SipahilerD Andrltu. • ,• Kukur9isecud-====:... "(jQ Panaya ODara itJSam,.arlq D 0 likores ~~lochori I KlobucarD D 0 Halka Bardti Loa ha O? ___ Buru. GULF OF CORINTH OTTOMAN ARCHIVE 1 - 30 "Albanian' village "Greek" village I I 1Koklnaria Likorea 1---1 0 Luta Baltu 0 Parapungi r 0-; I TalizeO Kakooi/Kaatorya Beroaare D Andonl Konda~ _ ~ +Beluaa ?DPoravank111 Kaznla Archondiha B. Archondltaa ,. 0 Oramiai 31 - 50 51 - 134 0 D 135 - 199 200+ LEVADEIA ? THESES Jews Greek 35 654 FAMILIES Unknown e1hnic Fig.15.- Population and ethnicity in Boeotia in 16878, after Ottoman census records. Site Vm4 (t) Fig. 16. - The occupation surface of the deserted medieval village ofVM4 in Later Turkish times, based on suiface sherds. N t\ 32 I Grab Sample = 4 Sherds The Archaeological Investigation of Deserted Medieval and Post-Medieval Villages in Greece The demographic development of Boeotla as reflected In the number of households of 16 Boeotla vlll111ges 3000 II] Households 2000 VI '1:1 .... 0 .c CIJ VI :::1 0 =1000 1466 1506 1570 1642 1688 1800 40 1896 Early Modern Late T wtish ( 1640-1800) Early Turkish (1460-1640) 1879 Boeotia Survey D E3 rn:l possible probable certain 30 20 VI ......, Ill VI ....0 ....1:: 10 :::1 0 c.! EByz C7th-9th MByz M/LByz-F ClOth-12th LByz/F F-T 1204-1400 AD 1400-1600 AD T !600-1800 AD T-Mod Cl 9th -early 20th Periods Settlement trends of the Early Byzantine (EByz), Middle Byzantine (MByz), La1e Byzantine/Frankish (LByzJF). Late Fmnkish-Early Turkish (F-T). Late Turkish (T), and Late Turkish-Early Modern Period (T-M) Fig. 17. - The comparison between demographic change for 16 Boeotian villages from the 15'h-J9'h centuries (based on research by Dr. M. Kiel) and the statistics of settlement numbers over time based on archaeological field survey in S. W. lages from the 15th-19th centuries, based on Dr. Kiel 's Ottoman archive research, and the statistics of settlement numbers over time based on the Project's surface field survey evidence (collated by Dr. K. Sbonias). Bibliography BINTLIFF J.L. 1991: The Roman countryside in Central Greece: observations and theories from the Boeotia Survey (1978-1987), in: G. BARKER 33 J.L. Bintliff & J. LLOYD (eds.), Roman Landscapes. Archaeo- logical Survey in the Mediterranean Region, London, British School at Rome, 122-132. BINTLIFF J. L. 1995: The two transitions: Current research on the origins of the traditional village in Central Greece, in: J.L. BINTLIFF & H. HAMEROW (eds.), Europe Between Late Antiquity and the Middle Ages, Oxford, Tempus Reparatum, 111130. BINTLIFF J.L. 1996a: The Frankish countryside in central Greece: The evidence from archaeological field survey, in: P. LOCK & G.D.R. SANDERS (eds.), The Archaeology of Medieval Greece, Oxford, 1-18. BINTLIFF J.L. 1996b: The archaeological survey of the Valley of the Muses and its significance for Boeotian history, in: A. HURST & A. SCHACHTER (eds.),LaMontagne des Muses, Geneva, 193-224. BINTLIFF J.L. 1996c: The two transitions: Current research on the origins of the traditional village in Central Greece, in: Epetiris tis Eterias Viotikon Meleton, B 'Diethnes Synedrio Viotikon Meleton, Livadhia, Athens, Society for Boeotian Studies, 605-614. BINTLIFF J.L. & SNODGRASS A.M. 1988a: The end of the Roman countryside: A view from the East, in: R.F.J. JONES et al. (eds.), First Millennium Papers. Western Europe in the First Millennium A.D., British Archaeological Reports S40 1, Oxford, 175-217. BINTLIFF J.L. & SNODGRASS A.M. 1988b: Mediterranean survey and the city, Antiquity 62, 57-71. JOCHALAS T. 1971: Ober die Einwanderung der Albaner in Griechenland, Beitrage zur Kenntnis Sudosteuropas und des nahen Orients 13, 89-106. KODER J. & HILD F. 1976: Hellas und Thessalia, Wien, Osterreichische Akademie der Wissenschaften. KIEL M. In press: The rise and decline of Turkish Boeotia, 15th-19th century, in: J.L. BINTLIFF (ed.), Recent Developments in the History and Archaeology of Central Greece, Oxford, Tempus Reparatum. LOCK P. 1986: The Frankish towers of Central Greece, Annual of the British School at Athens 81, 101123. LOCK P. 1995: The Franks in the Aegean, 1204-1500, London. LOCK P. & SANDERS G.D.R. (Eds.) 1996: The Archaeology of Medieval Greece, Oxford. STEDMAN N. 1996: Land-use and settlement in postmedieval central Greece: An interim discussion, in: P. LOCK & G.D.R. SANDERS (Eds.), The Archaeology of Medieval Greece, Oxford, 179192. Prof. Dr. John Bintliff Archaeology Department Durham University Science Site South Road Durham DHI 3LE UK 34 Rural Settlement in Medieval Europe- Papers of the 'Medieval Europe Brugge !997' Conference- Volume 6 Giovanni Di Stefano Villaggi tardo bizantini degli Iblei: primo medioevo siciliano La cuspide sud-orientale della Sicilia, fra i1 VII e l'VIII sec. d.C., e caratterizzata da un insediamento rurale sparso di tipo fortificato, cioe da veri e propri "Kastellia", e da un incastellamento, a "Kastra", diffuso negli speroni rocciosi, alla confluenza delle grandi vallate fluviali. Equesto il tipico popolamento che nell 'Isola precede la conquista araba e il fenomeno dell'insediamento rupestre: i cronisti che scrivono intomo al 750 d.C. (Ibn al-Atir, An-Nuwayri) descrivano uno scenario di villaggi di fattorie-fortificate diffuse ed insediate nell'altopiano. Nel secolo scorso, soprattutto, i resti di questi villaggi erano ben visibili sull 'altopiano ibleo e Paolo Orsi documento vari ruderi di case. In questa area estato ora possibile non solo effettuare un primo inventario dei dati archeologici schedati ma anche elaborare dei computi statistici sulla distribuzione delle segnalazioni. Tuttavia, occorrera, ancora, acquisire altre conoscenze sul territorio per tentare una sintesi completa ed articolata delle tendenze di distribuzione ed occupazione dell'area in epoca proto-araba. La tecnica edilizia impiegata nella costruzione di queste fattorie-fortificate e molto singolare: grandi e pesanti blocchi di calcare appena sbozzati, di forma parallelepipeda, grossolanamente messi in opera a piu filari sovrapposti. La pezzatura di questi blocchi di calcare dipende spesso dalla possibilita di sfaldare, in maniera regolare, il soprassuolo roccioso circastante. Ma, ovviamente l'impiego massiccio di grandi blocchi e per lo piu costante nelle strutture portanti dell' edificio, in modo da risolvere con soluzioni semplici ma efficaci i problemi statici di questa architettura privata. Cantonali, architravi, piedritti di porte e finestre sono, infatti, messi in opera in unica soluzione con ortostati molto solidi e massicci. Tra 1' altro, 1'assenza costante di legamenti cementizi in questi edifici ha probabilmente accentuato 1'uso e la diffusione di questa tecnica "megalitica". Non si sono ravvisati, tra 1'altro, neanche mattoni. Solamente blocchi di calcare accostati, messi in opera completamente a secco, sovrapposti in maniera quasi incoerente avvolte con un doppio paramento litico, ed emplekton all 'intemo. E lo spessore di questa muratura, certe volte considerevole, su cui si basa la statica dell 'elevato dell' edificio. Sicuramente la copertura di queste fattorie bizantine con tegole cotte, striate sulla superficie, che dovevano essere sostenute da una intelaiatura lignea, a doppio spiovente. Gli edifici avevano quasi sempre un solo corpo edilizio centralizzato, di forma quadrata, trapezoidale o rettangolare, molto allungata. In molti casi e probabile che la fattoria aperta su un cortile era anche recintata con un solido muro e con veri e proprie torri di avvistamento sulla campagna o di difesa. Avvolte sono presenti dei siloi, per la conservazione delle derrate alimentari. L'edificio era diviso intemamente, da vari tramezzi in muratura, in tre o quattro ambienti, spesso giustiapposti e non comunicanti fra di loro. Tra le fattorie sono spesso note grandi cisteme ipogeiche comuni, ma anche vere e proprie opere idrauliche (cisteme multiple, canali, etc.), come nella valle di Buttino, a Centopozzi. Piccoli edifici religiosi sono spesso noti in questi agglomerati: si tratta di architetture certe volte mol to modeste (con navatine precedute da esonartece) o di vere e proprie chiese di grande impegno costruttivo con ambienti a cupola voltati, come aS. Croce Camerma. Probabilmente, una concentrazione "urbanistica" eindividuabile in questi villaggi solamente in prossimita degli edifici religiosi. Ma spesso si tratta di semplici orientamenti delle singole fattorie che nelle adiacenze delle chiese rimangono costanti. Per il resto l'immagine complessiva di questi edifici eproprio quella di un agglomerato di fattorie disposte in maniera disorganica, senza un apparente coordinamento spaziale fra i singoli complessi. Ogni unita edilizia appare piuttosto autosufficiente pur se sfrutta alcune attrezzature comuni. Questi villaggi rurali sono per lo piu diffusi nell 'altopiano calcareo, fra Ragusa e Siracusa, si a a ridosso della costa, sulle propaggini dei pendii rocciosi, che sulla sommita montuosa degli Iblei, in relazione, ancora, con la viabilita tardo-romana che 35 G. Di Stefano ~HtntV•H• ..... · $""'~ ) J~ ~COM!SO CISTERNAZZI e BUTTINO o IIFAL TRABACCIIE POZZIII ~ 1'-J.GUS~ • SCROFAHI Cl CASSERO EIIHIGLIU~~~ I M kela >-IOOICA -~p~~~~!\~ '"'\ t!Uf•tt ' ' .. w.. Fig. 1.- Carta dell'area degli Jblei. tagliava la cuspide sud-orientale della Sicilia, da Agrigento a Siracusa, e in relazione alla viabilita minore, ai tratturi, fra 1'intemo e la costa. Non e improbabile che nella distribuzione nel territorio di questi agglomerati abbiano avuto un ruolo indifferente sia una mutata condizione climatica registrata proprio alla fine del mondo-antico che una diversa economia. Un inaridimento, costante, in tutto il bacino mediterraneo e, infatti, ben attestato e registrato, probabilmente dalle varie strutture idrauliche note. L'intenso popolamento rurale dell' altopiano puo, tra 1'altro, indiziare monoculture specializzate (la viticultura nelle fasce paralitaranee e nei pendii collinari degli Iblei, la cerialicultura nell 'altopiano ). 36 Queste fattorie richiamano, per le tecniche megalitche e per le forme tipologiche, alcune fattorie fortificate della cirenaica. Lungo la costa meridionale della Sicilia gli insediamenti tardo-bizantini si concentrano in prossimita degli approdi, soprattutto vicino Punta Secca, dove e molto attivo l'ancoraggio di Kaukana. Si conoscono circa 25 edifici, distribuiti lungo la fascia costiera di 300 metri, per lo piu attribuiti cronologicamente fra la seconda meta del quarto e il VII sec. d.C. Gli edifici sorsero su un'area mai prima antropizzata, pertanto la conformazione urbanistica dell' abitato non ricalca precedenti impianti ellenistici o di eta romana, dando luogo ad un aggregato dalla forma alquanto inaspettata e inconsueta rispetto alla regolarita, tradizionale, della citta classiche e romane. Villaggi tardo bizantini degli Iblei: prima medioevo siciliano 0 10 30 ~Omt. scala 1: 1000 Fig. 2.- Cont. da Pianicelle (dis. Gianni Giacchi) . . I, _, . , .. ' I N Fig. 3.- Cont. da Costa (dis. Gianni Giacchi). 0'--~---=::Sm t . 37 G. Di Stefano L' appellativo di chorion, con cui I' abitato app are ricordato ne lie fonti, corrisponde filologicamente, ad un insieme di case e terreni agricoh. E stata possibile, una lettura dell 'impianto: sicuramente gli edifici 22 e 18 sono due complessi edilizi religiosi. Rispettivamente, si tratta di un vero e proprio convento, sede di una comunita eremitica, isolata geograficamente in una piccola isoletta costiera, e di una chiesetta cimiteriale a tre navate. Ma nonostante tutto, pen), la visione dell 'insieme urbano conferma ed acuisce l'aspetto d'isolamento delle singole unita, peraltro raramente aggregate. Proprio in prossimita dell a chiesetta gli edifici 17, 7, 8 e 2, tutti molto vicini, appaiono raggruppati e in qualche modo in comunicazione con uno spazio attorno all a chiesa. Pure 1'edificio 19, tra I' altro apparso no una abitazione, ma un edificio pubblico (un bazar o una locanda?) sembra in qualche modo gravitare proprio in questo baricentro comune. Forse la dimensione degli edifici e in relazione all 'uso degli spazi circostanti: un muro a doppio semicerchio, il n° 2, con una cisterna in prossimita dell'edificio 12, e forse una recinzione, 0 una sistemazione a terrazze del terreno o il limite di una proprieta agricola, di un 'orto o, piuttosto, un recinto per gli animali. Negli edifici piu complessi, gli ambienti si articolano su ali laterali attorno ad un cortile, chiuso verso I' esterno, di forma semicircolare, da quale si accede, per mezzo di larghe scale in muratura, al piano superiore Proprio in queste unita edilizie piu complesse sono riscontrabili lunghe file di grandi ambienti, · forse dei veri e propri magazzini aperti a sud, su ampi cortili. Avvolte sono giustapposte nella stessa unita parti funzionali distinte. Appare logico ipotizzare, anche in ragione di queste differenziazioni formali e funzionali, I' esistenza di vari gruppi sociah: contadini e forse commercianti. Peril mediterraneo centrale I 'importanza di Kaukana per lo studio dell 'habitat urbano bizantino, non appare affatto inferiore ai coevi abitati dell' Africa e del medio-oriente. Anzi, e sorprendente sia l'affinita "urbanistica", che delle singole tecniche edilizie, proprio con i coevi villaggi dell a Siria del nord, dell' area palestinese della Transgiordania, o del limes libico e tunisino. In modo particolare si puo richiamare l'abitato di Ghirza o i villaggi di Behyo, Quatua e Refeda. Bibliografia BEJIOR G. 1986: Gli insediamenti della Sicilia romana: distribuzione, tipologia e sviluppo da un primo inventario dei dati archeologici, in: GIARDINA A. (a cura di), Societa romana e impero tardoantico. Ill (Le merci e gli insediamenti), Bari, 463-519. Dr STEFANO G. 1985: La Regione camarinese in eta romana, Modica. Dr STEFANO G. 1994: Distribuzione e tipologia degli insediamenti di eta repubblicana ed imperiale sull 'altopiano Ibleo, in: Le ravitaillement en ble de Rome et des centres urbains des debuts de la Republique jusqu 'au Haunt Empire. Actes du Colloque International de Naples (1 991), NapoliRoma, 237-242. FALL! CO A. M. 1971, Villaggi tardoantichi e bizantini della Sicilia orientale noti all'Orsi e loro attuale consistenza, in: Atti If Congr. Naz. Archeologia Cristiana (1969), Roma, 177-183. MESSINA A. 1993: Tre edifici del medioevo siciliano, Sicilia Archeologica XXVI (82), 61. MOLINARI A. 1994: Il popolamento rurale in Sicilia tra V e XIII secolo: alcuni spunti di riflessione, in: FRANCOVICH R. & NOYE G. (a cura di), La storia dell'alto Medioevo italiano (VI-X secolo) alia luce dell'archeologia. Convegno Intern. Siena 1992, Firenze, 361-377. ORSI P. 1931: Epigrafe cristiana di Palazzolo Acreide (Acrae): Contributi alla storia dell'altipiano acrennse nell'antichita, Rivista di Archeologia Cristiana VIII, 287-299. REBUFFAT R. 1988: Les fermiers du desert, in: L 'Africa romana. Atti V Conv. di studio Sassari 1987, Sassari, 33-68. SODINI J.P. 1993: La contribution de l'archeologie a la connaisance dum onde byzantin (IV-VII siecle ), Dumbarton Oaks Papers 47, 139-184. TATE G. 1992: Les campagnes de la Syrie du Nord du !le au VIle siecle, Inst. franc;:ais d' Arch. du ProcheOrient 133, Paris. WILSON R.J .A. 1990: Sicily under the Roman Empire. The archeology of a Roman Province, Warminster. Dott. Giovanni Di Stefano via B. Croce 6 97100 Ragusa Italia 38 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Andre Bazzana, Jolwny De Jl;feu/emeester & Andre Jl;fatthys Quelques aspects du peuplement medh~val du Vane de Ricote (Murcie, Espagne) Depuis quelques annees, les recherches d'archeologie medievale realisees dans la Peninsule iberique s 'orientent pour la plupart d'entre elles vers une approche globale des territoires amenages et exploites par l'homme: aux problemes d'une nature difficile, ou la montagne est partout presente et ou I' eau vitale est mal distribuee, s'ajoutent le vaste probleme historique de la confrontation de deux societes que la religion differencie mais qui s'opposent surtout quand on examine les bases sociales et socio-economiques qui les caracterisent. Deux evenements majeurs marquent les regions d'al-Andalus au cours du Moyen Age: c'est, d'une part, l'essor rapide et decisif d'une societe fortement orientalisee, dans sa culture comme dans ses pratiques quotidiennes -et le developpement de l'hydraulique agraire en fait, a !'evidence, partie-, d'autre part, les mutations lentes ou brutales selon les cas qui suivent le choc d'une conquete feodale deja largement entamee au Xle siecle (1085, prise de Tolede) mais qui s'impose en deux temps, dans la premiere moitie du XIIIe siecle (1238, prise de Valence) puis a !'extreme fin du XVe siecle (1492, prise de Grenade). La recherche, dont on presente ici quelques resultats recents, s'inscrit dans le contexte de cette approche des societes andalouses et se place a lajonction de deux axes thematiques refletant chacun une problematique historique specifique: !'etude des territoires, cornme mode d'organisation et de mise en valeur des paysages mediterraneens, et !'analyse des structures irriguees qui contribuent a fayonner ces paysages et a leur donner leur originalite technologique et ecologique. C' est la collaboration -scientifique, economique et technique- engagee depuis plusieurs annees entre la Region wallonne, la Region autonome de Murcie, la Casa de Vehizquez (Madrid) et !'Unite Mixte de Recherche 5648 du C.N.R.S. (Lyon) qui a permis de developper 1' etude du peuplement medieval du Valle de Ricote, noyau geographique et administratif imp !ante sur le moyen Segura; 1'expression recouvre, apres la conquete chretienne du XIIIe siecle, les territoires d' Abanin, Blanca, Ojos, Ricote et Villanueva de Rio Segura 1• Comme no us le verrons plus loin, le territoire de Ricote couvrait egalement, a 1'origine, celui de Cieza; et c 'est en epoque islamique deja, que 1'ensemble primitif de Ricote fut partage en deux territoires castraux: Riqut et Siyasa. Sur un territoire relativement vaste, qui mesure un peu plus de 800km 2, le hisn 2 primitif de Ricote associe des sites ruraux, composes de hameaux et de villages, et des sites defensifs parmi lesquels figure peut-etre le premier Siyasa; les terroirs montrent la presence, des la haute epoque, d'une complexe organisation d'hydraulique agraire. Dans cette micro-region, s'est done developpee une etude des structures conservees a travers deux epoques du peuplement: celle correspondant a la presence musulmane puis mudejare et celle correspondant aux mutations de 1' epoque chretienne. S 'agissant du territoire primitif de Ricote- incluant I' actuelle zone de Siydsa- on tentera done, dans les pages suivantes, de donner une description des donnees archeologiques qui ont pu, a ce jour, etre recueillies, puis de presenter une premiere et encore tres incomplete etude du territoire de Cieza, en insistant -car c' est la I' objet de recherches recentes- sur la vallee du Rio Segura et les terres irriguees qui la jalonnent. Rodriguez Llopis 1988,26-27. Sur la signification de valle ou de va/1, voir, s'agissant de la region voisine de Valence, Bazzana 1992, 159-162. Le valle de Ricote forme, lui aussi, un ensemble geographiquement coherent, un espace naturellement delimite par un encadrement montagneux qui suit le f1euve Segura et qui se referme, au nord et au sud, par deux goulots rocheux ou des fortifications medievales controlent la voie de passage; clans cet espace, les principaux noyaux du peuplement (entre autres, les villes et les villages actuels) se rassemblent sous un nom unique qui correspond sans doute a celui de la structure castrale primitive. 2 hisn (pi. husun ): une fortification et son territoire castral. 1 Le fleuve, la montagne et l'homme Dans la zone de Cieza I Ricote, le Rio Segura passe a travers un paysage de montagne mediterra- 39 A. Bazzana, J. De Meulemeester & A. Matthys neenne, paysage fort accidente, parseme, dans la vallee, de buttes temoins. Le fleuve est encastre entre les sierras. Dans le paysage, apparaissent divers elements physiques visibles, comme les chateaux, les hameaux, etc ... , d'autres elements venant de l'archeologie agraire: parcellaires anciens, sites particuliers, comme le grenier d' Abanin. Les structures hydrauliques font naturellement partie des elements archeologiques que 1'on peut reconnaitre dans le paysage rural: sources amenagees, barrages, canaux d 'irrigation, machines hydrauliques, etc., marquent non seulement le developpement historique des systemes d'irrigation agraire mais aussi I' evolution de !'usage des terres dans 1'agriculture regionale; 1' examen archeographique de ces elements permet en meme temps certaines interpretations chronologiques. D'autres donnees sont plus discretes, ainsi les informations exploitables sur les limites des territoires anciens qu'il faut chercher dans une assez longue histoire. A. Peuplement et territoire Dans le monde rural medieval, le village peut etre defini comme le regroupement d 'hommes et de femmes au sein d'un habitat qu'ils implantent, autour de certaines polarites qui sont principalement le chateau et 1' edifice religieux3 . Si ce demier est, le plus souvent, le plus ancien, on ne dispose malheureusement que tres rarement de renseignements sur la date de son apparition; en general, il faut attendre la creation de 1' autre point d' ancrage du village qu' est le chateau pour qu'on soit assure de !'existence du village qui, normalement, l'accompagne: c'est, en effet, la mise en place des cadres de pouvoir -done de la structure batie qui en abrite les detenteurs ou les representants- qui marque habituellement la demiere etape du processus de fixation du peuplement. C'est du moins la le schema le plus commun, celui que des fouilles archeologiques recentes viennent en partie remettre en cause; il arrive, en effet, que des groupements d'hommes soient anterieurs a la creation du premier edifice religieux et du chateau. Sans doute, les realites sontelles differentes en milieu chretien -ot1l'eglise paroissiale est edifiee sous le contr6le des autorites religieuses ou la'iques- et en milieu islamique: la, !'emergence de la mosquee se revele plus spontanee. Pour 1' archeologue qui analyse prioritairement les vestiges materiels, il est tres difficile d'identifier le processus "d'encellulement" -pour employer laterminologie de Robert FossierC qui conduit au village: pour que le village existe veritablement, il faut, audela des aspects materiels analysables, que s 'y etablisse un "etat d'esprit villageois"; or, on sait bien, avec Adriaan Verhulst, que la "definition large du 40 village depasse 1'approche archeologique et rend difficile !'application de celle-ci en dehors des elements purement materiels et geographiques" 5 . Des lors, que peut done faire 1'archeologue, sinon se limiter a demontrer !'existence d'un habitat -plus que d'un "village"- et des poles d'attraction que sont, dans le monde chretien, l'eglise et le chateau, et a dater les evolutions perceptibles? En terre d 'Islam, la problematique reste globalement la meme, accordee cependant, dans le cadre d 'al-Andalus, aux caracteres d'une societe non feodale: ainsi, "chateau" sera pris non dans son sens de residence fortifiee privee, mais dans celui de hisn, c 'est-a-dire de fortification (parfois simple refuge) contr6lant un territoire castral organise et exploite par la communaute rurale. B. Les grandes unites du pays age S 'agissant des paysages, on distingue nettement les terres de regadio du fond de vallee formant la huerta -un veritable jardin dans lequel les arbres fruitiers dominent- tandis que, plus haut, sur les terres de secano, ou terres d'agriculture seche, se cultivaient traditionnellement les cereales et la vigne. Prise dans le sens transversal, au Moyen Age la vallee se structure du fleuve a la ligne des cretes par la succession des elements suivants: zones humides de fond de vallee, terres de regadio, les terres irriguees, ligne des acequias islamiques, habitats, terres de secano (d'extension relativement faible), paturages de versants et de montagne, monte et garrigues. Ce sont la, presente d 'une autre maniere les trois elements habituels du systeme agraire andalous 6 , et en general mediterraneen: les terres de parcours -vastes terrains de paturage recouvrant les zones montagneuses, vers lesquelles conduisaient les drailles et qui produisent le bois pour la construction et les chantiers navals, le sparte et divers produits d'alimentation; les terroirs de secano, domaine de la "culture seche"' de 1' arboriculture et des cereales; enfin les terroirs de regadio, ou les techniques de la petite hydraulique agraire permirent une exploitation intensive des huertas et I 'introduction de cultures d'origine orientale 7 . Pour sa part, 1'implantation du peuplement se dessine schematiquement comme la juxtaposition Pour le monde rural chretien, voir les developpements que consacre acette notion J.-M. Pesez 1992. Fossier 1992, 207-208. Verhulst 1992, 10. Perez Picazo & Lemeunier s.d., 67-68. Bolens 1990, 9-29. Quelques aspects du peuplement medieval du Valle de Ricote (Murcie, Espagne) Fig. 1.- Localisation du Va/le de Ricote. 0 - - 2lan I ,. ,, / \ I_ r ... ) I I I ' I 1 I/ \I r ,/" ..... ---" / I I / I Rlcote• Ojos • _ /'_1' ___ ,.., ..... •Ulea ,--'~:~nueva • de Rio Segura Archena ... I I I ( I ( ) I r' d'une serie de grandes cellules, composees chacune de trois unites 8 : le centre comporte le chateau et un noyau de peuplement generalement fortifie, et appele en epoque chretienne villa: ce terme renvoie non seulement a la realite materielle du chateau, mais surtout a la capacite defensive de la population, protegee par son enceinte; celle-ci enveloppe un habitat fortifie mais, surtout, une societe consciente de sa cohesion et de plus en plus independante du noyau castral; une zone etroite de culture est pratiquement limitee aux secteurs de regadio; clans la huerta de Cieza, abandonnee au XIIIe siecle mais tres vite repeuplee, les chretiens ne furent cependant pas en mesure de cultiver de nouveau toutes les terres abandonnees; un enorme territoire est, enfin, le domaine d'une agriculture pratiquee temporairement et, surtout, des recoltes naturelles; cette zone est parsemee de tours de guet et de refuge, d'ou doit etre donnee l'alertes aux pasteurs, aux agriculteurs et aux defenseur du chateau. Le Valle de Ricote comprend un site archeologique tres particulier et unique, pour 1'instant, clans le monde hispano-musulman: au contact de la zone de regadio et au sommet du Cabezo de la Cobertera (butte temoin en bordure du fleuve), la vallee conserve en effet, les vestiges d 'un grenier forti fie du type des agadirs berberes 9 . Dans son ensemble, ce site du Cabezo ne presente guere de ressemblances avec les autres sites islamiques etudies jusqu, a ce jour; les plus proches seraient cependant ces corn- Bazzana, Cressier & Guichard 1988. Pour les fouilles et leur interpretation ethno-archeologique voir De Meulemeester & Matthys 1995. 41 A. Bazzana, J. De Meulemeester & A. Matthys partiments de plan cane adosses a 1'enceinte du Monte Marient ou visibles sur le site castral de Ux6 10 • C'est moins par recours a la documentation medievale que par reference aux comparaisons ethnographiques avec les greniers maghrebins que l'on peut expliquer ces vestiges. Au Maghreb, le grenier collectif fortifie est ne d'une climatologie irreguliere pouvant entra1ner mauvaises recoltes et disette, auxquels s'ajoutait la continuelle menace du pillage: eloigne de I 'habitat quotidien, 1'agadir ou grenier collectif, est ainsi un edifice ou les Berberes emmagasinent recoltes et autres biens; c 'est souvent, aussi, une forteresse situee en un lieu escarpe. Il semble que les conditions naturelles et l'economie du pays aient impose la necessite de stocker et que s'y soit ajoutee celle de parer aux ravages de la guerre; de la seraient nes les greniers collectifs dont la presence suppose des communautes suffisamment organisees et egalitaires, ou le pouvoir appartient aux chefs de families dans le cadre de la tribu, de la fraction ou du village. Faut-il supposer la presence dans la vallee du Rio Segura, d'une telle societe, tirant ses traditions de celles des "republiques berberes"? Meme si la richesse de la huerta devait reduire les risques de mauvaise recolte, l'insecurite de la premiere moitie du XIIIe siecle etait suffisante pour conduire a proteger dans un agadir les biens des familles. L 'habitat -groupe ou disperse, mais non forti fiedes utilisateurs de 1, agadir eta it probablement situe entre regadio et secano, mais necessairement dans la vallee, entre les noyaux actuels d' Abaran et Blanca, distants seulement de 3km. Le probleme fondamental du site d'Abanin reste celui de son origine, vraisemblablement almohade, de cet etablissement original: "Le grenier fortifie nalt de la coi'ncidence d'une certaine economie rurale et de la guene, que celle-ci so it effective ou seulement a 1'etat de menace permanente" 11 ; le grenier du Cabezo de la Cobertera conceme done bien une population agricole locale et voisine- confrontee au probleme de sa defense et de sa survie, face a une menace reelle ou imaginaire. La ceramique paralt dater ce grenier de la premiere moitie du XIIIe siecle; il fut abandonne au moment de la conquete chretienne ou, au plus tard, de la rebellion mudejare de 1264. Le site, en ruine, fut partiellement reoccupe vers le milieu du XVe siecle: archeologiquement parlant, cette nouvelle phase marque le repeuplement du Valle de Ricote par des mudejars venant du nord, de Hellin. 10 Monte Marinet: Bazzana 1990, 248, 370; Bazzana & Guichard 1988, 18; Bazzana 1992, 261; Ux6: Bazzana, CRESSIER & Guichard 1988, 212-216, 256; Bazzana 1990, 248; Bazzana 1992, 260. 42 2 Riqiit, le hisn et le territoire La huerta du Rio Segura, entre Cieza, au nord, et Ulea, au sud, est protegee et surtout controlee par quatre chateaux qui dominent la vallee depuis leurs cretes rocheuses (fig.l). Au sud de la ville actuelle de Cieza, le fleuve passe a travers un goulot qui est surveille par le chateau de Cieza. A environ 2500m en aval du meandre, le Segura passe a nouveau a travers un goulot, plus etroit encore que celui de Cieza. Sur la crete qui surplombe le fleuve, rive gauche, fut implante le chateau de Blanca. Le village musulman se developpa au pied du chateau et sur sa pente septentrionale. A hauteur du village d'Ulea, a environ 6km en aval de Blanca, la vallee du Segura s 'ouvre assez largement et la fin du trace encaisse du fleuve est controlee par les chateaux de Ricote, rive droite, et celui d'Ulea I Oj6s, rive gauche un peu en avaJI 2. Plus loin, le controle est assure par le chateau d'Archena. A. Le chateau S 'agissant du site de Ricote, plusieurs elements peuvent etre reunis. D' abord, la localite fortifiee arabe est deja mentionnee a la fin du IXe siecle, a 1' occasion d 'une campagne des troupes omeyyades dans la kura de Tudmir, alors en rebellion contre Cordoue 13 ; ce tres ancien texte, sur lequel nous reviendrons, montre en effet !'importance du site et pose le probleme des limites de son territoire et de son organisation. Veritable nid d'aigle de la Sierra de Ricote, le chateau est construit sur une crete rocheuse de forme allongee, presque perpendiculaire a la vallee du Rio Segura, que le site domine d'une hauteur d'environ 350m. Il controle a la fois le village de Ricote, situe dans une petite vallee laterale et la vallee du Segura ou sont etablies les prises d'eau destinees aux irrigations locales. Le texte d'Ibn Hayyan, que nous mentionnions plus haut, eclaire !'organisation spatiale du site. Ce hisn de Ricote est, en effet, un de ceux pour lesquels on possede une mention tres ancienne; comme Tijola, a Almeria, Ricote (ou Riqllt) est signale des 896, date a laquelle le site est attaque par 1, armee emirale. On decouvre acette occasion qu'il est constitue d'une double structure emboltee: une qasaba ou fortification installee au sommet du relief, et une zoned 'habitat et de refuge dont on sa it peu de choses 11 Jacques-Meunie 1951, 185. Pour !'instant, seuls les chateaux de Cieza, Blanca et Ricote ont fait l'objet d 'une prospection et de !eves topographiques. 13 Carrnona Gonzalez 1990, 27. 12 Quelques aspects du peuplement medieval du Valle de Ricote (Murcie, Espagne) I I / \ \ \ ' ( J ) ( / J .) ./ 0 - - 2km Fig. 2. - Les territoires primitifs de Cieza et de Ricote. sinon qu'elle dispose d'une enceinte defensive, ou hizdm 14; ce double niveau de fortifications atteste !'existence d'un noyau important de peuplement, complete peut-etre par celui qui s'installe a l'exterieur de 1'enceinte. Par la suite, les textes arabes ne sont pas tres prolixes et le meme site apparait sous des noms divers -al-Sukur = Ricote 15 , Wadi Riqut = Rio Ricote 16- sans qu'on apprenne rien de plus a son propos. Le plan simplifie qui en a ete etabli montre comment les constructeurs ont adapte le chateau au terrain nature!, dont il epouse etroitement la forme; on distingue bien aussi la division spatiale du site en deux parties, qui constituent d'une part un secteur defensif, de superficie restreinte mais a forte valeur militaire et strategique, d 'autre part une assez vaste basse-cour a double fonction d'habitat et de refuge; des trouvailles ceramiques de surface attestent une occupation au Xle siecle. Ces caracteres font du chateau de Ricote un exemple classique parmi les fortifications musulmanes d'al-Andalus 17 . Ricote est de nouveau mentionne clans les textes a partir du XIIe siecle, puis, vers la fin de 1'epoque islamique, le chateau joue encore un role important clans les rebellions anti-almohades ( 1228-1241) -entre autres celle que dirige Ibn Hud- puis clans la revolte mudejare anti-castillane de 1264-1266: ce fut, ace moment, le demier bastion musulman de la region; il fut ensuite abandonne. B. Les limites territoriales Il est difficile de delimiter avec certitude les territoires anciens puisqu'on ne peut pas toujours tenir compte des limites communales actuelles. Les limites territoriales d' epoque islamique n 'ont pas 14 Ibid.; voir Bazzana, Cressier & Guichard 1988, 55. IS Bazzana, Cressier & Guichard 1988, 67. Guichard, 1990/91. Voir, Bazzana, Cressier & Guichard 1988, passim. 16 17 43 A. Bazzana, J. De Meulemeester & A. Matthys toujours ete conservees et les centres du peuplement chretien (medieval ou modeme) correspondent souvent a des rearrangements du territoire castral arabe dont les terroirs ont ete repartis de fas:on plus egalitaire entre les differentes aldeas dependantes de la villa qui fonctionne comme chef-lieu 18 ; pour le Valle de Ricote, il s'agit de la villa du meme nom. On sait cependant que, dans al-Andalus, les communautes islamiques dum onde rural s 'organisaient dans le cadre d'un territoire qui regroupait des terres de categories et de qualites complementaires, quelques noyaux de peuplement et un chateau -le hisn-, centre et symbole de la communaute, et qui pouvait etre utilise comme refuge en cas de guerres ou de difficultes. Comment retrouver les limites des anciens territoires? Les limites actuelles ne sont utilisables que si des documents contemporains de la conquete chretienne peuvent demontrer qu' elles correspondent aux anciennes: c'est tres souvent le cas a Valence mais on ne peut generaliser les resultats qui y ont ete obtenus19; meme la localisation actuelle d'un centre de peuplement de 1'importance de Cieza est contestable, dans la mesure ou elle repond a des mutations historiques recentes. Quant aux unites territoriales, tres differentes les unes des autres par leurs dimensions spatiales 20 , il existe des exemples aussi bien de fortes modifications comme de conservation de ce qui existait avant la conquete. Dans quelques zones d'alAndalus, on a pu demontrer le mode d 'organisation et de distribution spatiale des territoires castraux: !'observation concrete du terrain comme une lecture attentive de la documentation ecrite le permettent et, si les textes arabes ne mentionnent que le nom du chateau, les documents chretiens sont, sur ce point, beaucoup plus precis. Dans la vallee moyenne du Segura, le val de Ricote montre une organisation complexe qui semble, a partir d'un territoire unique, eclater en plusieurs cellules juxtaposees: Abanl.n I Blanca, Oj6s I Ulea et Ricote proprement dit; au contraire, un peu plus au nord, le territoire de Cieza para1t avoir conserve ses limites anciennes et passe, sans modification profonde de structure, a 1'ordre militaire de Santiago. Entre les seigneuries de Cieza et de Ricote, cette limite eta it la me me a 1, epoque mudejare et la limite actuelle reflete apparemment une situation ancienne. Ainsi, la limite communale entre Cieza et Abad.n suit un trace plus ou mains rectiligne, determine sur la rive droite du Segura par une ligne de crete et sur la rive gauche par la Rambla del Moro 21 . En revanche, ce n'est que dans sa partie septentrionale que la limite entre Blanca et Abaran suit une ligne naturelle de crete; ensuite, depuis les hauteurs de la Sierra de la Pilajusqu'au fleuve, elle para1t mains rationnelle: sur la rive droite, elle s'approche d'abord du pied du Cabezo de la Cobertera, avant de partager ce site de grenier fortifie en deux, puis elle presente un trace tres artificiel qui, sans doute, reflete le partage, en epoque chretienne d'un territoire villageois plus ancien. Les autres limites communales entre Abaran I Blanca d'une part, Ricote d'autre part, suivent a nouveau une ligne de crete, ce qui correspond a un schema tout a fait habitue! qui pourrait remonter au IXe siecle. Dans le reste du Valle de Ricote, le village de Oj6s semble aussi etre ne d'une amputation du territoire d'Ulea: les deux centres, eloignes seulement de quelque 1500m, sont separes, sur la rive gauche, par une ligne de crete qui passe -elle aussi et de maniere tout a fait anormale- a travers le site du chateau d'Ulea! Ajoutons que !'on peut remarquer, dans le Valle de Ricote, que les habitats situes pres du fleuve s 'etendaient, a I' origine, sur les deux rives: Abaran I Blanca et Oj6s I Ulea. L'actuel Ricote ne touche le fleuve que sur 200m a peine, au pied du relief sur lequel fut implante le chateau. Pour 1'instant, 1'analyse des limites communales actuelles permet de penser que, en epoque islamique, mais en un moment non determine, la vallee du Rio Segura faisait partie d'un seul hisn 22 , deja mentionne ala fin du IXe siecle 23 ; au nord se trouvait 1'actuel territoire de Cieza, au sud celui du Valle de Ricote 24 , dont faisait partie les territoires actuels de Ricote, Aban'm, Blanca, Oj6s et Ulea, separes seulement a 1'epoque chretienne 18 toires castraux valenciens". Rambla: lit torrentiel aecoulement sporadique; voir Herin, s.d., 13. 22 Bazzana 1992, 223; plusieurs exemples, que fournit la zone situee au Nord de Valence, conquise par le roi Jaimel entre 1234 et 123 8, montrent comment les nouveaux conquerants ont conservees les limites territoriales anciennes au moment ou commence le processus de feodalisation de la region. 23 Carrnona Gonzalez 1990,27. Guichard 1983, 87-93. Bazzana 1992, 221 donne des exemples ou les limites territoriales restent inchangees; lorsqu'il y a des modifications, elles sont de trois sortes: I. amputation du territoire musulman pour la creation de nouvelles entites, 2. distribution entre plusieurs seigneuries des terres d'un hisn, 3. parfois, le regroupement de plusieurs husun. 20 Voir l'exemple du Shark ai-Andalus, ou apparaissent de nombreux husun: BAZZANA 1992, 287-307; Guichard 1990/91, document 43, "Donnees comparees sur quelques castra et terri19 44 C. L 'etude du regadio Dans tout le Sud-Est espagnol, 1'irrigation est elle une necessite geographique? I! semble que non. En effet, si nous laissons de cote le secteur semi-deser- 21 24 Ibid. Quelques aspects du peuplement medieval du Yalle de Ricote (Murcie, Espagne) tique du Sud de la region valencienne (avec Orihuela et Elche), le milieu nature! permettait pendant le Moyen Age la culture de quelques arbres xerophiles -figuiers, oliviers, caroubiers- ou de plantes qui achevent de murir au debut de l'ete (orge et ble par exemple). Ceci signifie clairement que le regadio est un fait culture!. Le resultat de differentes recherches menees aussi bien en Espagne qu'au Portugal ou au Maroc, montre que, au Moyen Age, les agriculteurs islamiques utilisaient simultanement quatre modeles technologiques pour recuperer de 1'eau; il s'agit d'irrigations qui utilisent: - soit les sources alimentant des bassins, comme les marges des Baleares qui d'ailleurs, rappellent des systemes semblables trouves au Yemen 25 ; - so it des puits dont 1'eau est tiree a1' aide d'un recipient attache a une simple corde ou a un balancier, systeme mediterraneen par excellence, repandu aussi dans toute !'Europe du Sud-Est26 ; - soit les roues elevatoires, comme la noria (de I' arabe nd 'ura) avec sa roue verticale installee sur un canal ou sur la riviere elle-meme, ou la sdniya (arabe, en catalan cenia ou sinia), roue a double tambour supportant des cordes auxquelles sont attaches les godets de terre cuite 27 ; soit, en fin, les divers systemes d' acequias ou canaux qui transportent 1' eau deviee d'une riviere par un barrage, ou de qanats' sortes de galeries a pente legere, ventilees par des puits verticaux alignes 28 • Selon les travaux les plus recents, le grand developpement du regadio se situe en pleine epoque islamique, vers les IXe et Xe siecles; puis, apres le XIIIe siecle et en epoque chretienne, Valence et Murcie connaissent une nouvelle phase d'expansion; une autre se place a la fin du Moyen Age, si importante que certains auteurs ont parle d'un veritable 25 Pirenne 1977. Voir un exemple medieval dans Butzer et alii 1985, 492493 et fig.?; id., 495-496 et fig. I!. 27 Voir Colin 1932, 22-60; Colin 1933, 156-157; Torres Balbasl940, 192-208; Torres Balbas 1942, 461-469; Schioler 1962, 480-486; Schioler 1973; Bazzana & Montmessin, sous presse. 28 Bertrand & Cressier1985, 115-135; Bertrand & Cressier 1986, 569-5 80; Barcel6 et alii, 1986. 29 Voir Rodriguez Llopis 1988; Lemeunier, sous presse. 30 Bolens 1974; Bolens 1981; Bolens 1989, 71-87; Bolens 1990. 31 M. de Epalsa 1988; V. Lagardere 1991. 32 Vol americain de l'annee 1956, vols plus recents et de meilleure qualite de "Hacienda" vers 1974, puis de IRYDA. 33 Barcel6 1989. 34 Communication de M. Barcel6 au Castrum5 a Murcie (Castrum5. Archeologie des espaces agraires, Madrid-Rome, sous presse). 26 "boom" hydraulique, qui commence vers 1480. Les grands reseaux de regadio qui s 'organisent alors s'etendent al'echelle de toute une region et sont assez differents des microsystemes qu 'avaient mis en place les Musulmans. a) Problemes de methodologie Une autre difficulte, latente mais bien reelle, des recherches sur les irrigations medievales est la rarete des documents les concernant; 1' epoque chretienne fournit cependant -principalement a partir du XVe siecle- une documentation exploitable, qui montre un developpement rapide apartir de I' etablissement de la paix en 149229 , et qui permet de mieux conna1:tre I 'histoire sociale des epoques anterieures. La rarete est quasi to tale pour 1' epoque islamique ou les geographes pas plus que les chroniqueurs ne se livrent jamais a une description precise des espaces agraires et de leurs modes de mise en valeur. On cherchera vainement des textes arabes relatifs aux regadios de Cieza ou de Ricote, mais on pourra mobiliser quelques informations, d' ordre general, sur I' existence meme des reseaux et sur leur fonctionnement, dans la litterature agronomique etudiee par Lucie Bolens30 , ou dans les recueils (seule trace ecrite d'un droit coutumier, essentiellement oral) de consultations juridiques31 • Comment, dans ces conditions, men er 1'enquete et tenter de restituer une cartographie des structures d'irrigation d'epoque islamique? La methodologie employee s'appuie sur les methodes de travail de "1 'archeologie extensive" et privilegie done les prospections de terrain, 1' analyse de photographies aeriennes 32 et la modelisation des resultats obtenus lors des etudes de cas les plus productives (celles qui concernent le debut de l'epoque chretienne). Deux observations restent afaire. La premiere concerne les travaux hydrauliques -captage de 1' eau par barrages, deviations, canaux et micro-systemes de conduitsqui, une fois construits, marquent definitivement le paysage, introduisant ce que Miquel Barcel6 33 appelle le "principe de rigidite", veritable "loi" qui interdit les modifications de detail ou les transformations partielles de la structuration de l'espace agraire: a Cieza, Abaran I Blanca ou Ricote, certains elements du systeme hydraulique islamique (Xe-XIIIe siecles) res tent reconnaissables a travers celui de I' epoque chretienne (XVe et XVIe siecle, principalement); ni la conquete feodale du XIIIe siecle ni les phases successives de depeuplement et repeuplement des XIVeXVe siecles n' ont pu effacer les traces des epoques anterieures. Enfin, il convient de bien connaitre les principes elementaires qui regissent le fonctionne45 A. Bazzana, J. De Meulemeester & A. Matthys ment des espaces irrigues clans 1'Islam medievaP 4 . Toute possibilite d' acces a 1'eau est mise a profit et les conditions technologiques sont des plus simples: elles ne requierent aucun investissement majeur, ni n' imposent le recours a un specialiste de 1'hydraulique; elles sont le fruit d'un savoir paysan, d'une tradition transmise a travers les siecles, a laquelle se mele cependant la connaissance theorique des principes de base du nivellement et des ecoulements par gravite; I 'un des moyens les plus simples -en meme temps que le plus commun et le plus repanduconsiste, en amont des terrains a irriguer, a deriver en partie, grace a un barrage, les eaux du fleuve vers la bouche d'entree d'un canal d'irrigation: la pente naturelle du fleuve etant plus forte que celle du canal, le gain en altitude est suffisant pour amener l'eau, a travers un ensemble de canaux ramifies souvent construits en terre, vers les champs, ou des systemes peu sophistiques mais efficaces de vannes et d 'obturations temporaires permettent 1'irrigation des parcelles. b) Le systeme actuellement visible On distingue trois secteurs principaux, que 1'on decrira sommairement avant de detailler celui du territoire actuel de Cieza. - Le premier tron<;on dessert la huerta de Cieza; il se developpe sur les deux rives du fleuve, les premieres "prises" d' eau etant situees a une dizaine de kilometres a l'Ouest de la ville. On retrouve encore, aussi bien sur la photographie aerienne que sur le terrain, la boucle d'une ancienne acequia, vraisemblablement d'epoque islamique, qui circonscrit les terrains bordant le meandre juste en amont de Cieza. Sur la rive gauche comme sur la rive droite, les eaux canalisees rejoignent le Segura en aval de Cieza, juste avant la limite territoriale qui la separe d 'Abanin: frontiere, en epoque islamique, entre les hisn/s de Siyasa et d' Abanin. - Le deuxieme tron<;on conceme le territoire de Blanca I Abanl.n. Deux acequias principales prennent leurs eaux, respectivement rive gauche et rive droite, a la limite du territoire avec Cieza et parcourent la huerta jusqu'a la petite ville de Blanca. Sur la rive gauche, I' Acequia de Blanca a son origine a la prise dite "del Menju", sur 1' embouchure de la Rambla del Moro; au long de son cours, elle est aujourd'hui equipee de trois norias, celle de Don Garcia, la Principale et celle de F emandez 35 . La tradition orale locale 36 date l'etat actuel de la noria de Don Garcia du XIXe siecle, mais fait remonter sa construction a 1' epoque d 'Isabelle la Catholique a la fin du XVe siecle; les autres sont plus recentes. Sur la rive droite, 1'Acequia de Charraca, conserve deux norias, toutes deux moder- 46 nes. Dans une des vallees laterales importantes, formee par la Rambla de Benito, cette acequia court au sommet d'un aqueduc qui traverse une petite vallee laterale; une analyse de la ma<;onnerie de briques et des modes de mise en reuvre indique qu'il fut construite a la fin du XVe siecle ou clans la premiere moitie du XVIe siecle, c'est-a-dire au moment d'une nouvelle colonisation du Valle de Ricote; il servait a irriguer de nouveaux terrains plus eloignes du fleuve. Des deux cotes du fleuve, les canaux principaux suivent la limite entre les terres les plus basses et les premieres collines, au relief plus accidente; ainsi, sur la rive gauche, on voit le canal contoumer systematiquement les petites hauteurs sur lesquelles -sous forme de batiments isoles ou de petits hameaux- est implante 1'habitat. Au sud d' Abaran, 1'Acequia de Blanca est prolongee par 1'Acequia de Abaran, qui suit les premieres pentes sur lesquelles est installee la vieille ville d' Abaran et ses extensions plus recentes; le canal dessert encore un moulin avant de rejoindre le f!euve. Il est presque certain que ces acequias sont, au moins pour une partie de leur trace, d'origine arabe 37 . Des deux cotes du Rio Segura, les canaux franchissent le goulot de Blanca. Rive droite, la Charraca irrigue encore quelques terres agricoles avant de rejoindre la riviere. Rive gauche, 1'Acequia de Blanca passe sous la ville actuelle: le trace tortueux de la rue principale --qui sert de limite entre la partie d'origine arabe et la partie chretienne de la villa de Blancacorrespond au trace de 1' acequia qui suit la limite entre les hauteurs situees au pied du chateau ou se trouvait 1'habitat islamique, et les terres basses, occupees seulement apres la conquete. Le canal rejoint le fleuve en aval de la ville de Blanca. - Le troisieme tron<;on est constitue par les canaux de Ricote-Oj6s-Ulea. Au pied du chateau de Ricote, qui ainsi domine et controle les prises d 'eau, un barrage alimentait deux canaux, un de chaque cote du fleuve; il est recouvert par 1'amenagement modeme de la centrale electrique, qui ferme aujourd 'hui le goulot de Ricote. On ne reconnait dans ce secteur qu'un seul trace de canal principal, avec, sur la rive gauche, une seule noria; on constatera d 'ailleurs que la vallee n'est pas tres large a cet endroit et ne presente pas non plus de vallees laterales, ce qui rend inutile la construction d' acequias supplementaires. Le parcellaire y a ete fortement modifie, mais la presence, sur le trace actuel, d'irrigations d'epoque musulmane parait assuree. Comme a Blanca, les deux Martinez Soler & Banegas Ortiz 1994. Note en novembre 1994 aupres d'un huertano. 37 Compte tenu de la topographie du terrain, ils n'ont pu etre modifies ou deplaces. 35 36 Quelques aspects du peuplement medieval du Valle de Ricote (Murcie, Espagne) [A) • Boom hydraullque • et ~poque moderne [B) Epoque chr~tlenne (raccordements) 0 [C) Epoque lslamlque 500m Fig. 3.- Interpretation- apartir d'une photo aerienne- du secteur irrigue du moyen Segura en amont de la vi/le de Cieza. Les acequias des differentes epoques (A, Bet C) et le parcellaire en lamelles des premiers terrains irrigues (C); 1. localisation d 'une eventuelle noria irriguant les terres situees en amont; 2. passage en tunnel du canal. canaux depassent le goulot d'Ulea I Oj6s. Sur la rive gauche, 1' acequia passe sous les maisons d 'Ulea pour alimenter un moulin situe 1' extremite orientale du village; puis elle rejoint le Segura. Sur la rive droite en revanche, 1' acequia poursuit son cours vers la ville de Villanueva del Rio Segura, puis en aval de celleci. La presence d'une autre acequia entre le fleuve et le canal venant de Ricote I Oj6s permet de considerer celui-ci comme plus recent. De toute fayon, clans cette partie du val de Ricote, la longueur des acequias signale des amenagements d'epoque chretienne. Plusieurs problemes se posent l'examen des systemes hydrauliques du Valle de Ricote. Ils sont anciens, mais sont-ils medievaux ou antiques? Les phases chronologiques de leur mise en place, peuvent-elles etre etablies et qu'avait-il avant le premier reseau? Comment ces systemes se sont-ils maintenus atravers les siecles et quelle confiance peut-on faire aux observations realisees aujourd'hui sur des structures aussi fragiles? La reponse a cette demiere question releve de la reflexion methodologique: on sait que, tout au long des epoques modeme et contemporaine, les canaux ont du etre repares, voire reconstruits; leurs traces cependant sont restes les memes. Pour 1'observateur d 'aujourd 'hui, trois systemes se superposent, les plus recents venant reutiliser des portions, conservees ou remodelees selon un trace identique: a a - le reseau de developpement actuel ou sub-actuel, qui utilise des eaux "extemes" provenant des nappes phreatiques (remontees par pompes electriques) ou des adductions artificielles (par tuyaux souterrains et tunnels) venant d'autres bassins hydrauliques (surtout celui du Tage), - le reseau modeme, decrit par les textes chretiens et dont le trace remonte a plusieurs epoques mais principalement aux XV e-XVIe siecles 38 , - enfin le reseau d'epoque islamique, dont nous allons tenter de retrouver quelques portions et d' en comprendre le fonctionnement. 3 Le reseau hydraulique de Cieza I Ricote On sait que 1'hydraulique, qui fait partie des connaissances agronomiques, a ete etudiee et divulguee par les agronomes andalous 39 ; les donnees que procure cette documentation peuvent etre completees par quelques textes juridiques40 qui, de favon partielle et occasionnelle, expliquent les modes habituels 38 Communication personnelle de Guy Lemeunier. Bolens 1981. 40 Glick 1970; de Epalsa 1988, 13-19; Barcel6 1989, VIII-L; Bolens 1989, 71-87; Lagardere 1991, 83-122; Lagardere 1992, 213-225. 39 47 A. Bazzana, J. De Meulemeester & A. Matthys d'usage et de gestion de !'eau. Cependant, une autre categorie documentaire facilite la comprehension du fonctionnement des systemes islamiques: depuis quelques annees, c'est l'archeologie qui montre !'organisation spatiale et les caracteres technologiques de I'hydraulique agraire andalouse 41 . A. Problbnatique gemirale On sait que, pour anciens qu'ils soient, les travaux hydrauliques s' inscrivent dans les paysages et laissent des traces qui peuvent traverser les siecles. Dans la huerta de Murcie comme dans la partie basse du cours du Segura (entre Orihuela et la mer), plusieurs etudes ont ete effectuees recemment. Certes, I' archeologie hydraulique de la huerta murcienne proprement dite reste, en bonne partie, a faire; toutefois, on commence a disposer d'informations concemant les axes principaux de la structuration du paysage agraire. S'il est impossible, par manque de donnees, de poser globalement le probleme des espaces agricoles de toute une region, on peut en revanche proposer de nouvelles problematiques a propos des regadios medievaux, de leurs origines et de leurs phases d) evolution. demarche prend appui sur les methodes de travail de ce que I'on appelle "I 'archeologie extensive"42 , c 'esta-dire que sont privilegiees les prospections au sol, I' analyse de photographies aeriennes et un essai de modelisation des resultats obtenus a partir des etudes de cas les plus productives; la carte permet la representation et la mise en evidence des informations recueillies mais c'est la photographie aerienne verticale (avec, par exemple, des agrandissements au 11 5000) que se trouve etre le document de base de la recherche: elle nous procure, en effet, les indices qui permettent de diriger et d'orienter les prospections, et foumit quelques donnees de chronologie relative eclairant !'evolution du paysage. Le mode de raisonnement, a partir des traces revelees par les cliches aeriens recents -qui revelent souvent, on le sait, une realite disparue- tente de remonter aux siecles anterieurs afin de retrouver, dans la mesure du possible, I' etat medieval des structures etudiees; un principe de base est que le trace des installations hydrauliques agraires repond a une veritable "loi" qui interdit les modifications et les transformations partielles: dans la majorite des cas, le paysage reste dans l'etat ou il etait au debut, seulement touche par les mutations majeures qui sont, dans notre exemple de Cieza I Abanin, la conquete feodale du XIIIe siecle ou les phases successives de depeuplement I repeuplement des XIVe-XVe siecles. B. Considerations methodologiques Ce type de recherche suppose quelques difficultes methodologiques. La premiere resulte directement des realites du terrain prospecte. I! s'agit d'abord de zones de vallee densement occupees et avec de nombreuses constructions. Ensuite, la huerta connait une mise en culture dans laquelle orangers et citronniers dominent largement; la prospection de tels vergers, peu ou pas laboures s'avere difficile. De plus, il s'agit d' etudes urgentes puisque les destructions actuelles, dues a I' extension urbaine, a la reorganisation des axes de circulation et, surtout, a I' amenagement, depuis les annees quatre-vingt de nouveaux canaux d'irrigation, arrachent ou effacent toutes traces des anciens systemes; seul le trace, laisse plus ou moins intact, subsiste. Une autre difficulte, latente mais bien reelle a I'heure de rassembler des informations sur les irrigations medievales, est la rarete de la documentation ecrite: rarete relative pour I' epoque chretienne, rarete quasi totale pour l'epoque islamique. Comment, done, faire cette enquete et tenter une reconstitution theorique des structures d'irrigation d'epoque islamique? La methodologie suivie est simple mais i1 est bon de la preciser, ne serait-ce que pour que le lecteur en saisisse les possibilites mais aussi les limites. La 48 C. Les sources chretiennes nous eclairent sur les regadios du bas Moyen Age Les travaux recents des geographes, tout comme ceux des historiens de I' epoque modeme ont mis au jour !'organisation spatiale et territoriale des regadios ciezans. Mais le probleme subsiste de savoir quelles sont les modifications d'epoque chretienne et quel etait le premier regadio musulman. Pendant I' epoque chretienne, la premiere evolution que I' on constate met en evidence, se Ion G. Lemeunier, une forte extension des perimetres irrigues et une consolidation des infrastructures 43 ; le meme auteur pen se que, avant, c 'est-a-dire a la fin de l'epoque islamique, l'economie agricole de Madlna Siyasa reposait sur la mise en valeur d 'un regadio alimente par les eaux du Segura, grace au canal "de 41 Voir, par exemple, Kirchner & Navarro 1994, 159-182. Voir, comme synthese rapide des problemes poses par les techniques de regadfo et de maltrise de !'eau au Moyen Age, Bazzana 1994, 317-335. 42 Bazzana & Guichard 1988. 43 Lemeunier, s.p. Quelques aspects du peuplement medieval du Valle de Ricote (Murcie, Espagne) 1'Andelma" (aujourd'hui appele Acequia de Landerma); sur la rive gauche, quelques hameaux exploitaient les eaux provenant des sources de Ascoy, d'El Ojo et de Bolvax 44 . Anterieures a la realisation de prospections archeologiques, ces idees ne sont pas fausses et mettent bien en relation -et avec raison-la vie de la madina islamique et !'existence d'un systeme elabore d'irrigation. I1 convient cependant de completer et, surtout, nuancer ces affirmations: on verra plus loin qu 'il n' existait pas un seul regadio, au pied des pentes descendant de Siyasa, mais que plusieurs zones basses, dans les meandres du Segura etaient aussi utilisees clans un systeme original et complexe qui dessinait, tout au long du fleuve, une sorte de cha\'ne de terres irriguees; de meme, il faudra imaginer qu'on ne peut penser, pour l'epoque anterieure au milieu du XIIIe siecle, a !'existence d'une seule -et longue, evidemment- acequia sur la rive droite; de meme que pour la rive gauche, il ne s'agissait pas d'un canal unique mais d'un succession de plusieurs canaux, courts et charges chacun d'irriguer un espace de petites dimensions. Le "boom hydraulique" commence vers 1'annee 1480 et se prolonge pendant un peu plus d'un siecle; apres la construction de la Acequia del Homo, dans la premiere moitie du XVIe siecle, on peut considerer que le systeme est pratiquement acheve: il ne va pas se modifier ensuite sauf quand viennent a se developper -mais seulement dans notre XXe siecle- de nouveaux moyens techniques (puits profonds, pompes electriques, transvases, etc.). Apres la conquete chretienne, les seigneurs castillans avaient interet non seulement a repeupler les territoires conquis (ou, au moins, a essayer de maintenir sur place les populations musulmanes), mais aussi a mettre en etat de marche le systeme d'irrigation, base de prosperite economique pour toute la region. Pour cela, les seigneurs ne tarderent pas a investir pour assurer aterme une croissance de leurs revenus: quand, en 1483, le Commandeur de l'Ordre de Santiago signe avec les mudejares de Hellin une charte de repeuplement du village d' Abaran, illeur promet de reparer et d' entretenir les canaux qui prennent de 1'eau dans le fleuve. D. Le regadio medieval de Cieza L'etude des structures de la petite hydraulique agraire qui apparaissent sur le territoire de Cieza montre 1'existence de sources naturelles exploitees depuis les premiers siecles du Moyen Age: elles se trouvent associees a quelques petits noyaux d'habitat situes sur les pentes de la vallee mais assez loin du fleuve; en epoque islamique, une partie des eaux d'irrigation '- .. ........_ \ \ \ ) . trace hypothettque (epoque C) D micro-strig~tions (parcelbire mis en place en cpoque is la mique) [C) Epoque lslamlque [B) Epoque chretlenne (raccordements) Fig. 4. - Evolution du systeme hydraulique et de !'implantation des canaux atravers les phases principales du developpement du reseau des acequias. provenaient de ces sources que l'on trouvait principalement sur la rive droite; certaines auraient ete mises en etat de production par les agriculteurs musulmans; quelques vestiges, dont parlent certains documents, etaient encore visibles et fonctionnels aux XVe et XVIe siecles; ce type d'amenagements suscitait alors une grande surprise: on les disait "construits avec une technique incroyable". Peu de traces de 1'utilisation des sources se conservent; en revanche il reste possible d'etudier les traces des canaux et d'en tirer quelques elements chronologiques. 44 Ibid. 49 A. Bazzana, I. De Meulemeester & A. Matthys a) Structuration et chronologie On est conduit a distinguer trois phases chronologiques: une phase A, qui correspond au developpement du systeme des canaux durant le "boom hydraulique" du XVIe siecle; une phase B, caracterisee par les premieres reparations d'epoque chretienne; enfin, une phase C, qui correspond au reseau d'epoque islamique. L' examen des photos aeriennes et la prospection au sol montrent 1' existence, sur les deux rives du Segura, d'un double reseau d' acequias. Apartir des premieres "prises", situees a 1'Ouest de la ville, il faudrait distinguer deux series de canaux (epoques A et B). - Au Nord du Segura, sur la rive gauche, une acequia suit la limite entre les terres basses, disposees le long du fleuve, et les premieres pentes des collines voisines qui offrent un relief plus accidente (epoque B); le trace de cette acequia presente une particularite sur laquelle on devra s'interroger: pourquoi, a plusieurs reprises, s'eloigne-t-elle du fleuve -ce qui permet d' irriguer une zone assez vaste- pour ensuite se rapprocher de nouveau du lit du Segura, comment si elle essayait de le rejoindre? En realite, elle s'en eloigne une nouvelle fois afin d'alimenter une nouvelle zone irriguee. Une autre acequia -dite "Del Homo", du nom d 'un ham eau proche de la prise (construite en 1600)- prend ses eaux a trois kilometres en amont du premier canal et se trouve situee sur les premieres pentes des collines qui bordent la vallee (epoque A); elle presente un parcours plus tourmente qui lui permet de penetrer toutes les petites depressions ou vallees laterales, ou de bonnes terres sont disponibles: la surface irriguee est des lors beaucoup plus ample; elle acheve son cours clans le fleuve, au lieu-dit "Del Menju". - Au Sud, sur la rive droite, apparait la meme disposition spatiale de deux acequias distinctes: il s' agit de la Acequia Landerma ( ou canal "de 1'Andelma", clans la documentation medievale) et de la Acequia Nueva; elle sont separees par des distances variables selon les cas (parfois seulement quelques dizaines de metres); le differentiel d'altitude etant de l'ordre de quelques metres, ce qui est bien suffisant pour augmenter sensiblement 1'espace irrigue. I1 s'agit done d'une double serie de deux canaux qui apparait sur la carte, sur les photos aeriennes et sur le terrain. I1 est clair que nous devons regrouper typologiquement, d'une part les deux acequias les plus proches du fleuve (acequias basses), d'autre part les deux autres, implantees sur les pentes des collines (acequias hautes); il convient d'interpreter cette double realite en fonction du "principe de rigidite" dont nous avons parle et selon lequell 'une comme 1'autre repondent a des logiques differentes. Si l'on veut conna'itre leurs dates de construction, il faut recher50 cher deux epoques distinctes (1 'une etant plus recente que !'autre): les acequias hautes correspondraient alors a 1'epoque A -la plus proche d' auj ourd 'hui dont les donnees historiques indiquent une construction se situant aux XVe et XVIe siecles; les acequias basses a une epoque B ou C, anterieure a la precedente et qu' il n' est pas illogique, si 1'on tient compte de ce que 1'on sa it des irrigations musulmanes de Cieza, de situer avant la conquete chretienne. De toute maniere, les acequias hautes, dont les prises sont installees en amont de ce1les qui alimentent les acequias basses, ne peuvent etre que posterieures a 1'epoque islamique; la presence meme de ces acequias hautes est en contradiction complete avec ce que l'on sait du droit islamique des eaux: les traditions juridiques interdisent que l'on s'approprie les eaux d'un fleuve (ou d'un canal) au-dessus d'un barrage deja existant, car ce serait un veritable vol de 1'eau a ceux qui y etaient insta11es les premiers 45 . Sur la rive droite, un canal "ancien", c'est-a-dire d'epoque B ou C et appelee Acequia Landerma -il figure clans les textes medievaux sous le nom de "Canal de 1'Andelma"46- parcourt les terres plus pres du fleuve; il est, pour ainsi dire, double quelques metres plus ha ut par la Acequia Nueva; entre le "vieux canal" et le fleuve, le parcellaire, vu sur les photos aeriennes, presente une disposition caracteristique en micro-strigations paralleles ou en leger eventail. C'est ce que l'on remarque bien, parmi d'autres exemples, avec les parce11es situees clans le meandre de la rive gauche, en aval de la ville: la, cependant, une double orientation des parce1les pourrait signal er un essai de restructuration realise a une epoque inconnue. Une rapide prospection au sol montre que ces parce11es avaient initialement une largeur moyenne d'environ 7m. Sur l'une comme sur !'autre rive, les eaux canalisees retournent au Segura en aval de Cieza, juste avant d'arriver ala limite territoriale qui separe cette ville du territoire d 'Abanin: limite ou frontiere mineure d'epoque islamique tardive, entre les chateaux de Siyasa et de Blanca. S'agissant des acequias basses, l'examen des photos aeriennes revele un phenomene particulier: les canaux suivent un trace presentant quelques anomalies; en completant notre information par des donnees de terrain, cinq series d'elements apparaissent. On a deja note que le canal de 1'Andelma, en un trace date de l'epoque B, semblait dessiner, a partir de la prise Del Menju, une courbe de grande ampleur; a la fin de cette courbe, il se rapprochait du fleuve, passant a peine a quelques metres du lit actuel. 45 46 Bolens 1989; Lagardere 1991 et 1992. Lemeunier, s.p. Quelques aspects du peuplement medieval du Valle de Ricote (Murcie, Espagne) I1 parait clair, d'autre part, que ceci ne se produit pas de maniere accidentelle mais que, au contraire, se repete sur les deux rives du fleuve, le long des deux acequias basses seulement: celles-ci quittent le Segura pour, ensuite, s' en approcher une nouvelle fois avant de s' eloigner de nouveau. Ainsi, elles dessinent une serie de festons, qui delimitent des superficies variables. - Aucun indice de ce type n' apparalt au long des acequias hautes qui, de leur cote, presentent un trace paraissant dependre beaucoup plus (et peut-etre uniquement) du relief: clans ce cas, c'est la presence plus ou moins pregnante des col lines et c 'est la vigueur des pentes qui conditionnent 1' emplacement du canal par rapport au fleuve. - Un feston se termine toujours avant qu 'un autre ne commence a se developper. Que se passe-t-il done clans ce secteur de la vallee? Ce que suggerent !'analyse des documents cartographiques et photographiques anterieurs aux recents travaux de regularisation du Segura et ce que confirment les prospections archeologiques, c 'est que no us sommes la en presence de vestiges tenus mais, bien reels d'une ancienne organisation hydraulique. - Ainsi, juste en aval de la ville actuelle de Cieza, le canal appele aujourd'hui Acequia Landerma s'interrompait, durant une phase primitive (d'epoque C?), pour retourner au fleuve; alors, c 'est un autre prise d'eau, situee plus en aval, qui permettait le developpement d'un nouveau canal qui se developpait ensuite vers le Sud-Est. 11 se passait la meme chose rive gauche, ou, anterieurement aux phases A et B, la acequia arrivant de l'Ouest s'interrompait juste avant d'arriver a Cieza, pour reprendre plus loin son cours, la ou le canal actuel prend le nom de Acequia Marcelino: sur la photo aerienne comme sur le terrain, on remarque encore la presence d'une ample courbe que dessine une ancienne acequia, vraisemblablement edifiee en epoque islamique. Elle vient delimiter les terres qui bordent le meandre situe en amont de Cieza; la prise est aujourd'hui detruite mais on peut encore en retrouver des vestiges; le debouche du canal reste, lui, bien visible. Sur tous ces espaces, delimites par les acequias dont nous venons de par! er, se developpent des champs cultives (jardins irrigues) qui presentent une distribution parcellaire caracteristique, au moins clans la Peninsule iberique et au Maghreb, de 1' epoque islamique 47 : parcelles etroites et allongees, qui forment des micro-strigations souvent disposees en eventail; ces parcelles se disposent perpendiculairement au trace 47 48 Guy 1977; Bazzana et alii 1986. Bazzana 1986, 15-28. du canal et/ou du lit du fleuve. Evidemment, ne n'est qu 'entre la Acequia de 1'Andelma et le Segura, oud'une maniere plus generate- clans les secteurs situes entre les acequias d'epoque Bet le fleuve que l'on peut faire ces observations; au contraire, le parcellaire est totalement different entre, par exemple, la Acequia de Landerma et la Acequia Nueva: la, il s' organise selon une maille carree qui correspond a un parcellaire plus recent (epoque B, avec modifications en epoque A). Parfois cependant, cette maille vient recouvrir des zones theoriquement inclues clans le systeme de regadio ancien (c 'est-a-dire irrigue a partir des acequias les plus anciennes): ceci indique que, bien que le canal primitif ait conserve son trace anterieur ( d'epoque C), le parcellaire lui a ete reorganise et redistribue, peut-etre apres une phase d' abandon et de ruine des structures de distribution de !'eau; clans 1'histoire de la region, c 'est la periode fin XIIIe siecle - XIVe siecle qui correspond le mieux a ce moment. D'une maniere generate, clans les systemes d'irrigation d'al-Andalus et comme on a pu !'observer clans le cas de Siyasa, il n 'y a pas de canal unique, apte a repartir les eaux sur 1' ensemble du territoire castral; au contraire, apparait une sorte de succession de petites unites separees les unes des autres par des espaces non irrigues: chacune dispose de sa propre prise d'eau. De plus, ces micro-irrigations se distribuent tout au long de la vallee, ou 1'on voit se dessiner une chaine de zones vertes, bien delimitees par les acequias; le droit islamique con forte 1' existence ce ces petites unites -entites fermees entre la prise et l'exutoire- qui, s'agissant d'un fleuve comme le Segura, disposent de la quantite d'eau necessaire. Enfin, en relation topographique directe avec le regadio ainsi etabli, existe un habitat d'epoque islamique, situe a cote de la zone irriguee et a quelques metres au-dessus du canal majeur; c'est ce qui a deja ete mis en evidence clans la petite vallee du Rio Albaida, a Valence 48 ou le mode le est parfaitement clair puisque achaque zone de regadio correspond un habitat actuel ou une alqueria medievale, aujourd'hui abandonnee. b) Conclusions: vers une archeologie agraire du Valle del Segura Dans cette premiere approche du probleme des irrigations medievales de Cieza, nos conclusions sont, a !'evidence, provisoires. Nous avons pose, a partir de 1' exemple du Valle de Ricote, un certain nombre de problemes, mais toutes les reponses n' ont pas encore ete trouvees; nous esperons qu'il sera possible d 'all er plus loin clans 1'interpretation des vestiges de structures agraires medievales mais cela suppose 51 A. Bazzana, J. De Meulemeester & A. Matthys d'autres etudes de cas, dans le cadre d'un travail d 'equipe. I! convient aussi de rester bien conscient du fait que, pour I' instant, si la problematique est claire, les reponses ne sont que partielles et que, pour progresser, il conviendra d'approfondir la reflexion methodologique. Cependant, telle qu 'elle a ete menee -pendant une trop courte periode et avec des moyens limites1' enquete realisee foumit des renseignements utiles et enrichit la reflexion qu'il convient de mener sur les terroirs et les structures agraires du Moyen Age islamique. Deja, !'etude des regadios de Cieza -et le meme resultat pourrait etre obtenu en prolongeant la recherche plus en amont, sur le territoire de Calasparra- montre bien la difference entre les deux types de societes -musulmane d'abord, puis chretiennequi se succedent dans ce monde rural du moyen Segura. En epoque chretienne, le fonctionnement des communautes rurales s'inscrit dans le cadre d'une societe feodale: aussi trouve-t-on de nombreuses traces de !'intervention des pouvoirs d'Etat -ou, avec les Ordres militaires, semi-etatiques- et des autorites locales, dans le contra le et 1'etitretien, mais aussi dans la construction des prises et des acequias abandonnees. De la conquete du milieu du XIIIe siecle jusqu'a nos jours, les acequias sont conyues pour constituer un systeme continu se developpant sur une distance moyenne (qui peut atteindre, dans le cas du Valle de Ricote, plus de 1Olan de longueur). Il est clair que, si I' on engage des travaux sur de longue distance, il fa ut aussi en vi sager I' aide de techniciens et d' ouvriers specialises, aptes a calculer les pentes et les gabarits, a construire des canaux en materiaux resistants, a edifier eventuellement des aqueducs etc. L' intervention de I 'Etat est, des lors, quasiment necessaire49. Au contraire, dans le cadre de petites communautes rurales semi-independantes -fonctionnant sur les modes d'une societe tributaire, ou les responsabilites sont collectives-, les canaux d'epoque islamique s 'organisent sur de courtes distances et se succedent, en chapelet, le long du fleuve ou s'etendent les zones irriguees en forme d'amande s'effilant la ou se trouvait, en amont la prise d'eau, en avall'exutoire permettant un retour au fleuve des eaux non utilisees. Le chapelet des terroirs irrigues disposes le long du fleuve est a I 'echelle du village (aldea, alqueria) ou, mieux, du hameau de quelques maisons, qui s'installe a proximite meme des terres mises en culture et qui reclament tous les soins; le systeme est, de plus, en accord avec le droit des eaux en pays islamique, 49 Voir l'exemple bien connu de la Acequia Real del Jucar, Valence. 52 a qui interdit aux paysans de prendre de I 'eau en amont d 'un canal deja existant. Ainsi, chaque proprietaire ou communaute doit elaborer son propre systeme d'irrigation avec sa propre prise d'eau et son propre exutoire a la riviere. Les canaux secondaires a leur tour divisent le terrain en parcelles de largeur reduite (quelque 7m). On peut pens er que 1'origine de ce mode d 'amenagement est liee aux capacites techniques des constructeurs, c'est-a-dire les paysans euxmemes: il s'agit d'ouvrages simples, qui ne necessitent guere de connaissances techniques; les canaux peuvent meme etre simplement creuses dans la terre; quant aux prises d'eau, elles font appel a la mise en reuvre de materiaux facilement accessibles -le bois (pieux, fascines, branchages divers) et la pierre (galets de riviere, blocaille, graviers, etc.)- qui composent une ma9onnerie bien adaptee ason usage. La fragilite en est evidente, mais il n' est pas question de tenter de !utter contre les violentes inondations mediterraneennes et, a Murcie comme dans le Maghreb actuel, on reconstruit periodiquement les ouvrages hydrauliques que les eaux ont pu detruire. Deux problemes restent aapprofondir: d'une part, celui de ]'organisation agraire de ces regions en epoque antique (principalement romaine et romaine tardive) et au moment du premier developpement du peuplement medieval, pendant les premiers siecles du Moyen Age, d'autre part celui de la chronologie precise de l'amenagement des reseaux medievaux: il s'agit sans doute du Xe siecle, moment privilegie de la "revolution agricole" selon L. Bolens. Le reseau des acequias musulmanes fut herite mais sans doute dans un etat d'abandon et de destruction (partielle) assez generalise- par les Chretiens. La remise en bonnes conditions de production n'intervient pas avant la fin du Moyen Age: seule la conquete de Grenade par les Rois Catholiques, en 1492, supprime definitivement les menaces que faisait planer le tout proche royaume nasride sur les terres murciennes. Dans la huerta de Cieza, pendant le XVe siecle, avant meme que se manifeste le "boom hydraulique" -qui s 'initie vers 1480 et se poursuit pendant tout le XVIe siecle- s 'organise en une succession de terres irriguees ou se developpe une vegetation verdoyante; la remise en etat de la vallee impose le recours a des techniques lourdes et couteuses, la recherche d, investissements reguliers et prolonges, ce qui, dans la majorite des cas, impose une intervention de I 'Etat. Les irrigations murciennes d, epoque islamique etaient d 'une tout autre nature: leur origine n'etait ni technique ni politique, mais sociale. I! s'agissait, jusqu'au XIIIe siecle, d'une hydraulique sans hydrauliciens, sans specialiste en hydraulique ni manieurs de capitaux; elle correspondait a la mise en valeur de terres irriguees -selon Quelques aspects du peuplement medieval du Valle de Ricote (Murcie, Espagne) quelques precedes elabores et complexes, mais utilisant des techniques simples- par de petites communautes campagnardes. Bibliographie BARCEL6 M. 1989: El disefio de espacios irrigados en al-Andalus", in: El agua en zonas ciridas: arqueologia e historia, Almeria, VIII-L. BARCELO M., CARBONERO M.A., MARTi R. & ROSSELLO-BORDOY G. 1986: Les aigiies cereades. Els qanat(s) de l 'ilia de Mallorca, Palma de Majorque. BAZZANA A. 1986: Terroirs et peuplement au Moyen Age dans l'Espagne musulmane: une etude de cas dans la vallee du Rio Albaida (Valencia), Photointerpretation. Images aeriennes et spatiales 84/ 3, 15-28. BAZZANA A.et alii 1986: Photo-interpretation. Images aeriennes et spatiales, numero special, n° 84-2 et 84-3. BAZZANA A. 1990: Maisons rurales du Shark Al-Andalus. Essai de typologie, in: BERMUDEZ LOPEZ J. & BAZZANA A., La casa hispano-musulmana. Aportaciones de la arqueologia - La maison hispano-musulmane. Apports de l' archeologie, Publicaciones del Patronato de la Alhambra y Generalife, Granada, 247-267 et 370-372. BAZZANA A 1992: Maisons d'al-Andalus. Habitat medieval et structures du peuplement dans l 'Espagne orientate, Collection de la Casa de Velazquez n° 37, Madrid. BAZZANA A. 1994: La pequefia hidraulica agricola en al-Andalus, in: Ciencias de la naturaleza en alAndalus, Ill- Textos y estudios, Madrid (C.S.I.C.), 317-335. BAZZANA A., CRESSIER P. & GUICHARD P. 1988: Les chateaux ruraux d 'al-Andalus. Histoire et archeologie des husun du Sud-Est de l 'Espagne, Collection de la Casa de Velazquez n° 19, Madrid. BAZZANA A. & GUICHARD P. 1988: Archeologie extensive clans la region valencienne (Espagne), in: NOYE GH. (ed.), Castrum 2. Structures de l 'habitat et occupation du sol dans les pays mediterraneens: les methodes et l 'apport del' archeologie extensive, Collection de 1'. Ecole Franyaise de Rome 105 - Publications de la Casa de Velazquez, serie archeologie IX, Rome-Madrid, 3-28. BAZZANA A. & MONTMESSrN Y. sous presse: Na'ura et saniya clans l'hydraulique agricole d'al-Andalus, in: La maftrise de l 'eau au Moyen Age dans la Peninsule iberique et au Maghreb al-Aqsa, Madrid. BERTRAND M. & CRESSIER P 1985: Irrigation en amenagement du terroir clans la vallee de 1'An- darax (Almeria): les anciens reseaux de Ragol, Melanges de la Casa de Velcizquez XXI, 115-135. BERTRAND M. & CRESSIER P 1986: Antiguos sistemas de irrigaci6n en el Valle de Andarax, Arqueologia medieval espaiiola, I Congreso, Ill, Saragosse, 569-580. BOLENS L. 1974: Les methodes culturales au Moyen Age d'apres les traites d'agronomie andalous: traditions et techniques, Geneve. BOLENS.L. 1981: Agronomes andalous du Moyen Age, Geneve. BOLENS L. 1989: L'irrigation en al-Andalus: une societe en mutation, analyse des sources juridiques, in: El agua en zonas ciridas: arqueologia e historia, Almeria, 71-87. BOLENS L 1990: La revolution agricole andalouse au XIe siecle, in: L 'Andalousie du quotidien au sacre. XIe-XIIIe siecles, Collected Studies Series 337, Aldershot-Hampshire, 9-29 = Studia Islamica, XL VII, 1978, 121-141. BUTZER K.W. & E., MATEU J.F. & KRAU P. 1985: Irrigation Agrosystems in Eastern Spain: Roman or Islamic origins?, Annals of the Association of American Geographers 75, n° 4, 479-509. CARMONA GONZALEZ A. 1990: Recorrido por la Geografia Hist6rica, in: Guia Islamica de la Region de Murcia, Colecci6n Ibn Al' Arabi, Consegeria de Cultura, Educaci6n y Turismo, Murcia, 13-29. COLrN G.S. 1932: La noria marocaine et les machines hydrauliques clans le monde arabe, Hesperis, XIV, 22-60. COLIN G.S. 1933: L'origine des norias de Fes, Hesperis XV, 156-157. DE EPALSA M. 1988: El agua en el derecho musulman, in: Agua y poblamiento musulmcin I Aigua i poblament musulma, Benissa, 13-19. DE MEULEMEESTER J. & MATTHYS A. 1995: Un grenier collectif fortifie hispano-musulman: Le Cabezo de la Cobertera (Vallee du Rio Segura/ Murcie). Bilan provisoire d'une appproche ethnoarcheologique, in: Ethno-archeologie mediterraneenne- Madrid 1991, Publications de la Casa de Velazquez, Madrid, 181-196. FosSIER R. 1992: Villages et villageois, in: Villages et vi!lageois au Moyen Age, Paris, 207-214. GLICK Th. 1970: Irrigation and society in medieval Valencia, Harward-Massasuchetts. GUICHARD P. 1983: Orient et Occident: peuplement et societe, Castrum I. Habitats fortifies et organisation de 1'espace en Mediterranee medievale, Lyon, Travaux de la Maison de 1'Orient 4, 87-93. GUICHARD P. 1990/91: Les Musulmans de Valence et la Reconquete (XIe-XIIIe siecles), Damas, Institut fran9ais d'etudes arabes, 2 vols. GUY M. 1977: Manuel de photo-interpretation, Paris. 53 A. Bazzana, J. De Meulemeester & A. Matthys HERIN s.d.: Les huertas de Murcie. Les hommes, la terre et I 'eau dans I 'Espagne arride, Aix-enProvence. JACQUES-MEUNIE Dj. 1951: Greniers-Citadelles au Maroc, Publications de 1'Institut des Hautes Etudes Marocaines LII, 2 vol., Paris. K!RCHNER H. & NAVARRO C. 1994: Objetivos, metodos y pnictica de la Arqueologia hidraulica, in: Arqueologia y territorio medieval, I, Aetas del coloquio "Problemas en arqueologia medieval", Jaen, 159-182. LAGARDERE V. 1991: Droit des eaux et des installations hydrauliques au Maghreb et en Andalus aux XIe et XIIe siecles, dans le Mi'yar d'al-Wansharisi, Cahiers de Tunisie XXXVII-XXXVIII, n° 145-148, 83-122. LAGARDERE V. 1992: Agriculture et irrigation dans le district (iqlim) de Velez-Malaga. Droit des eaux et appareils hydrauliques, Cahiers de civilisation medievale XXXV, n° 3, 213-225. LEMEUNIER G. sous presse: L'irrigation aMurcie au debut de 1'epoque moderne, Castrum5. Archeologie des espaces agraires, Madrid-Rome. MARTINEZ SOLER J.J. & BANEGAS 0RTIZ J. 1994: Las norias de Abarim, Aban1n. PEREZ PICAZO M.T. & LEMEUNIER G. s.d.: Els estudis d 'historia agraria a la regia murciana: I' estat de la qilesti6, Estudis d 'historia agraria 4, 63-84. PESEZ J.-M. 1992: Archeologie de la maison paysanne, in: Villages et villageois au Moyen Age, Paris, 181-192. PIRENNE J. 1977: La maftrise de I' eau en Arabie du Sud antique. Six types de monuments techniques, Memoires de 1'Academie des Inscriptions et Belles Lettres II, Paris. RODRIGUEZ LLOPIS M. 1988: Sefiorios y feudalismo en el Reino de Murcia, Murcia. SCHI0LER Th. 1962: Las norias ibericas, Revista de dialectologia y tradiciones populares XVIII, 480-486. SCHI0LER Th. 1973: Roman and islamic water-lifting wheels, Biblioteca Universitatis Hauniensis 28, Copenhague, Odense University Press. TORRES BALBAS L. 1942: La Albolafia de Cordoba y la gran noria toledana, Al-Andalus VII, 461-469. TORRES BALBAS L. 1940, Las norias fluviales en Espafia, Al-Andalus V, 192-208 (= Obra dispersa I, Madrid, 1981, 209-222). VERHULST A. 1992: Villages et villageois au Moyen Age, in: Villages et villageois au Moyen Age, Paris, 9-13. Andre Bazzana Directeur de recherche au C.N.R.S. Unite mixte de recherche 5648 (Lyon) France Johnny De Meulemeester Service des Sites et Monuments nationaux Luxembourg UMR 5648 (Lyon) Andre Matthys Inspecteur-general de la Division du Patrimoine Ministere de la Region wallonne Namur Belgique 54 Rural Settlement in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Christopher Dyer Recent developments and future prospects in research into English medieval rural settlements The Medieval Settlement Research Group has recently issued a new policy statement on all aspects of rural medieval settlements, which includes recommendations for conservation and presentation of sites, but its main concern is to define research priorities. This paper is designed to expand on the points made in that outline of policy. It reviews recent work and advocates future developments, reflecting the thinking behind the Group's formulation of policy, though inevitably it will be coloured by the individual interests of its author. The policy statement reflects the ever broadening conception of settlement studies, which in the view of the Group should encompass a wide range of sites and landscapes, not just every type ofhabitation (villages, hamlets, farms, seignorial residences etc.), but also the associated boundaries, fields, water control systems, woods, pastures, communications etc. Although the Group is concerned to focus attention on the medieval period, it recognizes that medieval occupation of a place was often an episode within a much longer period of use of the land, stretching back into prehistory and extending into modem times. In the recent past both survey and excavation has been concentrated on deserted settlements, and there are good reasons for this because these sites provide for the researcher invaluable opportunities to explore a time capsule of medieval rural occupation, undisturbed and unpolluted by modem features. Deserted sites are also more likely to be threatened at the present time with total destruction, making rescue excavation necessary. We are now facing up to the archaeological potential of all settlements, not just the villages which have suffered heavy shrinkage, or the hamlets that have dwindled down to a single farm, but also the still inhabited places, even villages which are now as fully occupied as they were in the high middle ages. The archaeological contexts in such places may have been damaged by modem activity, but they have the advantage of providing evidence about settlements which survived, and therefore represent the successful majority of places which escaped desertion. The broader view of settlement study involves contributions from the practitioners of a variety of disciplines, not just archaeologists, geographers and historians, but also environmental scientists, placename scholars, architectural historians, small finds and pottery specialists and many others. As well as striving to make use of every variety of evidence, multidisciplinary study also involves bringing to the subject approaches deriving from other subjects, such as anthropology. We only need to consider the recent work on the distinction between public and private space within houses and villages to appreciate the insights that settlement study can provide into mentality. The full implications of settlement planning for reconstructing the mindset of medieval people, which can in the twelfth century, for example, be linked with other evidence for a new rationality and orderliness in government and religion, is another example of the potential of settlement studies to contribute to intellectual and cultural history. This paper is divided into four sections. The first two are concerned with defining the distribution of hamlets and farmsteads, and with the analysis of regional differences in settlements and landscapes. The third deals with new insights into the lives of settlements and their inhabitants, including their late medieval decline. And finally research techniques, both methods for gathering evidence and for developing interpretations will be discussed. 1 Understanding dispersed settlement Most people in medieval England, in common with much of the population of continental Europe, lived in dispersed settlements rather than nucleated villages. This is true of the overwhelming majority in the period before c. 850, as our sample of excavated sites is heavily weighted in favour of small hamlets and isolated farms. The predominance of small and scattered settlement is still the case in the period after the formation of villages, because in the thirteenth and fourteenth centuries the areas of dispersal in55 C. Dyer eluded most of the west of the country, but more important the south-east, from the Wash to the Channel, where population densities were especially high. We still do not !mow enough about the relationship between the early medieval dispersed settlements and those of the later middle ages, but there are sufficient examples to suggest an overall discontinuity. Most !mown hamlets and farms of the preConquest period were abandoned within that period, and many of the late medieval dispersed settlements seem no older than the villages- their earliest phase of occupation seems to lie between the eleventh and thirteenth centuries. This is not to support the older interpretation of the dispersed settlements as part of the process of internal colonisation in the twelfth and thirteenth centuries, because we no longer believe that vast areas of wood and waste were occupied for the first time at that late date. It seems more plausible to argue that the settlement pattern everywhere went through a great upheaval in the tenth, eleventh and twelfth centuries, which led to a general creation of new settlements, not just in those regions which were heading towards 'villagisation', but also in those developing a scattered pattern ofhamlets (Dyer 1996). Such generalisations are however premature, because not enough basic research has been done on dispersed settlements to draw conclusions even for individual regions, let alone the whole country. The listing of medieval settlement sites in our basic sources of information, the county Sites and Monuments Records (SMRs), is patchy and incomplete. Their recording of 'deserted medieval villages' is often very full, though there is sometimes a tendency to describe any site with earthworks of abandoned settlement as a 'deserted medieval village'. For the pre-Conquest period the SMRs usually contain a complete record of finds, including any indications of settlement. But the only type of later medieval isolated settlement which is listed systematically is the moated site, which is of course a valuable source of information, but means that thousands of unmoated farmsteads are omitted. Gradually local and regional survey projects are beginning to gather the data. Very thorough work on upland landscapes, such as Bodmin Moor in Cornwall (Johnson & Rose 1994) has produced a complete catalogue of sites, and in parts of the lowlands of the county earthworks of abandoned farms and hamlets have been systematically recorded (MSRG 1994, 33). In Suffolk, a county previously thought to be devoid of earthwork remains of deserted settlements, evidence has been collected through an Earthwork Reconnaissance Survey (MSRG 1995, 39-40). Norfolk has long been the scene of research into settlements through intensive field walking of whole 56 parishes, and this plotting of pottery scatters is often combined with documentary research (Davison 1990). Another approach, for which a pilot study has been conducted in Hampshire, involves the use of existing archaeological and historical data and the gathering of information about both nucleated and dispersed settlements from the detailed maps of the early nineteenth century (MSRG 1995, 7-10). Parallel studies have been proceeding in Wales and Scotland, notably a survey of remains of houses and shielings on the Black Mountain in Carmarthenshire (MSRG 1994, 19), and various projects in the Scottish highlands, including the plotting of 'medieval or later' settlements from early maps ( MSRG 1994, 7-11 ). A-ll of these are welcome initiatives, but the longterm goal must be the assembly of data for whole counties which will enable us for the first time to see the totality of the medieval settlements. This means using a combination of evidence, as no single source of information will be adequate. Earthworks must be recorded both from aerial photography and detailed ground level observation, as the remains of a single farmstead will often escape detection from the air. Field walking is essential for detecting the many sites damaged by ploughing. And the documentary research must include, as well as early maps and county wide records such as tax records, the more detailed local archival material which is likely to mention individual farms and groups of cottages. The documents are especially valuable for establishing whether the existing settlements, most of which will not have early architectural evidence, are of medieval origin. Only when a larger sample of the country has been surveyed in this way will it be possible to classify the various types of hamlet by their plan forms, which are as varied as village plans, and, like the villages, include both irregular and regular layouts. We will also be in a better position to make judgements about the chronology, origins, development and abandonment of dispersed settlements. Such work will also allow comparisons to be made between regions. One suspects that the present statement that, for example, Kent and Lancashire both have dispersed settlements will need to be reformulated once we have learned more about the details of the distribution and form of their hamlets and farmsteads, as in every other respect the landscape and history of those counties are different. We ought to be able also to observe the relationship between the dispersed settlements and their fields and other resources. At present only a limited number of such sites have been properly excavated. Once their overall distribution has been plotted, it should be easier to make decisions about the preservation of threatened sites, and their excavation in advance of development. Recent developments and future prospects in research into English medieval rural settlements 2 Regional differences in settlement and landscapes The Monuments Protection Programme of English Heritage encountered a real problem in dealing with medieval rural settlements because they were so numerous, and presented so many difficulties in deciding which sites were 'important' enough to merit preservation. Accordingly a research project has been defining regional patterns so that individual sites can be judged in relation to rationally defined regions (MSRG 1994, 12-17). The maps that have been produced, whatever their initial purpose, are of course of significance for settlement studies in general. They are based on nineteenth-century maps, so they contain some data deriving from modem industrialisation, and will be influenced by rural depopulation in late medieval and modem times, but they are reflecting primarily the underlying patterns deriving from the great formative period of rural settlement between c.900 and 1200 or 1250. They indicate the major regional differences, such as the 'central province' cutting across the country from Northumberland to Dorset, in which nucleated villages predominated, and more subtle variations between much smaller districts. Workers in each locality might wish to revise its detailed conclusions, as would apparently be the case for example in the counties of central southern England discussed in a recent volume (Aston & Lewis 1994). Nonetheless it is valuable to have consistently applied criteria being used to define regional differences over the whole country. The same type of source material, nineteenthcentury maps, have been used on a study of four east midland counties, though here the data has been mapped alongside distributions of population, farming systems, and archaeological sites derived from Sites and Monuments Records (Lewis, Mitchell-Fox & Dyer 1997). The comparison between indicators of pre-medieval activity, soil types, population density, agricultural systems, lordship and other variables has been used to sharpen the basis of the discussion ofthe explanation of the origin and development of different settlement forms, above all between nucleated villages and dispersed settlements, and that debate will no doubt continue. Perhaps we exaggerate the differences between settlements and landscapes. We are accustomed to use a shorthand classification of 'nucleation' and 'dispersal', though it has always been recognised that in some areas dispersed settlements cluster into hamlets of varying size, and that in others most people lived in single farms. There seem to be many intermediate forms which blur the distinction between the broad categories. Among the various types of nucleated villages some seem more dispersed than others, notably the polyfocal village which has a number of nuclei only loosely connected together. Similarly among the hamlets some seem more village-like than others, notably the various 'rows' arranged along a road or on the edge of a green which, with a little expansion or infilling could be turned into villages (Wrathmell 1994). The rows in particular often have a regular layout, making them resemble even more closely both towns and larger villages. Detailed survey work, and documentary study, throws up indications even in districts where the nucleated village seems to have reigned supreme of isolated hamlets or farms, sometimes on the edges of village territories (Taylor 1995). Some may be survivals of earlier dispersed patterns, others may be new foundations of the twelfth or thirteenth centuries. Perhaps further research into these intermediate forms will help to resolve the still open questions about the local divergences between nucleated and non-nucleated settlements. There is general agreement that the period of village formation (c.9001200) coincided with that of the origins of 'midland' field systems, the early phases of urbanisation, the proliferation of small parishes, and changes in the structure of estates and forms oflordship. Thereafter the consensus breaks down. Chronology remains a problem, with some researchers emphasising the late ninth century or even earlier, whereas others point to the many village excavations which have produced no dating evidence before the twelfth century. Discussion of the role of estates and lordship has been given greater precision with the suggestion that peasant settlers were gathered on the 'inland' attached to the manor houses of pre-Conquest estates (Faith 1992). The association between the fragmentation of the primeval estates and village formation has been made doubtful by the growing body of evidence that smaller territorial units existed within the great estates, and that some large estates were being put together at the time that they were supposedly breaking down (Hadley 1996). The idea that villages were born out of the need to reorganise farming systems at a time of expansion and marketing opportunities remains attractive. Villages are closely associated with regular field systems which seem to have emerged by the tenth century, and there is a logical connection between compact groupings of habitations and the efficient exploitation of open fields (Fox 1992). However, the scientific evidence does not seem to show any great expansion in cultivation or signs of environmental stress at this period (Murphy 1994; Bourdillon 1994). Explanatory models in this field as in others are moving away from heavy determinism towards an 57 C. Dyer emphasis on mentality and choice. The catastrophe theory, that villages were formed at a moment of crisis when earlier methods of farming collapsed, or when new lordship or state power imposed themselves on a previously loosely organised countryside, is being replaced by the assumption that changes happened gradually through some evolutionary process. The social history of the period could be seen as one in which ties of kinship declined in importance, and when the protective role of the community, encouraged by the state, played a more important part in people's lives. The village, especially in its more regularly planned form, was the material embodiment of community, and expressed an idea about how people should live and organise themselves. Villages developed alongside one another through imitation of models, and the form of settlement helped to define, along with many other features of daily life, the various regional cultures. For example, the great regularity ofDorset settlements resulted in a uniformity of the tofts in each village, and from one village to another (Taylor 1994). Discussion about village origins was advanced by a well-organised but brief conference session at Leicester in 1992 (Fox 1992), but we have lacked debates on this issue- more the statement of different positions, which sometimes ignore the existence of contrary opinions. Perhaps more direct exchanges of view will develop in the future. 3 Life of settlements, and their development and decline Although so much thought and speculation has been devoted to the problems of settlement origins, the great bulk of our evidence relates to the functioning and development of settlements, and indeed some of the most innovative work has been done in this field. Changes in the shape and size of villages was one of the themes of settlement research in the 1970s, but this field of enquiry has been advanced with the application of techniques of plan analysis. Both deserted and surviving villages can be shown to have experienced replanning, the addition of planned or irregular extensions, the provision of greens or market places, or the infilling of such open spaces, as well as more conventional shrinkage or migration (Everson, Taylor & Dunn 1991 ). A specific problem has been drawing the frontier between villages and small towns, and this was the subject of a conference in which urban and rural specialists exchanged views, and established a remarkable amount of agreement (MSRG 1993, 7-14 ). In the pre-Conquest period plan58 ned towns and nucleated villages were emerging at about the same time- was this a simultaneous development, or did villages imitate urban forms? In the twelfth and thirteenth centuries both villages and towns were planned and replanned, and sometimes the distinction between a market village and a market town with an agrarian dimension in its economy is difficult to draw. Within the settlements, although relatively few peasant houses have been excavated recently, work continues on their interpretation. Their origin in the Germanic or Romano-British building traditions has formed a dimension in the debate on the size and character of the Anglo-Saxon migrations (Hamerow 1993). In the late medieval period there have been useful dialogues between excavating archaeologists and architectural historians, which began with Wrathmell 's reinterpretation of the Wharram Percy peasant houses. The discussion should continue with more certainty now that precise dates of standing buildings have been established, and they can be seen to be securely medieval, and indeed in one pocket of Berkshire/Oxfordshire small buildings in the cruck tradition can be taken back to the decades around 1300 (Vernacular Architecture 1990, 47; 1992, 58) Standing buildings rather than excavated structures have been used to investigate the use of domestic space, and to interpret the buildings and changes in their form in relation to changing perceptions of privacy and social distinctions within the household (Johnson 1993). More work needs to be done of this type using published house plans and records of the distribution of artefacts. One insight into the social use of dwellings has come from the study of temporarily occupied shielings, which with the help of historical accounts of more recent transhumance allows us to visualise the movement of young people out of the household to create a distinctively transient society on the summer pastures (Herring 1996). Regrettably historians of peasant society, who have access to documentary evidence for families and households which can sometimes reveal such residential arrangements as the provision of cottages for the children who were unable to inherit the main holding, do not relate their findings to the material evidence (Razi 1993 ). Environmental studies, for long focused on urban sites, are beginning to make a real contribution to rural settlement studies. The preliminary conclusions are now available on the animal bones from West Cotton (Northamptonshire), a hamlet site occupied mainly in the thirteenth and fourteenth centuries (Albarella & Davis 1994). They shed light on rubbish disposal, which contrasts with the pits or collective tips used in towns, as food bones seem to have been Recent developments and future prospects in research into English medieval rural settlements left in the open to be gnawed by an apparently sizeable population of dogs. Comparison between bone assemblages from late medieval rural settlements and those from high status sites such as castles, and from towns, reveal some of the characteristics of the peasant diet, which contained relatively little game, a relatively low percentage of pork, and a high proportion of mutton (Albarella & Davis 1996). There is a growing body of information about animal husbandry and breeding, from the responses to the market for animal products as early as the eighth century, to the development of larger animals at the end of the medieval period. Botanical studies from the early middle ages continue to add to the data on the extent of cultivated land, which points to considerable regional variations, but which supports in general the idea that the migration period in some parts of the country involved no great upheaval in the agricultural economy (Rackham 1994; Fenton-Thomas 1992). New information about the types of cereals grown, such as the spread of free-threshing varieties of wheat, has implications for farming systems which we have yet to take fully into account (Campbell1994) Some ofbest botanical evidence for the later middle ages has come from excavations of moated sites, and there is a wealth of data from sites such as Wood Hall (North Yorkshire ) relating to the local environment and the specific activities carried out on the site (MSRG 1995, 36-38). Settlement archaeologists, like those involved in the Wood Hall project, are aware of the rieed to make the fullest use of environmental data in their interpretations, rather than regarding the specialist reports on bones and pollen as mere appendages to the main task of recording structures and contexts. We expect to gain much more from environmental studies in the future, both through the accumulation of new data, and from developing a dialogue between scholars primarily concerned with settlements and those with expertise in analysing botanical and bone evidence. A major trend in historical thinking in recent years has been to highlight the effects of commercialisation, on society as a whole, but in particular on the peasantry (Britnell 1993). In the early middle ages agricultural systems were devised in order to give a surplus of animal products, presumably in response to market demand, and this could have been an ingredient in the emergence of divergent forms of fields and settlements in the tenth to twelfth centuries. The growth of more specialised regional farming systems helps to explain the extension of settlement on pastoral uplands or wetlands, especially in the thirteenth century. The involvement of the rural population in industries made settlements viable in places with relatively poor agricultural land, such as the Kentish weald or the forests of Staffordshire. Thanks to new work on the hinterland of London, and on the regional links of provincial towns such as Exeter, we can appreciate the involvement of the inhabitants of rural settlements, such as those on the Devon uplands and West Cotton (Northamptonshire) in the supply of urban markets (Campbell, Galloway, Keene & Murphy 1993; Kowaleski 1995). The commercial influences on production have left relatively little trace in the material culture of rural sites. It is rather in the evidence for consumption that the archaeological evidence for commercialisation is most striking. Excavators have always appreciated that pottery and small finds originated outside the village, and the widespread distribution of Ipswich ware and Thetford-type ware in East Anglia shows that rural consumers were obtaining the products of urban industries in the seventh to tenth centuries. In the later middle ages we can begin to reconstruct the mechanics by which peasants travelled to small town markets to acquire commodities supplied from more remote points of manufacture. Now we appreciate that a wider range of goods than was previously supposed were acquired by exchange -not just the sea fish of which bones are found on inland rural sites, but also some of the meat represented by finds of animal bones. Building materials, especially timber, are likely to have been bought in many cases, and craft specialists such as carpenters would often have been hired from outside the village. The decline of rural settlements in the later middle ages can be interpreted also in the context of market relations. This was not a matter just of reduced demand for grain and the increasing market for animal products, but also shifts in the land market, which led to the engrossing of holdings and amalgamation of tofts. The frontiers between private and public space were being renegotiated as village communities and their common fields were converted into farms (Dyer 1997). The causes of the abandonment of rural settlements at the end of the middle ages has been discussed for many years, but that does not mean that the problem has been resolved. The shift away from deterministic explanations ofvillage creation applies to the study of desertion as well. The new emphasis on dispersed settlements requires more work on their varied history of desertion, shrinkage and survival. The complexity of their history at the end of the middle ages is suggested by growing evidence that new settlements were being created even when the population of the whole country was shrinking or at least stagnant (e.g. Fox 1996). 59 C. Dyer 4 Research methods Bibliography As the age of large scale research excavations seems to recede, we need to reassert their value. Large scale excavations in the last decade have not indicated any diminishing of returns: for example West Heslerton (North Yorkshire) has shown that large village type settlements can be found before the ninth century; W estbury/Tattenhoe (Buckinghamshire) revealed the growth and decline of midland hamlets; a market village with distinctive features was excavated at Dassett Southend (Warwickshire); and West Cotton, as well as producing important environmental evidence, threw new light on the relationship between manor houses and peasant settlements. There is still scope also for thematic studies concentrating on types of site over a large area, like the search for earthwork remains of rural settlements in East Anglia or Cornwall, or identifying agricultural buildings, such as the late medieval sheepcotes of the Cotswold Hills which are also being found in other parts of the country (Dyer 1995), or the documentary research into the distribution of places with distinctive names like the places called 'hide', suggesting the survival of small units of land and occupation even within village dominated regions (MSRG 1995, 19). The future development of the subject will be well served by large scale survey projects, such as those designed to identify dispersed settlements advocated above, and those concentrated on specific regions like the fenland survey (Silvester 1993) and the valuable work on the reclaimed wetlands of the Severn estuary and Somerset which has revealed the remarkable achievements of Roman drainage schemes, their decay in the post Roman period, and the early medieval programme of drainage and recolonisation (MSRG 1994, 18; Rippon 1994). The most fruitful results are still most likely to come from the intensive study of carefully selected slices of landscape - large parishes or groups of parishes, which coincide with medieval administrative and agricultural units. These should ideally be subjected to multi-disciplinary research involving teams of specialists, not just in order to recover the maximum amount of all kinds of evidence, but also to interrogate that material from every possible perspective, and to bring to it insights deriving from a wide range of subjects. In that way a lively area of enquiry, concerned with a central aspect of the medieval past, will continue to inform and stimulate us into the next millennium. ALBARELLA U. & DAVIS S.J.M. 1994: The Saxon and medieval animal bones excavated 1985-1989 from West Cotton, Northamptonshire, English Heritage, Ancient Monuments Laboratory Report, London. ALBARELLA U. & DAVIS S.J.M. 1996: Mammals and birds from Launceston Castle, Cornwall: decline in status and the rise of agriculture, Circaea 12 , 1-156. ASTON M. & LEWIS C. (ed)1994: The medieval landscape ofWessex, Oxbow monograph 46, Oxford. BoURDILLON J. 1994: The animal provisioning of Southampton, in: Rackham (ed) 1994. BRITNELL R.H. 1993: The commercialisation of English society 1000-1500,Cambridge. CAMPBELL B., GALLOWAY J., KEENED. & MURPHY M. 1993: A medieval capital and its grain supply, Hist. Geog. Research Ser. 30. CAMPBELL G. 1994: The preliminary archaeobotanical results from Anglo-Saxon West Cotton and Raunds, in: Rackham (ed) 1994, 65-82. DA VISON A. 1990: The evolution of settlement in three parishes in south-east Norfolk~ East Anglian Archaeol. 49. DYER C. 1995: Sheepcotes: evidence for medieval sheep farming, Medieval Archaeology 39, 136-164. DYER C. 1996: Rural settlements in medieval Warwickshire, Trans. Birmingham and Warwicks. Archaeol. Soc. 100, 117-132. DYER C. 1997: Peasants and farmers: rural settlements and landscapes in an age of transition, in: D. Gaimster & P. Stamper (eds), Age of transition, Oxford. EVERSON P.L., TA YLOR C. C. & DUNN C.J. 1991: Change and continuity: rural settlements in north-west Lincolnshire, Royal Comm. on Hist. Monuments, London. FAITH R. 1992: Estates, demesnes and the village, in: Fox (ed.) 1992. FENTON-THOMAS C. 1992: Pollen analysis as an aid to the reconstruction of patterns of land use and settlement in the Tyne-Tees region during the first millenia BC and AD, Durham Archaeol. Journal 8, 51-62. Fox H. 1992: The agrarian context, in: Fox (ed) 1992,36-72. Fox H. (ed) 1992: The origins of the midland village, Leicester. Fox H. 1996: Cellar settlements along the south Devon coastline, in Fox (ed) 1996. Fox H. 1996: Seasonal settlements, Vaughan papers, Leicester. HADLEY D.M. 1996: Multiple estates and the origin 60 Recent developments and future prospects in research into English medieval rural settlements of the manorial structure of the northern Danelaw, Journal Hist. Geog. 22,3-15. HAMEROW H. 1994: Migration theory and the migration period, in: Vyner (ed) 1994, 164-177. HERRING P. 1996: Transhumance in medieval Cornwall, in: Fox (ed) 1996. JOHNSON M. 1993: Housing culture. Traditional architecture in an English landscape, London. JOHNSON N. & ROSE P. 1994: Bodmin Moor: an archaeological survey. Vol 1 The human landscape to c 1800, English Heritage and the Royal Comm. on Hist. Monuments, London. KOWALESKI M. 1995: Local markets and regional trade in medieval Exeter, Cambridge. LEWIS C., MITCHELL-FOX P. & DYER C. 1997: Village, hamlet and field. Changing rural settlements in central England, Manchester. MSRG, Medieval Settlement Research Group Annual Reports. MURPHY P. 1994: The Anglo-Saxon landscape and rural economy: some results from sites in East Anglia and Essex, in: Rackham (ed) 1994. RACKHAM J. (ed) 1994: Environment and economy in Anglo-Saxon England, Council for British Archaeol Res. Report 89. RAzr Z. 1993: The myth of the immutable English family, Past and Present 140, 3-44. RIPPON S. 1994: Medieval wetland reclamation in Somerset, in: Aston & Lewis (ed) 1994,239-253. SILVESTER R. 1993: The Fen land project in retrospect, East Anglian Archaeol. 50, 24-39. TA YLOR C.C. 1994: The regular village plan: Dorset revisited and revised, in: Aston & Lewis (ed) 1994,213-218. TA YLOR C. C. 1995: Dispersed settlements in nucleated areas, Landscape Hist. 17, 27-33. VYNER B. (ed) 1994: Building on the past, Royal Archaeol. Institute, London. WRATHMELL S. 1994: Rural settlement in medieval England: perspectives and perceptions, in: Vyner (ed) 1994, 178-194. Christopher Dyer The University of Birmingham School ofHistory Edgbaston Birmingham B 15 2TT UK 61 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Jlllark Gardiner Trade, Rural Industry and the Origins of Villages: some Evidence from South-East England The origin of English villages has exercised historians and archaeologists for at least one hundred years. Early studies of the subject were much influenced by the German historian, August Meitzen who argued that the villages were found by Anglo-Saxon incomers and replaced the earlier dispersed pattern of Celtic and Roman settlement. He claimed that hamlets and isolated farms persisted in those areas reached late or never occupied by the Anglo-Saxons. This model of settlement development persisted for a remarkably long time, mainly because the means for its examination were not available. It was not until the growth of medieval archaeology during the late 1950s and 1960s that the view that villages were founded within a few generations of the Adventus Saxonum became less and less credible. Deserted medieval villages, which were being examined in increasing numb\!rS, rarely showed any evidence for underlying 6th-, 7th- or 8th-ce~tury occupation. Many villages indeed showed little signs of activity before the 12th century. At the same time Anglo-Saxon settlements were being discovered on sites which had not been occupied by later medieval villages. By the late 1970s a new orthodoxy was arising, and was summarised in Christopher Taylor's study, Village and Farmstead (Taylor 1983). He took a long perspective, placing the growth of villages within the context of the changing patterns of settlement since the end of the last Ice Age. He argued that villages were not age-old settlements, but the results of particular economic and social circumstances which led to their formation in large numbers, mainly in the period 1100 to 1300. The periods before and after were marked by greater numbers of dispersed settlements. That realisation opened new perspectives on the study of the medieval village. Firstly, it allowed villages to be seen as one of a range of rural settlement types, which included on the one hand hamlets and isolated farmsteads and, at the other extreme, settlements which were little different from small towns. Villages were no longer recognised as the 'standard' type of rural settlement, and other forms as lesser variants. Secondly, it raised questions about what were the particular social and economic conditions which produced village settlements in the period after the Norman Conquest. These were issues which archaeology alone could not answer and it encouraged a new approach which integrated settlement studies into a wider analysis of medieval rural society. In the most recent work two aspects of villages have been examined. Who was responsible for their foundation and planning? And can the variation in the distribution and plan of villages be explained in terms of regional differences? At an early stage in the examination of English villages their plans had revealed a degree of regularity which could not be explained through the process of gradual development. Many villages had quite clearly been laid out as discrete settlements from the beginning. With the re-evaluation of the date of the formation of villages, it became difficult to determine a context in which the settlement pattern in a parish or vill might have been totally re-organised. It was no longer possible to imagine that the village had been planned by a successful Anglo-Saxon warrior for his followers. In the north of England, where planned villages have been most extensively studied, it was suggested that the new settlements were established after the Harrying of the North, the systematic devastation of the area in the 11th century by the Norman army. The dislocation created by this event would have allowed villages to have been created on new sites according to systematic plans. That context now seems less likely since planned villages have also been found in southern England where there was no comparable devastation. Even in the north, the effects of the Norman armies are now believed to have been less than was once thought. The underlying assumption, that planned works must necessarily be the results of seigneurial action, has not passed without comment. Were decisions necessarily made by the lord of the manor or his officials? There are many examples of the community of the vill acting independently to reorganise the pattern of fields and to maintain bridges. The creation and planning of a settlement was therefore likely to 63 M. Gardiner have been well within the capacity and authority of such a body (Dyer 1985; Harvey 1989). The distribution of planned settlements is only one aspect in the regional variation in settlement form which has long been noted. One hundred years ago the historian, Frederic Maitland (1897) published two contrasting maps to illustrate the differences in his study of Domesday Book. More systematic work on the subject has been undertaken over a long period by Brian Roberts (e.g. Roberts 1987; Roberts & Wrathmell 1995) who has sought to classify the settlement forms and identify regional types. Roberts' work, which is primarily descriptive, now offers a tool which has yet to be employed to interpret the variation in settlement form. However, a recent study of settlement in four counties in the East Midlands by scholars at Birmingham has sought to integrate the study of settlement form into a wider understanding of medieval society (Lewis, Mitchell-Fox & Dyer 1996). A fundamental problem faced by historians studying village origins is that the process of settlement nucleation happened so early that it is rarely recorded in documentary sources. Detailed local records survive in large numbers from the mid-13th century, but by then most English villages had already been established. It is useful to turn to those few areas of England where village formation occurred at a very late date to gain an insight into settlement change. The process in these places may be recorded in documentary sources, although it cannot be claimed that they are representative of village formation more generally. The fact that villages developed late of itself means that they are likely to be atypical. Nevertheless, these areas may provide some insights into the nature of settlement nucleation, a process which is otherwise poorly understood. Settlement development in yhe Sussex High Weald The Wealden district of south-east England is an area of poor soils and relatively late settlement. It still remains one of the most heavily wooded areas in the country, because the soils are generally unsuitable for arable agriculture and in some parts too barren even for good pasture. The Weald can be sub-divided into the High Weald which has the poorest soils and an outer fringe of clay lands known as the Low Weald. Historically, the most wealthy areas of the SouthEast lay towards coastal fringe. The major towns were either situated on the larger rivers or had developed around ports which served both coastal trade and trade with Continental Europe. Most of these towns 64 had already developed by the end of the 11th century. There were many fewer towns inland, though during the first half of the 13th century new urban centres grew up at, for example, East Grinstead and Hailsham. In the early 13th century the pattern of settlement in the High W eald beyond these towns was dispersed. Isolated farmsteads were surrounded by their own fields and there were extensive tracts of common land (Brandon 1969). Permissive settlements Village development took place in the Weald mainly during the second half of the 13th century. Trade in the interior of the Weald was the stimulus for settlement growth and villages grew up particularly at the sites of markets. The sequence here was the reverse of that commonly found in medieval England: trading places developed at unoccupied sites and subsequently attracted permanent settlement. The village of Mayfield exemplifies that pattern. Work probably began on the construction of the archbishop of Canterbury's palace at Mayfield shortly after 1260. The palace was certainly built before 1279, and was further enlarged in c.131 0 (Davis et al. 1969, 10). It cannot be coincidental that a market charter was granted in 1261. The market may not have flourished immediately for no rents from stalls are recorded in a survey of c.l285 survey (Calendar of Charter Rolls 2, 38; Redwood & Wilson 1958,4760). The earliest evidence for a settlement on the site comes from a list of 1389. A fire that year swept through the village, and damaged the church and buildings on 85 'shop-places'. The list shows that most tenants held more than one shop-place and it seems probable that the small plots, originally laid out for stalls, had been grouped together to create larger areas for more substantial buildings. In the mid-15th century cottages occupied two shop-places and bigger houses stood on the former sites of three or more stalls (Cooper 1869, 16-7; LPL ED 2068, 1900, 1080; ESRO AMS 5512, ff. 5v., Sr., lOr., etc.). The site, though adequate for the original stalls set out in front of the church and along the street, was inconvenient for a village (fig. 2). The palace and church on the north side of the High Street limited growth; the buildings in front of the churchyard still occupy cramped sites. In the 15th century barns and gardens belonging to those houses lay on the opposite side ofthe road in 'forelands' on the south side ofthe street. The land here falls sharply to the south, restricting the scope for development and as result the plan of the village was linear. stretching along the edge of the high ground. Trade, Rural Industry and the Origins of Villages: some Evidence from South-East England N 0 [t] • Towns less than 50 50 • 99 100. 149 e Villages • Other places e With a market 150 and above 0 Without a market 0 The allocation of the Mayfield rents to 'shopplaces' clearly indicates the village developed from an unoccupied site to become a centre of trade and subsequently grew into a settlement. Wadhurst, which lay in the same manor as Mayfield, had a similar pattern of development, though it did not reach the size of its neighbour. Wadhurst was granted a market charter in 1253 and stallage for a market and fair was recorded in the c.1285 rental, though there is no evidence for a nucleated settlement at that date (Calendar of Charter Rolls 1, 432; Redwood & Wilson 1958, 30-47). A triangular market-place in front of the church is clearly identifiable on early 19thcentury maps, though by then it was largely infilled by permanent buildings (fig. 3). A timber-framed market hall, which survived until it was demolished in the mid-19th century, also stood within the market-place, suggesting a settlement of some pretension. A survey of 1498 records that some of the 'shop-places' within the market-place were occupied by houses; other dwellings lay immediately in front of the churchyard. Here too the stalls, which had occupied these plots, had been superseded by houses. The largest building plots lay on the land to the south of the market-place. These do not seem to have been planned, regular plots, but had expanded from street frontage backwards and into the existing pattern of km 10 fields (LPL ED 1900; ESRO AMS 5512, ff. 34r.53r.). The development ofTicehurst village a few miles to the east along the same road was very similar. Early 17th-century maps show a market-place partially infilled with permanent buildings and a market hall. There were further houses on the west side of the market-place and, as at Wadhurst, a number of cottages were squeezed in on the north and west sides of the churchyard (ESRO SAS/COlD 2; ESRO TD/E 2; ESRO ACC 4683; fig. 4). A rental ofHammerden of c.l295 mentions a piece of land in the marketplace measuring 52 by 13 feet (15.8 by 4 m), a forge and two other houses. Charters of about the same date granting land to Hastings Priory mentioned two further houses close to the church. A century later Ticehurst was described as a villa and there were anumber of houses and cottages there which by 1499 had increased to 11 cottages, one shop and some other plots of land (HMAG Priory Ch. 3-6, 13; ESRO SAS/ CO/B 71, 72). Not all such trading places were so successful. Wartling was granted a weekly market at a comparatively late date, 13 37, and the first reference to shops occurs in the court rolls of that year. An account of 1340-1 records rents of 3s. and unpaid rents of 14d. and 15d from two other stalls (Calendar 65 M. Gardiner of Charter Rolls 4, 431; BL Add. Rolls 32722, 32601 ). Later 14th-century court rolls show that the shops, a forge and a stable stood around a marketplace close to the church, though there is no evidence of any houses. The shops were little more than booths and did not provide accommodation. The largest measured 14 by 10 feet (4.2 m by 3 m). Shops of 7 feet by 5 feet (2.1 m by 1.5 m) and 6 feet by 4 feet ( 1.8 m by 1.2 m) are also mentioned. The shops were permanent structures for a thief gained access to one through the window of the door of the building. Seven shops are listed in the rental of 1366. From the 1370s onwards the demand was strong and new sites were let for shops and the lord was able to charge increased rents (BL Add. Rolls 32641, 32642, m. 1r., 32692, m. 13v., 32695, 31508, 32693, 32712, m. 2r.; ESRO SHE 7/16; BL Add. Rolls 32691, 32697, 32700). Mayfield, Wadhurst, Ticehurst and Wartling developed initially as places of trade around a market-place. The first three were on the main ridgetop routes through the W eald; Wartling was on a road leading northwards across the marshes to the upland. The village of Heathfield by contrast had no regular pattern of shop-places, acquired a market charter comparatively late and was apparently not situated on any major roads. In spite of these disadvantages, Heathfield developed as a place of both manufacture and trade. The first extant rental attributed to 1253X62 mentions four shops, one of which lay next to the churchyard. Two of the shops were constructed on former farmland. The rental evidently lists the tenants in a broadly topographical order, for others in the same part of the document, Adam de Ecclesia and Robert Vicar, bear names suggesting their proximity to the church. Other tenants mentioned in that part of the rental include Geoffrey Chaloner, Geoffrey Weaver (Textor) who had one house, Nicholas Weaver with one house and 1Y:z acres (0.6 ha) and Godfrey Mercer who may have lived nearby. It seems probable at this date that most of these surnames were not hereditary, and can be taken to indicate livelihood (McKinley 1988, 56-8, 60-1 ). Heathfield, therefore, was probably a settlement of textile workers and merchants. The village was not granted a market charter until 1316, when there was an already well-established trading centre (Calendar of Charter Rolls 3, 306; Peckham 1925, 97; WSRO Ep. Vl/1/3, f. 47r.). None of the villages considered above may be described as planned settlements in the sense the term is generally used. Seigneurial planning, insofar as it occurred at all, was limited to regularising the plots used for the market stalls and to purchasing market charters. Whether these actions initiated the markets or merely served to encourage the growth of already 66 developing commercial activity is considered further below. It was evidently not the intention of the lords of Mayfield, Wadhurst, Ticehurst, Wartling and Heathfield to found settlements beside their churches. They did not lay out regular plots of a size suitable for houses. The settlements which subsequently emerged were fitted in the restricted spaces available around or at the site of the market. These villages, excepting Wartling where no permanent occupation was established, may be described as permissive settlements. Their development from market sites was evidently not hindered, even if it was not actively facilitated their lords. The attitude of lords to the development of permissive settlements is well illustrated by the village of Rotherfield. It was situated near on the top of the hill around the crossing point of three routes in the centre of an area of demesne. The tenements 'on the hill' as they are described in a survey of 1346/7 then comprised 34 plots of land with buildings. The lord ofRotherfield made small areas ofland available for the development of the village, but very few of the plots were more than a rood or two (0.1-0.2 ha) in extent. A plan of the manor of 1597 suggests that buildings in the village lay in three areas: to the north of the church on the opposite side of the road, to its east around the road junction and to its west. The tenements to the north of the church and those on the east side of the junction appear to have been cut out ofthe demesne (Brandon 1969, 149; ESRO SAS/LB 34; ESRO GIL 32; fig. 5). The earliest surviving court rolls of the 1450s mention one rood (0.1 ha) of leased demesne land near the junction which, according to a later annotation, was occupied by house, barn and orchard. Other buildings in the village were constructed on land named Courtmede, on the former rabbit warren and on land taken from the highway. Shops were situated close to the churchyard according to a rental of c.1500 (ESRO SAS/Aber 84; ESRO SAS/Aber 68). Planned Settlements The planned settlements in the Weald, unlike the permissive villages, were intended from the outset to provide housing as well as a site of trade. Their tenements were larger and the villages occupied less restricted locations. Trade was foremost in the minds of the founders of planned settlements, as the early grants of market charters or the urban attributes suggest. For example, the plan ofBurwash clearly points to its planned character. The village has a linear plan and stands astride a major ridge-top route through the Weald (fig. 6). A series of tenements ran back from Trade, Rural Industry and the Origins of Villages: some Evidence from South-East England 1582 1610 3•430 N t 0 Fig. 4. 0 I00 metres I 00 yards 67 M. Gardiner the street frontages to a rear boundaries parallel with the road. A market charter was granted in 1252 when the manor of Burwash was in the king's hands and there was certainly a village here in 13 29, when seven houses in the villa ofBurwash were granted to the manor of Woodknoll. There were at least eight stalls in the market-place in 1423 (Calendar of Close Rolls 1251-3, 82; Peckham 1946, 354; BL Lansdowne MS. 269, f. 90v.; ESRO ASH 198). Burwash was a demesne manor of the lord of the Rape of Hastings and by the early 13th century was the centre of its administration. The size of the planned settlement suggests that it was intended to be a small town, but, if so, it failed to achieve an appropriate level of prosperity. Planned settlements are only found within the manors of the larger lords, and in particular those belonging to monasteries. Boreham Street was established by Bayham Abbey and Sedlescombe Street by Battle Abbey, while Robertsbridge Abbey founded a village at Robertsbridge itself. Boreham Street was probably the place indicated as a settlement on the road between Lewes and Winchelsea on the 'Gough' Map attributed to c.1360 (Parsons 1958,2, 16, n. 1). The origins of the village may date to 1251 when the abbey was granted a market and Midsummer fair, however, that right had not been exercised when it confirmed in 13 25 (Calendar of Charter Rolls 3, 479; Calendar of Patent Rolls 1324-27, 176). A probable starting date of the settlement is suggested by a series of charters from the 1270s recording grants of land between half acre and two acres (0.2 to 0.8 ha) in extent to hold 'in free burgage' (BL Cotton MS Otho Aii, nos 378-81). If the grants in free burgage signified the ambition of the abbot, then he was to be disappointed for there is no evidence that the settlement was ever more than a few houses with a chapel. A late 13th-century rental lists tenants with the surnames Baker, Weaver (two) and Skinner, which may indicate occupations (Ray 1931, 174; BL Cotton MS Otho Aii, ff. 73v., 77v.). The east side of the small village Sedlescombe Street lay in the lordship of the abbot of Battle, the west in the manor of Bricklehurst. The settlement was situated on the road to Hastings just north of the bridge over the River Brede. In the mid-16th century the whole settlement comprised about 18 houses (Martin & Martin 1979, 71). It may have been a similar size in 1433 when a survey of the Battle portion alone recorded 10 messuages and one garden with a further building. On the east side the holdings were a regular size, each measuring one acre or a half, and paid rent at the rate of 3s. per acre (0.4 ha). There were a similar number of tenements in c.131 0 and the rents of 1s. 6d. and 3s. mentioned in the court rolls in 68 the 1290s take the history of the settlement back to the late 13th century. No settlement, however, is mentioned in charters of the 1240s recording the acquisition of land by Battle Abbey. It seems that the Battle part of the settlement had been formed during the second half of the 13th century out of part of the demesne field called Blakelond. The plan of the west side of Sedlescombe Street in the manor of Brick1ehurst is less well documented, but its less regular plan suggests it may have developed organically, taking advantage of the activity on the opposite side of the road (PRO E315/57, ff. 14r.-14v.; PRO E315/56, ff. 5v.-8r.; LIL Hale MS. 87, ff. 56v.-58v.; HEH BA vol. 5, nos. 162, 163, 174; ESRO FRE 520/ii). The development of the villages ofRobertsbridge and Salehurst is more complex than those already considered. It is improbable that there was a significant settlement when the Cistercian monks chose the site ofRobertsbridge for their abbey in c. 1180 for they preferred remote locations for their religious houses. The village is more likely to have developed after the monastery was relocated to Elham in the Rother valley to the east (D'Elboux 1944, 7, 124). A grant of a market and fair was made in 1225 during Henry Ill's stay at the monastery, but was cancelled the following month, because it was said to have threatened existing markets in the neighbourhood. It cannot be a coincidence that Salehurst and Robertsbridge were granted, or in the case of the latter, regranted, market charters in the same year, 1253, for the two grew up as rival settlements on the opposite banks and two separate crossing points over the Rather valley (Rotuli Litterarum Clausarum, 2, 14, 16; Placita de Quo Warrante, 759; Calendar of Patent Rolls 1461-7, 408; Calendar of Charter Rolls 1, 416; cf. Saul 1986, 164-5). The abbot ofRobertsbridge had high expectations of his settlement and evidently intended to establish a borough. He withdrew his tenants from the hundred of Henhurst, formed his own hundred and created the posts of constable, ale-conner and street-driver (Rotuli Hundredorum 2, 217-8). The names of tenants recorded in a rental of the 13th or early years of the 14th century included Waiter Textor (Weaver), Henry and Robert Fuller and Henry Tinctor (Dyer) who may have been textile workers. Waiter le Chaper (the Trader), the heirs of Peter Faber (Smith) and Gilbert le Tomur (Turner) are also mentioned (ESRO SHE 7I 3; CKS U1475/M242). Thus by c.1300 Robertsbridge seems to have been a flourishing manufacturing and trading centre. The development and subsequent decline of the rival market of Salehurst on the north side of the Rother valley is more difficult to trace. In 1349 its lord, James de Etchingham claimed that the construe- Trade, Rural Industry and the Origins of Villages: some Evidence from South-East England 2262 !•710 N 293 ·ooo t Rotherfield 2295 2300 8'178 1'604 Fig. 5. 0 0 ( . 100 yards 202 2·613 \ I 00 metres 6·730 562 587 Fig. 6. tion of a dam across the Rother downstream at Knellesjlete had restricted the movement of ships and caused the destruction of his 'market town' of Salehurst (Calendar of Patent Rolls 1348-50, 78). The assertion was perhaps rhetorical, for the village may already have suffered from competition from Robertsbridge. Whatever the reasons, Salehurst declined into a small settlement, while Robertsbridge flourished. Some Minor Settlements The settlement at Bodiam was granted a charter for a market and fair in 13 83 but failed to become a flourishing centre (Calendar of Charter Rolls 5, 281 ). It was situated on a road leading to a bridge over the Rother, a crossing point which had been used since the Roman period. Four or five long, regularly spaced tenement boundaries may still be traced in the grass behind the 19th-century houses to the west of the castle and the pattern is also shown on a map of 1672 69 M. Gardiner (ESRO AMS 569113/1; ESRO BAT 4435 (26)). The settlement is almost entirely surrounded by demesne and was very probably a planned settlement laid out by the lord of Bodiam manor. A likely context for this is the mid-1380s when Sir Edward Dallingridge obtained a licence to crenellate, began work on Bodiam Castle, established a nearby water-mill and obtained his market charter. Dallingridge appears to have invested his profits from the French wars in a model settlement with castle, village, mill and market, but his ambitions were modest with regard to the village: only a small number of tenements were laid out (Whittick 1993, 122; Taylor, Everson & Wilson-North 1990, fig. 4). A final example of a lateestablished settlement is Dallington which did not obtain a market charter. The first evidence for buildings there is in 13 83 when a grant was made to construct a stall on a piece of land measuring eight feet by 10 feet (2.4 by 3.0 m) next to the churchyard. A shop place is mentioned there 50 years later. The site evidently failed to develop beyond a few stalls (ESRO SHE 7116; BL Add. Roll 32691, m. 5v.; BL Add. Roll. 32716, m. 4v.; ESRO AMS 6270/90/6). The Pattern of Village Development in the Sussex Weald Settlements have been divided into two categories, permissive and planned. Seigneurial action played some part in the growth of almost every settlement, whether passively in permitting encroachment upon the highway or the erection of market stalls, or more actively in obtaining a market charter or leasing areas of demesne. However, relatively few of the settlements in this area, compared to those in other parts of England, can be identified as planned. The intention of most lay lords in the Weald was more modest, to foster markets within their manors and so to increase their revenue. The consequence was that when those markets grew into permanent settlements, there was often limited scope for expansion. The planned settlements of Robertsbridge, Sedlescombe Street and Bodiam occupied more spacious sites than, for example, the villages ofMayfield and Rotherfield where land had to be obtained by encroachment upon the highways or through grants of demesne. The relationship of trading settlements to the road system is likely to have been of considerable importance, although it is difficult to identify the main routes in the Weald, as elsewhere in medieval England. The physical geography of the region dictates that the ridge-top routes were likely to have been major lines of communication, and the most impor- 70 tant of these were the ridges running from Heathfield to Battle and Hastings, and the second from Heathfield through Burwash towards Etchingham and beyond. The 'Gough' Map marks the line of a third road from Hailsham through Boreham Street to Battle and Winchelsea (Pelham 1931, 182-4). The settlements located away from the major highways were seriously disadvantaged. It would have been difficult for Dallington to the south of the ridge-top road towards Battle, and Brightling (see below) to the north, to develop into important centres of trade. Some villages succeeded in spite of their location. Heathfield lay some distance from any significant highway, but was well situated for local textile working, craft of considerable importance as the number offulling mills in the vicinity attests (Gardiner 1996, 134). A location on higher ground above the crossing-point of a stream or river was commonly chosen for the site of settlements. The larger centres of Robertsbridge, Northbridge Street and Sedlescombe Street have already been discussed to which may be added Newenden just over the border in Kent (Graham 1952, 77). Some hamlets in Sussex lay in similar positions. Bodiam and Boreham Street were both located near crossings. The hamlet or small village of Saltcote Street (Playden) lay either side of the road to the ferry over the channel of the River Rother and a chapel stood just beyond the west end of the settlement (Johnston 1967). The free tenements paid a rent of 4d or a multiple of that sum. The place-name suggests that salt extraction was practised and anumber of fishing boats were based at the hamlet (Dulley 1969,42, 56; Sussex Topographical Surveys: Playden Parish (1993); CKS U47/42/M12). Fishing may also have provided the livelihood for the minor settlements at Northeye and Bulverhythe, both limbs of the Cinque Ports (Burleigh 1973, 72; Searle and Ross 1967, 52). The latter place even obtained a market charter, though it never developed into a significant settlement (Calendar of Charter Rolls 3, 13 7). Location may also be invoked as the cause of failure of two Wealden markets. Frant was granted a market charter in 1296, and when a survey was made ofthe manor in 1635 there were a few houses around the green, including a smithy and one shop-place (Calendar of Charter Rolls 2, 467). Other houses on the west side of the green which lay in Rotherfield manor are shown on a map made about 40 years earlier. The location, although on a road northwards to Tonbridge, was not propitious. The settlement was set amid acidic heathland described as 'waste' in the mid-14th century and its hinterland was inadequate to support a market village. Parrock in Hartfield occupied a similar location on the northern slopes of Ashdown Forest. Pottery, the greater part dated to the Trade, Rural Industry and the Origins of Villages: some Evidence from South-East England 13th and 14th century, and iron slag suggest a possible industrial settlement (Tebbutt 1975). Quo Warranto proceedings of 1279 record the claim of a market charter suggesting the aspiration to establish an altogether more ambitious settlement, but proximity to the town of East Grinstead and the poverty of the soils in the surrounding area seem to have prevented its development (Placita de Quo Warranto, 754). The villages and hamlets of the Weald may in some senses be regarded as substitute towns. They supplied to a limited area those goods which otherwise would have to have been bought at the predominantly coastal urban centres. This is particularly apparent by examining the villages and hamlets in this area (fig. 1). The distribution shows a marked concentration inland away from the larger towns. Hastings, Winchelsea and Rye in the south-eastern corner of the county evidently dominated trade and manufacturing in their vicinity to such a degree that they prevented the growth of rival markets. The smaller settlements inland served the needs of communities beyond easy travelling distance of the coastal ports. A number of the planned settlements were seemingly founded with the intention of creating urban centres. The urban officials at Robertsbridge, the tenements held in free burgage at Boreham Street and the plan of Burwash have already been mentioned. The settlement ofLamberhurst over the border in Kent was described as a town in the late 13th century, the extent of which was delimited by two crosses (Placita de Quo Warrante, 364). All these effectively were failed towns, though they flourished as villages with some urban attributes. The discussion has shown that the period 1250 and 1300 was crucial for the development of the network of markets and villages in the Weal d. Markets established later generally failed to flourish and their associated settlements did not develop into places of any significance. The example of Wart ling has already been discussed and Brightling, which lay further north, is a further instance. Two stalls are recorded at Brightling in 1422-3, when one was no longer tenanted and the other below the churchyard paid a reduced rent. A fair held there in the early 15th century produced a very small toll and was evidently of little commercial significance. It was described in 14 72 as a vicus or hamlet (ESRO ACC 3612/5, unlisted roll 5; ESRO ASH 198, 200, 200A; BL Add. Roll 31359). The development of villages in the Weald has been traced from the mid-13th century using documentary evidence. There is little written evidence before that date, although some conclusions may be drawn from analysis of the topography. It has been argued that many villages developed around the site of a market-place adjoining a church. The churches at Ticehurst and Wadhurst were situated at the side of the market-place, set back from the main road. The areas which came to be occupied by the markets must therefore pre-date the foundation of the churches, taking them back before c.11 00 by which time most of the Wealden churches had been established. At Rotherfield the situation is rather different. The church stood at the eastern end of a clearly defined funnelshaped area ofland and indeed appears to have been founded within it. In all these places the churches were evidently secondary to the existing open spaces. These open areas or 'focal places' appear to be very ancient and may have served as meeting points or trading places long before churches were established. Trade, industry and village growth elsewhere Trade in much of medieval England developed around existing settlements, and in turn encouraged their growth. However, Salzman (1928, 207-208) has drawn attention to assemblies (congregationes hominum) which took place to trade in an informal manner. He cited the example of a Sunday gathering in the late 13th and early 14th century at Crosthwaite church near Derwent Water in Westrnoreland at which people came together to buy and sell corn, flour, peas, beans, flax, yarn, meat and fish. This was not a regularised market because no toll or stallage was paid, nor apparently did it take place at a village. The Crosthwaite gathering must have resembled many early markets, which were not regulated, but took place spontaneously when people were gathered together at church. Sunday was the most common day for markets before the early 13th century when under ecclesiastical pressure many changed to a weekday and moved out of churchyard (Britnell 1993, 84-85). The development from congregationes to licensed markets and their associated permanent settlements may been very much more common than documentary sources suggest. Unregulated trade on sites without settlement will rarely attract documentary notice. The markets of the Sussex Weald and at Crosthwaite discussed here may be unusual only because they were very late examples which left some trace in the written record. The pattern in Essex seems to have been very similar to the Sussex Weald. In both counties villages contained many who derived their living from craftwork (Gardiner 1996; Poos 1991, 36-37; Hilton 1977, 172-174). Both had high density of markets, 5.1 per 100 sq. miles in Essex compared to 5.5 in the Rape ofHastings at the eastern end of Sussex (Britnell 1981 b, 21 0). The develop71 M. Gardiner ment of village settlement has yet to be examined in Essex, but the present study may be instructive. The origin of village is unlikely to have a monocausal explanation. The growth of nucleated settlement took place over a number of centuries as the economy of England was undergoing very rapid change. Trade and craftwork were one of the factors which stimulated settlement nucleation. As such the historical and topographical evidence from the W eald is informative not merely about the growth of villages more generally, but may provide insights into the way towns developed in England at an earlier date. References Unpublished Sources BL British Library (London) CKS Centre for Kentish Studies (Maidstone) ESRO East Sussex Record Office (Lewes) HMAG Hastings Museum and Art Gallery HER Henry E. Huntingdon Library (San Marino, California) LIL Lincoln's Inn Library (London) LPL Lambeth Place Library (London) PRO Public Record Office (London) WSRO West Sussex Record Office (Chichester) Sussex Topographical Surveys: Playden Parish (copies in ESRO and the Sussex Archaeological Society library, Lewes). Printed Sources BRANDON P.F. 1969: Medieval clearances in the East Sussex Weald, Transactions of the Institute of British Geographers 48, 135-153. BRITNELL R.H. 1981a: Essex markets before 1350, Essex Archaeology and History 13, 15-21. BRITNELL R.H. 1981 b: The proliferation of markets in England, 1200-1349, Economic History Review 2nd series, 34, 209-221. BRITNELL R.H. 1993: The Commercialisation of English Society 1000-1500, Cambridge. BURLEIGH G.R. 1973: An introduction to deserted medieval villages in East Sussex, Sussex Archaeological Collections 111, 45-83. Calendar of Charter Rolls (1903-27), 6 vols, London. Calendar of Close Rolls (1892- in progress), London. Calendar of Patent Rolls (1891- in progress), London. Cooper W.D. 1869: Mayfield, Sussex Archaeological Collections 21, 1-19. D'ELBOUX R.H. (ed.) 1944: Survey of the Manor of Robertsbridge, 1567-70, Sussex Record Society 47, Lewes. 72 DAVIS F.N. et al. (eds) 1969: The Register of John Pecham, Archbishop of Canterbury 1279-1292 1, Canterbury and York Society 64, Oxford. DULLEY A.J.F. 1969: The early history of the Rye fishing industry, Sussex Archaeological Collections 107, 36-64. DYER C. C. 1985: Power and Conflict in the Medieval Village, in: D. HOOKE (ed.), Medieval Villages: A Review of Current Work, Oxford, 27-32. GARDINER M.F. 1996: The geography and peasant rural economy of the eastern Sussex High W eald, 1300-1420, Sussex Archaeological Collections 133, 125-139. GRAHAM R. 1952: Registrum Roberti Winchelsey, Cantuariensis Archiepiscopi AD 1294-1313 1, Canterbury and York Society 51, Oxford. HARVEY P.D.A. 1989: Initiative and authority in settlement change, in: M. ASTON, D. AUSTIN & C.C. DYER (eds), The Rural Settlements of Medieval England, Oxford, 31-43. HILTON R.H. 1977: Bond Men Made Free: Medieval Peasant Movements and the English Rising of 1381, London. JOHNSTON G.D. 1967: Ferries in Sussex [continued], Sussex Notes and Queries 16, 307-308. LEWIS C., MITCHELL-FOX P. & DYER C.C. 1996: Village, Hamlet and Field: Changing Medieval Settlements in Central England, Manchester. McKINLEY R. 1988: The Surnames of Sussex, Oxford. MAITLAND F. W. 1897: Domesday Book and Beyond, Cambridge. MARTIN D. & MARTIN B. 1979: A specialised building in Sedlescombe, Historic Buildings in Eastern Sussex 1 (iii), 71-72. PARSONS E.J.S. 1958: The Map of Great Britain circa A.D. 1360 Known as the Gough Map: An Introduction to the Facsimile, London. PECKHAM W.D. (ed.) 1925: Thirteen Custumals of the Sussex Manors of the Bishop of Chichester, Sussex Record Society 31, Lewes. PECKHAM W.D. (ed.) 1946: The Chartulary of the High Church of Chichester, Sussex Record Society 46, Lewes. PELHAM R.A. 1931: Studies in the Historical Geography of Medieval Sussex, Sussex Archaeological Collections 72, 156-184. Placita de Quo Warrante, London, 1818. Poos L.R. 1991: A Rural Society after the Black Death: Essex 1350-1525, Cambridge. RAY J.E. (ed.) 1931: Sussex Chantry Records, Extracted from Documents in the Public Record Office, Sussex Record Society 36, Lewes. REDWOOD B.C. & WILSON A.E. (eds) 1958: Custwnals of the Sussex Manors of the Archbishop of Trade, Rural Industry and the Origins of Villages: some Evidence from South-East England Canterbwy, Sussex Record Society 57, Lewes. ROBERTS B.K. 1987: The Making of the English Village, London. ROBERTS B.K. & WRATHMELL S. 1995: Terrain and Rural Settlement Mapping: The Methodology and Preliminary Results, (report for English Heritage), London. Rotuli Hundredorum, London, 1812-18. Rotuli Litterarum Clausarum, London, 1834. SALZMAN L.F. 1928: The legal status of markets, Cambridge Historical Journal2, 205-12. SAUL N. 1986: Scenes from Provincial Life: Knightly Families in Sussex 1280-1400, Oxford. SEARLE E. & Ross B. (eds) 1967: Accounts of the Cellarers of Battle Abbey 12 75-1513, Sussex Record Society 65, Lewes. TAYLOR C.C. 1983: Village and Farmstead: A History of Rural Settlement in England, London. TA YLOR C. C., EVERSON P. & WILSON-NORTH R. 1900: Bodiam Castle, Sussex, Medieval Archaeology 34, 156-157. TEBBUTT C.F. 1975: An Abandoned Medieval Industrial Site at Parrock, Hartfield, Sussex Archaeological Collections 113, 146-151. WHITTICK C.H.C. 1993, Dallingridge's Bay and Bodiam Castle Millpond- Elements of a Medieval Landscape, Sussex Archaeological Collections 131,119-123. Mark Gardiner Department of Archaeology The Queen's University of Belfast Belfast BTI 1NN Northern Ireland 73 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Carenza Lewis Medieval Settlement in Hampshire and the Isle of Wight The Royal Commission on the Historical Monuments of England (RCHME) is widely recognised as expert in the detailed field survey of the earthwork remains of medieval settlements, exemplified in projects in Northamptonshire and Lincolnshire (RCHME refs). This paper describes work that was carried out in Hampshire and the Isle ofWight (fig 1) in England to investigate a new approach to the study and recording of medieval settlements within RCHME. This paper provides a review of the aims and methods used and describes some of the results of the work in Hampshire and the Isle ofWight. Aims and Principles In England, a long history of archaeological investigation into medieval settlement sites, combined with a wealth of documentary evidence and maps means that a considerable amount of information regarding medieval settlements exists. However, these various strands of evidence have too often remained separate, which has limited their ability to increase our understanding of the nature and evolution of settlement in the middle ages. Furthermore, this separation of historical and archaeological evidence has resulted in the omission of many documented settlements from record systems designed to identify and protect historic sites. Recent RCHME work in Hampshire and the Isle ofWight was intended to address these problems by bringing together and assessing archaeological, historical and geographical evidence for medieval settlement over a wide area. The approach develops that used in recent Birmingham University research into medieval settlement in the east midlands (Lewis & Mitchell Fox 1992, 1993; Lewis, Mitchell Fox & Dyer 1996). The approach was firstly to create a database with a separate record for every known medieval (410-1540 AD) settlement, and secondly to analyse and map the database information to identify and study patterns in the development and pattern of settlement in the medieval period. The methodology devised for the project required that the database record for each settlement should include a range of historical, archaeological and geographical data, which would enable the extent of knowledge about each place to be immediately apparent and its likely character, status and type to be assessed. Following completion of the database, information on it had to be mapped, analysed and interrogated to produce a descriptive and interpretational review placing the evidence in its wider context. One of the strengths of the approach was that it could cover a large area rapidly, so that any county would take just three months to complete. Following completion, the database information had to be supplied to national and local archaeological record systems so that the information could inform managerial and research initiatives. A detailed account of the findings also had to be prepared and a summary published in the annual report of the Medieval Settlement Research Group (MSRG). Project methodology A number of different archaeological, architectural and documentary sources were used in the pilot, which was conducted by two full-time staff (one historical researcher and one archaeological investigator) and one part-time volunteer with no previous experience who helped with data inputting. Field assessments were carried out over one week by one archaeological field investigator. One ofthe benefits of the approach used in the Medieval Settlement Project was the speed and economy of effort with which it could produce useful results. The project comprised six main areas of activity summarised below. 1-3 and 5 were carried out in succession, concurrently with 4 (historical research). 6 (analysis) was carried out following completion of 1-5. 75 C. Lewis I) Historic place-name data input The first task undertaken was the entry to the database of all documented historic places likely to have been settlements in the medieval period (4101540 AD). One of the major inadequacies of existing heritage records (NMR and SMRs) is that most medieval settlements which are of documented medieval date but for which no archaeological evidence has yet been recorded are omitted. One of the aims of the Medieval Settlement Project was to remedy this deficiency. For each place documented before 1540, eight items of information were input to the database as the first stage of the pilot. These comprised the modem place-name, the earliest form or spelling of the name, the meaning of the place-name, the date at which the name was first documented, the national grid reference (if identifiable, to six figures), the modem parish and county within which the place lies, and the project database record identification number. This created 2,289 records on the database and took 12 working days to complete. Place-name surveys compiled by Kokeritz (1940) for the Isle of Wight, and by Gover (unpublished typescript 1960) and Coates (1989) for Hampshire provided information for this stage. Modem Ordnance Survey maps at 1:50,000 were used to locate these places and provide grid references. Some additional grid references were found from 1:25,000 Ordnance Survey maps. Despite this fairly intensive effort, not all places could be located: some are only named on 1:10,000 maps, others only on older 6" maps, and others are even more obscure. Limited time did not allow the examination of any of these maps for missing grid references, consequently c. 500 historic places on the database still lack complete grid references. 2) Archaeological data Four main sources were used to identify the physical evidence for medieval settlement in Hampshire and the Isle of Wight, namely the Sites and Monuments Records (SMRs) for each county, the National Monuments Record (NMR) held by RCHME and the archive of the Medieval Village Research Group (MVRG). All of these records are now at least partially computerised, but have been built up cumulatively over a period of time, and in most cases each newly discovered piece of information has been added as a separate record as it comes to light. Thus a standing medieval building is likely to be recorded in a separate record to an excavated building or a 76 medieval pottery find from within the same settlement: rarely is such a place classified anywhere as a settlement. The archaeological evidence for medieval settlements is presently very dispersed and is thus almost impossible to identify, retrieve, quantify or assess. One aim of the RCHME research was to collate and synthesise this disparate information and create supplementary records of settlements. Each of these would show the range of evidence for the whole settlement, enabling the level of knowledge about each place to be immediately apparent, and its likely character, status and type to be easily identified. i) SMRs: The Hampshire SMR provided a print-out of all medieval (410-1540 AD) records classified under the following categories: farm, moat, manor, settlement site, building material, finds, cruck-framed building, hearth, house platforms, occupation hollows, pottery finds, kilns, sunken-floored buildings, parish church, deserted village, shrunken village, village and chapel. It was hoped that this would retrieve all records likely to contain settlement data (the large number of terms which had to be used to recover settlement information illustrates the problem of trying to identify settlement information in existing record systems). The printout (including the descriptive text) was searched thoroughly and all information about medieval settlements was collated, synthesised with the place-name data and input to the database. For the much smaller county of Isle of Wight all records of early or late medieval date were rapidly scanned for settlement evidence at the County Archaeological Unit, following which a full print-out of all records found to contain information relevant to medieval settlement was provided for detailed assessment and entry to the database at NMRC. ii) NMR: From the NMR a short print-out of all records of medieval date (including both archaeological and architectural records) in Hampshire and the Isle of Wight was obtained. This was scanned to identify all evidence indicative of medieval settlement, which was then collated and added to the database. iii) MVRG: A printout of the indexed information from the MVRG (held by RCHME) was used to cross-reference sites which had been recorded by the Medieval Village Research Group. The detailed archive information (which has not been computerised) was rapidly reviewed for relevant information regarding extent of earthwork remains. Medieval Settlement in Hampshire and the Isle ofWight Fig.l. Project area RCHME As the aim of the pilot was to identify, review, collate and assess the evidence for medieval secular settlement, information about standing domestic buildings of medieval date was included, and in some cases this provided the only physical evidence to support the documented medieval date of a settlement. The presence of a medieval church was recorded as an associated monument, but was not assumed necessarily to indicate the certain presence of a settlement. Pevsner and Lloyd's survey of the buildings of Hampshire and the Isle of Wight (Pevsner & Lloyd 1967) was used to establish the date of some churches and other buildings where this information was not included in the SMRs or NMR. Information about other types of medieval buildings such as religious establishments, castles, hospitals etc was only included if the sites lay within settlements, in which case they might have affected the status and development of the settlement. Any of this information which related to places recorded during the place-name data survey was ad- ded to the database record for that place. Any evidence about a site not already on the pilot database was added as a separate new record. Approximately 270 new settlement sites were added to the project database from evidence contained in the SMRs and NMR. Any additional information necessary to clarify or support the archaeological evidence was included in the free text memo field. Work collating and synthesising the pertinent information from the major existing archaeological records took 16.5 working days. 3) Nineteenth century settlement form With the exception of deserted or very extensively shrunken settlements, historical and archaeological information reveals little or nothing about the size or layout of medieval settlements. However, an understanding of the likely morphological form of settlement in a region is often crucial in establishing its archaeological potential. For example, a place with a 77 C. Lewis ~: Fig. 2a. ·. . ...... ..-: . i .J.~ ...... ,l Nucleated cluster settlements in Hampshire and the Isle of Wight in 1801 i 10 km. large documented medieval population which now exists only as a single building may reasonably be posited as the possible site of a deserted medieval village if it is in an area where all other surrounding settlements are nucleated. If, however, all neighbouring settlements are of dispersed form it is more likely that the medieval record of a population under a single place name represents an administrative convenience. In this case it is unlikely that the named place was ever a large nucleated settlement, but was perhaps merely the site of the manor house with the peasant/ tenant tofts scattered across the settlement territory. The source used to characterise the form of the places recorded in the pilot project was the first edition 1" Ordnance Survey (OS) map published for the region by Colonel Mudge in 1810. This was selected because it was the earliest map which provided consistent and reliable coverage across the whole of the pilot area while also providing an adequate degree of detail. Earlier maps at the County Record Office were of insufficient detail, quality or coverage. (This OS map also confers an advantage for the future in that it covers the whole of England, allowing for consistency across a national project) For each historic place recorded on the pilot database from the place name survey and the archaeological records which could be identified on the 1810 OS map (a total of 1,518), the form of the 78 lsoo settlement in 1810 was entered onto the database, using a simple range of morphological types which included compact/nucleated cluster, regular row, interrupted/irregular row, common-edge settlement and isolated farmstead. These (with the exception of isolated farmsteads) were subdivided into small, medium and large settlements, so that a cluster ofless than 15 houses was classified as a small compact/ nucleated cluster, rather than a hamlet, avoiding use of that term which has never been firmly defined. In the course of the 1810 map survey a number of other places were noted which, despite the apparent lack of historic place-name evidence for their medieval existence, seemed likely on other grounds to have formed part of the medieval settlement landscape. A total qf.l93 of these were added (with the standard data range of name, parish, NGR etc data) to the database as new records, classified as settlements of postmedieval (but not medieval) date. 4) Historical Data Most of the historical data examined related to levels of population, wealth and agricultural development, but also included simple types of information (where available) about the economic and institutional status of settlements and also the agrarian Medieval Settlement in Hampshire and the Isle ofWight Fig. 2b. . .•. I J.!)g ·+···· Regular street/row settlements in Hampshire and the Isle of Wight in 1801 . 12 ..5 km . ~00 organisation associated with them. This information substantiates the record created from the place-name and archaeological sources, and enables the varying size and importance of the recorded settlements in the medieval period to be identified, assessed and compared locally and regionally. Additional data relating to other aspects of settlement such as social and manorial structure was also reviewed and provided additional background information for the final synthesising report but was not included in the database because of its less reducible character. i) Domesday Book: Domesday Book was used to extract and collate data pertaining to settlement for three categories: 1. Tenurial- including the number and size of holdings and the type of lord (royal, lay or ecclesiastical). 2. Agrarian and economic resources- including the numbers of ploughs (separated into demesne and tenant) and the extent of other resources including meadow and woodland. 3. Population- the number of people recorded at each place according to their various categories. The total of all recorded plough teams and population, and an adjusted population figure (allowing an average of 4.5 people per recorded family for all categories except figures for slaves which were assumed to represent individuals) were then calculated for each ' lOO ecclesiastical parish to allow the mapping of population densities across the region. Some additional calculations have also been made using the Domesday Book data, including the relative proportions of demesne and tenant plough teams for each place, and the numbers of slaves and their ratios to demesne ploughs (there has been speculation that this should be 2: 1, indicating that slaves acted as demesne plough men, but unusually, this is often much higher in Hampshire). The data for population, lordship and ploughs was input to the database. Other information was used to produce county distribution maps and to inform the county report. The time taken processing the Domesday Book data was three weeks. ii) 13271133411428/ay subsidies: The second task was to collect and synthesise data from the lay subsidies of 1327, 1334 and 1428. The 1327 subsidy was chosen because it is the earliest (and appears to be the only) record for Hampshire to provide lists of named tax payers for each place. As this roll has never been published the original, held in the Public Record Office was examined. The numbers of tax-payers listed for each place were counted and the overall tax assessments recorded. The names of the individuals were also scanned for evidence for additional contemporary settlement sites or other topographical information. These were added to the database. 79 C. Lewis The 1334 subsidy, though only listing the total sums paid by each settlements (rather than listing sums paid by individuals) was included in the pilot because it is one of the few medieval taxation records providing local evidence surviving for the whole of the country, and will therefore be important for making comparisons within the national project. The published edition of the 1334 data was used for this (Glasscock 1975). The figures for 1327 and 13 34 were input to the database for each settlement and totalled by ecclesiastical parish to enable the densities of population and taxable wealth to be mapped across the region. The time taken collecting and processing the fourteenth century taxation data was two weeks. The list of parishes with less than ten households in the lay subsidy of 1428 is a useful indicator of those settlements which were particularly small after the Black Death, particularly in Hampshire where the 13 77 Poll Tax returns are not available. This information was also input to the database. iii) Nomina Villarum: Data has also been taken from the 1316 Nomina Villarum, providing evidence for the whole county of the number and type of lords then holding land in each vill. iv) 1524 lay subsidy The latest taxation records employed were those of the Lay Subsidy levied in 1524-5 (Sheail 1968). The numbers of tax-payers for each place were assessed and input to the database and calculated by parish for mapping purposes in one week. v) Settlement status The identification of evidence relating to the status of settlements in the early or later medieval period involved trawling a miscellaneous collection of sources, in a more speculative search for evidence about institutional and economic status and agrarian organisation. Examination of pre-Conquest sources, including charters and the Anglo-Saxon Chronicle, allowed identification of Anglo-Saxon royal vills, assembly sites and minsters. Post-Conquest sources examined include the calendars of royal charter rolls, the Victoria County Histories (VCH 1900-1912) and Beresford and Finberg's handlist of boroughs (Beresford & Finberg 1973), which has produced a list of over sixty medieval markets, boroughs and fairs. All Hampshire cartularies readily available in print have been scanned, principally for evidence of field systems, as well as additional evidence relating to settlement status. Published manorial records have helped illuminate social structure and the agrarian economy. 80 5) Field Assessment 56 sites were selected for field assessment to ascertain the presence and extent of earthwork remains of former settlement. Most of these were historically attested sites in dispersed regions, which were revealed by the desk-top survey to be particularly poorly understood and often classified inaccurately or on inadequate evidence. For example, many sites have in the past been classified as deserted villages from historical sources alone, and it is now recognised that many places in non-nucleated regions which are documented as having taxable populations may have been merely the site of the manor within an area of scattered hamlet and farmstead habitation and never existed as nucleated villages: such sites require field verification to establish whether there really is any physical evidence for more extensive former settlement - if none is evident, such sites should not be classified as deserted or shrunken villages. 6) Analysis Following completion of the data collection and database entry, maps were generated from the database using a computerised geographical information system. This facilitates evaluation of the geographical distribution of various settlement phenomena, such as dispersed and nucleated settlement types, deserted and shrunken settlements. It is possible to produce distribution maps of any query carried out on the database. The potential of this for the study of medieval settlements is only just beginning to be explored (Lewis & Mitchell Fox 1996, RCHME in preparation). All collected data was analysed using the database and maps- these are, independently and together, a very powerful but easily accessible research tool. The following pages provide an indication of the sort of analysis that can be conducted using the project data and maps. Results: Medieval settlement in Hampshire and the Isle of Wight The project data, available on the database and as a series of maps enables a wide range of issues pertaining to the nature and development of the medieval rural landscape to be explored and assessed. Four issues are considered to provide examples of the sort of assessment that can be carried out using the RCHME project data. Medieval Settlement in Hampshire and the Isle ofWight ~i Fig. 3. . ...... ·. ... . ·.. ... .. .. .I I I .\!?.() i ·····r··r Dispersed settlements (Common-edge and interrupted rows) in Hampshire and the Isle of Wight in 1801 10 km. 1) Settlement pattern A computerised mapping software system was able to provide distribution maps of different settlement types in 1810. These showed clearly that nucleated villages dominated the major river valleys of the chalkland, and the chalk massif of central Hampshire (fig 2). Dispersed settlement types are very uncommon in these areas (fig 3). In contrast, it is clear that the north, east and south of the region is largely characterised by dispersed settlement; nucleation in these areas is associated with psuedo-urban function (the presence of a market, etc), or recent (nineteenth/twentieth century) expansion. Essentially the dispersed regions correspond to the area away from the chalk. It is notable that there is some similarity in the distribution of evidence for dispersed settlement and that for settlements documented for the first time only in the later medieval period (fig 4). The database confirms this observation: 64% of dispersed settlements (interrupted row and common-edge) are documented for the first time post-1 086, whereas for nucleated settlements (clusters and rows) the figure is only 40%. Both figures, of course, reflect the expansion in written records in the fourteenth and fifteenth centuries which led to many smaller settlements being recorded for the first time. But this does seem I I soo to indicate a greater degree of expansion in dispersed, woodland regions than in the nucleated champion regions. The roots of this may lie in the relatively underexploited nature of woodland!heathland regions in earlier centuries. In the pre-Conquest period, 32% of all nucleated sites are referred to in documents, but for dispersed settlements the figure is only 8%. Interestingly, however, Domesday Book records similar percentages of both types (26% of all dispersed settlements are recorded there, compared to 28% of nucleated settlements), suggesting that woodland regions were 'catching up' fast in the later preConquest centuries. Certainly by 1086, they do not appear to have been vast uncharted and unadministered wastelands. Equally, the woodland and heathland regions seem to have seen a particularly great expansion in settlement in the post-Conquest era, an observation which is supported by documentary references to assarting on many of the later medieval manors, including those of the Abbey ofWinchester, whose pipe rolls documents many details of estate activities. The correlation of earlier documentary evidence with that for the nineteenth century settlement pattern suggests that, while some elements of the later pattern are of late origin (much of the present settlement around the Solent, for example, post-dates even 81 C. Lewis Fig. 4. .. . .... . '. .... ··.: : . .·..... .·..... .. . ·.:. ·:.-::.. . . ... ... . ... .. . ........ ·.· .. .. .. ·.... · ..... . ·.·..· \ ~ Settlements in Hampshire and the Isle of Wight documented for the first time post. 1086 25 km. ! 400 the early nineteenth century), the post-medieval patterns of dispersed and nucleated settlement may in many cases, give a fair indication of the medieval pattern. Another interesting figure which can be noted is that only 18% of common edge settlements are named in Domesday Book, while 20% are named for the first time in the fourteenth century. For interrupted rows, in contrast, 30% are name din 1086, and only 8% appear for the first time in the fourteenth century. While a documentary reference does not, of course, data the origins of the settlement, the difference between the two types of dispersed settlement may indicate that common-edge habitation is a later form of settlement. Only further archaeological investigation could confirm of refute this suggestion. 2) Field assessment and identification of deserted sites Few sites which can confidently be classified as deserted are known in the region (fig 5). The possibility of using historical evidence to identify or predict likely deserted sites can be tested using the information recorded on the database. For example 119 sites are recorded in Domesday Book but omitted from 1334lists: those with very limited settlement in 1810 are often classified as deserted medieval vil- 82 ' 500 lages (DMVs). However, in Hampshire and the Isle of Wight only five such sites actually have any recorded archaeological evidence for shrinkage. The others therefore must be classified, on the basis of known evidence rather than presumed status, simply as settlements or documented place names. Sites whose population fell markedly between 1086 and 1327 might also be expected to be prime candidates for desertion in the later fourteenth and early fifteenth centuries, and to exhibit evidence for shrinkage. The recorded population of 10 sites fell from 50% above average in 1086 to more than 25% below average in 1327. However, these were all still paying tax in 1524 (albeit with a low average population of 13.5 compared to 29 overall for the region), and none have recorded evidence for shrinkage or desertion. These sites could usefully be targeted for future field assessment, to ascertain whether or not they have any earthwork evidence for contraction. However, in the meantime, it must be suspected that these falls in population in fact reflect, at least in part, the widespread fragmentation of Domesday manors into smaller taxation units in the centuries following the Norman Conquest. Chilcombe, for example had 9 churches and a recorded population of c. 213 in 1086: the fall to just 4 in 1327 must be due in part to the break-up of this large estate of the Bishop of Winchester into smaller taxable units. Medieval Settlement in Hampshire and the Isle of Wight r-~---- ---~,----- - - - - - - - - - - - ---·---· I ~~ Fig. 5. ,' I .· .. Deserted and shrunken settlement~ in Hampshire and the Isle of Wight 10 km. During the project, 56 sites were selected for field assessment. Most were either isolated farmsteads whose names were documented in the medieval period or sites classified on archaeological record systems as deserted, in areas of predominantly dispersed settlement, for which archaeological evidence was particularly poor. Field assessment revealed that one in four had earthwork remains of settlement, but in no case were these extensive or substantial. The presence of other earthworks, including remains of ridge and furrow field systems, at a number of sites suggests that settlement evidence, had it existed, should have been preserved. The absence of large areas of abandoned former settlement suggests that the regions where settlement was dispersed in the nineteenth century were probably, likewise, areas generally lacking large nucleated villages in the medieval period. As a substantial proportion of sites visited had previously been classified as deserted, the field assessment indicated that extreme caution should be exercised in classifying settlements as deserted on historical evidence alone. 3) The early medieval period- shift and success Of more than 2, 700 sites recorded in the pilot, only 299 have any evidence for occupation in the I lsoo early medieval period. For 210 of these the evidence is purely documentary; and only 39 actually have definite archaeological evidence for settlement. No archaeological evidence at all is known for early medieval settlement from the Isle of Wight. While these figures may seem to indicate something of a dearth of evidence for settlement of this period, the region does include the sites of Chalton Down (in south Hampshire) and Cowdery's Down (in north Hampshire), both of which have been extensively excavated (Addyman & Leigh 1973, Cunliffe 1973; Millet & James 1983). Overall, only a handful of attested early medieval settlement sites were deserted before the Norman Conquest - even in the case of early (ie pre-7th century) Anglo-Saxon sites, most known sites continued in occupation in the post-Conquest era. There are however 21 sites which produced pottery finds but which could not be confidently classified as settlements, and most of these are from places which were not flourishing in the later period. These include Up Sombome, where seven sherds of (possibly residual) Saxon pottery were recovered from a posthole; a deserted site in Faccombe parish where 6 sherds of grass-tempered ware were found in a remote downland situation, similarly sited pits associated with animal bone and Saxon pottery in Romsey Extra parish, and a single sherd found in 83 C. Lewis fields in Farringdon parish. The last 3 are from sites which are unoccupied in the post-conquest period, but the slight nature of the archaeological evidence means that they cannot realistically be classified as early medieval settlements, but as merely as pottery finds. Other pottery find spots are even less likely to represent settlement: some may be funerary vessels (such as at Compton and Hucklesbrook Farm (in Ellingham, Harbridge and lbsley parish), many others such as at Lymington, Odiham or Barnes High (Isle of Wight), are dubious or unprovenanced. However the distribution of such evidence, and other deserted early medieval settlement sites does however generally favour the chalk downland and suggests an abandonment of these areas (which were extensively occupied in the Roman period) during the AngloSaxon period. This is supported by excavations at Chalton and Cowdery's Down, both downland sites which were abandoned by the mid-seventh century. In other areas there is little or no evidence for deserted early medieval settlements. This is particularly significant in east Hampshire and the Avon Valley in west Hampshire which have been subject to intensive fieldwalking programmes (Shennan 1985; Light et a/1994;). The few abandoned early medieval sites which have been found in the Avon valley showed strong continuity with Roman pottery distribution. Most are also close to later medieval settlements, suggesting a process of gradual shift within the densely occupied river valleys. While the downlands seem to have been abandoned for settlement, elsewhere the assumption, based on present evidence, must be that settlement was either very sparse (which seems unlikely), or very conservative, mostly underlying or adjacent to later settlements. This contradicts current orthodoxy, based mostly on work carried out in the midlands, where it is thought that small dispersed early medieval sites were abandoned in large numbers, probably around the ninth century, and replaced by nucleated villages set within regular open field systems. Settlement change in Hampshire seems to have been of a more limited extent, mainly comprising a much earlier (ie pre mid-eighth century) abandonment of the downland. Riverine settlement was always an enduring.feature of the landscape in Roman, Anglo-Saxon and later medieval periods, but downland settlement was a feature only of particularly expansive periods, such as the Roman (and the high medieval, see below). It is of course, quite possible that Hampshire, whose terrain and political history over the Anglo-Saxon period were both very different to the midlands, did indeed have a correspondingly different process of settlement evolution during this period. 84 In fact, many early medieval settlements in the region seem, rather than being abandoned, to have survived and flourished. Examination of the project data suggests that there is a significant link between early occupation and enduring success. More than a third (12 out of33) of early medieval settlement sites were towns or market villages in the later medieval period and three-quarters (25 out of33) were medium or large settlements in the early nineteenth century. Overall, 12 of the 50 late medieval towns and market villages have archaeological evidence for early medieval settlement, and this must suggest that many other similar sites may also have Saxon precursors. Even early-established sites which did not become markets seem to have been larger than average. The average population for these sites in 1086 was 49 (compared to 25 overall), in 1524 it was 72.5 (compared to 29 for all settlements). Only 5 have evidence for significant shrinkage or desertion in the later medieval period (2 are shrunken, 3 deserted). While large or urban sites are perhaps more likely to have been excavated, (in advance of development) introducing a possible bias to the sample, the fact that documented early medieval sites show a similarly higher than average population of 40 in 1086 supports the general suggestion of a link between Sax on occupation and post-Conquest longevity and prosperity. Interestingly, 19 out ofthe 33 sites with evidence for early settlement were also documented in the preConquest period- an unusually good correlation! 4) Later medieval settlement- variety in desertion In many areas the downland remained devoid of settlement from the early Saxon period onwards. In others, particularly in central and north Hampshire, settlements such as Hatch near Basingstoke (Fasham et a/1995), were established on downland in the late pre- and early post-Conquest period. Few later medieval settlements with conclusive evidence for desertion are known in Hampshire and the Isle ofWight, but the figures for those which can be identified reveal some interesting patterns. The average population for all sites recorded in Domesday Book was 25; for settlements which were later deserted (excluding Old Highclere, to which we will return below) it was only 15. In 1327 the average number of taxpayers for all recorded sites was 16, but only 9 for later deserted settlements. It seems that a significant number of deserted sites were ones which had always been smaller and poorer than average. However, the data suggests considerable variation in the progress of depopulation. Lomer, for exampl,e recorded 11 individuals in 1086, 7 in 1327, fewer than Medieval Settlement in Hampshire and the Isle of Wight ten households in 1428, but still had 6 (8 are listed in the second survey) taxpayers in 1524. With early medieval occupation attested by pottery finds and a documentary reference in AD 802, Lomer, although always small, seems to have maintained quite a stable population for most of the middle ages, and must have declined to its 1810 farmstead status in the postmedieval period. Apparently more erratically, Westbury (in East Meon) recorded a reasonable population of 13 individuals in 1086, contained only one taxpayer in 1327. The population recovered by 1524 to 9, but fell again in the post-medieval and was just a single farm by 1810. Other sites show a different pattern again. Durton (in Arreton), for example, recorded with 2 occupants (but with 2.5 ploughs, a very much higher ratio than normal) in 1086. Although, unusually, it rose in value between 1066 and 1086, it lacked a church, had vanished as an independently taxed unit by 1327, and was just a single, unremarkable farmhouse in 1810 (Winter 1984, 187). Durton has been classified as a deserted village by the MVRG and the NMR on the basis of its disappearance from documentary sources, but the evidence suggests that it may never have been more than a single farm. Contraction, if it did occur, probably dates to the eleventh century or earlier. Old Highclere, in contrast again, stands out as the site of a probable Saxon minster and with 69 recorded inhabitants, had a much larger population than any other deserted site in 1086. Significantly, its abandonment was not a result of decline: it was forcibly depopulated (or, in effect, relocated to the existing site of (New) Highclere) when a park was created in the thirteenth century. Other large settlements which were later deserted are Merdon and Newton. Newton was a borough in the thirteenth century; while Merdon was associated with a castle and paid the highest sum of all deserted settlements in 1334, but was reduced to 9 taxpayers by 1524, and just 2 farmsteads by 1810. The decline of these settlements may be related to their commercial failure. The evidence collected in Hampshire and the Isle of Wight points to considerable variation in the processes by which settlements came to be deserted. It is notable that there is no evidence for the sort of sweeping depopulation seen across vast swathes of the midlands. This suggests that depopulation generally occurred on a settlement-by-settlement basis, for specific site-related reasons, and not as part of a widespread, chronologically distinct process involving rural recession. Conclusion The RCHME project in Hampshire and the Isle of Wight has increased our knowledge and understanding of a range of aspects of the historical process of settlement between AD 410-1540 in the project region, which will be of value to those involved in curating, managing, studying, recording or surveying medieval settlement sites. The project has clarified the present extent of knowledge and highlighted a number of lacunae which exist in the archaeological evidence, particularly in areas of dispersed settlement, thus identifying priorities for future work. The RCHME research has also created consistent, standardised records for all known or suspected medieval settlements which will improve the ability of county archaeologists to manage and protect settlement sites. Despite planning legislation, such sites are often particularly threatened by development ranging from modern village infilling to agricultural building construction, and most are at present excluded from archaeological record systems such as the National Monuments Record (NMR) or county Sites and Monuments Records (SMRs) and can thus fail to trigger appropriate mitigatory action. The corpus of information available as a result of the RCHME project should also benefit the academic community for whom, it is hoped, the database, printouts, maps and report will provide evidence for, and stimuli towards, new research directions. As always, the most difficult part of any research is working out which questions should be asked of the data. The preceding pages have outlined a few of those which have suggested themselves during the project and which may stimulate debate, but detailed pursuance of research themes is beyond the scope of the RCHME project. One of the great merits of this approach, however, is that the data collected will henceforth be available for anyone to revisit and re-examine. It is also hoped that the information will be of interest to the general public, many of whom live in the settlements which have been studied. From a European perspective, it is important to note that one of the reasons why the RCHME work was carried out in southern England was to establish whether the methodology developed for the east midlands would work in a different region - and ultimately on a national scale. The work in Hampshire and the Isle of Wight has established that it would. Extending the scope of such research beyond the original east midlands focus allows comparison of a standardised set of information for very different regions which is vitally important if we are to understand the interplay of various factors in the evolution of medieval settlement in England. 85 C. Lewis Could such an approach usefully be carried out in other European countries? RCHME research in southern England indicates that the basic principle of methodically synthesising and analysing as much as possible of the information pertinent to medieval settlement from diverse sources to assess the present state of knowledge, create a consistent record and identify priorities for future work - could feasibly and usefully be applied to any European country, all of which have a historic pre-modern period. Such studies carried out in other European countries would allow evidence and ideas from different countries to inform and stimulate each other across the continent, for an era when cultural and territorial boundaries were in almost constant flux, when ideas could travel widely, and European integration was less an issue than a fact of life. However limited and problematic the evidence for medieval settlement might be, assembling and assessing it in this manner is one way of making the most of it, wherever the work is being carried out. Acknowledgements The work this paper describes is the result of a coordinated effort by several people. First and foremost, Patrick Mitchell Fox carried out the historical research without which the project could not have been attempted and Katharine Moore worked voluntarily data-processing, both with great dedication and enthusiasm in difficult circumstances. Staff at the county archaeological and record offices for Hampshire and the Isle ofWight were generous in supplying information from their records and expert local knowledge. In particular I must thank Michael Hughes for his invaluable support, advice and local knowledge. Also Hampshire County Council who generously provided a grant to support the project. The project owes its greatest debt to everyone who has worked on medieval settlements in the region and contributed to the store ofknowledge from which the project has drawn. References BERESFORD M.W. & FINBERG H.P.R. 1973: English medieval boroughs, Newton Abbot, David and Charles. COATES R. 1989: The place-names of Hampshire, London, Batsford. CUNLIFFE B. 1973: Chalton, Hants: Evolution of a landscape, Antiquaries Journal 53, pt 2, 173-90. ADDYMAN P.V. & LEIGH D. 1973: The Anglo-Saxon village at Chalton, Hampshire: second interim 86 report, Medieval Archaeology 17, 1-25. FASHAM P., KEEVIL G. & CoE D. 1995: Brighton Hill South (Hatch Warren), Wessex Archaeology Report 7, Wessex Archaeology, Salisbury. GLASSCOCK R. 1975: The lay subsidy of 1334, Oxford University Press). GovER J.E.B. 1960): Hampshire place-names, vols 1 and 2, unpublished typescript. KoKERITZ H. 1940: The Place-Names of the Isle of Wight, University ofUppsala. LEWIS C. & MITCHELL FOX P. 1992: Medieval settlements in Bedfordshire and Northamptonshire: an interim report, Medieval Settlement Research Group Annual Report 7, 15-20. LEWIS C. & M ITCH ELL Fox P. 1993: Medieval settlements in Buckinghamshire and Leicestershire: an interim report, Medieval Settlement Research Group Annual Report 8, 21-27. LEWIS C. & MITCHELL Fox P. 1996: Settlement in Hampshire and the Isle of Wight, Medieval Settlement Research Group Annual Report 10. LEWIS C., MITCHELL Fox P. & DYER C.C. 1996: Village, Hamlet and Field, Manchester University Press). LIGHT A., SCHOFIELD A.J. & SHENNAN S.J. 1994: The middle Avon Valley survey: a study in settlement history, Proceedings ofthe Hampshire Field Club and Archaeological Society 50, 43-101. MILLET M. & JAMES S. 1983: Excavations at Cowdery's Down, Basingstoke, Hampshire, 1978-81, Archaeological Journal140, 151-279 PEVSNER N & LLOYD D. 1967 (reprinted 1973): Hampshire and the Isle of Wight, Middlesex, Penguin. RCHME (in preparation): Medieval settlement in Hampshire and the Isle of Wight. SHEAIL J. 1968: The regional distribution of wealth in England as indicated in the 1524-5 lay subsidy returns, unpublished PhD thesis, London University. SHENNAN S. 1985: Experiments in the collection and analysis of archaeological survey data: the East Hampshire Survey, Collis, Sheffield. VCH 1900-1912 (reprinted 1973): A history of the county ofHampshire, vols 1-5, London, Dawsons. Carenza Lewis RCHME National Monuments Record Centre Kemble Drive, Swindon SN2 2GZ UK Rural Settlement in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Terry Barry Recent research in medieval rural nucleated settlement in Ireland Since the last Medieval Europe Conference, some five years ago, there has been continuing research in the field of rural settlement studies in Ireland generally. Although this paper will concentrate on giving an overview of research on nucleated rural settlement in the High Middle Ages, that is from the twelfth to the sixteenth century, it will also briefly examine the latest research on the earlier period from the coming of Christianity in the fifth century up until the Norman invasion of 1169-70. In the first millennium AD Ireland was dominated by a dispersed pattern of rural settlement, the main element being the ringfort or defended farmstead of the free element in early Irish society. There were at least 45,000 examples of this settlement form found all over the Irish countryside, mainly constructed between 600AD-900AD (Stout 1997). The ringfort itself is an annular settlement form delimited by an external ditch with an internal bank, with an average diameter of 30 metres. In the eastern half of the country where the soil cover is thicker they are often known as raths, while in the western, more rocky areas, they are known as cashels as their banks are often constructed of dry stone. Recent research by Stout, using County Offaly as a detailed study area, has shown that ringforts in densely settled upland areas were frequently located more than 2 km away from ecclesiastical sites, while marshland zones were dominated by well-defended ringforts situated in strategic locations (Stout 1996). In some cases they are also accompanied by souterrains, underground passages which were possibly either used for storage or as hiding places for the community who lived in the ringfort. The chronology of these enigmatic features is still not fully established although there is a secure date for at least one example at Coolcran, County Fermanagh where the oak posts which originally supported a roof of oak planks have been dendrochronologically dated to 822+9 AD (Williams 1985). However, this may not be a typical souterrain because the great majority of surviving examples seemed to have been constructed of dry stone. Along with the ringforts there were also possibly over 1,200 crann6gs, which were lake dwellings often constructed on artificial islands made out of soil, timber, stones and even from the occupation remains of the settlement itself (Edwards 1990). They often have their origins in prehistory, but some have important early medieval horizons. One of the most spectacular later examples is located at Moynagh Lough in County Meath, currently being excavated by Bradley. There was also substantial prehistoric settlement which predated the construction of the crann6g in the first half of the seventh century. The crann6g itself had five occupation layers which lasted from then to the end of the eighth century and revealed evidence for possible workshops which included metalworking. Many spectacular finds of the period have also been found in these levels, such as some very unusual artefacts such as two Merovingian glass vessels and a bronze spatula, which may have been part of a cosmetic set (Bradley 1996, 70). There were also many hundreds of ecclesiastical enclosures, not all necessarily of monastic origin, many of which have been identified by aerial photography (Edwards 1990). They are most densely distributed in a central band stretching from Dublin in the east to Clare on the western coast. As well as these we have many examples of early medieval monasteries ranging from the small isolated examples like Skellig Michael located on a rock off the County Kerry coast, to the larger examples such as Clonmacnoise on the bank of the River Shannon in County Offaly that arguably grew into some kind of town by the eleventh century. There has also been an attempt in recent years to locate evidence for rural nucleated settlements in preNorman Ireland. Geographers such as Evans have suggested, based on an interpretation of the early law tracts, that these might have been the dwellings of those poorer members of early Irish society who lived in separate settlements from the free element of that society who inhabited the ringforts (Evans 1964). Research has centred around the 'clachan', which has been identified by Proudfoot as 'a cluster of farm houses and associated outbuildings usually grouped without any formal plan' (Proudfoot 1959, 110). A 87 T.Barry drain N T ---:: 0 so m Fig. 1. ·Plan of Pipers town Deserted Medieval Village, County Louth. few of these settlements have survived to this day in areas as diverse as Donegal in the north-west, to the southern tip of Kilkenny in the more prosperous south-eastern part of the island. Examples can be dated from at least as early as the seventeenth century in the cartographic sources of the period, but all attempts to prove a medieval origin for them have been fruitless to date. The archaeological excavation of the deserted clachan at Murphystown, County Down, the only example to have been scientifically excavated, produced no conclusive evidence of a medieval origin for this settlement (Barry 1994, 21 ). To complete the picture of the rural settlement pattern of pre-Norrnan Ireland, there is mounting 88 evidence to show that some of the major prehistoric monuments such as the hillforts and promontory forts were still being occupied in this period. For instance, the interior of the Iron Age promontory fort at Dunbeg on the Dingle Peninsula was occupied by the remains of a dry stone beehive structure which had two tenuous occupation layers that were radiocarbon dated from the end of the ninth to the beginning of the eleventh century AD (Barry 1981, 311-317). As well as these earlier sites which were re-occupied there were probably also unenclosed dispersed settlements existing in this period which are very difficult to locate because of the non-existence of an identifiable bank or ditch. Recent research in medieval rural nucleated settlement in Ireland MERCANTILE TOWNS • Major D Lesser SMALL TOWNS A Castle Cores /:,. Ecclesiastical Cores .., Dual Cores .... .. ... RURAL -BOROUGHS o Rura~Boroughs 111 Rura~Castle-Boroughs e Rura~Eccleslastlca~Boroughs A Rural-Boroughs identified from pop. evidence .., Rural-Boroughs identified from payment of sporadic taxations • MARKET VILLAGES 0 80km l!ll2.!::! e Ecclesiastical Towns + Market VIllages Fig. 2.- Distribution map of Anglo-Norman nucleated settlements in Ireland (After Graham 1993). It was undoubtedly·the Anglo-Normans who introduced a full network of nucleated settlements into the two-thirds of Ireland that they conquered. Before they came in 1169-70 the only nucleated settlements were either the ports, mainly located along the east coast of the country, which were founded by the Scandinavians in the tenth and eleventh centuries or some of the larger monasteries such as Glendalough in County Wicklow whose wealth and ecclesiastical fame drew a large population concentration around them (Edwards 1990). This hierarchical network of nucleated settlements that the Anglo-Normans introduced into Ireland can best be seen in Graham's map (Fig. 2) where all types of Anglo-Norman nucleated settlement are shown, from the largest cities to the smallest villages. This paper will concentrate on these vil89 T.Barry !ages, but before discussing them further it is important to explain the term 'rural borough' which appears in the legend of the map. This was first used by Glasscock in 1970 to identify Anglo-Norman rural nucleated settlements in Ireland no larger than contemporary English villages but which were given borough status in Ireland. These borough charters were granted by the great Anglo-Norman nobles in order to attract settlers from the overcrowded regions of lowland Britain to settle their relatively underpopulated colony in Ireland (Glasscock 1970). Despite all the recent large-scale archaeological investigation of the Anglo-Norrnan levels of cities such as Dublin or Waterford there have been few archaeological excavations of Anglo-Norman deserted nucleated settlements in Ireland, especially of their villages (Barry 1994). Indeed, between Glasscock's investigation ofthe deserted manorial village ofLiathmore, County Tipperary in 1968-9 and that at Piperstown in Co. Louth in 1987 there has only been one other archaeological excavation of part of a possible deserted medieval village located beside Bourchier's Castle, Lough Gur, County Limerick. Here, Cleary excavated two medieval houses and four huts which may have also dated to the medieval period, which arguably were part of a village clustered around the thirteenth-century Fitzgerald Castle which was located under the later Bourchier' s Castle (Cleary 1983). At Liathmore, which has yet to be fully published, Glasscock found the base of a round tower of the earlier monastic site as well as further evidence of seventeenth-century settlement there that Leask and Macalister had also located during an earlier excavation. Disappointingly for him he was unable to find any evidence of medieval occupation on this important site (Glasscock 1970). Piperstown was chosen because it was the only known Deserted Medieval Village in County Louth, Ireland's smallest county, and because there had not been a modem excavation of such a settlement type in the eastern half of the country. It had not been recognised as a possible Deserted Medieval Village by the Archaeological Survey of Ireland until 1974 because it lacked an identifiable church site and because the motte was very small and irregular (Buckley & Sweetman 1991). There is, however, the ruins of the medieval church of St. Mary at Drumshallon located around 1/2 km to the south of the present village earthworks. This church might well have also functioned as the parish church of the village community of Piperstown. The village is located on the margins of high ground, some 300-400 metres OD, but it is surrounded by lower more agriculturally rich lands on all but its western side. It first enters the documentary record in 1316 when 14 cottages at 90 'Pippardeston' were 'burnt by the Irish' during the disorders which were attendant upon the Scottish Bruce Invasion of Ireland (1315-18). Thus it probably had its origins earlier in the previous century when the Pipards settled this part of Ireland in the decades following the Anglo-Norman Invasion (Duffy 1997, 99). Obviously the village had shrunk in size during the Middle Ages as today only four possible house platforms can be identified on the landscape (Fig. 1). Because of a general lack of funding only one season was completed, and only one house platform was fully excavated as well as one section which was cut through the westerly perimeter bank of the village. Excavation of the platform did produce evidence for a very tenuous late medieval or post-medieval house with external dimensions 8 m in length by 5.6 m wide. It was divided into two by an internal partition, with the inhabitants living in the western end and the animals on the eastern side of the partition, with a drain to take away their waste. The medieval finds included 29 sherds of medieval cooking ware as well as some metal artefacts of the same period. The medieval metal finds were a thirteenth century hunting arrow head as well as a Jews harp of possible medieval date. Two small knife blades, two nails and a small horseshoe were also found but they were not diagnostic enough to be dated precisely. The other metal finds were either unidentifiable or were of post-medieval date. Indeed the greatest number of pottery sherds, over 250 in all, came from the first half of the eighteenth century when the village was finally deserted. The fact that it was deserted in the post-medieval period is largely in line with what is suspected for many other village desertions in Ireland which were not deserted until after the fifteenth century, such as Liathmore, County Tipperary or Caherguillamore, County Limerick (Barry 1994). Generally speaking the artefacts recovered at Piperstown would indicate that the dwellers in this particular house did not enjoy a high standard of living because no sherds of imported medieval fine ware were found nor any medieval coins or fine quality metal objects, such as were located at Caherguillamore in County Limerick, when two medieval houses were excavated in the 1940s (6 Riordilin & Hunt 1942). Nevertheless, after the excavation was completed a number of fifteenth and sixteenth century coins were found on the site, an indicator perhaps of the thriving nature of the village during the later Middle Ages. The tenuous nature of the house, with its cheap and readily available building materials, would also suggest that its inhabitants were not very wealthy. Indeed it was probably very similar to the mud walled cabins that were still to be found in Recent research in medieval rural nucleated settlement in Ireland different parts of rural Ireland in the last century and even in the first few decades of this present century. It was a simple one-storied structure with few if any window openings, and where the smoke from the hearth would just have filtered out through its thatched roof. It was probably occupied at the end of the Middle Ages in Ireland although it was difficult to be entirely sure about the exact length of occupation because of the general lack of securely stratified dateable finds associated with the house's occupation levels. The general paucity of finds also meant that it was difficult to be sure about the socio-economic basis of the village community here as represented by the inhabitants of this particular house. It was probably significant that only a small number of animal bones were discovered, which may indicate that although animals were kept there that they were not such an important feature of their diet. Alternatively, the small number of animal bones recovered were in such poor condition that they could not be readily identified, so it is possible that the acidic nature of the soil had completely destroyed the remainder of these bones. The final possibility is that the animal bones may have been dumped some way from the house platform, and so were not detected during the excavation. Analysis of the drain which ran through the house produced some possible husks of grain seeds, while the low-lying lands surrounding the village on all but its western edge were probably under arable in the Middle Ages. It is likely that this was where the village's open-field system was located in the medieval period. As well as these cereals which probably were an important part of their diet the village's inhabitants obviously supplemented this by hunting deer and other game, as is shown by the hunting arrow head located during the excavation. Beyond these tantalisingly limited pieces of evidence we cannot be completely sure of this important aspect of everyday life in this medieval village. This probable desertion date is very different to that found in many areas of medieval England, especially in the Midlands, where the major desertion period seems to have been in the fifteenth and sixteenth centuries when arable farming gave way to an emphasis on pastoral farming in the changed economic conditions of the later Middle Ages (Beresford & Hurst 1971 ). In Ireland the major period of desertion would appear to have been in the seventeenth century when war and famine combined to finally destroy marginal settlements which had existed since the expansionary thirteenth century. These factors even affected important settlements such as the prosperous town ofNewtown Jerpoint which controlled a major bridge across the River Nore in the heartland of the Anglo-Norman colony in Leinster. At the height of its prosperity in the thirteenth century it had a population of at least 250 family members belonging to its governing class of burgesses, as well as from among its free tenants. It is arguable that its full population size was at least twice this figure because the surviving medieval documents do not record the numbers of people who did not own any land in the town. But by the later Middle Ages it had been limping along for a century or so after the dissolution of the nearby wealthy Cistercian abbey of Jerpoint in 15 3 8. This event meant the end of its important trading relationship with the religious community there which had started with the foundation of the town in 1200 AD, and which was central to the economic well-being of the town. It would appear that a severe plague in County Kilkenny in the first half of the seventeenth century, together with the presence of Cromwellian armies in the locality in the 1640s all led to its final abandonment . In this discussion of current research into the medieval rural nucleated settlement pattern it is important to note that one of the newest developments has been the attempt to examine the Gaelic Irish component in a more systematic way. The Discovery Programme Ltd. which was set up to encourage large-scale archaeological research projects in Ireland has just employed someone to produce a programme for future research on all types of medieval rural settlement in Ireland, including the heretofore elusive Gaelic settlements, both dispersed and nucleated. It has been determined by the Directors of the Discovery Programme that this important yet often neglected aspect of the rural settlement pattern be given priority in this investigation. Until recently almost all medieval rural settlement research has been focused upon the Anglo-Norman element within it. It is scarcely surprising that the concentration of research in this field up to the present has been on the AngloNormans because they left behind them many manorial and central government documents which give the modern researcher much additional useful information on their settlement pattern. The Gaelic Irish sources, on the other hand, are usually either literary or annalistic writings which are more difficult to interpret if we wish to understand their settlement forms in any detail (Duffy 1997). Secondly, the settlement forms introduced into Ireland by the Anglo-Normans have also been studied closely in the other parts of Europe that were controlled by them, so this gives an excellent basis for comparative research. But the new emphasis on Gaelic Ireland in future research may well help to rectify this imbalance in our contemporary knowledge of medieval Irish settlement generally. 91 T. Barry In conclusion, although nucleated settlements were usually the most prominent components of much of the rural settlement pattern within the Anglo-Norman colony it is arguable that there was also a greater number of dispersed settlements than was originally thought. For instance, the 750 or so moated sites, many of them the defended manor houses of the Anglo-Norman lesser nobility of the thirteenth and fourteenth centuries, were dispersed widely over the colony and especially on its periphery (Barry 1994, 84-93). There is also increasing evidence for the continued occupation of ringforts throughout the later Middle Ages, particularly in the Gaelic-Irish controlled areas of the north and west outside the control of the Anglo-Normans. Nevertheless, the Anglo-Norman's greatest and perhaps most enduring contribution to the settlement pattern of Ireland was the establishment of rural nucleated settlements all across the eastern half of the island. References BARRY T.B. 1981: Archaeological excavations at Dun beg promontory fort, County Kerry, 1977, Proceedings of the Royal Irish Academy 81 C 12, 295-329. BARRY T.B. 1994: The archaeology of medieval Ireland, London. BERESFORD M.W. & HURST J.G. (eds.) 1971: Deserted Medieval Villages: Studies, London. BRADLEY J. 1996: in: I. BENNETT (ed.), Excavations 1995, Dublin. BUCKLEY V.M. & SWEETMAN P.D. (eds.) 1991: Archaeological Survey of County Louth, Dublin. CLEARY R.M. 1983: Excavations at Lough Gur, Co. Limerick: part Ill, Journal of the Cork Historical and Archaeological Society 88, 51-80. DUFFY S. 1997: Ireland in the Middle Ages, Dublin. EDWARDS N. 1990: The Archaeology of Early Medieval Ireland, London. EVANS E.E. 1964: Ireland and Atlantic Europe', Geographische Zeitschrift 52, 224-41. GLASSCOCK R.E. 1970: Moated sites and deserted boroughs and villages: two neglected aspects of Anglo-Norman settlement in Ireland, in: N. STEPHENS & R.E. GLASSCOCK (eds.), Irish Geographical Studies, Belfast, 162-177. GRAHAM B.J. 1993: The High Middle Ages: c. 1100 to c.1350, in: B.J. GRAHAM & L.J. PROUDFOOT (eds.), An Historical Geography ofIreland, London, 58-98. 0 RiORDAIN S.P. & HUNT J. 1942: Medieval dwellings at Caherguillamore, County Limerick, Journal of the Royal Society of Antiquaries of Ireland 72, 37-63. PROUDFOOT V.B. 1959: Clachans in Ireland, Gwerin 2, 110-122. STOUT M. 1996: Early Christian settlement and society in Ireland with particular reference to ringforts, unpublished Ph.D. thesis, Trinity College Dublin. STOUT M. 1997: The Irish Ringfort, Dublin. Acknowledgement Professor Brian Graham of the University of Ulster, for permission to reproduce Figure 2. Abbreviation: OD = Ordnance Datum. Dr. Terry Barry Department of Medieval History Trinity College Dublin 2 Ireland 92 Art and Symbolism in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference· Volume 5 Marco Milanese, Fabrizio Benente & Franco Campus Il progetto Geridu. Indagini archeologiche in un villaggio medievale abbandonato della Sardegna 1 Introduzione L 'analisi della tematica dei villaggi abbandonati ha registrato, in Italia, importanti contributi a partire dagli anni Cinquanta e Sessanta, con lo sviluppo di ricerche a carattere prevalentemente regionale. Fanno eccezione alcune sintesi di piu ampio respiro, fra cui il noto saggio di C. Klapisch Zuber e di J. Day (Klapisch Zuber & Day 1965). Questa problematica, dopo la precoce attivita di alcune missioni straniere ne! Lazio, e stata recepita agli inizi degli anni Settanta dalla nascente archeologia medievale italiana, in un clima di aperto confronto fra archeologia e geografia storica. Esperienze metodologicamente mature ed importanti, come quella del Gruppo Ligure di Ricerca sulle Sedi Abbandonate (Quaini 1973,712-713, 737-744) e quella del G.R.A.M. di Palermo non sono tuttavia riuscite, per motivi diversi, a sviluppare le premesse. Scavi importanti come quello dei villaggi medievali di Zignago (Boato et alii 1990) e di Brucato (Pesez 1984) sono stati comunque portati a termine ed editi da gruppi di ricerca operanti in Liguria e Sicilia. E pero innegabile che si sia verificato, a partire dal decennia scorso, uno spostamento complessivo degli interessi degli archeologi medievisti italiani verso altre tematiche, come l'incastellamento e la transizione tra la tarda antichita e l'alto medioevo, mentre gli interrogativi storiografici fondamentali sui quali dovrebbe lavorare in modo sistematico l'archeologia del villaggio (diacronia dei modelli insediativi, continuita e rotture, cultura materiale, tempi, modalita e cause degli abbandoni) rimangono sostanzialmente irrisolti o non verificati. 2 I villaggi medievali abbandonati della Sardegna 2.1 La Sardegna, con oltre 800 villaggi documentati agli inizi del XIV secolo (Day 1988, 18), euna delle regioni italiane dove il problema dei villaggi medievali abbandonati si e posto all 'attenzione dei ricercatori con maggiore incisivita ed e stato piu volte segnalato come uno dei nodi interpretativi fondamentali di tutta la storia economica, demografica e sociale dell'isola in epoca tardo medievale (Tangheroni 1976, 244). La mobilita dell 'insediamento, una delle principali caratteristiche dell' abitato rurale sardo (Day 197 6, 204), ha determinate le rilevanti dimensioni quantitative che i1 fenomeno assume in questo territorio. Pur non essendo questa la sede per una trattazione sistematica di tale complessa tematica storiografica, si richiameranno almeno alcuni punti particolarmente significativi. 2.2 Gia alcuni storici sardi d'eta moderna descrivevano la presenza dei resti dei villaggi abbandonati nelle campagne sarde (Para 1586; Aleo 1677). L'esistenza di questi centri scomparsi non e mai uscita totalmente dal possibile campo di osservazione immediate di storici, geografi ed eruditi locali, per la consistenza dei resti sui terreno (es. Day 1984, 26), delle fonti scritte e di quelle orali (Day 197 6), nonche per la lunga durata del fenomeno degli abbandoni, che si protrae sino all' eta moderna ed all' eta contemporanea, secondo dinamiche registrate analiticamente dalle fonti statistiche (Corridore 1902). Stante quindi una continuita di riflessioni e studi, che ha visto contributi, talora fondamentali, anche da parte di ricercatori stranieri (Le Lannou 1941, 106 ss.), una stagione decisiva per lo studio dell'insediamento rurale della Sardegna e quella che, nei primi anni Settanta, si econcretizzata negli ormai classici atlanti dei villaggi abbandonati sardi (Dau 1973; Terrosu Asole 1974). 2.3 La spinta demografica verificatasi a seguito della colonizzazione rurale dei secoli XI e XII determina in questo periodo, in Sardegna, la nascita di quattordici nuove diocesi (Day 1984, 21) e fu connotata da grandi opere di dissodamento dell 'incolto, probabilmente gia completate entro la prima meta del XII secolo (Meloni 1994, 53). Il processo tendente alla dispersione della popolazione rurale sui territorio dovette entrare in crisi gia nel tardo XII secolo 93 ;vr. Milanese, F. Benente & F. Campus epoca alia quale possiamo datare, sulla base delle fonti scritte, i primi casi di abbandono (Day 1976 a, 230-231). I! momento "classico" degli abbandoni dei villaggi medievali della Sardegna, in cui si verifica la maggiore ondata del processo di spopolamento delle campagne, tuttavia costituito dai secoli XIV e XV: le fonti statistiche redatte dai re d' Aragona, anche sulla base di precedenti fonti fiscali pisane (Le Lannou 1941, 105), permettono di stimare che soltanto 352 dei circa 805 villaggi sardi noti attorno al 1320 erano sopravvissuti ne! 1485; risulta, quindi, una valutazione possibile di circa 453 sedi umane scornparse. Secondo J. Day (Day 1988, 18), anche nel XIV secola, I 'inurbamento continua ad essere motivo dello spopolamento dei villaggi, come potremo verificare oltre, anche per il caso specifico di Geridu. Per quanto riguarda la grande pestilenza della meta del secolo, possiamo controllarne gli effetti sulla base di un inventario del regno sardo (Bofarul 185 6) voluto da Pietro IV d' Aragona ne! 13 58: a questa sembra imputabile un calo della popolazione isolana stimato attomo a! 50%, mentre il numero dei villaggi cancellati in seguito a! passaggio del morbo non ricalca questa indicazione statistica . Su questo contesto demografico comunque gia largamente sofferente, si innestano alcune congiunture, che vengono spesso chiamate in causa per spiegare, fra gli ultimi decenni del XIV secolo ed il successivo, la scomparsa di oltre 300 villaggi sardi: nuove ondate di pestilenza (1376, 1398, 1404, 1410, 1424, 1476), carestie (1374, 1421), la guerra di liberazione condotta dai re d' Arborea dal 1365 a! 1420, e quella successiva del marchese di Oristano, terminata nel1478 con la battaglia di Macomer, che determinerebbe un'ulteriore fase di abbandoni (Day 1987a, 177; Day 1988, 19). e 3 11 villaggio medievale di Geridu (Sorso, Sassari) 3.1 L' area interessata dai resti del villaggio medievale di Geridu ubicata in comune di Sorso, lungo la Strada Provinciale n. 25 Sassari-Sorso (Fig. 1), a circa 2 Km da quest'ultimo centro, in direzione sud (I.G.M., 1:25.000, F.180, Ill, N.O., "S. Andrea"): la strada attuale interseca i resti sepolti del villaggio, che risulta pertanto spezzato in due vaste porzioni. La sola struttura del villaggio di Geridu sopravvissuta a! di sopra del piano di campagna attuale, costituita da una modesta porzione d'elevato della chiesa di S. Andrea, secondo una situazione "classica" del villaggio medievale, I' esistenza dei cui resti sepolti spesso segnalata, in aperta campagna, e e e 94 soltanto dalla presenza di una chiesa diruta o restaurata. Ancora agli inizi del XIX secolo, la chiesa si trovava in discrete condizioni, prima di essere utilizzata come cava di materiale da costruzione per la nuova chiesa parrocchiale del vicino centro di Sorso. Noto nelle fonti scritte a partire dai primi decenni del XII secolo, il villaggio di Geridu era ubicato ne! Giudicato di Torres (Logudoro), ne! territorio della curatoria di Romangia, sicuramente di limitata estensione (98,5 Kmq), ma forse la piu densamente popolata curatoria medievale di tutta la Sardegna, secondo i dati demografici desumibili da Day 1987b, 318. Dalle stesse statistiche s1 evince inoltre che attarno agli anni venti del Trecento, oltre a Geridu, vi erano almeno altri cinque villaggi nella stessa curatoria di Romangia: Sorso, Sennori, Taniga, Uruspe e Genor, mentre P!aiano, Domusnovas e Cleu risultano gia abbandonati. Dalle medesime fonti, Geridu si pone con chiarezza come il centro piu popolato della Romangia, certamente uno dei maggiori aggregati demici della Sardegna settentrionale, con una stima possibile, attomo a! 1321 (Le Lannou 1941, 105) o al 1323-4 (Terrosu Asole 1979, 32; Day 1987b, 318), di 326 uomini validi o fuochi, contro i 213 di Sorso, i 140 di Sennori, i 90 di Taniga e i 37 di Uruspe e di Genor. Sembrerebbe pertanto possibile ipotizzare un numero di abitanti oscillante tra 13 50 e 1600 ea; cia, tuttavia, con la massima prudenza dovuta alle note difficolta di interpretazione del valore demografico reale del singolo fuoco. Dopo lo sbarco catalano-aragonese in Sardegna nel1323, Geridu venne assegnato a!la citta di Sassari e, in seguito, variamente infeudato: questa situazione dovette avviare la lenta agonia demografica del villaggio, di cui abbiamo numerosi riscontri nelle fonti scritte, ed incoraggia, nell 'impossibilita di far fronte all a crescente pressione fiscale, I' inurbamento in direzione di Sassari. Ne! 1344 gli abitanti di Geridu lamentano l 'eccessivo carico fiscale, proprio a causa dell'emigrazione verso Sassari (Day 1987, nota 68; Day 1984a, 18). La convergenza di queste congiunture negative e della violenta pestilenza della meta del secolo, giustificano il drastico calo demografico registrato dal gia ricordato censimento catalano del 1358, secondo i1 quale a Geridu erano rimasti soltanto 70 uomini validi o fuochi (Day 1987c, 318). Geridu continua probabilmente a vivere ancora ne! 1391, quando venne infeudato a Galcerando de Santa Coloma (Casula 1977, 48-49), con le vicine ville di Taniga e Sorso. Secondo J. Day, I 'abbandono di Geridu sarebbe completato definitivamente entro il 1427 (Day 1973, 121, n. 218; Day 1987, nota 68), 11 progetto Geridu. Indagini archeologiche in un villaggio medievale abbandonato della Sardegna data in cui le terre di Geridu vennero annesse a quelle di Sassari. Poiche le direttrici di ricerca sinora sviluppate in tema di insediamento rurale medievale della Sardegna sono state fondate in modo pressoche esclusivo sulle fonti scritte, soltanto sporadici sono stati i tentativi di discussione sulla struttura materiale degli abitati (Meloni 1994), per la scarsita dei riferimenti documentari noti utili per una riflessione in questo sensa. Aprire il fronte della ricerca archeologica sistematica, cambiando radicalmente la natura delle fonti e la loro capacita informativa, pone la riflessione storiografica a diretto contatto con gli aspetti materiali dell a societa rurale medievale e con I' organizzazione spaziale e socio economica di una comunita di villaggio del medioevo della Sardegna. In questo sensa, con I' aspirazione di costruire un modello di insediamento (che, sec ondo tendenze sottolineate dalle ricerche europee ne! settore dei villaggi abbandonati, ci attendiamo soggetto a profonde trasformazioni diacroniche e non cristallizzato nella configurazione iniziale del si to) o piu semplicemente un caso da sottoporre a discussione, sono state avviate ne! 1995 le indagini nell'area del villaggio medievale abbandonato di Geridu, situato nel settore nord-occidentale dell' isola, in prossimita dell a citta di Sassari. 4 Le strategie dell'intervento archeologico 4.1 Le indagini archeologiche condotte a partire dal 199 5 nell' area del villaggio medievale di Geridu nascono ne! quadro complessivo della collaborazione tra la Soprintendenza Archeologica per le Province di Sassari e Nuoro e l'Universita degli Studi di Sassari e I 'Universita degli Studi di Genova (Cattedra di Archeologia Metodologia della Ricerca Archeologica), con l'appoggio del Comune di Sorso. La progettazione delle indagini archeologiche e stata concepita tenendo presenti i seguenti dati: 1. le informazioni gia acquisite grazie ad alcune indagini di emergenza condotte in diverse aree del sito dalla Soprintendenza Archeologica; 2. le informazioni stratigrafiche deducibili da alcune sezioni occasionali esposte da sbancamenti di enormi dimensioni per lavori agricoli o edilizi; 3. le osservazioni condotte nelle prime ricognizioni effettuate nel si to nell 'invemo-primavera 1995, che avevano evidenziato la vasta superficie probabilmente occupata dai resti del villaggio, in seguito valutata preliminarmente attomo ai 14 ettari di estenstone. Tutta la superficie di un'ampia particella catastale (133: mq. 7230), ubicata nei pressi della chiesa di Fig. 1. b~ 0 ::som "CC m Sant' Andrea e denominata area 3000, si presentava interessata da un' omogenea distesa di scaglie litiche, tegole frammentate e rari reperti ceramici, nonche da una micromorfologia della superficie che faceva ipotizzare la presenza di strutture sepolte sottostanti (fig. 1). I primi saggi eseguiti nel 199 5 hanno consentito di sviluppare in modo piu maturo la strategia complessiva della ricerca e di individuare con chiarezza gli obiettivi ed interrogativi fondamentali, finalizzati ad un utilizzo storiografico delle fonti archeologiche, che possiamo cosi riassumere: 1. Determinazione di una ipotesi di estensione dell'area occupata dal villaggio, da attuarsi con indagini di superficie, a causa del "costa zero" con il quale si sono dovute confrontare le prime campagne di scavo. 2. Valutazione della potenza dei depositi stratigrafici e delloro significate per la microstoria del sito. 3. Quando e in quali modi e stato abbandonato il villaggio di Geridu ? Si e trattato di un processo di lunga durata, articolato nel tempo ? 4. A quando risalgono le prime fasi del villaggio e come sono caratterizzate archeologicamente (cultura materiale, modi di costruire, etc.)? Esiste frattura o continuita con l'insediamento romano di eta imperiale certamente presente in una zona di Geridu ancora da riconoscere con precisione? 5. Qual'era l'organizzazione socioeconomica del villaggio e quali risvolti essa aveva nella topografia del sito tardo medievale? Esistevano quartieri produttivi distinti da quelli residenziali e la vicinanza alia chiesa costituiva un fattore di privilegio sociale? 95 M. Milanese, F. Benente & F. Campus 6. Qual' era il livello dell a cultura materiale del si to tardo medievale e delle tecniche di costruzione? La vicinanza alia costa e gli interessi, in questa zona, della famiglia genovese dei Doria, determinarono un rapporto mercantile privilegiato con la Liguria? Dopo i primi saggi e ricognizioni, svolte ne! 1995, due consistenti campagne di scavo sono state effettuate ne! 1996 (aprile, dicembre-gennaio 1997), ne! corso delle quali sono state aperte grandi aree (per una superficie di ea. 700 mq. in corso di scavo) che permettono di impostare le prime valutazioni sull'organizzazione spaziale dell' insediamento. Ne! dicembre 1996, grazie all 'adesione a! pro getto dell'antropologo G. Mullen, e stata aperta una grande area di scavo nei pressi della chiesa del villaggio, in una zona sicuramente interessata dalle necropoli tardo medievali del sito. Marco Milanese 3. disponibilita del proprietario del terreno a permettere I' occupazione ed i lavori di scavo. Nella prima fase d 'intervento (dicembre 1995) caratterizzata dall 'esigenza di una valutazione del potenziale stratigrafico dell 'area, si e proceduto all' apertura di due settori di scavo (31 00, 3500) di estensione limitata, con una superficie di 40 mq. ciascuno (5 metri allineamento sud-nord, 8 metri allineamento est-ovest), allineati e distanziati fra loro di 15 metri. Nelle due successive fasi di intervento (aprile 1996 e dicembre-gennnaio 1997), finalizzate ad una lettura estensiva deg!i aspetti stratigrafici e planimetrici degli edifici precedentemente individuati, i settori di scavo sono stati ampliati fino a raggiungere un'estensione di 182 mq. (settore 3100) e 162 mq. (settore 3500). Di seguito si presentano in maniera sintetica i risultati delle indagini condotte nei due settori e si procede ad una prima valutazione globale delle problematiche di scavo che sembrano caratterizzare !'area 3000. 5 I documenti stratigrafici. Lo scavo dell'area 3000 5.1 Settore 3100 (fig. 2) Nelle fasi preliminari del progetto di indagine archeologica del villaggio abbandonato di Geridu e stata adottata una gamma diversificata di strumenti conoscitivi, che hanno consentito una prima valutazione globale delle problematiche stratigrafiche e delle tematiche diacroniche del sito. In quest'ottica, all'analisi di archeologia di superficie ed alia lettura di stratigrafie verticali esposte da lavori di sterro, si eaffiancata I' analisi stratigrafica di due settori di una prima area campione (area 3000). Ne! quadro di tale progetto, quindi, I 'indagine stratigrafica dell 'area 3000 e, piu in generale, le prime due campagne di scavo condotte nei settori 3100 e 3500, hanno assolto a! compito di un prima approfondimento stratigrafico (dicembre 1995), cui ha fatto seguito una piu estesa analisi del tessuto insediativo precedentemente evidenziato (aprile 1996, dicembregennaio 1997). La scelta dell'area 3000 come prima area d'indagine e motivata da una serie di concause: 1. posizione centrale rispetto all' insediamento, prossima all' edificio di culto principale, la chiesa di Sant' Andrea, ma non cosi vicina da essere interessata dalla presenza di aree cimiteriali; 2. presenza in superficie di un'alta concentrazione di materiale fittile e materiale da costruzione (materiale lapideo e coppi delle coperture), inequivocabili segni di un'intensiva presenza antropica, associati a vistose anomalie del rilievo, caratterizzato da dossi ed avvallamenti che ne alterano il naturale degradare verso nord; 96 I! settore di scavo 3100 eubicato in posizione centrale rispetto alia particella catastale contrassegnata dal numero 133, a circa cinquanta metri dal limite della strada comunale Sorso-Sassari. Dopo una prima campagna di scavo ( dicembre 1995), volta alla valutazione del potenziale stratigrafico del settore e conclusasi con I' individuazione di due edifici (3000/1 e 3000/2), 1' area oggetto dell' analisi stratigrafica, con le successive campagne di scavo, e stata ampliata a! fine di individuare l'intero sviluppo planimetrico degli edifici e di indagame in estensione gli interni (fig. 2). Negli ultimi anni (genericamente "dieci-venti anni" secondo la fonte orale raccolta) il sito e stato oggetto di tentativi di aratura, attuati mediante un aratro a dischi che incide il terreno per una profondita media di 30 cm. circa. Questo fatto potrebbe giustificare la presenza del pietrame e dei coppi sull' intera superficie dell'area 3000 (ma anche su buona parte del sito). Durante la seconda campagna di scavo, inoltre, si eosservato che i terreni incolti ubicati nelle vicinanze e destinati a! pascolo, vengono periodicamente interessati da arature leggere. La rimozione dello strata di humus ha consentito di porre in luce su tutta !'area un suolo (3027) caratterizzato da una fitta presenza di frammenti di pietra calcarea in disfacimento, con superficie irregolare e pendenza da sud verso nord. Questo strata deve essere ricondotto a fasi recenti di utilizzo dell 'area, caratterizzate da tentativi di riduzione a coltura e di aratura. I1 progetto Geridu. Indagini archeologiche in un villaggio medievale abbandonato della Sardegna -i- + 0 I ! 0 I ' 0 ,, 30711 J 'i -i- Fig. 2. Nella parte nord del settore, lo strato 3027 copriva direttamente le rasature delle strutture murarie perimetrali dell' edificio 3000/1 ed il relativo strato di crollo intemo (us 3007), mentre nella parte sud occultava una serie di unita stratigrafiche (cfr. infra) for- 0~1~ matesi in seguito all a frequentazione dei resti dell' edificio 3000/2 dopo il suo crollo ed abbandono. Si presentano di seguito le sequenze documentate nei singoli edifici. 97 M. Milanese, F. Benente & F. Campus 5.2 Edificio 3000/1 5 .2.1 Questo primo edificio, localizzato nella parte nord del saggio, e delimitato dai muri 3004, 3005, 3033 e 3035, realizzati con un doppio filare contrapposto di bozze di calcare locale, con un riempimento centrale della cortina eseguito a sacco. In corrispondenza degli angolari si osserva una piu accurata disposizione delle singole pietre con utilizzo di pietre di maggiori dimensioni, talora squadrate. Il legante argilloso utilizzato epiuttosto povero, tanto da far apparire la muratura in alcune zone quasi completamente a secco. La faccia a vista intema delle murature conserva tracce di un intonaco, costituito da fango argilloso, steso direttamente sulla superficie delle murature forse con lo scopo di coibentare gli ambienti. Le strutture murarie che delimitano 1' edificio, ed in particolare 3035 (lato est) e 3005 (lato ovest) sono conservate in elevato, rispetto a! piano d'uso intemo, da un massimo di circa 8 corsi (verso sud) ad un massimo di 2 corsi (verso nord) e la loro rasatura o crollo sembra aver condizionato successivamente l'andamento dell'attuale piano di campagna, che declina appunto da sud verso nord. L'edificio era ad un solo piano (domus terrestris), aveva una pianta a sviluppo rettangolare ed una superficie intema di ca.43 mq. Presentava in origine una copertura a doppio spiovente con palo ligneo centrale alloggiato su un grosso concio dalla superficie spianata (us 3054). L'edificio aveva un piano d'uso intemo in terra battuta (us 3079) ed era suddiviso in due ambienti da una tramezza, costituita da due zoccoli in muratura (us 3066-3067) su cui si innestava un elevato in materiale deperibile. Una porta con cardine in pietra consentiva, all' intemo dell a casa, il passaggio tra i due ambienti. Il vano ovest, ad uso cucina e residenziale, risulta di ea. 22 mq. Sulla sinistra dell 'ingresso era presente un focolare in argilla concotta, con soletta rialzata rispetto a! piano d'uso. Nell'angolo tra i muri 3004 e 3005 vi era un'area ad uso dispensa testimoniata dalla presenza di almeno tre giare, trovate frammentate sui piano d 'uso e dal rinvenimento di grano ed altri reperti paleobotanici carbonizzati. In prossimita della porta intema era presente una cassa in legno di cui sono stati trovati gli elementi in ferro (serratura, un elemento angolare di sostegno e, quella che probabilmente era la maniglia ). Il secondo vano, cui si accedeva da un' am pia apertura posta sui fronte nord della casa, aveva uno sviluppo intemo di ea. 21 mq. La mancanza di reperti ceramici, la presenza di un ferro da muratura con gancio e di un elemento litico forato (in sardo sa loriga), originariamente inserito in parete allo scopo 98 di legarvi un animale, consente di ipotizzame una funzione di stalla per il ricovero di un numero limitato di capi di bestiame. Sui fronte nord dell' edificio si aprivano due distinti accessi ai due ambienti. L'accesso a! vano cucina avveniva tramite una piu stretta soglia (us 3064) di ea. 1,1 m. e il cardine ed altri elementi li tici so no stati trovati ne! crollo (3007) in prossimita dell' apertura. L'accesso a! secondo vano, in posizione centrale rispetto al muro 3035, avveniva attraverso una soglia in calcare finemente battuto (us 3090) della larghezza di ea. 1,2 mt. 5.2.2 L'edificio estato abbandonato repentinamente, forse a causa di un incendio che ha provocato anche il crollo del tetto. Gli oggetti metallici trovati in situ fanno pensare, anche a causa della loro dispozione caotica, ad un abbandono totale e repentino. La compresenza tra oggetti d'uso domestico, utensili agricoli (falcetto) e diverse armi, dovrebbe ricondurre l'incendio del tetto e la parziale distruzione dell' edificio ad un evento bellico, o comunque doloso ed intenzionale. Successivamente al crollo del tetto I' edificio non venne riparato, ma prima del crollo dei muri perimetrali I'intemo venne frequentato. I coppi sembrano infatti avere subito un'azione di calpestio e all'intemo vennero realizzate almeno due buche (us 30213023). Gli elementi di crollo dei muri perimetrali coprivano, inoltre, uno strato (us 30 15) costituito da limo argilloso di colore giallo distribuito prevalentemente lungo i muri 3004, 3005 e 3035, il cui limite sfuma a nord sulla sottostante us 3046, strato di crollo dei coppi di copertura dell' edificio. La formazione dello strato di limo puo essere dovuta a! dilavamento dellegante dei muri perimetrali, o all' eventuale dilavamento di un intonaco in argilla avvenuto dopo il crollo del tetto. La limitata estensione di us 3015 ed il suo sfumare su us 3046 avvalorerebbero questa ipotesi. 5.2.3 La deposizione dello strato di crollo (us 3007) e la formazione del successivo strato 3027 chiudono la sequenza stratigrafica relativa all' edificio 3000/1, che puo essere cosi periodizzata (Tav. 1). 5.3 Edifici 3000/2- 3000/3. 5.3 .1 Nella parte centrale del settore, gli edifici 3000/1 e 3000/2 erano separati da una intercapedine di ampiezza non superiore al metro, in cui, a! di sotto di uno strato di crollo erano presenti strati di terreno argilloso (us 3018 e 3062), che in fasi successive hanno costituito il piano di calpestio ne 11' intercapedine tra le due case. Il progetto Geridu. Indagini archeologiche in un villaggio medievale abbandonato della Sardegna Tav. 1. Per. fa se principali attivita cronologia I V b a b c d costruzione e vita dell' edificio 3000/1. costruzione del muro 3036, relativo ad un vano laterale. incendio, crollo del tetto, formazione di 3046. formazione dello strata limoso 3015 per dilavamento dei muri perimetrali. attivita di discarica su 3046, realizzazione delle buche 3021-3023. crollo dei perimetrali all'intemo dell'edificio- formazione di 3007. abbandono dell'area. aratura e tentative di utilizzo agricolo. XII -XIVsec. Ill II I - - Nella parte sud del settore 3100 il suolo 3030, tagliato dalla buca 3029, copriva uno strata di crollo esteso su tutta !'area (us 3003=3039), da cui emergevano le rasature dei muri 3006, 3013, 3044 e 3050, 3090 che insieme delimitano diverse fasi di utilizzo dell' edificio a pianta rettangolare 3000/2. L 'us 3030 si appoggiava, infine alla struttura muraria 3031, prossima allimite sud del settore, che dovrebbe costituire il perimetrale nord di un terzo edificio (3000/ 3) da indagarsi con la prosecuzione dello scavo. 5.3.2 L'edificio 3000/2, localizzato nella parte sud del saggio era delimitato in origine dai muri 3006, 3013, 3050 e 3090, realizzati con la medesima tecnica documentata per l'edifico 3000/1 (cfr. supra). La casa, ad un solo piano (domus terrestris) aveva pianta rettangolare e sviluppo planimetrico intemo di ea. 30 mq. L'indagine si efermata, ne! gennaio 1997 ad una fase di abbandono (us 3093) dell'originale piano d'uso che non consente ancora una lettura dell'articolazione intema della casa. L'edificio, forse dopo un crollo parziale (us 3080), sembra avere avuto una seconda fase di utilizzo che ha comportato, con la costruzione del muro 3044, una riduzione della superficie intema a mq. 20. In questa seconda fase una soglia (us 3051) realizzata nell' angolo trail muro 3044 ed i1 muro 3050 consentiva, da sud, l'accesso all'intemo dell'edificio. La soglia e stata pasta in opera tagliando in parte us 3051, ed era costituita da due lastre di pietra poste in verticale, con funzione di batti-porta e da un elemento litico con foro sagomato per I 'alloggiamento del cardine della porta posto all'estemo del muro 3050. 11 piano di calpestio all'intemo dell'ambiente era costituito dall 'interfaccia superiore dello strata di crollo 3080, caratterizzato da una marcata presenza di frammenti ceramici ed ossa animali. La mancanza di aree di fuoco, di partizioni interne e 1'estrema irregolarita della superficie d'uso rendono problematico ricondurre questa seconda fa se dell' edificio ad una destinazione d'uso abitativo. 1° meta XIV fine XIV XX sec 5.3.3 Ancora piu articolate e complesse sono le vicende che segnano il degrado progressivo e l'abbandono definitivo dell' edificio. Lo strata di crollo 3060, che segna l'abbandono della seconda fase d'utilizzo, era coperto da uno strata di terreno argilloso giallastro (us 304 7), tagliato a! centra da una grande buca (3049). E ipotizzabile che 3047 costituisca uno strata formatosi all 'intern a di 3000/2 per azione di dilavamento seguita all'abbandono dell' edificio e che si a stato interessato da successive attivita, quali, ad esempio, la realizzazione di una sequenza estremamente articolata di buche. Oltre alla buca centrale 3049, si e infatti individuata una buca (us-3017) che, realizzata dopa !'abbandana di 3000/2, ne ha tagliato l'angolo tra i muri ovest e sud. Le dimensioni ed i1 profilo della buca, la stratigrafia intema, formatasi per progressivo collasso delle pareti, fanno supporre un'originaria funzione di vase a per la raccolta dell' acqua. I1 riempimento che sigilla la buca (us 3034) conteneva maiolica arcaica pisana con decoro del I-IV gruppo (Berti & Cappelli 1994,218, 256-259; Berti & Tongiorgi 1977, 10°-12° gruppo) pertinenti all a terza fase produttiva pisana databile a partire dalla seconda meta del XIV secolo. Una terza buca di grandi dimensioni (us-3045), individuata in prossimita del limite sud del settore, intaccava i1 riempimento 3034 della buca us-3017, e presentava due strati di riempimento (us 3059 e us 3040). Tale buca, per le caratteristiche del suo riempimento (us 3040), si puo configurare come scarico di rifiuti domestici e residui di pasta. La sti-uttura muraria 3031, che costituisce illata nord dell 'edificio 3000/3 risulta realizzata sul riempimento 3040 di us-3045. Ne consegue che tale riempimento ne data post quem la realizzazione e deve essere segnalata in tal sensa, la presenza (in 3040) di un "alfonsino" minuto coniato ad Iglesias, emesso a! nome di Giacomo II d' Aragona e, quindi, entro il 1327. 99 M. Milanese, F. Benente & F. Campus 5.3.4 Un successivo strato di crollo (us 3039) segna il progressivo deterioramento di quanto rimaneva in elevato dell'edificio 3000/2 e, in prossimita dell'angolo tra i muri 3006 e 3013, tale strato era tagliato da una esile struttura in pietre (30 19), di forma rettangolare, realizzata in appoggio ai muri stessi per contenere una raccolta di coppi originariamente integri (us 3009). La presenza dei coppi, raccolti intenzionalmente e stoccati nell 'angolo di un edificio in abbandono, trova confronto diretto con la situazione documentata per l'edificio 4 del settore 3500 (cfr. infra). Nello strato di crollo era, infine, presente un avvallamento riempito da uno scarico di rifiuti domestici (us 3012), costituito da terreno ad alta percentuale di cenere, con frammenti di ceramica da fuoco e con un 'alta percentuale di residui faunistici con tracce evidenti di macellazione. 5.3 .5 La sequenza in questa parte del settore echiusa dalla formazione dello strato 3030 e potrebbe essere cosi periodizzata (Tav. 2). 5.3.6 Analizzando ora i dati desunti dallo scavo dei tre edifici individuati ne! settore 31 00 si puo dedurre che la dinamica del collasso delle strutture e dell' abbandono sembra aver avuto tempi differenziati. In particolare, 1'edificio 3000/1, non presenta 1' intensa fase di frequentazione post-abbandono che caratterizza invece 1'edificio 3000/2, costruito ne! periodo IV ed utilizzato nelle fasi a e b del periodo Ill come edificio ad uso promiscuo (ricovero animali ?), area di discarica e di parzia1e raccolta di materiale edile per un eventuale reimpiego. La presenza, nell 'edificio 3000/2, di uno strato di rifiuti (us 30 12) con reperti faunistici con vistose tracce di macellazione, frammisti a carboni e ad oggetti ceramici indicherebbe che, ne! periodo II, fase b, qualche edificio vicino, probabilmente l 'edificio 3000/3, era ancora abitato. Questa fase, sulla base di un primo esame dei materiali ceramici, con la presenza di maioliche arcaiche pisane con decorazioni a croce in verde e raggi in bruno (Berti & Cappelli 1994: decori I-IV gruppo), sembra delimitata alla seconda meta del XIV secolo. Tale cronologia coinciderebbe con il post quem indicato per la costruzione del muro 3031, perimetrale nord dell'edificio 3000/3. L'abbandono del villaggio non sarebbe quindi repentino ed assoluto, ma progressivo, con fasi differenziate di collasso degli edifici. Tale abbandono comporterebbe, inoltre una fitta fase di frequentazione dei crolli con reimpiego di strutture ancora parzialmente in elevato e raccolta di materiale edilizio per un eventuale riuso, correlata a tentativi di ripresa dell' attivita edilizia su cui ancora bisogna indagare. Fabrizio Benente 5.4 Settore 3500 (fig. 3) I1 settore 3500 (fig. pianta Geridu 00/02) e ubicato nella porzione settentrionale dell 'area 3000 ed ha attualmente un'estensione di 162 mq. Gli elementi determinanti che hanno portato a preferire questa zona per aprire il secondo settore di scavo sono del tutto simili a quelli del settore 3100, Tav. 2. Per. fa se IV III a b c II I 100 principali attivita cronologia castruziane e 1"fase dell' edificia 3000/2 2° fase di utilizza di 3000/2 castruziane del mura 3044 e della saglia 3051 frequentaziane della superficie di us 3080 cralla parziale e farmaziane di 3060 attivita successive all' abbandona di 3000/2 farmaziane di 3047 realizzaziane dell a buca 3017 realizzaziane della buca 3045 realizzazione della buca 3049 cralla di quanta ancora in elevato dei perimetrali di 3000/2 castruziane di 3000/3. deposita dell a strata di rifiuti 3012. farmaziane di 3030 realizzaziane della buca 3029 ab bandana aratura e tentativo di utilizza agricola XIII - 1o meta XIV sec. 1° meta XIV 2° meta XIV 2° meta XIV fine XIV XX sec. I! progetto Geridu. Indagini archeo1ogiche in un villaggio medieva1e abbandonato della Sardegna I l,. L I I I =·~1-- ~~~~ '"~~OB I I I I I 0 . \ O~df Y'>~J C)-1 ' '-1..! 0 0 0 8 o0 0 0 0 oo 3SB3 0 0 0 I I i I J515 AMB.o ,, anche se e opportuno segnalare che in questa zona 1' andamento naturale dell a collina, orientato da sud verso nord, appare piu regolare rispetto al settore 3100 (cfr. supra). Al di sotto dell 'humus e stato posto in luce uno strato di terreno abbastanza omogeneo di colore marrone scuro, ricco al suo intemo di frammenti ceramici, di coppi e scaglie litiche di piccole dimensioni esteso su tutto il settore (us 3501=3527) ed interpretabile come il prodotto delle attivita agricole condotte nell'area in tempi recenti. Immediatamente al disotto di tale unita stratigrafica e stato possibile documentare le rasature delle murature di quattro edifici (3000/4,5,6,7) che insieme alle interfacce superiori dei crolli intemi ed estemi agli ambienti venivano a costituire un livello omogeneo caratterizzato da evidenti segni di usura delle superfici ricollegabili alle sopraccennate attivita agricole. Ad un momento immediatamente precedente e invece da collegare la deposizione e 1'accumulo di focolari (us 3504, 3531, 3530) costituiti semplicemente da pietre poste in circolo, concotte sulle superfici, correlabili con una frequentazione occasionale ., . dell'area, che doveva presentarsi come una distesa di pietre e macerie con un sottilissimo strato d'humus. 5.4.1 Edificio 300014. Localizzato al centro del settore 3500 e al momento quello meglio documentato dal punto di vista planimetrico e stratigrafico. Obbiettivo raggiunto solo con tre ampliamenti del settore nelle due ultime campagne di scavo. Presenta forma rettangolare con una larghezza intema dim 6 (3506) e una lunghezza dim 11 (3514, 3507); le murature, che risultano orientate sull 'incrocio degli assi est-ovest- i lati corti- e nordsud, hanno uno spessore variabile tra i 65 e 63 cm e conservano solo pochi corsi: tre nellato meridionale e due in quello settentrionale. La tecnica costruttiva consiste in un doppio filare contrapposto di pietre di calcare locale, del tutto analoga a quella documentata nel settore 3100 (cfr. supra). L'edificio 3000/4, dopo l'asportazione degli strati di crollo, presentava nella sua parte centrale uno strato uniforme di frammenti di coppi posti di piatto (us 3559) interpretabile come il residuo della copertura. 101 M. Milanese, F. Benente & F. Campus Lo strata di coppi si presentava in realta come il frutto di tre azioni: la prima (us. 3505) rivolta allo stoccaggio in file parallele verticali di coppi integri, con 1' attenzione di porre la parte piu larga dei singoli pezzi in basso, concentrando i materiali in prossimita dell 'angolo sud/ovest (us 3505). La seconda caratterizzata da una disposizione caotica dei frammenti di grandi dimensioni, ancora utilizzabili, strata localizzato su buona parte del lata sud (us 3519=3546). Infine il livello dei coppi di piatto descritto in precedenza ed interpretabile anche come lo scarto diretto delle due precedenti, ed interessato successivamente da fenomeni post deposizionali di calpestio e assestamento. La formazione degli strati di coppi, sulla base dei reperti numismatici rinvenuti, e databile a partire dal secondo quarto del XIV secolo. La rimozione dei coppi ha consentito di porre in luce uno strata molto compatto e uniforme, di colore marrone scuro (us 3520=3543) e con presenza in superficie di numerose chiazze di carbone, interpretabile come illivello di battuto pavimentale dell'ambiente nell 'ultima fase di utilizzo domestico. L'analisi globale di tutti gli elementi emersi sul battuto consente gia una prima definizione dell' articolazione intema dell' edificio. Trattasi di una cas a terranea con copertura a doppio spiovente organizzata con una trave di colmo su cui poggiano i travetti, coperti a loro volta da uno strata di canne o di paglia su cui poggiavano direttamente le tegole, tutti elementi documentati nello scavo della porzione est dell'ambiente (us. 3578, 3580), la trave centrale esostenuta, oltre che sui lati brevi, su un palo alloggiato in un'apposita struttura in pietra (us 3587). L'accesso all' ambiente avveniva dal lato sud, in prossimita dell'angolo tra i muri 3506 e 3507, dove e presente un grosso blocco in calcare che fungeva probabilmente da battiporta. La parte ovest dell' edificio era destinata alla cucina ed alla prima conservazione dei cibi. Indicazioni in tal senso provengono dalla presenza di un focolare (us 3555), pasta in prossimita della soglia su una soletta di argilla concotta e dalla presenza di alcuni oggetti ceramici integri e da numerosi frammenti di ceramica acroma destinata al contatto diretto del fuoco; l'angolo nord-ovest, dove sono stati recuperati parti di anforacei, era destinato alla prima conservazione delle derrate o dell' acqua per I 'uso quotidiano. La zona est era suddivisa in due parti: l'angolo nordest racchiuso da una tramezza lignea (us 3565, 3567,3593) con un strata d'uso caratterizzato da una concentrazione di pietre di piccole dimensioni (us. 3560) su cui va a sfumare la strata nero descritto sopra (us. 3543) e forse destinato al ricovero di un animale. Nell'angolo sud-est si apre un taglio circolare (us 3581) profondo 1,5 m ea. interpretabile come pozzetto adi102 bito ad una conservazione stagionale delle derrate alimentari come formaggi o granaglie o legumi secchi. Lo scavo dei diversi riempimenti (us 3582, 3586) del pozzetto, costituiti da strati di terra mol to depurata ricca al suo intemo di semi carbonizzati, ha permesso di verificare in realta che il taglio ericavato nell'angolo di due murature ortogonali tra loro (us. 3563, 3574) palesemente pertinenti ad un edificio preesistente al 3000/4, e del quale solo attraverso le prossime campagne di scavi si potra chiarire la cronologia e I' organizzazione planimetrica rispetto agli altri edifici. 5.4.2 Edificio 300015. Su questo ambiente abbiamo al momento scarse informazioni a parte i limiti dell 'angolo sudlest costituito dai muri 3533 e 3509 coperti direttamente dallo strata di riporto agricolo, che occultava anche un possibile strata d'uso intemo (us. 3570). Le murature sono realizzate con la medesima tecnica costruttiva descritta in precedenza. Anche da questi pochi dati e comunque gia possibile trarre alcune considerazioni preliminari che prendono lo spunto dal fatto che il muro 3533 e allineato con quello dell'edificio 3000/ 4, e che tra le due case sia stato previsto uno spazio di rispetto (us 3515) destinato alla raccolta delle acque piovane provenienti dai due tetti adiacenti. 5.4.3 Edificio 300016. Il data piu significativo di questo edificio e che non rispetta ne gli spazi aperti esistenti tra gli edifici 3000/4 e 3000/5, ne l'indipendenza presente tra gli edifici appena descritti. Infatti il nuovo ambiente e realizzato allato di una costruzione precedente sfruttando alcune murature gia esistenti- porzione meridionale del muro 3506- e appoggiando un nuovo muro (usm 3516) ad una cortina preesistente (usm 3506). In questo modo la nuova costruzione sembra occupare quasi completamente uno spazio destinato, in un prima tempo, o al passaggio oppure a spazio aperto di servizio degli ambienti. Tutto cio permette di ipotizzare che il nuovo ambiente si tratti semplicemente o di un ampliamento connesso all' edificio 3000/4, oppure che la nuova struttura sia sorta in modo del tutto autonomo quando ormai si era perso completamente il significate di ordine e indipendenza tra le diverse abitazioni del villaggio. Inoltre la possibilita di costruire una sequenza di cronologia relativa tra gli edifici 3000/4 e 3000/6, il crollo di quest'ultimo edificio obliterava tutta la parte ovest dell a sequenza descritta per 1'ambiente 3000/4, permette di correlare ancora meglio le diverse fasi di vita tra i due ambienti. AI suo intemo e stato possibile documentare la presenza di un livello di terreno nero, ricco di frammenti di ossa e ceramica, interpretabile come un bat- 11 progetto Geridu. Indagini archeologiche in un villaggio medievale abbandonato della Sardegna Tav. 3. Per. fa se attivita principali cronologia V IV a a Costruzione del 1°edificio Rasatura delle murature del 1o edificio e costruzione 4, 5 dell' ambiente Vita dell'edificio 4, costruzione di un divisorio in materiale deperibile e apertura del pozzetto (us.3581) Abbandono edificio 4, incendio, crollo del tetto Costruzione e vita di 3000/6 Serie di attivita all' interne dell' ambiente 4 stoccaggio dei coppi integri, assestamento delle macerie, demolizione dell~ murature Crollo dell'ambiente 5 Occupazione temporanee, asportazione di materiale edile, focolari superficiali Formazione del terreno agricolo e utilizzo agricolo con arature meccaniche 2° meta XIII sec XIV sec. b Ill a b c II I 1° quarto XIV sec. 2° quarto sec. XIV 2° meta sec. XIV fine XIV XX sec. tuto pavimentale su cui poggia un focolare (us. 3577), non ancora indagato in maniera esaustiva. 5.5 Area 3000: analisi dei dati e prospettive d'indagine 5.4.4 Edificio 3000/7 Localizzato per ultimo nell' angolo sud est del settore, nel corso della campagna svolta nel dicembre 1996, e stata posto in luce solo una limitata porzione di un angolo (us 3589). L'unica annotazione possibile e che l'edificio ha il medesimo orientamento planimetrico degli edifici 3000/4 e 3000/5 e che risulta separato dal prima da una stretta stradina (us. 3573) del tutto analoga per ampiezza e orientamento con quella del settore 3100. La sequenza stratigrafica relativa al settore 3500 puo essere dunque cosi periodizzata (Tav. 3). Da un punto di vista strettamente cronologico, le fasi centrali offrono margini sufficienti di affidabilita, anche se devono essere indagati ancora i contesti pertinenti agli ambienti 6 e 7. 11 prima periodo connesso all a vita dell' edificio preesistente all' edificio 4 e al momento datato dal terminus ante quem costituito dall'abbandono di quest'ultimo. Gli ultimi periodi possono essere inseriti in archi cronologici abbastanza lunghi e con limiti di separazione alquanto sfumati tra loro: soprattutto per quanta riguarda 1'operazione di spoglio sistematico delle strutture murarie superstiti e la formazione dei focolari, azioni che possono essere iniziate anche in momenti immediatamente successivi al definitivo abbandono degli edifici e potrebbero essersi protratte sino alla prima meta dell' ottocento con la formazione progressiva del pascolo, 1' aratura e invece collocabile con certezza, grazie alle fonti orali raccolte, negli ultimi vent'anni di questo secolo. Franco Campus La periodizzazione, proposta congiuntamente e criticamente analizzata per i due settori di scavo, potra in futuro essere oggetto di migliorie e di modifiche sostanziali. Ancora di difficile collocazione, almeno nelle sue fasi iniziali e il periodo V registrato nel settore 3500, ma che comunque aggiunge gia nuovi elementi sulla complessita dell' evoluzione iniziale del villaggio. I limiti di separazione tra periodo IV e periodo Ill sono sicuramente meno netti, la distinzione in fasi potra essere meglio organizzata e correlata, con l'ampliamento dello scavo. La fase insediativa del periodo IV sembra essere pianificata e programmata, caratterizzata dall' adozione di moduli abitativi di grandi dimensioni (superiori ai 40 mq.), dall'utilizzo delle medesime tecniche costruttive, dalla presenza di una copertura in coppi. Gli edifici sono tutti orientati allo stesso modo (murature parallele agli assi nord-sud ed ovest-est) sono indipendenti tra loro, con presenza di spazi intramuranei per lo scolo delle acque. Questa organizzazione pianificata dell' insediamento sembra perdersi progressivamente, forse a partire dagli inizi del XIV secolo. Nella prima meta del secolo i due edifici del settore 3100 e quell a centrale del settore 3500 sono distrutti, 3000/1, 4 vengono definitivamente abbandonati, 3000/2, ridotto, e destinato ad usi non abitativi, uso che non puo essere escluso anche per 1' edificio 3000/6. Nelle ultime fasi del periodo Ill la crisi diventa piu marcata: gli edifici non crollano sincronicamente, ma aree in abbandono convivono con abitazioni ancora in uso e con case di nuova costruzione (3000/3 3000/6). Viene attuato lo 103 M. Milanese, F. Benente & F. Campus spoglio sistematico dei crolli, con fenomeni di raccolta e stoccaggio delle tegole di copertura ancora integre. Gli edifici in abbandono, onnai privi di copertura diventano spazi aperti destinati alia discarica dei rifiuti domestici e forse alia custodia degli animali. Ne! periodo II, probabilmente nella seconda meta del XIV secolo, avviene 1'abbandono definitivo dell' area, che vanifica tutte le precedenti attivita di stoccaggio e raccolta dei coppi, che rimangono accatastati senza poter essere riutilizzati, pur rimanendo per un lungo periodo in superficie e ben visibili. Attivita di frequentazione o di spoglio possono essere anche avvenute nel periodo I, contemporaneamente all 'attivita di demolizione e spoglio condotta sulla chiesa di Sant' Andrea nel XIX secolo. Sono attestate nell 'ultimo ventennio le attivita legate a tentativi di riconversione ad uso agricolo dell'area, tentativi che hanno come obbiettivo minimo di trasformare I 'area in pascolo invemale, considerato i1 sottile strato di humus a disposizione. I1 prosieguo dello scavo, 1'ampliamento e il collegamento dei settori 3100 e 3500 dovrebbero consentire I' acquisizione dei dati cronologici mancanti su11' avvio della fase insediativa medievale mentre l'analisi dei contesti pertinenti a! periodi IV e V consentira una piu chiara definizione dell 'articolazione intema degli edifici. Rimane ancora del tutto aperto il problema dell'organizzazione della viabilita intema al villaggio e della presenza di eventuali spazi aperti nel tessuto insediativo, ma lo scavo de11' area 3000 dovrebbe costituire un campione valido, almeno per I 'area centrale dell 'insediamento del villaggio. Fabrizio Benente 6 I reperti 6.1 Premessa La notevole quantita di reperti emersi nelle prime campagne di scavo, la cui analisi e ancora in corso, non pennette di presentare in questa sede una riflessione sui rapporti quantitativi intercorrenti tra le diverse classi di materiali. Tuttavia, pur rimandando questo obbiettivo ad una fase piu matura della ricerca, si ritiene utile presentare una prima analisi volta a focalizzare i limiti del campione presentato, si sottolinea che esso estato composto con materiali provenienti prevalentemente da contesti di crollo, caratterizzati da reperti in fase e residuali, mentre i piu attendibili contesti chiusi attribuibili a! XIV secolo non sono stati sinora presi in esame. Alcuni reperti residuali emersi da strati di pieno XIV secolo sono attribuibili al XII-XIII secolo e pro104 pongono prime infonnazioni su questa fascia cronologica, per la quale non disponiamo sinora di alcun contesto chiuso. I reperti metallici ed il vetro, sono stati volutamente esclusi da questa relazione preliminare e saranno pertanto presentati nell 'edizione dello scavo, corredati dei dati di associazione. Marco Milanese 6.2 Quadri tipologici e circolazione delle ceramiche La Sardegna, a partire dall 'XI secolo, e a! centro delle rotte commerciali tirreniche e mediterranee, ideali vettori di diffusione delle ceramiche prodotte in Sicilia e ne! Nord Africa e, a partire dal XIII secola, provenienti da Savona, da Pisa e dall 'Italia meridionale (Campania, Puglia). Per quanto riguarda i contesti architettonici (Berti, Hobart & Porcella 1990, 153-168; Hobart & Porcella 1993, 139-160) i quadri tipologici ed i fenomeni di circolazioni sembrano essere ben delineati, mentre per quanto conceme i contesti di consumo sembra ci siano ancora ampie possibilita di approfondimento. I1 quadro tipologico delle produzioni ceramiche emerso da questa prima fase d'indagine documenta 1' ampia circolazione di tipi ceramici importati dalla Liguria (graffita arcaica savonese, protomaiolica savonese, ingobbiata monocroma), dalla Toscana (maiolica arcaica pisana, boccali privi di rivestimento depurati), dall 'Italia meridionale tirrenica (spiral ware), dalla Spagna (giare e grandi contenitori, vasellame da mensa di produzione catalana, ispano moresche) e dal Nord Africa (ceramica smaltata e decorata a cobalto e manganese, invetriate monocrome). Le produzioni locali sembrano, invece, destinate a coprire il fabbisogno di vasellame da cucina e di altri tipi funzionali. L'attestazione di ceramiche smaltate decorate a cobalto e manganese, di produzione tunisina, con ceramic a campana tipo "spiral ware", associate a produzioni savonesi duecentesche (ingubbiata monocroma, graffita arcaica tirrenica) trova confronti con situazioni gia note in Sardegna, con reperti rinvenuti nel cagliaritano (AA.VV. 1993a, 32) e con "bacini" documentati nei contesti architettonici (Berti, Hobart & Porcella 1990, 153-168; Hobart & Porcella 1993, 148-149). Per quanto riguarda la Sardegna Settentrionale, le attestazioni di San Nicola e Santa Barbara a Sassari, dove le produzioni liguri sono associate con prodotti dell 'Italia Meridionale (Santa Barbara) e con ceramiche islamiche (San Nicola), costituiscono il confronto piu eclatante per i materiali rinvenuti a Geridu. I1 progetto Geridu. Indagini archeologiche in un villaggio medievale abbandonato della Sardegna Si propongono di seguito alcuni spunti di indagine emersi da un primo esame dei materiali: 6.2.1 Liguria Accanto alla circolazione della graffita arcaica tirrenica, che e ormai fenomeno ben noto e su cui sembra inopportuno soffermarci in questa sede (Hobart & Porcella 1993, 149; Berti & Cappelli 1994, 151-168; Varaldo 1995) e particolarmente significativa l'attestazione delle produzioni savonesi piu rare e marginali, come 1'ingobbiata chiara e monocroma, ma soprattutto le graffite arcaiche tirreniche monocrome e la protomaiolica savonese. Queste tipologie, ben note nei contesti liguri di XIII secolo (Andora, Savona, Casteldelfino, Genova, San Fruttuoso di Camogli), sono di solito scarsamente attestate in termini di esportazione. La graffita arcaica tirrenica monocroma, ovvero una tipologia con caratteristiche morfologiche e decorative assimilabili a quelle della graffita arcaica tirrenica, ma priva della decorazione dipinta in verde e giallo, e attestata negli scavi di Caste! Delfino (Milanese 1982a, 90, fig. 4 e 7; Milanese 1982b, 87, n. 26), documentata negli scavi di San Fruttuoso di Capodimonte (ex inf. A. Gardini), nota da reperti da scavo genovesi (Anfrews & Pringle 1977, 123, n. 86) e savonesi ed e anche attestata da bacini della chiesa di S. Cecilia a Pisa (Berti & Tongiorgi 1981,tav. CCXIII, n. 354; tav. CCXIV, n. 333; Berti 1993, 126), di San Francesco a Lucca (Berti & Cappelli 1994, 157, n. 12, fig. 118) e da bacini della chiesa di Santa Marta a Novara (Cortellazzo & Panto 1993, 35-36, 45, fig. 9).Trattasi di una produzione a circolazione piu limitata e a carattere minore, avviata da qualche atelier savonese nel corso del duecento, esauritasi probabilmente agli inizi del XIV, senza alcun rapporta di continuita o filiazione con la piu tarda graffita monocroma ligure (Benente, Gardini & Sfrecola 1993, 21; Benente 1996a, 253). Sono stati rinvenuti almeno due frammenti di protomaiolica savonese, uno dei quali trova confronti con la decorazione attestata da un frammento proveniente dagli scavi del Priamar a Savona (Varaldo 1990, fig. 2, csd 10376, fig. 6). Entrambi sono, infine, confrontabili con un bacino pisano della chiesa di San Giovannino, che reca un volatile dipinto in bruno e verde al centro del cavetto (Berti & Tongiorgi 1981, tav. CLXI, n.626). Risulta infine significativa 1'estrema scars ita delle piu importanti produzioni savonesi della seconda meta del XIV secolo e del XV secolo (maiolica arcaica, graffita monocroma) (Benente 1991; Benente 1996a; Benente, Gardini & Sfrecola 1993); scarsita da correlarsi con la cronologia proposta per l'abbandono del villaggio. 6.2.2 Toscana Sono attestati alcuni frammenti riconducibili alla classe delle brocche e boccali nudi, ad impasto depurato, di tipo pisano, recentemente oggetto di una revisione globale (Berti & Gelichi 1995, 191-240). A questa classe di oggetti non deve essere associata la funzione di contenitori da trasporto, ma piu probabilmente quella di contenitori da dispensa per liquidi o per aridi. La loro presenza e la loro circolazione, sovente in associazione con maiolica arcaica pisana, trovano confronti in area alto-tirrenica, in centri con approvvigionamento ceramico eterodiretto (Berti & Gelichi 1995, 237). La maiolica arcaica di produzione pisana risulta ben documentata ed in particolare sembrano avere buona circolazione le forme chiuse, sia i prodotti a piede svasato duecenteschi, sia le produzioni della prima meta del Trecento. Tale presenza rimanda a modelli di circolazione gia documentati in molti centri di consumo alto-tirrenici in contesti del XIIIprima meta XIV secolo. Si fa riferimento, in particolare, all 'associazione tra le forme aperte di produzione savonese e le forme chiuse di produzione pisana, la cui attestazione in diverse realta regionali (Benente & Gardini-Sfrecola 1993, 15-21; AA.VV. 1993a, 15-22; AA.VV. 1989, 87-88, tav.V) andrebbe riletta in uno studio volto alla definizione delle pratiche di commercializzazione del vasellame da mensa tra Toscana, Liguria, Provenza, Sardegna e Corsica. 6.2.3 Italia Meridionale Gli scavi di Geridu fomiscono testimonianza della circolazione, come ceramica d'uso, di produzioni del tipo spiral ware, gia attestate in Sardegna nei contesti architettonici (Hobart & Porcella 1993, 149-150) e prodotte in diversi centri dell 'area campano-laziale tra la fine del XII ed il XIII secolo (Molinari 1990, 362). La spiral ware, ben documentata a Genova (Palazzo Ducale) e a Savona (Priamar), nota a Pisa (Hobart & Porcella 1993, 149, nota 57) ampiamente attestata nell'Italia meridionale (Fontana 1984, 121122; Molinari 1990, 362-363), rinvenuta anche nel Nord Africa (Vitelli & Riley 1979, 96-101) e nel Mediterraneo orientale (Pringle 1984, 460), costituisce una delle classi ceramiche ad ampia diffusiotie, il cui vettore di circolazione potrebbe essere stato la rotta tirrenica e mediterranea delle navi genovesi. 6.2.4 Islam Occidentale Non essendo ancora stati analizzati i contesti iniziali d'uso degli edifici scavati (cfr. supra), soltanto pochi frammenti, in giacitura residuale, rimandano a ceramiche d'importazione islamica, prodotte in Tunisia, nel Maghreb e nella Spagna. 105 M. Milanese, F. Benente & F. Campus Trattasi di un frammento di ceramica smaltata decorata a cobalto e manganese, di alcune invetriate monocrome e di un frammento di giara decorata a stampo la cui presenza e cronologia possono essere ricondotte alle prime fasi di vita del villaggio di Geridu. mente connessa allo sfruttamento agricolo della zona, a meglio definime almeno i1 quadro cronologico, che si puo preliminarmente delineare fra I e V-VI secolo d.C., sulla base di materiale residuo presente nei contesti medievali. 6.2.5 Spagna Accanto alle tipologie ceramiche duecentesche sembra inserirsi, con la conquista catalano-aragonese, la circolazione delle ceramiche da mensa e dei grandi contenitori da trasporto e da dispensa di produzione spagnola (Francovich & Gelichi 1984, 28-39). Ceramiche smaltate dipinte in bruno e verde, prodotte in Catalogna (Telese Compte 1992, 93-100) o, piu probabilmente, a Paterna (Paz Soler 1992, 1322) sembrano avere buona diffusione in questa fase, trovando confronti con le piu !imitate attestazioni provenzali (AA.VV. 1989, 87, fig. 61, tav. V, nn. 38) e liguri (Mannoni 1975, 107, tipo 83, nn. 2, 3). Le ceramiche decorate in blu o in blue lustro compaiono nei tipi databili entro il XIV secolo, mentre sono assenti i prodotti valenzani del XV secolo. La presenza di queste ceramiche deve essere correlata alla presenza catalana in Sardegna e, piu in generale, ai traffici commerciali dei mercanti catalani con Provenza, Liguria, Sicilia e Mediterraneo Orientale (Castellaccio 1983). Fabrizio Benente 7.4 L'interrogativo riguardante i tempi e i modi dell 'abbandono trova prime risposte nelle sequenze documentate nello scavo dell 'area 3000, che indicano che i depositi archeologici di Geridu sono altamente informativi per 1'interpretazione delle attivita di spoglio, dei processi di abbandono e di crollo degli edifici, grazie all 'articolata pluristratificazione esaminata in precedenza che ha perrnesso di evidenziare fenomeni anche molto particolari, come il recupero e 1' immagazzinamento di materiale da costruzione .. I documenti archeologici sinora analizzati testimoniano con chiarezza che lo spopolamento di Geridu si verifico gradualmente attraverso tutto il XIV secolo: il campione esaminato sottolinea pertanto la lunga durata di questo fenomeno e ne caratterizza analiticamente la dinamica, attivata da una serie di concause (fra cui l'eccessiva pressione fiscale e le pestilenze dell a meta del secolo ), gia discusse sulla base dei dati storiografici, che perrnettono di seguire il calo demografico e la scomparsa dell' insediamento fra la fine del XIV e gli inizi del XV secolo. 7 Conclusioni 7.1 I dati sin qui discussi hanno gia consentito conclusioni parziali su differenti tematiche, ma sembra comunque opportuno discutere sinteticamente in quale misura le prime campagne di scavo nel villaggio di Geridu abbiano portato risposte agli interrogativi iniziali della ricerca. 7.2 I1 problema dell' estensione del villaggio medievale estato affrontato con campagne di ricognizione, che, nonostante i limiti esposti, hanno consentito di ipotizzare che il sito di Geridu possa occupare una superficie di circa 12-14 ettari. I1 ricorso a differenti metodologie di prospezione ed alla fotointerpretazione aerea, sinora impediti dal gia ricordato costo zero di queste fasi iniziali della ricerca, portera ad una piu esatta formulazione di questa risposta e ad una sua piu soddisfacente articolazione qualitativa. 7.3 Sui problema dell a continuita fra il si to romano certamente presente nell 'area di Geridu ed il villaggio medievale, le campagne di scavo e ricognizione hanno contribuito, oltre a confermare la gia indiziata esistenza di una frequentazione romana probabil106 7.5 I dati sinora raccolti permettono prime risposte all 'interrogativo concemente 1'organizzazione spaziale del villaggio, stante il gia evidenziato problema della sua estensione topografica. Sono stati ad oggi evidenziati i resti di 15 edifici (7 nell'area 3000; 4 nell'area 5000; 1 nell'area 2000 e 3 nei saggi di tutela della Soprintendenza), tutti riferibili alla fase due-trecentesca del sito, con la sola eccezione dell' edificio 5000/3. Anche se i1 campi one dell 'insediamento sui quale si sono sin ora concentrate le ricerche e forzatamente limitato, gli edifici della fase tardo-duecentesca e trecentesca posti in luce dalle campagne di scavo sembrano rispondere ad una programmazione generale dell' organizzazione spaziale del villaggio e dell' orientamento degli edifici, probabilmente funzionale a contrastare il forte vento di maestrale che spira particolarrnente nei mesi piu freddi. Le case sono suddivise tra loro da passaggi di diverse dimensioni talvolta cosi stretti da non consentire neppure il transito di una singola persona e sembrano, pertanto, esclusivamente funzionali ad una raccolta delle acque piovane e forse ad una parcellizzazione dello spazio fiscale (Bucaille & Pesez 1980, 77). Le case di Geridu presentano moduli rettangolari, con superfici interne utilizzabili variabili da 30 a 43 Il progetto Geridu. Indagini archeologiche in un villaggio medievale abbandonato della Sardegna a 66 mq. La copertura ea doppio spiovente con palo centrale destinato a reggere il trave maestro del tetto. Tutti gli edifici sinora scoperti sembrano rispondere con chiarezza alla tipologia di domus terrestris, limitata quindi al solo piano terreno, quale e stata evidenziata negli studi regionali svolti sulle fonti scritte. I1 caso dell 'edificio 3000/1, datato al primo quarto del XIV secolo, e particolarmente informativo per un'analisi funzionale degli spazi intemi della casa del villaggio: naturalmente non sappiamo ancora quanta questo esempio .possa essere rappresentativo di tendenze piu generali, infatti altri edifici in corso di scavo come il 3000/4, invitano alla massima prudenza in questa fase delle indagini. Le evidenze disponibili suggeriscono tuttavia 1' appartenenza dell' edificio 3000/1 al noto modello della maison mixte (es. Chapelot & Fossier 1980, 228): esso risulta infatti suddiviso in due ambienti, grazie ad un tramezzo precario forse di canne, legno o vimini (quale quello attestato in una casa mista quattrocentesca romagnola, che separava la stalla dalla camera da letto: Librenti & Zanarini 1991, 41/43) certamente dotato di una porta in legno atta a separare la stalla dalla cucina. I due ambienti, comunicanti tra loro all 'intemo dell 'abitazione, erano tuttavia muniti anche di autonomi accessi dall 'estemo, maggiore quello della stalla, minore quello della cucina. Per soste brevi o per i periodi caratterizzati da clima migliore 1'asino, sicuro ospite della stalla domestica, era legato a sa loriga, un blocco litico murato nella facciata dell'edificio e forato per consentire di leg are 1' animale all' estemo dell a casa. La cucina presenta un focolare soprarelevato in argilla, sulla sinistra dell'ingresso, mentre un angolo del vano era destinato a dispensa, testimoniata da alcune anfore contenenti fave selvatiche e grano carbonizzato. Questi resti ci rimandano alle attivita economic he legate allo sfruttamento dei campi, testimoniate anche da alcuni falcetti per la mietitura. Aderente al tramezzo precario chiare tracce rimandano ad uno dei piu emblematici elementi dell 'arredo della casa contadina medievale, la cassa in legno che non manca mai neppure nelle case piu povere (Bresc 1976, 114/127) e che nella citta di Sassari era invece spesso importata da Pisa e da Venezia, nella piu ricca versione dipinta. La casa 3000/1, distrutta da un incendio nel secondo quarto del XIV secolo e pertanto un documento archeologico di particolare interesse per la ricostruzione antropologica: essa si presenta di !imitate dimensioni, probabilmente buia e piena di fumo (Bucaille & Pesez 1980, 80), "povera di cose e affollata di uomini", con una cucina polifunzionale che "accoglier ... spesso alla rinfusa gli arredi che le competono e in piu attrezzi agricoli, scorte granarie, giacigli, e nello stesso tempo servira alla famiglia per mangiare, riscaldarsi, dormire, lavorare" (Mazzi 1980, 143). Le strutture murarie, per quanto lo scavo abbia sinora potuto evidenziare, sono costituite da una doppia cortina di pietre calcaree locali sbozzate, cavate seguendo i piani naturali di sfaldamento della roccia e il riempimento intemo e formato da scaglie litiche di lavorazione. I principali materiali da costruzione, il calcare tenero elveziano e 1' argilla, sono sta ti chiaramente reperiti in loco, cosi come sembra ragionevole ipotizzare (anche se mancano ancora i riscontri archeometrici) per le teule, i coppi di copertura presenti nel sito in enormi quantita. A questo proposito, occorre ricordare che il vicino villaggio di Bosove disponeva nel XII secolo di un proprio teularius, (Meloni & Dessi Fulgheri 1994, 164-165), figura la cui esistenza anche a Geridu e ampiamente supportata dall' evidenza archeologica: queste osservazioni su fonti di differente natura suggeriscono un mutamento di prospettiva rispetto alle ricostruzioni dei sistemi di copertura delle case rurali medievali di questa zona della Sardegna elaborate sulla base delle fonti scritte secondo le quali le coperture avrebbero di norma utilizzato materiale vegetale deperibile e, solo raramente, le teule in terracotta (Meloni & Dessi Fulgheri 1994, 51-52). 7.6 Le prime osservazioni riguardanti la cultura materiale di Geridu hanno permesso di apprezzare la vivac ita dei rapporti mercantili che gravitano nell' area di Sassari fra XIII e XIV secolo. Il quadro delle importazioni si presenta infatti particolarmente articolato ed ha evidenziato nella Toscana (Pisa), nella Liguria (Savona) e nell 'area catalana le direttrici quantitativamente privilegiate delle merci ceramic he presenti nel sito. A tale proposito si deve sottolineare il rilievo che le fonti e la letteratura storiografica attribuiscono alla presenza genovese e ligure nel Logudoro, particolarmente nell 'ultimo ventennio del XIII secolo (es. Loddo Canepa 1952, 49 ss.), a conclusione di un lungo periodo di tensioni e scontri con Pi sa, che ebbe in Porto Torres ed in Sassari importanti basi operative di carattere mercantile (Deliperi 1931, 3 ss.). La presenza commerciale ligure in questa zona della Sardegna trove nei mercanti genovesi i suoi piu attivi riferimenti, grazie anche ai vasti possessi fondiari, nel nord-ovest dell' isola, dell a famiglia Doria (Deliperi 1935, 6; Artizzu 1985, 147; Meloni 1990), saldamente attestata nelle vicine curatorie di Anglona e della Nurra ed in altre piu meridionali. Proprio all'iniziativa dei Doria si deve l'incastellamento del punto strategico di Caste! Genovese (Zirolia 1899), I 'attuale Castelsardo, porto di rilievo sulla costa set107 M. Milanese, F. Benente & F. Campus tentrionale. Tuttavia, forse ad una presenza diretta di mercanti savonesi in questo porto ed a Sassari, documentata a partire dal XIV secolo (Varaldo 1981, 349 ss.), e probabilmente da riferire il ritrovamento, a Geridu e nella stessa citta di Sassari, di prodotti ceramici di Savona. Informazioni puntuali per la ricostruzione dei sistemi economici e del paesaggio sono gia emersi dalle prime indagini sui resti faunistici da contesti chiusi tardotrecenteschi: una base quantitativa particolarmente solida ha consentito di individuare attendibilmente una dieta camea assai varia, basata principalmente su uno sviluppato allevamento di ovicaprini per macellazione e per la produzione di latte e lana, mentre asini e bovini di piccola taglia risultano adulti e quindi utilizzati per lavoro. La presenza di daini e cervi suggerisce infine 1'esistenza di vasti boschi e, quindi, un paesaggio di Geridu medievale assai diverso da quello attuale ( caratterizzato da oliveti e vigneti) e che soltanto un adeguato programma archeobotanico potra definire con un soddisfacente livello di analiticita. Marco Milanese Bibliografia Albisola = Atti del Convegno intemazionale sulla ceramica. AA. VV. 1965: Villages desertes et histoire economique, Xle-XVIIIe siecles, Paris. AA.VV. 1993a: Un gout d 'Ita lie, Aubagne. ALEO G. 1677: Succesos generales de la Is/ay Reyno de Sardefia, Cagliari. ANDREWS D. & PRINGLE D. 1977: Lo scavo dell'area sud del convento di San Sivestro a Genova (19711976), Archeologia Medievale IV, 47-207. ARTIZZU F. 1985: La Sardegna pisana e genovese, Sassari. BENENTE F., GARDINI A. & SFRECOLA S. 1993: Ligurian Tablewares ofthe 13th to 15th Centuries. New Archaeological and Thin Section Data, Medieval Ceramics 17, 13-23. BENENTE F. 1996a: Graffita monocroma, in: La citta ritrovata. Archeologia urbana a Genova 19841994, a cura diP. MELLI, Genova, 252-253. BENENTE F. 1996b: Maiolica arcaica, in: La citta ritrovata. Archeologia urbana a Genova 19841994, a cura diP. MELLI, Genova, 248-249. BENENTE F. 1991: Note sulla maiolica arcaica a Savona e in Liguria tra XV e XVI secolo, Albisola XXIV, 91-108. BERTI G. 1993: Materiali post-classici, in: Pisa. Piazza Dante, a cura di S. BRUNI, Pontedera, 00. 108 BERTI G. & CAPPELLI L. 1994: Ceramiche medievali e post-medievali (Museo Nazionale di Villa Guinigi), I. Dalle ceramiche islamiche alle "Maioliche arcaiche" Secc. XI-XV, Ricerche di archeologia altomedievale e medievale 19-20, Firenze. BERTI G. & GELICHI S. 1995: Le "anforette" pisane: note su un contenitore in ceramica tardo-medievale, Archeologia Medievale XXII, 191-240. BERTI G., HOBART & PORCELLA M. F. 1990: Protomaioliche in Sardegna, Albisola XXIII, 152-168. BERTI G. & TONGIORGI L. 1977: Ceramica Pisana. Sec. XIII-XV, Pisa. BERTI G. & TONGIORGI L. 1981: I bacini ceramici medievali delle chiese di Pisa, Roma. BOATO A. et alii 1990: Scavo dell'area est del villaggio abbandonato di Monte Zignago. Zignago 4, Archeologia Medievale XVII, 355-410. BOFARULL Y MASCARO P. 1856: Repartimientos de Ios reinos de Mallorca, Valencia y Cerdefia, Collecion de documentos ineditos del Archivo de la Corona de Aragon, tomo XI, Barcelona. BUCAILLE R & PESEZ J.M. 1980: L'habitat paysan en Bourgogne viticole du XIVe au XIXe siecle. Approche antropologique, Archeologia Medievale VII, 73-82. BRESC H. 1976: L'habitat medieval en Sicile, in: Atti del Colloquia Internazionale di Archeologia Medievale, Palermo, 186-197. BRESC G. & H. 1976: La casa del "borgese": materiali per una etnografia storica della Sicilia, Quaderni Storici 31, 110-129. CASTELLACCIO A. 1983, Aspetti di storia italo-cata/ana, Sassari. CASULA F.C. 1977: Carte Reali Diplomatiche di Giovanni I it Cacciatore, re d 'Aragona, riguardanti l 'It alia, Padova. CHAPELOT J. & FOSSIER R. 1980: Le village et la maison au Moyen-Age, Hachette. CORRIDORE F. 1902: Storia documentata del/a popolazione di Sardegna, (1479-1901), Torino. CORTELLAZZO M. & PANTO G. 1993: Bacini in Piemonte, Albisola XXVI, 31-50. DAY J. 1973: Villaggi abbandonati in Sardegna dal trecento al settecento. Inventario, Paris. DAY J. 1976: Villaggi abbandonati e tradizione orale: il caso sardo, Archeologia Medievale Ill, 203-239. ·DAY J. 1976a: L'insediamento precario in Sardegna nei secoli XII-XVIII, in: Atti del Colloquia Internazionale di Archeologia Medievale, Palermo, 228-242. DAY J. 1984: La Sardegna e i suoi dominatori dal secolo XI al secolo XIV, in: DAY J., ANATRA B. & SCARAFFIA L., La Sardegna medioevale e moderna, Torino. I! progetto Geridu. Indagini archeologiche in un villaggio medievale abbandonato della Sardegna DAY J. 1987: Uomini e terre nella Sardegna coloniale. XII-XVIII secolo, Torino. DAY J. 1987a: La ristrutturazione demografica della Sardegna nei secoli XIV-XV, (trad.it.) in: Day 1987. DAY J. 1987b: I. Dati demografici 1316-1485, in: Day 1987. DAY J. 1987c: Quanti erano i sardi nei secoli XIVXV?, (trad.it.) in: Day 1987. DAY J. 1987d: Malthus smentito? Sottopopolamento cronico e calamita demografiche in Sardegna nel tardo medioevo, (trad.it.) in: Day 1987. DAY J. 1988: Gli uomini e il territorio: i grandi orientamenti del popolamento sardo dall'XI al XVIII secolo, in: Storia dei Sardi e della Sardegna, vol. II, 13-48. DELIPERI N. 1931: Lo sviluppo del commercia sardo ne !la prima meta del secolo XIII, Sassari. DELIPERI N. 1935: Ordinamenti mercantile tributari in Sardegna prima della conquista aragonese, Sassari. FARA G.F. 1580: Opera I-III, a cura diE. CADONI, Sassari, 1992. FARA G. F. 1586: De Corographia Sardiniae, (manoscritto del1586), Cagliari (1835). FONTANA M.V. 1984: La ceramica invetriata al piombo di San Lorenzo Maggiore, in: La ceramica medievale di San Lorenzo Maggiore in Napoli, a cura di FONTANA M.V. & VENTRONE VASSALLO G., Napoli, 49-176. FRANCOVICH R. & GELICHI S. 1984: La ceramica spagnola in Toscana ne/ bassomedioevo, Quaderni dell 'Insegnamento di Archeologia Medievale della Facolta di Lettere e Filosofia dell 'Universita di Siena 3, Firenze. HOBART M. & PORCELLA M.F. 1993: Bacini ceramici in Sardegna, Albisola XXVI, 139-160. KLAPISCH-ZUBER C. & DAY J. 1965: Villages desertes en Italie, in: Villages desertes et histoire economique, XIe-XVIIIe siecles, Paris. LE LANNOU M. 1941: Pastori e contadini di Sardegna, (3 ed. italiana, 1992), Cagliari. LIBRENTI M. & ZANARINI M. 1991: Strutture materiali e forme insediative nel territorio bolognese in eta medievale, in: Archeologia e Insediamento ntrale in Emilia Romagna ne/ Medioevo, a cura di S. GELICHI, Bologna, 23-206. LODDO CANEPA F. 1952: La Sardegna attraverso i secoli, Torino. MANNONI T. 1975: La ceramica medievale a Genova e nella Liguria, Bordighera. MAZZI M.S. 1980: Arredi e masserizie della casa rurale nelle campagne fiorentine del XV secolo, Archeologia Medievale VII, 137-152. MELON! G. 1990: Insediamento umano nella Sardegna settentrionale. Possedimenti dei Doria alla meta del XIV secolo, in: Atti del XIV Congresso di Storia dell a Corona d 'Aragona, SassariAlghero. MELON! G. 1994: La Sardegna rurale in un importante documento del XII secolo: liberi, servi, commercia e potere nel mondo giudicale di una societa isolana, in: MELON! G. & DEssl FULGHERI A., Mondo rurale e Sardegna del XII secolo, Napoli, 13-122. MELON! G. & DESSi FULGHERI A. 1994: Mondo rurale e Sardegna del XII secolo, Napoli. MILANESE M. 1982a: Alcune problematiche della ceramica savonese della prima meta del XIII secolo alla luce delle acquisizioni dello scavo di Caste! Delfino, Albisola XV, 89-104. MILANESE M. 1982b: Lo scavo archeologico di Caste! Delfino, Archeologia Medievale IX, 74-114. MOLINARI A. 1990: Le ceramiche rivestite bassomedievali, in Crypta Balbi 5. L 'esedra della Crypta Balbi ne! Medioevo (XI-XV secolo), a cura di L. SAGUI & L. PAROL!, Roma, 357-484. PAZ SOLER M. 1992: Valencia, in: AA.VV., Mediterraneum, Viterbo, 11-68. PESEZ J.M. 1973: Ricerche e prospettive di lavoro intomo ai villaggi abbandonati, Quaderni Storici 24, 767-806. PESEZ J.M. (dir.) 1984: Brucato. Histoire et archeologie d'un habitat medieval en Sicile, Coli. Ecole Fran9aise de Rome 78, Roma. PRINGLE D. 1984: Pottery as evidence for trade in the Crusader States, in: I comuni italiani ne! Regno Crociato di Gerusalemme, Genova, 451- 47 5. QUAINI M. 1973: Geografia storica o storia sociale del popolamento rurale?, Quaderni Storici 24, 691-744. TANGHERONI M. 1976: Archeologia e storia in Sardegna. Topografia e tipologia. Alcune riflessioni, in: Atti del Colloquia Internazionale di Archeologia Medievale, Palermo, 243-250. TELESE COMPTE A. 1992: Catalogna, in AA.VV., Mediterraneum, Viterbo, 90-120. TERROSU ASOLE A. 1974: L'insediamento umano medioevale e i centri abbandonati fra il secolo XIV ed il secolo XVII, in: At/ante del/a Sardegna, Roma, 50-51. TERROSU ASOLE A. 1979: La nascita di abitati in Sardegna dall 'Alto Medioevo ai nostri giorni, Suppl. al fasc. II dell' At/ante dell a Sardegna, a cura di A. TERROSU ASOLE & R. PRACCHI, Roma. VARALDO C. 1990: Protomaiolica e imitazioni negli scavi del Savonese, Albisola XXIII, 69-78. VARALDO C. 1995: La graffita arcaica tirrenica, in: Actes du VI Congres International sur la ceramique medievale en Mediterranee, Aix-en-Provence 1995 (c. d. s.). 109 M. Milanese, F. Benente & F. Campus VITELLI G. & RILEY J.A. 1979: Medieval Spiral Ware from Carthage, Papers of the British School at Rome XL VII, Roma, 96-101. Nota storica intorno a Caste! Genovese e all'epoca degli statuti di Galeotto D 'Oria, Sassari. ZIROLIA G. 1899: Marco Milanese Universita degli Studi di Genova Dipartimento di Archeologia e Filologia Classica & Universita degli Studi di Sassari Dipartimento di Storia Fabrizio Benente Universita degli Studi di Pisa Scuola di Specializzazione in Archeologia Franco Campus Universita degli Studi di Roma "La Sapienza" Scuola di Specializzazione in Archeologia 110 Art and Symbolism in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 5 M 0Angels Ruf, Toni Vila, Xavier Soli, Xavier Llovera, Cristina Yaiiez & Josep Maria Bosch Le site du Roe d'Enclar (Andorra): evolution des strategies d'occupation et exploitation d'un territoire du Veme a la fin du VIIIeme siecle Introduction Avec cette communication, le Service charge de la recherche (Servei de Recerea Historiea) du Patrimoni Cultural d 'Andorra presente une synthese des resultats obtenus apres onze ans de campagnes de fouilles clans le site du Roe d'Enclar (1979-1993). Ce travail de recherche archeologique a surtout apporte des informations sur une de ces vallees pyreneennes orientales meconnues pour les periodes du BasEmpire et pre-feodale (IVeme-IXeme siecles). Etant donne 1' etat de la recherche, cette etude est exceptionnelle. Actuellement, les recherches sont en train de concentrer leurs efforts sur ces siecles de transition. Elles sont toutefois surtout centrees sur les zones proches des cotes. L'occupation des hautes vallees pyreneennes du Tet, le Tech, 1'Aude, 1'Ariege, le Fluvia, le Segre, le Valira, et la Noguera Pallaresa est encore tres peu connue par rapport aux regions proches des centres urbains plus anciens comme Narbo, Emporiae, ou Gerunda. Dans ce sens, il faut signaler le travail de synthese sur la romanisation clans le Nord-Est de la Catalogne (Casas et al. 1995) ou les etudes micro-regionales de la region de Narbonne dirigees par Favory et Fiches (1994). L'etude de notre site nous a amene a dresser un etat des connaissances dans toutes les vallees avoisinantes, aussi bien sur le versant nord que sur le versant sud de cette partie orientale des Pyrenees. Dans cet ensemble, la region de la Cerdagne est parmi celles ou la romanisation a ete un peu etudiee (Rico 1992; Olesti 1990). Mais pour ce qui conceme la transition entre 1'Antiquite tardive et la periode prefeodale, le Roe d'Enclar constitue le seul site connu clans une vaste region eloignee des grands centres politiques de !'Empire romain, et puis du regne wisigothique de Tolede. Devant !'importance du site nous avons veille a appliquer une approche interdisciplinaire pour obtenir un maximum d'informations des objets et des structures exhumees. Les datations au 14 C, les analyses des mortiers, des ceramiques et des scories de metaux sont autant de donnees qui ont permis de restituer !'occupation du site. Ce travail a ete publit~e en 1997 sous le titre: Roe d 'Enclar. Transformacions d'un espai domina! (segles IVXIX), dans la collection de monographies du Patrimoni Cultural d' Andorra. Situation geographique Situe au coeur des Pyrenees, 1' Andorre est un pays extremement montagneux dont 1' altitude moyenne est de 2.000 m. L'orographie s'organise autour d'une grande vallee centrale orientee Nord-Sud creusee par la riviere Valira. Perche sur un plateau a une altitude de 1.225 m, le site du Roe d 'Enclar domine la plaine centrale. Depuis cet emplacement, le site veille ainsi sur 1' entree sud du pays qui se fait etroite a cet endroit. I1 foumit en outre une bonne visibilite vers l'interieur du pays. Cette situation met le site a proximite d'un vaste reseau de communications fluviales. Le Valira qui coule aces pieds se jette ensuite, en-dehors des frontieres andorranes, clans la riviere du Segre. Le cours de celui-ci permet de relier la plaine meridionale de Lleida avec le versant sud des Pyrenees, puis a travers les ports de montagne il etablit un lien avec les vallees qui menent a Toulouse par 1' Ariege et a la Mediterranee par les vallees de 1'Aude et de la Tet. L'evolution du site Le site du Roe d'Enclar presente cinq moments d 'occupation all ant de la protohistoire jusqu 'au XIXeme siecle. Dans un premier temps, le site connalt une occupation protohistorique (entre 1' Age du Bronze ancien et 1' Age du Fer) dont temoignent surtout la ceramique et les objets lithiques, notamment des polissoirs et des haches. La periode d'occupation suivante se divise en deux phases. La premiere se situe pendant le processus simultane d'iberisation et de romanisation de ces vallees. Cette occupation n' a pas laisse de 111 M.A. Ruf, T. Vila, X. Sole, X. Llovera, C. YMiez & J.M. Bosch Segle IV l '\ "--'------" Zona ' "-----._ -----------. de Feixes 'I ........ ..... ' \ , ... \. .... , . '\ t ,. / I 1.,"' I I I I I I ' ' ' .. - ..... ,. - I ' .,.. .. ,. I I , / .' \ ' ' Zona de Feixes ',> I I I I \ \ \ I\ 1 \ J lll\)l i ..._\\\\Jifl \ )I \ ~ ( ,' ' I : I f \ I Camf '. principal ~: I I ''- ,' 1 \ ' ,' ' ,. • Camf principal ' ' : '. . '" ' ' ' I I I I I I I I ' I lnstal.lacions per a l'elaboraci6 del vi Zona de Feixes N {) Zona de Feixes ~ o 10 Fig. 1.- Roe d'Enclar: plan general des vestiges de la deuxieme phase (!Verne siecle) (re/eve: Alex Net). 112 20m Le site du Roe d'Enclar (Andorra) vestiges et on ne peut, de ce fait, definir de fa<;on precise que! put etre 1'usage du site. La deuxieme phase a lieu au IVeme siecle et se distingue par la transformation du Roe d 'Enclar en un espace voue a la culture de la vigne (fig. 1). La troisieme grande etape correspond aun moment de transition entre le mode romain et la periode feodale (fin du IVeme-IXeme siecle). Nous pouvons anouveau observer deux phases clans cette periode de !'occupation. Dans un premier temps, on construit une fortification sur la partie haute du plateau. Sa fonction etait de control er les lieux de passages NordSud traversant la vallee d' Andorre et d' am~ter ainsi les eventuels mouvements de peuples germaniques vers 1'Hispanie. Le second moment est marque par la formation, puis la consolidation d'un petit d'habitat perche (fig. 3 et 6). L' avant-demiere etape co'incide avec 1' avenement du feodalisme clans ces vallees pyreneennes. L'ancienne fortification tardo-romaine d'Enclar devient alors un chateau sous 1' emprise du comte (IXemeXIIIeme siecle). La cinquieme etape correspond a une longue periode pendant laquelle les autorites communales recuperent le Roe d'Enclar comme zone de paturage (XIIIeme-XIXeme siecles). L'exploitation viti-vinicole (IVeme siecle) Au cours du IVeme siecle, le pui du Roe d'Enclar est voue ala mono-culture du vin et ason elaboration. Il est tres probablement rattache un habitat plus important place dans la plaine en contrebas du rocher, dans la vallee centrale. Le site constitue alors un exem- a ple du modele d'exploitation viti-vinicole propre au Bas-Empire (IVeme siecle): la culture et la fabrication du vin se font au meme endroit alors que !'habitat est place ailleurs. Plusieurs elements nous ont amene a definir cette phase d'occupation. Les vestiges muratLx de la zone vouee a la culture, le fait d'avoir trouve le poids du pressoir, une cuvette revetue d 'opus signinum et les restes de vitis vinifere sont autant d'indices de la vocation agraire du site acette epoque. Les stntctures de culture No us avons decouvert tout d' abord les murs qui temoignent de 1'exploitation agraire. Il s 'agit de tout un systeme de terrasses allongees et etroites tres adequates pour la culture de la vigne. La superposition des structures murales et !'absence de references documentaires ad'autres activites agraires du IXeme au XXeme siecle nous amene penser que ces structures sont a her a l'etape du Bas-Empire que connait le site. a Le poids du press air A ces structures s'ajoute la presence d'une partie significative de 1' outillage qui servait a presser le raisin. Il s'agit d'un bloc de granit (larg. 80 cm, long. 86 cm, hauteur 39-56 cm) dont les cotes opposes sont creuses par deux rainures verticales d 'une trentaine de centimetres de long (L. 34 cm, 1. 9 cm et L. 43 cm, 1. 14, prof. 8 cm), et dont le centre possede une concavite circulaire (diam. 16 cm et profondeur 5 cm). Le poids du rocher est de 550 ou 600 kg (fig. 2). Le petit bass in Le second vestige qui nous indique que la production de vin se faisait sur place est un petit bassin qua- Fig. 2.- Le poids du pressoir (a droite) et reconstruction du systeme d'ancrage de la vis de pressoir ala pierre (a gauche) (dessin Josep M. Bosch). I I 113 M.A. Ruf, T. Vila, X. Sole, X. Llovera, C. Yariez & J.M. Bosch drangulaire (1. 180 cm, L. 230 cm) bati en pierre et impermeabilise par un double revetement fait de deux types d' opus signinum: I' un recouvre les parois, I' autre le fond du bassin. Ce bassin dont la capacite est de 49 litres etait utilise pour recueillir le mo(lt du pressoir. Les operations liees a 1'ecrasement du raisin et a la fermentation du mout avaient lieu dans un autre reservoir dont nous avons observe les restes du cordon hydraulique et du revetement en opus signinwn. Le site du Roe d'Enclar appara1t comme faisant partie d'un ensemble visant a exploiter le territoire. Les caracteristiques des structures agraires trouvees sur le site nous revelent une exploitation avec des elements de culture romaine. Dans ce modele d'exploitation, !'elaboration du vin se fait a l'endroit meme ou il est cultive, et loin du lieu de residence; il existe un mode le oppose dans lequel le lieu d' elaboration du vin est aussi celui de la residence, a I' ecart de la zone ou est cultivee la vigne. Le castellum d'Enclar (fin du IV erne-Veme siecle) La recherche archeologique sur le haut-plateau du Roe d'Enclar a permis de decouvrir un ensemble de pieces metalliques correspondant a un uniforme militaire et un armement dont la chronologie se situe du milieu du IVeme siecle au milieu du Veme. A cela s' ajoute un ensemble de structures murales dont les caracteristiques constructives et la disposition sont significatives. Tout ceci nous porte a considerer que nous avons affaire aux vestiges d'une infrastructure militaire de 1'Antiquite tardive. Nous pouvons la mettre en relation avec une politique generale destinee a convertir la limite provinciale romaine entre I 'Hispanie et la Gaule en une barriere pour arreter les eventuels mouvements de peuplades germaniques vers le Sud de 1'Empire. Les temoins directs ou indirects de la volonte de faire des Pyrenees une frontiere sont les claustra des Pyrenees occidentales, mentionnes au cours de la lutte de Didymus et de Veranianus contre Cons tan tin Ill (408) (Velazquez 1991 ), et, dans le Pyrenees orientales les castella de Caucoliberi (Cotlliure, Rosse116), de Vultaria (Oltera, Rossel16 ), les deux qui se trouvent a 1'en droit appele dans les textes Clausuras (Les Cluses, Vallespir), celui de Sardonia (Opol, Rosse116) et enfin celui de Lybiae (Liibia, Cerdanya) (Barbero & Vigil 1984). Les bronzes Nous avons decouvert un ensemble de bronzes hispano-romains dates de la fin du IVeme siecle et du Veme siecle. Certains d'entre eux sont a vocation exclusivement militaire, ce qui renforcerait la fonction du site a cette epoque. Il s'agit, par exemple, de 114 boutons indispensables pour unir les differentes parties en cuir des equipements de monture. Ils sont propres aux cavaliers hispano-romains: type IIIc (grand) (fig. 5/2), type IIId (petit) (fig. 5/3). Un passant de ceinture tres rare dans la region dans laquelle nous nous situons, constitue un des elements remarquables du site (fig. 5/6). C' est 1'objet militaire par excellence du Veme siecle, les paralleles les plus proches se rencontrent dans le Nord de la Gaule, a 1'Est du Rhin, en Germanie, Belgique, dans I' a ire danubienne, en Illyricum et ponctuellement en Hispanie et dans le Nord de 1'Afrique. Dans ces regions, il est date entre les annees 380-420 (fin IVeme-debut Veme). Dans notre cas, il est bien evidemment plus tardif (Rovira 1997). Sa presence confirme le caractere militaire de 1' occupation de ce moment. Les structures militaires L'etude exhaustive des mortiers des murs du site a permis de distinguer differents groupes. L'un d'entre eux est apparu tres clairement, il s' agit de 1'opus signinum dont nous avons parle plus ha ut. Un second groupe qui se detache nous a permis de mettre en relation deux structures quadrangulaires qui se distinguent par leur construction. La premiere se trouve juste a cote du chemin d' acces principal a 1'Ouest du site (T4) et la deuxieme est situee au point le plus au Nord du piton, pres d'un second acces (T2) (fig. 3). Une troisieme structure dont le mortier est different mais que nous situons dans la meme logique, est constituee de deux murs en angle; elle est placee au Sud a un point avec une tres bonne visibilite. Cette structure est amortie par la construction de 1, eglise a la fin du VIIeme siecle (T3). A!'Est, le flanc du rocher est inaccessible de par sa topographie meme. I1 faut remarquer le soin mis dans la construction de ces structures; il est exceptionnel dans 1'ensemble du site. Leur disposition sur le plateau (aussi bien leur relation stratigraphique avec les amenagements posterieurs, que leur emplacement aux endroits les plus vulnerables du relief) et leur morphologie nous amenent a penser a une fonction de tours de guet. De par leur disposition, il est fort probable qu 'il y eut un mur perimetral qui les ait uni en renforyant leur role militaire. Nous n 'avons pas pu fouiller ce contour de fayon a confirmer cette hypothese de travail. Malgre cela, de par la situation strategique du roe her a 1'entree de la vallee principale, et etant donne les troubles qui surgissent dans les regions voisines a cette epoque, il nous semble que le site du Roe d'Enclar a pu etre choisi comme point militaire. I1 nous semble qu'il est possible qu'il s'agisse d'un castellum. Il est difficile de determiner le degre d'importance du site dans 1'organisation du territoire etablie par les romains dans cette partie interieure des Pyrenees, Le site du Roe d'Enclar (Andorra) 'I ' ' I ' '' ' ', ' ' Segles V-IX Zona de Feixes I .. ~ ' ' 'I ' ' ' ' ' ' ' ' ,.' ' # ... .. .. .. ~ ' I '.• ' '' .......... ' ' '" ' ' ' ' Camr principal ' ',_. ' ' : '. 1 Zona de Feixes \ ' .' 1 I '\'\\\) ', ' \ ,1\\)\ :----..\\\ \J(I I ' ' .\)I ' ' ' I Camf · principal / ' '' ( I r \\( D t2 I ' ' ' ' ' ' '' ',f=l t4 ' ' ' ' ' N \ Zona de Feixes -EL) ~ I I\ :1 I ~\ \\ \\ \ ~ ~ L___jl___J ''--- o 10 20m ' Fig. 3.- Roe d'Enclar: plan general des vestiges de la troisieme phase (fin du !Veme-!Xeme siecle) (releve: Alex Net). 115 M.A. Ruf, T. Vila, X. Sole, X. Llovera, C. Yaiiez & J.M. Bosch Fig. 4.- Quelques exemples de la vaisselle de table composee de sigiltees estampiltees paleochretiennes (dessin Julio Ariza). etant donne le manque de sites connus pour cette periode. En tout cas, nous considerons que le castellum d 'Enclar est plutot lie aux autres claustra batis pour preserver 1'Hispanie de la penetration des peuples germains. Occupation du site de hauteur (Veme-IXeme siecle) Le materiel importe qui accompagne cette occupation constitue un ensemble exceptionnel compose de ceramique fine, de bronzes et de verre, pour lequel nous ne disposons pas d'exemples comparables dans les zones voisines. Il constitue done une reference obligee pour !'etude de 1' Antiquite Tardive dans cette partie des Pyrenees. Le lot de materiel correspondant aux activites domestiques du site est compose en grande partie de vaisselle de table, puis de recipients destines au magasinage. La vaisselle de table est surtout composee de sigillees estampillees paleochretiennes, aussi bien cuite en atmosphere reductrice qu' oxydante. De par leur typologie, elles se rattachent 116 majoritairement au groupe de la Narbonnaise. La principale caracteristique de cet ensemble est la variete aussi bien pour ce qui est des pates que pour leur aspect. Le riche eventail d'estampilles et leur distribution sur les pieces rappellent les ceramiques languedociennes. Quelques unes seulement sont proches de la production proven9ale. Leur datation se situe au Veme siecle. Le second type de vaisselle de table est moins represente. 11 s 'agit de vaisselle d' origine africaine, la sigillee D, des formes Hayes 61/ Lamboglia 54 datant de la periode 325-400/420. Nous disposons aussi d'une lampe a huile d'imitation de type Atlante X/Hayes IIB. Sa chronologie vades annees 450 a 600 (fig. 5). Les ceramiques africaines ont surtout servi pour le stockage. 11 s'agit d'amphores nordafricaines tardives, certaines allant jusqu'a la fin du VIeme siecle. Le numeraire de cette etape est repn!sente par des monnaies dont la date d'emission est a situer entre le milieu du Illeme siecle et le milieu du IVeme. Mais ils ont pu etre utilises au cours du VIeme siecle (Gomis 1997). A la vue de !'ensemble de ce materiel: ceramiques, bronzes, verres, on peut etablir une chronologie Le site du Roe d'Enclar (Andorra) de cette etape du site allant de la fin du IVeme siecle au Vleme siecle. Certains elements, comme les amphores entre autres, nous permettent de donner une chronologie pour cette etape jusqu' au Vleme siecle. Ceci est confirme par les datations de 14 C. L'habitat dont temoigne ce materiel est accompagne d'inhumations que nous avons date par le 14 C du Veme au Vli~me siecle. Ainsi nous pensons que la transformation du Roe d'Enclar en castellum et la presence de cette culture materielle sont des indices suffisants pour penser a 1'existence d'un habitat stable a l'interieur du castellum des le debut du Veme siecle. Celui-ci est le fruit de la mise en place meme du castellum et de l'insecurite cree par les deplacements des peuples germaniques vers le Sud. Pendant le Veme et en partie le VIeme siecle, cet habitat vit une certaine vitalite due a son role defensif, mais cet elan s 'estompe pendant le Vleme siecle lorsque cette partie du territoire devient une zone peripherique au sein du regne de Tolede. Malgre cette decadence, 1'habitat perdura jusqu'au IXeme siecle (fig. 7). Le materiel livre par 1'habitat des Veme-Vleme siecles temoigne des contacts commerciaux de cette partie des Pyrenees avec la Narbonnaise. Ces contacts s' effectuent par les voies fluviales qui relient les deux versants des Pyrenees a 1'Est. Les voies commerciales ~- Fig. 5.- Quelques bronzes hispano-romains dates de la fin du JVeme siecle et du Veme siecle. 3 et 2: boutons indispensables pour unir les differentes parties en cuir des equipements de monture, 6: passant de ceinture (dessin Julio Ariza). transmettent les nouvelles des troubles qui se produisent au sud de la Gaule pendant les Veme et Vieme siecles et 1'on peut penser que 1' occupation du site qui se deroule alors est lie ace climat d'insecurite. - _.-_ l I I I I CP- ----. \ L -I - I \ \ ll____j 0 2 Fig. 6.- Roe d'Enclar: plan des vestiges de !'habitat stable cree a l'interieur du castellum (re/eve: Alex Net). 117 M.A. Ruf, T. Vila, X. Soh!, X. Llovera, C. Yar'iez & J.M. Bosch 11 appara!t clairement par les datations au 14 C que I 'habitat se prolonge de fa9on continue. Les datations effectuees sur les tombes et les silos rattachent cette etape du site la periode qui va du VIIeme au IXeme siecle. Cette consolidation se concretise avec la construction d'une eglise avec son espace funeraire la fin du VIIeme siecle. Autour s'installeront les structures de stockage et 1'habitat. Comme dans les regions voisines du Sud de la France (Demians 1994; Charmasson 1970; Lebecq 1990, 21; Favory 1994, 2'18; Lugand & Pellecuer 1994, 25 8), dans la partie la plus septentrionale de 1'Italie (Brogiolo 1995) ou dans le Nord-Ouest de 1'Hispanie (Quiroga & Lovelle 1993, 33-34), le perchement de !'habitat que no us avons decrit pour 1'etape anterieure se poursui t. Le site suit ainsi la tendance generale de cette periode. Le materiel est caracteristique de ce type de sites perches. Il est lie aux activites domestiques, et est constitue de ceramique commune cuisson reductrice et d'objets en fer (couteaux). Pendant cette periode les ceramiques importees sont absentes du lot. C'est la production locale cuisson reductrice qui predomine. Cette ceramique modelee et regularisee la toumette est lie !'usage culinaire. La forme predominante est l'oule, bien qu'il y ait des cruches ou des bols. Grace 1'etude analytique des matieres premieres et des recipients nous avons pu etablir deux types de productions locales. L 'une est rattacher aux alentours proches du site et la vallee centrale du pays, }'autre provient de la vallee orientale bien foumie en argiles. La presence des ceramiques de ce second groupe produites a plusieurs kilometres du site nous a permis d 'illustrer les echanges dont cette ceramique commune est l'objet l'interieur du territoire dans une periode peu connue. Au IVeme siecle, les indices concemant la fabrication du vin et les objets associes temoignent d'un des aspects de tout un systeme de vie qui aurait, du moins, des traits romains significatifs. Nous ignorons quel etait exactement son fonctionement puisque nous n'avons pas trouve le centre auquel etait lie !'exploitation viti-vinicole de ce moment. Neanmoins, les vestiges mettent en evidence toute une strategie d'amenagement des versants et du haut-plateau pour la culture de la vigne et la fabrication du vin. A 1•etape suivante, il est fort probable que la situation d 'instabilite qui se produit entre les V erne et VII erne siecles fut un element decisif dans la creation d'un castellum et 1'apparition d'un habitat perche sur le site. Cette infrastructure militaire chargee de veiller sur les lieux de passage devait peu peu ete suivie d'un deplacement progressif et partiel de la population, attiree par le perchement du site. Malgre la presence dans un premier temps d' elements de culture romaine, nous considerons que cet habitat perche est indigene et qu'il maintenait des relations commerciales, tout au moins entre les V erne et VIIeme siecles, avec la region narbonnaise grace la proximite de la Strata Ceretana (voie secondaire du grand axe cotier, la via Domitia). A la fin du VIII erne siecle, 1' expansion de 1'etat carolingien vers le Sud place la vallee d' Andorre sous le pouvoir politique d 'une nouvelle aristocratie. La donation de la vallee d'Andorre par l'empereur Charles le Chauve au comte Sunifred, en l'an 843, correspond bien cet avenement. Ces changements politiques se traduisent par la construction du chateau d'Enclar. Ace moment, le comte utilise a sa faveur la tradition de lieu d'habitat et la situation strategique du site, qui devient alors le nouveau centre du pouvoir comtal et feodal. Conclusion Bibliographie Pour 1'instant, le site pyreneen du Roe d'Enclar est un des rares exemples des sites qui soient connus dans cette partie des pyrenees pour cette periode de transition (IVeme-Xeme siecles). 11 est en ce moment tres difficile de reconstruire 1' evolution de 1'occupation du sol partir de ce site seulement. Dans l'attente d'autres recherches !'evolution du site du Roe d 'Enclar nous permet de voir un exemple de continuite dans 1' occupation des lieux perches dans cette partie des Pyrenees. Les caracteristiques morphologiques et la situation geographique dominante par rapport la plaine de la vallee centrale d' Andorre ont influence les differents usages du site aux differentes periodes: au Bas- Empire, 1'epoque prefeodale et feodal e. BARBERO A. & VIGIL M. 1984: Sabre Ios origenes sociales de la Reconquista, Ariel, Barcelona. BROGIOLO G .P. et alii 1995: Citta, castelli, cam- a a a a a a a a a a a 118 a a a pagne nei territorio di frontiera (secoli VI- VII), Edizioni all 'Insegna del Giglio, Mantova. CASAS J., CASTANYER P., NOLLA J.M. & TREMOLEDA J. 1995:El man rural d'epoca romana a Catalunya, l 'exemple del nord-est, Centred 'Investigacions Arqueologiques de Girona, num. 15, Girona. CHARMANSSON J. 1970: Un oppidum du bas-empire: Lombren (Avenejean; Gard), Archeologia, num. 36, Paris, 54-61. D.A. 1997: Roe d'Enclar. Transformacions d'un espai dominant (segles IV-XIX), Govern d' Andorra, Le site du Roe d'Enclar (Andona) Andorra la V ell a. DEMIANS G. 1994: L 'oppidum de Saint-Blaise du V au VIIs. (Bouches-du-RhOne), Maison des Sciences de 1'Homme, Paris. FAVORY F. et alii 1994: Lunel-Viel et son territoire, clans: Les campagnes de la France mediterraneenne dans l'Antiquite et le !taut Moyen Age, Maison des Sciences de 1'Homme, Paris, 1994, 163-245. FAVORYF. & FICHES J.L. 1994: Les campagnes de la France mediterraneenne dans l 'Antiquite et le haut Moyen Age. Etudes microregionales, Maison des Sciences de 1'Homme, Paris. GOMIS M. 1997: Les monedes, clans: Roe d'Enclar. Transformacions d 'un espai dominant (segles IVXIX), Govern d' Andorra, Andorra la V ell a, 515526. LEBECQ S. 1990: Les origines franques (V-IX siecle), Editions du Seuil, Paris. LUGAND M. & PELLECUER C. 1994: La region de Meze et la villa des Pres-Bas a Loupian (Herault): contribution a 1' etude du litoral languedocien, clans: Les campagnes de la France mediterraneenne dans l 'Antiquite et le haut Moyen Age, Maison des Sciences de l'Homme, Paris, 1994, 246-278. NOLLA J.M & RoD A I. 1995: El sector meridional dels Pirineus a l'epoca antiga. Unes reflexions, clans: Cultura i medi. De la Prehistoria a I 'Edat Mitjana, Xe Co·loqui Internacional d'Arqueologia de Puigcerda, Puigcerda, 507-515. QUIROGA J.L. & LOYELLE M.R. 1993: Poblamiento rural en el noroeste de la Peninsula Iberica (ss. VXI): Una introducci6n al estudio del poblamiento rural entre la Antigtiedad tardia y la Alta Edad Media en Galicia a traves de un am\.lisis microregional, Boletin de Arqueologia Medieval 7, Madrid, 21-52. Rico C. 1992: Les Pyrenees entre la Gaule et les provinces iberiques a l 'epoque de la domination romaine, Tesi Doctoral, Universitat de Tolosa IITolouse Le Mirail, inedit. ROVIRA M. C. 1997: Els accessoris dels vestits, clans: Roe d'Enclar. Transformacions d'un espai dominant (segles IV-XIX), Govern d' Andorra, Andorra la Vella, 143-150. VELAZQUEZ I. 1991: Notas sobre la descripci6n de tierras pirenaicas en fuentes literarias de epoca tardoantigua y visigoda. (Aspectos literarios fundamentalmente), clans: Aetas del Congreso Internacional Historia de los Pirineos, Universidad Nacional de Educaci6n a Distancia, Madrid, 4 77488. M" Angels Ruf, Toni Vila, Xavier Sole, Xavier Llovera, Cristina Y Miez & Josep Maria Bosch Servei de Recerca Historica Patrimoni Cultural d' Andorra Carretera de Bixessarri s/n Aixovall Principat d' Andorra 119 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Maruska Federici-Schenardi & Robert Fellner L'habitat rural du haut moyen age de Develier-Courtetelle (Jura, Suisse) Introduction Dans le canton du Jura, au nord-ouest de la Suisse, un service d'archeologie a ete cree en 1985, suite a un projet de construction autoroutiere. D 'emblee, un programme de sondages sur le trace de 1' autoroute A 16 a ete planifie. Sur un tron9on de 15 km ainsi explore entre les villes de Delemont et de Porrentruy, plus de quinze sites ont deja ete decouverts et fouilles (Schifferdecker 1994). Ce chiffre n'est pas definitif puisque la prospection de la demiere moitie du trace autoroutier vient seulement de commencer. L'habitat rural merovingien de Develier-Courtetelle a ete decouvert lors de sondages realises entre 1987 et 1993. L'etendue considerable de ce site (environ 1 km de long et 3,5 ha de surface) a implique la creation de deux equipes actives pendant quatre campagnes de fouille de neuf mois chacune (1993-1996). L'etude vient d'etre lancee et devrait se terminer par une serie de monographies a paraitre des l'an 2000. Situation topographique et geologique Le site de Develier-Courtetelle est situe a 450 metres d'altitude, clans une plaine alluviale large d'environ 250 metres, occupant le fond d'un vallon lateral de la vallee de Delemont. Dans cette plaine coule le ruisseau "La Pran" domine, au sud, par le "Bois-de Chaux", colline formee de molasse de l'Oligocene. Dans ce bassin sedimentaire, 1'epaisseur des depots quatemaires recouvrant la molasse varie de 3 a 4 metres. 11 s'agit essentiellement de formations fluviatiles (graviers, sables, limons) entre lesquelles viennent s 'intercaler des sols enfouis. Deux occupations anterieures a celle du Haut Moyen Age ont ete rencontrees; l'une de l'Age du Fer, l'autre de l'epoque gallo-romaine. Ces deux niveaux, presents ponctuellement seulement, ne renfermaient que peu de mobilier et de rares structures. La couche archeologique du Haut Moyen Age est le plus recent des sols en- fouis et la seule occupation d'une etendue vraiment importante. Situee a une profondeur oscillant en general entre 40 et 60 cm, elle se developpe le long du ruisseau "La Pran". Son etat de conservation change considerablement d'un endroit a l'autre; parfois presque completement erodee, elle peut s'epaissir et presenter plusieurs phases sedimentaires. L'activite ininterrompue du ruisseau "La Pran" a considerablement influence la conservation des traces d'occupation. D'une part, la formation des meandres posterieurs a 1'occupation en a detruit de grandes parties mais, d'autre part, l'enfouissement rapide des vestiges situes clans des bras morts du ruisseau a permis la conservation de matieres organiques telles que bois ou macrorestes vegetaux non carbonises. L'organisation spatiale de !'habitat Plusieurs ensembles de structures s 'echelonnent le long du cours d'eau. 11 s'agit soit de fermes, soit de zones a vocation specifique (Peytremann 1995) vouees notamment a l'artisanat et au stockage. Ace stade de la recherche deja, d' evidentes differences ont ete relevees entre les moities orientale et occidentale du gisement, situe de part et d' autre du ruisseau "La Pran". Dans la partie ouest du site, un systeme de fosses orthogonaux semble delimiter les differentes zones d'activite, en separant par exemple un quartier d'habitat d 'une aire artisanale. Plus a 1' est, des espaces depourvus d'amenagement et de mobilier s'intercalent entre les differentes concentrations de structures. De maniere provisoire, on peut distinguer six fermes, quatre zones liees a 1'industrie du fer, deux secteurs de stockage et deux types d, amenagement plus specifiquement rattaches au ruisseau. Des enclos reserves au betail semblent egalement etre presents a plusieurs endroits de }'agglomeration rurale. Deux tombes isolees, probablement merovingiennes, ont ete mises aujour a proximite immediate de }'habitat, mais la necropole qui devait etre as121 M. Federici-Schenardi & R. Fellner N .( .. -· ·~ I .,, ~ :1\ . ~- Deveher ;:.::::::::::- ~_,_-- ~ 453 . . .• ·-~--~ 'l.:·~:c;-:.;;-;;;=~~···~~..;:-:·:::;::::;::.;;;; • • .I ..--11 - 111!1111 --·· I 10 km 5 0 .....,....~~., l! .. ::·:~:·.::.· ·::::::---~ • f\ ...,.... ..1,. - . ·~. "~i;i"''.:/ . . " ·:..... ._J:E-- e1- j'J'f I .....\, -· .. . I I I a u. Delernont .- ·__.o-- u fl/ / ·~1 '"' -/,:-:""'"-' .··· I we ---l a .... ·· WKmm Extension de la fouille 1993-96 Fig. 1. - Situation geographique et extension du site de Develier-Courtetelle. 122 L'habitat rural du haut moyen age de Develier-Courtetelle (Jura, Suisse) Fig. 2.- Exemples de plans de ferme dans la partie orientate (1) et occidentale (2) du site. + D 0 0 • 1 · Trou de poteau • Trou de poteau potentiel 0 Fosse 122> Foyer \ '-\ Empierrcment ~ Foss6 tmiTiil Maison t:ZZl Cabane en Fosse H Biltiment annexe .. 2 0 tOm sociee a cette occupation n 'a pas ete decouverte. I! faut neanmoins signaler la presence d'un cimetiere merovingien quelques centaines de metres, decouvert au 19e siecle (Gerster 1976). L'etude n'est pas encore suffisamment avancee pour que !'organisation interne du village et l'interactivite des differentes parties de ce dernier puissent etre presentees ici. Nous nous limiterons done une breve description des premiers acquis. a a Les fermes Les differentes fermes de 1'habitat rural de Develier-Courtetelle presentent chacune une maison d 'habitation entouree de plusieurs batiments annexes tels que cabanes en fosse ou petites constructions a 4 ou 6 poteaux generalement (fig. 2). Cette association, typique du Haut Moyen Age, est particulierement lisible dans la partie orientale du site, ou chaque 123 M. Federici-Schenardi & R. Fellner Fig. 3.- Exemples de plans de maison d 'habitation dans la partie occidentale (1, 2) et orientate (3, 4) du site . + • ® ) ~==---= !Ill ~ • ' CP • • Trou de piquet • == ® ~ @ @ • • • ,::; • 0 ~ (t)~ t/1 @ '~ ' ~~~ ~ 11/iiP !3 • • • (j 3 .:-<i) (j) ~ @ •• @ 4 5m ferme presente une seule phase de construction. Plus a 1'ouest, la presence de constructions superposees brouille la vision de 1' organisation. Les maisons d 'habitation Les maisons d 'habitation presentent differents modes de construction. Dans la partie orientale du site, elles possedent toutes un plan a une seule nef. Dans deux cas, il s'agit de constructions poteaux en bois mesu- a 124 Fosse Foyer Fosse ~ \:') ;::, ~ 2 a Trou de poteau C• Trou de poteau potentiel 0 ez> (j) Ill rant environ 6,5 x 8 m (fig. 3.3, 4); clans un troisieme cas, on a affaire a un batiment sur solin de pierre entoure d'un portique sur deux cotes. Le corps central mesure 7 x 9 m; le portique atteint 2 m de large. Dans ces trois exemples, les proportions de la partie habitee sont tres semblables (env. 1,2/1 ). De plus, chaque maison contient un foyer situe aproximite de la paroi sud. Dans la partie occidentale du site, les maisons d'habitation semblent toutes presenter un plan adeux L 'habitat rural du ha ut moyen age de Develier-Courtetelle (Jura, Suisse) nefs. Une construction, asablit':re basse, est materialisee par deux fosses paralleles (fig. 3.1) et mesure environ 7,5 x 3,5 m. Elle comporte un foyer central. Deux auh·es maisons, apoteaux en bois, font respectivement 6,5 x 4,5 m (fig. 3.2) et 9 x 5 m. L'architecture des maisons de la partie occidentale est plus leg ere que celle de 1'est: le diametre des poteaux est plus petit et l'espacement entre ceux-ci est plus grand. L 'erosion du niveau de circulation pourrait etre a I' origine de 1' absence de foyer dans les deux batiments apoteaux. Les ea banes en fosse Quinze structures de ce type ont ete decouvertes sur le site. Ces fosses de forme plus ou moins rectangulaire presentent des dimensions comprises entre 2,5 x 2 m et 3,5 x 3 m, soit une surface variant entre 5 et 10m2. La profondeur conservee de ces structures oscille entre 10 et 50 cm. A une exception pres, ces cabanes en fosse contiennent, a 1' interieur de 1' excavation, une serie de trous de poteau. Le type a quatre poteaux d'angle, parfois doubles, semble etre la solution la plus couramment choisie. Le modele a deux poteaux fait defaut. Dans deux cas, un avant-toit materialise par deux trous de poteau a ete pers;u. La fonction primaire de ces structures n'a pas pu etre clairement definie. Leur remploi en tant que zones de rejet ou de combustion a par contre ete observe aplusieurs reprises. Les deux tiers de ces constructions se concentrent dans la partie occidentale du village. Ce nombre eleve trouve peut etre une explication clans la presence de plusieurs phases de construction clans cette aire du site. Les bdtiments annexes Depourvus de foyer, les batiments annexes sontrestitues d'apres les alignements des trous de poteau sauf dans un cas ou il s'agit d'une construction a sabliere basse. La plupart sont des structures carrees aquatre poteaux d'angle mesurant entre 2 et 3 m de cote. D'autres structures, rectangulaires et de dimension superieure, sont constituees de 6 a 12 poteaux. Les differentes fonctions de ces batiments seront definies a un stade ulterieur de la recherche. Les zones a vocation specifique A proximite plus ou mains immediate des fermes se trouvent des ensembles de structures associees une serie d'activites diverses. On peut notamment distinguer des quartiers lies a1' industrie du fer et des zones apparemment destinees au stockage des recoltes ou au parcage du betail. a Le travail du fer La vallee de Delemont est riche en minerai de fer, et des bas fourneaux datant du Ha ut Moyen Age ont ete fouilles dans cette region (Eschenlohr & Semeels 1991 ). Ce type de construction qui temoignerait de la premiere phase de la cha'ine operatoire du travail du fer, la reduction du minerai, n'est pas atteste dans !'habitat de Develier-Courtetelle. Par contre, les scories produites lors du raffinage et du forgeage de ce metal sont tres abondantes et se concentrent a proximite de petites fosses marquees par des traces de combustion. On distingue deux categories de structures liees a ce travail, a savoir les bas foyers de raffinage et les bas foyers de forge. Les premiers se definissent par la presence de scories en forme de calotte et les seconds par celle de battitures (petits eclats de fer produits lors du martelage). Dans la partie occidentale du site, un quartier destine au travail du fer et delimite par un systeme de fosses a ete observe en bordure d'une zone d'habitat. Cet ensemble se caracterise par deux bas foyers de raffinage et un bas foyer de forge associes ades trous de poteau epars. A 1'est ont ete relevees trois zones de travail, dont une est constituee d'un bas foyer de raffinage et d'une aire de forge associes aune concentration tres importante de scories d 'un poids approchant deux tonnes. La metallurgie du fer a joue un role preponderant clans l'economie du village; celle du bronze, bien plus limitee, est attestee par la decouverte de quelques petits creusets portant des traces de ce metal. Le stockage La presence de petites constructions aproximite des maisons indique qu 'une partie des produits agricoles a ete stockee a 1'interieur des fermes. Neanmoins la mise au jour de plusieurs ensembles de constructions en dehors des zones d'habitat suggere 1, existence de quartiers reserves au stockage et au parcage du betail. Un ensemble de ce type se dessine notamment clans la partie centrale du site. Une cabane en fosse et une petite construction quadrangulaire sont associees ades alignements de trous de poteau definissant probablement des enclos. A proximite immediate, un empierrement traversant !'ancien lit du ruisseau a ete interprete comme gue. Les amenagements du ruisseau Dans les anciens meandres du ruisseau, au-dessous du niveau de battement de la nappe phreatique, on a mis au jour plusieurs concentrations de bois datant de !'occupation. Il s'agit de deux sortes d'amenagement qui semblent egalement avoir ete destinees a des 125 M. Federici-Schenardi & R. Fellner Fig. 4. - Ceramique apate rugueuse (1, 2, 3), sableuse (4) et fine (5). ~ I I 2 I I ' - - - - - - - t . · .. 4 0 vocations specifiques qui pour le moment demeurent enigmatiques. I1 s' agit d 'une part de trois bass ins surcreuses artificiellement clans un ancien lit du ruisseau et proteges du ruisseau contemporain par une digue en gravier. Ils sont relies entre eux et au ruisseau par un canal d' adduction. 126 2cm 5 D'autre part nous avons affaire, plus a l'ouest, a des amenagements ponctuels de la berge, par endroits assainie !'aide d'empierrements et stabilisee par des rangees de piquets relies entre eux par un tressage. Ce type de construction se rencontre encore actuellement clans les ruisseaux de la region et fonctionne effectivement comme protection de la berge contre !'erosion. a L'habitat rural du haut moyen age de Develier-Courtetelle (Jura, Suisse) Fig. 5. - Ceramique tounu?e ancienne (1, 2, 3) et d montage mixte Tg I . - () I 1 (4, 5). 2 rf' I I 3 4 Le mobilier Parmi le mobilier archeologique mis au jour lors de la fouille de Develier-Courtetelle, les categories suivantes peuvent etre distinguees: poterie, terres cuites, objets en metal, scories de fer, verre, industrie lithique, bois et ossements animaux parfois travailles. Cet inventaire, representatif d'un habitat du Haut Moyen Age, est expose succinctement ci-dessous. La poterie La poterie, en general tres fragmentee, est subdivisable en cinq classes d'apres la pate: - la ceramique fine, cuisson majoritairement reductrice, est composee de gobelets, vases biconiques et rares pichets (fig. 4.5); - la ceramique rugueuse, a cuisson souvent oxydante, est constituee de pots a cuire ovoldes et de quelques formes ouvertes (fig. 4.1, 2, 3 ); a 127 M. Federici-Schenardi & R. Fellner - la ceramique sableuse, a cuisson generalement reductrice, est uniquement composee de pots a cuire (fig. 4.4); - la ceramique a montage mixte, a cuisson oxydoreductrice, est egalement uniquement composee de pots acuire (fig. 5.4, 5); - la ceramique toumee ancienne, a pate jaune, provenant vraisemblablement de 1' Alsace, a cuisson oxydante elevee, est representee par des pots acuire aprofil globuleux et epaulement marque (fig. 5.1' 2, 3). D'apres la typologie, le mobilier ceramique se situe entre le 6e et le 9e siecle. L'etude preliminaire de la repartition spatiale de la ceramique a permis de reconna1tre plusieurs ensembles co'incidant souvent avec une ferme. Ceci devrait permettre de comprendre !'evolution chrono-typologique de ce mobilier. Les terres cuites Les tuiles constituent la categorie dominante. Il s' agit d' elements de recuperation et de reutilisation provenant probablement d 'un etablissement gallo-romain situe a proximite du site. Les briques recyclees sont plus rares. Les elements de clayonnage sont peu nombreux. Une seule concentration importante a ete mise aujour pres d 'une structure de combustion. Ceci incite a penser que l'argile n'intervenait pas souvent dans la construction. L'inventaire des terres cuites comprend, outre ces elements architecturaux, quelques fusa'ioles. Les objets en metal Les artefacts en fer sont de loin les plus nombreux. Les objets en bronze sont relativement bien representes, tandis que 1'argent, a deux exceptions pres, n' a ete utilise que pour la technique de damasquinage rencontree sur des elements de ceinture. L'outillage, en fer, se compose de lames de couteaux et d'objets lies au travail du bois ou du metal comme des ciseaux, des meches a cuillere et une petite enclume. Des eperons et des pointes de fleche en fer (fig. 6.3, 5), ainsi que des boutons decoratifs de fourreaux de scramasaxe en bronze sont des elements de 1'equipement militaire. Parmi le mobilier vestimentaire, les elements de gamiture de ceinture et de courroie predominent. Fabriques en fer, parfois damasquine (fig. 6.2), en bronze et rarement en argent, ces objets rappellent le mobilier retrouve dans les necropoles. Les quelques fibules en bronze appartiennent al'habillement feminin. On releve en particulier la presence d'une fibule en forme de cheval (fig. 6.6) et de deux fibules ansees, dont une incrustee de verre rouge (fig. 6.1). Une 128 fibule conique amedaillons ocules date de la periode gallo-romaine. En outre, une figure anthropomorphe en t6le de fer interpretee provisoirement comme ex-voto (fig. 6.4) a ete trouvee en bordure de I'habitat. Il n'est pas exclu que cet objet soit plus tardif. I! reste asignaler plusieurs fragments de clochettes en t6le de fer recouvertes de cuivre ou de bronze et differents types de clous et fragments de tiges en grande quantite. Parmi le mobilier datable, les pieces attribuables au 7e siecle predominent. La majorite des objets metalliques sont actuellement en cours de restauration et n'ont pas encore ete etudies. Les scories de fer Entre trois et quatre tonnes de scories de fer ont ete retrouvees. Les scories en forme de calotte, souvent fragrnentees, sont les plus nombreuses. De plus petite taille, les scories coulees proviennent vraisemblablement des operations de raffinage et forgeage et non pas de la reduction du minerai. Les battitures, eclats minuscules de fer produits lors de martelage, sont particulierement frequentes pres des aires de forge. Le verre Les recipients en verre et les objets de parure en pate de verre forment les deux categories reconnues. Les premiers sont tres fragrnentes; les tessons de coupes hemispheriques ou coniques predominent. Quelques rares fragments aparois plus epaisses proviennent de bouteilles carrees romaines. Parmi les objets de parure, on compte deux fragments de bracelet et plusieurs dizaines de perles de forme, de couleur et de qualite de fabrication variees. Le type de perle le plus courant est de couleur jaune opaque, en pate de verre fritte. Generalement globulaire, ces petites perles peuvent aussi etre composees de plusieurs segments. Les objets lithiques La vaisselle et les outils ont aussi ete fabriques a partir de la pierre. Ainsi, de nombreux fragments de marmites fac;:onnees dans de la pierre ollaire temoignent de l'interet que l'on portait a cette matiere refractaire ideale pour la cuisson des aliments. Quelques aiguisoirs et polissoirs en gres molassique ont probablement un lien avec la fabrication des objets en fer. La presence limitee d'eclats de silex s'explique par une utilisation exclusive de ces pieces comme briquet. I! s 'agi t en partie d' artefacts prehistoriques reutil ises. L 'habitat rural du ha ut moyen age de Develier-Courtetelle (Jura, Suisse) Fig. 6. - Objets metalliques: 1 jibule en bronze, 2 elbnents de garniture de ceinture en fer et argent, 3 pointe de flee he en fer, 4 figure anthropomorphe en fer, 5 eperon en fer, 6 jibule en bronze, 7 boucle d 'oreille en bronze. Echelle 1:1, sauf(4) et (5). I' o 4 2cm =- 0 o 2cm =-5 6 Des fragments de meules en gres ou conglomerat completent cet inventaire. L 'ambre L' ambre, importe de la Baltique, est aus si atteste: huit perles de forme allongee ont ete decouvertes sur le site. L 'origine exogene de cette matiere ain si que celle de la pierre ollaire, provenant des gisements alpins, indique !'existence d'un commerce longue distance. a Conclusions Le ruisseau "La Pran" semble jouer un role d'e!ement ordonnateur clans I' organisation spatia le du village de Develier-Courtetel!e, qui peut ainsi etre qualifie d'habitat lineaire, selon le vocabulaire adopte par H. Hamerow (1995). Les questions d'ordre chronologique, economique et social constitueront le sujet central de I' etude 129 M. Federici-Schenardi & R. Fellner pluridisciplinaire qui a recemment debute. Elles interviendront a deux niveaux, c 'est-a-dire aus si bien au sein des differentes unites composant le village que clans un cadre plus large. Pour ce qui est de la datation, quelques jalons ont deja ete poses: d'une part par la typologie, d'autre part par une serie de datations C14. Si !'occupation semble avoir dure du 6e siecle au 8e, voire au 9e siecle, elle semble atteindre son apogee au courant du 7e siecle. Sur le plan economique, il appara1t clairement qu' on a affaire avant tout a un village de forgerons. I1 s'agira neanmoins encore d'estimer le degre d' importance de cette activite metallurgique ain si que celui des autres domaines de la production. L 'extension de la surface exploree, ainsi que la qualite des vestiges, rendent le site de DevelierCourtetelle exceptionnel a 1'echelle suisse, ce qui justifie une etude approfondie. Le present document a ete redige sur la base des donnees contenues clans les rapports de fouille annuels disponibles a 1'Office du patrimoine historique, Porrentruy. Depuis 1993, 8 rapports, so it environ 800 pages au total, ont deja ete consacres a 1' etude de cet habitat. Bibliographie ESCHENLOHR L. & SERNEELS V. 1991: Les bas fourneaux merovingiens de B01icourt, Les Boulies (JU, Suisse), Cahiers d' ArcMologie Jurassienne 3, Porrentruy, Office du Patrimoine Historique et Societe jurassienne d'emulation, 143 p. GERSTER A. 1976: Romische und merowingische Funde in Develier, Helvetia archaeologica 7, 3038. HAMEROW H. 1995: Shaping settlements: early medieval communities in Northwest Europe. In: BINTLIFF J. & HAMEROW H. (dir.), Europe between Late Antiquity and the Middle Ages, BAR International Series 617, Oxford, 8-37. PEYTREMANN E. 1995: Les structures d'habitat rural du Haut Moyen Age en France (Ve-Xes.). Un etat de la recherche. In: Lorren Cl. & Perin P. (dir.), L 'habitat rural du Haul Moyen Age (France, Pays-Bas, Danemark et Grande-Bretagne). Actes des X!Ve Journees internationales d 'Archeologie merovingienne, Guiry-en- Vex in et Paris, 4-8 fevrier 199 3, Memoires de 1'AF AM VI, Condesur-Noireau, 1-28. SCHIFFERDECKER Fr. 1994: Sous la Transjurane. Prospection et sondages entre Porrentruy et Delemont, Archeologie Suisse 17, 31-35. Maruska Federici-Schenardi & Robert Fellner Office du patrimoine historique Section d'archeologie Hotel des Halles 2900 Porrentruy (JU) Suisse 130 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference - Volume 6 Armelle Querrien L'exploitation du sol et du sous-sol autour d'un bourg castral du Berry, du Xle au XVe siecle Le village de Moulins-sur-Cephons, clans le departement de 1'Indre, est un ancien bourg castral, qui s, est developpe a 1, interieur des basses-cours du chateau a motte edifie au bord de la Cephons au milieu du XIe siecle. La motte fut construite au moment ou les guerres opposant les comtes de Blois aux comtes d' Anjou se deroulaient sur les confins de la Touraine, du Berry et du Blesois; elle etait situee a la frontiere entre les terres relevant de Deols et celles relevant du comte d' Anjou. Son abandon coincide avec la fin de la guerre de Cent Ans. La fouille de la motte a ete accompagnee de nombreuses recherches sur le paysage et !'exploitation du territoire relevant de la motte, mettant en jeu differentes sources et techniques. 1 Les sources a. Les sources archeologiques forage du puits autour de 1050. Plusieurs niveaux d'occupation et de remaniement ont ete mis en evidence au sommet, le plus ancien ayant ete conserve grace a une surelevation de la plate-forme par une couche de remblais d' 1 m d'epaisseur. La plate-forme de la motte, occupee des la deuxieme moitie du Xle siecle, l'etait encore durant la guerre de Cent Ans, les couches de destruction de la motte renfermant du materiel du XVe siecle et des monnaies emises entre 1419 et 1431. C'est egalement a cette periode que les Grands Fosses qui delimitaient la zone fortifiee annexee a la motte ont ete combles. Tout le mobilier 1 et le materiel osseux mis au jour, et les nombreux prelevements effectues clans les structures et clans les stratigraphies ont ete etudies avec la participation de divers specialistes, etudiant les vestiges du monde vegetal tels que les pollens, graines, fruits et fragments de bois. Une reconstitution du paysage des environs de !'atelier de potier et de la motte a ete entreprise grace aux differents resultats obtenus. La motte etait de dimensions imposantes puisque, jusqu' au XIXe siecle, elle avait 45 a 50 m de diametre a la base, 22 m au sommet, et une hauteur de plus de 7 m. Elle n'a conserve que la moitie de son volume, 1'autre moitie ayant alors servi de carriere de terre. Malgre les lacunes dues a son exploitation, le site s'est revele d'une grande complexite. Un atelier de potier du debut du XIe siecle, date par archeomagnetisme et par sa production, a ete decouvert, enfoui sous la motte. Un puits, fore depuis la plate-forme de la motte peu apres son edification, s'enfonyaitjusqu'a 7 m sous le niveau du sol en place au Xle siecle; a sa base se trouvait un cadre en bois de plan carre, le "rouet", forme de quatre poutres de chenes encastrees, destine a assurer la stabilite de la construction mayonnee edifiee au dessus. Une datation par dendrochronologie et une autre par carbone 14 ont permis d'etablir le -- - - Sous la direction d' A. QUERRIEN, Un village nuidieval en Bas Berry, Moulins-sur-Cephons, Catalogue d 'exposition, ARHAMIS, Moulins-sur-Cephons, 1988, 84 p. limitc du dcpartcmcnt de l'lndre Situation de Moulins sur Cephons (Indre, France). 131 A. Querrien b. Les sources documentaires Les plus anciens documents ecrits concemant la seigneurie de Moulins sont posterieurs a 1200. Les premieres mentions de la castellania, du castrum et de I'aula de Moulins ont au moins 150 ans de retard sur les donnees archeologiques. Pourtant la motte etait le centre d'une seigneurie qui avait rang de chatellenie et dont le dominus exerc;:ait le droit de haute, moyenne et basse justice. Un document anterieur montre qu'il y avait la en 966 un grand domaine lai'c, dont un manse fut donne a l'abbaye SaintSulpice de Bourges: "cedo ... mansum meum indominicatum qui est in pago Biturigo in vicaria Bulbiensi in villa Mollinis super jluvium Naonis hoc est una cum casualis et suprapositis urtiferis et cultiferis, terris, camp is, pratis, mancipiis et omnibus que aspiciunt ad ipsum 2... ". Ce manse devait se trouver au nord de Moulins, a Balzeme, ancienne paroisse sous le vocable de Saint Sulpice. Le centre de la villa a ete identifie a Moulins-sur-Cephons, qui est le seul lieu dont le toponyme correspond clans tout le bassin du Nahon, la Cephons etant un affluent du Nahon, et qui est a 9 km a vol d'oiseau de Bouges, centre de la vicaria Bulbiensis. La villa etait sans doute situee sur la rive droite de la Cephons, la ou une enceinte de la fin du troisieme millenaire avant notre ere, encore occupee a I' Age du Fer et aI' epoque gallo-romaine a ete partiellement fouillee. L 'atelier pourrait en avoir ete une dependance, etablie sur la rive gauche, de !'autre cote d'un ancien gue. A partir du XIIIe siecle les sources se multiplient, notamment grace au cartulaire de Levroux et aux archives de I 'abbaye cistercienne du Landais et du prieure fontevriste de Jarzay, situes au maximum a 5 km de la motte, le demier etant clans la seigneurie de Moulins. Un aveu et denombrement de 15003 , redige pour assurer les droits des seigneurs de Moulins clans la periode de reconstruction du regime seigneurial qui a suivi la guerre de Cent Ans et les troubles de la fin du XVe siecle, donne une description tres precise de la terre de Moulins a la fin du Moyen Age. Deux autres aveux et denombrement du XVIIe et d'un Etat de la terre de Moulins du XVIIIe siecle 4 nous sont L. BUHOT DE KERSERS, Essai de reconstitution du cartulaire A de l'abbaye Saint-Sulpice de Bourges, Memoires de la Societe des Antiquaires du Centre XXV, 1912, I 05 et I 06, charte de 966. 3 Aveu d' lmbert de Batamay, mars 1499 avant Paques, ancien style, c'est-a-dire 1500, Archives departementales de l'lndre, H 839, copie sur papier du XVIIe sii:cle. Toutes les indications de f" sans precision d'origine renvoient ace document. 132 egalement parvenus. L'ensemble de ces documents sont autant de jalons necessaires a la mise en oeuvre de la methode regressive a laquelle les analyses de parcellaire et les tentatives de reconstitution du paysage ont recours pour remonter le temps a partir de documents comme les photographies aeriennes de I' IGN, les cartes au 1/25000, et les cadastres du XIXe siecle. Ils permettent de suivre de nombreux evenements, dont les changements de toponymes, les partages de parcelles, certains defrichements et retours a la friche, des modifications du reseau routier. La conservation du cadastre de 1812 et de ses etats de section, et sa confrontation avec I'aveu de 1500 ont permis de retrouver la situation et les surfaces des terres du domaine seigneurial a la fin du XVe siecle, de retracer la limite de la chil.tellenie et de reconstituer en grande partie le terroir de Moulins au XVe siecle, avec ses hameaux, ecarts, moulins, chemins, ponts, gues, etangs, bois, pres, terres cultivees, vignes. La repartition des terres cultivees, pres, vignes, tail lis et forets, telle qu'on la perc;:oit clans l'aveu de 1500, ne connalt que des modifications de detail par rapport au cadastre de 1812. Les indications recueillies clans les textes du XIIIe siecle, bien que fragmentaires, montrent une permanence de I' organisation du terroir de Moulins entre 1200 et 1500. La comparaison entre les analyses des pollens actuels et celles etablies pour ce site occupe du XIe au XVe siecle fait remonter cette organisation au moins au XIe siecle. C'est a cette periode de mise en place de la chil.tellenie qu 'une reorganisation du parcellaire autour du cheflieu de la chil.tellenie suivant un schema radioconcentrique a ete attribuee lors d'un etude recente 5 . Ce schema a succede a une organisation orthogonale heritee de I' Anti quite et conservee durant le ha ut Moyen Age. Une etude des structures d 'habitat et d'occupation du sol a demontre la remarquable stabilite de cet heritage du Moyen Age. C'est seulement au milieu du XIXe siecle que !'evolution du reseau routier et le debut de l'exode rural commencent a modifier l'apparence du terroir villageois d'ou plusieurs ecarts disparaissent. Les etangs sont asseches a la fin du XIXe et les moulins abandonnes au debut du XXe. Archives dep::rtementales de l'lndre, H 839, copie du XVIIe sii:cle de I' aveu de 1623. Archives privees du chateau de Saint-Aignan (Loire et Cher), aveu de 1646. Archives nationales, minutier central, L XXXIX, 388, Etat de la terre de Moulins en Berry, 1728. 5 A. QUERRIEN, Parcellaires antiques et medievaux du Berry, Journal des Savants, 1994, fasc. 2, 235-366, 4 depliants hors texte, 40 cartes et figures. L'exploitation du sol et du sous-sol autour d'un bourg castral du Berry, du XIe au XVe siecle Le bourg castral de Moulins d 'apres les sources archeologiques, historiques et le cadastre de 1812. 1 matte; 2 grands fosses; 3 eglise; 4 chatelet d 'entree du castrum; 5 maison seigneuriale intra muros; 6 7 Maison des Salles; moulin du Pont. 2 Le paysage a. Lajlore L'etude palynologique 6, par !'analyse des pollens et des spores de fougeres apportes par le vent et deposes dans les sediments du site, a permis de restituer le paysage vegetal des environs du site, depuis l'etablissement de !'atelier de potier jusqu'a nos jours. Les faibles taux de pollens d'arbres ont montre que, des !'an mil, le paysage etait decouvert et voue aux prairies et aux cultures. Differentes etudes archeologiques ont par ailleurs mis en evidence 1'anciennete de !'occupation de cette region; de nombreux vestiges prehistoriques, protohistoriques et gallo-romains ont ete trouves en prospection sur le territoire de la commune, et deux importants etablissements du chalcolithique mis au jour. Mais seules les structures et le mobilier ont ete publies. Une tentative d'etude palynologique dans !'enceinte du chalcolithique s'est soldee par un echec, les prelevements etant steriles. Aucun element de comparaison avec les epoques anterieures au Moyen Age n'est done possible actuellement, alors que les defrichements ont dO. etre importants des le neolithique. A Levroux, 5 km au L'etude palynologique a ete faite par Bui Thi Ma'i, CRACNRS, Valbonne; elle sera developpee, ainsi que les autres etudes presentees ici dans la publication finale de la fouille dont la redaction est en cours. sud-est, les etudes faites sur le village de la Tene aboutissent a !'image d'un paysage deja decouvert. Les arbres autochtones sont surtout representes par des pins, des aulnes, des noisetiers et des chenes pedoncules. D'autres arbres ont ete introduits et cultives comme le tilleul, les noyers et les chiHaigniers, pour la consommation des fruits, le marronnier (niveau du debut du Xle siecle ). La vegetation herbacee dominante est celle des prairies de fauche et des patures (graminees et cichoriees du type du pissenlit) dont les frequences varient probablement se Ion 1'importance des coupes ou de la frequentation du betail. Les cultures cerealieres sont attestees par les pollens de cereales, associes a ceux des "plantes messicolescompagnes" que !'on trouve communement au milieu des champs de ble, de seigle, ou d'orge. Ce sont, par exemple, les marguerites (anthemidees), les coquelicots (papaveracees ), les centaurees, les plantains et des cruciferes. Certaines especes attestees par d'autres sources n 'apparaissent pas; certaines pollinisent peu, d'autres sont coupees pour leur consommation avant leur floraison. Mais surtout, pour la plupart des herbacees, I'analyse se limite le plus souvent a determiner la famille; elle peut rarement identifier I' espece. Ainsi en est-il pour !'ensemble des cereales qui ne peuvent que rarement etre distinguees les unes des autres; toutefois un certain nombre de pollens de seigle ont pu etre isoles de !'ensemble des pollens de cereales des niveaux de 1' atelier de potier du debut du Xle siecle, grace a un etat de conservation remarquable. 133 A. Querrien LAFLORE carpo- palynologie logie ESPECES LlGNEUSES aubepine aulne bouleau buis charme chataignier chene cormier epic ea epine vinette frene genevrier griottier hetre lierre marronnier ne flier nerprun noisetier noyer orme pecher peuplier pin platane pommier ou poirier prunellier prunier creque prunier cultive sap in saule sureau hieble sureau noir tilleul - - - xylologie - - X X X X X - X - X X X X X X - - X X - X X - X - - X - - X X X X X X X X X - X X - X X X X X - - X X - t extes - X X X - X X X X - X X X - X - - - X X X - - - - - in d. = in determine( e) N.B.: !'attribution au XVe siecle du pepin de fraisier et de la semence de fenouil mineralises trouves dans le Grand Fosse n 'est pas certain e. La xylologie 7 et la carpologie 8 ont montre quelles especes poussaient sur place ou etaient apportees par 1'homme volontairement ou involontairement. De nombreux fragments de branchages, ne presentant aucune trace de travail, ont ete trouves dans la zone immergee du puits; ils y ont ete jetes depuis la plateforme de la motte et proviennent de ses environs 134 carpo- palynologie logie PLANTES CULTIVIlES seigle froment epeautre orge vetue millet avoine cereales ind. pois feverole fenouil vigne cultivee chanvre !in PLANTES SAUV AGES alliaire officinale bryone dioique capselle cirse ou chardon cruciferes ind. fetuque des pres fleole des pres fougeres fraisier gesse sans feuille graminees ind. grande marguerite la!che distique lampsane commune legurnineuses ind. lentille d' eau myagre perfolie nielle des bles onoporde acanthe ortie royale oseille/patience paturin commun renoncule scelerate renouee des oiseaux renouee liseron ronce setaire verte/verticillee silene vesce ind. vulpie ind. X X X - X X X - X X - X X xylologie - X - X - X - X - X X - - - - X X X X X X X X X - - - - X X - X X X - X X X X X X - - X - X X X X X X X X X X X X X - - X - - X textes - - - - - - - - - X X X X X X - - immediats. Ils appartiennent a quatre essences dont la plus representee est le chene (358 fragments); L'etude xylologique et anthracologique a ete faite par Anne Dietrich. L'etude carpologique a ete faite par Marie-Pierre Ruas, CNRS, Toulouse. L' exploitation du sol et du sous-sol autour d 'un bourg castral du Berry, du Xle au XVe siecle viennent ensuite le noisetier (29), l'aulne (3) et le saule (2). Une vingtaine d'especes cultivees ou consommees par I 'homme ont ete recensees grace a des grains conserves apres carbonisation ou a des elements conserves dans I' eau du puits (noyaux, pep ins, coques, graines). Ce sont des cereales (seigle, froment, orge vetue, millet, avoine), des legumineuses (pois, feves), des fruits (prunes, griottes et peches, raisins, pommes et peut-etre poires, sureau, mures, noisettes et noix), une pi ante textile (chanvre, atteste egalement par les pollens). Une vingtaine de plantes sauvages ont egalement ete identifiees: plantes qui se melent aux recoltes (nielle des bles, renouee des oiseaux), plantes de prairies (fetuque et fleole des pres, grande marguerite). Ces especes, par leurs caracteristiques ecologiques, sont des temoins de la nature des terrains ou elles se sont developpees. Ainsi, la presence permanente d'eau dans les Grands Fosses, au XVe siecle, avant leur comblement, mise en evidence par des analyses micromorphologiques, est confirmee par la presence de Ientille d'eau et renoncule scelerate. De plus, la flore trouvee dans Ies echantillons de ce fosse montre la presence de prairies de fauche et de pacages aux environs immediats du fosse. En revanche, les especes mises en evidence dans les echantillons provenant du puits sont des plantes de terrains incultes pietines, comme 1'etait la plate-forme de la motte. Un tableau general des especes determinees par ces differentes methodes, complete par les mentions de flore relevees dans les textes, a ete dresse. C'est ainsi que le !in, dont aucun vestige n'a ete trouve, est mentionne en 1500. Ce tableau met en evidence la complementarite des methodes utilisees. En effet, aucune pi ante n' est representee par to us ses elements constitutifs. Deux facteurs principaux interviennent sur la disparition des restes et doivent etre pris en compte dans !'interpretation des resultats. En premier lieu, la conservation dans le sediment varie selon les cas. Ainsi les bois ne se sont conserves que dans la zone immergee du puits, renfermant des vestiges datant du comblement de la plate-forme, done du XVe siecle; ailleurs ils se sont decomposes. D'autre part, les resultats d 'une discipline comme la xylologie sont directement lies au comportement de I 'homme et aux selections qu' il opere, alors que la palynologie ~chappe a cette influence. Les informations foumies par la palynologie correspondent a I' image globale de la vegetation locale et regionale, A. QUERRIEN, La viticulture en Bas-Berry au Moyen Age. «Cl os et vignes separees» autour de Levroux et de Moulins-surCephons, in: La vigne et le vin dans le Centre de la France, Revue de I 'Acadernie du Centre, 1995, 97-113. tandis que les graines, les bois et les charbons donnent une vision des plantes utilisees par I'homme et croissant vraisemblablement dans le voisinage. b. L 'agriculture et l'arboriculture Au debut du Xle siecle, !'atelier de potier est environne de champs ensemences en cereales, notamment en seigle, de prairies de fauche et de pacages. Des les premieres phases d' occupation de la motte (XIe-XIIIe siecle), la diversite des cereales cultivees apparait avec deux bles d'hiver (froment, seigle), et trois de printemps (avoine, orge, millet). I! y avait une rotation des cultures, entrainant une large repartition des travaux agricoles au long de I 'annee et la diminution des risques dus aux mauvaises recoltes. Cette rotation est confirmee par la presence de plantes adventices des cereales d'hiver (nielle des bles et myagre perfolie), et d'autres des sarclages et cultures d'ete (capselle, setaire et renouees). La polyculture permettait de disposer de diverses denrees stockables (grains, pois, feves ). Les rentes du seigneur s 'elevaient en 1500 a 17 setiers et 10,5 boisseaux de froment, 4 se tiers de seigle, 3 muids 10 se tiers 15,5 boisseaux d' avoine, 3 a 4 setiers de pois et feves W 41 r); ses deux metairies lui valaient chacune 4 a 5 muids de ble. La viticulture, assez developpee aux abords de Levroux jusqu'ala crise du phylloxera, vers 1885, occupait de petites surfaces dispersees dans le territoire de la seigneurie; le vin produit devait etre de qualite mediocre et consomme sur place 9 . Le seigneur exers:ait un droit de banvin de quarante jours, une fois par an, a la date de son choix; ce droit est maintenu jusqu'a la Revolution. En 1500, outre la production des vignes de son domaine qui pour 4 arpents lui rapportent de 12 a 15 pipes de vin par an, les dimes lui «valent de 8 a 10 pipes de vin par an», la pipe etant estimee a environ 220 litres (f' 26r). Les pepins retrouves dans la partie immergee du comblement du puits proviennent de grappes dont les grains ont vraisemblablement ete manges sur la motte. La presence d'arbres fruitiers dans des vergers, des potagers ou des vignes des environs est attestee par des restes de fruits egalement consommes au XV e siecle, mais les textes n 'en parlent pas. Dans les paturages situes le long des cours d' eau paissaient les vaches, les boeufs et les chevaux, tandis que les troupeaux de moutons parcouraient librement les terres non cultivees. Bois et forets, situes en limite du terroir, appartenaient essentiellement au seigneur. L 'exploitation des arbres pour la construction et pour le chauffage etait tres strictement reglementee, de meme que les "pessons" et la "glandee" pour l'alimentation des pores (F 29v). 135 A. Querrien LA FAUNE trouves en fouille mentionnes clans l'aveu de 1500 X X ANIMAUX DOMESTIQUES boeuf mouton ch{wre pore cheval ane chat chien poule oie X X X X X X X X - X X X - X X - X ANIMAUX SAUV AGES li<':vre lapin de garenne cerf sanglier chevreuil cygne X X - X X X X X - - X lNSECTES abeille ANIMAUX INTRUSIFS rat musaraigne campagnol taupe crapaud X X X X X - MOLLUSQUES moule des rivieres escargot de Bourgogne X - X X - X X POISSONS carpe anguille c. Animaux domestiques, e!evage, chasse et peche Aucun depotoir n'a ete retrouve au cours des fouilles, mais seulement des dechets culinaires disperses et peu nombreux. L'etude de la faune 10 a done ete limitee a un aperc;:u des especes presentes sur le site. L'essentiel appartient a des animaux domestiques, boeufs, moutons, chevres, cochons, poules, oies, eleves pour la consommation de leur viande, des laitages et des oeufs. Quelques os de cheval ou d'ane ont ete retrouves dans les niveaux anciens (XIe-XIIIe s.). L'aveu de 1500 indique que chaque annee le seigneur a droit a"sept vingt gellines" des "habistans et demeurant" en sa terre de Moulins, exception faite des exemptes de trois hameaux (f<' 11 v et 12r); ses 136 dimes lui rapportent "de quarante acinquante aignaux et dix a douze cochons (fO 26r)", et ses rentes 8 "oysons" (fO 41r) et 75 chappons (fO 42r), 76 gellines (en plus de celles de la gelliniere; fO 42r). I1 a en outre "le proffit du bestial qui peut estre nouris" dans sa metairie de Grange Rouge (f<' 38r). Les chats et les chiens sont representes par plusieurs individus, principalement du XVe s. Parmi les restes de chien figurent ceux d'un animal dont la taille au garrot etait inferieure a 20 cm. Ce squelette appartenait a un petit chien de compagnie, tels ceux que 1' ont voit sur la table du Due de Berry, dans I 'illustration du mois de janvier des Tres Riches Heures du Due de Berry. C'est un animal rare, qui correspond au statut aristocratique du site. Des animaux intrusifs ont ete reperes: des rats, des campagnols, des musaraignes, des taupes. La plupart des restes de ces micro-mammiferes ont ete trouves dans des pelotes de rejection de chouettes tombees dans le puits; ces pelotes comprenaient egalement de nombreux restes de batraciens. La chasse est evoquee par quelques os de lievres, de cerfs, d'oiseaux sauvages (cygne). Les os de lapin sont egalement peu nombreux, alors que l'aveu de 1500 note 1' existence de deux garennes seigneuriales a "connils". Le meme texte indique que sangliers et chevreuils etaient aussi chasses. Le seigneur avait un "maistre venneur" qui aidait les habitants du village de Thouez, situe aI' oree des bois du seigneur, a faire et diviser 1es «hais» necessaires pour que le gibier ne se perde pas, sinon ils lui devaient "pour chacun serf perdu ung thoreau en l'aage de trois ans, pour une biche une thaure en l'aage de deux ans, pour ung sanglier ung pourseau en l'aage de deux ans et pour une sangliere une truiee puinee en l'aage de deux ans (f<' 17 rv)". Quelques restes de carpes ont ete retrouves en fouille dans les niveaux du XVe si eel e. L' elevage de la carpe etait alors tres developpe dans la region et les 8 etangs du seigneur de Moulins en comptaient plus de 25000 en 1500. L'aveu de 1500 mentionne en outre les anguilles que les meuniers pouvaient pecher dans les biefs des moulins. Mais seules ces deux varietes de poisson sont attestees au Moyen Age, alors que la Cephons, comme le Nahon, portait bien d' autres especes qui etaient pechees et consommees, mais qui sont aujourd'hui detruites par la pollution de la riviere due aux produits chimiques utilises par les megisseries de Levroux 11 • L' aveu de 1500 evoque 10 L'etude de la faune a ete faite par Marie-Christine Marinval- Vigne, I' etude malacologique par Patrice Rodriguez. 11 La Federation de Peche de I'Indre enregistrait les especes suivantes, en 1985, dans le Nahon: carpes, gardons, truites fario, loches, vairons, anguilles, lamproies, epinoches, bremes, L'exploitation du sol et du sous-sol autour d'un bourg castral du Berry, du Xle au XVe siecle ces autres especes, mais sans les preciser, quand le seigneur rappelle son droit de pecher une nuit par an, de son choix, "toutes les anguilles et poison qui se pourront prandre esdits moulins et pescherie d'iceulx" (f' 13v). I1 semble qu'on ait egalement consomme des escargots dits de Bourgogne (Helix pomatia Linne): 132 individus de grande taille, calibres, ont ete retrouves en fouille. Un tableau met en vis a vis la liste des especes retrouvees en fouille et de celles mentionnees clans les textes. Aux especes precedentes, il convient d'ajouter les abeilles des ruches pour la cire et le miel.; en 1500, la prevote est affermee moyennant 25 livres de cire par an en 1500 (f' 38v) et les rentes rapportent 17 livres de cire (f'41 v), soit 42 livres en tout. d. Les amenagements hydrauliques et au seigneur d'Entraigues, avant de passer aux mains du seigneur de Moulins au XVIe ou, au plus tard, au debut du XVIIe siecle. Cette situation resta inchangee jusqu 'a la Revolution. Aucun etang ne fut vendu ni baille. Le long de la Cephons et de ses affluents etaient installes de nombreux moulins appartenant au seigneur du lieu et aux proches etablissements religieux. Plusieurs d' entre eux rem on tent au ha ut Moyen Age, et ont entra!ne !'adoption du toponyme Moulins atteste au milieu du Xe siecle. Entre le XIIIe et le XVe siecle, une quinzaine de moulins a fonctionne clans la seigneurie. La majorite etait des moulins a ble. Quatre etaient a tan et servaient a pulveriser l' ecorce de chene utilisee pour la preparation des cuirs; ils ont fonctionne jusqu' apres la demiere guerre, mais I 'utilisation des produits chimiques pour la teinture des cuirs a m is fin a I 'utilisation de I' ecorce clans I' industrie du cuir, et entraine leur abandon. Deux etaient a draps, dont l 'un etait en ruine avant 1623. L 'utilisation de martinets hydrauliques semble posterieure au Moyen Age clans cette region 13 • La terre de Moulins est traversee par la riviere de la Cephons et ses affluents. Elle renfermait de nombreux etangs qui ont presque tous ete asseches au XIXe siecle. On ne sait a quand remonte leur construction, les premieres mentions les concemant etant du )QIIe sieel e. Les etangs, onereux a mettre en oeuvre et a entretenir, constituaient une source de revenus importants qui interessa particulierement les seigneurs de Moulins. En effet, i1s etaient empoissonnes et peches regulierement, I' exploitation en etant faite directement par le seigneur et ses officiers 12 . Ainsi tous les grands etangs, de 22 a 60 ha, et presque tous les petits appartenaient au seigneur ou furent acquis par lui. Les seuls qui ne lui ontjamais appartenu sont celui du prieure fontevriste de Jarzay et les deux petits de la celle de Grandmont. Au XIIIe siecle, le seigneur de Moulins accorda un droit de peche sur ses etangs aux moines du Landais, mais on ne sait ni de combien, ni de quels etangs il s 'agit. En 1500, le seigneur avoue huit etangs qui sont peuples de 25000 carpeaux. Le fief de la Ferriere en a 3, attestes en 1623, appelles l'Etang de la Ferriere, l'Etang Vieil et l'Etang Neuf; ils sont alors passes clans la main du seigneur du Moulins, la Ferriere ayant ete incorporee au domaine seigneurial. De meme l'etang des Ysambert, mentionne clans un bail de 1438 (ADI, G 131 Grand Livre Noir du chapitre Saint Sylvain de Levroux, f' 13 ), appartenait au chapitre de Levroux Au fond du puits et lors de son creusement au milieu du XIe siecle, un cadre en bois, ou rouet, forme de quatre poutres de chenes, a ete dispose pour stabiliser les parois en pierre montees jusqu'a son sommet; c 'est le seul vestige de bois conserve de cette periode. Dans les niveaux inferieurs du puits, restes en eau depuis leur comblement, des pieces de bois gorgees d'eau ont ete retrouvees. Leur etude a permis d'identifier les arbres selectionnes au )QVe-XVe siecle par les artisans locaux. La majorite des objets et debris retrouves sont en chene. Ce sont des fragments de bois de construction (planches de moins d' 1 m de long, bardeaux, lattes), ou de mobilier (fonds de seau, douelle de tonneau, pointes, outil). Une planche sur laquelle ont ete prelevee des morceaux de forme circulaire ou ovalaire de 3 a 5 cm de diametre pourrait avoir servi a la fabrication de boutons, de pions ou de jetons. chevesnes, ablettes, perches, brochets. 11 n'y avait alors pas un seul poisson dans la Cephons (source: Rivieres Nahon et Cephons, diagnostic de qualite et programme de rehabilitation des milieux, octobre 1994). Les ecrevisses etaient egalement nombreuses dans la Cephons avant cette pollution. 12 La premiere mention d 'un maitre des eaux et fon-:ts a Moulins remonte a l'aveu de 1623. 13 A. QUERRJEN, La mise en oeuvre de l'energie hydraulique en Berry. Les moulins du bassin de la Cephons, in: Melanges oflerts aJ.M. Pesez, aparaitre. 3 L'utilisation des artisanats matii~res premieres et les a. Bois travailles et bois de chauffe 137 A. Querrien Les techniques de mise en oeuvre ont ete analysees. La taille se fait sur quartier, les planches perpendiculaires au fil se tordant beaucoup moins que les autres. Les outils utilises sont la hac he, I 'herminette ou le couteau. La fouille a mis aujour deux fragments de lames de scie, l'une acadre, !'autre egoi'ne. De nombreux fragments de noisetier ont egalement ete trouves; les jeunes branches, tres souples, ont pu servir aconfectionner des parois tressees. Enfin, les qualites du saule, bois tendre et fibres courtes ne donnant pas d'echardes, en faisaient un materiau bien adapte au toumage, comme l'atteste un fragment d 'ecuelle et un autre d 'un objet indetermine. Les bois choisis comme combustibles ont ete recenses a partir des prelevements de charbons de bois effectues clans les couches de destruction de !'atelier de potier et clans les niveaux d'occupation de la plate-forme: chene, aulne, hetre, peuplier, epinevinette. Excepte le chene qui, en buches, donne un feu long et de bonnes braises, ce sont de petits branchages qui brulent rapidement. Le faible nombre d'essences inventoriees (5) est le fait d'une selection par I' homme. Le hetre, par exemple, represente une part importante des charbons de bois, alors qu' il est absent des bois gorges d'eau. L'exploitation forestiere pour le bois de chauffage est attestee clans differents documents, que ce soit lors de conflits a propos de droits d 'usages concedes au chapitre de Levroux ou a d 'autres etablissements religieux, ou pour proteger le seigneur contre le chapardage: "quand aulcuns est trouve dedans mes bois de Dizeu couppant derosbant ou charoiant aulcuns bois dudit lieu, en ce cas les chevaux charrettes qui sont trouvez en iceux bois sont a moy confisquez, et s'ils n 'avoient charrette ne chevaux ils soient trouvez coupant et faisant dommages esdits bois, je !es puis constituer prisonniers et iceulx detenir jeusque a ce qu 'il aye baille caultion solvable, se soient soubsmis a I' ordonnances de ma justice et doibvent estre comdampnez en amande arbistraire a la discretion de mon bailly dudit lieu de Molins (f" 19v)". Mais aucune precision n'est donnee sur les arbres. Xylologie et anthracologie montrent !'exploitation des especes ligneuses des bois, des haies et des bords de riviere des environs du site; de ce fait, elles foumissent une liste d 'especes (7) beaucoup moins complete que la palynologie (23). Ce tableau resterait incomplet sans la mention de I' exploitation permanente des ecorces de chene pour les nombreux moulins atan (8) qui s 'echelonnaient le long de la Cephons et ou etaient moulues !es ecorces de chene necessaires aux tanneries de Levroux. Ces ecorces etaient decollees avec une levrette (cuillere en fer, au manche en bois ou en fer), au mois de mai, 138 quand le bois «sue», clans !es bois des environs et en particulier celui de la Ferriere, a I' ouest de Moulins. b. Textiles et cuir Un artisanat textile foumissait des draps de laine, de !in et de chanvre. I! est atteste par des redevances clans les aveux et denombrements apartir de la fin du XVe siecle: en 1500, le seigneur rec;:oit 40 a50 aulnes de toile pour ses dimes en !in et chanvre, et «demie cens de laines» pour les dimes de lainages (f" 26r). I! peut se deduire des etudes de faune et de flore. Des pollens de chanvre ont ete trouves en grand nombre (7 %) clans le sediment du cercueil monoxyle (620890), correspondant soit a un depot intentionnel au moment du depot du corps clans le cercueil, so it aune pluie pollinique au moment du transfert sur la motte au milieu du XIe siecle. La tres forte concentration de ces pollens a la base des Grands Fosses combles au XVe siecle ( 18 %) indique que le rouissage devait y etre pratique, ou a proximite. Plusieurs chenevieres sont indiquees clans les textes et portees sur le cadastre de 1812. Un seul objet de fouille se rapporte acet artisanat: une petite fusaiole en calcaire trouvee clans le puits. P!usieurs moulins a draps fonctionnaient le long de la Cephons dont celui du seigneur de Moulins, juste au nord de la Maison des Salles, en activite en 1500 et en ruine en 1623. Le travail du cuir etait tres developpe clans cette region. Levroux etait un grand centre de tanneries, et la Cephons portait plusieurs moulins atan. La fouille du puits n'a livre qu'une semelle de chaussure d'enfant en cuir, un bouton fait de deux brins de cuir entrelaces et un fragment de cuir non identifie, mais seul le fonds du puits offrait des conditions de conservation pour ce type d'objet et le puits, comble au XVe siecle, avait auparavant ete entretenu et sans doute cure a p!usieurs reprises, et n' etait pas un depotoir. Deux objets trouves en fouille pourraient se rattacher acet artisanat: une tige de fer du type alene (fin XIe-XIIIe s.), et un outil en chene cylindrique, avec une extremite taillee en biseau, evoquant une forme acuir (XVe s.). c. La metallurgie du fer Les fouilles ont mis au jour de nombreux residus d'une production metallurgique 14 locale representee des la premiere phase d'occupation de la motte au L'etude des elements relatifs ala metallurgie a ete realisee en collaboration avec Prof. Paul Benoit de I'Universite de Paris l. 14 L'exploitation du sol et du sous-sol autour d'un bourg castral du Berry, du Xle au XVe siecle .-· • Cour La Pomerie I I I I I I '-I ,/ I I I I ' I I I \ \ ', ,, !km I L 'Habitat d'apres l 'aveu de 1500 I H. DavuJ * bois etang + riviere • • chemin village etablissement religieux residence aristocratique metairie limite de la paroisse lieu-dit habite limite de la chil.tellenie moulin seigneurial bourg castral et siege de la paroisse moulin autre Xle siecle, jusqu'a son comblement au XVe siecle. Les elements retrouves sont des scories lourdes de forme massive, de 12 cm de diametre environ, qui ont la forme d 'un culot de four, et des scories beaucoup plus legeres, dont certaines presentent un aspect vi- treux et correspondent a ce que les metallurgistes actuels appellent des laitiers. Ces scories peuvent provenir de la reduction du minerai dans un four, ou du travail a la forge. Il est en effet souvent difficile de distinguer les scories provenant de ces differentes 139 A. Querrien operations. Dans les techniques employees au Moyen Age, leurs compositions chimiques sont comparables: il s'agit surtout d 'oxydes et de silicates. L'abondance de silice clans les scories, la diffusion de metal en formation visible clans un des culots scies, tendent a faire penser que des activites de reduction ont eu lieu, a proximite immediate de la motte, clans des structures de petite dimension. Ce travail a tres bien pu etre effectue clans une forge d'ou proviendrait une partie des scories. Ni le travail d'extraction, ni le travail de forge n 'apparaissent clans Ies textes, pas meme clans les redevances. Pourtant la toponymie temoigne d'une importante activite metallurgique sur le terroir: ]es Minieres, la Ferriere, Ies Forges, l'etang et le moulin de la Forge, le Foumeau, la Molline. La Ferriere et la Molline sont mentionnees clans l'aveu de 1500. Mais certains de ces toponymes sont posterieurs au Moyen Age. Ainsi les Minieres apparaissent clans les textes au XVIIe siecle: "la grange de deffunt Andre Villede appelee de present les Minnieres" (aveu de 1623, f' 23v et 27r). L'etang et le moulin de la Forge ont change de nom apres 1646, date a laquelle ils s 'appellent encore Roe (aveu de 1646, Saint-Aignan). Le minerai de fer etait exploite par poche, a ciel ouvert, clans une large zone situee essentiellement au sud-ouest de la motte; une aire de grillage du minerai a ete fouillee clans les bois de Thouez, a !'est de Moulins, mais aucune datation n'a ete obtenue 15 • Les prospections menees clans un rayon de six kilometres autour de la motte ont permis de mettre en evidence !'existence d'un certain nombre de sites ou du minerai est present. Sur les 5 echantillons envoyes a !'analyse, 4 ont des teneurs et des qualites suffisantes pour etre utilises comme minerai. L 'un a ete trouve clans les terres de remblai de la plate-forme, un autre a 300 m (oolithes ferrugineuses); les autres proviennent soit du lieu-dit les Minieres a 800 m, ou les agriculteurs trouvent encore des blocs d 'hematite, soit du bois de la Ferriere a 2,5 km. L'analyse montre que ces minerais se composent essentiellement de silice et d'oxyde de fer, c'est-a-dire qu'ils sont relativement faciles areduire. Ils appartiennent ala categorie des minerais d'alteration superficielle: ils proviennent de 1'action de 1'atmosphere, clans des conditions climatiques de type tropical, sur les couches superficielles de 1'ecorce terrestre. Ces minerais ont d'ailleurs ete retrouves clans des petits gisements de 15 P. TOURNAlRE, Bois de Thours (sic), Chronique des fouilles medievales, Arc!teologie medievale XV, 1985, 306. 16 L'analyse des elements chimiques a ete effectuee au Centre de Recherches Petrographiques et Geochimiques du C.N.R.S. a Nancy (C.R.P.G.). Les donnees sont traitees informatiquement et comparees avec des resultats obtenus par ailleurs (Programme 140 surface ou de tres faible profondeur. Faciles a reconna1tre, faciles a exploiter, composes d'un minerai aise a reduire, ils correspondent au type de gisement que le Moyen Age a mis en valeur. Les analyses chimiques 16 ont m is en relation minerai et scories. Bien que les teneurs en fer des echantillons de minerai varient, ils appartiennent tous a un meme type. Oxyde de fer et silice en sont les principaux composants; l'alumine est a peu pres totalement absente et la chaux tres rare. La teneur en chaux des scories, superieure a celle des minerais, prouve 1'usage de calcaire ajoute comme fondant a la charge du four. Ce minerai ainsi defini par ses elements majeurs, est parfaitement compatible avec les scories retrouvees. L 'etude des elements moins abondants, elements mineurs ou traces, confirment cette affirmation. Aussi sans qu'on puisse dire avec exactitude de que! site d'extraction provient le minerai, il est certain que la metallurgie de Moulins-sur-Cephons a trouve sur place, clans un rayon tres restreint, les minerais necessaires. d. Materiaux de construction et carrieres Au sommet de la motte, le premier talus est couvert de silex, et le deuxieme de calcaire de Buzanc;;ais, les deux parements tranchant totalement 1'un par rapport a 1'autre. Les silex du premier tal us se ramassent a meme le sol clans les champs des environs, surtout au nord du village, sous la forme de galets; il semble qu'ils aient ete casses lors de leur utilisation sur la motte. Certains eclats laissent penser que des elements d 'un atelier de taille contemporain de 1'enceinte du chalcolithique ont ete utilises pour ce parement medieval. Le calcaire de Buzanc;;ais provient lui de carrieres a ciel ouvert, comme il en existait de nombreuses sur le plateau de la Champagne, au sud de Moulins (carte geologique de la France, BRGM, au 1/50000, Levroux, J?h). Dans la basse-cour principale se dresse I' eglise dont les deux phases de construction sont reperables au simple examen du chevet. La premiere eglise est construite, a la fin du XIe ou au debut du Xlle siecle, en gres glauconnieux vert et gres ferrugineux rouge, au dessus d'une assise en calcaire et gres durs. Une deuxieme nef lui est accolee au sud, a la fin du XIIe ou au debut du XIIIe siecle. La pierre blanche alors ARTEMISE-SCORIES). Les analyses foumissent uniquement le pourcentage des elements qui composent l'echantillon, quelque soit leur forme; ainsi elles donnent la quantite de fer globale, aussi bien le fer metallique que les oxydes ou les silicates de fer contenus clans l'echantillon. L'exploitation du sol et du sous-sol autour d'un bourg castral du Berry, du Xle au XVe siecle a utilisee, une craie silex de provenance plus lointaine (C 3a), met en evidence cet agrandissement et est adaptee aux sculptures que la nouvelle partie, plus soignee, reyoit pour decorer les pilastres du choeur et le nouveau portail. La construction de la plate-forme de la motte du XIVe-XVe siecle a utilise des gres roses et des gres ferrugineux (C 1-2a), provenant du nord de Moulins, du calcaire mameux huitres, du calcaire piquete de Levroux (J 6-7a) et du calcaire de Buzanyais, provenant du sud. Tous ces materiaux ont ete exploites dans des carrieres ciel ouvert de petites dimensions, situees dans les environs. Les memes types de materiaux se retrouvent dans les autres batiments du village de la fin du Moyen Age. La tour du chatelet d'entree du bourg, attribuee au XIVe siecle, est faite principalement de moellons de gres ferrugineux, et d' autres de gres rouge; elle comprend aussi des calcaires et des silex. Les restes de la maison seigneuriale intra muros du debut du XVe siecle allient calcaires et gres. A 1'interieur, au premier etage, une fenetre a coussiege et une cheminee sont faites de calcaire aise sculpter, sauf les corbeaux de la cheminee qui sont en gres. Une porte surmontee d'une accolade, qui a ete reutilisee dans le reconstruction de cette maison est en gres vert. Dans la Maison des Salles, nouvelle et demiere residence des seigneurs de Moulins du milieu du XVe siecle, on a utilise du calcaire de bonne qualite, propice la taille, pour la cheminee de la gran de salle de 1'etage et pour les fenetres meneaux et coussieges. La plupart des constructions etudiees ont utilise des materiaux heterogenes et locaux extraits pres du village et aptes la fabrication de moellons. Gres et calcaires etaient issus de banes superficiels, faciles d'acces et abandonnes apres epuisement du gisement. Seuls les calcaires silex blonds et noirs utilises pour 1'eglise avaient une origine plus lointaine: ils provenaient des carrieres a ciel ouvert et souterraines a Luyay-le-Male connues autrefois pour leur tuffeau, 16 km au nord-ouest de Moulins. La documentation ecrite ne se preoccupe guere de ces activites. L'aveu de 1500 mentionne la "perriere" (f" 24r) pres du chemin de Moulins a "Coqu", vers l'ouest, ou la carte geologique indique une ancienne carriere de craie a ciel ouvert, et la "sabelonniere ou a a a a a a a a s'estire le sable" (f" 3v) pres de la Fontaine SaintSulpice, Balzeme, au nord. En 1766, une piece de terre appelee les "sablieres" est "partye en culture et partye en friche" la Charonnerie, 1'est de Moulins. La pierre devait etre reservee aux constructions aristocratiques et aux edifices religieux. Lorsque la maison seigneuriale intra muros est abandonnee et detruite, !'exception de son pignon nord, la maison de notable qui est construite sa place 17 a sa base en moellons et l'etage en pans de bois et torchis, selon la technique qui devait etre couramment utilisee pour les maisons paysannes. Les materiaux necessaires ce type de construction etaient disponibles sur place et peu onereux a se procurer et a mettre en oeuvre. a a a a a a e. Tuileries et poteries Plusieurs types d'argile ont ete exploites autour de Moulins. La region est en effet riche en argile provenant des couches secondaires du cenomanien. Cette argile commune, qui donne des pates de couleur soutenue, orange, rouge gris-noir, apres son passage au four, etait destinee ala fabrication de tuiles, carreaux et briques, dont les couches de destruction du XVe siecle ont livre de nombreux exemplaires sur la motte. La repartition des toponymes tels que 1'Argillerie ou les Ardiller (2), les Fosses, la (ou les) Tuilerie, suit la couche cenomanienne sur la carte geologique, sauf au sud de Levroux, ou les be so ins de la ville ont suscite la creation d 'une tuilerie dans les fauxbourgs de la ville, la matiere premiere etant apportee et non extraite sur place. Mais ces etablissements ne sont pas dates. Les tuiles devaient etre reserves aux edifices religieux et aux maisons de notables. Bardeaux, branchages et chaume etaient utilises pour les couvertures des maisons paysannes et des batiments agricoles, comme la bergerie du seigneur Beaugibier qui, en 1500, est "couverte de bardeau et les comiers de thuille (f" 34r)". En revanche, les potiers du debut du Xle siecle utilisaient une argile specifique, donnant une pate claire la cuisson. Cette argile comporte une forte proportion d'alumine (Alz0 3), et de faibles pourcentages de Fep 3 et de Kp, caracteristiques des argiles kaolinitiques 18 . Or les affleurements de kaolinite sont a a a A. QUERRIEN, Une maison de notables du XVe siecle a Moulins-sur-Cephons (Indre), in: J.-M. Pesez (dir.), Cent maisons aparaltre. 18 Les analyses chimiques ont ete realisees par Daniel Dufournier (CNRS, CRAM, Caen), et les analyses physiques par le regrette Claude Lorenz (Universite de Paris VI) et Annie Blanc (Laboratoire des Monuments Historiques). Tableau resumant les resultats des analyses chimiques des ceramiques de I'atelier de potier de Moulins; moyenne et ecart-type de la serie, pour 30 echantillons analyses: SiO AI 0 Fe 0 TiO CaO MgO Na 0 K 0 MnO moyenne: 70} 23;8) 3,J5) 0,71 2 0,45 0,43 o,ib 0,68 0,02 ecart-type: I ,9 I ,9 0,30 0, I 0 0,12 0,07 0,07 0,17 0,007 17 nuidieva/es, 141 A. Querrien exceptionnels dans cette region, si on les compare aux possibilites d'extraction des nombreuses argiles locales cenomaniennes. Les niveaux d'argile ou des carrieres ont pu etre ouvertes ont ete reperes sur la carte geologique. Des prW:vements ont ete effectues et des echantillons ont ete cuits. Des analyses physiques (lames minces) et chimiques ont permis de comparer ces argiles acelle des poteries et d'identifier le gisement. Il s'agit d'une argile tertiaire, qui etait extraite a 5 km au sud-ouest de !'atelier de Moulins, dans un lieu appele le "Terrier Blanc". Ce choix correspond a des constatations faites dans d 'autres parties du Berry ou les depots tertiaires foumissent des argiles refractaires et des argiles a poteries. Le materiau etait de qualite et permettait d'obtenir une pate de couleur claire a la cuisson, preference que 1'on observe dans beaucoup d' autres regions a cette epoque. La distance entre 1' argiliere et 1'atelier n 'etait pas un reel handicap, malgre les difficultes des transports de l'epoque, car il suffisait d'une ou deux charretees de terre pour qu'un potier puisse travailler pendant un an. L'emplacement de !'atelier a ete determine par I' existence prealab le d' un groupement humain. Les textes sont muets au sujet de l'artisanat ceramique. L'archeologie a non seulement mis en evidence 1'atelier du XIe siecle, mais elle a prouve qu 'un artisanat ceramique continuait, si ce n'est sur place, du moins dans les environs immediats. L'analyse des pates des ceramiques montre en effet que la meme argiliere a servi du XIe au XVe siecle. Ces etablissements artisanaux, comme pour la metallurgie, posent le probleme des rapports entre artisanats et seigneurie. s 'agissait-il d' etablissements autonomes, ou bien dependaient-ils directement de la seigneurie? La construction de la motte a-t-elle entra'ine la fin de I' atelier, ou bien celui-ci s 'est-il developpe plus loin? Que! etait le statut social de ceux qui travaillaient a produire les ceramiques? Moulins etait un petit chef-lieu de chiltellenie, dont le territoire etait reparti de maniere equilibree entre terres cultivees, paturages, bois taillis et forets. Sa population fournissait l'essentiel des produits necessaires ases besoins, que ce soit dans le domaine agricole ou artisanal. Poteries, tuiles, objets en fer, en cuir ou en bois, toiles de laine, de chanvre et de lin etaient fabriques sur place. Mis apart quelques objets de dinanderie, bijouterie et tabletterie, et des produits manufactures provenant d'ateliers specialises pour l'armement et l'equipement du cavalier, la grande majorite du materiel mis au jour tors des fouilles provient de maniere certaine, ou peut provenir, d'une fabrication locale. Les materiaux necessaires sont disponibles et exploites dans les environs du village, et, dans le cas de la poterie, la cha'ine materiauproduit fini a pu etre reconstituee. Toutefois, il ne semble pas qu' il y ait eu de verrerie aMoulins, mais un etablissement verrier, dont on ne connait pas 1'origine, se trouvait a 6,5 km au nord-est, dans les bois du seigneur de Levroux. Certaines trompes en terre cuite et quelques vases des niveaux recents de la motte furent peut-etre achetes ades marchands, dans les foires et marches des environs, ainsi que des anneaux en alliage cuivreux et en bronze dore. Une bague en argent gravee d'un tau de Saint Antoine entoure de trois etoiles a sans doute une origine plus lointaine. Mais ce sont des exceptions qui contraste avec l'homogeneite de !'ensemble du materiel retrouve. L 'enquete me nee a Moulins illustre 1' interet de combiner diverses sources pour tenter de pallier les insuffisances des unes et des autres. La documentation ecrite, d' origine seigneuriale lai:que et religieuse, est orientee vers les interets seigneuriaux. L'archeologie met en lumiere des documents d'essence differente. Dans une region ou la documentation fait presque totalement defaut avant le XIIIe siecle, elle reste la source privilegiee pour la decouverte et 1' etude des sites anterieurs. Elle 1'est egalement pour aborder les differents aspects de !'exploitation des terroirs et de la transformation des ressources naturelles. Armelle Querrien C.N.R.S. 103 bd. De Magenta 75010 Paris France 142 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 /sabelle Catteddu Le site medieval de Saleux "les Coutures": habitat, necropole et eglises du haut Moyen Aget Situe en rive gauche de la Selle (seul affluent navigable de la Somme ), ce site de gage sur 4ha a livre de nombreuses structures d 'habitat, cabanes excavees, silos, fosses, biHiments et trous de poteaux, s'organisant autour d'une importante necropole (plus de 1500 individus ont en effet ete degages et etudies). C'est au centre de cette necropole, qu'ont ete decouverts les restes de fondation de deux edifices religieux successifs: 1'un sur poteaux, 1' autre sur fondations calcaires. L'importance de ce site a ete renforcee par la decouverte en bordure de la riviere, d'un amenagement de berge avec structures en bois conservees 2• Les differents indices chronologiques qui nous ont ete fournis lors de la fouille temoignent d'une creation aux alentours du VIIeme s. ap. J.C. et d'une occupation s 'etendant jusqu' au Xleme s., sans discontinuite apparente. L'habitat etait subdivise par des fosses en plusieurs parcelles geometriques au sein desquelles se developpaient les structures. La mise en place de la necropole et du noyau initial de l 'habitat semble etre simultanee clans le courant du VIIeme s .. Si les secteurs degages clans ce fond de vallee restent occupes jusqu'au Xleme s., ils subissent des modifications importantes clans le decoupage de l'espace deja habite. En effet, au terme de notre etude, nous avons pu reconnaitre une unite principale en bordure de riviere au VIIeme s. Une petite necropole s 'organisait a 1'ecart, au SO, autour d'une tombe privilegiee en sarcophage, protegee par un petite edicule en bois. Au cours des VIIIeme-IXeme s., une eglise en bois est construite autour de ce meme edicule. On assiste alors a un accroissement de la necropole et a l'amenagement d'un enclos funeraire, comble autour La repartition des structures d 'habitat semble repondre a une volonte d'amenager des zones a vocation specifique. Alors que les structures de stockage se concentrent clans la partie E du site, les batiments sur poteaux s 'etendent plus nettement au N. Des traces d'activites domestiques ont ete reconnues. Si aucun four (artisanal ou domestique) n'a ete retrouve clans les limites de notre decapage, quelques foyers isoles ont toutefois ete releves malgre un etat d' arasement assez important du site. Temoins de 1'activite domestique: un mobilier ceramique bien represente, des dechets de scories,du rnobilier en os et lithique bien caracterise. Plusieurs fonds de cabane ont livre des amenagements internes et un mobilier temoignant de la pratique du tissage: fusa!oles, broches, ou latteurs a pointe, peignes carder, aiguille en os, lissoir en verre et pesons. Ces fonds de cabane presentaient des cavites regulieres marquant les emplacements d 'un metier a tisser vertical. L' artisan at de 1' os est bien atteste. Comme clans beaucoup de sites ruraux, cette activite devait etre pratiquee familialement clans chaque foyer, au fur et a mesure des besoins. Des objets varies sont fabri- Le site des coutures a ete decouvert lors de sondages realises sur le tron9on de I' autoroute A 16, reliant Paris a Amiens. La fouille s'est deroulee durant 9 mois, de mars anovembre 93, sur la commune de Saleux en Picardie, a5 km au SOd' Amiens. Cette fouille a fait l'objet d'une convention signee entre l'amenageur Sanef, I' Afan, et l'Etat (Service Regional de I' Archeologie de Picardie). 2 C'est clans ce contexte, que nous avons fait intervenir un maximum de specialistes en sciences paleoenvironnementales et connexes, et ce, des le debut de la fouille. du IXeme s. mais toujours present clans le paysage. L'habitat suit lui-aussi un mouvement d'extension. Au cours des IXeme et Xerne s. 1'habitat s 'etend sensiblement au N et au S. Le Xeme ou XIeme s. voient la construction de 1'eglise sur fondations calcaires, en plusieurs etapes. La necropole continue de s 'accroitre et l 'habitat englobe progressivement la necropole en s'approchant de l'eglise. L'habitat a 143 I. Catteddu ques en os et en come: tabletterie, manche de couteau, peigne, elements vestimentaires ... L' elevage et la culture sont bien illustres par les echantillons carpologiques 3 et archeozoologiques 4 • La diversite des especes vegetales exploitees (polycultures) indique une gestion equilibree des terres arables et des ressources naturelles, forestieres. Le mode de stockage choisi semble etre le silo excave. L 'exhaustivite de 1'echantillonnage carpologique a permis 1' obtention d'un spectre tres etendu et une bonne caracterisation des apports des vegetaux sur le site. On constatera la predominance du ble tendre, ma is egalement la presence de seigle, d, orge vetue, d' avoine, d' epeautre et de legumineuses. M is a part I' avoine, ces cere ales etaient sans doute conduites en cultures d'hiver comme en temoigne le cortege important d 'adventices d 'hiver. La cueillette jouait egalement un role comme l' atteste la grande variete des petits fruits sauvages cueillis dans la foret. La majorite des apports de plantes reflete soit des cultures, soit des milieux ruderalises locaux, ou des paturages ou paissait le be tail. Le milieu forestier n 'est atteste que par les produits de la cueillette. En ce qui cone erne 1' elevage, on constate une importante representation du boeuf et des caprines, puis du pore et surtout du cheval. Les proportions de boeuf et d'equides correspondent a un important besoin en matiere d'utilisation de force animale. Pour la foumiture de produits cames, le faible apport du au pore est compense par un elevage de bovins largement toume vers la production de viande. Le role des caprines etant de produire du lait ou de la laine. Le mouton vient par ailleurs en seconde position a la place du pore habituellement a cette place. Une grande stabilite des choix economiques peut etre vue au travers de I'alimentation et de 1'elevage, saufpeutetre pour les caprines. La berge L 'habitat eta it limite a1'E par la riviere de la Selle et par une zone humide qui livra a la fouille des amenagements complexes. Pour une meilleure comprehension de ce secteur, nous avons fait appel a une equipe de plongeurs afin de realiser une fouille subaquatique 5 • La presence d'un ancien meandre reconnu comme une zone instable et marecageuse, et la proximite du versant E constituent des zones d'instabilite peu pro- Etude de V. Mateme, AFAN. Etude de J.H. Yvinec, C.R.A.V.O. Compii:gne, F. Groupe de Recherches Archeologiques Subaquatiques de la 144 pices a l 'installation humaine. Toutefois les vestiges decouverts ala fouille attestent une volonte de controle de ces contraintes et une reflexion sur les transformations permettant a la fois une gestion de I' environnement et une exploitation ades fins economiques. Cette volonte est cependant soumise a des reflexions complexes qui mettent en jeu des mecanismes au sein d'unites fonctionnelles, clans la vallee de la Selle. Au moins cinq actions anthropiques ont ete reconnues: - une chenalisation sur environ 100 m et une zone gueable. A cet en droit 1'endiguement lateral aura it du entrainer des risques de debordements, mais la presence de biefs dans ce contexte de chenalisation perrnet de trouver cet equilibre; une anse naturelle (ancien meandre de la riviere) remaniee par un remblais anthropique: celui-ci joue un role d'endiguement vertical et facilite l'acces a la riviere; l'amenagement d'un bief d'amont, partant de la riviere vers le site et les rangees de pieux; un bief d' aval, partant d 'une vanne et deux rangees de pieux bordes de planches posees de chant, vers la riviere. Ces biefs seront endigues lateralement al'aide de remblais afin d'eviter les divagations laterales et les debordements vers 1'habitat; l'amenagement d'un vase d'expansion (bassin) endigue lateralement, qui s 'elargit entre les deux biefs et qui debouche sur la vanne. Cette demiere perrnettait de control er le debit de 1'eau qui s 'engouffrait entre les alignements de pieux bordant le depart du bief aval. On remarquera ici la volonte d'utiliser la force tractrice del' eau, d' accelerer le cours d'eau et de provoquer une force mecanique. Si les amenagements decouverts aSaleux, ont ete realises en partie, afin de faire face aux contraintes topographiques du lieu et de les controler, les volontes economiques sont evidentes (peche, vivier, moulin ... ). D'autre part, la chenalisation associee au gue (exhaussement du lit) augmente la force tractrice de 1' eau utile aune navigation avalante. Un gue de cette importance annonce egalement un lieu de passage tout aussi consequent. Les echantillons dendrochronologiques 6 synchronises ont restitue une moyenne de 148 annees consecutives. La date d'abattage etait comprise entre 763 et 778. Les resultats palynologiques7 sont tres constants. L 'espece la mieux representee est le houblon, associe a un couvert d'aulnes, saules et frenes typiques des milieux humides voire des tourbieres. Somme, equipe dirigee par M. Sueur. V. Bemard, Besanc;:on, AFAN, F. A.M. Munaut, Universite Catholique Louvain-La-Neuve, B. 6 Le site medieval de Saleux "les Coutures": habitat, necropole, et eglises du haut Moyen Age X X / X X \ X X \ X X X X .... ·v·S· q,: .• 10 . X t~ . ,, . I . ..,"' f/) Q) .... :J :::> '5 w 0 X ....J () <( Cl) V) Q) ....J " <D 8' E z E a.: 0 ~ I <I; (") 0 ,; .c u " m 0 X ~ ::>" ,..."' '0al 0 ::: (\j OJ '" (.) -' 0 145 I. Catteddu En dehors de cette zone fluviale, les environs de 1'habitat carolingien devaient avoir 1'aspect d 'un paysage ouvert ou croissaient quelques pins epars aux abords de prairies et champs. Prairies et cultures cerealit':res sont bien re presentees dans 1'environnement du site. La necropole Elle regroupait 1192 tombes pour un total de 1500 a 2000 individus, s'organisant a l'interieur d'un espace delimite par un fosse d 'enclos s 'integrant parfaitement dans le reseau fossoye de !'habitat et du parcellaire environnant. Les tombes etaient disposees autour des edifices religieux, mais sont egalement recoupees par ceux-ci. Six sepultures ont ete releguees au milieu de 1'habitat. L'etude de cette necropole comprenait egalement celle des modes d'inhumation et de 1'architecture des tombes, ainsi que des aspects anthropologiques 8 et paleopathologiques9 • L'importante densite des inhumations et le nombre de reductions, les nombreux recoupements et la rarete du mobilier ont gene considerablement les tentatives de chronologies absolues ou relatives et 1'etude generale de ces sepultures. Les pratiques funeraires Plusieurs types de contenants ont ete reconnus: cerceuil cheville, coffrage de bois maintenu par des calages de pierre ou de terre, et monoxyle. Les coffrages et cercueils sont de forme rectangulaire ou trapezoi'dale et sont plus souvent etroits que larges. La presence d 'un monoxyle atteste 1'usage de ce mode d'inhumation a Saleux. Quelques inhumations en linceul ont ete proposees. Les sepultures en pleine terre sont nombreuses. Les fosses anthropomorphes se presentent sous deux types, soit une logette externe creusee dans le sediment a 1'extremite de la fosse, so it un calage a 1'interieur de la fosse ou du coffrage de bois. Les calages sont plus nombreux pour les inhumations en pleine terre. I! existe cependant des cas de calage cephalique dans les coffrages de bois ou lorsque les individus sont inhumes en linceuls. L 'etude de 1'espace de decomposition met en evidence 1' incidence de la nature structurelle de la tom be sur la position des squelettes, en particulier des membres superieurs. La position flechie des membres Etude d'identiftcation de N. Moreau et L. Staniaszek (AFAN) 146 inferieurs semble correspondre a des positions des membres superieurs asymetriques. Elle semble egalement relative ades espaces vides. Les mouvements peuvent etre dus au transport du coffre. Quelques cas asymetriques en espace colmate peuvent laisser sceptique. Les attitudes donnees aux defunts sont-elles rectifiees apres le depot? Les observations faites sur le marquage des sepultures soulevent la question de la preservation de la sepulture. A Saleux, plusieurs groupes de sepultures partiellement juxtaposees ont ete mis au jour. La disposition des individus et les relations stratigraphiques suggerent que des marquages au sol (le plus sou vent en matiere perissable ), indiquaient 1' emplacement des inhumations. Les reductions et ossuaires sont tres nombreux et se rencontrent dans !es secteurs les plus densement occupes de la necropo!e: au S, a1'0 de la chapelle et dans la nef. Ils se composent des os d'un ou plusieurs individus. Deux ossuaires plus importants sont a noter a l'E de la necropole. Typologie et organisation des differents modes d 'inhumation Le sarcophage decouvert au centre de la necropole est unique dans cet ensemble. Les contenants de bois sont en majorite anterieurs aux sepultures en pleine terre. Ils se repartissent sur !'ensemble de la necropole mais se rarefient en peripherie de celle-ci. Les contenants sont egalement nombreux aproximite immediate de la chapelle avec une concentration notable au S du choeur. Des coffrages ont ete identifies avec certitude, en fonction des calages lithiques, les autres bloques avec de la terre n 'ont pu etre differencies. Les sepultures en pleine terre se repartissent sur 1'ensemble de la necropole, plus clairsemees cependant dans les zones peripheriques. Il faut egalement noter que les sepultures de ce type ceinturent le secteur funeraire. Elles presentent des variations de formes de fosses et de positions de membres superieurs. On remarquera egalement la posteriorite des individus dont les mains sont en avant du thorax sur les autres sepultures en pleine terre. Elles se trouvent surtout dans la peripherie E et S du choeur. Les fosses anthropomorphes ou a calage cephalique semblent posterieures aux sepultures en espace vide. Etude des Docteurs C. Obry, E. Marasset, F. Berlemont, Clinique V. Pauchet, Amiens, F. Le site medieval de Saleux "les Coutures": habitat, necropole, et eglises du haut Moyen Age Problematique de la chronologie de la necropole de Saletu L'handicap majeur dans !'etude de la necropole de Saleux a reside essentiellement dans la problematique chronologique. Si plusieurs sepultures ont pu etre datees assez precisement par le mobilier funeraire, les edifices de culte, le mobilier de comblement et les structures d 'habitat ont foumi des references pour 1' etablissement d 'une chronologie relative. La chronologie relative des tombes entre elles repose surtout sur les recherches menees autour des orientations et des edifices de culte. Ce sont elles qui ont, le mieux, permis de degager la plupart de nos conclusions. Cette etude doit encore etre affinee et renforcee, a la fois a travers un travail interne a la necropole, mais surtout a 1' aide de comparaisons avec des sites offrant la meme problematique, ce qui a ce stade de notre travail et dans les delais impartis n' etait absolument pas realisable. A Saleux, !'utilisation des orientations est importante dans la chronologie ou la topochronologie de la necropole, mais elle ne peut etre systematique. Un role directeur dans la dynamique spatiale est a attribuer sans conteste aux edifices religieux et au sarcophage. Et cela, tant au niveau des orientations, que de 1'organisation spatiale en general. Toutefois, de nombreux facteurs humains (topographie, regles intemes aune population. 00) ont pu influencer 1'orientation de certaines tombes. Un grand nombre d' elements topographiques importants nous echappent en effet totalement: haies, palissades, levees de terre, chemins, etc ... A Saleux, 1'organisation spatiale des sepultures no us a toutefois permis de distinguer des acces aux edifices, et plusieurs espaces de circulation. La population presente aSaleux Les etudes anthropologiques et paleopathologiques ont livre des informations sur la representation de la population, mais aussi sur les agressions environnementales, sur la prise en charge, la malnutrition, cyclique ou non, !'hygiene ... L' etude des indicateurs non specifiques de stress (hypoplasie de l'email dentaire et lignes de Harris) sur la population de Saleux revele des carences assez repandues (carences et maladies infectieuses ou meme traumatiques) l'origine de troubles de croissance. L' etude biologique de la necropole, favorisee par un excellent etat de conservation et la presence d' un echantillon complet, a pemis de composer des images precises de cette population. Ainsi la determina- a tion sexuelle fait appara!tre un pourcentage d 'homme et de femmes apeu pres equivalent face a un nombre important d' indetermines. Ces resultats sont relativement normaux dans le cas d'une population nature lie. L 'age au deces revele une proportion plus importante d'immatures par rapport aux adultes. L'age de mortalite le plus represente chez les enfants etant place entre 2 et 5 ans. Si l'etude osteometrique montre une grande heterogeneite au sein de la population, quelques traits dominants ont toutefois pu etre degages. L'etude pathologique foumit de nombreux elements d'information. La frequence des atteintes de la sphere bucco-dentaire par exemple, va dans le sens d 'une hygiene plutot mediocre. Les cas d 'arthroses et d'enthesopathies sont plus qu'abondants. Ces lesions degeneratives touchent le rachis, alors que certains individus sont encore tres jeunes. Les fractures des avant-bras, et sur le cote gauche, s'ajoutent a celles des membres inferieurs. Certaines sont assez remarquables et ne sont pas toujours tres bien traitees. Les cas d'arthrite sont egalement illustres. Cette etude confirme que la population de Saleux souffrait de carences et de troubles de croissance. A Saleux, !'etude de !'organisation du cimetiere implique non pas une vue statique mais dynamique de !'ensemble. Ces recherches ne peuvent etre menees que dans le cadre d 'une fouille rigoureuse, et suppose une bonne connaissance chronologique des tombes. En effet, ceci exige la reconnaissance de la fas:on dont les tombes appartenant aun meme ensemble topographique ont ete disposees au cours du temps: s 'agit-il de simples juxtapositions ou existe-til des surfaces reservees pour un groupe? La fa<;:on dont les diverses reductions de corps ont ete effectuees peut-etre significative. L' approche paleopathologique de type epidemiologique, prenant aussi en compte les differents indicateurs non specifiques de "stress" a quanta elle ete realisee. 11 est evident que ces regroupements familiaux ou groupes selon l'age ou le sexe, ou la condition sociale ne vont pas concemer toutes les tombes. Les excellentes etudes realisees a ce sujet, par des historiens, retracent a partir des textes et des fouilles, les differents comportements de l'homme devant la mort, et !'application des diverses mentalites aux necropoles. Nous n'oublierons done pas que les emplacements vont dependre, aussi, de facteurs plus materiels: la topographie du cimetiere, le sens de son developpement, le hasard, ou d'une raison inconnue, des facteurs topographiques, historiques, culturels, socioeconomiques, et surtout religieux ... 147 I. Catteddu Les edifices de culte Si tous les elements precites ont leur place, les eglises et 1'ensemble du contexte religieux occupent sans conteste a Saleux une place primordiale Au Vlleme s. et probablement durant une partie du VIIIeme s. les tombes s'organisent autour d'un sarcophage. Certaines tombes ont conserve le mode d' organisation en rangees. Le sarcophage eta it protege par un edicule en bois dont subsistent les traces. Son caractere unique au milieu de toutes ces tombes attire !'attention d'autant qu'il occupe une position plus elevee par rapport au reste du site et qu'il sera au centre des edifices religieux posterieurs. Son orientation est par ailleurs celle des sepultures les plus anciennes. Plusieurs d'entre elles ont livre du mobilier funeraire datable des Vlleme et VIIIeme s. Directement au N du sarcophage, une sorte de seuil etait amenage dans le tuf. La tombe primitive avait par ailleurs ete soigneusement videe de tous ossements. Quelques elements de mobilier funeraire etaient eparpilles au pied et en partie dans le sarcophage (agrafe a double crochet, bague a chaton lisse, perle en pate de verre, elements de tabletterie avec decor de croix inscrite ). Selon Charles Bonnet, 1'apparition de ces structures etablies a proximite ou sur des sepultures est a placer plus ou moins tard dans le Haut Moyen Age, selon que 1'on se trouve dans la vallee du Rhone et au sud des Alpes (des le Veme s.) ou dans les pays plus au N. (fin Vleme- Vlleme s.). Lors de !'edification des eglises ou des chapelles, elles sont transformees en chapelle annexe ou inclus dans les murs. Le souvenir est maintenu. A Saleux, c 'est vers le VIIIeme s. qu'une eglise en bois est construite a !'emplacement meme de l'edicule et autour de celui-ci. L'acces etait probablement axial, a la fac;ade occidentale: deux doubles inhumations sont amenagees juste dans le passage, egalement restitue par un sentier d'acces borde de tombes. A !'entree E. de !'edifice, Charles Bonnet suggere la presence d'un autel. En effet, un espace de 0,7 m X 0,5 m est depourvu de toute inhumation, !'ensemble de 1'espace etant densement occupe. Apres 1'etablissement de cet edifice en bois, on assiste a un changement d'orientation radical des tombes. Cellesci vont desormais respecter 1'orientation de 1'eglise en bois. Toutefois il n' est pas exclu que certaines tombes continuent pendant un temps indefini de suivre celle du sarcophage. La duree du second edifice semble couvrir egalement les IXeme et Xeme s. De nombreuses traces de reamenagement ont ete reconnues. La nef en bois rectangulaire a fonctionne un moment avec un choeur carre en pierre, avant de passer a un nouveau plan parfaitement oriente, et a un 148 batiment sur fondation calcaire. C'est a nouveau les positions des sepultures qui suggerent 1, emplacement des chemins d 'acces. Cet edifice est date par comparaison autour de la fin Xeme s.- An Mil (meme si ce plan est connu plus tot). Il est parfaitement oriente E-0 et va entra!ner a son tour une reorientation des tombes. Ainsi apres fixation de 1'habitat, un lieu de culte s'implante sur la necropole fortifiant ainsi la christianisation de la population tout en repondant aux besoins spirituels immediats de la communaute. Et plus concretement encore, elle contribue a !'organisation de 1' espace construit. A Saleux, 1' eglise ne constitue pas dans un premier temps, 1' epicentre. El le est a la fois situee a proximite des champs et d'une ou plusieurs maisons, de jardins ou cultures proches de 1'habitat. Mais !'evolution de !'habitat, tel qu'il appara!t grace aux fouilles, placera progressivement 1, eglise au centre. Les periodes qui nous concement sont difficiles a definir notamment dans un espace micro-regional aux nombreuses facettes. Les facteurs marquants de la fin du VIIeme et le VIIIeme s., c'est-a-dire la disparition du depot funeraire et 1'abandon de nombreuses necropoles, sont associes, semble-t-il, a une nouvelle rupture dans 1'occupation du sol. Les problemes rencontres dans le domaine de la datation du mobilier ceramique ne sont pas etrangers aux difficultes de comprehension de !'organisation des habitats. Tandis que de nombreux sites d'habitat sont abandonnes a ce moment precis, de nouveaux apparaissent a proximite de territoires deja fortement humanises. On assiste ensuite a une nouvelle rupture aux alentours de la fin du Xeme s. ou du Xleme s. apres avoir observe une croissance du nombre des unites d'exploitations, puis un passage des habitants de ces habitats disparus, dans des villages ou hameaux. Entre ces Vlleme et Xleme s., on cherchera a mieux comprendre, en depit d'une problematique chronologique tres presente, la dynamique d'un village et surtout sa naissance, et ce, a travers 1'habitat, la necropole et 1, eglise. Autant de parametre dont 1'etude est plus que jamais d'actualite. Nous esperons des lors, que le site de Saleux, et sa richesse d'informations, pourront apporter ql,.lelques elements de reflexion a cette longue recherche. En effet, si au niveau regional, le site de Saleux constitue un inedit, il re! eve une dimension egalement nationale et europeenne, du fait de la grande rarete d'etablissement de ce type. lsabelle Catteddu lng. en archeologie a I'AFAN, responsable d'operation 69 rue du General de Gaulle, 59!33 Phalempin France Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Rene Proos V enray - 't Brukske, an early medieval settlement on the sandy soils of Limburg Late in 1993, during investigations in the wake of new major roadworks east ofVenray, a large number of traces of medieval settlement were discovered. Most of these consisted of buildings, wells and pits, dating from between the end of the 5th to the beginning of the 8th century, and the period from between the middle of the 1Oth to the beginning of the 12th century. The present paper will focus mainly on the early medieval settlement-traces. Up till now, we know very little of the settlements from the period between the end of the Roman occupation to the beginning of the later Middle Ages. Excavations of this kind of find-spot were quite rare in the Meusevalley: most of what we know stems from three settlements excavated in the 1970s and 1980s, namely Neerharen-Rekem 1, VoerendaaF and Gennep 3 • Of these sites, two (Neerharen-Rekem and Gennep) did not survive the 5th century, the third (Voerendaal) does seem to continue into the 8th century but has not yet been published in any detail: On the basis of the scanty historical evidence available, it seems probable that around the middle of the 5th century the emerging kingdom of the Franks extended its territory - which originally stretched from Salland, Twente and Westfalia down into the Betuwe- into the area formerly known as Toxandria. The total extent of this region in late-Roman times is not well-known; what seems to be certain is that large parts of the present Dutch provinces of North-Brabant and Limburg were part of it. On a recent map compiled by Brulet, charting all the known late-Roman fortresses and defensive works, it is plain to see that main concern of the Empire was the upkeep of long-distance connections, be it land- or water-based 4 . The important East-West road running from Cologne to Boulogne was heavily fortified, and the Meuse~ valley north of Maastricht also counted a number of fortresses. The area to the north of the road and the De Boe et al. 1992, 493-496 Willems 1992,526-533. Heidinga & Offenberg 1992. west of the river-valley appears remarkably empty, devoid of roads and settlements. We will not be far of the mark when assuming this to be the area in which the forebears of early Frankish kings like Childeric and Clovis had their power base. Prior to 1977, however, archaeological proof of the existence of early Franks in this area was virtually lacking. Between that year and 1990, a total of four settlements dating from the 4th and 5th centuries have been unearthed. Neerharen-Rekem and Gennep have already been mentioned and to these we may add Donk near Herkde-Stad in Belgium 5 and Geldrop in the Netherlands 6. And to these four we can now in turn add Venray't Brukske, albeit that only part of this settlement could be excavated, due to the particular circumstances surrounding this major road-project. The find-spot is situated to the east ofVenray, on a site just before the point where the motorway crosses the swampy grounds of the Oostrumsche Beek, a small brook watering into the Meuse. On early 19thcentury maps, one can clearly distinguish a large 'bump' on this spot, presumably a low rectangular hill of about 600 by 300 m. Most of this hill is now gone and only a small part of it is still recognisable east of the A73 motorway. The discoveries were made in a road-trench running north-south, about 300 m long and generally not wider than 40 m. The extreme narrowness of the excavation meant that in most cases only part of the early medieval buildings could be unearthed; in most cases, however, this sufficed to identify and study the houses. A number of structures - mainly wells could be fairly well dated through the use of 14C and dendrochronological techniques. Three chronologically separated entities could be distinguished in the excavation: the first is a midRoman cremation cemetery, partly overlain by the traces of the second entity, a settlement from the Brulet 1993. Van Impe et al., 1992,559-562. Bazelmans & Theuws 1990,33-37. 149 R. Proos Venray - 't Brugske: Plan of the major buildings. Scale 1:500. Venray - 't Brugske - North. . :. 10th-12th centuries. To the north of these and separated from them by what must have been a lowlying and moist gully, lay the early medieval settlement. The settlement belonging to the Roman burial ground has not been found yet, but judging from the period of the cremations (which runs well into the 3rd century), we may assume that the settlement was deserted by about the middle of that century. In the course of the first half of the 5th century, new people arrived on this spot which is well-suited to habitation, located as it is along the edge of a streamvalley. The surrounding countryside offered lots of possibilities for exploiting up arable lands and meadows. The uncultivated land around the edges of the cultivated area supplied the newcomers with wood for buildings and fuel. We have not actually found any buildings from this period yet, but one of the wells already remarkable for containing only wheel-turned late-Roman pottery - was positively dated between 400 and 440 AD by means of the 14C method. Two wells date from the second half of the 5th century; they are located quite close to the Roman burials, having 150 0 ( JY Q. •'. .0 .. 0 8 'llo • I been dug right through each other. They both contained only wheel-turned pottery, presumably imported from the Rhineland. But here as well, buildings dating from the same period are missing. Venray - 't Brukske, an early medieval settlement on the sandy soils of Limburg .;st CO•© ~ ~ a ~ 0 C9 t ' il c~ . ~ \ ~ Q. ~ 0@ ~0 0 0 @ Ei> <11 ~ ·~ 0 - •• 0 \) --- --- --Venray- 't Brugske: house 1. This image of the settlement slowly changed in the course of the 6th century. Both a small building, probably a granary or a barn, as well as two separately located a-typical pits of a hitherto unknown function, contained pottery datable to the first half of the 6th century. The only grave found so far dates from around the middle of that century. It must have been a woman buried here, because in the grave a large number of beads were found, along with a plain iron buckle, a small knife and a large grey pot. Unfortunately, both the beads and the top of the pot went missing soon after the excavation, so the dating of the grave rests solely on the shape of the buckle and the shape and fabric of the pot. During the second half of the 6th century the picture becomes clearer. A large three-aisled building appeared, orientated along a north-south axis, with entrances set in the middle of the long gables and opposite one another. Associated with the house are a large well and a small pit-dwelling, a so-called Grubenhaus, and presumably also a palisade which became derelict in the next phase. A small building of unknown function was built against the palisade. • D 0 ' • ~~ ~ 0 ~ <Q) • © ~ Venray- 't Brugske: house 3. In the course of the first half of the 7th century the settlement consisted of three two-aisled buildings, again orientated north-south. The northern house is accompanied by two wells, both dated dendrochronologically to 629/630, whereas one of the southern houses has two Grubenhauser, as well as a small building, probably a barn, located in the immediate vicinity. The possibility that the third and smallest twoaisled building is older cannot be ruled out; the few fragments of pottery found in this house were handmade and are difficult to date. The building has been allocated to this phase because of the similarity in orientation with its neighbour. A drastic change appears to have taken place · during the course of the second half of the 7th century. New house with an east-west orientations were built and several small pits appeared here and there, filled with iron slag, burnt bones and burnt pottery. A large well, dated dendrochronologically to 693, also took its place in this phase. The house with the rounded end-gable (House #6) must have burnt down to the ground quite quickly; its sudden demise 151 R. Proos " • i4) ;'f.J •• Cl i~~~~ • ·~ ' • ~ oO • tlfiP ~,·~:.i 0 0 • ,;!f:; \'.~ • €lP :~'P..i tft~~l ~ Ill X • •• '•~ ';~ ,.t e , 0 0 e .. Venray- 't Brugske: house 4. • 0 '.. 0 4l • 0 ·~ .:; ·-···~ . I- l T -ITT allowed us to reconstruct its plan because the postholes concerned contained large amounts of charcoal. Finally, we arrive in the 8th century. Only one house remains in this part of the settlement. It is a building of the wide three-aisled type, the likes of which have come to light more recently in excavations in the south of the Netherlands. The rest of the settlement has probably been relocated in a northwesterly direction by way of the neighbouring Antoniusveld. On this Antoniusveld, quite an extensive excavation-campaign took place in the summer of 1996. A large number of house-plans, this time dating mainly from the late Merovingian period to the 12th century, were discovered here. The Houses The space available does not allow me to deal with every building in detail here. However, a couple of the most noteworthy houses do merit closer inspection. Apart from being one of the oldest houses in the settlement, house #4 also posed a number of difficulties when it came to reconstructing its groundplan. Although the building, orientated on a north-south 152 axis, has probably been excavated in its entirety, the groundplan as it is presented here is by no means certain. We are looking at a three-aisled, so-called multi-purpose building, housing both the living accommodations for people and a stable for cattle, as well as storage-space for the harvest. It must have been about 20 m long and 7 m wide, built by means of a large number of relatively small posts forming the frame of a wattle-and-daub wall. Towards the middle of the house two entrances can be discerned, set immediately opposite one another. The southern half of the building seems to have been used for housing the cattle; at least three sets of postholes seem to indicate the same number of cattleboxes. A few posts located within the building, immediately south of the entrances, presumably indicate a wall which separated the stable from the living quarters of the humans. In the north-eastern gable, a small part of the wattle-and-daub construction that once made up the walls has survived. The main problem with the house lies in its northem part, because a large part of the western wall, presumed to have stood here, has disappeared. But the southern part also poses a problem. It looks as if behind the short gable, a second wall was placed here later on, thus creating a small room of about one by Venray- 't Brukske, an early medieval settlement on the sandy soils of Limburg • ~ • 'lli 0 ~~ 1 :.; • • "' () (J~~ :P ~ 'J~ 9ij l!i1l c CJ ~; :ill "' --- -Venray- 't Brugske: house 6. Venray- 't Brugske: house 5, six meters. Perhaps we should see this room as a sort of storage-room; its small dimensions suggest it cannot have had any housing- or stable-functions. Archaeological parallels for House #4 are hard to find. Construction-wise, there are quite a few similarities with the well-known houses of type Odoom B7, in the north of the Netherlands; especially a house excavated in Dalen (prov. ofDrenthe) 8 shows a lot of similarities with House #4; the main difference, however, is the repeated occurrence of four entrances in houses of the Odoom-B type, whereas House #4 only seems to have two. Buildings of the Odoom-B type are dated by Waterbolk in the 6th/7th centuries; Huijts9 also sticks to this periodisation, although Kooi, backed by the results of recent excavations in Peelo, recently proposed a slightly earlier date for both the Waterbolk 1973. Dalen House C, see Huijts 1992, 137, afb. 139. Huyts 1992, 137. 153 R. Proos X ~~ ~ S> • X ..• ~ 7 I I ~ ~ ~ .;:.l\ ;;: i ~ T Venray- 't Brugske: house 9. Odoorn Band Odoorn C types ofhouses 10 . Anyway, the ceramics found within House #4 point at a constructional phase in the second half of the 6th century, but no later than around 600 AD. House #3 was probably built in the same period, or perhaps a little later. It is a two-aisled building, its main axis also lies north-south, and it measures 14 by 5.5 m. A possible entrance is located in the middle of the eastern long gable, at the point where two postholes are closely set together. A second entrance could have possibly been located in the middle of the northern small gable. However, it is quite possible that a third entrance was located in the western long wall; the traces of it must have been destroyed at the time when House #2 was built. In other words: House #3 did not exist any more by the time House #2 was erected. The ceramics of House #3 show the building was put up late in the 6th century or - even more probable- in the first half of the 7th century. House-plans comparable to House #3 are hard to come by in the Netherlands. North of the Central 154 Riverhine area and along the coast, a totally different building-tradition had been adopted in the 7th century. When pushed, one could see Building 26 from Dommelen 11 as a parallel, but this building is about half a century younger. More easily comparable house-plans, in terms of both size and chronology, come from Germany 12 , France 13 and Belgium 14 . These buildings vary in length between 7 and 14 m, while their width lies between 4 and 7 m. All date from the 7th century. House #2 lies to the east of House #3 and part overlaps it. It was orientated on a east-west axis and about 6 m wide. The reconstructible length now measures 10.5 m, but it was most certainly longer than that. It is a one-aisled building; two possible entrances show themselves in the north- and in the south wall respectively; they are set opposite one another and are located on the spot where the postholes were set together more closely. It seems as if House #2 was supplied with an extension on the northern side. These extensions or 'kubbingen' (to use the Dutch phrase) do show up more often in early medieval buildings. They are usually associated with specific activities or functions, such as for instance housing cattle 15 . Ceramics associated with the house point to a construction in the second half of the 7th century. One-aisled houses occur in large parts of northwestern Europe during the 7th and early 8th centuries16. Mostly, when fire-places are lacking, they are interpreted as barns, but this explanation does not hold very easily in Venray. The only fire-place found here and identified positively as such was located outside any identifiable configuration; a second possible fire-place lies within House #5, all other buildings lack a fire-place. In other words, the absence of a fire-place within a partly excavated building- be it because ofpost-depositional processes or because of ° 1 Kooi 1991-92 and Kooi 1993-94. Theuws 1991. 12 For instance at Speyer (Bemhard 1982); G1adbach (Wagner, Hussong & Mylius 1938; Sage 1969); Lauchheim (Stork 1995) and Burgweinting (Osterhaus 1987). 13 Juvincourt-et-Damary (Bayard 1989). 14 Rogge 1981. 15 See for instance 'maison 8' at Juvincourt-et-Damary (Bayard 1989) or 'Haus 19' ofKirchheim near Miinchen (Christlein 1981, Geisler 1988), both dating from the 7th century. 16 For example in Warendorf (Germany), where one-aisled houses of 14 by 4.5 m and 16 by 5 m were found (Winkelmann 1958); Speyer (Germany), where a one-aisled house of I0 by 6 m has been found (Bemhard 1982); Vreden (Germany) where a oneaisled house of 14 by 5 m stood in the 8th century (Reichmann 1982) and West Stow (GB) where a 7th-century one-aisled house (Hall 2) of 9 by 4.5 m was located (West 1985). 11 Venray - 't Brukske, an early medieval settlement on the sandy soils of Limburg Venray - 't Brugske: out-building B5. Venray - 't Brugske: hay-stack. 0 the limitations of the excavation circumstances cannot be used in this instance as a negative proof against the existence of one-aisled houses. Another argument against a barn-function for House #2 is the occurrence of a separated space of about 3 by 5 m in the eastern part of the building. Barns usually do not have an internal spatial division, while buildings with a combined living- and stable- function usually do. House #5, the next door neighbour of House 2, had a three-aisled ground plan, orientated east-west. It must have been about 11 m wide; it is difficult to reconstruct its original length, but the building must have been at least between 10 and 12.5 m long. An entrance was not discernable and may have been located in the unexcavated eastern end of the house. Judging from the pottery, the house had been built around the beginning of the 8th century. A building fairly well comparable in terms size and buildingtradition was unearthed a couple of years ago in Weert, about 35 km. to the south ofVenrayi7. Other closely related house-plans are known in Wijk-bij-Duurstede/De Geer 18 , Gassel 19 , Dommelen20 and Geldrop 21 • In Germany, this type ofbuilding has been excavated in Breberen 22 and Marl-Sinzig 23 • 17 18 19 20 21 22 23 Roymans 1995, 20-21, fig. 17-18. VanEs 1994,233, afb. 195. Verhoeven & Vreenegoor 1991. Theuws 1991, 360-362, fig. 11. Bazelmans & Theuws 1990,33-37, afb. 14-15. Bohner 1958, 461, Ab b. 18. Trier 1981, 219-220. Closely related types have also been To conclude, when looking at the settlement as a whole, it might well be termed remarkable to see the number and the variety of house-plans within a period of about two centuries. Furthermore, the high percentage of wheel-turned, imported pottery, both in 5th- as well as in 6th- and 7th-century wells and pits 24 stands in marked contrast to the situation north of the great rivers. A readily available explanation for this phenomenon cannot be given at this moment. It may, however, well be wise to keep in mind that we have to deal with influences from all sorts of directions in this northern part of the Meuse-valley during the period in which early states were formed. The location of the region, right next to one major river and in close proximity to another, guaranteed a constant influx of new ideas and goods. We may not be far off the mark when assuming the northern part of the Meuse-valley to have been some sort of cultural turntable, functioning simultaneously both as a buffer and as a serving-hatch. Literature BAZELMANS J. & F. THEUWS (eds.) 1990: Tussen zes gehuchten: de laat-Romeinse en middeleeuwse bewoning van Geldrop- 't Zand, Amsterdam. BERNHARD H. 1982: Die fri.ihmittelalterliche Siedlung Speyer "Vogelgesang", OFFA 39, 217-233. B6HNER K. 1958: Die frankischen Altertilmer des Trierer Landes, Berlin. BRULET R. 1993: Les dispositifs militaires du BasEmpire en Gaule septentrionale, in: Fr. V ALLET & M. KAZANSKI (ed.), L 'armee romaine et les Bar- bares du 3e au 7e siecle. Actes du Colloque inter- excavated in Brenz a.d. Brenz (Baden-Wilrttemberg, Knaut 1990) and Irlbach (Bavaria, Osterhaus 1993). Certainly remarkable is the fact that two of the buildings in question (Breberen and Brenz a.d. Brenz) seem to have been early churches (cf. Bohner 1958; Knaut I990). 24 Taken in its entirety, about 95 % of the pottery found in the settlement turned out to be imported. 155 R. Proos national organise par le Musee des Antiquites Nationales et l'URA 880 du CNRS (SaintGermain-en-Laye, 24-28fevrier 1990), 135-148. DE BOE G., DE BIE M. & VAN lMPE L.: NeerharenRekem. Die komplexe Besiedlungsgeschichte einer vor den Kiesbaggem geretteten Fundstatte, in: Spurensicherung. Archiiologische Denkmalpjlege in der Euregio Maas-Rhein. Ausstellung aus An/ass der Offnung der Grenzen innerhalb der Europiiischen Gemeinschaft zum 1. 1. 1993. Aachen, 14. Juli-13. September 1992, Kunst und Altertum am Rhein 136, Mainz, 477-496. Es W.A. VAN 1994: Wijk bij Duurstede-De Geer, in: W.A. VANEs & W.A.M. HESSING (eds.), Romeinen, Friezen en Franken in het hart van Nederland, Amersfoort, 231-233. HEIDINGA H.A. & OFFENBERG G.A.M. 1992: Op zoek naar de vijfde eeuw. De Franken tussen Rijn en Maas, Amsterdam. HUIJTS C.S.T.J. 1992: De voor-historische boerderijbouw in Drenthe. Reconstructiemodellen van 1300 v66r tot 1300 na.Chr., Arnhem. IMPE L. VAN, HUYGE D., VAN Laere R. et al. 1992: Archaologische Untersuchungen im Demertal und seiner Umgebung, in: Spurensicherung. Archiiologische Denkmalpjlege in der Euregio Maas-Rhein. Ausstellung aus An/ass der Offnung der Grenzen innerhalb der Europiiischen Gemeinschaft zum 1.1.1993. Aachen, 14. Juli-13. September 1992, Kunst und Altertum am Rhein 136, Mainz, 550-572. Koor P.B. 1991-92: Project Peelo: het onderzoek in de j aren 1977, 1978 en 1979 op de es, Palaeohistoria 33-34, 165-285. KOOI P.B. 1993-94: Het project Peelo: het onderzoek in de jaren 1981, 1982, 1986, 1987 en 1988, Palaeohistoria 35-36, 169-306. Koor P.B. 1994: Een opgraving op de Westakkers te Dalen, Nieuwe Drentse Volksalmanak (1994), 39(131)- 53(145). OSTERHAUS U. 1992: Ausgrabungen in Burgweinting, Stadt Regensburg, in: Auf Spurensuche, Regensburg, 12-16. REICHMANN Ch. 1982: Landliche Siedlungen der Eisenzeit und des Mittelalters in Westfalen, OFFA 39, 163-182. RoYMANS N. (ed.) 1995: Opgravingen in de Molenakker te Weert: campagne 1994, Amsterdam. ROGGE M. 1981: Een Merovingische nederzetting te Avelgem-Kerkhove (West-Vlaanderen), in: A. VAN DOORSELAER (ed.), De Merovingische beschaving in de Scheldevallei: handelingen van het internationaal colloquium Kortrijk, 28-30 oktober 1980, 67-102. SAGE W. 1969: Die friinkische Siedlung bei Gladbach, Kreis Neuwied, Dusseldorf. 156 STORK I. 1995: Furst und Bauer, Heide und Christ. 10 Jahre archiiologische Forschungen in Lauchheim/Ostalbkreis, Stuttgart. THEUWS F. 1991: Landed property and manorial organisation in Northern Austrasia: some considerations and a case study, in: ROYMANS N. & F. THEUWS (eds.), Images of the past. Studies on ancient societies in Northwestern Europe, Amsterdam, 299-397. TRIER B. 1981: Neue Ergebnisse der archaologischen Hausforschung in Westfalen, in: A. VAN DOORSELAER (ed.), De Merovingische beschaving in de Scheldevallei: handelingen van het internationaal colloquium Kortrijk, 28-30 o/dober 1980, 211-221. VERHOEVEN A. & VREENEGOOR E. 1991: Middeleeuwse nederzettingen op de zandgronden in Noord-Brabant, Overdrukken ROB 395, Amersfoort. WAGNER K.H., HUSSONG L. & MYLIUS H. 1938: Frankische Siedlung bei Gladbach, Kr. Neuwied, in: Germania 22, 180-190. WATERBOLK H.T. 1973: Odoom im Fruhen Mittelalter: Bericht der Grabung 1966, in: Neue Ausgrabungen und Forschungen in Niedersachsen 8, 25-89. WATERBOLK H.T 1991: Das mittela!terliche Siedlungswesen in Drenthe: Versuch einer Synthese aus archaologischer Sicht, in: H.W. B6HME (Hrsg.), Die Salier. Siedlung und Landesausbau zur Salierzeit, Tl. I, Sigmaringen, 47-108. WEST S. 1985: West Stow: The Anglo-Saxon village, East Anglian Archaeology 24, Norfolk. WILLEMS W.J.H. 1992: Die kaiserzeitliche villa von Voerendaal, in: Spurensicherung. Archiiologische Denkmalpjlege in der Euregio Maas-Rhein. Ausstellung aus An lass der Offnung der Grenzen innerhalb der Europtiischen Gemeinschaft zum 1.1. 1993. Aachen, 14. Juli-13. September 1992, Kunst und Altertum am Rhein 136, Mainz, 526-534. WINKELMANN W. 1958: Die Ausgrabungen in der fruhmittelalterlichen Siedlung bei Warendorf (Westfalen), in: W. Kramer (Hrsg.), Neue Ausgrabungen in Deutschland, 492-517. ZIMMERMANN W.H. 1992: Die fri.ih- bis hochmittelalterliche Wi.istung Dalem, Gem. Langen-Neuenwalde, Kr. Cuxhaven. Archaologische Untersuchungen in einem Dorfdes 7.-14. Jahrhunderts, in: H.W. BOHME (Hrsg.), Die Salier. Siedlung und Landesausbau zur Salierzeit 1, Sigmaringen, 37-46. Rem~ Proos Rijksdienst voor Oudheidkundig Bodemonderzoek Kerkstraat 1 3811 CV Amersfoort The Netherlands Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Dries Tys Landscape and Settlement: the Development of a Medieval Village along the Flemish Coast 1 Taking into account the importance of the excavation of the shrunken village ofWal-Raversyde it was worthwhile to subject the landscape in which the village emerged and developed to a historical examination. More particularly, we paid attention to the exploitation of the physical landscape and the use of the cultural landscape thus created, looking also at the evolution of the physical presence of humankind in that landscape from the High Middle Ages to the Early Modern Times. 1 A brief word about methodology We focused on the settlement itself, using a retrogressive and interdisciplinary reconstruction of the historical landscape in and near Wal-Raversyde. The starting point of this reconstruction was the mid19th-century land registry plan of the Section Raversyde of the municipality ofMiddelkerke. At the same time, this plan provided us with the circumscription of the research area. Starting from there and making good use of the many historical sources (18th-century drawn and painted maps and various 18th- and 17th-century copies of parochiallandbooks from 1628 and 1534) we succeeded in reconstructing in some detail the 17th- and 16th-century organisation of the plots in the research area. On the strength of the data found in the extensive archives of the abbey of SaintPeter's in Gent, which owned by far the larger part of the area, and of the many geological and particularly archaeological data (surface finds on the beach and systematic excavations by the Institute for the Archeological Heritage), we were able to further reconstruct the evolution of the larger part of the landscape going back as far as 992 A.D. I wish to thank my brother M. Tys, for the English translation of this text, S. Moemaut for the drawing of the figures I, 4 and 5 and Prof. Dr. F. Verhaeghe for his help in general. 2 The formation of the landscape, the High Middle Ages At the end of the lOth century, the landscape seems to have been very much as it originated after the socalled post-Roman inundation-phase between the 4th and 8th century A.D. (formerly referred to as the Dunkirk !!-transgression; Baeteman & Denys 1997, 8). During these inundations, the causes of which are yet to be fully explained, a tidal channel which separated a strip of land from the mainland was formed in the peat-moor between the present towns ofNieuwpoort and Ostend (Prof. Baeteman pers. comm.). After the flooded coastal area evolved into salt-marshes, the separate strip of land eventually formed the socalled geographical entity of Testerep, on which the research-area is located (Fig. 2). The salt-marshes on Testerep were used to herd sheep. The wool of these sheep was probably an important factor in the development of the textile industry in the rising Flemish towns of that time and also constituted the basis of the economic power and wealth of the major abbeys. The donation in 992 and 995 of two terrae ad oves 2 or 'sheepland' on Testerep to the Ghent St.-Peter's abbey laid the foundation of the domain of the St.Peter's abbey on Testerep. This domain is the most important part of the landscape examined and can be reconstructed without problems on the map of the research-area (Fig. 1). Until the 11th century, these sheeplands on Testerep were constantly threatened by flooding from the tidal channel on the landside. Eventually this problem was controlled, probably at the latest around 1100 A.D., by the construction of the so-called Kaaidijk ('quay-' or 'embankment'-dike). This Kaaidijk divided the research area into two zones: a now wellprotected oudland ('oldland') zone on Testerep and FA YEN A. 1906, Liber tradition urn Sancti Petri Blandiniensis. Livre des donations faites I 'abbaye de Saint-Pierre de Gand, depuis ses originesjusq 'aux X!ieme siec/e, avec des additionsjusq'en 1273, Gand, 1906, 95-nr. 100 and 96-nr. 102. a !57 D. Tys 0 fOO 200 300 400 .100m NOORDZEE LEFFINGE Fig. 1. - The main parts of the research-area. Above the "Kaaidijk" we find the "oldland", the original Testerep. Underneath the "Kaaidijk" we find the "new land", the reclaimed tidal channel. The largest part of the area is formed by the domain of the Abbey ofSt.-Peter's (Gent), which exists of original land on the "oldland" and recovered land on the "new land". an unprotected wet or flooded one (which would become the nieuwland or 'newland') along the channel. Recent geological findings show that Testerep was not ravaged by floods of any great importance after the early Middle Ages (Pieters 1993, 251 and 255). The investment in a defensive dike screening the 'oldland' suggests that the latter area had an economic value well worth protecting. Besides, these lands no longer consisted of salt-marshes, but now consisted of salty and other meadows. However, in contrast to prof. Verhulst's (1995, 91) theory, we feel that this never put a stop to specialised sheepherding, since the domain ofSt.-Peter's was transformed into a berquaria (a sheepfarm that was let in exchange for a certain tribute), probably at the end of the 11th century and at least until 1133 3 • The open, unprotected zone along the tidal channel and the channel itself evolved into a cultural landscape in the 12th century. The 'newland' (or novae terrae) zone of Testerep was partly formed by the gradual and passive drying out of the wet and flooded lands ofSt.-Peter's in the first quarter of the 12th century4 and partly by the active reclamation of the tidal channel by Count Philip of Alsace between 1165 and 1173, which united Testerep with the mainland again (Verhulst 1995, 53-54). What remained of the old tidal channel was a large ditch called Groat Ge!eed (Fig. 2). The reclamation of the channel meant an important extension of the agricultural area. In the meantime, the research area went through an important evolution during the course of particularly the 12th century. Although the sources leave somewhat to be desired for this period, we can establish that during that phase: 1) specialised sheepherding made way for cattlebreeding and agriculture, 2) the population of the area grew considerably leading to the emergence of the first villages on Testerep, even before the reclamation of the channel was completed, 3) the St.-Peter's domain was leased by several individual agricultural enterprises 5 and 4) cropfarming began to intensify. Why did sheepherding disappear to be replaced by cattle-breeding and farming? Did cattle-produce and industrial crops bring in more on the urban markets? DE HEMPETINE T. & VERHULST A., De oorkonden der Graven van Vlaanderen ljuli 1128-september /191). Regering van Diederik van den Elzas Ouli 1128-januari 1168), Brussel, Idem. See the description of the domain in the Liber lnventarius Omnium Bonorum from 1281 (Rijksarchief Gent, fonds St.Pietersabdij, f050v 0 and 114v0 -ll5r0 ). 1988, 51-54, nr. 25. 158 Landscape and Settlement: the Development of a Medieval Village along the Flemish Coast ... 'W« D Z. ,r.... d" ler .81rup Fig. 2. -Rough reconstruction-map ofTesterep and the tidal-channel which separated Testerep form the mainland. After the reclamation of the tidal-channel remained a large ditch, called the "Groot Geleed". Was it perhaps the growth of the population which led to an increase of the demand for cereals (and thus also for fertiliser)? Or could it have been the arrival of English wool, which may have been of higher quality? These are questions which our sources do not answer and which are part of a complex problem. In any case, the development outlined above, testifies to the favourable conjuncture in the High Middle Ages, which according to prof. Thoen (1988, 10751077), resulted in Flanders from the early commercial orientation and limited impositions on the economic surplus. 3 Settlement in the 13th and 14th century: the origins of Walraversyde It is, however, within the context of this new agricultural structure that Walraversyde originates in the second half of the 13th century. The oldest written testimony of the village dates from 1290, when les poisonniers de Wiltravenszeide are indebted to the city ofYpres 6 . The place-name means the hyde, landing-place or wharf, of a certain Walraf and is comparable to the toponyms Lombardsyde and Koksyde. This 13th-century hyde could be situated along the bank of what remained of a small tidal channel that ran into the sea in the research-area, or as Chocqueel called it une petite baie naturelle (Chocqueel 1950, 88 and E. Cools, pers. comm.). The settlement along WYFFELS C., Analyse des reconnaissances des dettes passees devant les echevins d 'Ypres (1249-1291. Editees se/on le manuscrit de t Guillaume Des Marez. Brussel, 1991, nr. 5204. 7 Rijksarchief Gent, Fonds St.-Pietersabdij, Goederenbeheer Brugse Kwartier, Rek. 806d. Reeks I, the original landing-place should be the one located on what is now the beach. Until the end of the 1970, the remains of this village could be seen on the beach, where they were thoroughly researched by A. Chocqueel, E. Cools and A. Mortier. This hyde possibly originated at the same time as the other hydeplaces along the Flemish coast around the mid-13th century, albeit within the framework of the seignorial policy of Countess Margareth ofConstantinopel who aimed at promoting the Flemish ports. Anyway, it is worth noting that the original village was located on Countal territory, west of the domain of St.-Peter's. According to the findings on the beach, the village moved more to the east during the 14th century (E. Cools, pers. comm.), while a rudimentary landbook of the domain from 1357 proves that in that year part of the village was located inside the borders of the domain 7 . Unfortunately, there is no other historical information about the 13th- and 14th-century village to confront with the archeological finds on the beach. Therefore, as we will see later, our knowledge of the 13th- and 14th-century Walraversyde is very weak in comparison with our knowledge of the 15th- and 16th-century village. 4 Settlement in the 13th and 14th century: the agricultural structure of the domain ofSt.-Peter's We have more data concerning the agricultural inhabitation of the domain of St.-Peter's behind the village. This settlement was densely built and populated. According to the landbook of 1357 some 50 farmsteads and houses were to be found in an area of approximately 194 hectares. Many of them probably had a moat. Like Veurne Ambacht this area had a great density of isolated Einzelhofe in the mid-14th 159 D. Tys century 0/erhaeghe 1981, 109-111). Besides, most of the moated sites in Veurne Ambacht had originated between the end of the 12th century and the mid-14th century (ibidem, 106). This might be an indication for the development ofmoated farmsteads and houses in the domain. At the same time, the agricultural lands were cut up rather strongly amongst many, mainly smaller, leaseholders (see Fig. 6), whilst the majority of the agricultural enterprises in that area were small (between 1.5 and 4 hectares) and submarginal (less than 1.5 hectares) ventures. This evolution was probably influenced in a positive way by the favourable economical conjuncture of the High Middle Ages, with high incomes and low taxes and other liabilities. This situation favoured particularly the smaller, more intensive and commercially oriented agricultural enterprises, that had strongly grown in number before 1357 (Verhulst, 1990,67-74 and 85). The excessive fragmenting of the holdings and the increasing liabilities in the 14th century may have caused such problems that pauperization and 'proletarisation' took place within the group of small enterprises, which resulted in a growing number of submarginal holdings (ibidem, 84-85 and 115). From that time onwards, the sources do not refer any longer to these small holdings as farmsteads. But the term used is plaetsen daer zy up woenen 8 or 'places where they live', in other words cottages. Also the appearance at that time of the parochial institution called the 'table of the poor' (in both the parishes of SintMariakerke and Middelkerke) is an indication of a certain pauperization of the agricultural population. Possible solutions for the subsistence-problems of the inhabitants of the submarginal and small ventures consisted of performing wage work on the larger holdings, the exploitation of the clay and peat in the underground, or maybe even fishing at sea. It is not impossible that the development of Walraversyde was influenced by the 14th-century subsistence problems in the domain and its surroundings. 5 The crisis of the late 14th and early 15th century and its consequences on landscape and settlement The situation changed completely as a result of several closely connected events at the end of the 14th century. The village as well as the domain suffered heavily from a severe crisis between 1360 and 1420. In this period the conjuncture in Flanders was Idem. Idem, Rek. 837/2 and I. 314, F 5 r0 -7 r0 • 160 not as favourable as before, which was generally caused by falling incomes and rising impositions (Mertens 1967,51-52 and Thoen 1988, 1030). More specifically, the crisis hit the area a first time with the outbreak of the war between the town of Gent and the Count of Flanders, between 1379 and 1385, which had major consequences for the Flemish coast. The sources tell us that the fields and the farmsteads at the domain of St.-Peter's were temporarily abandoned during the years of war 9. Together with the burdens and destructions caused by war, this meant a great loss of income for the different holdings in the area. The war and the intensive use of the dunes in the decades before the war had weakened the strength of the dunebelt so much that the village and the domain suffered from severe sand-drifts. The weakening and drifting of the dunes made Walraversyde and the domain very vulnerable to storms and floods. Therefore it is not surprising that the flood of 1393 became a large disaster for the whole of the central part of the Flemish coastal area. The domain of St.-Peter's alone lost no less than 35 ha to the inundations and sanddrifts resulting from the flood of 1393. The dunes even drifted all over the village and into the domain, where they formed a new but very weak dunebelt. The village had no other option than to be- probably in its totality- rebuilt behind the new dunebelt, on lands which belonged to the domain. The remainders of this relocated and newly built Walraversyde are nowadays the object of the systematic excavations of this site, directed by Marnix Pieters of the Institute for the Archeological Heritage. What remained of the old, drifting dunebelt, was washed away by the sea, so that the ruins of 13th- and 14th-century Walraversyde came to lie on the (new) beach before the dunes, where they remained visible until 1980. This severe crisis had important consequences for both the village and the domain. During the first years of the 15th century, the rentholders of the domain were not able to pay their rents because of their poverty ('mits den aermoede van den volke') 10 . The structural character of the consequences of the crisis is shown in the landbook of the domain from the year 1463 11 • This landbook shows how the agricultural structure of the domain had been changed completely in comparison to the situation in 13 57. Only 6 out of the 50 14th-century farmsteads and houses still existed in 1463, together with several moats and ruins of vanished farmsteads and cottages. Thus, an important phenomenon of Hofivustungen had occurred in the domain in the years between 1357 and 10 11 Idem, Rek. 832a, F 8 V0 • Idem, Rek. 841 p. Landscape and Settlement: the Development of a Medieval Village along the Flemish Coast 0 20 40 60 80 Duinen lOOm Duinen Zeedijk Zeedijk ;;;: --r.s<6!7 r::l- r.:, - t-;.,- 61- I •.. 'CI.EJ i.!!.J I~~; 60 I :~~ I 1 ~ ~J I I I I 97 I 1 98 I I OD Fig. 3.- Reconstruction-map of the village of Walraversyde according to the landbook of 1534. The squares show us how many houses stood on esch plot in the village, while the dotted squares symbolise the houses which had disappeared by the year 1534. The mill was situated on plot M and the brewery of the willage on plot BR. The rectangular symbols stand for farmsteads and the Roman numbers indicate the different streets, paths and rivelets in the village. On the northside, the village is edged by the sea-dike. 1463. The fragmenting of the domain of a century earlier had equally disappeared and the land was now concentrated in the hands of four larger landowners and their agricultural enterprises (more than 10 ha.) (see also Fig. 6). The smaller and the submarginal enterprises disappeared almost completely from the area. An explanation for this structural change can be found in the crisis we mentioned earlier. It is possible that high impositions, high costs (cf. the damage of the war and the storm), important losses because of the temporary leaving of the domain during the crisis, a high mortality and a low reproduction forced most of the farmers into such a poverty that particularly the peasants of the small and the submarginal ventures were forced to sell and abandon their holdings and/or that most small farms and cottages fell into decay. In the first decades of the 15th century, the abandoned and ruined holdings may have led to such a supply of land that its value dropped; this in turn favoured the concentration of the abandoned lands into the hands of four larger ventures. The question which arises is what happened to the inhabitants of the abandoned farms and cottages. Some probably perished, others might have joined the ur- ban proletariat of, for instance, Bruges, and maybe a few settled in the new village ofWalraversyde, working as fishermen or earning a living as wageworkers on one of the larger farms. In the 15th century, Walraversyde probably had a mixed economy which included not only seamen, but also craftsmen and wageworkers. There was probably a close relation between the large and probably concentrated 15th-century village and the open agricultural area with only a few isolated farmsteads (most of which still exist today) behind the village. At the top of the social structure in the village probably stood the masters of the fishing-fleet, who probably not only owned their ship but also worked as traders and even pirates. These 'skippers' had an important part in the foundation and enlargement of the chapel of St.-John, which was built in the new village around 1435, and also in its possessions (Vlietinck 1889, 14-15). In the 15th century, the fishing fleet of Walraversyde became one of the five major sea-fishing fleets in Flanders (the others were those of Dunkirk, Newport, Ostend and Heist). Apart from the historical data, the recent excavations in this 15th-century Walraversyde also offer us vital information 161 D. Tys about the village. The fruitful confrontation of the archeological and historical data makes it possible to reconstruct successfully not only the topographical situation and spatial pattern of the village but also the social-economical conditions of its inhabitants (e.g. Pieters, Ervynck, Van Neer & Verhaeghe 1995). 6 Difficulties between c. 1470 and 1570: the shrinking of Walraversyde At the end of the 15th century, the area had to deal with the violence of war once again. The war between the regent Maximilian of Austria and a coalition between Bruges and Gent (between 1483 and 1493) took part for a great deal between Nieuwpoort and Ostend. The domain and the village lay right between the fighting parties and were probably deserted for several years. Once again, the consequences of this desertion were important, though not as important as in the previous century. The historical sources tell us that many houses had been abandoned and were 'delapidated and ruined' (syn vervallen ende te rui'ne) as a direct result of the conflict 12 . Another source, the landbook of the parish ofMiddelkerke of 1534, describes in detail how many houses stood on each plot in Walraversyde at that time 13 (see Fig. 3). It also describes more particularly which part of the village had been abandoned in the previous years. It is striking that in 1534, in comparison with the location of the village in the 15th century, almost the complete north-eastern part of the village had disappeared (see Fig. 3). This is also the part of the village which has been the subject of the recent systematic excavations by the Institute for the Archaeological Heritage (see Pieters 1993, 1994 and 1995). This archeological research proved - independently from the historical evidence - that the houses in the north-western part of the village were probably abandoned before the end of the 15th century (idem 1994, 295). Thus the historical and archeological data do not contradict one another and their confrontation confirmed the hypothesis that the whole of the northeastern quarter of Walraversyde was deserted and ruined before the first years of the 16th century because of the war against Maximilian of Austria. It seems that this part of the village was abandoned rather systematically, but this problem needs further investigation. Anyway, the result of the events of the . late-15th century.was that Walraversyde became a 12 Idem, Rek. 840h, p. 415. Private collection Van Der Heyden (Leffinge), ommeloper van de parochies Middelkerke en Mariakerke, 1666, copy of an 13 162 shrunken village, and, as we said earlier this process was already engaged by the beginning of the 16th century. Although the war and the temporary abandonment of the area probably led to a great loss of income for the domain, it seems that the larger farms had overcome this difficult period without too many problems. On the contrary, in the 16th century the property of this larger holdings had increased even more in comparison with the situation in the 15th century (see Fig. 6). This means that the agricultural structure which had come into existence in the previous century had been consolidated, notwithstanding the emergence in the domain of a few smaller farms. It is very striking that the lands of this larger farms were equally concentrated topographically around the farmsteads themselves. It is possible that we are dealing with an early example of re-parcelling. The shrunken village ofWalraversyde did not recover in the 16th century. Quite to the contrary, the village had again to deal with limited sand-drifts, the village-brewery closed its doors and the fishing-fleet suffered from the increased danger at sea. Indeed, during the 16th century the Channel and the North Sea were terrorised by different belligerent parties and the Flemish fishermen were obliged to arrange for armed convoy-ships to protect their fleet (Coornaert 1970, 140-145). 7 The Eighty Years' War and its consequences: the end of the village of Walraversyde The lack of safety at sea was only a prelude to the disaster which followed with the Eighty Years' War, which started in 1567. Throughout the coastal area, the war situation lasted for 40 years. It started with a blockade of the Flemish ports by the Watergeuzen. As a reaction, mercenary troops were sent to the coast, where they started to plunder and loot themselves (Geldhof 1982, 55-56). During the following years, the war hit the coastal plain very hard. Walraversyde and the domain were almost completely deserted between 1571 and 1581 (see Fig. 5). 1581 was the year in which the dikes that surrounded the Calvinistic bastion ofOstehd were breached in order to defend the town. From then onwards and for several years, the domain was flooded by the sea 14 , which made the area uninhabitable. During the following years, war raged on and ravaged the vicinity original from 1534. 14 Rijksarchief Gent, Fonds St.-Pietersabdij, Reeks Goederenbeheer Brugse kwartier, rek. 845a, f" 37 V 0 • I, Landscape and Settlement: the Development of a Medieval Village along the Flemish Coast 0 20 40 60 80 lOOm Duinen Fig. 4.- Reconstruction-map of the remains of the village ofRaversyde after the Spanish War, according to the landbook of 1628. In comparison with the situation in 1534, the largest part of the village had disappeared. Only a small number of village-houses around the chapel remained. In the north-west of the village, the dunes had craned the sea-dike and bured several houses. Along these dunes a new path was drawn, called the "slaghe van den dune". ofOstend completely. In 1598, the sources tell us that the domain lay waste, because of the 'destruction of these parts through troubles and war' (verwoestinghe van selver kwartiere deur de troubelen ende oorloghe15). While in the last decade of the 16th century the larger part of Flanders and its coast started to recover, Ostend and its surroundings remained a battlefield and this until 1604. Only in that year, Spinola 's troops succeeded in capturing the last stronghold of the Calvinist troops in the Southern Netherlands. For the first time in 40 years (some) peace returned to the region. During the siege of Ostend, the abandoned village ofRaversyde, as it was called from the beginning of the 17th century onwards, served as military camp for Spinola's cavalry (Vlietinck 1889, 39-40). The first year in which the St.-Peter's abbey started to collect again the rents in the domain behind Raversyde, and thus also the first year in which the domain was taken into use again, was 1610 16 • In 1628, a new landbook of the parish of Middelkerke shows the effects of the war on the domain and the village, as well as the first attempts of recovery in the area 17 • The landbook describes how almost all the houses in the village from before the conflict had disappeared (Fig. 4). Most of the plots on which the village of Raversyde once stood, were now fields and meadows. The plots just behind the old seadike and the seadike itself had been covered by the sand of the drifting dunes. The present Duinenweg ('Dunes' road') still follows the edge ofthis late 16th and early 17th-century sand-drifts (Fig. 4). The village ofRaversyde had not, however, disappeared completely. Around the mill and the ruined chapel a few houses remained. In contrast to some of the other villages which had suffered from the war, the shrunken village of Raversyde did not recover. A lot of research has yet to be done on this problem, but with some caution we would venture to bring forward some interesting explanatory elements. According to us, one of the most important explanations is to be found in the enormous economic damage caused by the war along the coast in general and in Raversyde 15 17 16 Idem, F 67 v0 • Idem, rek. 845c, F 27 V 0 • Rijksarchief Brugge, Ommelopers verzameling Mestdagh, nr. 923, 1628. 163 D. Tys % 100% 100 80 60 40 20 0 1576 1577 1578 1579 1580 1581 Fig. 5.- The evolution of the un-paid rents (in%) between 1576-1581, including the part ofthe domain that had been deserted and abandoned in these difficult years. in particular. Because of the war, Raversyde lost its most important economical activity, namely its fishing at sea. Other ports had taken over the part of Raversyde, while none of the by now more than 30 years older fishermen from before the war returned to the village after the hostilities had ended. With the disappearance of sea-fishery, one of the most important factors which could have had a positive influence on a possible recovery of the village had been lost. Another factor which could have had a positive influence was the chapel of St.-John. Several attempts of the few remaining villagers to renovate the ruined chapel failed. The church governors of Middelkerke played a dubious part in this matter. During the war, they had been given the care over the 50 ha owned by the chapel (English 1960, 230). The economic value of these lands was great, certainly in the years of recovery after the war. So, when the remaining villagers ofRaversyde demanded that the lands of their chapel be returned to their community, in order to allow them to start rebuilding the chapel, Middelkerke refused and even tried to hinder any attempt at a renovation of the chapel and thus also the recovery of the community of the village (idem, 231-233). Middelkerke succeeded in its attempts and kept the lands. The chapel was never rebuilt. Raversyde remained an agrarian hamlet in close relationship with the old domain behind the shrunken village. 8 New impulses to a memory In 1735, the old chapel, used as a shed, collapsed during a winter storm. Only the tower remained, until it was pulled down in 1860. By this time, Raversyde had become almost legendary, nothing more than a memory. This changed at the end of the 164 19th century, when Raversyde and its surroundings finally got new impulses from rising coastal tourism. At the same time, some local historians and antiquarians started to become interested in the history and the remains on the beach of the old fishermen's village (see Vlietinck 1889). In the 20th century, this led to further archeological research on the remains of the village on the beach. Particularly the work of A. Chocqueel, who gathered a lot of surface finds between 1930 and 1950 (see Chocqueel 1950) and of Mr. and Mrs. Cools, who brought the site to the attention of universities and media, was very important in this respect. Although around 1980 the remains on the beach disappeared under sand-depositions as a result of the building of new breakwaters, the site of Raversyde remained within the sphere of scientific interest; eventually, this finally resulted in the abovementioned systematic excavations of the remains of the village behind the dunes since 1991 (Pieters 1997). At the end of the 20th century, these remains of the landscape and settlement ofRaversyde will be preserved in a touristic and educational park, ready to confront the 21st century. 9 Conclusions, problems and questions The historical study of the landscape and settlement of Raversyde and its surroundings was fruitful and led to several interesting problems. We succeeded in a partial but not unimportant reconstruction and interpretation of the social and particularly economic history of the inhabitants and the users of the landscape and settlement studied. Our work reveals the existence of an interaction between landscape and settlement and the social and economic situation and evolution of its users. We saw how the feudal owners of the salt-marshes started to manipulate the possibilities of the natural landscape and tried to manage the exploitation thereof as they saw fit. The consequence of the construction of a defensive dike was that the area evolved into a cultural landscape with the potential to change from sheepherding to other agricultural activities. We also saw how the feudal owners abandoned direct exploitation of the domain, leaving the area to a number of more individual users, who availed themselves of the opportunity to start new agricultural enterprises. The potential of the landscape as well as the important socio-economic evolution of the second half of the 12th century influenced the emergence of a new agrarian structure with new villages and other forms of rural settlement. As said earlier, we do not know the exact causes of this important but complex development and many problems remain to be solved. Landscape and Settlement: the Development of a Medieval Village along the Flemish Coast 1463-1464 1357 6,5% 1,1% 14,6% 32,5% ,5% 46, 11,6% 1,3% ea. 1534 1,1% 1628 9,8% 50,8% Fig. 6. - The evolution of the tribute-ratio in the domain ofSt.Peter's in 1357, 1463, 1534 and 1628. 1 I'll 2 11 3 !?a 4 0 llilll 5 6 holding less than 1 gemet (0.44 ha). holding between 1 and 5 gemeten (between 0.44 and 2.2 ha). holding between 5 and 10 gemeten (between 2.2 and 4.4 ha). holding between 10 and 20 gemeten (between 4.4 and 8.8 ha). holding more than 20 gemeten (more than 8.8 ha). village of Raversyde. We saw how within the context of this new agrarian structure, the number of mainly smaller farms and holdings grew significantly. Probably already by the end of the 13th century, several more 'proletarian' holdings emerged, the own lands of which did not, however, raise enough surplus to allow them to survive. We can ask ourselves what caused this late 13th- and 14th-century 'proletarisation'. Again, not enough is known about the causes of this development and these need to be studied more closely from both a historical and an archaeological point of view. The same holds true for the submarginal farms and holdings themselves. At any rate, the submarginal and small farms in the coastal plain seem to have been structurally vulnerable. The first severe crisis in the coastal plain after the phase of prosperity of the High Middle Ages resulted in the complete ruin of the submarginal and smaller farms and holdings. After this crisis only the larger farms, which controlled the larger part of the area, survived. Thus, a new agricultural structure, in which the agricultural areas available had been concentrated in the hands of a few isolated larger farmers, emerged in the beginning of the 15th century. In the following decades and centuries this concentration-movement, which started only much later in the other parts of Flanders (Verhulst 1990, 137-140), went on until after the Spanish War when almost all of the lands were concentrated in the hands of the larger landholders. We can ask ourselves many questions concerning this evolution, questions which are closely related to those concerning the origin and evolution of the 'proletarised' and fragmented agricultural structure before 1400. What were the fundamental causes of these structural changes? Maybe further geological, pedologal, archaeological and historical research related to the topographical situation of the late 13thand 14th-century submarginal and small holdings (see 165 D. Tys Verhaeghe 1984, 154) and an investigation of the particular situation of the impositions in the coastal plain may contribute to a better understanding of the causes of 'proletarisation', progressing concentration, and so on. In this context, maybe the possible role and the social and economic background of the (urban/patrician?) landowners in the research area can be examined. This debate should not let us forget the close relationship between the landscape and settlement in the domain and in the village ofWalraversyde. We saw how a small channel -a remnant of the earlier naturallandscape- probably was at the origin of the fishermen's village of Walraversyde. As a result of a complex set of conditions including warfare and some natural phenomena, this 13th- and 14th-century village was abandoned and (probably systematically) rebuilt at the end of the 14th century. Maybe some of the inhabitants of this new and presumably larger village formerly lived on the deserted 'proletarian' small and submarginal farms and houses. Therefore, it is not impossible that the structural changes discussed earlier also had an influence on the rebuilt village. We can ask ourselves what the exact relation was between the 13th- and 14th-century village and the fragmented agricultural structure on the one hand as well as between 15th-century Walraversyde and the later, more concentrated agricultural structure on the other. Wars seem to have played an important role in the evolution ofWalraversyde. At the end of the 15th century, the war between Maximilian of Austria and the Flemish towns resulted in a first shrinking of Walraversyde, while the events of the Eighty Year's War hit the village very hard and made Raversyde disappear as a village, leaving a small hamlet that would never recover from these woes. The question is whether or not these wars can be considered as the fundamental causes of the evolution of Wal-Raversyde as we outlined it. In this respect, we can ask ourselves why it seems that only Walraveryde suffered so severely from these wars. Put otherwise, why did the other villages in the area recover after the Eighty Year's War and why did Walraversyde not? Are the causes for the shrinking ofWalraversyde in the 16th and 17th century of a specific type or of a more general and structural nature or even a combination of both? Are these causes to be found on a economic level (agriculture and/or fishery) or do we have to look for them elsewhere? What seems obvious is that only further historical and archeological research can be of help and among the priorities, we may mention: the 15th- and 16th-century village behind (and underneath) the dunes, 166 the chapel, the remaining houses of the 17th-, 18th- and 19thcentury hamlet, the effects of the late-14th-, late-15th- and late16th-century wars on other villages in the area, the evolution of the medieval and post-medieval agricultural structure of the coastal plain on a more than local scale and the moated sites and Hofwiistungen in the area. As we have seen, it is only through an interdisciplinary approach, linking and confronting of the archeological and historical data that we may arrive at a better understanding of the above-mentioned problems and discussions about the landscape and settlement in and around Wal-Raversyde. Bibliography AUGUSTYN B. 1992: Zeespiegelrijzing, transgressie- fasen en stormvloeden in maritiem Vlaanderen tot het einde van de 16de eeuw. Een landschappelijke, ecologische en klimatologische studie in historisch perspectief, Brussel. BAETEMAN C. & DENYS L. 1997: Holocene shoreline and sea-level data from the Belgian coast, in: ERONEN M., FRENZEL B., PIRAZZOLI P. & WEISS M. ( eds), Sea-level changes during holocene times, Paleoklimat Forschung I Paleoclimat research 21, in press. CHOCQUEEL A. 1950: Les civilisations prehistoriques et anciennes de la Flandre Occidentale d'apres l'examin d'objets leur ayant appartenu, Bruxelles. COORNAERT M. 1970: De verdediging van de kust van Noord-Vlaanderen vanaf 1300 - VI, Rand de Poldertorens 12, 4, 138-148. ENGLISH 1960: De kerk van Raversyde, De Biekorf 61,229-234. GELDHOF J. 1982: De politieke en religieuze situatie in het Brugse Vrije, 1578-1584, in: Brugge in de Geuzentijd. Bijdragen tot de Geschiedenis van de Hervorming te Brugge en in het Brugse Vrije tijdens de 16de eeuw. Herdenking Oostvlaamse synode (8 en 9 mei 1582). Brugge, mei 1982, 55-70. GOTTSCHALK M.K.E. 1977: Stormvloeden en rivieroverstromingen in Nederland I, !I & Ill, Assen. MERTENS J. 1970: De laat-middeleeuwse landbouweconomie in enkele gemeenten van het Brugse Vrije, Publikaties van het Belgisch centrum voor landelijke geschiedenis 28, Gent-Leuven. PIETERS M. 1993: Archeologisch onderzoek te Raversyde (stad Oostende, prov. West-Vlaanderen). Interimverslag 1992, Archeologie in Vlaanderen 2-1992,247-264. Landscape and Settlement: the Development of a Medieval Village along the Flemish Coast PIETERS M. 1994: Laat-middeleeuwse landelijke bewoning achter de Gravejansdijk te Raversijde (stad Oostende, prov. West-Vlaanderen). Interimverslag 1993, Archeologie in Vlaanderen 3-1993, 281-298. PIETERS M. 1995: Een 15de-eeuwse sector van het verdwenen vissersdorp te Raversyde (stad Oostende, prov. West-Vlaanderen). Interimverslag 1994, Archeologie in Vlaanderen 4-1994, 219-236. PIETERS M. 1997: Raversyde: a late medieval fishing village along the Flemish coast (Municipality of Ostend, Province of West-Flanders), in: DE BOE G. & VERHAEGHE F. (eds.), Rural settlement in Medieval Europe. Papers ofthe 'Medieval Europe Brugge 1997' Conference Vol. VI, Zellik, 1997. PIETERS M., ERVYNCK A., VAN NEER W. & VERHAEGHE F. 1995: Raversyde: een 15de eeuwse kuil, een lens met p1atvisresten, en de betekenis van de studie van de site en haar bewoners, Archeologie in Vlaanderen 4-1994,253-277. PREVENIER W. & BLOCKMANS W. 1983: De Bourgondische Nederlanden, Antwerpen. THOEN E. 1988: Landbouwekonomie en bevolking in Vlaanderen gedurende de late Middeleeuwen en het begin van de Moderne Tijden. Testregio: de kasselrijen van Oudenaarde en Aalst, Gent (= Publikaties van het Belgisch Centrum voor Landelijke Geschiedenis, 90). THOEN E. 1994: Le demarrage economique de la Flandre au Moyen Age: le role de la campagne et des structures politiques (XIe-XIIIe siecles). Hypotheses et voies de recherches, in: VERHULST A. & MORIMOTO Y. (eds.), Economie rurale et economie urbaine au Moyen Age, Publikaties van het Belgisch Centrum voor Landelijke Geschiedenis 108, Gent-Fukuoka, 165-184. TYS D. 1996: Een historische landschapsstudie van middeleeuws en later (Wal)Raversyde (einde 1Ode tot begin-17de eeuw), Univesiteit Gent, onuitgegeven licentiaatsverhandeling. VERHAEGHE F. 1981: Moated Sites in Flanders, features and significance, in: HOEKSTRA T.J., JANSSEN H.L. & MOERMAN I.W.L. (eds.), Liber castellorum. 40 variaties op het thema kasteel, Zutphen, 98-121. VERHAEGHE F. 1984: The late medieval crisis in the Low Countries: the archaeological viewpoint, in: SEIBT F. & EBERHARD W.E. (eds.), Europa 1400. Die Krise des Spiitmittelalters, Stuttgart, 146-171. VERHULST A. 1959: Historische geografie van de Vlaamse kustvlakte tot omstreeks 1200, Bijdragen voor de Geschiedenis der Nederlanden 14, 1, 1-37. VERHULST A. 1964: De hospites van de abdij Voormezele te Bredene en het superaratum of overhert. Bijdrage tot de bewoningsgeschiedenis van de Vlaamse kustvlakte, in: Hulde-album Archivaris Dr. Jos De Smet, Studia Historica Gandensia 10, Brugge, 331-339. VERHULST A. 1966: Histoire du paysage rural en Flandre de l'epoque romaine au XVI!Ie siecle, Bruxelles. VERHULST A. 1990: Precis d 'Histoire Rurale en Belgique, Bruxelles. VERHULST A. 1995: Lands chap en Landbouw in Middeleeuws Vlaanderen, Brussel. VLIETINCK E. 1889: Walraversijde. Een gewezen visschersdorp op de Vlaamse Kust, Oostende. Dries Tys (F.W.O. Vrije Universiteit Brussel) Krommenelleboog 47 C 9000 Gent Belgium 167 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Marnix Pieters Raversijde: a late medieval fishermen's village along the Flemish coast (Belgium, Province of West-Flanders, Municipality of Ostend) Since the spring of 1992, a team of the Institute of the Archaeological Heritage (lAP) of the Flemish Community working in close collaboration with the Flemish Employment and Vocational Training Service (VDAB) and with the Provincial Government of West-Flanders has been carrying out archaeological excavations at Raversijde. The excavation site is partly located in the Provincial Domain ofRaversijde and yielded remains of a late medieval fishermen's village known in historical sources as 'Walraversijde'. So far some 3/4 of a hectare ofthis medieval settlement has been investigated, revealing the ground-plan of about twenty houses with their associated infrastructure and a variety of mobile artefacts. The contribution by Dries Tys (see elsewhere in the present volume), concerning among other things the historical and environmental setting of the area, is referred to and used as a starting point for the archaeological exploration of the 15th-century habitation zone under excavation behind the so-called Gravejansdijk, a 15th-century dike (Fig. 1). After a brief and mainly geophysical analysis of the environment of the site, the different types of archaeological structures will be described and commented on. In a next paragraph the material culture of the inhabitants is briefly looked at together with a screening of the different categories of organic finds and their interpretative possibilities. Finally, in a last paragraph, a synthesis of the presently available information is made in order to highlight the specific nature of the settlement and to outline future research. Environmental setting The excavations are located in a relatively flat polder area, 3 to 4 m above sea-level. Today, the site is separated from the sea by dune-belt culminating locally at an altitude of 20 m above sea-level and about 100 m wide at its base. A massive, 3 m high and at least 20 m wide man-made clay-accumulation situated just landwards of the actual dunes, constitutes the body of a late medieval dike. The initiative for the lay-out of this dike is said to have come from John the Fearless, Duke of Burgundy (1405-1419). The dike formerly marked very sharply the transition to the low-lying polder area. Since post-medieval times, however, important amounts of wind-blown sand have been deposited on a 50 to 60 m wide strip running parallel to the dunes, thus levelling out the transition from dunes to polder. Close to the dunes, this sandy layer- which is devoid of medieval structures and tapers out away from the dunes- can reach a thickness of about 1 m. At a distance of about 100 m from the dunes, the sandy top-layer is completely assimilated into the actual plough-layer. The upper part of the polder soil is generally heavily clayey. These clays can fairly well be used to produce bricks. Their CaCO -content gives rise to the variable but mostly yelloJ.,ish colours of these typical 'polder-bricks'. Furthermore, as saline peat is used as fuel, spontaneous glazing occurs frequently in the kiln and gives some bricks a partially greenish glazed outlook. More sand-containing deposits are met at a depth of I to 1.5 m. In turn, these deposits cover strongly stratified clayey deposits, constitute the most easily and regularly available source of drinking-water within the habitation zone. Furthermore, a 1 m thick peatlayer is present at a depth of2 to 2,5 m (Fig. 2). This important source of fossil fuel, already partly exploited in the Roman period, has been thoroughly exploited in medieval and post-medieval times. The peatdigging played a very important role in shaping the landscape and explains several characteristics of the present-day topography. Before being inhabited, the investigated zone was used for agricultural activities. A buried plough-layer is encountered systematically underneath the buildings. Close to the dunes, evidence for a phase of peatdigging has also been registered underneath the buildings and traces of peat-digging have equally been observed below the so-called 'Gravejansdijk'. 169 M. Pieters The 15th-century inhabitation So far nineteen buildings with brick wall-foundations have at least in part been investigated. For only half of these, a complete or nearly complete groundplan could be examined. These ground-plans are preserved in the soil mainly as a network of robbery trenches. From the study of the few preserved walls it can be deduced that these foundations have been constructed using recycled materials. The very heterogeneous mixture of all kinds of bricks - some of with preserved remains of plastering- illustrates this very clearly. The bricks not only show variable dimensions (23-28 by 10-14 by 5-7 cm) but some of them also have different morphological characteristics (curved or asymmetrical bricks and bricks with which have been profiled in a specific way). Furthermore, we found that these foundations had been built without the use of lime mortar, although the robbery trenches regularly yielded whitish mortar fragments. The latter suggest that the walls (or at least some parts of the walls) were plastered at a higher level. In addition to providing a better isolation, a whitish plastering of the walls also enhances the brightness within buildings which only had a few small windows or even none at all. A kind of beige clayey sand is used as cement. The foundation ofthe walls is generally very shallow. Sometimes, the picture seems to be one of the walls not having any subsoil foundations at all, but having been built from the ground-level upwards, In one of the buildings, the original living-floor is still present and shows that in fact the base of the corresponding walls is at the most only 20 cm deeper. This means that if we take into account the 15 cm thick layer of clay that has been brought in the building during or after its construction, the walls did not have any real subsoil foundation. Generally, only one to three courses ofbricks have been preserved in situ. In a few cases, however, up to eleven courses can be found, reaching a height of up to 80 cm. The walls are generally 35 to 50 cm thick and consist of a kind of cast work with only the vestments constructed with complete or half bricks. The lowermost layers can reach a thickness of up to 70 cm (2.5 bricks). In the zone which has been investigated most intensively, the buildings are grouped in three different areas separated from one another by ditches which are 3 to 4 m wide (Fig. 2). The infilling in these ditches can be separated into two parts which are different genetically. The upper part results totally from human intervention by man and consists mainly of ashes mixed with all kinds ofhuman refuse. The lower part is strongly clayey and corresponds to a water170 lain deposit. The origins of these ditches probably date back to the phase of the agricultural use of the area. The infilling of the ditches with all kinds of refuse certainly happened at a time when the draining function the ditches was no longer required. Anyhow, their orientation (NW-SE/NE-SW) apparently has determined the orientation of the buildings. The latter run parallel or at square angles to one another. The predominant ground-plan has a rectangular shape. In its basic form, it reflects a 12 m long and 6 m wide volume. Internally, the available space is subdivided into two rooms, a large one occupying two thirds of the total surface and a smaller one occupying the remaining third. The rectangular lay-out varies essentially in terms of its length which ranges from 12 to 17.5 m. Apart from one exception with a width of 8 m, the width of the constructions is relatively stable at 5,5 to 6 m. When looking at the ground-plan of the different buildings (Fig. 2), it is obvious that three of them diverge from the above described basic rectangular model. Building 4 with its circular annex is a singleroom building. The buildings 1 and 13, the two largest ones of the group, are more complex. In both cases, however, the composing basic ground-plan model can be recognised in the overall plan. Both buildings also display other characteristics which single them out. First of all, they have a brick-floored fireplace, which in the case of building 1 most probably supported a stove or oven. The localisation of a heating device close to an internal wall suggests that these walls at least were sufficiently heat-resistant. Secondly, both buildings are equipped with a few supplementary conveniences such as a brick-built well, a vaulted latrine, a kind of a small, and a circular underground cold storage facility(?) or floored spaces, both inside and outside the building. In the area closer to the dunes, outdoors brick pavements which are 1 to 1.5 m wide and adjoin the walls are the rule rather than the exception. This is probably due to the better conservation of the archaeological structures in this area and this allows us to assume that the buildings generally had such outdoors brick pavements adjoining the walls. One of these pavements has semicircular (radius 4.5 to 5 m) lay-out and a considerable surface (about 30 m 2); it consists of a layer of canted bricks and is partially bordered with natural stones. The brick surface shows traces of significant wear and one is tempted to assign to this pavement a special function; but up till now, no hard archaeological evidence confirms this hypothesis. So far, only one building has yielded information concerning the entrance. Its location in the southeastern corner of the building is understandable Raversijde: a late medieval fishermen's village along the Flemish coast (Belgium, West-Flanders, Ostend) lilllilli] 1 ~2 0 lOOm Fig. 1.- Location of the excavated part (darker zone) of the late medieval habitation zone (greyish area) with indication of the church-site. 1. Excavated; 2. Approximate village area. from a climatic point of view, since less than 10 % of the winds in this part of the country blow from this direction. Some buildings have buttresses, external as well as internal ones. At present, no regular pattern can be discerned as to their position. While the internal buttresses may have been intended as direct supports for parts of the roof, the external ones had to counter the lateral pressure exerted by the roof-construction. As fragments of tiles and slates are very scarce among the archaeological finds, the roofs were probably thatched. The observation of lots of charred remains of galingale (Cladium mariscus) among the plant remains could mean that this plant, a member of the family of the Cyperaceae and well-known in the region, has been used as thatching material. Besides the floored fireplaces, circular or oval surfaces burnt in situ are met frequently in the buildings. They are often flanked by pits filled in with ashes or heated sediments, mostly sands. On three occasions dug-in pots have been found. As a rule, these were filled with ashy sediments. In one of them a coin was found at the bottom of the pot as if it had been hidden there deliberately. No evidence for the housing of cattle (such as traces of trampling or high phosphate accumulations) can be detected in the area actually under study. If cattle-raising belonged among the activities of the villagers it surely did not take place in that part of the settlement which has already been excavated. Judging from the regular and very homogeneous nature of the general lay-out, it appears that the area 171 M. Pieters nmrD lfiL_j [ill I 1111111 2 3 f:§3 4 C'lJ Fig. 2.- General simplified ground-plan of the excavated area with indication of the 1: ditches, 2: buildings, 3: wells and 4: peat-exploitation pits. with the ditches was inhabited only for a limited period oftime. The occurrence of only a few and mostly slight modifications of the buildings is a second argument in favour of this hypothesis. On the other hand, at least three phases can be observed in the area closer to the dunes and this may point to a longer period of inhabitation. Besides the buildings with a brick base, at least three wooden buildings are present in archaeological record of Raversijde which has been investigated up to now. The numerous post-holes in the zones concerned do not, however, allow us to identify a clearly defined ground-plan. A combination between both types of construction- wood and brick- is also possible. On one occasion a brick base building with a wooden annex has been documented. On several places, 25 to 70 cm wide eaves-drips occur; these gullies drain the rain-water dripping from the roof to the ditches or to some large manmade depressions in the area. To the south of building 4 a sewer connects three gullies and drains the collected water to the ditch. Another vital condition for the inhabitants is the supply of drinking-water. Apart from three wells constructed with bricks, the water-supply is predominantly ensured by means of barrel-wells. So far, nearly 50 of them have been investigated. In general they are about 2 m deep and reach the sandy aquifer. 172 Only one of them reached a depth of 4 m and perforated even the above-mentioned peat-layer. Below the water-table the barrels- originally herring-barrels- are very well preserved. These constructions which are rather characteristic for a fishermen's village are also very useful as actual archaeological indicators. As these barrels are made from oak, which according to the dendrochronological analysis grew in the Gdansk area (Poland), they are extremely useful as a dating tool. So far about 22 of them have been analysed and provided us with dates ranging from the late 14th to the late 15th century. Remarkably enough, the exceptionally deep well also provided the oldest date (1389-1401). A possible hypothesis could be that this well had been installed as a kind of a test to explore the subsoil when settling the area. These deeper structures also are a valuable source of information on the material culture of the inhabitants, particularly where the more organic and perishable part of it such as wooden and leather objects are concerned. During the period when the wells were in use, lots of small animals (amphibians, small mammals and insects) were trapped in them. Their remains are very helpful in reconstructing the local environmental conditions. When a barrel-well got out of use, apparently no recycling whatsoever was planned for. The remaining inconvenient was simply filled in. Raversijde: a late medieval fishermen's village along the Flemish coast (Belgium, West-Flanders, Ostend) When analysing the spatial distribution of these wells while at the same time taking into account the dendrochronological information, two groups of five wells each deserve special attention. It looks as if in both groups the wells constitute a chronological sequence having replaced each other at given time-intervals. This in its turn could shed some light on the duration of the occupation of the associated buildmgs. The bottom fill of some of these wells clearly suggests that the inhabitants preferred unpolluted water. Frequently pure sand-bodies have deliberately been introduced into the wells in order to cover the base which had already been partially polluted by organic matter (leaves, branches, small animals). As the small diameter of the wells excludes a thorough cleaningas is possible with the brick wells which have a larger diameter (80 to 100 cm) - the introduction of sand could be a means to ameliorate temporarily the quality of the water. The barrels themselves are relatively small: 72 to 75 cm high and with a diameter halfway between top and bottom of 51 to 62 cm. These dimensions represent a volume of something in between 110 and 150 litres. The barrels frequently show marks on the stave with the generally more or less square bunghole. When the original lid of the bung-hole was not preserved, the bung-hole was closed with another stave, with a piece of leather or with the bottom or cover of a barrel before the barrel was sunk into the soil. So far, only in one case a bottom or cover of a barrel was left in its original position and used as well-bottom. The wells lead us to another type of structure: latrines. So far five such structures have been investigated. Two of them are constructed with bricks and are clearly integrated into the architectural concept of the respective buildings (buildings 13 and 14). Three other latrines have a barrel as an underground receptacle. These barrel-latrines are not recycled barrelwells as can be inferred from their shallow depth. Two of them are paired and obviously represent a communal latrine arrangement. From the analysis of the remaining typical filling is clear that the latrines were used exclusively as toilets and not as household refuse dumps as is frequently the case in urban contexts. Besides these shallow barrel-latrines, two shallow circular brick structures with an inner diameter of about 40 cm and an approximate depth of 1 to 1.2 m have been investigated as well. These structures were completely filled in with brick debris after having been abandoned; they are provisionally identified as cold stores. One of them contained a grey ware pitcher standing at the bottom of the structure. When appropriately closed at the upper part, such a structure could indeed serve perfectly well as a cold store. A rectangular brick structure, 1.4 m long and 0.8 m wide, can be interpreted as the lower part of a smoke oven to treat fish. The cement in the joints between the bricks of the floor is coal-black, while the cement in the joints of the walls has the normal beige colour. Recently, a second and nearly identical structure has been unearthed but in this case, the cement in the floor and in the walls did not show the same differences in colour. Roads have not yet been found. Only a pathway metalled with natural stone and close to building 13 has been unearthed. When looking at the general ground-plan two large features have not yet been discussed. One of these has been excavated nearly completely and proved to have a surface of more than 500m2 • When looked at in section, it becomes clear that these features have a complex origin. After the extraction of the peat and part of the clay and the partial filling in of the pit, the remaining depression was recycled as a household refuse dump and as a drainage basin. By doing so, the inhabitants in fact lived around their proper refuse dump. Material culture and organic finds The small finds include a wide range of materials. Among the artefacts, ceramics constitute the bulk of the finds. The ceramics mainly consist of common red and grey wares produced in the region roughly in the 15th century. The red wares largely predominate (with percentages up to 80 %). The following forms occur: cooking pots, skillets, bowls, plates, dishes, pitchers, jugs, storage jars, dripping pans and lids. More exceptional are anthropomorphic whistles, colanders, fire-covers and oil-lamps. Among the grey wares now and then potsherds occur which are covered with a sticky black substance, which can be identified as tar or pitch. These objects probably have something to do with a known treatment of fishingnets with tar or pitch. Brickwares also figures among the ceramics but only in very small quantities. Among the imported wares, two European regions are very well represented: Germany and the Iberian Peninsula, more specifically Andalusia. Stoneware from Siegburg, Langerwehe and Raeren is present on the site. Most of the stoneware products cannot, however, be readily assigned to a specific production centre: they can only be identified in a general way as coming from the Rhine area and/or from the adjoining Aachen-Raeren region. Among the Siegburg products, green glazed examples occur regularly. The Iberian pottery in173 M. Pieters eludes the well-!mown majolica but coarser products from the Mediterranean region are also present among the finds. The actual picture emanating from the ceramics points to people with a moderate standard of material living. Further research on the ceramics as well as crossfitting and spatial analysis of the finds should allow us to detect socio-economic differences - if any between the inhabitants of the excavated part of the village. A first approach of the infilling of the ditches already highlighted the promising heterogeneity of the material. Among the non-ceramic items, objects related to fishing dominate the finds collection: sinkers in lead, limestone and brick, a sounding-lead, barbed iron fish-hooks between 4 and 12 cm in length, smaller copper-alloy fish-hooks, disgorgers copper-alloy, antler marlinespikes, wooden netting-needles and cork floaters. One of the netting-needles is nicely decorated with a very typical drawing of three entangled fishes. Other finds can indirectly be related to fishing activities: the presence of pit-coal and of lots of natural stones. The Belgian coastal area is very poor in natural stones. Independent from historical research, the petrologic study of an assemblage of nearly 100 natural stones excavated in one of the ditches concluded that they came from the northeastem coast of the British Isles somewhere between Northern England and the valley of the Humber. The pit-coal probably came from Newcastle-upon-Tyne, since the fishermen of Walraversijde participated in this late medieval trade. Apart from objects related to fishing and therefore typical for coastal, estuarine or riparian sites, other artefacts such as spindle whorls and weights allow us to identify more widely spread domestic industries such as spinning and weaving. Apart from these, no other crafts can at present be discerned in the archaeological finds. Dress accessories such as buckles, pendants, badges and combs (in wood as well as in ivory) form another important group of small finds. Objects related to leisure (such as dies, marbles, mans and tops) or to religious beliefs (such as pilgrims badges, bracelets and clay figurines) also occur regularly. Among the organic finds - particularly wooden ones - brooms are equally well represented. Some exceptional finds such as a fragment of one of the oculars of a pair of spectacles, a measuring-rod and a visor, are to be related to the categories of people frequenting the village. The pair of spectacles and the measuring-rod are probably related to the presence of merchants while the visor undoubtedly belonged to an armed fisherman. This is not contradicted by the fact that the authorities regularly summoned people from Raversijde to assist in the defence of the coast. From the small finds, it can be concluded that the area under study had already been left at the end of the 15th or in the beginning of the 16th century. The Fig. 3.- Decorated container in so-called 'brick-ware' typical for the Belgian coastal area. Scale 42%. 174 Raversijde: a late medieval fishermen's village along the Flemish coast (Belgium, West-Flanders, Ostend) numismatic evidence is quite clear. Most of the coins date from the period of John the Fearless (14051419) and of his son and successor Philip the Good (1419-1467). The calcareous and humid soil environment results in very good conservation conditions for bone material; sometimes even the cartilaginous parts are conserved. In the zone along the dunes with the best conservation conditions even a few starfish-skeletons have been preserved. The first archaeozoological analyses reveal the consumption of sheep, cattle and pigs. The meat menu is further supplemented with chicken, goose and duck and occasionally with rabbit and red deer. The villagers also consumed lots of shellfish such as mussels, oysters and whelks. For the moment no consumption remains of crab or lobster have been found. It is clear that in the actual context special attention has been paid to the fish remains. Apart from eel and carp, the fish remains nearly exclusively consist of sea-fish. The already identified species are cod, glounder, haddock, herring, plaice, tuna, thomback and whiting and different species of sharks. Typical for the site is the presence of occasional catches as sharks that so far have not been found on inland sites. A second feature is the joint occurrence of small and large specimens joint. Normally, the smaller specimens are absent from pure consumer sites. One of the pits -the first which has been studied in any great detail -contained a lenticular deposit of fish-bones. Thousands of plaice remains, belonging to approximately 130 individuals have been investigated and it was shown that the skeletons are incomplete. Only head and tail elements, together with stomach contents, had been deposited; cut marks behind the head and near the caudal fin occur frequently. The high concentration of these remains and the absence of any admixture of other fauna! remains or archaeological objects indicate that the deposit reflects a single event. It is believed that these fish bones are the remnants of plaice processed for future consumption and export inland. A first study of macro-botanical remains revealed the presence of the following consumption plants: wheat, oats, barley and broad b~an. All these species could be grown in the vicinity of the site. The analysis of the parasite remains of the infilling of one of the barrel-latrines revealed the presence of eggs of Trichuris sp., Ascaris sp. and Trichostrongylus. The presence of the last parasite is most probably related to the presence of sheep in the immediate surroundings. All parasites discovered indicate a precarious - at least when compared to our present standards- hygienic level. . .: ·, ,'• .. .. • .\ :' .. i··:'! ·:: ~; . ·t . r, ,. : ... ] t··: Fig. 4.- Wooden netting-needle decorated with three entangled fishes. Scale 1:1. Finally, an entomological study of the beetle remains from one of the barrel-wells pointed to a grassy environment. Synthesis At the end of the 14th or in the beginning of the 15th century, a formerly agricultural zone of Walraversijde became a building ground. The buried plough-layer and at least part of the ditches are the reminders of this agricultural phase. It is plausible 175 M. Pieters that at that point in time, the settlement which had been heavily damaged by numerous storm-tides, was partly relocated landwards in the safer environment behind the 'Gravejansdijk'. Not only the regular and homogeneous lay-out of the building-ground under study, but also the chronological data provided by the barrel-wells and by the small finds suggest this area to have been inhabited mainly during the 15th century. The abandonment of the area studied is probably related to the uprising of the Franc of Bruges against Maximilian of Austria in the last quarter of the 15th century. This uprising was very severe, particularly in the countryside. It was a period of military troubles combined with pestilence and ensuing socio-economic problems. The buildings investigated can be described as follows. The walls have a brick base with a minimal height of 80 cm. Although we suggest that above their brick base, the walls of these buildings were essentially in wood or wattle and daub, it cannot be excluded that some walls were constructed in brick up to the roof. Whatever the case, the roof was thatched. The whitish mortar remains found regularly in the robbery trenches probably indicate that the walls had a whitish finish. The buildings predominantly have a rectangular ground-plan, which in its basic form consists of a 12 m long and 6 m wide volume internally subdivided into two rooms. The largest of the latter generally contains the fireplace. Some indications such as the absence of internal points of support for the roof and the presence of a few internal buttresses could be interpreted as pointing to a cruck-construction, but we could just as easily imagine a roof which was in essence supported by the walls (internal as well as external ones). Two of the buildings two can be singled out by their larger size and by the presence of several additional conveniences as a brick well and a brick latrine. It is possible that those buildings belonged to people of a slightly better socio-economic status than the average inhabitant. This must, however, still be checked through confronting the architecture with its associated refuse. Evidence for the housing of cattle could not be found. The buildings do not have a surrounding garden or enclosure as seen frequently in medieval villages (e.g. in Wharram Percy). Nevertheless, we may assume that at least some villagers engaged in agricultural/horticultural or stockbreeding activities in the immediate vicinity of the village. The fact that fishing and associated activities played a very important part in the life of the villagers may explain the dense built-up of the area with buildings standing very close to one another. The rather elementary houses ofWalraversijde fit in very 176 well in with people whose activities took place mostly out of doors. The large depression resulting from peat- and clay-digging activities places the lay-out of the area under study in a specific context. It is quite possible that the building activities on this terrain started with the digging of a large pit in order to extract peat and clay to be used for the construction of the buildings which in fact seem to be located round about the remaining depression. The typical barrel-wells are quite an important part of the archaeological story as they provide us with the possibility to follow the chronological evolution of the inhabitation. Apart from their importance in terms of chronology, they also allow us to gain a measure of insight in the environment of the area. The material culture is quite characteristic, first and foremost all because of the presence of numerous objects and finds directly related to fishing. An important amount of other finds are indirectly related to these fishing activities.That fishing was an important part of the activities can also be deduced from the remains of fishes themselves. The presence of natural stones coming from the northeastern coast of the British Isles, of pit-coal, of a considerable amount of imported wares from the Mediterranean (such as ceramics from Andalusia), of cork floaters and of ivory combs also has to be mentioned. Similarly, some finds such as a fragment of a pair of spectacles point to special characteristics of the site. A first detailed analysis of the material culture of the inhabitants allows us to suggest that in socio-economic terms, the inhabitants seem to have a standing which was slightly above the rural average in the region. Much archaeological research still remains to be carried out. First of all and as the excavations are still in progress, the full exploitation of the archaeological data, including the cross-fitting and spatial analysis of the small finds is yet to come. The same holds true for the contributions from the different partner sciences. So far, the partner sciences have in fact led only to first assessments. A first general assessment concerning the zone with the ditches is planned in the course of 1998. Secondly, as the archaeological strategy is in fact regulated by the rhythm of the realisation of the touristic infrastructure, an important part of the village- including the church and the adjoining churchyard, the windmill, the brewery, the main street and different satellite habitations outside the nucleated area- has not yet been touched by excavations. We hope that this important hiatus can be taken care of by the next edition of MEDlEY AL EUROPE. Raversijde: a late medieval fishermen's village along the Flemish coast (Belgium, West-Flanders, Ostend) References Archeologie du village deserte, Cahiers des annates 27, 1970, 2 vol, Paris. BERESFORD M. 1954: The Lost Villages of England, London (reprint of 1987 by Alan Sutton). BERESFORD M. & HURST J. 1990: Wharram Percy. Deserted Medieval Village, London, English Heritage. CHAPELOT J. & FOSSIER R. 1980: Le village et la maison au moyen age, Hachette. CHOQUEEL A. 1950: Les civilisations prehistoriques & anciennes de la Flandre occidenta!e d'apres l 'examen d 'objets leur ayant appartenu, Bruxelles. DE GELAEN E. 1973: De gevolgen van de opstand tegen Maximiliaan van Oostenrijk voor het Vlaamse platteland, Handelingen van de Koninklijke Geschied- en Oudheidkundige Kring van Kortrijk (Nieuwe Reeks) XL, 183-244. DEGRYSE R. 1983: Vlaamse kolenhandel en de Schonense kaakharing te Newcastle upon Tyne (13771391 ), Handelingen van het Genootschap voor Geschiedenis CXX-314, 157-188. DE PAEPE P. & PIETERS M. 1994: Petrology and Provenance ofUnworked Stone from the Medieval Fishing-Village at Raversijde (mun. of Oostende, prov. of West Flanders), Archeologie in Vlaanderen IV, 237-251. FosS !ER R. 1995: Villages et Villageois au Moyen Age, Paris. PIETERS M. 1992: Archeologisch onderzoek te Raversijde (stad Oostende, prov. West-Vlaanderen). Interimverslag 1992, Archeologie in Vlaanderen II, 1992,247-264. PIETERS M. met een bijdrage van DE BUYSER F. 1993: Laat-middeleeuwse landelijke bewoning achter de Gravejansdijk te Raversijde (stad Oostende, prov. West-Vlaanderen). Interimverslag 1993, Archeologie in Vlaanderen III, 1993,281-298. PIETERS M. met bijdragen van F. BOUCHET, A. ERVYNCK & M. VAN STRYDONCK 1994: Een 15de-eeuwse sector van het verdwenen vissersdorp te Raversijde (stad Oostende, prov. WestVlaanderen). Interimverslag 1994, Archeologie in Vlaanderen IV, 219-236. PIETERS M., ERVYNCK A., W. VAN NEER & VERHAEGHE F. 1994: Raversijde: een 15de-eeuwse kuil, een lens met platvisresten, en de betekenis voor de studie van de site en haar bewoners, Archeologie in Vlaanderen IV, 1994, 253-277. PONCELET L. & MARTIN H. 1947: Hoofdtrekken van het Belgisch Klimaat, Verhandelingen Koninklijk Meteorologisch Instituut van Belgie XXVI. TYS D. 1996: Een historische landschapsstudie van Middeleeuws en Later (Wal)Raversyde (eindeJOde tot begin-17de eeuw), onuitgegeven licentiaatsverhandeling Universiteit Gent. TYS D., this volume: Landscape and settlement. The development of a medieval village along the Flemish coast. VAN DOORSELAER A. & VERHAEGHE F. 1974: Excavations at the XIVth century village of Roeselare (Sint Margriete) (East Flanders, Belgium), Dissertationes Archaeologicae Gandenses XV, Brugge. V AN NEER W. & ERVYNCK A. 1993: Archeologie en Vis, Herlevend Verleden 1, Zellik. VAN NEER W. & PIETERS M. in press: Evidence for processing of flatfish at Raversij de, a late medieval coastal site in Belgium, in: Proceedings of the 71h Meeting of the International Council of Archaeozoology. V AN UYTVEN R. 1980: Crisis als cesuur 1482-1494, in: Algemene Geschiedenis der Nederlanden 5, Haarlem, 420-435. ZADORA-RIO E. 1995: Le village des historiens et le village des archeologues, in: MORNET E. (ed.), Campagnes medievales: l'homme et son espace. Etudes offertes aRobert Fossier, Paris. Mamix Pieters Instituut voor het Archeologisch Patrimonium Doomveld 1, bus 30 B-1731 Asse-Zellik Belgium 177 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference - Volume 6 Christopher Loveluck Flixborough - the character and economy of a high status Middle Saxon settlement in northern England Introduction Between 1989 and 1991, excavations at Flixborough on the south bank of the Humber estuary, in northern England (Fig. 1), revealed remains of a Middle Saxon settlement which have proved exceptional to date. The finds have included the foundations of over twenty buildings and other structural evidence, together with an extremely rich artefact assemblage and a vast collection of vertebrate remains. Although the full extent of the settlement was not uncovered, the quality and range of evidence from the site has provided a hitherto unprecedented opportunity toestablish archaeological criteria for defining the nature and character of high status Middle Saxon settlements, both within the area of the Humber estuary N t 0 Kms 20 Wharram Percy • Garton on the Welds • • Driffleld • Beverley NORTH SEA Barrow • Burnhom GH Rlby e Hollon le Fig. 1. -Location of Flixborough and other Anglo-Saxon sites situated in the hinterland ofthe Humber estuary, discussed in the text. 179 Chr. Loveluck and more widely in England, and to illustrate nearly all the facets of the economy of such a settlement, especially animal husbandry, craftworking and industrial activities, and trade and exchange links. The Anglo-Saxon remains at Flixborough were first indicated as a result of an archaeological evaluation in advance of sand quarrying in 1988. During this work, Mr Kevin Leahy (Keeper of Archaeology, Scunthorpe Museum, North Lincolnshire) uncovered eleven unfurnished inhumation burials. Following inconclusive geophysical survey results, further evaluation was undertaken as a control exercise by the Humberside Archaeology Unit (now the Humber Archaeology Partnership). This resulted in the discovery of substantial, well-preserved Anglo-Saxon settlement remains. As a consequence of these finds, English Heritage funded the excavation of a sample of the probable Anglo-Saxon settlement area, with a view to defining the character, date and importance of the site. The settlement evidence was predominantly Middle Saxon, dating between the 7th and early 1Oth centuries AD. There are, however, additional indications that the area in the immediate vicinity of the excavations constituted a continuous settlement focus from the Romano-British period to the High Middle Ages. The following account provides a summary of the preliminary analysis of the Flixborough remains, undertaken to assess their importance. The detailed post-excavation analysis and publication programme is being canied out over the next five years (Loveluck 1996). Settlement morphology and occupation sequence The heart of the Middle Saxon occupation at Flixborough was situated on a windblown sand spur, overlooking the floodplain of the River Trent, 8 kilometres to the south of the Humber estuary. The foundations of a series of at least twenty buildings were uncovered, many of which were constructed on longlived building plots, superimposed over the foundations of their predecessors (Fig. 2). Periodically, however, the use of the excavated area for habitation was interspersed with major phases of refuse dumping. This cyclical exploitation of the spur for housing and dumping has provided unprecedented stratified deposits from a Middle Saxon rural settlement. The refuse and demolition spreads, in particular, yielded exceptional artefact and animal bone remains, both in regard to their quality and quantity. In addition to the buildings and refuse dumps, other components within the settlement structure include a boundary ditch, running on an east-west alignment, from the western limit of the occupation area; metalled path180 ways, linking different building plots; and activity areas, associated with craftworking, ovens or disposal pits. The extent of the superimposition ofbuildings and other structural features seems to be a reflection of the concerted planning of settlement layout, whether by an individual or collective authority. This spatial organisation was probably influenced by the unstable soil conditions towards the steeper slopes of the spur, thus promoting the re-use of a limited number of suitable building plots. Hints of the size of the area covered by AngloSaxon settlement evidence have also been provided during the excavations and further evaluation and survey work. The partial recovery of the foundations of several buildings, running beneath the eastern edge of the excavated area, together with the discovery of pits and gullies, containing Middle Saxon ceramics, on the gentle southern slope of the spur, demonstrate that the Anglo-Saxon settlement area continued to the east and south of the excavated remains, probably encompassing the cemetery found in 1988 (Fig. 3). In addition, a scatter of Mid to Late Saxon pottery, collected during fieldwalking to the north of the excavations, and the identification of a strong magnetic anomaly immediately to the east also suggest Anglo-Saxon activity to the north and east of the excavated area. Preliminary stratigraphic analysis and a spot-dating exercise, based on the ceramic evidence from the site, has allowed the provisional dating of the major phases of activity within the excavated area. Other than sporadic traces of a Romano-British settlement in the vicinity, suggested by several pits and various finds, the earliest evidence for a settlement located on the spur is provided by the Anglo-Saxon remains. Between the 7th and early 9th centuries, three major phases of construction are apparent from the superimposed buildings and other structures (Fig. 4). A major episode of refuse dumping then occuned during the first half of the 9th century, after the vast majority of earlier buildings had been demolished. Subsequent to this dumping phase, a new series of buildings was constructed on the same alignment as their antecedents, between the mid to late 9th century. These buildings were in turn demolished, with the onset of a second major dumping phase, characterised by the deposition ofvast quantities of animal bones. This horizon was followed by the accumulation of dark soils over large parts of the excavated area. These dark deposits contained pottery, bone and metalwork, among other finds and may have resulted from the levelling of middens on the periphery of an adjacent habitation focus, during the early decades of the 1Oth century. The discovery of an oven and pits, containing 12th-century Beverley ware, and a 13th- Flixborough- the character and economy of a high status Middle Saxon settlement in northern England 50 m 0 site north Later Medieval Ditch 0 0 Middle Saxon Ditch Fig. 2.- Provisional Plan of the main Anglo-Saxon settlement features recovered during the 1989-/991 excavations. century ditch cuting across the north of the excavated Middle Saxon settlement remains, suggests that the excavated area remained on the periphery of a settlement during the Anglo-Norman period. ·The Late Saxon and Medieval focus may lie immediately to the east of the excavations, around the now ruined 13th-century church of All Saints. After the 14th century, however, the spur previously occupied by the Middle Saxon settlement was inundated by windblown sand, up to 2 metres deep in places, and therefore abandoned for settlement purposes. 181 Chr. Loveluck fieldwalked Middle Saxon material / / / / / ./"/ ~'/ ,.. .... " / cropmarks I I I I I I I I / / / / N I I I I I I I I I I I I I I I I I I I \ site of .-1-"-1--~~­ Little or \ ' 1 1 N. Conesby X ! Deserted 1\ I II V Medieval t f\ Village (~ \ \ ~/ ~~~ ~"( I .f, I I \ I I 1988-91 I ,,." / I / I // /1'/ I / / / / / / I / I / / / / . s1le of +All Saint's Church I I I I I I I I I I I I I I I I I I I I I f f E /1 " I \ I I I I I \ I ~ ................\ \.~~~Grave \ Yard \ \ I I I I ,// / f 0 "'f' I : :/ I 1/ Middle Saxon features 1994-<:>95 ~ .. ~/ 11 I Jl I 11 I /1 ESCARPMENT I tl Ill Ill IJ I IJ I 11 I If I lt I f I 100 200 M Fig. 3.- Distribution ofAnglo-Saxon settlement and cemetery remains from Flixborough, discovered between 1988-1995. The buildings The buildings at Flixborough represent a significant addition to the relatively small body of information on rural architecture in the Middle Saxon period. Complete or partial building foundations and structural remains such as daub, charred timber fragments and structural ironwork allow for the identification of a series of construction techniques, between the 7th and lOth centuries AD. Variants of post-hole, continuous trench and sill beam foundations are evident among the buildings. All can be characterised as having a "hall" -type design and were constructed on a north-west to south-east alignment. Their dimensions 182 range from 9 metres by 5.30 metres to 19.70 metres by 6.50 metres. Internal fired clay hearth bases were also present in the majority of cases, often located in the eastern halves of buildings. ·with one exception, however, the foundation plans did not indicate the location of doorways. Figure 5 illustrates one example of each of the different architectural styles exhibited among the building plans. Buildings 1, 10 and 13 are exceptional in that they are the only examples of their foundation types. Overall, however, there does not appear to be any chronological progression in the use of particular earth-fast foundation styles, during different phases in the occupation sequence. Flixborough - the character and economy of a high status Middle Saxon settlement in northern England I I : A / I I slto north I olte north I I I I I I .I .. I Ditch A I 1, " I ··~I J Ditch I I I / I I I I I I I / / I I I I / I I '1, --- Buildin~.~ • I I I I ~ I ( I I I / I I I I I I 0 4 I 0 I I ·' I I I I I I I I I \ 7th-Early 8th century I I I I I I I I / I ---_.-"" ............ I I I Early-Mid 8th century PHASE 3.20 PHASE 3.10 I Ir ollo A north :. • Ditch Ditch I I I / ~ (;, 8 ' I /Building / I / I 13 I / / / I I I I I I / I I I I I I I I I I I I I I I I I 2 I I I I I I I I I century Mid-Late 8th century PHASE 3.31 I r I Dllch / I I A ulta norlh I / I / / I I I I I 4 I I I I I I I I I I I Mid-Late 9th Century PHASE 3.40 Fig. 4.- Provisional Phasing of the Anglo-Saxon occupation sequence. 183 Chr. Loveluck Flixborough 1989-91 building plans -r s 4 o ~0,Q~ J;~l e .J ~;:; a b c d r;tl 6:!::>~, ,o ~ 0 Q~ e 0 f 30m Detailed discussion of the buildings would be premature at this stage in their analysis, though some noteworthy points can still be made about the remains of individual buildings- especially numbers 1, 7, 10 and 13 (see Fig. 5). Among this group, building 1 has been identified as a possible church or mortuary chapel, based on parallels at other Anglo-Saxon sites such as Yeavering, Whithorn and more locally, Burnham (Hope Taylor 1977, 73-74; Hill 1991, 18 and Coppack 1986, 39-41, see Fig. 1). Its remains consist of a rectangular building, 13.60 by 6.20 metres in size, with a cobble and padstone wall footing at ground level, possibly for a timber sill, and a series of opposing post-holes in its long walls. The position of a doorway was also located in the middle of the southern long wall - the only doorway location currently identified at Flixborough. Four east - west aligned graves were cut into the floor of this building, containing three adults and an infant, while an additional burial was located outside the building, to the south. Building 1, however, does not seem to have remained an ecclesiastical building throughout the lifetime of the Middle Saxon settlement. Construction of hearths and the accumulation of occupation deposits over the original floor surface indicate its use as a dwelling, during the later stages of its exist184 HEARTH ~ Fig. 5. - The different foundation styles exhibited by the buildings: a) Building 5, Continuous trench and stone postpad; b) Building 1, Gravel and stone cobble sill for walls and external post-holes; c) Building 13, post-holes; d) Building 3, Post-holes in trench; e) Building 7, Continuous trench and stone cobble post-settings; f) Building I 0, Continuous trench and wooden sill beam. ence. Building 1 was subsequently replaced by building 10, which was also rectangular in shape, with a continuous trench earth-fast foundation, containing the ghost of a timber sill beam. It was 11.80 metres by 5.80 metres in size, with a fired clay hearth base in its eastern end. Both buildings 1 and 10 provide rare examples of earth-fast and raised sill beam bases for the support of timber superstructures, in Middle Saxon contexts. The cobble and padstone footing for building 1 is similar to the gravel and dry stone wall footings ofbuildings also thought to have had timber superstructures at Middle Saxon sites, such as Hartlepool, Dunbar, Whitby and Whithorn (Daniels 1988, 204; Holdsworth 1991, 315-316; Rahtz 1976, 461; Cramp 1993, 65-66 and Hill 1991, 18-23). Building 1 at Flixborough, however, is larger than all the stone-founded buildings at the latter sites, with the exception of"hall" 1Ob at Whithom which appears to be of a similar size (Hill 1991, 21 ). Buildings 7 and 13 also merit description because of their exceptional nature. The former is the largest building found at Flixborough, being 19.70 metres by 6.50 metres in size. It was rectangular and possessed a continuous trench foundation with limestone postsettings, regularly spaced along its long walls. The character of the foundations and the size of this build- Flixborough- the character and economy of a high status Middle Saxon settlement in northern England ing. Unlike other buildings from the site, large quantities of daub with wattle impressions were also discovered along the wall lines ofbuilding 13, suggesting the existence of wattle and daub wall panels for this building. It also shared a characteristic with buildings 5 and 7, in that it did not contain a hearth. ing are almost identical to those exhibited by building A from the Mid to Late Saxon phase of the settlement at Raunds-West Cotton, in Northamptonshire (Cadman & Foard 1984, 81 ). Buildings with very similar characteristics and dimensions were also found at Wicken Bonhunt, Essex and North Elmham, Norfolk (Rogerson & Wade 1973, 143; Marshall & Marshall 1991, 37-41; Wade-Martins 1980, 137). The lack of evidence for external post-holes, to provide foundations for raking timbers, suggests that the weight of the roof was placed on the long walls of building 7. Unusually among the Flixborough buildings, this hall does not possess a hearth on a floor at ground level. This absence may be accounted for by the possibility that building 7 had a raised floor or more than one storey. Building 13, in contrast, provides the only building at Flixborough with mainly post-hole foundations. It is 13.50 by 6.50 metres in size and is broadly rectangular, with double rows of post-holes along the long walls, representing the walls and raking timbers (Millett & James 1983, 242-245). Large post-holes, presumably for roof support, are also evident positioned on the central long axis of the build- The artefact and environmental evidence Moving from a consideration of the structural evidence to the artefact and biological remains from Flixborough, the value of the exceptional collection of finds is greatly enhanced by their discovery in stratified contexts, good preservation conditions and by the recovery strategies employed during excavation, particularly with regard to environmental remains. The circumstances of site formation, involving use of the spur for habitation phases and major refuse dumping episodes have ensured the survival of stratified deposits, containing an exceptional range and quantity of Middle Saxon artefacts, and a vast collection of fauna! remains in the form of animal e - I • . I OF=) "'"<:. . rgr:: Fig. 6. - Selection of 7th- to 9th-century gilded copper-alloy and silver dress accessories, recovered during the excavations (Scale 1:1). 185 Chr. Loveluck --• ; I( I ; • --11 • I l - I -· ~ -~~- • \) Fig. 7.- Selection of 7th- to 9th-century gilded copper-alloy and silver pins from the site (Scale I: 1). bones. The preservation of the animal bones was especially promoted by a calcareous component within the sand, forming the spur, and by the subsequent inundation of the settlement, beneath further sand deposits. As a consequence, some of the dumps and occupation horizons had never suffered acid leaching, resulting in the exceptionally good preservation ofbone and shell (Canti 1992, 18). It is difficult to summarise the nature and wealth of the artefact and fauna! assemblages due to their sheer size, and they are best discussed within the preliminary consideration of the economy and character of the settlement below. Nevertheless, it is appropriate to note that thousands of well-preserved artefacts were recovered, including decorated dress accessories (Fig. 6 and 7); domestic utensils; horse riding gear, harness and possible horseshoes; an exceptional collection of iron tools, relating to agriculture and a range of crafts (Fig. 8); textile-manufacturing evi186 dence; regionally-produced pottery, and a large quantity of lead artefacts and melt. Industrial debris from iron, lead and fine metalworking was also present, together with less conclusive evidence of glassworking in the form of a glass tessera, cull et and possible glassworking residues. The site also yielded imports from southern England, e.g. the largest collection oflpswich-type ware, north of East Anglia, as well as a large quantity of imports from northern France, the Low Countries and Germany. The imports included wheel-thrown ceramics, glass vessels, coinage and lava quem stones. In addition, iron, copper-alloy and silver styli were recovered, found in the fills of ditches or pits and within rubbish dumps. Two inscribed artefacts were also found, comprising a lead plaque with the inscribed names of seven individuals and an alphabet ring, showing the first eleven letters of the la tin alphabet (Fig. 9). The lead plaque was pierced by a series of rivet holes along its outer Flixborough - the character and economy of a high status Middle Sax on settlement in northern England ~·~.""'0: ~-.~i: :. .... .~ ; '· ....:· •': ,1: ; '; ~· ,,; ~~; .. :·... · .··,·: ...... ~ ·.~·:· -11111111 \ Fig. 8.- Carpentry, metalworking and leatherworking tools, found during the excavations (Scale 1 :2). edges which suggest that it may have come from a reliquary. It had been discarded within an occupation deposit, probably associated with building 10, from the mid to late 9th-century phase of occupation. The alphabet ring was unstratified. It is evident from the artefactual remains that there is a significant degree ofresiduality among the finds. A small quantity ofRomano-British pottery and metalwork was recovered, together with a collection of Early Anglo-Saxon dress accessories, including. 6th-century small. long brooches and a great squareheaded brooch and 7th-century annular brooches and a "safety pin" -type brooch (see Fig. 6). The occurrence of the Romano-British and early Anglo-Saxon material probably indicates the existence of as yet undiscovered Romano-British and Early AngloSaxon settlement foci, in the vicinity. A high level of residuality is also demonstrated in the deposition of finds manufactured in the 8th and 9th centuries, with ' large quantities of artefacts being deposited a considerable length of time after their manufacture- this is particularly the case with the refuse deposits. The latest Anglo-Saxon finds from the area of the excavations consist of Late Saxon wheel-thrown ceramics, comprising Torksey and Torksey-type wares and Lincoln wares, dating from the mid 9th to early 1Oth century. A lOth-century silver penny ofEdward the Martyr (975-979 AD), recovered as an unstratified find adjacent to the excavation, may also indicate a nearby Late Saxon phase of occupation. The recovery procedures relating to environmental remains were arranged with the Environmental Archaeology Unit, University of York, prior to the excavation. Three methods of collection were undertaken, involving the hand-collection of vertebrate, shell and certain plant remains, and the taking of "bulk-sieved" and "general biological analysis samples" (Dobney et al. 1992, 24-26). The overall hand187 Chr. Loveluck . ti'< ...... . Fig. 9. - The inscribed lead plaque and copper-alloy alphabet ring, together wit a silver stylus from Flixborough. (Scale 1: 1). collected bone assemblage weighs 5000 kg and a preliminary assessment suggests the presence of approximately 35,000 identifiable bones and a further 140,000 bone fragments. A broad species diversity was indicated among this material, with domesticated and wild animals represented (Dobney 1994, 190-193 ). The range and importance of this assemblage is discussed within the consideration of the provisioning of the settlement below. Partial analysis of a small proportion of the bulk-sieved samples (86 out of a total of 1759) also demonstrated that the remains of birds were present in the majority. Fish bones were also common and in some cases abundant in certain contexts, providing indications of the consumption of a wide range of riverine and marine fish (Dobney 1994, 194-195). In contrast to the excellent bone preservation, the soil conditions proved hostile to the survival of more delicate organic remains. Hand-collected plant and invertebrate remains comprised only small quantities of hazelnut shell, plum stones, egg shells of fowl and snail shells, although a large quantity of sea shellsmainly oyster - was collected, along with a few pieces of charred timber and a significant amount of charcoal. Within the examined bulk-sieved samples, 188 plant remains were limited to small amounts of widely dispersed charcoal fragments, together with sporadic charred seeds and herbaceous stem fragments (Dobney et al. 1994, 213-215). Insect remains were extremely sparse in the samples investigated, although snail shells were more numerous. These snail remains provide the only substantial source of information on environmental conditions in the vicinity of the settlement. Despite the more limited plant and invertebrate remains, they still allow for some appreciation of agricultural and horticultural practices, alongside the exceptional evidence for animal husbandry and exploitation of wild fauna! resources, provided by the vertebrate remains. The economy The combined evidence yielded by the environmental and artefact remains from Flixborough provides an exceptional opportunity for the investigation of the full range of activities constituting the economy of the Middle Sax on settlement. The inhabitants of the settlement were sustained by the products of a mixed agricultural regime, in- Flixborough- the character and economy of a high status Middle Sax on settlement in northern England volving a combination of arable cultivation and animal husbandry. This was supplemented by additional food procurement strategies, such as wild fowling, fishing and the hunting of other wild fauna. Arable cultivation and grain processing are indicated by the recovery of an iron coulter, charred seed grains of cereals and pulses, and a large collection of rotary quem fragments. The majority of the querns were imported from the Eiffel region of Germany. The tending or importation of orchard fruits is also suggested by the presence of plum stones. Preliminary analysis of a sample of the vertebrate remains indicates that cattle, sheep/goat, pigs, geese and chickens predominate among the domesticated animals, although significant quantities of horse and a small number of cat bones were also recovered. Detailed analysis will provide information on the extent of reliance on different livestock for dietary purposes, age ranges and seasonality relating to animal slaughter and butchery techniques. Conclusions will also be drawn on the extent to which the inhabitants relied on the consumption of stock raised within the confines of the settlement's immediate territory or whether it was provisioned by a combination of native production, renders and/or exchange. The wild fauna and flora exploited by the settlement were derived from a variety of environmental habitats. The bones of a large number of now extinct European Cranes, together with those of various ducks, suggest wildfowling in the Trent floodplain; while the remains of deer, hare, woodcock and hazelnut shells indicate exploitation of woodland or more open landscape. Large quantities ofboth riverine and marine fish, as well as oyster shells and net-sinkers, also demonstrate access to foodstuffs from the adjacent river and the open sea. The majority of these wild resources may have been derived from the immediate hinterland of the settlement, but as with domesticated livestock, they could also have been imported from other landholdings as renders, exchanged commodities or as resources for which there were common rights of access (Hooke 1981, 36; Loveluck 1994,333334). As a whole, the patterns of consumption of domestic and wild faunal resources at Flixborough will ultimately be compared with those derived from the study of analogous material at other Mid - Late Sax on sites. These will include rural settlements, such as Wicken Bonhunt; craft and trading centres, such as Fishergate in York, and larger nascent towns, such as Hamwic, with a view to examining the applicability of assigning interpretations of site status and character from vertebrate remains (Crabtree 1994, 50; O'Connor 1994, 141, and Bourdillon 1994, 123-124). The products of the agricultural economy provided most of the raw materials for the settlement's craftworking activities. Well-preserved tools and industrial debris have been identified relating to textile manufacture, carpentry, leatherworking, ironworking and non-ferrous metalworking (Fig. 8). All four stages of textile manufacture are represented: iron spikes from heckles or carding combs reflect fibre preparation of wool or flax; spinning and weaving is demonstrated by spindle whorls, pin-beaters and over 750 loom-weights, and the embellishment of cloth is indicated by the presence of shears, needles and thimbles. Carpentry is reflected primarily by an exceptional collection of woodworking tools, although dressed charred timber fragments also provide hints of carpentry techniques. The tools comprise axes and adzes (including T-shaped varieties), shaves, spoon bits for drills, rasps, wedges and chisels. Many of these artefacts were recovered as a hoard, housed in two large lead tanks (Leahy 1994, 352). Specialist iron tools also provide the evidence for leatherworking, namely "slickers", for cleaning and stretching tanned hides; "lunette" knives and "creasers", for the finishing of leather products (Ottaway 1994, 105). The metalworking evidence is provided by a combination of tools and manufacturing debris. Both iron smelting and smithing remains were recovered. Smelting was indicated by the presence of smelting slags, while smithing was indicated by smithing slag and partly worked objects. A collection of tools associated with ironworking includes punches, files and a small pair of tongs - although the latter item could also have been used for non-ferrous metalworking. The one fragment of iron ore recovered suggests the exploitation of bog iron rather than the nearby Frodingham ironstone deposits. Copper-alloy and leadworking evidence is provided by sheet metal offcuts, melt and scrap artefacts, together with seven fired clay mould fragments. Further analysis of the moulds should demonstrate whether they were used for the casting of copper-alloy or other fine metals. In addition to the evidence for the character of the agricultural regime and specialist craftworking activities, the range of imported commodities found at Flixborough shows that its inhabitants were extensively integrated within exchange networks with other parts of England and continental Europe. The pottery wares, coinage and raw materials, such as lead, indicate regular contact with areas linked by the East Midlands and Humber river systems, while the presence of four variants ofMaxey-type ware reflects either its regional exchange, or a common pottery tradition throughout Lincolnshire and Northamptonshire, in the 8th and 9th centuries (Vince & Young 1994, 56-62). Northumbrian stycas and Late Saxon Torksey-type and Lincoln pottery wares also demonstrate inclusion within trans-Humber and Lincoln189 Chr. Loveluck shire exchange networks between the 9th and early lOth centuries. In addition, the large quantity of lead recovered suggests links with the Peak District and Upper Trent valley. The Humberht charter, dated to 835 AD, records the regular export of Peak District lead to Canterbury, as a render to the Archbishop (Hart 1981, 111 ), and the relative abundance of lead on Mid - Late Saxon settlements at Flixborough, Riby, Lurk Lane-Beverley and Thwing (see Fig. 1) probably reflects the role of the Rivers Trent and Humber in the transport of this commodity to other parts of eastern England at this time. The discovery of the large Ipswich- type ware assemblage at Flixborough, consisting of a variety of types and forms, also reflects the role of the Humber estuary as a trading interface with south-eastern England. Other sites with Ipswich-type ware in the hinterland of the estuary include Riby, Bolton le Clay, Barrow-uponHumber, Lurk Lane-Beverley and Wharram Percy (Didsbury 1994, 237; Hayfield forthcoming; Watkins 1991,71-73 and Didsbury pers comm.). The collection of commodities derived from continental Europe also reflects trade and communication routes along the east coast of England. From the end of the 7th century, the inhabitants of Flixborough were in receipt ofwheelthrown Seine valley pottery, together with other wheelthrown red and black-burnished wares from northern France or Belgium, while small quantities of Badorf ware were also imported during the 9th century (Hodges 1981, 68- 84; Blackmore & Redknap 1988, 235). Other imports include the previously mentioned Eiffel lava quem stones and fragments of 89 glass vessels. The only comparable collection of vessel glass from Northern England is that derived from Fishergate, the Middle Saxon trading and craftworking settlement at York (Hunter & Jackson 1993, 1333-1339). The silver coinage from Flixborough also reflects links with southern England and north-western Europe. The coins include early 8th-century sceattas, ultimately derived from Frisia, and West Saxon broad penny coinage, minted between the mid 9th and late 1Oth centuries AD. Only one Mercian penny was identified - a late 8th-century issue of Offa. The West Saxon coinage, including issues of Aethelwulf (855865 AD), Aethelberht (858-865 AD), Alfred (871880 AD) and Edward the Martyr (975-979 AD), appears to have been the only coinage reaching Flixborough after the mid 9th century. These West Saxon issues indicate the maintenance of links with southern England at a time often characterised by Scandinavian disruption. Surprisingly, Anglo-Scandinavian coinage was not recovered, despite ceramic evidence suggesting occupation into the 1Oth century. 190 In total, the Flixborough settlement has yielded an exceptional range and quantity of imports in comparison with most Middle Saxon rural settlements. However, this apparent wealth may partially be a reflection of the fact that it is the most extensively excavated Middle Saxon settlement in the hinterland of the Humber. It has already been demonstrated that Flixborough is not unique in the region, with regard to its range of imports derived from other parts of England, and this is also the case with imports from continental Europe. Rescue excavations at Riby, in north Lincolnshire, also recovered northern French black-burnished ware pottery, Eiffel lava quems and a Frisian "porcupine" sceat from Middle Saxon settlement remains, within a series of enclosures (Steedman 1994, 222; Didsbury 1994, 246-249; Booth 1994, 272). The similarity in the distribution patterns of continental imports on both banks of the Humber is also demonstrated by the presence of lava quems and glass vessel fragments from the Middle Saxon monastery at Lurk Lane, Beverley (Foreman 1991, 106; Henderson 1991, 124), the recovery of a Merovingian pottery vessel from a 7th-century cemetery in Driffield (Mortimer 1905, 294), and by the occurrence of northern French black-burnished ware and Tating ware at Wharram Percy (Slowikowski 1992, 29). Taken as a group, all the above settlements benefited from integration within the same long-distance exchange networks, running along the North Sea coast to the Humber estuary and its feeder rivers (Loveluck 1994, 312). The character and status of the settlement The apparent similarity between the archaeological remains from Flixborough and material derived from excavations at documented Middle Saxon monastic settlements has previously led to the suggestion that Flixborough was a monastery (Whi twell 1991, 247; Yorke 1993, 146). Prior to any acceptance of this interpretation, however, it is first necessary to briefly examine the development of the criteria for defining the nature of Middle Saxon rural settlements. With a few exceptions, before the 1980s the vast majority of Middle Saxon sites which had been subject to excavation were major monastic centres, such as Monkwearmouth, Jarrow and Whitby (Cramp 1969, 21-66; Cramp 1976, 229-241; Peers & Radford 1943, 2788). As a corrollary to their documented possession of royal patronage, the structures and finds recovered from these monasteries were seen as characteristic of high-status monastic settlements. Analysis of the remains seemed to corroborate the textual evidence provided by Bede and other clerics, confirming that Flixborough- the character and economy of a high status Middle Sax on settlement in northern England these religious foci were also craftworking centres, enjoying extensive contacts with continental Europe. One of the legacies of the above excavations has been to indicate broad categories of evidence which are thought to characterise Middle Saxon monastic settlements. These include the possession of a characteristic settlement morphology, involving a planned layout within an enclosure, centred on a religious cult focus; the use of new building media, such as stone architecture; indications of specialist craftworking and long-distance exchange; and evidence ofliteracy, represented by styli or inscriptions. Before the mid 1980s, all the excavated sites viewed as monasteries also possessed textual evidence to reinforce their identification, thereby conditioning the interpretation of the archaeological remains. Flixborough, with its planned layout, a probable church, evidence for craftworking and luxury imports, together with a collection of styli, would fulfil most of the above criteria for defining a monastic settlement. However, if the Flixborough remains are examined alongside those from other Middle Saxon rural settlements, such as Wicken Bonhunt, Essex; Saint Peter' s-Northampton; Raunds-West Cotton, Northamptonshire, and more recently excavated sites at Brandon-Staunch Meadow, Suffolk and Riby, north Lincolnshire, it is evident that the long-held archaeological criteria for identifying monastic settlements must be re-assessed in order to advance the archaeological interpretation of the full spectrum of Middle Saxon rural settlements. Sites like Flixborough, Brandon, Saint Peter'sNorthampton, Wicken Bonhunt and Riby provide a contrast to monastic settlements, such as Jarrow, Whitby and Hartlepool in that there is no documentary evidence to influence interpretation of site character and status. The former sites are also materially wealthy rural settlements- although Flixborough and Brandon appear particularly rich- exhibiting a series of similarities in settlement morphology and structural characteristics; evidence for craftworking; trade and exchange and in three cases, literacy. The evidence for common traditions in aspects of spatial organisation can be illustrated in the widespread use of major enclosure boundaries to structure settlement layout at Flixborough, Brandon, Riby and Wicken Bonhunt (Fig. 2; Carr, Tester &Murphy 1988, 373; Steedman 1994, 221; Wade 1974, 175-176). Indeed, the weight of evidence from these sites and others, such as Raunds, Saint Peter's-Northampton, North Elmham, Goltho and Cheddar, suggests that many Middle Saxon rural settlements had at least, a loosely planned morphology, within or associated with enclosures or linear boundaries (Hamerow 1995, 16; Cadman & Foard 1984, 81-83; Williams 1984, 27; Wade-Martins 1980,54-55, Beresford 1987, 24; Rahtz 1979, 49-51 ). In the light of this observation, the previous identification of the Flixborough boundary ditch as a monastic vallum must be viewed with extreme caution. The nearest architectural affinities to the buildings at Flixborough, in terms of foundation style and building size, are also found at sites such as Raunds-West Cotton, Wicken Bonhunt and North Elmham, which have been interpreted as manorial centres for secular or ecclesiastical magnates. In contrast, the use of a gravel and stone footing for the probable church at Flixborough is akin to smaller buildings in the monastic settlements at Hartlepool and Whitby. However, buildings with gravel, cobble and dry-stone footings do not appear to be exclusive to monasteries. The documented Northumbrian royal centre at Dunbar, in Lothian, also possessed a phase of buildings with these footings (Holdsworth 1992, 43-44). The use of the presence of imported commodities as evidence for monastic character is also questionable. Like Flixborough, the putative magnate settlements of Wicken Bonhunt and North Elmham have yielded quantities of wheelmade Frankish pottery, from northern France and Belgium (Rogerson & Wade 1973, 143; Hodges 1980, 424-426). Other suggested high status centres, such as Saint Peter'sNorthampton and Riby, have also yielded imported Frankish pottery or other imports (Oakley & Hunter 1979, 298; Steedman et al. 1994). As a consequence, the occurrence of imported commodities on Middle Saxon rural settlements should not be relied upon as an indicator of monastic character. Similar reservations should also be held over the use of specialist craftworking evidence as a trait linked to monasteries, outside proto-urban centres. It is sensible to expect that Middle Saxon "secular" magnate settlements would also have supported dependent artisans, and that they would have been fully integrated into regional and longer distance exchange networks especially if they were located in key trading zones, such as the Humber Estuary. In conclusion, the similarities in structural characteristics between Flixborough and the Middle Saxon sites seen as manorial centres, together with the common occurrence of imported commodities from continental Europe and evidence for craftworking on these settlements, lead to the conclusion that the Flixborough remains are more likely to represent a wealthy manorial centre rather than a monastery. Nevertheless, the recovery of the collection of styli does suggest a significant ecclesiastical component within the social fabric of the settlement. This may have taken the form of resident clerics to serve the religious needs of the inhabitants and provide assist191 Chr. Loveluck ance in estate administration. An ecclesiastical, though not a monastic component, has also been suggested for the Saint Peter's settlement, at Northampton. This site has produced a stylus and the remains oflarge earth-fast timber buildings, succeeded by two mortared stone buildings. One of the stone buildings appears to have been the Middle Saxon precursor of Saint Peter's church. The settlement is interpreted as part of the administrative centre (caput) of a royal estate, with an associated Minster church (Williams 1984, 27). Following the brief summary of the archaeological evidence from Flixborough and the preliminary thoughts on its interpretation, outlined in this paper, the current programme of detailed analysis involving a multi-disciplinary team, funded by English Heritage, will result in a far more comprehensive characterisation of the settlement and its economy. The publication of the site will hopefully coincide closely with the completion of the analysis of the West Heslerton Anglo-Saxon settlement, with its Middle Saxon phase, together with the prospective publication of the Brandon, Thwing and Wicken Bonhunt Mid- Late Saxon sites. In conjunction, with the longstanding information on monastic settlements and the recently published reports on the Hamwic and Fishergate proto-urban centres, the detailed analysis of Flixborough, West Heslerton and other rural sites will enable a complete re-assessment of the nature of the Middle Saxon settlement hierarchy and the relationships between its constituent parts. Acknowledgements Thanks are extended to John Marshall for his preparation of the illustrations. References BERESFORD G. 1987: Goltho: The development of an early medieval manor, c. 850-1150, English Heritage Archaeological Report 4, London. BLACKMORE L. & REDKNAP M. 1988: Saxon and Early Medieval Imports to the London area and the Rhenish Connection, in: GAIMSTER D.R.M, REDKNAP M. & WEGNER H. (eds.), Zur Keramik des Mittelalters und der beginnenden Neuzeit im Rhein/and, British Archaeological Report, International Series 440, Oxford, 223-239. BOOTH J. 1994: Coins, in: STEEMAN K., Excavation of a Sax on Site at Riby Cross Roads, Lincolnshire, Archaeological Journal151, 212-306 (272). BOURDILLON J. 1994: The animal provisioning of Saxon Southampton, in: RACKHAM J. (ed.), Environ192 ment and Economy in Anglo-Saxon England, CBA Research Report 89, York, 120-125. CADMAN G. & FOARD G. 1984: Raunds: Manorial and Village Origins, in: FAULL M.L. (ed.), Studies in Late Anglo-Sa;wn Settlement, Oxford, 81-100. CANTI M. 1992: Research into Natural and Anthropogenic deposits from the excavations at Flixborough, Humberside, Ancient Monuments Laboratory Report 53/92, English Heritage. CARR R.D., TESTER A. & MURPHY P., 1988: The Middle Saxon settlement at Staunch Meadow, Brandon, Antiquity, Vol62, No. 235, 371-377. COPPACK G. 1986: St. Lawrence's Church, Bumham, South Humberside: The excavation of a Parochial Chapel, Lincolnshire History and Archaeology 21, 39-60. CRABTREE P.J. 1994: Animal Exploitation in East Anglian Villages, in: RACKHAM J. (ed.), Environment and Economy in Anglo-Saxon England, CBA Research Report 89, York, 40-54. CRAMP R.J. 1969: Excavations at the Saxon monastic sites ofWearmouth and Jarrow, Co. Durham: an interim report, Medieval Archaeology XIII, 2166. CRAMP R.J. 1976: Monastic Sites, in: WILSON D.M. (ed.), The Archaeology of Anglo-Saxon England, Cambridge, 201-252. CRAMP R.J. 1993: A Re-consideration of the Monastic site ofWhitby, in: Spearman R.M. & Higgit J. (eds.), The Age ofMigrating Ideas- Early Medieval Art in Northern Britain and Ireland, National Museums of Scotland Publication, Edinburgh, 6473. DANIELS R. 1988: The Anglo-Saxon Monastery at Church Close, Hartlepool, Cleveland, Archaeological Journa/145, 158-210. DIDSBURY P. 1994: The Pottery, in: STEEDMAN K., Excavation of a Saxon Site at Riby Cross Roads, Lincolnshire, Archaeological Journal 151, 212306 (228-249). DOBNEY K. 1994: Animal Bone Assessment Report, in: WHITWELL J.B. et al., Flixborough Middle Sax on Settlement Excavations 1988-91 -Material Assessment Report, Humberside Archaeology Unit report, for English Heritage, 189-200. DOBNEY K., HALL A., KENWARD H. & MILLES A. 1992: A working classification of sample types for environmental archaeology, Circaea 9, 24-26. DOBNEY K., HALL A., KENWARD H. & MILLES A. 1994: Environmental GBA and BS samples Assessment Report, in: WHITWELL J.B. et al., Flixborough Middle Saxon Settlement Excavations 1988-91 -Material Assessment Report, Humberside Archaeology Unit report, for English Heritage, 210-224. Flixborough - the character and economy of a high status Middle Saxon settlement in northern England FOREMAN M. 1991: The Stone and Fired Clay, in: ARMSTRONG P., TOMLINSON D. & EVANS D.H., Excavations at Lurk Lane, Beverley, 1979-82, Sheffield Excavation Reports 1, 105-114. HAMEROW H.F. 1995: Shaping Settlements: Early Medieval Communities in Northwest Europe, in: BINTLIFF J. & HAMEROW H.F. (eds.), Europe Between Late Antiquity and the Middle Ages, British Archaeological Report, International Series 617, Oxford, 8-37. HART C.R. 1981: The North Derbyshire Archaeological Survey, to AD 1500, Sheffield. HA YF!ELD C. forthcoming: The Pottery, in: FOREMAN M.F. (ed.), Excavations at St. Chads, Barrow-upon-Humber, Lincolnshire. HENDERSON J. 1991: The Glass, in: ARMSTRONG P., TOMLINSON D. & EvAns D.H., Excavations at Lurk Lane, Beverley, 1979-82, Sheffield Excavation Reports 1, 124-130. HILL P .H. 1991: Whit horn 4 - Excavations 19901991, Interim Report, Whithorn Trust, Stranraer. HODGES R. 1980: Characterisation and Discussion of Identified Imported Sherds, in: WADE-MARTINS P., Excavations in North Elmham Park, 19671972, East Anglian Archaeology Report 9, Vol II, Norfolk Museums Service. HODGES R. 1981: The Hamwih pottery: the local and imported wares from 30 years' excavations at Middle Saxon Southampton and their European context, Southampton Archaeological Research Committee Report 2 = CBA Research Report 37, London. HOLDSWORTH P. 1991: Dunbar, Current Archaeology No. 127, VolXI, No. 7, 315-317. HoLDSWORTH P. 1992: A multi-period settlement on the Lothian coast, Medieval Europe 1992, Rural Settlement- Pre-printed Papers, Vol 8, York, 4146. HOOKED. 1981: Anglo-Saxon Landscapes of the West Midlands: the Charter Evidence, British Archaeological Report, British Series 95, Oxford. HOPE TA YLOR B. 1977: Yeavering- An Anglo-British centre ofearly Northumbria, HMSO, London. HUNTER J.R. & JACKSON C.M. 1993: Glass, in: ROGERS N.S.H., Anglian and Other Finds from Fishergate, The Archaeology of York- the Small Finds Series 17/9, York, 1331-1344. LEAHY J. 1994: The Flixborough Hoard, Current Archaeology No. 14, 352. LOVELUCK C.P. 1994: Exchange and Society in Early Medieval England, 400-700 AD, unpublished Ph.D. thesis, University ofDurham. LOVELUCK C.P. 1996: The Anglo-Saxon Settlement and Cemetery Remains From Flixborough - Revised & Summarised Assessment and Updated Project Design, Humber Archaeology Partnership report, for English Heritage. MARSHALL A. & MARSHALL G. 1991: A Survey and Analysis of the Buildings of Early and Middle Anglo-Saxon England, Medieval Archaeology XXXV, 29-43. MILLETT M. & JAMES S. 1983: Excavations at Cowdery's Down, Basingstoke, Hampshire, Archaeological Journall40, 151-279. MORT!MER J.R. 1905: Forty Years' Researches in British and Saxon burial mounds of East Yorkshire, London. 0AKLEY G.E. & HUNTER J.R. 1979: The Glass, in: WILLIAMS J.H., St. Peter's Street, Northampton, Excavations 1973-1976, Northampton, 296-302. O'CONNOR T. 1994: 8th-11th century economy and environment in York, in: RACKHAM J. (ed.), Environment and Economy in Anglo-Saxon England, CBA Research Report 89, York, 136-14 7. OTTAWA Y P.J. 1994: Ferrous Metal Assessment Report, in: WH!TWELL J.B. et al., Flixborough Middle Sax on Settlement Excavations 1988-91 - Material Assessment Report, Humberside Archaeology Unit report, for English Heritage, 96-106. PEERS C. & RADFORD C.A.R. 1943: The Sax on Monastery ofWhitby, Archaeologia 89, 27-88. RAHTZ P. 1976: Appendix C: The building plan of the Anglo-Saxon monastery ofWhitby Abbey, in: WILSON D.M. (ed.), The Archaeology of AngloSaxon England, Cambridge, 459-462. RAHTZ P. 1979: The Sax on And Medieval Palaces At Cheddar, British Archaeological Reports, British Series 65, Oxford. ROGERSON A. & WADE K. 1973: Wicken Bonhunt, Note, in: Medieval Britain 1972, Medieval Arcltaeology XVII, 143. SLOWIKOWSK! A.M. 1992: Anglo-Saxon and Medieval Pottery, in: MILNE G. & RICHARDS J.D., Wharram: A Study of Settlement on the Yorkshire Wolds, VII- Two Anglo-Saxon Buildings And Associated Finds, York University Archaeological Publications 9, 27-38. STEEDMAN K. 1994: Excavation of a Saxon Site at Riby Crossroads, Lincolnshire, Archaeological Journal151, 212-306. VINCE A.E. & YOUNG J. 1994: Pottery Assessment, in: WHITWELL J.B. et al., Flixborough Middle Saxon Settlement Excavations 1988-91, Material Assessment Report, Humberside Archaeology Unit Report for English Heritage, 56-67. WADE K. 1974: Wicken Bonhunt, Note, in: Medieval Britain 1973, Medieval Archaeology XVIII, 175176. WADE-MARTINS P. 1980: Excavations in North Elmham Park, 1967-1972, East Anglian Archae193 Chr. Loveluck ology Report 9, Vol I, Norfolk Museums Service. WATKINS G. 1991: The Pottery, in: ARMSTRONG P., TOMLINSON D. & EYANS D.H., Excavations at Lurk Lane, Beverley, 1979-82, Sheffield Excavation Reports 1, 61-103. WHITWELL J.B. 1991: Flixborough, Current Archaeology No. 126, Vol XI, No. 6, 244-247. WILLIAMS J.H. 1984: A Review of Some Aspects of Late Saxon Urban Origins and Development, in: FAULL M.L. (ed.), Studies in Late Anglo-Saxon Settlement, Oxford, 25-34. YoRKE B. 1993: Lindsey: The Lost Kingdom Found?, in: VINCE A. (ed.), Pre-Viking Lindsey, Lincoln Archaeological Studies 1, Lincoln, 141-150. Dr Christopher Loveluck Humber Archaeology Partnership The Old School Northumberland A venue Kingston upon Hull HU2 OLN UK 194 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 M. A. Aston The Shapwick Project, Somerset, England Introduction Shapwick is a parish of 1284 hectares in the middle of the county of Somerset in south-west England mid way between the small market towns of Bridgwater and Glastonbury. Today there is one village running north-south for two kilometres in the centre of the parish and the land is worked from six farmsteads, one in the village and five out in the fields (Fig. 1). The population is just over four hundred. About one third of the parish is low lying fen land, part of the Somerset Levels, while the rest is on the rising land of the Polden Hills running west out to the river Parrett (Fig. 2). The geology of the area is recent peat and alluvial deposits in the north of the parish while in the south the base is liassic rocks, principally limestones and clays. The parish was owned by Glastonbury Abbey from 729ad to 1539; from probably the twelfth century the land was split between two manors, the larger belonging to the abbot and the smaller to the almoner. From the sixteenth century there have been a large number of owners of the two manors. They often seem to have been at odds with each other over land matters and it may have been this litigation that produced the profusion of documents and maps which have survived for the post-medieval period. Now the parish and each of the farms, though not most of the village, is owned by Lord Vestey and it is through the kind co-operation ofhis land agent for the estate, Bill Robbins, that the project has been able to take place so successfully. The project was set up with three main objectives. Firstly to test the hypothesis that Shapwick is a planned village, where the village and its medieval common fields were laid out to replace a more scattered pattern of hamlets and farmsteads with their own field systems. Secondly to develop further the methodology and the field techniques for what has come to be called landscape archaeology, especially map analysis, field work techniques and geophysical and geochemical prospecting. It was also intended to use the project to teach students of all ages and on all sorts of courses, and also where appropriate to involve the local people and keep them informed of progress with the research. Background As a result of the last fifty years of research on rural settlements in Britain a general model can be proposed for the development of villages and farmsteads from the post-Roman to the post-medieval period (500 to 1500ad). This model suggests that from a variety ofRomano-British settlement types a predominantly dispersed pattern of hamlets and farmsteads developed by the Anglo-Saxon period. Even though large nucleations existed in the late Roman period nothing like this seems to have existed in later centuries. Over much of the country the dispersed pattern has persisted through to the present day, although it is possible to demonstrate that there have been considerable changes in detail. Elsewhere the settlement pattern was dominated by much more nucleation in the middle ages and later; usually this is associated with the most fully developed common field systems. The 'village belt' of England runs from east Somerset and Dorset in the south, up through the midland counties of Oxfordshire, Warwickshire, Leicestershire and Northamptonshire, and through to the north east, Yorkshire, county Durham and Northumberland. Again this general picture is very variable in detail. Villages are now seen as rather late developments, being founded only in certain areas. The date range is quite wide with some archaeologists suggesting villages may have been founded in the ninth or tenth centuries, others saying the eleventh and twelfth centuries and some historians seeing the process going on to the thirteenth century. While we can be more specific for certain areas and for particular examples, in general we seem to be looking at the period 800 to 1200ad as the main period for village foundation. How this was achieved was very variable but two important strands can be distinguished - the planned 195 M. A. Aston Fig. 1. - General map of the location of Shapwick. Location of Shapwick ewens SHAPWICl< • Brldgv.<~ter e,l Glo<tonbury est~t 8TAUNTON i Sweet track ,.li / // LEVELS 0 working farms lkm village and the polyfocal (or agglomerated) village. The former was the result of planting a new village, deliberately laid out to a regular plan on a new site, the latter the growing together of a number of separate centres, such as the church, manor house, mill, a green or a ford, with the addition of planned blocks of properties. There is little room in this model for the process of colonisation with new farmsteads built on former forested or marshy lands. This clearly happened and there are some clearly documented examples but much 196 of the colonisation model was based on an assumed chronological development for placenames and the first documentary references to particular settlements; neither of these forms of evidence need necessarily be related to the date of foundation of a settlement. With the new model there is an assumption that hamlet and farm sites are much older and more persistent than was formerly thought; villages are added in some areas at specific times and replace the pattern of older hamlets but there is otherwise relatively little foundation of absolutely new settlements except for a few The Shapwick Project, Somerset, England LZr]g~----------~L~0------------~41~-----------4~2------------~L~l------------~LL~-----------;'~z SHAPWICK PARISH in 1904 LO )9 39 ]8 Ftom Ordoonc.e $U(vey 6" 1m1te Mops Sec.ond Ed1hon 190!. MA•lon December 1993 37 J~9------------7,,0,-----------~,I------------~,~2------------7.o~----------~,~,--------~--~4S 37 Fig. 2. -Map of the parish of Shapwick in 1904. medieval farms and large numbers of post-medieval farms following enclosure. For some time the author was engaged in looking for a parish or area where some of the ideas in this model could be tested in the west of England and specifically in the historic county of Somerset. A number of parishes were looked at and in the end Shapwick was picked because of the research already carried out by Nicholas Corcos. He had suggested that the village looked as if it had been planned from the evidence of eighteenth century maps, that the common fields were very regular with similarly sized holdings in each of the two medieval common fields, and that certain furlong names as recorded in a survey of 1515 were of a 'habitative' type (cote, wick, worth, ton and so on), that is they looked as if they should really be attached to settlements rather than areas of fields, suggesting a different field system if not a more dispersed pattern of settlement. On the basis of this evidence Shapwick was chosen in 1988 as the basis for a ten year project which we are now some eight years through. The Shapwick Project There have been two main objectives to the project, one involving a close examination of the village and the other to look in great detail at the development of the landscape of fields in the parish. While these exercises have often run concurrently often more work is going on in one than the other. The Village The village has been examined to see if there is any evidence that it was planned at a particular date. To do this it is necessary to identify changes in the plan particularly those of recent centuries, allow for 197 M. A. Aston -----.. ----, SHAPWICK VILLAGE Fig. 3.- Plan of the village of Shapwick in 1764-5. 1764-5 estate maps 0 0 100metres 500 feet these, and remove them in order to get at a more original plan. To do this a great deal of map analysis has been undertaken, together with an examination of all the historic buildings in the village. This been carried out along with morphological analysis of the village's plan of streets, lanes and properties as well as a series of small scale 'opportunistic' excavations undertaken as circumstances presented themselves. 198 Map Analysis (Figs. 3 and 4) When the project began it was known that there was at least one map earlier than the tithe map of 1839 but as work progressed it became clear that there were very many maps of the eighteenth century including several of the village and others of the early enclosed areas of the fields and of the last remaining The Shapwick Project, Somerset, England Fig. 4. - Plan of the village of Shapwick in I 885. SHAPWICK VILLAGE 1885 OS 25inch SH Shopwick House KL Kent Lone HF OH Home Form Down House CF SMC NF Church Farm SI Mary's Church s New Farm School KF LF King's Farm Lawn Farm LH V Lawn House Vicarage SM HF Smithy Hilt Form M Shapw1ck flour mill a Quarry Spring SP .... .. ..... ...... ...... ...... f ,o 1' ' f ' ' p pump w well .....' ...... '! . f 0 0 •• 100 metres 500 feet bits of the common fields with furlongs and strips. From these map sources it could be shown that there had been substantial alterations to the village plan in the decades before around 1800. These consisted principally of the removal of a large area of the north end of the village, with several streets closed off and many houses demolished to create a park to the east and south of the main manor house. A by-pass road was constructed to the west and eventually many of the roads left as cul-de-sacs to the east of the park were stopped up and removed. Elsewhere in the village a more direct north-south route was created by cutting a new road through some houses and plots on the eastern side of the southern half of the village. The fine series of maps enable us to see these changes and many other alterations in great detail. 199 M. A. Aston I I L.______J I L Fig. 5. - Reconstruction drawing of the manor house roof (lane Penoyre). Historic Buildings Analysis One building in the village, the second smaller manor house, now a school, had been the subject of a building survey before the project began. The Somerset Vernacular Buildings Research Group were therefore approached and asked if they would consider surveying all the other buildings in the parish. This they agreed to do and they have now finished this phase of the work and published a monograph on all the buildings of the parish dating to before the end of the nineteenth century. Some surprising results came from this work. The only thatched cottage in the village, Forsters, which was known to be late medieval with an inserted floor and stacks in the sixteenth and seventeenth centuries, and unlike most houses in the village seems to have survived substantially unaltered, turns out to have late medieval smoke-blackened thatch still remaining on the north side of the roof. This is of rivet wheat (triticum-) and is one of only a small number which have been identified in southern England by John Letts in an English Heritage project. Most of the rest of the housing stock in the village was rebuilt or substantially altered in the decades around 1800 when there seems to have been an attempt to make an estate or 'closed' village along with the emparking scheme. The most interesting building is the main manor house now a hotel. Since the late eighteenth century 200 this has been assumed to be a seventeenth century building; it has transom and mullion windows. Clearly it has been altered in the eighteenth, nineteenth and twentieth centuries. However on examination it was found to have late fifteenth century roofs over the main hall area and each of the cross wings. Furthermore, a formerly detached building to the west was shown to be a detached medieval kitchen with smoke blackened timbers. Various buildings were added to link up these units in succeeding centuries. Dendrochronological dates have been obtained, by Dan Miles, for the medieval buildings; the hall (Fig. 5) produced a felling date for the oak trees of spring 1489, and the detached kitchen of spring 1428. A survey of 1515 says that the manor house was inside a moat and there is indeed a moat to the north of the house near the gardens. This is very small and it was always felt that it was too small for the site of the abbot's manor house. Now that a fifteenth century house has been identified within the hotel house this must be the house that had a moat around it. There is little trace of this on the surface but a very detailed earthwork survey by James Bond has identified much of its course (Fig. 6), the area of the island is recorded in 1515, and although it is not clear from geophysical surveys, a trench dug for an electricity cable located the moat edges, probable bridge supports or retaining walls, and showed from the very rich collection of finds in the moat that it was probably backfilled in around 1625; it seems to have constructed made in the thirteenth century. Also, if this is the house mentioned in 1515, which seems to be the case, then we know that it had a garden. This cannot have been in front of the house to the south as a great barn, probably the farmyard, and part of the village lay in that area; it is most likely to have lain to the north where there were two fishponds, the smaller moated site and later on the banquetting house. In order to test out this idea a geophysical survey was carried out in the field north of the house. While this revealed rectilinear anomalies which probably do represent garden features, the dominant features are curved lines and linear anomalies on a different alignment to the gardens. An excavation was carried out in 1993 which showed that beneath a layer of topsoil there was an extensive late Iron Age and Romano-British site with house foundations and gulleys and a mass of pottery, bone and metalwork; this probably represents a 'lake village' type of site such as those at Glastonbury and Meare. Geophysical survey has so far failed to define this site in any direction though it is unlikely to go much further to the north or south. The examination of the manor house and the adjacent structures provides a very good example of the The Shapwick Project, Somerset, England multidisciplinary nature of this project, the methodology and the wide variety of techniques being employed. Excavations within the Village area A number of excavations have been carried out within the village. The earliest was on the site of the new sports hall at the school. This showed how thorough the demolition ofbuildings had been when the park was created in the late eighteenth century; the earliest evidence of occupation was twelfth century pottery. Another opportunity was provided when the local water authority put in a new sewerage scheme which involved a wide trench across the park through the area of the former village. Excavations were carried out before construction began and these revealed the foundations of a number of post-medieval houses together with evidence of a boundary bank between the village closes and the common fields which was in existence in or by the tenth century; there were also a number of features with tenth century pottery and some evidence for twelfth century buildings. Excavations in July 1996 failed to locate any buildings of the tenth century adjacent to the line of the pipe trench but did identify boundaries on the same alignment as most of the boundaries in the village today, associated with pottery of this early date. Very limited excavation has taken place at the south end of the village but nevertheless this suggests that although the area was in use by the tenth century, judged from the few pottery finds of that date, no structures have been identified so far, The most extensive excavations took place in 1994, 1995 and 1996 in the middle of the village between the church and the smaller manor house. There, it is clear that the village alignments were in existence by the tenth century and that one building of that date at least was aligned on one of the lanes on N L1 SHAPWICK, Somerset: EARTHWORKB AT SHAPWICK HOUSE 0 Metres so CJeond l'ft4 Fig. 6. -Plan of the earth works of the moat at the manor house (lames Bond). 201 M. A. Aston 42 Nr-----------~4~0----------~41____________~42__________~4~3--------·--~44~--------~45 SHAPWICK PARISH ·· ... in 1754 and 1762 '•, '•, ···. ·· .... '•, 41 ·· .. 41 ·· ... '•, ··· ... 40 39 39 38 38 Comp1ted lrom lhGo ten mops of 1754 from SCRO 00/SG 36 c/206 '----'f.-,--'\_ of 1754 ond the nine mops of post 1762 of ouo~ of common field SCRO 00/SG 39 c/206 cl760 J/39 40 M As.ton Oe<ember 1993 41 42 43 44 J1 45 Fig. 7.- Map of the parish ofShapwick in 1754 and 1762. the same orientation. The were then many changes in the early middle ages before this part of the village was abandoned and became an open croft. Test Pits It is obviously very difficult to carry out a lot of archaeological excavation in a village that is still occupied and in use with houses, gardens, lawns, drives and so on. The project has therefore adopted the strategy of digging one metre square test pits, where we were allowed to, in peoples gardens. The stratigraphy is recorded together with any finds, a section is recorded and the nature and depth of the bedrock noted. Twenty one of these were dug in the spring of 1995 and a further 12 in spring 1996. Surprisingly, since this did not seem like a very productive exercise when we were carrying it out, the results have proved to be very useful, enabling us to map occupation in different parts of the village at different times. 202 Pottery of the tenth century occurs from one end of the village to the other while twelfth-thirteenth century pottery is only found in the central area. Little is found after the fourteenth century and it is assumed that it is being taken out to the fields along with the manure. Analysis of the Village Plan All of this research is used to generate models of the how the village may have developed and we are on the third version as our ideas change and we obtain more information. Currently it is felt that the whole of the area of the medieval village was used, and had pottery abandoned in it by the tenth century. However as we shall see (below) it would be very difficult for us to recognise any occupation earlier than this date as there is no earlier Saxon pottery in Somerset as in so many parts of the west of England. There does not appear to be anything Roman under ""l a?" CO () "" ~ BEERWA Y FARM, SHAPWICK, SOMERSET. MAGNETOMETER SURVEY OF THE OLD CHURCH SITE. ~ "';::;· !::.. N Greyscale of magnetometer data overlain on earthwork survey. "' ;::: ~ ~ ~ ::;.. ""0 ~ '"' ~ ;::: ;:; ~ ~­ ~ t;] ::: ~ c:;· ~ ~ ..., ~~ ;1 ~ C/.l ::r "' "d ::E ;:;· ;>;" ....'"t:l .3. ~ "~ C/.l 0 -4.5 ·1.5 1.4 3 .... 4.3 ~ -------__ _ nT "' "~ ------ -:...::....::-=- tr1 0 ~m ::l ANCIENT MONUMENTS LABORATORY. IJQ j;J N 0 w a ::l P- M. A. Aston -25-40 1]]40-45 ~45-5 ~50-55 155-75 1>75 • • • • • Fig. 9.- Plot of heavy metal over the old church site. lOO metres the village, except perhaps at the north end. Analysis of the pattern of streets, lanes and property boundaries suggests that these were fitted in between several pre-existing parallel roads running east west along the Polden Hills. While the layout of the lanes is different at each end of the village, with large open enclosures to the south and a pattern of narrow ladderlike lanes in the north, there is at the moment no clear evidence that these were laid out at different times. While some of the lanes may have been planned along with the properties between, it is likely that the layout of the larger units of the village were conditioned by pre-existing boundaries, possibly by the lynchets and field boundaries of a prehistoric or Romano-British field system across the area (see below). Within the pattern of roads and lanes the tenements seem to be of very regular and consistent sizes. Little research has been carried out so far on this aspect of the village plan, but it looks as if a regular module of length of one hundred feet or multiples thereof has been employed to lay out the properties between the lanes. In the middle ages the 'rope' of twenty feet was the standard measurement in Somerset and so the modules may have been based on this. More analysis of the village plan is clearly desirable. 204 The Parish A great deal of research has been carried out in the rest of the parish, principally on the 'upland' part away from the low-lying peat and fen land of this part of the Somerset Levels. Most of this work has involved fieldwork on ploughed land. Map Analysis The fine series of maps means that it is possible to analyse the changes in the pattern of fields, woodlands and lanes over the last quarter of a millennium. The earliest map of the whole parish is of 1787; the tithe map dates to 1839 and then from the 1880s there is a fine series of large scale Ordnance Survey maps (for the early 1900s and the 1930s) before the last modem metric mf'p of 1971. However it has proved possible to construct a map of the 1750s and 1760s from a series oflarge scale map lets Fig 7)- ten maps of 1754 showing areas of enclosed land and nine maps of post-1762 showing the remaining areas of strips in the common fields. This is important for several reasons. It gives us a clear view of what at The Shapwick Project, Somerset, England t Roman pottery • Fragments: o1 4-5 6-8 17 Fig. 10.- Sladwick pottery distribution and geophysical survey. least part of the common fields looked like, including the likely size and density of the strips, and it hints that much of the pattern of the enclosed fields is de- rived from the pre-existing pattern of furlongs and strips by a process of piecemeal enclosure which is well documented in the parish. 205 M. A. Aston BC AD 500 0 1000 • ?\ \ \ s \ \ ?__._?\ \ s \ \ \ ? 2000 1500 ASHCOTT \?~"" """"""""'\ CATCOTT \ 0"" """" """"s: Black land ENWORTHY + ~ ~ ~ _-_- ---~ ~ ~~A;bc=h=e-s_t=e=r======~?~~~~~?----t:•_o_o~t CHURCH! : 0 0 0 0 0 re occupied sites 0 0 0 0 0 0 0 Old Church ?----0---?---' SLADWICK Fig. 11. - A model for the development of the settlement pattern in the parish of Shapwick. Field Names Furthermore, and of the greatest importance, is that these maps contain large numbers of field names. The best source for these is the tithe map of 1839 but earlier maps contain additional or previously used names so that a very detailed picture can be built up of the minor names used to describe features in the landscape of the parish. There are no maps before the middle of the eighteenth century but there many documents which mention the same or additional field names. Since one of the main objects of the project is to locate any early settlements from the occurrence of 'habitative' names, it is clearly of great importance to locate these as closely as possible on the ground within the pattern of the later fields. This is being done by a complicated process which has been developed using the survey of 1515 which lists every tenant, tenants holding, plot ofland and its area, and comparing all of these aspects with names and areas on the tithe and modern maps. As we !mow the whereabouts of some of the names recorded in 1515 from later sources, and as the clerk recorded the plots mainly in a clockwise sequence, we can often suggest where a furlong or field name not recorded on later maps may be located; frequently the areas of the furlongs can be compared with modem hectares to help in this process. This rather long-winded process is necessary in order to locate those 'habitative' names which have not survived as later field names (most of 206 them) and so enable us to carry out further research on those fields using archaeological techniques (see below). Botanical Survey A full survey has been made of the hedges in the parish and work has begun on the hedgerow trees and the woodlands, only one of which, Loxley Wood, is of early origin. The theory that hedges can be dated by the number of woody shrubs in them does not seem to work in Shapwick - many of the hedges created in the eighteenth century enclosures have the highest numbers of species. Under the direction of Mike Martin further analysis has been undertaken of the complete species count for all hedges, both sides of the hedge with all species listed. These data have been subjected to a multi-variate analysis, Twinspan, which seeks to identify the most different group of species from the rest, then the next most different and so on. This shows that the hedges along the streams, along the roads and on the parish boundary are different from the rest but generally it does not help with dating. At the moment we do not !mow the date when many of the hedges were created and we cannot see if there is a direct relationship between numbers of species and date of hedge. It is clear that the reasons for variable numbers of species in hedges are numerous and at the moment little understood. ~ a;'Q" - •RON AGE POST-ROMAN and ANGLQ-SAXON N ~ A c (1) ~ levels -<: (1) ~ ;:: .--·· ·- -----.3·-··--. _ _ (1) ;:::: ROMANO- BRITISH B Levels - (:-_:_~~#;~~~)::'~-­ /I) '0(~"" •±"" '""''"' HJ.WO""' ~ ;:;. (1) V, levels ~ ~ ;:: (1) ~ EDBUSCOTT 1::. ;:::: !::>.. iS"' GASHCOTT ;:::: !::>.. t; (1) '"1:::! ~ ~ ~ ;=· ;:;. (1) '"1:::! ...1::.<::;· ;::,-.. ~ §;F {3 ;t ;::;· ?;- EARLY MEDIEVAL LATE MEDIEVAL E D Levels \,..Ll-'- 4D PRE-MEOIEVAL SETTLEMENT .:{--. PRE-MEDIEVAL fiELDS Levels I +CHURCH ;1 1111 MANOR 0 (I) 9 9 WOODLAND ·>: ::r" .§ *MILL MARSH =MEADOW / / ./. flll E A S\. T b!MEOIEVAL SETTLEMENT :e()" ?;" ...."'Cl ~0 §:;. (I) 0 30 .... "' -~ 3 Kll OMfTRES M Asron August 1996 t'I1 ;:l (IQ N 0 --..! iiO ;:l Q. M. A. Aston Field Survey, Finds Analysis Most effort as far as fieldwork is concerned has been expended in field by field survey as areas are ploughed for crops. From the beginning it was decided to cover as much of the 'upland' part of the parish as possible (some 800 hectares) by fieldwalking as land was ploughed, to collect all material indicative of human activity. It was decided to collect even the most recent material, partly because this had not been done before and we were interested to see what it might indicate, but also because many of the people taking part in the fieldwalking exercises had little experience of the finds of different periods. By asking them to pick up everything they have inadvertently brought in much material of great interest and importance that they might have ignored or discarded if they had been allowed to be selective. Initially a number of areas were sampled with ten metre squares, but very rapidly we switched to lines 25 metres apart with 25 metre long stints fitted rather randomly into the fields using the field boundaries as the base lines. For each stint a collection time of ten minutes is allowed and it is clear, from observation of fieldwalkers working, that a strip about 2.5 metres wide is examined; this means that in effect a ten per cent sample is being taken from the surface of each field. So far over eight years, 86 fields have been fieldwalked over an area of several hundred hectares. This represents over 60 percent of the 'upland' area and probably a larger area than has been walked in almost any other English parish as part of an archaeological field survey. All of the stints are numbered and the finds bagged. They are then washed and dried at a cottage provided by the estate and eventually transported to King Alfred's College, Winchester where Christopher Gerrard and his students and colleagues sort and analyse all the material. As well as the record of the fields, lines and stints, and the fieldwalkers involved, we also record the field conditions, wet, dry, ploughed and so on, weather conditions, sunny, dull or raining and details of slope, lynchets and drainage. We also feed in all that we know of the former history of the field from the early maps and other documents. All of these data have enabled Chris Gerrard, probably for the first time in a field survey project, to allow for the ideosynchrasies and foibles of individual fieldwalkers, as well as the effects of weather and field conditions, on the numbers and types of artefacts retrieved from the fields. The results of this work are still being analysed but it is clear that a number of new archaeological sites have been identified and other areas of activity defined. A number of fields have high densities of 208 flints, there are several new Romano-British settlements, and there are extensive scatters of medieval pottery and post-medieval finds all probably representing phases and areas of manuring. Of greatest interest are the finds of late Saxon material which might just hint at the possibility of pre-village settlements in the fields, under the common field areas and away from the present village (see below). The Early Church Site Rather unusually the church was moved in Shapwick in the fourteenth century. The old church site has never been forgotten by local people; there are eroded earthworks in the field and the field name on the tithe map in 1839 was 'old church'. The church may have been established there in the eighth century when the estate was granted to Glastonbury abbey, and it was probably the minster church for many settlements along the Polden Hills in the middle ages. In 1327 the abbey petitioned the bishop of Bath and Wells for permission to move the church and in 13 3 1 the new church in the village was consecrated by the bishop. The old church was dedicated to St Andrew (as is the cathedral at Wells) and the new one to St Mary (the dedication of Glastonbury Abbey). It was decided early on in the life of the project to use the church site to test out various techniques and ideas. It was used in 1990 to test out geophysical equipment to see if satisfactory results could be obtained from a site on this geology. Over the years many different techniques have been carried out in this field. Surveys of the slight earthworks have been conducted by the Royal Commission on Historic Monuments for England and by James Bond. These show that the church stood on a mound adjacent to a spring, all within a large embanked and ditched enclosure, with ridge and furrow of the common field system beyond. In 1993 six trenches were cut across these earthworks partly to confirm the site of the church but also to clarify the condition of the buried remains which are ploughed occasionally. These trenches hinted at Roman occupation in the vicinity, located the church and the probable priests house and the complex sequence ofboundary construction with several pre-medieval phases. Subsequent field walking across the field revealed prehistoric flints, Roman pottery and building material suggesting a villa, late Saxon and early medieval pottery perhaps indicating a Saxo-Norman manor and extensive later pottery deposits. In early 1996 a further geophysical survey by English Heritage produced an extraordinary pattern of anomalies which have not been investigated further so far but are very The Shapwick Project, Somerset, England difficult to explain with what we know of the site (Fig. 8). The Early Settlement Sites If we are to demonstrate that the settlement pattern in the parish changed and that there was a scatter of farms or hamlets out in the parish before the village was created, we need to find the post-Roman/ pre-medieval sites out in the fields. This is a difficult task for two main reasons; firstly so many of the fields are almost permanently pasture, and secondly, and much more problematically, for the critical period 400 to 900ad no pottery was used in this area - it is aceramic - and hence it is not available to be found in the ploughsoil in fieldwalking campaigns. This is a problem in many parts of the world where we have a 'dark age' - how do you recognise archaeological sites when there is no pottery on them? When this is combined with a society which used few inorganic objects, built in wood , apparently without postholes, rarely used stone and, in general, left few traces for archaeologists to find, it makes it very difficult to recognise human activity. But absence of evidence is not necessarily evidence of absence. In this project we have adopted a strategy which involves both geophysical and geochemical analysis of field samples. The basis for this is that it may be possible to detect anomalies in the ground by geophysical methods, where there are not necessarily any surface finds, while at the same time locating concentrations of phosphates and heavy metals which may indicate residues of human and animal activity. This builds on research by Mike Martin over many years in which he has shown that heavy metals such as lead, zinc, copper, cadmium and so on do concentrate in the soil where there is human occupation. Our ideas about the way heavy metals move through the food chain and build up in the soil are still evolving; it is hoped that further investigation of the course of heavy metals through the food chain will form the basis of postgraduate research beginning in 1997 or 1998. This approach has been applied to several fields where we were clear that there were archaeological sites of different periods and where we could see if there were concentrations or not. The field with the church site shows up clearly (fig 9) as does a Romano-British site at 'Abchester' and also the fields around the manor house. This has given us the confidence to try out the method on suspected sites of the period 400-900ad for which we have no archaeological evidence. Two such areas have so far been identified, a field called 'Sladwick' north of the village, and fields called 'Henry' west of the village. The latter was called 'Ennery' and 'Enworthy' in earlier times and so both are 'habitative' names of the type mentioned above with 'wick' and 'worth' endings indicating settlement sites. At 'Henry' nothing beyond a few pieces of tenth century pottery has been found in extensive fieldwalking across the area and the same is true for geophysical scanning over all the fields that have ever been in the area of 'henry' field names. A survey near Manor Farm however located a small area of anomalies which might be a settlement site. Soil sampling for heavy metals will be carried out over this area in 1997. More progress has been made with 'Sladwick' (fig 10). Here the field name survived to be recorded in medieval documents and on maps of the eighteenth and nineteenth centuries. The field is in a low lying area next to a low ridge, though there is slight rise at one end of it. Fieldwalking located a few RomanoBritish and late Saxon sherds but little else - though this might suggest an aceramic phase between Roman and late Saxon. Geophysical and geochemical surveys were carried out independently using the same twenty metre grid; they both produced features over a low mound at the western end of the field- the geophysics a set of rectilinear lines and blocks and the soil sampling a concentration of lead. Excavation in 1996 located two substantial walls built in a clear Romano-British style and forty of so Roman sherds in an area ten metres square. What does this represent? While at first sight this evidence might suggest a late Roman site there are really far too few Roman finds from it to make this a satisfactory explanation. But if the site was occupied in the fifth century, walls may still have been built in Roman fashion even if pottery was in diminishing supply. We have nothing as yet to compare with this site with as there are no lowly farmstead settlement sites of this date known in the west of England. It is clearly important to identify more 'habitative' names from the documentary sources so that this type of work can be carried out in other parts of the parish. Even if we can only locate names to the nearest hundred metres, this will be enough to bring in geophysical and geochemical surveys to locate any anomalies, as well as the use of conventional fieldwork techniques, to locate any sites. Ideally the whole parish should be looked at in this way so that a complete view, and one that was objective rather than one dictated by what we think we know already, could be obtained. Nevertheless we are encouraged by what we have called the 'Aston-Martin' method(!) and think that it could have wide application in finding the less obvious archaeological sites. 209 M. A. Aston Conclusions and Future Work In the remaining few years of the project fieldwork will continue in order to maximise the area covered and there will be further documentary and topographical analysis of the village plan and the preenclosure fields. We will have learned a great deal about how people used the landscape of the parish in the past and how the pattern of boundaries and properties evolved. It is perhaps too early to be categorical about the evidence but it does look as if the village was built up in some regulated fashion, probably from the tenth century, and that it replaced a scatter not so much of farmsteads but of a few hamlets. So far it looks as if these have their origins in the late Roman period. We assume they operated some sort of non-common field system, probably of infieldoutfield type, which may have been situated in the remnants of the pre-existing pattern of prehistoric or Romano-British fields. Two figures are shown here which attempt to show the evolution of the settlement pattern and the main landscape developments in the parish. Fig 11, the less conventional version, tries to show the origin, development and present form of the settlements in the parish- circles are farmsteads, thin hatched bands are hamlets and thick hatched bands are villages white symbols are desertions. The information needed to compile this diagram is very difficult to corn- pile. Just when were particular settlements first built and what form were they in? The question marks show the degree of uncertainty in almost every case. Fig 12 shows in a much more straight-forward way the current idea (August 1996) of how we think the parish developed from the (known) Iron Age sites to the Romano-British settlements located so far to the situation in the post-Roman and Anglo-Saxon periods. An attempt could also now be made to show the sites in the earlier prehistoric periods. The development of the village and its fields is shown in two stages, the earlier with small common fields and a lot of woodland and marsh and the later with full extent of common fields- the extent of woodland and marsh is not yet certain for the earlier periods. If the village and its common field system were developed in or by the tenth century, as seems likely, it is at least possible that this was the result of a deliberate policy decision by Glastonbury Abbey which owned the estate, as part of the reestablishment of regular monastic life in the old monastery. We might expect Dunstan, the new abbot, to not only refound the monastery but also look critically at the estates which supported it, like Shapwick, to see if they should be altered to make them more efficient at supporting the monks. From such policy decisions made such a long time ago did the appearance of much of the familiar English scene develop. M. A. Aston University of Bristol 8-10 Berkeley Square, Clifton Bristol BSS 1HH England 210 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Neil Price The Gamla Uppsala Project: Rescue and Research in an Early Medieval Ritual Landscape Gamla (i.e. 'Old') Uppsala, located a few kilometres north of the modem city of Uppsala in the province ofUppland, is well-known as one of Sweden's most spectacular monuments from the early medieval period. The focus of the site as it appears today is a 760 m-long gravel ridge which supports a cemetery of several hundred burial mounds dating from the 5th-1Oth centuries AD, at the eastern end of which are three huge barrows known as the Mounds of the Kings. These burial mounds, up to 10 m high and 75 m in diameter, were partly excavated in the last century and shown to be of Migration-period date. Adjacent to the cemetery ridge and the Kings' Mounds is another large mound, its flat top perhaps giving rise to the local tradition that it was the site of a royal Thing (assembly place); test excavations have tentatively suggested that the mound is also a burial monument. North of the so-called Thing Mound is the medieval church of Gamla Uppsala, once the seat of the archbishopric before its move to the site of the modem city ofUppsala. On artificial terraces built up around and near the church, excavations have revealed a number of massive longhouses dating from the Vendel and Viking periods. The possibility of a similar structure under the church is likely and widely accepted despite scant archaeological evidence, the stone foundations destroying most earlier deposits. Further excavations in the fields north of the church terraces have revealed more Vendel-period buildings, a road and massive boundary divisions spread over a large area. The surrounding landscape is a veritable necropolis, with up to 3000 Iron Age graves suggested through a combination of archaeology and reliable post-medieval maps to lie within a 2 km radius of the Kings' Mounds. The same area is known through excavation to have been the site of settlements spanning a similar date range. (The archaeology of the site is summarised in Lindqvist 1936, Damell 1990, Duczko 1993 & 1996, and Duczko et al. 1994.) Uppsala is mentioned in several medieval sources describing the earlier history of the pagan North, where the site is firmly linked to the Ynglinga dynasty ofkings. Named in numerous texts including Snorri Sturlusson's Heimskringla, the Ynglingatal quoted therein, and Beowulf amongst others, the Ynglingas are said to have ruled in eastern central Sweden and eastern Norway during the later Iron Age, forging these areas into powerful petty kingdoms (cf. Norr in Duczko 1996). Several of the kings are described as either dying or being buried at Uppsala, and the Kings' Mounds have been linked to a number of the dynasty's members including several of those mentioned in Beowulf. The longhouses on the church terraces have been interpreted as remains of the royal halls, surrounded by the lodgings of the king's retainers. Gamla Uppsala is perhaps best known as the alleged site of the famous pagan temple described by the cleric Adam of Bremen, writing in the 1070s. Purportedly quoting an eye-witness, he tells of a great building girded with golden chains, its main hall housing the images of the gods Odin, Thor and Frey. Men and animals were sacrificed to the deities at great feasts, their bodies hung in a sacred grove (cf. Alkarp 1993 and references therein for a recent discussion of the temple and its rites). It has been suggested that it is this building which once stood on the site later occupied by the medieval church. After years of conjectural reconstructions of what may have been the most important Norse temple in Scandinavia, it has recently been suggested that we should instead think more of a 'temple-hall', perhaps differing little visually from the other longhouses around the royal terraces (cf. Nordahl 1996). Combining all this data, much of it derived from Uppsala University's research excavations (see papers in Duczko 1993 & 1996) and recent rescue work (e.g. Karlenby 1994, Vinberg 1995, Roslund-Forenius 1996) a new picture has emerged of Gamla Uppsala over the last decade of scholarship on the site. We see a royal seat established in the Migration period and expanding during the Vendel period, focusing on massive kingly burials surrounded by a cemetery about which we as yet know very little. The site's 211 N. Price focus seems to have been a compact area of impressive buildings raised on terraces above their surroundings and looking out over other structures, roads and enclosures. In the Viking period there seems to have been a shift in the monument's meaning or purpose, with an increased emphasis on ritual at the expense perhaps of Gamla Uppsala 's royal connections (the king's interests may have shifted to Adelso/Birka and later to Sigtuna). This is the period of the temple and the aggressive paganism which so dismayed the German clergy. The surrounding landscape seems to have been increasingly dominated by cemeteries at this time. It is this ritual function of the site which was usurped by the Christian missionaries as the new religion became firmly established in eastern Scandinavia during the early 12th century, when the temple was replaced by the seat of Sweden's first archbishop. Gamla Uppsala in its landscape We now know that Gamla Uppsala is one of a number of similar sites throughout northern Europe -all centres of political and religious power during a period of emerging kingdoms and a transition from tribal - to state-based polities. Complexes of large burial mounds, often with associated buildings of various kinds, are a recurring feature of these monuments. In addition to Gamla Uppsala itself, Sweden possesses another such site at Hogom in the northern province of Medelpad. Similar patterns are seen at the Danish sites ofLejre and Jelling, and in Norway at Borre, A valdsnes and Bertnem. All these localities are linked by a strong ritual cohesion, apparently serving as the seats of local dynasties who seem to have had a mutually supporting idea of the material nature of the power they wished to create. In the British Isles these sites are most closely paralleled at Sutton Hoo, but we may also consider the Scottish centres at Strathearn and Dumbarton, and Tara and Armagh in Ireland, amongst several others. Almost identical "places of power" can be observed on mainland Europe too, especially among the Slav populations in Poland and the Baltic littoral. This phenomenon has been intensively studied in recent years, and been the subject of many conferences and books. The general consensus has concerned the study of these monuments in the context of their landscapes, those "sites" which we often perceive individually instead forming a cohesive whole representing the material expression of values held by the society in question. These values are naturally not static, and- in Tim Darvill 's words ( 1996)- what may be called the "cognitive map" of the landscape 212 "is always mutating, being contested, and redefining itself as a result of changing relationships between people and the way they see the world around them". In a similar manner, Martin Carver has compared the signals encoded in the monuments to a form of argument, in which different voices compete for a say in how a people should live and a society should function (e.g. Carver 1993 ). These approaches to landscape interpretation have been most developed within Neolithic and Bronze Age archaeology, amongst others by Richard Bradley who has himself written on Gamla Uppsala (1993, 94-95). He has argued that the first major monuments on the site that we know of- the Kings' Moundsmay themselves have been constructed in imitation or acknowledgement of an earlier tradition, such as the massive Bronze Age barrows which dot the Uppland countryside at sites such as Hagadalen. Bradley suggests that what were created as individual memorials to dead kings became instead a collective monument during the Viking period; thus he emphasises that in Adam ofBremen's description the huge earthworks are treated as a single phenomenon, providing a monumental setting for the religious festivities that take place at the temple. Through a sequence of successive interpretation and re-interpretation the landscape has maintained a ritual meaning but been transformed from a place of individual commemoration to a context for public assembly- the occasion being perhaps the fusion of a unified pagan religion with the growing power of the Uppsala kings. This ritual landscape at Gamla Uppsala, which has slowly been altered, adapted and re-created over millennia by the inhabitants of the area, is about to be changed again in perhaps the most dramatic manner in its history so far. A new rail development is to be constructed through the site, passing barely 200 m from the mounds and consuming vast areas of the surrounding ground. Constituting both an unprecedented destruction of the archaeology and simultaneously an opportunity to examine the hinterland of such a place in a way never previously possible, the East Coast Railway development is a major challenge to Sweden's archaeologists. This paper- and the talk for which it is a pre-circulated template presents a review of the ongoing rescue project set up in response to the proposed railway. The East Coast Railway development at Gamla Uppsala and its archaeological impact The development centres on a re-routing of the existing East Coast Railway which now runs northward from Uppsala city through Gamla Uppsala, The Gamla Uppsala Project: Rescue and Research in an Early Medieval Ritual Landscape passing approximately 40 m east of the Thing Mound. The track is being altered in conjunction with a wider programme of infrastructure development in central Sweden and is related in particular to a complete rebuilding of the E4 motorway, the new route of which will pass approximately 2 km east of Gamla Uppsala. From the city limits, the new railway is planned to run to the west of the existing track, through open fields. At the centre section of this new route, where the railway passes closest to the main monuments, it is proposed that the new line will be routed underground. The track will be carried through a 600 m tunnel running beneath the modern village of Gamla Uppsala and emerging again into fields to the north. The tunnel will not be of bore construction, but will be built of buried concrete sections, a technique necessitating a development corridor somewhat wider than the tunnel itself. From the tunnel mouth the line swings east to pass through a number of isolated farmsteads and their surrounding land before crossing the Samnan river (a tributary of the larger Fyris river) and then rejoining the existing East Coast Railway. Much of the above-ground track will be sunken into the landscape by means of substantial cuttings. The total length of the new track sections is approximately 3 km, from the outskirts of the city to just north of the Samnan. The width of the development corridor varies considerably, from a minimum of 50 m to almost 200 m at its widest point where the line includes areas for contractors' cabins, truck servicing and so on. Additional work is planned at points where the railway will cross existing roads and paths, resulting in several projections from the main development corridor. Around 14 hectares of currently undeveloped, open fields within a 1.5 km radius of the Kings' Mounds and the church/terrace complex will disappear within the railway corridor. A further 6 hectares ofland currently under roads, car parks, playgrounds and other modern features will also be included in the development (as discussed below, many of these modern features seal intact archaeological deposits beneath them). As will be readily understood from the above description, the archaeological impact of the rail extension is potentially massive. The human scale of the 14 ha of open fields can be better understood when one considers that this represents an area equivalent to 20 football pitches. When the draft development plans were first unveiled, it could be seen that the existence of archaeological features was possible over this entire area. In some parts the presence of sub-surface remains was either known through earlier excavation work or suggested on the basis of aerial photos and other forms of non-destructive reconnaissance undertaken in recent years.lt was clear by the early 1990s that an archaeological response of some gravity would be required. Archaeological work for the East Coast Railway 1990-95 The first stage of this response came in 1990 itself when the regional unit of the Central Board for National Antiquities (Riksantikvarieiimbetet UV Uppsala) carried out a preliminary desk-top assessment of the route through Gamla Uppsala. Although necessarily relatively superficial as the track specifications were then far from finalised, the report made plain the need for full archaeological investigations if the railway was to proceed (Gothberg 1990). Little happened for the next three years until 1994, when it seemed (mistakenly, as it later turned out) that the development was about to begin. At this time a draft design for a research-driven rescue project at Gamla Uppsala was published by the Institute of Archaeology at Uppsala University together with the consultancy practice Arkeologikonsult AB, with a contribution by Professor Martin Carver of York University (Duczko et al. 1994, Price 1994). The report was disseminated for comment and shortly afterwards the same organisations commissioned two flights over the development area to make an aerial photographic survey of the site. The research design had been prepared in the context of what the authors felt to be a widespread indifference within the archaeological community both to the importance of the monument and the potentially drastic effects that the development would have. Over the following eighteen months the entire project was plagued by increasingly bitter controversy within Swedish archaeology on the one hand, and within the general social debate on the viability and/or desirability of costly infrastructure projects on the other. Although centring principally on the political aspects of the railway itself- and in particular on the associated E4 motorway development - the administration of the archaeological response and the selection of appropriate contractors for the task were also hotly contested subjects. A discussion of these debates is inappropriate to the present paper, but their occurrence played a significant part in shaping the current project. Due perhaps to the increased urgency of the project following the development's shift to an active status in early 1996, this unfortunate situation seems to be for the most part resolved now, allowing archaeologists and other specialists from a wide range of 213 N. Price institutions to co-operate in the most effective fashion for the well-being of the monument. Beginning with the results of an Environmental Impact Assessment carried out in separate sections by the different institutions early last year (cf. Price 1996 for the University consortium's contribution), a major collaborative project has now been established which will hopefully set the tone for all the future work on the site. The project structure The new project is being run jointly by the Institute of Archaeology at Uppsala University and the Uppsala field unit of the Central Board for National Antiquities. Overall responsibility for the project therefore rests with Professor Bo Gdislund for the University and Sverker Soderberg for the Central Board, with the University further providing administrative and academic co-ordination in the form of Dr. Kent Andersson and Docent Wladyslaw Duczko respectively. The main project planning and fieldwork is undertaken by two separate units linked to the institutional partners: the author is director of the University team while that of the Central Board is jointly directed by Johan Anund and Hans Gothberg. For the evaluations, the digging staff of the University team were sub-contracted from the Arkeologikonsult consultancy. Like any such work in Sweden, the entire project is answerable to the County Archaeologist (a local government officer based in this instance at the county authority offices in Uppsala) who is responsible for the higher administration of the region's archaeology. As this is above all a rescue project, the above structure must naturally link with the developers themselves, who under Swedish law are obliged to fund all necessary archaeological work as determined by the County Archaeologist. For Gamla Uppsala the Swedish state railways Banverket have afforded the archaeologists every assistance over and above their legal obligations, and have been enthusiastic partners in shaping a project commensurate to the monument's importance and the degree of impact that the development will have on the archaeological remains. For the purposes of the evaluation the development area has been divided into separate sections for the two units; a reallocation of areas will probably be made before the main excavations if the final phase of the project goes ahead. The Central Board team excavated in and around the buildings of the modern village of Gamla Uppsala, at the heart of the site where the development corridor passes closest to the Kings' Mounds. Although small in area, this part of the site was judged to have the highest potential for 214 extensive archaeological deposits. The same team were to continue working northwards along the development corridor into an area of fields where prehistoric building remains had been contacted in earlier keyhole excavations, carried out for various purposes over the last decade or so (see below). The University team were allocated a larger area, approximately 70% of the development corridor, which was expected to produce more diffuse traces of settlement. This consisted of the entire corridor south of the present village, and also the area at the northern end of the development on both banks of the Samnan river. Objectives and methods Objectives For the purposes of the evaluations, part of the task of which would be to provide sufficient information for the production of a full research design for the final excavations, our objectives were relatively simple. The specific intention was to delimit the remains contacted within the development corridor and to obtain a preliminary overview of their character and dating. A deposit model should be produced in combination with a qualitative assessment of the strata, with an attempt made to predict areas of concentrated archaeology. However, in anticipation of the next phase of the project- the main excavations- a set of more general principles and objectives should be borne in mind, especially in relation to the wider landscape perspective emphasised above. We should be mindful of Martin Carver's memorable words from the draft research design of 1994, "all the landscape visible from the mounds must be regarded as precious" (in Duczko et al. 1994, 45). A clear focus must be placed on the chronological and cognitive development of the Gamla Uppsala landscape, with a natural weight falling upon the later Iron Age and the transformations of the early Middle Ages. Equally interesting is the early prehistory of the site: what was Gamla Uppsala before the Kings' Mounds were built? How was the land used in the Roman Iron Age, and even in the later Bronze Age when the Uppsala plain first drained? A particular feature must clearly be an assessment of the role played by Gamla Uppsala in the formation of the Swedish state, in so far as these questions can be answered within the constraints of a developmentoriented project. It may prove to be possible to expand the parameters of the project beyond the rail corridor, as in fact the developers have already suggested, for example through an associated training and research excavation for university students. The Gamla Uppsala Project: Rescue and Research in an Early Medieval Ritual Landscape All of this work should be published and disseminated to the highest international standards, and archived in such a fashion as to provide a lasting intellectual resource not only for scholars but also for the interested public. The final research design is as yet in the future, when these and other questions will be taken up and debated in full. later location. All recording was computer-driven, with EDM total station theodolites and portable computers used for fieldwork, later transferred to inhouse CAD and GIS workstations. All reports will be produced internally using DTP technology. The 1996 evaluations: preliminary results Methods Early in the project planning we experienced a small disappointment related to the use of nondestructive survey methods. Although recommended to different degrees by both the University and the Central Board in the research design, the use of geophysical and geo-chemical reconnaissance during the evaluations was deemed by the office of the County Archaeologist to be unlikely to contribute any additional information. The project was therefore instructed that the main focus of the evaluation strategy should be placed on the excavation of test trenches. The use of field-walking was similarly rejected after discussion, and in the event the only form of nondestructive survey that could be viably included within the project framework was the aerial photographic survey that had been completed in August 1994. The results of this survey could at least be used to assist in the location of test trenches in order to check the status of those features visible from the air. In the absence of such surveys, the field strategy consisted simply of excavating large numbers of short trenches approximately 2 m in width, arranged as appropriate in relation to the topography and the presence of known or contacted remains. Trenches were also sited in accordance with the recommendations of the academic advisor, Wladyslaw Duczko, tapping his extensive knowledge of the site and its environs. Trenches in the village area were more constrained by the modem buildings, and their configurations were necessarily more esoteric than those excavated in the open fields. A further aid to the test excavations was the use of CAD- and GIS-programs to overlay rectified versions of the earliest maps of the area onto modem plots of the development corridor, enabling us to predict with great accuracy the locations of buildings and graves. During the course of final preparations for the main excavations, as many aerial photographs as the budget permits will also be plotted in the same way. A selection of features was fully excavated and recorded in order to gain a representative sample of dateable material, the remainder being drawn in plan and given a basic description before being carefully reburied under a plastic shield and marked for ease of The evaluations were carried out in November 1996 at the onset of winter. Normally, little or no archaeological work is carried out in central Sweden so late in the year, due to the bad weather conditions of snow, sub-zero temperatures, poor light and fog. Although initially strongly resisted by both archaeological teams, it was soon clear that the schedule set by the developers and upheld by the County Archaeologist left no room for further delays. Accordingly the project had little choice but to proceed with the evaluations despite temperatures of up to ten degrees below freezing. In the event we believe that minimal information was lost due to the weather (with the exception of a small area in which the features were destroyed by frost action), but we certainly hope for a warmer start to the final phase of excavations! The following summary of the evaluation results is a strictly interim statement, and represents the situation in early January 1997. At the time of writing, none of the scientific analyses have been completed and no radiocarbon dates have been received from the laboratory. As discussed below, approximately one third of the test excavations are still to be completed, and will be carried out during the spring. In the absence of publishable plans at this stage, it will obviously be difficult for readers unfamiliar with the site to orientate themselves through the following notes, but the preliminary results printed here are primarily intended to serve as an interim reference for specialists interested in the site. Full visual materials will be presented with the main lecture at the conference. Information on the results of the test trenching carried out by the Central Board is presented here by courtesy of Riksantikvarieambetet UV-Uppsala. The first kilometre of the development as it stretches north from the limits of Uppsala city has been found to be empty of archaeological remains, following excavations by the University team. No monuments were previously recorded from this area, which is known to have been open fields as far back as the late 16th century- the date of our first reliable records. There is no particular reason to suggest that any form of prehistoric settlement or non-agricultural activity was ever located there, but the evaluations have in any case confirmed that modem ploughing in the area 215 N. Price has penetrated so deeply (up to 0.5 m in some places) as to have effectively destroyed any archaeology that once existed there. A Migration-period cemetery? By contrast to the southern extremity of the development corridor, archaeological remains were located in most of the other areas investigated. Our caution to "expect the unexpected" at Gamla Uppsala proved sensible, as a wealth of previously unknown features and sites were discovered. The first of these were found by the University team under the direction of the author during test trenching at the southern edge ofLilla Gardet, the level clay plain which borders the main cemetery ridge along its eastern side. No archaeological investigations had ever been carried out here previously, but programmes of environmental sampling combined with the evidence of early maps have shown this area to have been quite marshy in the early medieval period, its centre perhaps even occasionally inundated. The plain is a dramatic landscape, bounded as it is by the curve of the cemetery ridge to the west - Lilla Gardet dominates the eastward view from the Kings' Mounds and the larger grave-fieldand we had hoped to find some evidence of human activity on the dryer ground at the plain's eastern limit. On the aerial survey of 1994 it was noted that a number of circular features about 1-2 m across were faintly visible in the crop in the south-eastern corner of the plain, diametrically "opposite" the Kings' Mounds and approximately 700 m away from them. Our hopes for new information on the archaeology of Lilla Gardet were realised with the discovery of a cemetery at this part of the field. The burials consisted of circular stone-settings about 1.5 m in diameter, covering pits containing cremated bone and ash in pottery vessels. The ceramics could not be dated more precisely than to the Iron Age, and at time of writing we have not yet received the results of accelerator datings on charcoal from the graves. However, on the basis of their form and typology it is likely that the burials date from the Migration period: if this estimate is correct, this will be one of the first monuments contemporary with the Kings' Mounds to have been found at Gamla Uppsala. Only· a small number of graves were contacted and excavated in the evaluations, but to judge from at least one feature it is possible that the cemetery had its origins at an even earlier period: a sturdily stone-packed pit found in the centre of the surviving grave-field appears to be the foundation for a bautasten - a kind of standing stone commonly erected as a memorial over graves from the Roman Iron Age. 216 The burials had been badly damaged by ploughing, and in addition to the intact graves found in the test trenches we found traces of further stone-settings that had been almost totally destroyed. Although all the burials were found in locations which corresponded to the pale circles seen from the air, confirming that these represented graves, interestingly there were several such circles clearly visible on the ground which proved to have no stone-settings beneath. It is possible that there were originally other graves at these locations, but that after ploughing only the faint difference in crop growth marked where they once lay. A deposit model has been created for the plough-soil in the Lilla Gardet excavations, showing clearly that the surviving graves lie in an area where -probably by chance - the plough has penetrated up to 0.2 m shallower than in the rest of the field. The distribution of the stone settings found in the evaluations respects the dimensions of this pocket of shallower ploughing so precisely that there is no doubt that the cemetery must once have extended further than its present limits. How many graves have been destroyed by several centuries of cultivation in the field is impossible to estimate, but judging from the faint indications visible in the aerial photos it is possible that the cemetery was originally laid out along the eastern edge of Lilla Gardet for some considerable distance to the north. An area of almost 11,000 m 2 has been recommended for a full excavation to recover what remains of the cemetery. A settlement by the Kings' Mounds The most spectacular results from the evaluations were found by the team from the Central Board, working under the direction of Johan Anund in the centre of the modern village. During the Middle Ages the village at Gamla Uppsala was one of the largest in the whole of Uppland province, and its location can be precisely plotted from maps of 1640 and 1710, made when the late Medieval farms were still standing (cf. Dahlback et al. 1984, Rahmqvist 1986). From these earliest cartographic sources and documents such as tax records we know that more than 12 farmsteads stood on the site in the early 17th century. Two of these farms were thought to lie directly in the path of the rail development, and it was speculated that these late- and post-medieval buildings - together with more recent construction in the centre of a modern community- could have destroyed any prehistoric deposits in the area. In the event, while minimal traces of the medieval buildings were found in the form of layers of stone foundations, the few trenches that were able to be excavated around the The Gamla Uppsala Project: Rescue and Research in an Early Medieval Ritual Landscape lawns and yards of the modern buildings proved to be packed with prehistoric features. The northern section of the modern village- 250 m east of the Thing Mound - appears to have been particularly densely settled, with postholes so tightly distributed that only a few areas of clear ground could be seen between them. For the most part these postholes are cut directly into the natural sub-soil, but pockets of occupation deposits survive up to 0.3 m thick; more postholes have been found beneath these layers in the places where selective test pits were excavated through them. A marked decrease in the density of features further to the east may suggest a boundary of some kind, perhaps connected to a known Viking-Age cemetery which begins here, or alternatively linked to fence-lines reconstructed around a Roman Iron Age longhouse found near this spot in the late 1980s. All the new features are undated as yet, but ceramics of Viking-Age type, an early medieval bone pin and fragments of a medieval comb were all found in the test trenches. Slightly further away, approximately 250 m southeast of the Kings' Mounds, the trenching revealed large areas of surviving stratified occupation deposits up to 0.3 m thick. Almost the same density of postholes was found here too, together with the remains of stone-set graves. The chronological relationship of the burials and posthole buildings is unclear. The character of the central settlement seems to change about 100 m to the south of these remains (about 350 m south-east of the Kings' Mounds), with the posthole structures giving way to a series of sunkenfeatured-buildings (Grubenhiiuser). Some postholes continue here, but unlike the other areas there are no hearths. As for the northern area, no definite relative chronology can yet be established for these features. Test excavations at the edge of modern features which cannot be removed until the main phase of work- such as the main road through Gamla Uppsala village - indicate that these constructions seal prehistoric remains beneath them, and have not destroyed them as first feared. The finds of settlement traces in the modern village are not unexpected, but the sheer quantity of building remains was rather overwhelming. There is no doubt that just the test trenching alone has completely changed our perspective on the whole monument complex at Gamla Uppsala. The density of the remains clearly indicates that structures were re-built many times on almost the same spot, suggesting some form of pressure to respect established boundaries. At this stage of the project we can only speculate as to the importance of this in relation to the known presence of the royal estate at Gamla Uppsala, but it is tempting to make the obvious link. Although a final figure has not yet been drawn up owing to a delay in completing the evaluations in the fields north of the village (see below), it is already apparent that at least 23,000 m 2 will probably be recommended for full excavation in the central area. This large site is known to support several phases of activity of a varied nature, including different forms of structures and perhaps several graves, and furthermore to preserve considerable areas of intact stratified deposits. Given the likelihood of contacting further structures in the field area, it has been estimated that the final excavation area in this most important part of the site may be as large as 60,000 m2, including the central area mentioned above. (These estimates include the eventual excavation of areas which are currently inaccessible under roads, car parks etc.). As the remains excavated by the Central Board extend to the southernmost limit of their area, adjoining the existing railway which the new track will replace, the University has also recommended that a number of trenches be opened up within a 6,000 m2 area at the northernmost limit of their section. Although only a few postholes were found here during the evaluations on the opposite side of the railway to the Central Board trenches, it is clear that the settlement once extended at least this far. Even though ploughing has obviously removed most of the remains, the importance of the newly-discovered settlement is held to justify a further intervention in the fields. A settlement by the "King's ford" At the third site to have been located in the evaluations, the University team under the author's direction found indications of post-built structures by the Samnan river at a point approximately 800 m north-east of the royal terraces and the church. The buildings were found by a crossing place traditionally known as the Kungsvad (King's Ford), marked on the earliest maps with a road leading from the river to the church area (see Graslund 1993). A small settlement nearby bears the name Kungsgarden (King's Farm), but this is thought to be a modem name coined in reference to the fact that the crown had once owned the land; the farm itself does not appear on the 17th-century maps. Scatters of postholes at this site, at which no archaeological remains have been found previously, indicate clusters of buildings on a level plateau above the slope down to the river bank and the ford. No dateable artefacts were found, but several charcoal samples have been submitted for radiocarbon analysis; the results of the tests are expected towards 217 N. Price the end of January 1997. The site is one of several river crossings along this stretch of the Samnan, but the early link to the royal terrace area presents an intriguing possibility that the buildings found by the University represent some form of installation by a recognised royal travel route. Such a perception of prescribed movement zones linked to status has interesting implications for the way in which the territory around Gamla Uppsala was imagined by its inhabitants. The possibility that the remains formed part of a large settlement gains support from findings on the opposite bank of the Samnan river. From a point beginning a few metres outside the development corridor - and thus uninvestigated for the project the aerial photographic survey of 1994 picked up crop-marks which appear to show a large inhumation cemetery stretching for approximately 150 m eastwards along the heights above the northern riverbank. Grave-fields of this apparent type are most generally found within the Roman Iron Age (the cemetery is unlikely to be an unregistered Christian burial place, as no church is known from this spot), and it may be that the settlement by the ford is of greater antiquity than the royal connections of the Migration period. An area of nearly 7,000 m2 has been recommended for full excavation at the site of the Samnan settlement. Evaluations in Spring 1997 Due to a delay in arranging compulsory purchase compensation between the developers and one of the landowners, approximately one third of the development corridor was inaccessible for evaluation work in the winter of 1996. This incorporated a small section of the University area and a large stretch within the Central Board's area north of the village. Limited rescue excavations have already been carried out in this latter zone during the late 1980s and early 90s, in connection with cable-laying and the construction of a cycle track. On the basis of this earlier work we know that this area contains substantial post-built structures dating to the later Iron Age, and it seems likely that the settlement in the central area of the village extends several hundred metres north under the modem fields. It is not impossible that these features may in fact extend as far as the settlement found by the University team next to the ford, thus presenting a picture of a dense landscape of buildings covering the Uppsala plain for up to a kilometre around the royal terraces. The completion of the evaluation work in these areas is scheduled to start as 218 soon as the snow breaks in early spring 1997, and will thus have been completed by the date of the Medieval Europe conference. The future of the project At present (January 1997) the timetable for the project's future is uncertain. As noted above, the planning of the railway is linked to the controversial re-routing of the E4 motorway which will pass a few kilometres to the east ofGamla Uppsala, and the final go-ahead for this project has not yet been given. There seems little doubt, however, that the development will proceed as planned and archaeological preparations are being made accordingly. The preliminary start-date for construction of the railway has been set for 1999, before which the archaeological work will need to begin. At least two seasons of excavation are envisaged, working in step with the building contractors as successive areas of the site become available after roads and other features are diverted. As it appears now, the main phase of the Gamla Uppsala project is likely to involve one of the largest single~site rescue projects ever attempted in Europe, and unprecedented at a monument of this kind. We are intending to establish a series of workshops in association with the project, presenting and debating the findings at archaeological institutions and universities throughout Sweden. At a wider level, the present paper is one of a number of international presentations of the project at conferences and postgraduate seminars. The public are already involved in the Gamla Uppsala project through an exhibition on the evaluations at the provincial museum, and through public lectures; the site receives regular coverage in the media. For the main excavations we intend to operate regular site tours and talks, with the possibility of one or more popular science programmes for Swedish television. Plans are also being made for an illustrated popular book on the site. At a future date, we hope to arrange an international conference on issues relating to the site and its socio-political context in the early medieval world. It.is currently impossible to say how much work will have been carried out in addition to that presented above by the date of the conference, but the paper delivered at Medieval Europe will include illustrated summaries of the 1996 evaluations, the completion work carried out in spring 1997, a full review of progress up to October and more detailed information on the future of the project. The Gamla Uppsala Project: Rescue and Research in an Early Medieval Ritual Landscape References ALKARP M. 1993: Adam av Bremen och Gamla Uppsala, Dissertation in archaeology (CDUppsats), Uppsala University. BRADLEY R. 1993: Altering the earth: the origins of monuments in Britain and Continental Europe, Society of Antiquaries of Scotland, Edinburgh. CARVER M.O.H. 1993: Arguments in stone: archaeological research and the European town in the first millenium, Oxbow, Oxford. DAHLBACK G., FERM 0. & RAHMQVIST S. 1984: Det Medeltida Sverige 1:2. Tiundaland: Ulleraker, Vaksala, Uppsala stad, Riksantikvarieiimbetet, Stockholm. DAMELLD. (ed.) 1990: Gamla Uppsala:fornlamningar, utgriivningar, fynd, Uppsala University, stencil. DARVILL T. 1996: Billown, Isle of Man, Current Archaeology 150, 232-23 7. DUCZKO W. (ed.) 1993 & 1996: Arkeologi och miljogeologi i Gamla Uppsala. Vols. 1 & 2, Uppsala University Press, Uppsala. DUCZKO W., ELGH S., GRASLUND B. & PRICE N.S. 1994: Gamla Uppsala: ettforslag infor de arkeologiska undersokningarna pa Ostkustbanan och intilliggande undersokningsomraden, Arkeologikonsult AB, Upplands Vasby. GRASLUND B. 1993: Folkvandringstidens Uppsala, Uppland, 1993. G6THBERG H. 1990: Arkeologisk utredning: Uppland, Gamla Uppsala socken. Ostkustbanan dell: Uppsala - Storvreta, Riksantikvarieambetet, Uppsala. KARLENBY L. 1994: Ett tviirsnitt genom Gamla Uppsala socken, Riksantikvarieambetet, Uppsala. LINDQVIST S. 1936: Uppsala hogar och Ottars hogen, Kungliga Vitterhets Historie och Antikvitets Akademien, Stockholm. NORDAHL E. 1996: "templum quod Ubsola dicitur... " i arkeologisk belysning, Uppsala University Press, Uppsala. PRICE N.S. 1994: The Gamla Uppsala project proposal: supplement 1, Arkeologikonsult AB, Upplands Viisby. PRICE N.S. 1996: Gamla Uppsala Ostkustbananprojektet: Miljokonsekvensbeskrivning, Arkeo1ogikonsult AB, Upplands Vasby. RAHMQVIST S. 1986: Gamla Uppsa1a by- Upplands stOrsta. In: Fran 6stra Aros till Uppsala: Uppsala under tidig medeltid, Uppsala stads historia VII. ROSLUND-FORENIUS Y. 1996: Arkeologiskforundersokning, Gamla Uppsala, Gamla Uppsala socken, Uppland, Riksantikvarieambetet, Uppsala. VINBERG A. 1995: Ett hus fran iildre jiirnalder i Gamla Uppsala, Riksantikvarieambetet, Uppsala. Neil Price Institute of Archaeology University of Uppsala Sweden 219 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Sofia Andersson & Eva Svensson The local and regional arena of a Middle Age Swedish farm Introduction to a medieval arena The Slaamle farmstead, central to this paper, is located in the county of Varmland, north of lake Vanern in western Sweden. Today, the county of Varmland has about 300,000 inhabitants, with almost 140,000 of them in the Karlstad region (Karlstad is the largest city in the area). Almost 80 % of the county is forested and 8 % is arable. The forests are consequently of the greatest importance for the industry and people in the region. Fairly little is known of medieval Varmland. Written sources are scarce and most of our knowledge is based on archaeological and historical studies that was carried out during the 1980s and early 1990s. Older research focused on the construction and history of churches. The county belonged to the medieval diocese of Skara, which also included the province ofVastergotland. Similarities between churches can subsequently be traced in the two regions. But there are also a lot of differences, especially the building material. In Varmland, it was more common to built the churches in wood. There where no monasteries in the county, but the monastic system had great interest in the area, especially the salmon fisheries. Vanern and the streams and rivers in the region were rich in salmon. This was exploited by the crown and the monasteries from early times onwards. Several of the farms located along the rivers were subsequently owned by several Swedish monasteries and convents. Varmland also included a judicial district, with a provincial law. The code oflaw does not exist today, but we have indications that it resembled the law which has been preserved for the province ofVastergotland. There are remains of three medieval fortifications in Varmland: Amneholm, Edsholm and Saxholmen. The latter two have recently been the subject of archaeological investigations during the 1990s. These strongholds have been erected for different reasons, in different periods. There were no towns in the northern part of the lake Vanern area. The first town, Karlstad, was founded in 1564. Medieval markets and fairs are documented, most of them however in sources from the 16th and 17th centuries. Among these places we find: Tingvalla (today the city of Karlstad), Knusesund (today called Saffle) and Bro (today called Kristinehamn). Tingvalla is documented in written sources from the 13th century and it is said to be a central place and common meeting place for trade in Varmland. Archaeological investigations in the central parts of the city have not, however, revealed any agglomeration or settlement dating from before the 17th century. Knusesund and Bro are not mentioned in medieval records, but during the 16th century they are described as old market-places. Hence, there are source critical problems concerning early history of these places. That Varmland is a forested province is visible in the range of goods that are mentioned in the early period i.a. hides, timber and iron. The importance of the iron rose during the late medieval period. These products were also exchanged for corn, salt and fish. The arena of investigation: the Gunnarskog parish The Gunnarskog parish is characterised by the waterways running from north to south, with the settlement located close to the water. Most of the parish is covered with forest. The parish is mentioned for the first time in historical sources in 13 76 and was at that time an annex to the nearby Arvika parish (Ortnamnen i Varmlands lan 1942, 39). Next time the parish is mentioned is in the land register of 1503 (Samuelsson & Kallstenius 1939). Our knowledge of ancient monuments in Gunnarskog, as well as in the rest ofVarmland and Sweden, is based on the survey of ancient monuments carried out by The Central Board of National Antiquities. The first survey in Gunnarskog parish was done in the 1960s. Eleven ancient monuments were found at that time. The second survey was done by the authors in 1989 and the number increased to 800. A majority of the remains are settlements such as crofts, but also 221 S. Andersson & E. Svensson has changed and a lot of new remains have been found. For the moment there is a project called 'Forest and History' funded by the local department of forestry working in the area. The aim of the project is to find all kinds of deserted remains in the forest area and to inform the forest industry and thus preventing a destruction of ancient monuments. This has resulted in an increased awareness of these remains in the local community and it proves the importance of information and contacts during fieldwork. Archaeologists in the field work as a link between archaeology and the public. The Gunnarskog parish is considered as an area where permanent settlement started in the Middle Ages. This has been the traditional view in Swedish archaeology, that when there are no visible prehistoric remains such as graves, the settlement in the area is considered to have been established during the Middle Ages (Hyenstrand 1994, 8-9). This view has been challenged by research in the last decade. In Gunnarskog there are a number of prehistoric remains i.a settlement from the Stone Age, pitfalls for elk and charcoal pits connected with bloomery furnace iron production. Our research which will be presented below shows that the notion of a large scale colonisation during the medieval period in the central Swedish forested areas is too simplistic. In fact settlement can be traced back to prehistory. The excavations at Skramle Fig. 1. - Sweden with the location of Viirmland and the Gunnarskog parish. industrial remains, such as mills and sawmills are represented. A new category in the modern survey was the Stone Age settlements. This type of settlements is found predominantly by the shores and was not earlier known in the area. There was one system of pitfalls known after the first survey, a number increased to ten after the second survey. A number of single pitfalls were also found, both for elks and wolves. Dating of pitfalls for elks in Gunnarskog indicates that they were in use from at least 1400 B.C. up to modern times (Svensson 1995, 44). About 80 of the objects were protected by law. Today there are a lot more. One cannot take the result of the survey and equate it with the total number of ancient remains in the parish. After 1989 the picture 222 A problem concerning studies of medieval rural settlement in Sweden is the difficulty of locating settlement sites. The large-scale strategy applied on a national level at the registration of archaeological sites where farmsteads located on 17th century maps are believed to be situated on the same place as medieval settlements has not been successful when applied in regional and local conditions, i.e. confronted with reality. It has also been an apparent problem to locate medieval cultural layers and constructions during excavations of sites where medieval rural settlements were supposed to be located. A list of archaeological operations on such sites carried out during the years 1955-1992 shows that out of more than 400 sites touched by excavation only about half gave indications of medieval activities. Also worth noting is that when medieval remains occurred they were often fragmentary and difficult to date and understand (Ersgard & Hallans 1996, 47-111). In short there is a general lack ofknowledge of how medieval rural remains are located and recognised in the landscape, and it is our opinion that such a knowledge has to be built up from a local and regional level and be The local and regional arena of a Middle Age Swedish farm Fig. 2. - The parish of Gunnarskog. QLAKE EBD ~EAD/VllA!X f"ENTKH:O G 'V ~ H K V SYSTEM [f PITFALLS PREHSTORC GRAVE OCC~t-K:E [f SLAGG AID. WITH CHARCOAL PITS (LOCAL> MARKET REMAINS OF ROAD STMONj STot-l, RESTt.G PLACE N Tl£ t'fDfV AL RECORD 181 DESERTED FARMSTEAD/VllAGE; mN A!X 00 I'WE A!XS ~ AID. WITH PITFALLS 0 SHflNJ F CATCfM:NT CDNSTROCTlON OOT PITFAW F3 STRAY Ftll OOT STot-l AGE! s SETTl.EJ'IfNT SITE (•ST!l£ AGE"l T SO.APSTot;E rJJARRY 10 integrated in landscape analyses. During the years of excavation at the deserted farmstead of Skramle we have encountered, and are still battling with the problem of how to recognise and understand the remains. And we have had to learn the hard way that nothing is what it appears to be at first sight. The deserted settlement of Skramle was located in a somewhat irregular way. There was a tradition, written down in the 18th century, of the medieval farmsteads Skramle and Amot that were abandoned due to the Black Death in the middle of the 14th century and never resettled (Fernow 1977, 82). However the name ofSkramle was still used for a peninsula in the lake Gunnern and during the registration of archaeological sites of the parish of Gunnarskog, carried out in 1989 by the authors, efforts were made to locate the settlement site. But as nobody knew what to look for nothing was found. We returned to the peninsula of Skramle in 1990. Studies of old maps (17th-19th centuries) had added more probability to the theories that, somewhere on the peninsula, there really had been a settlement, abandoned before the oldest map (1641, LMV R3-97) came into existence. But even if the maps showed that there were was an area named Skramle that was in use but not settled, the maps did not indicate where a former settlement could have been located. As we had encountered a great interest for archaeology in the parish ofGunnarskog, we were accompanied by a group of local inhabitants on our return to Skramle in 1990. This was the beginning of a very fruitful co-operation that was later going to be 'institutionalised' into an association by the name 'Project Skramle'. And with a great deal ofluck we found the site. We took up two small interventions at random, and in the second intervention we hit a hearth that 223 S. Andersson & E. Svensson I I I I I I I "• I --~·· :/~.~~MHFN I ....... - GUNNARSKOG~-~~~~.- 6vr;noren ·············-·· ... ··. Sk~b·l~............. , rudsudden '\ o\hn.kat; '• ······ ·············· Fig. 3.- Historic Map Overlay, by Sofia Andersson, based on a map from 1709 (LMV R29-25:2). The original scale of the overlay 1:10 000, here diminished to 1:20 000. could be dated with 14C to the middle of the 15th century A.D. The fact that this was a hundred years later than the Black Death did not bother us. Once we knew that we had located the site we also started to see things that grew out of the ground and formed nice patterns of houses, ancient fields and other structures. The real excavations at Skramle started in 1992 and were originally supposed to have been concluded in 1996, but as the site has proved to be larger, more complex and of a scientifically much larger value than we had thought the project has been extended to 1998. As the project lacks sufficient financial support each excavation season is very short- from a week/ year to three weeks/year. The results are regularly reported in technical archives reports (Andersson & Svensson, unpublished reports), but so far only the results from the years 1990-1994 have been published (Anderson & Svensson 1995). So far eleven houses have been located and subjected to excavations to various extents. These houses can be grouped in three chronological phases; the 6th century (H IX), the 13th-early 14th century (H VVIII, X and maybe also H XI) and the 15th century (HI-IV) A.D. Around the houses there are remains of 224 ancient fields that have not been dated, but appear to have been used also after the abandonment of the site according to stratigraphy and finds. Probably at least some of the fields still in use did also belong to the medieval farmstead. The oldest house (H IX) from the 6th century A.D. appears to have been a kind of long-house (the size is not known as too small a part has been touched by excavation) with roof-supporting posts. As there is hardly any daub connected with this house the walls seem to have been of wood. A shallow hearth dug down in the ground in this house is among the constructions that have been located. The other houses appear to have been wooden constructions (most probably log-timbered). Most of the houses (H II-VI and H XI) are placed on a light southern slope with big stones and sometimes on terraces created to level out the slope. We do not know if there were several houses belonging to the 6th-century phase, but we know that for each of the two later phases there seem to be several houses with specialised functions. According to the general assumption, there was a change in settlement pattern around 1000 A.D., when the multifunctional longhouses were successively replaced by several smaller houses with specialised functions. This change took place over an extended period and was different in nature in different regions (Augustsson 1992, 64). Judging from the existence of hearths in four out of five houses belonging to the 13th-14th-centuries phase (H XI is omitted in this discussion as the evidence is still too limited), it is probable that Skramle at that time was a farmstead with several different households. The presence of cooking vessels and other items connected with daily domestic activities indicates that all four houses were dwelling houses, but we do not know whether they were contemporary. It is impossible to establish a relative chronology of the houses based on stratigraphical sequences, but it is likely that at least the houses H VII and H X were in existence at the same time as we have found remains of melted and slagged brass in house H VII and a buckle made out of the same brass in house H X (Nystrom, communication). The hearths in these four houses are of two kinds: the hearths in houses H VI and H VIII consisted of pits filled with fire-cracked stones and charcoal and in H VII and H X there were smoke-ovens with large bricks inside. The only house lacking a hearth (H V) is a large house that seems to be divided into three sections. We have interpreted this house as a combined barn, store, stall, etc. If the 13th-14th-centuries phase was crowded with dwelling houses and equipped with only a single outhouse, the situation was reversed in The local and regional arena of a Middle Age Swedish farm 0~ • Fig. 4.- Plan over the excavation-site of Skramle. the 15th-century phase. So far only the house HIll has been interpreted as a dwelling house (but it has not been excavated to any larger extent), whereas house HI, with a large hearth in a small house, probably was a cooking house and the small houses H II and H IV most likely served as storehouses. During 225 S. Andersson & E. Svensson the excavations in 1993 we made an astonishing discovery; one of the stones in the fundament to the large house- H V (M) belonging to the 13th-14thcentury phase- turned out to be a runestone with the older runic alphabet (juthark) dating from the 6th-8th centuries A.D. This was remarkable in two ways: first, only some twenty runestones with the older futhark are known in Sweden and, second, Skramle is situated in an area that should not, according to the present research status, have been colonised until the 12th century at the earliest (we had not yet located the 6th-century house)! The interpretation of the runic inscription is still being debated and will therefore be disregarded here. For the moment 476 finds have been registered; 458 ofthem belong to the 13th-14th-century phase, 2 to the 15th-century phase and 16 are more recent. The find material is dominated by simple objects of iron and stone, but there are also a few 'luxury' objects of bronze and brass such as a goblet, a buckle, a heraldic mount and a ring brooch. We have tried to sort the objects into different groups related to different functions and activities. The different groups are: domestic utensils (D), building details (B), objects related to husbandry and other animals (HA), handicraft (H), hunting and fishing (HF), agriculture (A), personal adornment (P) and others (0), comprising of pieces of flints and quarts, fragments and unidentified objects. Group D includes vessels, strike-a-light flints and grindstones. It is worth noting that no pottery has been found at Skramle, and apart from pieces of a probable goblet made of bronze all vessels were made of soapstone. Building details are mostly nails, but also hooks, hinges, plates and locks. The most frequent item in the HA-group is the horseshoe nail, but there are also some horseshoes, parts of snaffles and buckles. Almost all objects in this group relate to horses and very few objects indicate the presence of other animals. A unexpectedly large group is the handicraft group. One reason for this is that whetstones have been included here, but also because there has been some handicraft production at the site. There are both tools and refuse connected with the manufacturing of soapstone objects (probably mainly spindle whorls and vessels). Other important artefacts are spindle whorls, smoothing stones, scrapers, awls and the above-mentioned melted and slagged brass. On the other hand,the HF- and A-groups are very small; there are hardly any objects that can be connected to fishing, only a few arrows (mostly for crossbows) and a small number of sickles. Ironically, basic subsistence is badly represented among the artefacts. Until the excavation season of 1996, objects for personal adornment were both very few and of a very basic nature, such as simple iron buckles and 226 knives. But in 1996 a heraldic mount (not attributed to any family of nobility), a ring-brooch and an ornate buckle in brass were found in house H X. Local and regional trade: contacts and influences There are few studies on handicraft, production and trade in the medieval countryside in Sweden. Modem research, mainly within the field of history and economic history, has concentrated on the 16th century and the following era. Several researchers, influenced by Braudel and his concept of la tongue dun~e, argue that the economic conditions resemble the previous period i.e. the late medieval period (Magnusson 1996, 35-61, 98-105). Some scholars believe that research has underestimated the importance of the market economy in the medieval agrarian context (Winberg 1985, 156-158). Others claim that the so called natural economy, i.e. production intended mainly for the household, was the basis for the agrarian society (Osterberg 1977). Studies on handicraft in rural areas have been carried out in various regions of Europe, for instance in Germany by Helga Schultz and in England by Rodney Hilton (Gadd 1991, 53-56). A recent study on Swedish handicraft in the countryside shows that the crafts are underestimated in Swedish official records. These mentioned only those craftsmen who paid tax as craftsmen. During the Middle Ages, written law permitted craftsmen with land or with a domicile to avoid the relatively high taxes imposed on craftsmen by being taxed under some other title, i.a. as a farmer (Gadd 1991, 201-21 0). This study indicates that there was a production of handicraft in the countryside that never can be seen in the written sources. With this information, it is interesting to look into the archaeological material and see whether this picture can be verified or not. The investigations at Skramle indicate that handicraft for sale could have been very common in the Swedish medieval countryside, at least in the forested areas. As with rural studies as a whole, however, the problem is that we lack similar archaeological studies in Sweden that can complement the picture. Artefacts linked to handicraft can be observed in the archaeological material at Skramle. There are number of indications for 'over-production' of commodities. These products do not belong in the agrarian sector, such as cereals and livestock. Instead we find traces of an extensive soapstone production. Products such as spindle whorls and vessels have been made on the farm. Buildings, especially built for production within various types handicraft, were probably used. The occurrence of soapstone artefacts The local and regional arena of a Middle Age Swedish farm in archaeological excavations is usually related to contacts with Norway, since quarries for soapstone and production of soapstone items are well documented there. At Skramle we see that the use of soapstone quarries in the area probably was very important. The market for soapstone products was probably on a local scale. We can see that the material standard on the farm was very high during the period of soapstone production and during this period a lot of' outside' influences can be detected, partly in building techniques and partly in the finds material. At Skramle we relate the soapstone products to a local market, but there are other traces of more farreaching contacts. We have some indications for the currying of hides. For instance, smoothing stones and scrapers have been found. Hides were a product that was interesting for a larger market, extending beyond the local one. Either traders came to the region or the inhabitants of Skramle travelled themselves to a market-place to sell their products. We do not know for what kind of markets Skramle was producing its goods. The more important market-places, Tingvalla (Karlstad) and Knusesund (Saffle), were the ones nearest and they were used for some purposes, but Gunnarskog parish is also located near the Norwegian border and the foreign trade was probably very important at this time. Even if a Swedish town or market place was the closest, the farmer could sell his products on the other side of the border. This is particularly true in the case of a choice stands between an inland town and a port. To the farmer the prices were more important than the distance (Linge 1969, 30-43). The crown tried to stop this bordertrade during the late medieval period. Especially the trade in cattle was considered to be a threat to the Swedish economy. Amongst other things, the king aimed at strengthening the ties between town and countryside by making trade outside towns illegal. Obviously, this law was hard to enforce by the central government, as shown by the numerous exhortations and Royal Ordinances throughout the medieval period. In King Magnus Eriksson's town-law from the late 13th century it is stated that trade shall be conducted in the towns, both between townsmen and farmers and not in the countryside or in other places (Holmback & Wessen 1966, Kopmalabalken 23: 1). In a province such as Varmland, without towns and without a strong control from the state, the farmers probably had a larger possibility to evade the legislation and getting their products to an appropriate market and ensuring a high price. This relative wealth is visible in the diversity of the artefacts and their exclusive character during 13th and 14th centuries in Skramle. That some influences have reached Skramle can be interpreted in the artefacts discovered in connection with the investigations. But also the way the houses were constructed reveals external influences. In two houses we have found a kind of' smoke-oven'. This type of oven seems to be common in an urban setting. Whether it is common in the countryside is open to discussion. The question is problematic because of the lack of investigations. Studies on this topic have been carried out by Richard Blanton for different communities. He could observe that houses in the countryside often have been strongly influence by urban houses, especially in remote areas with farmers involved with long-distance trade (Blanton 1994). Is this the kind of influences we see in the 13th-14th-century houses in Skramle? Smoke-ovens are known in urban context both in Sweden and Norway at this time. One could argue that this iodic- 0 fls mllil HA rnH I;;;;;;J illlll HF I!! A rnmlp tiillill [Jo Fig. 5. - Relative importance of the different groups offinds; 13th-14th-century phase. 227 S. Andersson & E. Svensson ates a contact between an urban centre and the Skramle farmstead. The reservation is that we do not know how common smoke-ovens were in the countryside and in this area at this time. Another contact we can notice in the Skramle material is the melted and slagged brass. The brass production commences on a large scale in Sweden during the 16th century. Smaller production units have been located in the medieval towns, for instance in Lodose during the 14th century. It is not certain that the brass items were manufactured at Skramle. Maybe older brass objects have been melted down for casting new products. Thus, for instance, we have found a brooch made of brass which has possibly been produced at Skramle. Some other objects that can be characterised as 'luxury' objects have probably been bought on a market, since they are hardly of a local origin; these objects include a heraldic mount made of silver and a ring brooch made of silver and bronze. The contacts and influences at Skramle are one of the most important parts of our continued discussion in the project. We have shown that the preconception that a farm in a relatively remote area had little contact with the outside community can be questioned. The archaeological material is the only way to rectify this view and to find out what kind of economical situation and status a medieval farm located in a forest area really had, particularly when there are no written sources at all. Resource-utilisation and non-utilisation Skramle is situated in an area with waste forest outlands, and the forest and over time its resources have always been used in various ways. The possibility to use the outland has played an important role in making living in the parish of Gunnarskog possible, but the use of outlands could also be the basis for the production of goods for sale at a market and thus for the generation of wealth. It is important to point out that it was not the accumulation of wealth in itself that was interesting, but rather wealth as a mean for social reproduction consistent with a desired station. It was important to be a part of the society in general and its cultural values and practices also when living off the agricultural areas (Martens 1992, 5). What status did the farmers of Skramle wish to communicate? Probably the station of a wellsituated farmer-landowner, i.e. the dominating group within the local society and an important economic and political group in Viking age and medieval Sweden. It is also from this group that part of the nobility emerged (in Sweden the estate of nobility was institutionalised in 1280). With the rising power 228 of the state, king and church and the growing influences of European culture, the landed farmers had to chose sides. Some followed the king and could thus become part of the nobility. Others conserved old practises and values and considered the status of landed farmer and power within the local society to be the most important. The latter strategy appears to have been the most common one in the forest areas of central and northern Sweden (see Pettersson & Svensson 1996). Maybe the heraldic mount and the other luxury items from the late 13th century found at Skramle can be regarded as reflecting a landed farmer's wish to compete with the young nobility. Equally of interest is the fact that during the critical period of the late 13th century the farmers ofSkramle appear to have been cultivating the unusual and extravagant crops of wheat and oats (Ranheden, unpublished report) in a time when barley is believed to be the dominating crop in the region and wheat almost non-existent (Osterberg 1977, 217 -219). Resource-utilisation has to be studied from several points of views. Here the testimony from the surrounding landscape of the parish of Gunnarskog (see Fig. 2) and from the farmstead of Skramle will be confronted. There are four major categories of outland use that are traceable in the landscape as sites (outland use known to have been practised only in historical times are disregarded): pitfalls for elks, bloomery furnace iron-production with charcoal pits, soapstone quarries and shielings. Judging from the spatial organisation of these sites only a shieling can be connected with Skramle with some degree of possibility, all the other sites are located in areas belonging to other farmsteads in historical times. But the rights to the outland was a complicated matter: parts of the outland as well as single resources and constructions could be sold, donated, leased, etc. (Hvarfner 1960, 55-56, Backvall1920, 4-5). The organisation of the use of outland is another factor that complicates the picture. The fact that some of the systems of pitfalls stretch out between two farmsteads indicates that there was some kind of eoownership and co-operation between these farmsteads. On the other hand it is striking that only one farmstead is connected with iron-production and that several farmsteads appear not have participated in outland use (if outland use is restricted to the four known categories). In an as yet unpublished article one of the authors h:~s put forward a suggestion that the farmers in Gunnarskog practised a kind of specialisation in outland use and there was an exchange of goods on a local basis (Svensson, in print). This would implicate that the local area was of great importance (see above) and played an important role when forming subsistence bases and patterns for the individual The local and regional arena of a Middle Age Swedish farm farmsteads. Can such a theory be confirmed in the material from Skramle? The low number of artefacts connected with hunting and fishing at Skramle is striking. But even if the farmers of Skramle practised hunting and fishing it is not necessary that the objects related to this activities should be found at the farm; fishing implements could have been kept in buildings by the lake and hunting equipment could also have been kept outside the farmstead itself. On the other hand, the use of pitfalls did not demand extensive hunting equipment. When not being able to rely on the archaeological evidenc~ to illuminate the question of how important fish and wild game were to the daily diet at Slaamle, it would be natural to turn to the osteological evidence. Unfortunately the osteological material found at Slaamle is very sparse, only small fragments of burned bones seem to survive in the soil and most fragments are impossible to identify. The animals identified are sheep/goat, pig and cattle (Sjogren, unpublished report). The evidence of the sparse osteological material should thereby point in the direction of wild game not being part of the diet. But it is important to point out that it is possible that slaughtering and butchering could have taken place outside the farm in connection with the hunt (a system often practised today), and that meat from elks and other game could have reached Slaamle as pure meat both if the game was hunted by the farmer themselves and if it was bought from others. Hunting was also practised for other reasons than direct subsistence, e.g. for furs. Furs appear as an important economic resource in forest areas in Sweden, especially northern Sweden, and were exported to the European markets (Zachrisson 1996, 4). Cross- bows were used in the hunt of squirrels and other small fur-bearing animals. From the existence of such arrows and smoothing stones at Slaamle it seems likely that hunting for fur was part of the activities at Skramle, but it is difficult to judge its extension and importance. Moving from hunting to iron-production and smithing, we can start by stating that 229 objects of iron have been found at Skramle. Most of the objects appear to be of good quality and are well made. So far neither bloomery furnaces for iron-production nor a forge for further processing of iron into objects have been located at Skramle. Small amounts of slag (a little bit more than 2 kilos), probably after smithing, have been found at Skramle. Pieces of slag are found over most of the excavation area, but the small 'concentrations' that are found are located near the hearths of the dwelling houses. Maybe this is the result of minor mending of iron objects at home. Indications for such a practise are given by a knife that, when conserved, showed clear signs of having been the object of mending by an unskilled person (information from conservator Martin Andren, Kulturen, Lund). The absence of iron production and the practise ofhome-mending indicates that iron objects were bought and not made at Slaamle, maybe they were bought from the local iron-producers in the parish of Gunnarskog. A handicraft well visible at Slaamle is the manufacturing of soapstone objects. As has been stated above the production appears to have been directed towards a local market, as both the soapstone and the craftsmanship are of relatively low quality and the production seems to be directed towards a low number of different products. The low quality of the Out land Ill Non Outland Fig. 6. - Relative importance of objects related to outland production at Skramle (13th-14th-centwy phase) 229 S. Andersson & E. Svensson soapstone indicates that it was quarried in the vicinity, and there are two soapstone quarries located in the north and the south respectively of the parish of Gunnarskog. An analysis of the origin of the soapstone at Skramle has given evidence that the southern quarry was the one used most, but there are also indications that a minor quantity of the soapstone originates from the northern quarry (Nilsson 1995, 30). Whether the farmers of Skramle possessed parts of these quarries or bought the soapstone is not known. We will sum up these three categories of outland use and their importance at Skramle first. Among the different groups of finds into which the finds material of Skramle has been sorted and which have been presented above, there are no groups related to outland use - apart from the small group of hunting and fishing. If we instead make a group of objects related to hunting, fur working and soapstone production, the relative importance of outland production among the total finds material becomes more evident. As the fourth category of outland use, the shielings are part of the agrarian economy and thus part of the farmstead. In the county of Varmland shielings were used for grazing cattle during summertime, for hay making and sometimes also for agriculture (Eles 1975, 236-241). The relative importance ofthe shielings depended on how well these needs could be satisfied by the farmstead itself, and it is therefore not surprising that most of the shielings are found in the northern part of the parish of Gunnarskog where the agricultural lands of the farmsteads were of a lesser quality. There was however a shieling situated relatively close to Skramle, and as this was abandoned early it is impossible to attribute to a specific farmstead. The other farmsteads in the vicinity of this shieling however had their shielings in other locations, so it is possible that this particular shieling belonged to Skramle. The use of shielings was not without its problems as the long absence of the cattle caused a lack of manure for the fields by the farmstead. In northern Varmland, where shielings were common, the farmers had elaborate systems and dedicated a lot of work to compensate for the lack of manure (Morner 1952, 70). As the products of a shieling are the same as can be found at a farmstead, we have chosen to discuss the possible use of shielings by the farmers of Skramle from the point of view of manuring. When presenting the finds material found at Skramle the total domination of finds belonging to the 13th-14th-century phase is apparent. The cultural layers from this period are also of a different character than the cultural layers from the 15th-century phase. The earlier cultural layers are thick (some230 times almost 0,5 m thick) and greasy, but the 15thcentury ones are thin (hardly 0,1 m thick) and dry. There are several possible explanations to this; maybe there were fewer people living at Skramle during the 15th century (likely) or maybe this phase was a lot shorter than the preceding phase (not likely); maybe there even was a change in attitudes towards rubbish etc. The explanation of fewer people during the 15th century is however not enough to explain the thin cultural layers, as there were also more houses and larger areas in use during the 13th-14th-century phase. There appears to be a change in attitude towards rubbish for during the 13th-14th centuries there are concentrations of refuse close by the houses and these are lacking in the later period. Maybe there was a combination of a wish to alienate the refuse from the vicinity of the living quarters and a need for manure for the fields that resulted in the thin 15thcentury cultural layers. It is hard to use this tentative discussion as a conclusive argument for the use of shielings in the 15th century, but it can be used as an hypothesis for further investigations. We conclude this article by saying that it is important to look upon a farmstead in a larger context than the settlement itself. Medieval farmers in Sweden did not spend their entire life at home and they were part of different spheres of society. Here we have put the emphasis on the local and regional arenas, which we believe to have been of great importance to our farmers at Skramle, but there were other arenas that are still left to study. References ANDERSSON S. & SVENSSON E. 1990, 1992, 1993, 1995: Rapporter over arkeologiska delundersokningar av odegarden Skramle, RAA_' 595, Gunnarskogs socken, Unpublished reports (VM Arkiv). ANDERSSON S. & SVENSSON E. 1995: Skramle. Historien omen gard och en arkeologisk undersokning. Projekt Skramle 1990-94, Samha!lsvetenskap Arbetsrapport, Hogskolan i Karlstad, 10. AUGUSTSSON J -E. 1992: Medeltida husbyggande i Sverige, Bebyggelsehistorisk tidskrift 23, Stockholm, 55-85. BLANTON R. E. 1994: Houses and Households. A comparative Study, New York. BACKVALL L. 1920: Bliistugnen vid Bliistkiirringbiicken pa Sodra Braniis iigor i Dalby socken och dess iigare. ELES H. 1975: Satrar i V arm land, Sista !asset in: BERGENGREN G. (ed.), Studier tilliignade Albert Eskerod 9/5 1974, Stockholm, 231-252. The local and regional arena of a Middle Age Swedish farm ERSGARD L. & HALLANS A-M. 1996: Medeltida landsbygd. En arkeologisk utviirdering. Forskningsoversikt, problemomraden, katalog, Riks- antikvarieambetet Arkeologiska Undersokningar Skrifter 15, Stockholm. FERNOW E. 1977: Beskrivning over Viirmland .. Ny utgava med kommentar av Arvid Ernvik. (177 379), Karlstad. GADD C-J. 1991: Sjiilvhushall eller arbetsdelning? Svenskt !ant- och stadshantverk ea 1400-1860, Meddelanden fran Ekonomisk-Historiska institutionen vid Goteborgs universitet 64, Goteborg. HVARFNER H. 1960: Fangst, in: HVARFNER H. & JANSSON S. (eds), Fran Norrlandsiilvar och f}iillsjoar. Riksantikvarieiimbetets kulturhistoriska undersokningar i samband med kraftverksbyggen och sjoregleringar, Stockholm, 51-64. HOLMBACK A. & WESSEN E 1966: Magnus Erikssons stadslag. I nusvensk tolkning, Rattshistoriskt Bibliotek, sjunde bandet. Skrifter utgivna av institutet fOr rattshistorisk forskning ser. 1, Lund. HYENSTRAND A. 1994: Landets bebyggelsehistoria, in: Kulturminnen och kulturmiljovard, Sveriges Nationalatlas (Selinge K.-G. red.), 8-9. LINGE L. 1969: Griinshandeln i svensk politik under iildre Vasatid, Lund. LMV = National Land Survey of Sweden, Gavle (Statens Lantmateriverk, Gavle ). MAGNUSSON L. 1996: Sveriges ekonomiska historia, Fa1un. MARTENS I. 1992: Some aspects of marginal settlement in Norway during the Viking Age and the Middle Ages, in: MORRIS C.D. & RACKHAM D. J. (eds), Norse and later settlement and subsistence in the North Atlantic, Glasgow, 1-7. MbRNER A. 1952: Kort oeconomisk beskrifning Ofwer Wermeland ahr 1762, in: Viirmland forr och nu, Karlstad, 7-146. NILSSON S. 1995: Tiiljstensforemal pa Skramle. Hems!Ojdat eller kopt?, C-uppsats i arkeologi, Hogskolan i Karlstad. Ortnamnen i Viirmlands liin 6, Josse harad, Gunnarskog socken, 1942, 39-49. PETTERSSON S. & SVENSSON E. 1996: Jakten pa den fdrsvunna "Svensson", META 1996:1, Lund, 31-43. RANHEDEN H. 1995: Makrofossilanalys, Vr. Gunnarskog RA;{ 595 "Skramle ",unpublished report, VM Arkiv. SAMUELSSON S. & KALLSTENIUS G. 1939: En varmlandsk skattebok fran ar 1503, Nationen och hembygden If, 1939. SJOGREN J. 1995, Osteologisk analys, Vr. Gunnarskogs socken, RAA' 595, "Skramle ", unpublished report, VM Arkiv. SVENSSON E. 1995: Jiirnframstiillning i norra Viirmland, Samhallsvetenskap, Arbetsrapport 9 Hogskolan i Karlstad. SVENSSON E. 1996: Handiga skogsbOnders produktion och varuutbyte. Article under translation to English to be print in: Off the beaten track, Lund. ZACHRISSON I. 1996: Pelsverk fra nord- hardvaluta i Europa, Spar - fortidsnytt fra midt-norge 21, Trondheim, 1996. OSTERBERG E. 1977: Kolonisation och kriser. Bebyggelse, skattetryck, odling och agrarstruktur i viistra Viirmland ea 1300-1600, Lund. Communication with Inger Nystrom, Vastsvenska Konservatorsateljen, Goteborg. Sofia Andersson & Eva Svensson Department of Medieval Archaeology Archaeological Institute University of Lund Sweden 231 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Julian D. Richards Anglian and Viking Settlement in the Yorkshire Wolds Summary The Anglian and Viking town of York is now wellknown, but we know considerably less about rural settlement in its hinterland. This paper describes the interim results of a landscape project which aims to characterise and map early medieval rural settlement in the Yorkshire Wolds using information from aerial photography, geophysics, field-walking, metal detectors, and sample excavation, combined within a GIS. The project has revealed a developing settlement hierarchy from the Middle Saxon period, and nucleation at new centres coincident with the Danish settlement of Northumbria. Introduction The Yorkshire Wolds are rolling chalk uplands in Eastern Yorkshire. They are bounded to the north by the Vale of Pickering; to the west by the Vale of York; and to the south-east by Holdemess. One arm extends directly south as far as the Humber Estuary which, from around the 7th century AD appears to have formed the southern boundary of the early medieval Kingdom of Northumbria. To the north-east the Wolds extend as far as the North Sea coast at Flamborough Head. The underlying geology is chalk and so there is little surface drainage, although there are a number of dry valleys whose origin is uncertain, but which are thought to date from an ancient glaciation. There is one substantial watercourse, the Gypsey Race, which rises not far from the medieval settlement at Wharram Percy, and flows down the Great Wold Valley into the North Sea at the modem seaside town of Bridlington. Today the Wolds are lightly populated with settlement concentrated in dispersed farms and villages. The thin chalk soils are largely given over to arable farming although both cows and sheep are also grazed on the grass uplands. There are market towns at Malton and Driffield, lying off the chalk to the north-west and south-east respectively, but the nearest major urban centre is the City of York, which lies some 12 miles (c. 20 km) from the western Wolds edge. The Wolds appear to have been extensively farmed from at least the middle Bronze Age. There are numerous round barrows but the most significant feature is a series oflinear earth works that divide the landscape into territories and large enclosures. It was in the late Iron Age, however, that the Wolds became fully settled. Mapping of crop marks from aerial photographs reveals a landscape dissected by ancient trackways and partitioned by extensive field systems. Many of these form so-called "ladder patterns" comprising series of rectilinear fields or paddocks defined by ditches and often fronting onto a trackway, with occasional settlement enclosures. That this landscape is pre-Roman is clearly demonstrated south of Wharram-le-Street, where the Roman road south of Malton cuts obliquely across the field systems and trackways. Where such ladder settlements have been excavated, as at Wharram Percy, a Late Iron Age date has been confirmed, although they have generally been shown to continue in use into Romano-British times (Beresford & Hurst 1990, 87-92). The spacing of the settlements is generally every half mile (1 km), for example along the Thixendale valley, where there is a medieval village every 1 mile (1.6 km) and a Romano-British farm halfway between (Beresford & Hurst 1990, 92), The trackways look like cattle droveways and the paddocks may have served as animal enclosures although excavation of some of the medium-sized local villas and their associated corndrying ovens shows that cereal crops were also already important by the Roman period. The immediate post-Roman settlement pattern of the Wolds is much less well understood. Ancient pollen does not survive well on the chalk soils but most environmental archaeologists assume that the Wolds landscape remained largely cleared and that there was no extensive reafforestation. Nevertheless, there are no securely dated Early Anglo-Saxon settlement sites from the Wolds (Watkins 1983). Early Anglo-Saxon cemeteries have been excavated on the western Wolds edge at Sancton (Myres & Southern 233 J.D. Richards Fig. 1.- Location of Crop-mark sites Cottam A & B. 0 ' Cottam B r1 0 o((!J 0 ( 0 Cottam A <; ~ Cott•m DMV 1 o.-.-ac====~--~=====----- 1 km 1973; Timby 1993), and to the east at Sewerby (Hirst 1985), but only isolated burials are known from the Wolds tops. To the north the Anglo-Saxon cemetery and extensive settlement at West Heslerton (Powlesland 1986) lies just off the chalk, at the southern edge of the Vale ofPickering. During what is known as the "Final Phase" of pagan burial, in the 7th and early 8th centuries, there was a trend to highly visible burials in the Wolds, often reusing prehistoric barrows, such as at Garton and Uncleby (Mortimer 1905). Few Middle Saxon settlements have been excavated al234 though post-built structures have been discovered at Thwing where occupation debris, including sceattas and stycas, indicates activity from the 8th century onwards (Manby forthcoming). Northumbrian stycas found within some of the buildings at West Heslerton suggest that this site was also occupied in the 8th century. As witness to the 9th-century Danish settlement there are large numbers of Scandinavian type place-names but until recently there was scant archaeological evidence for Viking Age farmsteads in the Wolds. York may have been a thriving and Anglian and Viking Settlement in the Yorkshire W olds Fig. 2.- Cottam Site B: Metal detector finds. ARCIJNFO 7.0; J.D. Richards 1997; finds database: T. Austin, C. Buckley & J Kenny. + + + + + + + + ++ + + + + + + + 0 lOOm cosmopolitan trading centre from the early 8th century onwards (Kemp 1995; Hall 1994), but its hinterland was virtually unknown. Archaeological investigation oftwo Wolds settlements at Cottam and Wharram Percy, neither of which would have been suspected to have had an early medieval origin, has now begun to indicate extensive Wolds settlement in the Middle Saxon period, with nucleation and the development of new sites in the Viking Age. It is suggested that the majority of such settlements, however, continued as successful medieval and post-medieval villages and their origins therefore lie buried under present-day churches and houses. Cottam The existence of an early medieval settlement near Burrow House Farm came to light as a result of the discovery, over many years, of large numbers of Anglian and Anglo-Scandinavian artefacts by metal detector users. From 1987-89 some 200 man-hours of searching yielded over 60 pieces of 8th and 9th-century date (Haldenby pers comm). The find spots have been systematically recorded, and the objects have been published (Haldenby 1990, 1992, 1994). Several of the metal items are quite corroded, having suffered from agricultural disturbance, whereas much appears to have only been ploughed up in recent years and is still in a good state of preservation. To date the published finds include some forty simple pins, as well as disc-headed and racket-headed pins, a lead alloy brooch, over 30 ninth-century strap ends, a gilt mount, a fragment of rolled gold sheeting, over 20 ninth-century stycas, and numerous Roman coins. There is also evidence of domestic activity, including 8 lead spindlewhorls, and some 40 iron knife blades. The Anglo-Scandinavian finds include 2 Jellingestyle brooches and 2 so-called Norse bells. A chalk weight with a possible ship graffiti has also been 235 J.D. Richards 0 E9 0 0 oo E9 oo @0 0 lJ OQI @ 0 0 Fig. 3. - Cottam Site B: Datable metal objects. ARCIINFO 7.0; J.D. Richards 1997;finds database: T. Austin, C. Buckley & J. Kenny. recovered from the ploughsoil (Richards 1995). The metal detectorists did not make any systematic attempt to recover the non-metal artefacts, although they acknowledge that substantial quantities of both pottery and bone were observable in the ploughsoil. These finds are largely coincident with a sub-rectangular crop-mark enclosure, henceforth described as Cottam B, which does not appear to be linked with an associated field system. Instead, it appears to sit astride a trackway which skirts the dry valley and then runs south-east to a crop-mark ladder pattern enclosure, known as Cottam A, and then further south to the earthworks of the deserted medieval village ofCottam (Fig. 1). If the distribution of the metal-detector finds at Cottam B is plotted there appear to be two foci (Fig. 2). The southern concentration coincides with the 236 sub-rectangular ditched enclosure, whilst there is a second focus to the north which is less clearly associated with other features. Magnetometer survey, however, has revealed that there are further ditched enclosures in this area but in general their shallow depth means that they have little effect on crop growth. If the position of datable metal objects is plotted then it suggests that there is a difference in date between the two clusters, with most late 8th-century finds towards the south, and late 9th and 1Oth-century finds in the northern .::luster (Fig. 3). Field walking has confirmed the picture derived from the distribution of metal-detector finds (Didsbury 1990). This shows that there is a general background distribution of Roman potsherds across the field, whereas AngloSaxon sherds are concentrated towards the east. The Torksey ware sherds, which are not current before the Anglian and Viking Settlement in the Yorkshire W olds \ ,/ ... .J ...--.~ o Excavated finds 0 111 Exco.vated cu alloy finds lOOm Fig. 4.- Cottam Site B: Metal detector and excavation finds. ARCIJNFO 7.0; J.D. Richards 1997;finds database: T Austin, C. Buckley & J. Kenny. tenth century, are particularly focused towards the north-east, which is where the 1Oth-century metal finds were clustered. This localised shift has been further confirmed by excavation (Fig. 4). In 1993 two trenches dug across the southern enclosure revealed post-hole buildings and settlement debris of the 8th and early 9th centuries (Richards 1994). To the east of one of the buildings was a circular pit, c.l.5 m in diameter, towards the middle of which was an adult female skull, radiocarb0n dated to 1295 ± 60 BP (with a calibrated range of AD 664-775 at the 68% confidence limits). The fill of the pit also contained a Wessex silver penny of Aethelberht, dated AD 858-62, an AngloSaxon dress tag and two decorated comb fragments. The contents of the pit were sieved and yielded skeletons of frogs and water voles which must have stumbled into it whilst it lay open. By this stage, therefore, it is proposed that this area of the site had been abandoned. The settlement had been relocated to the north-east where a third trench dug in 1995 revealed a farmstead of the Viking Age (Richards in prep). This comprised a number of rectangular enclosures enclosing post-hole buildings either side of a trackway which would have been entered through a massive ditched and banked gateway. Occupation appears to have been relatively short-lived, perhaps spanning some 50 years or a single generation only, from the late 9th to early 1Oth centuries. It is proposed that at that point settlement shifted again, probably to the site of the deserted medieval villages at Cowlam or Cottam, although this theory is as yet untested by excavation. In 1996 excavation of the crop-mark enclosure at Cottam A revealed a more typical Romano-British 237 J.D. Richards farmstead. The Anglian and Anglo-Scandinavian artefacts recovered by metal detectorists in this area are interpreted as representing activity but not permanent settlement. It appears that the prehistoric trackway was still a landscape feature and would have been used by Anglo-Saxon or Scandinavian settlers travelling between their farms at Cottam B and any contemporaneous settlement under the site of the Cottam DMV. Such farmers may have driven their animals along this route and may have taken advantage of the opportunities provided for watering at the quarry holes which excavation has revealed were still open at Cottam A. In summary, the three seasons of excavation at Cottam have revealed a shifting and evolving early medieval settlement pattern. It is difficult to escape the conclusion that the sub-rectangular enclosures at Cottam B represent Anglian and then Anglo-Scandinavian farmsteads, whilst the more traditional ladder-pattern farmstead at Cottam A is confirmed as Late Iron Age and Romano-British. This work may therefore allow a reassessment of the typology of crop mark enclosures and a re-examination of the large number of undated enclosures known from the Yorkshire Wolds (Fig. 5). It is anticipated that far more may turn out to be of early medieval rather than oflron Age or Romano-British date. Wharram Percy The second site to be considered is better known for the campaign of excavations aimed at understanding the desertion of the medieval village, pioneered by Beresford and Hurst ( 1990). Excavations at Wharram Percy have revealed, however, that there was also considerable Middle Saxon activity spread over much of the area of the later medieval village (Fig. 6). In 1975, a two-post sunken-featured building was excavated at Site 39, on the northern fringes of the medieval village (Milne & Richards 1992). The finds from its fill included a Northumbrian sceat of c.750, and a sherd ofTating-type ware, of the late 8th or early 9th centuries. During 1980-84 two sunkenfeatured buildings were identified cut into a Roman hollow-way in Site 60, immediately south of the North Manor. These were originally described as 6th century (Hurst 1984, 82), largely on the basis of a strike-a-light for which 6th-century parallels were quoted, but it has been suggested that they could also be later (Milne & Richards 1992, 93). In 1989-90 a fourth sunken structure was found cut into a RomanoBritish field boundary ditch in Sites 94 and 95 (Milne & Richards 1992, 13-25). The associated midden deposit provided evidence for non-ferrous metal238 working, with both crucibles and clay moulds, including a mould fragment with interlace ornament dated on stylistic grounds to the late 8th or early 9th centuries. The midden also included a fragment of a stone cross-head dated to the 8th century. Further Middle Saxon finds have been made in other parts of the Guardianship area, including a second 8th-century cross fragment, a styca and a 9th-century strapend from Site 12, on the plateau immediately above the church, and a number of stycas and sceattas from the church excavation itself. The most intensive Middle Saxon occupation, however, was on the site of the later medieval South Manor house, where a Middle Saxon timber hall and smithy have been excavated (Stamper & Croft 1997). The South Manor site plays a critical role in any discussion of the nature of Middle Saxon occupation at Wharram Percy, and of its development into the post-Conquest village. Of all the Middle Saxon sites it is the only one which was still occupied in the postConquest period, and furthermore, as the site of one of the two manor houses, it clearly had a special status in the later 11th and 12th centuries. If continuity of occupation can be demonstrated then the South Manor provides valuable evidence for preConquest manorial origins comparable, for example, to that recovered from Raunds (Northamptonshire), or Goltho (Lincolnshire). There is no firm evidence for major early AngloSaxon activity in the South Manor area; most of the Anglo-Saxon pottery recovered from the South Manor site is probably 7th- or 8th-century. This is confirmed by the radiocarbon dating where the four samples from features associated with the smithy all fall within the calibrated date range of AD 600-1010 at the 68% confidence limits. A silver sceat of AD c. 700-710 is probably the earliest datable contemporary object, but two sword pommels and a 7th- or 8thcentury hilt guard are also probably contemporary with the Middle Saxon settlement, and are particularly diagnostic of its high status. During the Middle Saxon period we also have evidence for the first post-Roman laying out of boundaries and enclosure of the landscape at Wharram Percy. Two Middle Saxon boundaries were recovered from the South Manor excavations; both were east-west ditches, the second apparently replacing the first. The later ditch was contemporary with the smithy; the earlier was sealed and cut by smithing activity. It has been suggested that this redefinition of the boundary 1 m to the north indicates that there was relatively little pressure on land at this stage and that the Middle Saxon occupants of the South Manor site were thereby able to extend their holding (Stamper & Croft 1997). Anglian and Viking Settlement in the Yorkshire Wolds lOOm Fig. 5. - Cottam Environs: Undated crop-markenclosures. Topography from Ordnance Survey 1:50,000 map. Crown copyright reserved. ARC!JNFO 7. 0; J.D. Richards 1997; digitised by H. Fen wick. The Middle Saxon soils are believed to have been immediately sealed by "boney loams". This is thought to have happened quickly because the final Middle Saxon surfaces contained large sherds of pottery and complete bones, although the dating of the boney loams is unclear. One possibility is that at least the lower levels are Anglo-Scandinavian. Two radiocarbon dates for material interpreted as originating in the boney loams provide calibrated date ranges of AD 960-1030 and AD 880-1000, at the 68% confidence level. There is good evidence for Anglo-Scandinavian presence in the South Manor area, including schist and phyllite hones and a 9th- or 1Oth-century sword hilt guard, close in appearance to a smaller guard from Coppergate. Most significantly, there is a strap end and belt slide decorated in the Borre interlace style, current in Scandinavia in the 9th and 1Oth centuries, one of which was recovered from the boney loams. Caroline Richardson has noted that such belt slides are rare finds from the British Isles and the presence of Scandinavian dress accessories at Wharram "could indicate the presence of a Viking man at Wharram Percy" (in Stamper & Croft 1997). If not imported from Scandinavia then these items must at least have originated from within the Scandinavian community in Britain. It is thought that the first earth or turf built phase of the main manorial boundary bank may have been broadly contemporary, or slightly later than, the boney loams. It has also been suggested that the so-called lynchet bank which separates the medieval peasant 239 J.D. Richards tofts from their crofts is broadly contemporary with the boundary bank. It is suggested that this may be Anglo-Scandinavian. Many of the surviving earthworks at Wharram derive from the late medieval fragmentation of the village caused by its gradual desertion and the obliteration of disused croft boundaries (Hurst 1984, 83 ). It is clear, however, that the village originally consisted of a regular plan with two parallel rows of tofts and crofts. It is also very likely that the whole system of thirty tofts was planned from the start, and not assarted gradually from surrounding waste. Each toft and croft was, with certain exceptions, laid out on a module of 60 ft ( 18 m). This uniform layout suggests that the Danish Solskifte system (shown by documents to be in general use in the W olds area in medieval times) dates back to the laying out of the village and fields. There are three periods at which the planning of the Wharram area may have taken place: (1) at the time of initial settlement in the Middle Saxon period; (2) following the period of Scandinavian invasion and settlement; or (3) in the late 11th or early 12th century following the disruption of the Harrying of the North (Hurst 1984, 85). The third possibility is the most easily dismissed as it is thought unlikely that such major replanning could take place in the 11th or 12th century, when many different landowners would have been involved, without some documentary evidence in surviving deeds or charters. The first possibility is difficult to disprove but there is no archaeological evidence for it and it is also thought unlikely that the population of the Wolds would have expanded to its greatest extent in the 8th century (Hurst 1984, 86). From field walking in the Wharram area it is also clear that the pattern of Middle Sax on pottery finds supports the idea of scattered settlement at this stage whilst Late Saxon pottery is only found at the nucleated village sites (Hurst 1984, 82). In 1984 John Hurst suggested that the most likely date at which Wharram Percy became a nucleated planned village was during Scandinavian reorganisation in the 1Oth century (1984, 86). This likelihood is now reiterated. Conclusion At both Cottam and Wharram Percy there is evidence for the origins of early medieval settlement in the late 7th or 8th century with subsequent reorganisation in the 1Oth century, although we should not seek a single pattern of settlement evolution as at each site development took a different course. At Wharram there is a deal of circumstantial evidence 240 which supports both a 1Oth-century laying out of the village, and an Anglo-Scandinavian context for it. At Cottam the excavated Anglo-Scandinavian settlement was short-lived and it has been suggested that it was re-established as the Cottam DMV. The late 9th and 1Oth centuries was a time of tenurial change, due to the Viking settlement. A number of former large, often ecclesiastical, estates were being fragmented and passing into private ownership (Richards 1991, 30-31 ). In Yorkshire the Anglo-Saxon Chronicle informs us that in 876 "Healfdene shared out the lands of the Northumbrians, and they [the Scandinavians] proceeded to plough and to support themselves". The Yorkshire Wolds are one of the main areas where we believe that Scandinavian landlords took control of former large estates. In the former east Riding of Yorkshire 48% of place-names are Scandinavian influenced. In the Wharram area there is an almost complete replacement of English by Scandinavian place-names. This cannot denote completely new settlements filling in gaps as we now know that most of the landscape was already extensively farmed, at least from the Roman period. Existing settlements must therefore have been renamed after a redivision of the land by the new Scandinavian aristocracy. We have already noted that at Wharram Percy the Borre style belt slide and strap end from the South Manor site were probably manufactured in Scandinavia and arrived in the Wolds with a Scandinavian settler. To judge from the quality of the craftsmanship his status was relatively high. It may be of importance that both these objects were found in a Middle Saxon enclosure which the presence of a timber hall and smithy suggests was already of high status. Indeed this argues that whilst the Scandinavian takeover brought about the laying out and nucleation of the village, the site which was to become the focus of the nucleated plan was already a focal point in the Middle Saxon period. Scandinavian overlordship was based upon continuity of political gravity. New Viking landlords were also active in assimilating Christianity and erecting private chapels in the lOth century (Richards 1991, 99-100). At Wharram Percy a small timber church was established in the valley in the 1Oth century (Bell et al. 1987). Amongst the earliest burials in the church are the Anglo-Scandinavian style graves, marked by limestone slabs with head- and foot-stones, of two adults and a child. These may therefore represent the first Scandinavian lords of the Wharram manor. Radiocarbon dates for the burials excavated within the church now confirm that many were buried in the 9th or 1Oth centuries (Clark pers comm). Anglian and Viking Settlement in the Yorkshire Wolds / /' \.~ 0 ········ .-j .~·. ' ·.: ..·.···:.·:·:::::·:···.-.::: ··.:. •.. ~ ·:, -- N -::-::'-- I- ,. ,. f ......... . ,., .· / I '\ ' ................... , ' ,,,, 'Cl ·. \ . I I \ \ \ I \ \ ' ' ' [QJ Excavated areas 0 0 8th· 9th c. coins m w [!] 6th·6th c. structures 8th c. pottery group 8th/9th c. cross fragment 8th/9th c. metal working .... Fig. 6.- Plan of Wharram Percy showing Anglo-Saxonfinds. (After C. Philo in Richards & Milne 1994, fig. 44). 241 J.D. Richards Finally, it should be noted that the early medieval activity at Wharram Percy was only discovered because the village had ultimately been abandoned settlement and became the focus of a large campaign to examine the desertion of the medieval settlement which had unexpected results. Cottam was only discovered because of the chance association of Anglian and Anglo-Scandinavian metal artefacts with cropmark enclosures. It is suggested that there may be many more early medieval settlement sites awaiting discovery in the Yorkshire Wolds. References BELL R.D., BERESFORD M.W. et al. 1987: Wharram: A Study ofSettlement on the Yorkshire Wolds, Ill. Wharram Percy: The Church of St.Martin, Society for Medieval Archaeology Monograph 11, London. BERESFORD M. & HURST J. 1990: Wharram Percy: Deserted Medieval Village, London, B.T. Batsford Ltd. I English Heritage. DIDSBURY P. 1990: Fieldwork in Cottam and Cowlam Parish, Yorkshire Archaeological Journal62, 6367. HALDENBY D. 1990: An Anglian site on the Yorkshire Wolds, Yorkshire Archaeological Journal 62, 51-63. HALDENBY D. 1992: An Anglian site on the Yorkshire Wolds, Yorkshire Archaeological Journal 64, 25-39. HALDENBY D. 1994: An Anglian site on the Yorkshire Wolds- Part Ill, Yorkshire Archaeological Journal66, HALL R.A. 1994: Viking Age York, London, B.T. Batsford Ltd. I English Heritage. HIRST S. 1985: An Anglo-Sa.x:on Inhumation Cemetery at Sewerby, East Yorkshire, York University Archaeological Publications 4. HURST J.G. 1984: The Wharram Research Project: Results to 1983, Medieval Archaeology 28,77-111. KEMP R. 1996: Anglian settlement at 46-54 Fishergate, The Archaeology of York 711, Council for British Archaeology, York. MANBY T. forthcoming: Thwing, Paddock Hill: A Bronze Age and Anglo-Saxon site in East Yorkshire. MILNE G. & R!CHARDS J.D. 1992: Wharram: A Study of Settlement on the Yorkshire Wolds, VII. Two Anglo-Saxon Buildings and Associated Finds, York University Archaeological Publications 9. MORTIMER J.R. 1905: Forty years' researches in British and Saxon burial mounds of East Yorkshire, London. MYRES J.N.L. & SOUTHERN W.H. 1973: The AngloSaxon Cremation Cemetery at Sancton, East Yorkshire, Hull Museum Publications 218, Hull. POWLESLAND D. 1986: Excavations at Heslerton, North Yorkshire 1978-82, Archaeological Journal143, 53-173. RICHARDS J.D. 1991: Viking Age England, London, B.T. Batsford Ltd. I English Heritage. RICHARDS J.D. 1994: Cottam Evaluation, Yorkshire Archaeological Journal66, 57-58. RICHARDS J.D. 1995: An incised chalk weight from the Anglian Site at Cottam, Humberside, Medieval Archaeology 38, 167-169. RICHARDS J.D. in prep: Cottam: Anglian and AngloScandinavian Settlement on the Yorkshire Wolds. STAMPER P.A. & CROFT R.A. 1997: Wharram: A Study ofSettlement on the Yorkshire Wolds, VIII The South Manor Area Excavations, York University Archaeological Publications 10. TIMBY J. 1993: Sancton I Anglo-Saxon Cemetery: Excavations Carried Out Between 1976 and 1980, Archaeological Journal150, 243-365. WATKINS J.R. 1983: The Archaeology of Anglian East Yorkshire: a review of some published evidence and proposals for future fieldwork, East Riding Archaeologist 7, 25-39. Dr. Julian D. Richards Department of Archaeology The University of York The King's Manor York YOI 2EP UK 242 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Kaname Maekawa A deserted medieval village and the formation of a fortified town in Cambridgeshire, England Summary In 1994 and 1995, our group conducted research on a medieval village in Cambridgeshire, England. Using various archaeological and geophysical techniques, I was able to come to a number of interesting conclusions about the relationship between the decline ofmedieval villages in the area and the creation oflarger,fortified communities. Most significantly, I now believe that the inhabitants of numerous medieval villages eventually chose to or were forced to migrate to larger towns which were situated in and around massively designed castles. Most of these castles were equipped with large baileys or house plots which helped to facilitate defence against sudden attack. Introduction What I would like to outline in this paper is the process of nucleation and urbanisation of a middle ranking town from a medieval village. The discussion is centerd upon a specific example: the deserted medieval village at Swavesey in Cambridgeshire. There are essentially three reasons for the choice of survey area in 1994 Fig. 1. - The location of Swavesey and the distribution of medieval sites. (After Ravensdale 1984) 1 2 243 Kaname Maekawa Fig. 2. - General view from the air (after Ravensdale 1982). study has been to shed light on it. Finally, the date of this site has been roughly established by the small trial excavation undertaken by Cambridge Archaeological Unit in 1990. Historical background of Swavesey Swavesey parish forms part of the south-western hinterland of the fen basin, flanking the southern bank of the Ouse and stretching to the Roman Road running between Cambridge and Huntingdon (A14). It covers 1611 ha (3,982 acres). The southern part of the parish is upland, rising to 18m on the west next to the A14, the Roman Via Devana, falling gently to a small residual area of fen at the north and west, divided into Mow Fen and Mare Fen. Swavesey was flooded in 1874 and 1947. On the latter occasion, the floodwater appears to have reached the 5.4 m contour near the village. The high ground of the parish consists ofKimmeridge Clay (Chris 1992). In the early medieval periods, Swavesey was divided into three parts. At the north end is the church and the manor house, in the central zone, the nucleus of the settlement and a castle, and at the southern end an extention of the settlement and the deserted medieval village which posing interesting problems of the origin of the town (fig 1, 2). In this paper, I shall focus on this point. In the central zone, the dock and the market which reflect a period of prosperity in medieval times have survived as a village square. However it is not clear Swavesey, which was occupied from the iron age to the medieval period as the subject of our survey. Firstly, through this example, we expect to be able to grasp better the history of developed medieval towns and settlements in the regional historical continuum. Secondly, there is a high possibility of grasping the relationship between large and middle-sized towns on the one hand, and villages on the ether. This is still little understood and one of the achievements of this p ~ y ~ r 1- 'r \ I " I ; I "( ~G 9 I G 6 f r t I I I I I G 3 ' . I ' I I I I I Fig. 3. - The grid and rough plan (after Maekawa K. et al. 1996). 244 I I 0 ~ 20 40 m Fig. 4. - Plot of Resistivity survey (after Maekawa K. et al. 1996). A deserted medieval village and the formation of a fortified town in Cambridgeshire, England why the manor house was not in this central zone. Some say that the manor house was originally situated in the central zone island and was later moved to the northern island. According to the documentary evidence, Count Alan, Lord of Richmond, held it in 1086. A market and fair at Swavesey was granted to the then lord of the manor de alan la zouche in 1244. This was enlarged to an eight day fair in 1261. It was probably at this time that the town defences and streets were laid out. Swavesey had a port by about 1177, and the town was connected to the River Ouse by a navigation drain connected to the docks in the market place (Evans 1990). Our survey was made in the corner of the so-called castle close Black Horse Lane (fig 3). The site is situated on the west side of the central zone of Swavesey village. Along the west side, an earthwork of2m high with a double moat of 8m wide survives in a well preserved state. The north west corner of the earthwork is higher than the rest, and is thought to represent the motte on which a dongeon may have stood. The area of the close probably therefore represents an inner bailey. To the north of it, the area of Topleys Close by Amen Corner may well represent a second bailey, since it also has a moat and an earthwork. At the moment it is used as pasture for horses. Between these two baileys, there is a road which changes direction at right angles. It was probably an important road in the medieval period, linking Swavesey and the adjacent village of Fen Drayton. A number of fragments in Farrer's Feudal Cambridgeshire indicate how Swavesey developed during the thirteenth century: Fig. 5.- Plot ofMagnetometry survey (after Maekawa K. et al. 1996). 0 20 1: 1230 Swavesey acquired by Roger de la Zouch by exchange; 2: 1232 Roger granted fifteen oak trees for making lodgings at his manor of Swavesey; 3: 1244 Grant to Alan de la Zouch of market and fair at Swavesey; 4: 1261 Further grant, and enlarging the fair to eight days at Michaelmas; 5: 1267 Grant to Alan whose corn at Swavesey had been burned by the King's enemies. Topographical key elements (Fig.1) The church and the manor The church and the manor house are situated at the north end of the village of Swavesey on a low gravel 'island' separated from the medieval town of Swavesey to the south by the Navigation Drain which formerly allowed boats to reach the village from the River Ouse. It is not known when there was first a church at Swavesey, but a priory was established there before 1086. The nave is markedly broader at the tower end than at the chancel arch, and this may reflect the lay-out of an earlier building. The nave of the earlier church was rebuilt around middle fifteenth century. Dock and Town layout Mr. John Shepperson has reconstructed and mapped the areas inundated in the great floods of the past hundred years in Swavesey. According to his reconstruction and aerial photographs, Church End emerges 40 m Fig. 6. - Plot of Susceptibility survey (after Maekawa K. et al. !996). 245 Kaname Maekawa IUH ~ "'-.._ r----- r-- r------ r--...._ !'--- ...._ r-r--~'-----...__ r--, r---r-- ._.., -....: ____ ....... :::::: ·-H-- ;--. ...... ---- ..... 'I-,.... , ...._ -- t:r--r-- ~ • I 1 G 9 G 6 G 3 I, ,. :.& .: l ,. l ·- \ , .,, ' • I l.IM Fig. 7.- Plot of Ground Radar survey (1) (after Maekawa K. et al. 1996). !. Anomaly points; 2. the grid. as one island, and the central zone of the village (the putative medieval town and former Hale field) as another island. What he suggests is a dock area with four basins, including Swan Pond and Town Pond, and two others in between on the edge of church Green. The heart ofthe town is the market place, where the barges would have come tounloadto the navigation drain. Inside the town defences, the street layout seems to correspond approximately to a grid. I.IH .. ! OH The castle and Town defences The putative castle mound at Swavesey has probably been much reduced by gravel-robbing, and as it remains is comparable only to the smallest motte. The castle may have been the third in a chain including Burwell and Rampton in response to the anarchy of Stephen's reign, however,there is also a possibility that it was earlier, perhaps late 11th century in date. 246 Fig. 8.- Plot of Ground Radar survey (2) (after Maekawa K. et al. 1996). A deserted medieval village and the formation of a fortified town in Cambridgeshire, England Fig. 9. - Plan of Wharram Percy and a building reconstruction. I. general view; 2. reconstruction, after Beresford M. & Hurst J. 1990). ·· .. N \ 0--' ·50 ---IOOmottes --·.""' ·- '! peasant tolls & crolls Mh Ch ·--- boundaries ol:manor house church Pa parsonage F fishpond uncertain use • • 2 At the present, the town ditch and the earthworks on the west side of the settlement alone can be seen. However, in the 13th century, it is likely that on the east side, they went up to the dock area and continued to the fen. Concerning the western part, in 1984 and 1985, an excavation of the moats was carried on the 247 Kaname Maekawa Fig. 10. - Synthesis of various geographical methods and archaeological implication (after Maekawa K. et al. 1996). Resistivity survey Suceptibility survey * Ground rader survey 60m + + (SS02) l -y -1 -1 -4 Ditch 3 ~ /-. '.,.•~ -.',-.·.-: -·.·_ . !,.f", / / ,' I .~ ~ :}' --1 '-I .,.., ' ( S D 0 ' •·.;::. :-.r-.... .,...:;;.,.:,: t- ~:;{di ~ / -.... llouse plot 3 40m ,..t- ;.:~~j.~;.. 1- +/: House plot 2 / ,.." Ditch 2 : D 0 2 ) ,._1- Road I + House plot )' + '* i\j; (SSOl) + Manor House 0 60•n south side at the end of the village, where flat bottomed moats of 2m deep and 8m wide were discovered. Remains of 13th century were found inside the earthworks. Until 15th century, the inside of the moat was regularly cleared (Haigh 1984). In addition, surface finds from this area have been analysed by Dr. Hall (Hall 1974). He concludes that they divid principally into two groups dating to the late iron age, and the 11th or 12th (ie. the early medieval age). In addition, there were some finds of 14th or 15th century date. Dr. Ravensdale implies that the defensive works 248 including the bank and ditch and the castle were probably constructed in the last quarter of the thirteenth century. Ryder 's farm The farmhouse is situated outside the town wall, however, it has the socalled 'long cat slide' roof typical of an aisled hall, suggesting possible medieval origins. A small trial excavation by Cambridgeshire Archaeology suggests that Rider's farm may date back to the 13th century (Bray 1993). A deserted medieval village and the formation of a fortified town in Cambridgeshire, England o 100 lOO ....)00 feet !!Mlh+"·c.;+ '"""'=~=~=~ ' '' ' -Fig. 11.- Plan of Rampton castle (after Taylor C. C. 1973). Geophysical survey in the deserted medieval village Method (Fig. 3, 4, 5, 6, 7, 8) We conducted a geophysical survey which included a resistivity survey, magnetometry survey, susceptibility survey and ground radar survey (Maekawa K. et al. 1996). We set up a grid of 9 squares, each 20 m on a side, aligned with magnetic north. A pole in the south-west corner was established as the base point, and an x-axis to the east, and an y-axis to the north, with crossing points as shown. The grids and squares are numbered G 1, G2, G3 from the south-east corner and then GS, G6 from G4 next to G 1 and then again G8, G9 from G7 next to G4. Overall the surveyed area measured 3600 square meters. Around the site, the weather was dry and sunny for about one month up to including the survey period, therefore, the level of water under the ground was comparatively low. Results In the case of resistivity survey, taking account of the local geography, in the center of the cut-edged rectangular plot and circle plot, apparently higher anomaly was apparent. In each of grid squares GS and G8, a clean edged rectangular anomaly measuring 10 meters wide and 16 meters long was found. In the case of magnetometry survey, a sparse anomaly appered along the hollow which divides GS and G8, in the east corner of the rectangular block, and in the north west of G4. In the case of susceptibility survey, an anomaly exists along each of the hollows and on the south side of the rectangular block. As for the ground radar survey, rectangular anomalies can be seen on the top of each plot, as in G2, GS, and G8. In G3, G6, and G9 which are thought to be the rear part of each plot, no anomaly can be seen, in addition, suggest that clear anomaly can be seen in the recently filled ditch running from the middle of G7 to the southwest ofG4. Interpretation Let us turn now to the problem of interpretation, taking all the currently available archaeological material into account. 1) As can be seen from the sketch plan and aerial photograph, the landscape of this site is quite similar to that at Wharram Percy in Yorkshire which is famous as a deserted medieval village (Fig.9). At Wharram Percy, there was a Manor house at the north end of the village, and a church and parsonage at the south end. Between them, the main street ran and either side has rectangular house plots divided by ditches or earthworks. On the house plots, building remains fringe the road, and the area at the rear ofthe plot was given over to an agricultural use. The ultimate origin of this landscape may be sought in the Anglo-Saxon period and this type in general was established in the 249 Kaname Maekawa ·· ..... . IO~o _ _ _.;;...._ _ _ _ _ _ B _.:.o~oof"t c '>I'>.'. '_. • ' .. ·~. ' :,.._ _ _ _ __,.;;,s_o_ _ _ _ _..-,oo _,,.,, D + + I i Fig. 12. -Plan of Burwell castle (after RCHME 1972). middle of the 12th century. If we compare this case study with the results at D.M.V. site at Swavesey and aerial photographs, an intriguing possibility emerges. At Swavesey, underneath the ground where a farm house stands at the moment, it is possible a manor site has hurried, although the site of a church and parsonage is not discernible at present. Moreover, a road which is now not used runs from east to west along the exist at 20 meter intervals. Towards the rear of the plots, these features become less clear. 2) Secondly let us consider the implications of the geographycal results.The resistivity survey leads us to suggest that in the center of house plots, higher rectangular anomalies can be seen, and along the hollows resistivity becomes lower. In the case of the ground radar survey, anomalies partially arranged in a line can be seen. Taking account of the buildings excavated at Wharram Percy, it seems that medieval farm houses did not always have foundation stones on four sides, and there are examples which combine post holes and foundation stones. It seems reasonable to suppose that the anomalies surveyed mainly by ground radar represent buildings with postholes and foundation stones, which measured about 13 meters wide and 13 meters long in 02, about 5.5 meters wide 250 and 11.5 meters long in 05, and about 6 meters wide and 10 meters long in 08. 3) Thirdly, regarding the results of the magnetometry survey and susceptibility survey, the anomalies in the former case seem much higher in the south of09, in the west of05, between 01 and 02, and inside the presumed buildings. The anomalies in the latter case seem higher in 02, 05, 08 and in the hollow between 02 and 05. We may surmise that there is burned soil with magnetic mineral or iron materials such as nails there. 4) Lastly, taking account of the archaeological evidence of analogical examples in Britain, an attempt to synthesise a reconstruction is offered in Fig. 10. Here I have identified a hollow running north-south at the west end of the site as Road 2. I interpret the rectangular blocks as house plots. Outside of the presumed manor house, I take the hollow referred to as Ditch 1 to be a ditch which surrounded a possible manor house. The hollows which divide presumed House Plot 1 and 2, and 2 and 3 are referred as Ditch 2 and Ditch 3. I have termed the anomalies inside each of house plots 1, 2, 3 as Buildings 1, 2, 3. Because there are very few features behind the House Plots 1, 2, 3, we consider this was probably a cultivated area. A deserted medieval village and the formation of a fortified town in Cam!Jridgeshire, England Conclusions- Other examples in Cambridgeshire Burwell and Ramp ton (Fig. 11, 12) Burwell castle is firmly dated to the Anarchy of Stephen's reign, and Rampton castle (Giant's Hill) appears similar in plan to Burwell, with its wide, rectilinear moat. Rampton castle (Giant's Hill) is situated on the fen edge, and is believed to be the remains of an unfinished castle, dating from the mid-twelfthcentury Civil War ofKing Stephen's reign, and connected to the rebellions of Geoffrey de Mandeville. The castle has a rectangular mound with markedly curved south and east sides. It is surrounded by a deep flat-bottomed moat or ditch up to two metres deep. On the north side, it has three house plots which are surrounded by small earthworks. On the west side of it, it has three bailies divided by two hallows. The middle one has a house plot. On the northwest side, it has ridge and furrow. Burwell castle is situated on the fen edge as Rampton is. In its present form, it consists of a large rectangular moat, averaging 20 ft. wide at bottom and 9 ft. deep, isolating an area 200 ft. by 125 ft. On the north side, there are three house plots outside the enclosure which are surrounded by hollows. It is known that in 1143 Stephen ordered the building of a series of castles in this region to control the marauding Geofferey de Mandeville, and this is one of them. If Rampton and Burwell are from the same period, they might well be part of a system for commanding the ways that skirt the southern edge of the fens. What is more important here is the fact that both of the two have deserted medieval house plot forms around them as Swavesey does, and it is possible that this reflects nucleation initiated by a strong political power such as the King. Hypothesis Retuning to Swavesey again, it is evident that on the western side, the settlement is characterised by broader baileys surrounded by earthworks and a moat, on the eastern side, it consists of a town layout and docks. It might have been surrounded by moats for defensive purposes, however, no visible remains of such earthworks can be seen at the moment. The church and the manor house are isolated on the northern island. These facts may perhaps be in the assumption as follows. When the village at Box worth End was nucleated into Swavesey, there might have been a balance established between the old dock town and baileys with the castle mounds, these being dominated from a distance by the church and the manor house as a political center. The nucleation and the urbanization ofSwavesey in mid 13th century can be compared to the Incastellamento in southern part ofEurope which is, in general, considered not to have occurred in northern Europe including Britain (Wickham C. 1981). As mentioned above, field archaeology, geophysical survey and documentary evidence combine to suggest that the de la Zouch family redeveloped Swavesey as a commercial centre, re-siting the houses around the dock area at least in the 1240s. At the same time, the medieval village which was probably established before the Norman conquest, was nucleated into a new fortified town and eventually deserted completely. Acknowledgements The results of geophysical s.urvey at Swavesey discussed in this paper, were obtained by the joint efforts of those who participated in the Swavesey project. I am grateful to Professor Hideo Sakai, Professor Takao Uno (University of Toyama) and Dr. Simon Kaner (University of Cambridge). I would particularly like to thank Dr. Antony Clark (Consultancy in Archaeomagnetic prospecting, U.K.) for giving us the results of susceptibility survey. The final report of Swavesey Project is forthcoming and will be published by Oxbow books. Dr. Martin Morris (University of Chiba) was kind enough to check my English in this paper. I appreciate his kindness. References BERESFORD M. & HURST J. 1990: Wharram Percy Deserted Medieval Village, London. BRAY S. 1993: An Archaeological Investigation of a 13th Century Farmhouse in Swavesey, Cambridgeshire Archaeology. EVANS C. 1990: Archaeological Investigations at Swavesey, Cambridgeshire, Cambridge Archaeological Unit. HALL D. 1988: The late Saxon Countryside: Villages and their Fields, in: D. HooKE (ed.), Anglo-Saxon Settlements, London. HAIGH D. 1984: Excavation of the Town Ditch at Swavesey 1984, Proceedings of the Cambridge Antiquarian Society 73. HODGES R. 1995: Early Medieval Archaeology, London. MAEKAWA K., SAKAI H., UNO T. & KANER S. 1995: Swavesey Geophysical Survey at Black Horse Lane. MAEKAWA K., SAKAI H., UNO T. & KANER S. forthcoming: Swavesey, Oxbow Books. MAEKAWA K., SAKAI H., UNO T. & UEDA J. 1996: Geophysical Survey at Swavesey, England, News 251 Kaname Maekawa letter No. 17, Nara National Cultural Properties Institute (Japanese). RA VENSDALE J.R. 1982: History on Your Doorstep, London. RA VENSDALE J.R. 1984: Swavesey, Cambridgeshire: a fortified medieval planned markettown, Proceedings oftheCambridgeAntiquarian Society 72. Royal Commission on Historical Monuments in England 1972: North-East Cambridgeshire. SALZMAN I.F. (ed.) 1948: Ancient Earthworks, in: The Victoria History of the Counties of England. TA YLOR C.C. 1973: Cambridgeshire Earthwork Surveys, Proceedings of the Cambridge Antiquarian Society 54. University of London, Institute of Historical Research 1989: The Victoria History of the County of Cambridge and the Isle of Ely, Volume 9. WICKHAM C. 1981: Early Medieval Italy, London. WRATHMELL S. 1989: Domestic Settlement 2: Medieval Peasant Farmstead, York University Archaeology Publications 8. Kaname Maekawa Department of Archaeology Toyama University 3190 Gofuku Toyama City Toyama 930 Japan 252 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Udo Recker The Medieval Parish of Lohn Preliminary Report on a Research Project I Introduction Since the early 1950s, the appearance of the landscape in the western Rhineland between the cities of Aachen in the west, Cologne in the east and Monchengladbach in the north has changed dramatically. Large scale open-cast mining activities have destroyed a cultural landscape ('Kulturlandschaft') that had developed over for several hundreds of years. In the course of the last 40 years many villages and their boundaries have vanished and with them all material remams. As early as 1953, Bader and Hermbrodt drew attention to the danger linked with these procedures (Bader & Hermbrodt 1953). Nevertheless the local archaeological authority concentrated its activities on exceptional excavations like those at Morken (Bohner 1959; Hinz 1969) and the mottes and moated sites Husterknupp (Hermbrodt 1958) and Li.irken (Piepers 1981 ). Schwellnus stated that without any doubt the results of these occasional interventions were important but that one should never forget that in the meantime a dozen Rhenish villages and their contents have been destroyed largely unnoticed and much less documented (Schwellnus 1987, 113). In the mid-1970s, the idea won recognition that only a large-scale exploration enables us to gain an adequate insight into high and late medieval rural settlement structures. As a result the Rheinisches Amt for Bodendenkmalpflege (Rhenish Board for Archaeology) started a research project in the parish of Lohn. Although the realization of the original concept leave somewhat to be desired and the quality of the excavations gives the impression of rescue archaeology, the decision to try our hand at a project like this is to be welcomed. The parish of Lohn was located east of Aachen between Eschweiler and Ji.ilich. In medieval times it consisted of the six villages Erberich, Fronhoven, Hehlrath, Langendorf, Lohn and Pi.itzlohn as well as , of a manor called Hausen. With the exception of Hehlrath and one street of Fronhoven all these villages were demolished between 1973 and 1985. The Hausen manor has been rebuilt stone by stone in a suburb of Aachen. Up till now no final evaluations have been made. Only a number of preliminary reports have been published. The present report aims at providing an introduction to the research project and at presenting the initial results. II The "Medieval Parish of Lohn" - Project The project is meant as a contribution to the archaeology of rural settlements in the Middle Ages and Early Modem Times. It was set up in the Department of Pre- and Early History of the University of Bonn in 1993. Since 1996, the research work is financed by the 'Stiftung zur Forderung der Archaologie im rheinischen Braunkohlenrevier' ('Foundation for the support of archaeology in the Rhenish open-cast mining area'). The 'Medieval Parish of Lohn'-Project pursues several aims: on the one hand the archaeological evaluation of all features and finds, on the other hand the reconstruction of the development of a small-scale landscape ('Kleinlandschaft') in the western Rhine land in the High and Late Middle Ages as well as in Early Modem Times. Settlement structures should be analysed in their total cultural-historical range without any ostensible limitation. The oldest archaeological feature dates from the 9th century whereas the oldest finds go back to the 10th/11th centuries. The end of the 18th century with its world-shaking changes in the form of the French Revolution was chosen as the recent chronological limit of the project. The attempt to reconstruct medieval and early modem settlement structures and the development of the cultural landscape (' Kulturlandschaft") in high and late medieval as well as early modem times requires an interdisciplinary approach. Interactions exist with subjects like historical geography, history, art history, sociology and cultural anthropology. Whenever possible scientific analyses are used to back up the results. 253 U. Recker Archaeological aspects of the project So far, a total of 13 large-scale excavations and 7 rescue excavations or observations have been carried out. The largest group of archaeological finds consist of medieval and early modern pottery. Hence the evaluation of excavations was started with two pottery complexes from Lohn (Recker 1995) and Pi.itzlohn (Holtken 1995). A total of about 10,000 sherds and some complete vessels were analysed. Each fragment was examined according to a catalogue of 21 distinctive features. This way, 35 pottery ware types could be distinguished. Distinctions between the wares were made with regard to material and manufacturing technique, optical characteristics and optical criteria. No use was made of chemical and physical analyses. In addition, the shape of the vessels/fragments and - if present - the decoration were analysed. The results obtained were complemented with those obtained through the study of a third complex of about 7,000 sherds and a few vessels from Li.irken, a moated site located close to the western border of the parish. In doing so, a pottery spectrum of a period ranging from the 1Oth/11th to the 18th/19th centuries was covered. All ceramic remains from excavations within the parish will be classified according to this chronological framework. In a second step, all archaeological features are evaluated. By this means settlement structures should be reported/investigated per excavation in order to reconstruct the settlement structure of each village and of the whole parish. Scientific aspects of the project Supplementary to the conventional analyses of pottery, the determination of provenance is backed up by NAA (neutron activation analysis). At all excavation sites, wood was so badly preserved that it is impossible to get any dates by tree-ring dating. If possible, samples are analysed by radiocarbon dating. Bones will be subjected to an anthropological or osteological analysis. This may allow us to answer questions concerning human eating habits, nutritional conditions and diseases as well as those related to the kinds of domestic animals kept. Historical geographical and historical aspects of the project On April 25th, 1226 the archbishop of Cologne, Engelbert I von Berg, invested one of his vassals, the nobleman Gerhard von Randerath, with land in the Ruhr district. Four farmsteads are mentioned in this 254 source: 'Aldinhoven' (Aldenhoven), 'Ascwilre' (Eschweiler), 'Lon' (Lohn) and 'Inden'. His successor in office, Heinrich von Mi.illenarck, committed the fees to the convent of the Cathedral at Cologne. Since 1280, the parish has been under the wings of the sovereigns of Ji.ilich-Cleve-Berg. Several archives keep a multitude of written sources concerning the parish. Among them are parish registers, lists of peasants who cultivated land in Lohn and in the other villages of the parish, details on the location and the dimensions of plots, information concerning the types of grain or field crops grown, descriptions of the feudal duties, accounts, court records and so forth. Amongst other things, these sources will enable us to calculate the yields for selected years. Should the occasion arise, we will be able to reconstruct the correlations between the types of grain or field crops that were grown and the nutritional condition of the rural population. A category of sources of special interest are old maps of the parish and the surrounding area. The oldest remaining map in which Lohn is plotted dates from the year 1610. It shows the Duchies of Jtilich, Cl eve and Berg. The map was printed by Hessel Gerritz in Amsterdam and was originally part of the Blaeu Atlas. Similarly dated is a map printed by Claes Jansz Visscher. It has the same object and content as the first one and was also printed in Amsterdam. Several other maps of the 17th, 18th and 19th centuries are preserved. A French map drawn up in 1807 is of special importance. In 1801, the French Ministry of Finance decided to revise the law of taxation and gave the order to prepare a land register. In a first step 1,800 villages in France - including the western Rhine land- were selected to be surveyed. From 1803 onwards, all villages were surveyed. The French land register had not been entirely completed when the western Rhineland became part of Prussia in 1815. Hence it was continued by the Prussians from 1819 on. The map ofLohn was drawn to a scale of 1:5,000 and is the first exact plan of the village and its fields. Combined with the written sources, it will allow us to reconstruct the landed property of at least the 18th century. Whether it will be possible to go back into the 17th century or even further back is doubtful. Nevertheless an approximate distribution of the fields and the peasants related to them can be reconstructed by means of the field-names. Langerwehe, one of the major production centres of medieval and early modern pottery in Europe, is located only five kilometres southwest of Lohn. Because of this the parish has been a part of the local market for Langerwehe products. In this connection, it appears to be interesting to study in which quan- The Medieval Parish of Lohn. Preliminary Report on a Research Project tities pottery from other production centres occur in the parish. In the 16th and 17th centuries Lohn had its own court of lay assessors. The preserved court records afford an insight into everyday wranglings and rows. The court was appealed to in order to settle a sale of land, to call in hereditary tenancies, to regulate claims to inheritances and so forth. On the other side, the records shed light on the relationship between vassal or free peasant and the authorities respectively the feudal lord. The first mention ofLohn as an independent rectory can be found in the so-called 'liber valoris' of the early 14th century. Since the same century Lohn and the surrounding villages formed a parish. The ecclesiastical history of Lohn can be duplicated by papal documents of the 14th and early 15th centuries as well as by the registers of the archbishops of Cologne (Knipping 1909; Janssen 1973 and 1977). Ill Archaeological evidence The heart of the parish has always been the village of Lohn. This situation and the range of questions under consideration are good reasons to focus the following interpretations on this village. The centre of the settlement was formed by the parish church and a small market place from which the four main streets lead off. In addition, to these places another two spots within the village were of exceptional importance for the archaeological investigations. At the northern periphery of the village lay a plot of land that was called 'Alte Burg' ('old castle') or 'Brubbeleburg'. In the middle of the parcel, a hillock was located. This feature makes the fieldname even more interesting. The oldest building of the village was situated to the east of the market place. In 1678, the village was destroyed by a fire and according to oral history only three farmsteads were not burned down. It is said that the 'Hof Kaldenbach' (Kaldenbach farm) was one of these farmsteads. The parish church of Saint Silvester The excavations inside and outside the church have not yet led to a final and complete study. Nevertheless, it is possible to present a general review of the development of the church. The development can be subdivided into six phases: 1. a wooden structure of the 9th century; 2. a stone structure of the lOth century; 3. a church and steeple of the 12th century; 4. an enlarged structure of the 15th century which was destroyed in 1678; 5. a new construction of 1696 which Fig. 1. -Detail of sheets 77 (Aldenhoven) and 78 (Jiilich) of "Kartenaufnahme der Rheinlande 1:2 5000 durch Tranchot und v. Milffling "from 1805. was not changed for about 200 years; and finally 6. a new church of 1902/03 which was demolished in 1973. In the following the phases 1-5 are described; phase 6 is of no interest for the project. Phase 1 Beneath the nave of the modem church (phase 6), a wooden structure or rather eight postholes were uncovered. The excavator identified the structure as the first phase of the church (Dohm-Ihmig 1978, 542). According to her interpretation, we have to think of a church in form of a small hall with a rectangular choir. The hall was preserved to a length of about 7 m not including the choir. Because of the fact that the westem edge of the structure was disturbed, this does not tally with the original maximum length. Provided that the postholes have equal distances of nearly 2 m the hall should have been at least 8 m long. It was about 6.5 m wide. To the east, a choir of2 m in length and 3 m in width was attached. On top of undisturbed soil, a composition floor (Estrich) could be observed; it has been linked up with the wooden structure by Dohm-Ihmig (Dohm-Ihmig 1978, 544). The filling of the postholes contained a number of sherds ofredpainted pottery of the Pingsdorf-type and hard-fired grey wares. Inside and outside the church, several graves were excavated. At least ten of them relate to the wooden structure of phase 1. They are located east of the rectangular choir and all of them have a W-E-orientation. From a stratigraphical point of view, they are of the same age or somewhat younger than the wooden church. In the north-eastern corner of the choir, a posthole has been disturbed by the eastern part of a 255 U. Recker ·- ·-'" . -\, .,/ Fig. 2. - Phase 1 of the parish church (based on Dohm-Ihmig 1978, fig. 5) . 10m + >~- -~'-. i -· -·· i iI 11 Ii N 0 '""'""' grave, which means that the grave comes from outside the building. This grave and another nine burials are overlapped by younger walls which belong to the church of phase 2. Anthropological examinations are hampered by the bad preservation of the skeletons. There are several indications that the wooden structure was destroyed by fire. Two samples of charcoal that had been taken out of the postholes could be dated to the years 760 AD± 50 years and 850 AD± 50 years. According to this, the wooden structure was built in the first half of the 9th century. It can still not be said at which time it burned down. Whether we follow the interpretation ofDohm-Ihmig or not, there is one question that cannot be clarified. The first stone structure of the church (phase 2) dates from the 1Oth century. If we consider the durability of a wooden structure founded on postholes, the supposed end of the wooden structure does not coincide with the beginning of the stone structure. Phase 2 As mentioned before, the second phase of the church dates from the 1Oth century. In the first instance, a small hall of 3.3 m in length and 5 m in width was erected. In a joint of two walls, a fragment of a globular pot ('Kugeltopf) was found. A rectangular choir, 0.9 m long and 3.5 m wide, was attached to the hall. The preserved foundations are based on several layers of sand and gravel and were built with quarrystones, Roman tiles, limestone and sandy mortar. The lower parts of the walls contained a number of pieces of charcoal. At an undetermined point in time, this structure was rebuilt. Instead of the rectangular choir, a semi- 256 5m II I circular apse was attached to the church. At the same time or somewhat later, a larger hall was erected west of the entrance to the church. In the choir and apse as well as in the large hall, the same kind of rubble was observed. This seems to indicate that both parts of the building were erected at the same time. Phase 3 Further alterations can be demonstrated by archaeological and historical evidence. In the 12th century, the western close of the hall and the choir with the apse were removed. The latter were replaced by a larger choir with a semicircular ending; in the west a steeple was erected. The foundations of the new choir were based in part on the foundations of phase 2 which must have led to static problems in the following years. Because of this, two buttresses and six pillars were attached to the eastern end of the hall and the choir. It is impossible to ascertain the exact point in time when these measures were taken. The masonry was not interlocked with that of the hall. In general, the pillars were built in same way as the buttresses but their basal surface was only half the size and their foundations were not as deep as those of the buttresses. The same kind of stones and mortar were used. The dimension of the steeple base was c. 8 by 8 m. The quadratic structure was not attached to the westem edges of the longitudinal walls of the hall. Quite to the contrary, its eastern foundation lay 0.8 m west of the foundation of the western front of the hall. The soil between both foundations was undisturbed. In order to close the gap between the two buildings, the longitudinal walls of the hall had been lengthened in The Medieval Parish of Lohn. Preliminary Report on a Research Project Fig. 3. - Phase 5 of the parish church (taken from Dohrn-Ihmig 1978,fig. 3). a westerly direction. We owe a description of the steeple (Franck-Oberaspach & Renard 1902, 185) to the fact that it was not destroyed by the great fire in 1678. The Romanesque spire had three floors. While Dohrn-Ihmig (1978, 552) describes the building materials as quarrystones, Franck-Oberaspach and Renard (Franck-Oberaspach & Renard 1902, 185) speak of tuff. One could enter the church through a portal at the southern end of the steeple. On the first and second floor there were at least two windows with a round arch at each side. The portal and the windows were designed in Romanesque style. The roof had the shape of a frustum and was slated. Over the next centuries, the church and steeple were not changed or if something was changed it did not leave any trace in the archaeological record. Phase 4 In the 15th century, the northern longitudinal wall of the hall was removed and an aisle was added to the nave. Instead of the wall, three pillars were erected. At the eastern end of the church, a polygonal vestry was built. Moreover, the outside wall of the Romanesque choir was surrounded by a brick retaining wall. TI1e ground plan of this church was not changed over the next centuries. In 1678, French soldiers under the command of the Duke of Luxembourg invaded the Duchy of Jiilich-Cleve-Berg and burned down the village. Most parts ofLohn, including the church, were destroyed. Only the steeple withstood the fire. A new church was not erected before 1696, but as early as July 1679, one of three new church-bells was cast. Its inscription reads: 'S. Silvester I am called, 1678, October 4th, burned by fire, 1679, July 20th, totally indefatigable founded by Johannes Bourlet'. During the excavations in 1983, a large round pit was uncovered in which one of the three new bells had been cast. The pit was located close to the steeple of the 12th century and attracted attention due to the very heavily burnt soil surrounding it. The infilling of the pit contained a large number of burnt earth lumps as well. Phase 5 This brick-structure was erected in 1696 and referred to the ground plan of the burned-down church (phase 4). It consisted of a nave and a northern aisle with a steeple in the west. The windows of the nave and the aisle had a pointed arch, beyond that they were austere. The choir of the nave was semi-circular, that of the aisle was three-sided. East of the semi-circular choir, a polygonal vestry was located. A rectangular porch was added to the western front of the aisle. The overalllength of the building was 20 m and it had a width of 12.5 m. The ground plan of this church remained largely unchanged until its demolition in 1902. The 'Alte Burg' ('old castle') or 'Brubbeleburg' As mentioned earlier, the 'Alte Burg' or 'Brubbeleburg' was located on the outskirts of the village. 257 U. Recker The plot of land with the corresponding field-name lay about 400 m north-north-east of the market place and west of the radial road leading to Erberich. The plot- which was eventually used as a pasture- was excavated in 1983 (von Brandt 1985; Recker 1995). Amongst the hillocks, parts of a filled-in moat could be made out. It was observed to a linear extent of about 100 m. Although topographical and aerial maps of the area show a second moat south-west of the first one, it could not be located during the excavations. In total, an area of 8,300 m2 was examined. Even today the term 'castle' is still used generically for a variety of different objects. As early as in the 1960s, Sieper made a proposal for a three-part definition (Sieper 1962). According to this suggestion we have to differentiate between a fortified farmstead ('fester Hof), a fortified manor ('festes Haus') and a castle ('Burg'). Although this distinction presents many, it can still be accepted as a basis. If we follow Sieper's definition the 'Alte Burg' ('old castle') has to be identified as a 'fortified farmstead'. Except for the moat-system no other elements of defence could be observed. According to the features and the pottery spectrum, we have to distinguish four settlement phases. We can think of the first and fourth phases as being well-established, but the second and third phases and their full particulars are still in question. Phase I The first settlement phase can be dated to a period ranging from the 1Oth/11th to the 12th century. In this early stage the hillock of natural origin was fortified with an outer moat. A section of about 150 m of this moat could be uncovered. The preserved section runs from north to south and the section drawings show a two-phased filling. In a first phase, the moat had a pointed form, whereas in the second phase it was somewhat rounded. Because of the fact that the moat was not connected with a brook, it was only occasionally filled with water. East of this moat, three segments of a second inner moat could be located. This one was much smaller and was poorly preserved. Only a small number of pottery sherds could be found in the infilling of the moats. Most of them belong to the Pingsdorf-type wares. Because of the fabric, it can be considered certain that none of them were produced in the Vorgebirge west ofBonn. Particularly because of the temper in some sherds, they could be assigned to production sites in LangerweheJi.ingersdorf. Others may have been produced in South-Limburg (NL) or in East-Belgium. Both wares have been fired quite softly (Mohs 2-3), which makes it possible to distinguish them from a third group of Pingsdorf-type pottery that has been fired much 258 harder (Mohs 5). A fourth group of greyish, hard-fired (Mobs 5) earthenwares can be compared with pottery ofPaffrath-type or Elmpt-type. This pottery is wheelthrown but possesses elements of both wares. Some sherds have been fired so hard that on technical grounds one could identify them as proto-stonewares. The fabric is light grey and the shape corresponds with Pingsdorf-type vessels. It is assumed that this pottery is also of local origin. The range of different kinds of vessels illustrates that mainly cooking utensils occur. The so-called 'Tiillenkannen' (spouted pitchers) of the 12th century constitute an exception to the rule. Phases 213 Within these two periods several temporal gradations are possible. The final partition is still to be worked out. A group of older features within these two phases is represented by two west-east oriented rows of Grubenhauser. They disturbed the moats of the first phase. Each of the Grubenhauser contained a large amount of pottery mainly of the 12th/13th-14th/15th centuries. Most of the ceramic remains are yellow or greyish earthen wares, proto-stonewares or stonewares. It is possible that the Grubenhauser were linked with a fireplace or oven close by, but there is no indication as to what they were used for. The yellow earthenwares are represented by Pingsdorf-type vessels, but it can be considered that none of these were produced in Pingsdorf or in the Vorgebirge. Just like in the case of phase 1, some sherds could be assigned to production sites in Langerwehe-Jtingersdorf. This determination of provenance is based on the characteristics of the temper used. A second group of Pingsdorf-type sherds was produced in South-Limburg (NL) or East-Belgium. Both wares are baked very soft (Mohs 2-3). Among the finds are another two groups of hard-fired Pingsdorf-type pottery (Mohs 5-6). The third ware is comparable to pottery produced in the western Rhine land, South-Limburg or East-Belgium whereas the fourth ware has parallels among the pottery produced in the Vorgebirge. The greyish earthenwares found in the Grubenhauser can be subdivided into four groups. The first ware is represented by hand-made, hardfired (Mohs 3-4) pottery ofPaffrath-type. The pottery of the second and third wares is wheel-thrown and hard-fired (Mohs 5).1t is comparable to Paffrath-type and Elmpt-type pottery. A number of sherds has been fired so hard that they could be classified as very hard-fired earthenwares as well as proto-stonewares. The fourth ware can be identified as Elmpt-type pottery. The proto-stonewares and stonewares have mainly been manufactured in the Langerwehe- or The Medieval Parish of Lohn. Preliminary Report on a Research Project Cologne-Frechen-area. A small number of sherds can be identified as Siegburg stoneware. Difficult to classify in every respect are the orange-coloured leadglazed earthenwares. With the rise of hard-fired earthenwares and near-stonewares in the 13th century, only a slight change in vessel shapes can be observed. New shapes include tumblers and cans. A group of younger features is scattered over the whole excavation site. These can be dated to the 15th-16th centuries. The pottery spectrum is dominated by different kinds of stoneware. Possible production sites are Langerwehe, Raeren, Cologne or Frechen and Siegburg. Moreover there is a multitude oflead-glazed earthenwares. As mentioned before, the orange-coloured lead-glazed earthenwares are very difficult to classify. Green-coloured flat plates and dishes are typical products of the late 16th century. The fragments found in Lohn are probably from Frechen. The small number of earthenware cooking utensils can be interpreted as an indication of the use of metal pots and pans. Phase 4 This settlement phase can be dated to a period ranging from the 15th/16th to the 17th/18th centuries. In this period, a new rectangular moat was dug. Compared to the moat of the first settlement phase, its position was shifted to the west. Except for a small strip of its eastern section, the moat has been excavated completely. In the east, it was disturbed by the western ditch of the modem radial route. The flatbottomed moat was 8 to 16 m wide and had a maximal depth of 2.8 m. The area it enclosed measured 78 by 94 m. The questions of where and how an entrance to the interior zone was made possible could not be answered through archaeological means. The contour lines urge us to think of a causeway. In this case, two areas are suitable for such an entrance: one in the middle of the western section of the moat, a second in its south-eastern corner. South of the rectangular moat, sections of a second moat of the same period were found. This second moat was disturbed by a modem farmhouse built in the first decades of the 20th century and could not be excavated completely. As mentioned before the' Alte Burg' ('old castle') must have been a fortified farmstead. Only a few details concerning the moated area can be described. The features are restricted to the southern half of the excavation site. A rectangular cellar in the southwestern quadrant was built with brick stones and had a basal surface of3.5 by 9 m. Most of the brick stones were removed after the cellar was left open. Three brick rows and the landing of an outer staircase were found in situ. According to the pottery found in the infilling, the cellar can be dated into the 16th/17th centuries. East of the cellar, a paved track was uncovered. This was accompanied by two low walls. Typical of this settlement period is the occurrence of Langerwehe stoneware. More than 35 % of the pottery coming from the 'Alte Burg excavation was classified as stoneware of the Langerwehe-type. According to the shapes, mainly drinking-vessels and large storage-vessels were found. Furthermore, there are stoneware products coming from Raeren, Frechen and the Westerwald. In addition lead-glazed earthenware plates and dishes were found in large quantities. Polychrome decorated earthenware has been produced in the Netherlands, the Weser-Werra-area and the Lower Rhineland since the 16th century. In all probability, we have to take a local production into consideration as well, but most of the polychrome decorated pottery found in Lohn comes from the Lower Rhineland. Parts of the table-service are greencoloured and were manufactured in Frechen in the 16th/17th centuries. In contrast to phases 2/3, the orange-coloured lead-glazed earthenware can be dated in the 17th/18th centuries. Early yellowish-green-coloured earthenware was produced in the 18th century. Up to now, it is impossible to define an exact chronological frame for the brown-coloured lead-glazed pottery, which was found only in small numbers. The 'Hof Kaldenbach' (Kaldenbach farm) This farmstead is said to be one of the oldest buildings of the village. The earliest record of the 'Hof Kaldenbach' (Kaldenbach farm) dates back to the 15th century. In 1407, Gerhard von de m Berg was invested with the farmstead. According to Binding who investigated the historical building in the 1950s parts of it had been constructed before the great fire of 1678. He distinguished three phases. The kitchen, the laundry and two chambers on the first floor could be dated to the end of the 16th century. The cellar had a barrel vault and was constructed in the first half of the 17th century. The living-room and a bedroom upstairs were of the same date. In the 18th century, the parlour, a storage room and some bedrooms were erected. The remaining parts of the farmstead dated from the 19th and 20th centuries. In 1982, the buildings were investigated and surveyed. Afterwards the farmstead was demolished. The archaeological excavation started in the winter of the same year. There was no evidence of an older structure beneath the building of the 16th century. In the rear part of the plot, several pits, postholes and other features indicated that it had been in use at least since late medieval times. A trapezoidal cellar, 2.5-4 259 U. Recker by 9 m in size, contained a lot of hard coal, charcoal and fragments of brick stones. No parts of masonry were found. A second cellar was uncovered beneath a modern farm building. It was largely preserved and of rectangular shape. It had a floor space of 2.2 x 3.1 m. The walls were up to 2.1 m high and had been built with hewn stones, which were laid down very accurately along the inner sides of the walls. A floorcovering could not be observed. The access to the cellar was possible only from the outside. A small entry (0.35 m wide, 0.8 m high) with a threshold was located at the northwestern side of the cellar, but no remains of an outer staircase were found. According to the features, there had originally never been a structure on top of the cellar. The fillings of both cellars contained only small amounts of pottery, most of it coming from Langerwehe. Among the finds was a fragment of a Romanesque baptismal font, which may have been part of the interior decoration of the parish church (phase 3). It is uncertain what it was used for in a cellar at the back of a farmstead. Of importance are the relics of a drainage-system that drained off the rainwater as well as the waste water to a pond east of the farmstead. The drainagesystem and the pond were part of a late medieval and early modern sewerage. Within the village, waterpipes made of stoneware were used, whereas at the periphery open ditches were dug. The pond was built at the lowest point of the surrounding area and collected all the water from Lohn and Pi.itzlohn. The pottery spectrum confirms a settlement in late medieval and early modem times. According to the ceramic material a first phase can be dated to the 15th/16th centuries. It is followed by a second phase which corresponds to the oldest phase determined by Binding. sels have been analysed. The great number of pottery wares does not seem to coincide with the multiplicity of the pottery. Lohn as well as the entire parish is dominated by ceramic products from Langerwehe. About 60 % of all pottery was manufactured in Langerwehe. Some of the production centres in South Limburg (NL) and East-Belgium are also represented with large quantities of objects. Only a small amount of the pottery comes from Siegburg or from production sites in the Vorgebirge, in Westerwald or in South-Hesse. Only an extremely small number of pottery is of very high quality. Most of it consists of ordinary table-wares. A high percentage of the pottery is of poorer quality. Especially in the material from the 'Alte Burg' ('old castle') one can find a lot of evidence that pottery vessels of second-rate quality had been sold. We can infer from these facts aspects of the financial circumstances and the social status of the occupants of the 'Alte Burg'. It will be part of the problem to find out to what extent this statement can be generalised for the entire village or even for the whole parish. The demolition of the parish church in 1973 was the first visible sign of the eventual destruction of the village. Twelve years, later the parish and the medieval structure of the landscape were wiped out. The 'Medieval Parish of Lohn' -Project was planned to emphasise the importance of the archaeology of rural settlements in the Middle Ages and Early Modern Times. Acknowledgements The author is indebted to A. Brown, M.A., for reading the manuscript proof. Selected Bibliography IV Summary The parish of Lohn looks back on a thousand years ofhistory. The beginnings ofLohn date back to the 9th and 1Oth centuries. The wooden structure beneath the parish church and the 'Alte Burg' ('old castle') have been the initial cells. The excavations carried out in the parish in the 1970s and 1980s enable us to get a large scale impression of rural structures in high and late medieval times as well as in early modern times. Together with the preserved written sources, they allow us to have an inside look into the ordinary life of the medieval villagers. The chronological framework is based on the classification of the ceramic remains. So far, more than 20,000 sherds and a number of complete ves- 260 ASTON M. 1985: Interpreting the Landscape. Landscape Archaeology and Local History, London. BADER W. & HERRNBRODT A. 1953: Das Rheinische Braunkohlengebiet. Eine Landschaft in Not, Denkschrift des Rheinischen Vereins fi.ir Denkmalpflege und Heimatschutz, Neuss. BAUER I., ENDRES W., KERKHOFF-HADER B., KOCH R. & STEPHAN H.-G. 1993: Leitfaden zur Keramikbeschreibung (Mittelalter - Neuzeit). Terminologie - Typologie - Technologie, mit einem Beitrag von I. Endres-Mayser, Kataloge der Priihistorischen Staatssammlung Mi.inchen, edit. by H. Dannheimer, Beiheft 2, 2nd. edit., Kallmi.inz. BERESFORD M. & HURST J. 1990: Wharram Percy. Derserted Medieval Village, London. The Medieval Parish ofLohn. Preliminary Report on a Research Project K. BbHNER K. 1959: Das Grab eines frilnkischen Herren aus Morken im Rhein/and, Ko1n/Graz. VON BRANDT D. 1985a: Archii.ologische Untersuchungen in Lohn, Stadt Eschweiler, Kr. Aachen. In: Ausgrabungen im Rhein/and '83/84, Bonn, 205-211. VON BRANDT D. 1985b: Ausgrabungen und Funde 1983. Mittelalter und Neuzeit. Eschweiler, Kr. Aachen (11., 12.), Banner Jahrbiicher 185,497-499. VON BRANDT D. & GOEBELS J. 1984: Ausgrabungen und Funde 1982. Mittelalter und Neuzeit. Eschweiler, Kr. Aachen (1., 2.), Banner Jahrb. 184, 1984, 637-638. VON BRANDT D. & HERMANNS J. 1984: Ortskemuntersuchungen in Pi.itzlohn, Stadt Eschweiler, Kr. Aachen. In: Ausgrabungen im Rhein/and '81/ 82, Bonn, 208-210. VON BRANDT D. & SCHMIDT F. 1985: Ausgrabungen und Fun de 1983. Mittelalter und Neuzeit. Eschweiler, Kr. Aachen (5.-10.), Banner Jahrbiicher 185,497. D. VONBRANDTD., SCHMIDTF. &KEHRENG. 1985: Ausgrabungen und Funde 1983. Mittelalter und Neuzeit. Eschweiler, Kr. Aachen (4.), Banner Jahrbiicher 185, 496. DOHRN-IHMIG M. 1978: Ausgrabungen in der Pfarrkirche St. Silvester in Lohn, Kreis Aachen, Banner Jahrbiicher 178, 539-568. W. ERDMANN W., KOHN H.J., LODTKE H., RING E. & WESSEL W. 1984: Rahmenterminologie zur mittelalterlichen Keramik in Norddeutschland, Arch. Korrbl. 14,417-436. FRANCK-0BERASPACH K. & RENARD E. 1902: Die Kunstdenkmiller des Kreises Jiilich, Die Kunstdenkmii.ler der Rheinprovinz (Edit. by P. Clemen), Achter Band I, Di.isseldorf. HERRNBRODT A. 1958: Der Husterknupp. Eine niederrheinische Burganlage des friihen Mittelalters, mit Beitrii.gen von A. ZIPPELIUS, W. HERRE & G. SIEWING, Beihefte der Bonner Jahrb. Bd. 6, Koln/Graz. HINTON D.A. 1990: Archaeology, Economy and Society. England from the fifth to the fifteenth century, London. HINZ H. 1969: Die Ausgrabungen auf dem Kirchberg in Morken, Kreis Bergheim (Erft), Rhein. Ausgr. 7, Di.isseldorf. H6L TKEN Th. 1995: Die archilologische Dorfuntersuchung Piitzlohn (WW2), Kreis Aachen, unpubl. Magisterarbeit, Bonn. Institut ftir mittelalterliche Realienkunde 1984: Die Erforschung van Alltag und Sachkultur des Mittelalters. Methode - Ziel - Verwirklichung. Internationa!es Round-Table-Gesprilch, Krems a.d. Donau, 20. September 1982, VerOffentlichun- gen des Institus ftlr mittelalterliche Realienkunde Osterreichs 6 = Osterreichische Akademie der Wissenschaften, Phil.-Hist. Klasse, Sitzungsberichte, 433. Bd., Wien. JANSSEN W. (ed.) 1973: Die Regesten der ErzbischOfe van Koln im Mitte!a!ter. Fiinfter Band 1332-1349 (Walram von Jiilich), Bonn. JANSSEN W. (ed.) 1977: Die Regesten der ErzbischOfe van Koln im Mittelalter. Sechster Band 1349-1362 (Wilhelm van Gennep), Bonn. JANSSEN et al. 1979: Burgen aus Holz und Stein. Burgenkundliches Kolloquium in Base! 1977, mit Beitrii.gen von W. JANSSEN, W. MEYER, 0. 0LSEN, J. RENAUD, H. SCHNEIDER & K.W. STRUVE, Schweizer Beitrii.ge zur Kulturgeschichte und Archii.ologie des Mittelalters 5, Olten/Freiburg i.Br. KNIPPING 1909: Die Regesten der ErzbischOfe van Koln im Mittelalter. Drifter Band 1205-1304. Erste Hiilfte 1205-1261, edit. by R. Knipping, Bonn. KUNOW J., GIESLER J., GECHTER M., GAITZSCH W., FOLLMANN-SCHULZ A.B. & VON BRANDT D. 1986: Vorschlilge zur systematischen Beschreibung von Keramik, Kunst und Altertum am Rhein = Fi.ihrer RLMB 124, Bonn. MOMMSEN H., KREUSER A., WEBER J. & BOSCH H. 1987: Neutron activation analysis of ceramics in the X-ray energy region. In: Nucl. Instr. Meth. in Phys. Res. A257, 451-461. MOMMSEN H., KREUSER A. & WEBER J. 1988: A method for grouping pottery by chemical composition, Archaeometry 30, 451-461. MOMMSEN H., BEIER Th., HEIMERMANN D. & HEfN A. 1993: Charakterisierung der hochmittelalterlichen Keramik aus Langerwehe-Ji.ingersdorf durch Neutronenaktivierungsanalyse, Nearchos 1, 93-98. PERLMAN I. & ASARO F. 1969: Pottery analysis by neutron activation, Archaeometry 11, 21-52. PFOTENHAUER A. 1992: Dorfbegrii.bnis. Von Hii.usern und Menschen aufKohle, Monumente, Magazinfiir Denkmalpjlege in Deutschland 2, Heft 7/8, 34-38. PIEPERS W. 1981: Ausgrabungen an der A/ten Burg Liirken, mit Beitrii.gen von S. CORSTEN, H. HAMPERL, L. JONAS, F. KRETSCHMER, G. MULLER, H. REICHSTEIN & Th. SCHREIBER, Rheinische Ausgrabungen 21, Bonn. RECKER U. 1995: Die mittelalterliche undfriihneuzeitliche Keramik der Ausgrabung .. Alte Burg" in Lohn, Stadt Eschweiler, Kreis Aachen (WW5), unpubl. Magisterarbeit, Bonn. SCHWELLNUS W. 1977: Ausgrabungen auf dem Rittergut Hausen bei Lohn, Stadt Eschweiler, Kreis Aachen. In: Das Rheinische Landesmuseum Bonn, Sonderheft Januar 1977. Rheinische Ausgrabungen '76, Bonn, 161-163. 261 U. Recker SCHWELLNUS W. 1980: Hochmittelalterliche Siedlungsreste am Ortsrand von Fronhoven, Tagebau Zukunft-West. In: Das Rheinische Landesmuseum Bonn, Sonderheft. Rheinische Ausgrabungen '79, Bonn, 306-307. SCHWELLNUS W. 1987: Archaologische Untersuchungen in Dorfem des Rheinischen Braunkohlenreviers. In: Dorfer und Stadte. Ausgr. im Rheinland '85186, Bonn, 113-124. SAUERLAND H. V. 1902: Urkunden und Regesten zur Geschichte der Rheinlande aus dem Vatikanischen Archiv. Erster Band 1294-1326, Bonn. SAUERLAND H. V. 1912: Urkunden und Regesten zur Geschichte der Rheinlande aus dem Vatikanischen Archiv. Sechster Band 1378-1399, Bonn. SAUERLAND H.V. 1913: Urkunden und Regesten zur Geschichte der Rheinlande aus dem Vatikanischen Archiv. Siebenter (Schluss-)Band 14001415, Bonn. SIEPER W. 1962: Die Burg und ihre Bedeutung im Rahmen des mittel- und nachmittelalterlichen Befestigungswesen. In: Burgen und Schlosser 3, Heft 2, 37-54. STEINBRING B. 1996: Die Kleinfunde von der Hauptburg der Matte Alte Burg in Liirken, Stadt Eschweiler, Kreis Aachen, unpubl. Magisterarbeit, Bonn. Udo Recker, M.A. Rheinische Friedrich-Wilhelms-UniversiUit Institut fiir Vor- und Friihgeschichte Regina-Pacis-Weg 7 D- 53113 Bonn Deutschland 262 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Bruggc 1997' Conference- Volume 6 Birgitta Berglund Changes in the Power Structure around A.D. 1100 on the Northern Norwegian Coast. The importance of waterways and of the organisation of trade in building and in maintaining power 1 Introduction 2 Rural centres in North-Western Norway Along the coast in Helgeland in northern Norway I have identified several rural centres from Early Medieval times (Berglund 1995). They were politically independent chiefdoms before they became part of the Norwegian kingdom in AD 1000-1200. Then the king and the church built up new centres of their own. Some of the old centres disappeared totally, while others were soon used by the state to manage power in this province. In the province ofTnmdelag, south ofHelgeland, the king and church evidently used the old centres for their own purposes. There the continuity in power is much stronger than in Helgeland. In the present paper, I want to show how the waterways and the organisation of the trade influenced the different directions in managing power in the two neighbouring provinces. The result was, however, the same: the areas were incorporated in the Norwegian kingdom. In a broader perspective, these northern areas became part of the normal form of European societies at that time. The two provinces are known from the Icelandic Sagas. Helgeland in particular is also known for its large estates during the period A.D. 1700-1900. Both provinces are rich in medieval monuments, but archaeological research on the centres and their environment mostly concentrated on Helgeland. In the medieval town ofTrondheim (celebrating its 1000th anniversary this year, 1997) in the province ofTnmdelag, big excavations are however carried out. The Icelandic Sagas tell us much about the chiefs from Helgeland and Tmndelag as well as about the competition between the chiefs themselves and between the chiefs and the Norwegian kings. Other written sources from abo\}t 1430 onwards tell us a lot about the administration and land ownership in the area, not in the least about the holdings of the archbishops. The archbishop had his court in Trondheim ('Nidaros' in papal language). Along the northern Norwegian coast and along the Arctic Circle, in the province of Helgeland, I have identified five centres from the Viking Age the roots of which go back to the Early Medieval period. The two largest centres were politically, ideologically and economically independent areas comparable to small kingdoms. Here I want to stress Tj0tta which is one of these two most important centres. Tj0tta is situated at the mouth of a big fjord thus forming an important cross-road with the old maritime highway along the coast ('Norway' literally means 'the way north'). The fjord provides easy communication with the nomadic hunting people of the mountain areas to the east. Outside Tj0tta lies a big archipelago. The old maritime/sea-highway runs along the coast. This situation gives Tj0tta a strong position in the exchange of goods, trade, piracy and the collection of taxes. According to what Snorre Sturlason writes in the old Icelandic Sagas, the Viking chiefHarek came from Tj0tta. Snorre tells us that in A.D. 1030 Harek mobilized a whole army for the big battle ofStiklestad in Tmndelag, 400 kilometres south of Tj0tta. From that time onwards, the northern Norwegian coast was more closely tied to the kingdom ofNorway. The year 1030 has a strongly symbolic meaning in Norway because, according to the Sagas, the Christian king Olav Haraldson was killed in the battle of Stiklestad by another one of the chiefs who was fighting for the independence of the northern areas ofNorway. Afterwards the king was canonised as Saint Olav. King Olav was buried in Trondheim, about 90 kilometres south of Stiklestad. From that time onwards, the town became a destination for many pilgrims in northern Europe. The battle of Stiklestad was then regarded as the victory of Christianity in Norway. I have studied what the formation of the Norwegian state meant for the old centres in the north of Norway, especially in Helgeland. The king needed an apparatus to develop and maintain his power in the north. I have found that the king, and with him the 263 B. Berglund church, did not build up their new centres in the same areas as those where the old centres were situated. The resistance in the old centres was too strong. In contrast to the old ones, the new centres were specialised. The king had the real power and therefore he could have separate centres for administrative, religious and military functions. TjeJtta lost its function as a politically independent centre when the area became a part of the Norwegian kingdom and later (after 1450) of the kingdom of Denmark-Norway. But it survived as the centre of a big estate. Soon the king started to use the local landowner for his own purposes. Tj0tta became part of the biggest estate in Norway, owned by the ruler of the country under the Danish/Norwegian king, who reigned from Copenhagen. Around A.D. 1350, the ruler of the country gave the estate to the archbishop as a token of gratitude for the help the latter had given him in connection with his pilgrimage to Rome, prompted by the Black Death. In The Land-register of Archbishop Aslak Bolt from about 1432, Tj0tta is the farm with the highest income derived from land fees. Areas on the coast and in the fjords were attached to Tj0tta, according to the land-register. In later land-registers, we find that important islands with seal-grounds had also become attached to Tj0tta. For most of the period form A.D. 1500 to 1900, Tj0tta was the biggest estate in northern Norway, with many dependent farms from which it gained its income. Today, Tj0tta is the biggest farm in northern Norway. It is owned by the Norwegian state and used for experimental farming. The traditional specialisation in this area includes the cultivation of grassland and sheep-breeding. Recently, experiments with the cultivation of salmon have been started. (Today this region is one of the main production areas for salmon in Europe.) The roots of the large Tj0tta estate from the last century go back to Early Medieval times. I shall have a closer look at the role of the waterways, harbours and trade in the management of power in and over Tj0tta through the centuries. 3 Waterways, harbours and trade - Conditions for the power of Tjotta Tj0tta and the archipelago around it has been a landscape of power for at least fifteen hundred years. Tj0tta is the dominant island with extensive areas of arable land. Cultivation is marginal due to the short summer season. Tj0tta has a good, self-draining sandy soil, which is well-suited for cultivation. Hundreds of 264 islands surround Tj0tta. On most of them people had settled with their domesticated animals. On these islands there are large areas of grass, which grows quickly in this humid and relatively warm climate. Seaweed and fish remains were also used as fodder for the livestock. The sea of course provided fish, but also large sea-animals such as seals and whales. The area has many kinds of sea birds which provided eggs, meat and down. The archipelago also has plenty of good landingplaces. This is very important, both for the use of the resources from the sea and for trade and communication, and not in the least for the control of it. So the natural environment around Tj0tta is almost perfect for building up and maintaining power. The traces of this early medieval activity are still clearly visible mainly as impressive grave-mounds. They demonstrated the power ofTj0tta to the people passing by boat and to the people living in the archipelago outside Tj0tta. Both the size and the form of the graves were illustrations of power. In one corner of one of the large cemeteries, several boat-shaped graves form a fleet. Tj0tta wanted to demonstrate its ability to operate an entire fleet. In Early Medieval times and in the Viking Age, the houses of the most important site on Tj0tta were arranged around an open area like a court-yard. Several excavations have been performed on the site (Berglund 1995). This arrangement of the houses must have been planned and ready at the time when the first house was built. The houses are large ones with thick walls of turf. At one of the two entrances the houses almost look like a wall. This entrance is marked by a standing stone. At the other entrance, the site is shut offby means of a wall of turf. The place must have looked like a small castle, hidden between the hills standing on each side of the site. From these hills, early detection of arriving ships- whether carrying enemies or expensive cargo - was possible. The trade in hunting products from the far North had to pass close by as did the trade with hunting products from the mountains to the East. There is no reason to doubt the ability of Tj0tta to take advantage of this position. According to the Icelandic Sagas, the chief at Tj0tta often collected taxes from the Saami people. Due to the fact that the sea-level was higher when the court-yard site was built, it was possible to have harbours in two directions, one facing the outer archipelago anj the sea-lane to the West and one facing the fjords to the East. The most important reasons why the site was established exactly at this place were certainly the harbours and the possibility to look out over the sea from the hills. In an archipelago like this, it was important to have large ships and good harbours. No remains of Changes in the Power Structure around A.D. 1100 on the Northern Norwegian Coast 0 0 0 0 . a JAMTLAND N r 0 50 100km Fig. 1- The provinces, centres and other places in the provinces ofHelgeland and Tmndelag in Norway mentioned in the paper. 265 B. Berglund D (§ TJl?JTT~Furm-m~und s l t l :""" vs e Gravefie!d \? Gruvefield 6 " Standing stones N 0 100 I 200 300 m Fig. 2- The island ojTjf!tta with some of the most important ancient monuments. ships have so far been found in this area but a few harbours are known. Other imposing ancient monuments preserved on the island ofTj0tta are the courtyard site, cemeteries with standing stones, the farmmound (' terp ') with the stone church and two large boathouses. A few years ago, during diving operations, we found another harbour with a jetty protruding into the sea. This harbour faces the sea-lane and the archipelago west ofTj0tta. People who were sailing along this sea-lane must have seen it. A group of cooking pits are probably connected with the harbour. They are dated to the Viking period. In addition to the pits, the harbour consists of the jetty, two landingplaces and steep cliffs. It would have been possible to land here at high tide. Around A.D. 1000 the centre was moved from the court-yard site to the place where the farm is still located today. Here there is a big farm-mound site with thick cultural layers. I have excavated parts of the mound (Berglund 1995). The move to the new place 266 was probably made because as a result of the land rising, the distance to the harbours became too great. While the court-yard site was hidden between the hills, the new place was located fairly high up and its location was imposing, with the flat cultivated landscape in front of it and the mountains behind. From here it was easy to control traffic, both by boat and over land. When people from the outer islands came to Tj0tta with the products they paid to the landowner for the permission to live on the islands, they could land in the harbour on the western side. This was the shortest possible distance by boat. After landing they had to walk about 1.3 kilometres to the farm. That walk must have seemed long, leading through the flat landscape before the farmhouses of the landowners were reached. These houses were located on a higher and more imposing level. The new site also had a harbour facing the fjords and the eastern mountain-areas. During the 18th and 19th centuries, the Tj 0tta estate derives most of its income from farms in large Changes in the Power Structure around A.D. 1100 on the Northern Norwegian Coast Fig. 3 - The court-yard site at Tjetta with at least 12 visible house-grounds and 17 cooking pits. areas of northern Norway. Many of these farms were situated close to Tj0tta. It was typical for the northern Norwegian estates to have farms both in the outer archipelago and in the fjords. In this way they were economically independent, because they had all the goods and food they needed. From the fjords they got products such as wood, soapstone, tar and furs, and from the archipelago came maritime products such as fish, sealskin, food, seal-oil for lamps, and eggs and down from seabirds. Both the archaeological and the written sources make it clear that the estate had implemented this organisation this long before these recent centuries and probably already in Early Medieval times or at least in the Viking Age. On the islands around Tj0tta, I have identified several farm-mounds and carried out archaeological excavations in a number of them. These mounds cover smaller areas than the site of the big centre ofTj0tta. The archaeological material is quite different and of a much poorer character than that from the sites on Tj0tta. In the small settlements, there were for example hardly any iron objects, no imported material such as medieval ceramics, and only few examples of handicraft. And the food came from the areas near the dwelling-place. In contrast, the Tj0tta site yielded much iron, imported ceramics and other material, including products from different handicraft and food coming from a larger area. The imported material from Tj0tta shows that it was this centre which organised the trade. One of the reasons why a farm like Tj0tta could dominate the surrounding sites was the fact that it was easily accessed by boat. It was very difficult to hide in the surroundings or escape from the woodless archipelago when a threatening boat from Tj0tta appeared. Of course Tj0tta had a variety of positive and negative sanctions to use, but here I stress the importance of the waterways, harbours and trade. 4 Changes in the power structure around AD 1100 I have found that in Helgeland the king and the Church did not use the old centres of power to establish their own power. Instead they introduced their own centres. These centres were established on minor farms with low income and without big gravemounds and other symbols of power from Early Medieval times. As an example of the new centres in Helgeland, I want to stress the parsonage Alstahaug with its medieval stone church. Here the great baroque poet ofNorway, Petter Dass, was priest around A.D. 1700. He died 290 years ago, in 1707. Alstahaug was a very rich, if not the richest priesthood in northern Norway. In Alstahaug the soil hides thick cultural layers and I have excavated there during several seasons, the last one in 1992. The excavations revealed that much activity took place here from the 12th century onwards when the 267 B. Berglund church was built. It is interesting to note that we could not find any traces of human activity before that time. So the parsonage and the church were established at the same time in an area without any previous settlement. My opinion is that with his church organisation the king established this place as an ideological centre to break down the old Viking centres in the area. I think this had to be done because the resistance against the king was so strong. Before I explain this opposition, I will, however, first take a look at Tmndelag, the province south of Helgeland. In Tmndelag, the large fjord, 'Trondheimsjjorden ', bordered with arable land, is quite different from the landscape of Helgeland. Here we find the most important rural centres in the Early Medieval times far from the outer coast, but still within the fjord areas. There are also concentrations of rich finds from the Viking Age at the mouth of 'Trondheimsjjorden' (Sognnes 1988). A distinct and marked change in the powerstructure in Tmndelag around A.D. 1000 is that a town, known today as Trondheim, was established at the mouth of the river 'Nidelva' in the centre of the fjord area. Trondheim celebrates its first millennium this year, in 1997. The town was founded on a site which was not known as a centre in earlier times. However, it is located near the farm Lade, known from the Icelandic Sagas as the residence of the earls of Lade in the Viking Age. At that time, these earls controlled the trade in northern hunting products according to the Sagas. Nothing in the archaeological material from these centres contradicts the idea that the centres in Helgeland could organise their own trade. This is also supported by other stories in the Sagas. Historians and archaeologists (among the most recent Christophersen 1994) have discussed whether the town was established by the rulers ofLade or as a competitive centre to Lade. I have just started small excavations at Lade and hope that they will contribute to clarify the role of that farm. I think archaeological excavations at Lade are important in finding out whether the Sagas tell the truth. In Tnmdelag, the old centres certainly continued as centres even when the area became part of the Norwegian kingdom. The king and his church obviously used the old centres for their own purposes. There are several settlements with remains of strong symbols of power both before and after the creation of the Norwegian kingdom. In Helgeland, there was a marked break in the symbols of power at that time. There the king and the Church let other centres, different from the old ones, develop in order to exercise the power on their behalf. Why this great difference between Tmndelag and Helgeland? Perhaps it was easier for the old centres 268 in Tmndelag to demonstrate loyalty to the central power because the king and the earl often had their residence there. According to the old Icelandic Sagas, resistance was as strong in Tmndelag as in Helgeland. However, the centres of power in the North had more to loose by becoming part of the Norwegian kingdom. As independent chiefdoms, they could manage independent politics and trade with the northern hunting products. As a part of the Norwegian kingdom they had much to surrender to the central power without getting much in return. The foundation of Trondheim can be seen as an opportunity for the old centres to get an effective distribution of the iron produced in Tmndelag and in Jiimtland, the province east ofTmndelag in presentday Sweden. For Jiimtland too, the 'Trondlzeimsjjorden' has the nearest ice-free harbour. A traderoute over the mountains was already established here before A.D. 400 (Slomann 1950). The bloomery iron-making was going on far from the sea-lanes and the trade-route there (Magnusson 1986; Stenvik 1991 ). I think that the old centres controlling the bloomery iron-making welcomed a central apparatus for the distribution of the iron. Ships with iron and other products such as grain and wood from the inland areas ofTmndelag had to pass through the mouth of the dangerous 'Trondheimsjjorden' to come out to reach the main sea-lanes. There, the ship could easily be attacked. With the protection from the king, such a trade must have become safer. The big medieval harbour of king 0ystein, 'kong 0ysteins havn ', testifies to the fact that the mouth of 'Trondheimsjjorden' was an important strategic place (Jasinski 1995). Many of the old centres in Tmndelag certainly must have been happy with the establishment of central power and in this they felt quite differently from the old centres in Helgeland. There the king was a competitor for the northern hunting products rather than a protector of trade and transport as he was in Tmndelag. 5 Conclusion The extreme Atlantic coastal landscape around Tjotta encouraged the development of a centre of power there in Early Medieval times. The easy communications by boat helped the centre to control both the people and other resources. The centre placed people in the surrounding areas to harvest the rich resources. The position at the cross-roads between the sea-lanes and the mouth of a major fjord was very important for the development of the centre. The fjord it made possible to communicate with and control the Changes in the Power Structure around A.D. 1100 on the Northern Norwegian Coast /- " !Northern \ ( areas l \ "-._ -r J / I -l I :L :L I Vl <( <( :L <( w lz Vl LL 0 Vl IW ~ 0 0 g: :I: Vl LL IRON, GRAIN SOAPSTONE DOWN, EGGS IRON, GRAIN -l) c;i L~I L ~ .ZI w3 (/') ~-1 ~I Vl Z ~ g ~ w ~I ~~ :;ex w ~I ffi3 • LL ~I ~~ (/') Ow £1 t;;;c ~.1. ~ ~ ,.Y- 3 <! " 1/Southern\ \ areas \. ) I .......... Fig. 4 - Outline of local and long-distance exchange and trade organised by the centre of Tjetta. 269 B. Berglund resources of the nomadic hunting people living in the mountains. The position along the coast also made possible control over sea-trade. In the inland area, such a concentration of power was impossible. By A.D. 1000, Tj0tta had become established to the point that the ruler Harek - according to the Icelandic Sagas - played a substantial role in the conflict between the king and the aristocracy over the northern resource-areas. To maintain this power, the centre used different positive and negative sanctions. Organisation of trade could be used as a positive sanction. Threatening with weapons was a negative sanction. Imposing monuments, such as large graves (including grave-mounds, cairns and standing stones), specially shaped dwellingplaces, the imposing location of the dwelling-places and large harbours were used to keep the people subservient to the ruler of the centre. The image of power was strengthened by the visibility of these monuments to the many people passing Tj0tta. Later the king and his church organisation built up new centres in competition with the old ones, because the resistance against the Norwegian kingdom was so strong. In that manner, a centre such as Alstahaug was established. In Tmndelag, a town was established around A.D. 1000. Just like the new centres in Helgeland, the town was established at a site which did not belong to an early medieval centre . In Helgeland, the old centres along the coast strongly resisted the Norwegian kingdom since they did not need the king's protection. Their position close to the sea-lanes and the trade there made it possible for the chiefs to be independent of a king's protection. The old Helgeland chiefs saw the king as a competitor for the resources ofthe area. In the inner fjord area in Tmndelag, the situation was different. Here the chiefs had the advantage of the protection of a king when travelling with an expensive cargo on the dangerous 'Trondheimsjjorden'. In the inner fjord area in Tmndelag, the king was a co-operator when it came to resources and their transport. References BERGLUND B. 1995: Tjotta-riket- En arkeologisk undersokelse av maktforhold og sentrumsdannelser pa Helgelandskysten fra Kr.f tit 1700 e.Kr., Dr. philos avhandling, Fakultet for arkeologi og kulturhistorie/ A VH, Universitetet i Trondheim. CHRISTOPHERSEN A. 1994: Kaupangen ved Nidelva, Riksantikvarens Skrifter 7, Trondheim. JASINSKI M. 1995: Kong @ysteins havn pa Agdenes. Forskningsstatus og reviderte problemstillinger, Viking, Oslo. MAGNUSSON G. 1986: Lagteknisk jarnhantering i Jiimtlands !an, Jiirnkontorets Bergshistoriska Skriftserie 22, Stockholm. SLOMANN W. 1950: lvfedelpad og Jiimtland i eldre jernalder, Univ. i Bergen. arbok 1948, Hist.antikv. rekke 2. SOGNNES K. 1988: Sentrumsdannelser i Trondelag. En kvantitativ analyse av gravmaterialet fra yngre jernalder, Fortiden i Trondheim bygrunn, Folkebibliotekstomten. Meddelelser 12. Trondheim. STENVIK L. F. 1991: Iron production and economic Booms during 2000 years, in: A. ESPELUND (ed.), Bloomery ironmaking during 2000 years, Vol. I, Trondheim. Dr. Birgitta Berglund Norvegian University of Science and Technology Museum of Natural History and Archaeology Institute of Archaeology Erling Skakkes gt. 4 7b 7004 Trondheim Norway 270 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Julio A. Perez Celada Horticultura y molinos de agua en el curso medio del rio Carrion en la edad media El rio Carrion constituye la principal corriente de agua de la actual provincia de Palencia - perteneciente a la Comunidad Autonoma espaiiola de Castilla y Leon -, a la que recorre pnicticamente en su totalidad de Norte a Sur. Situado en la vertiente septentrional de la Cuenca del Duero e integrado en el sistema hidrognifico de este rio, el Carrion vierte sus aguas en el rio Pisuerga, tras nacer en la Cordillera Cantabrica, atravesar Ios paramos de raiia palentinos - formados por derrubios de aquella - y acceder despues, en su sector oriental, al ambito de las campiiias- ambitos arcillosos miocenicos consecuencia de la erosion fluvial de Ios paramos cabireos meridionales -que configuran la denominada tradicionalmente Tierra de Campos 1• Sin entrar en la consideracion de las virtualidades economicas de la Tierra de Campos, que hicieron de ella el principal "granero" cerealistico de la Cuenca del Duero, queremos remitimos a Ios caracteres que informan el valle del rio Carrion y a !as actividades productivas que se desarrollaron en el mismo a lo largo de la Edad Media y han perdurado en buena medida hasta nuestros dias. Cabe, en primer lugar referirse a la circunstancia de que el Carrion, como el Cea y el Pisuerga, ha excavado una cuenca planiforme que entraiia la peculiaridad de presentar una ribera oriental elevada sobre el curso del rio mientras que la accidental se caracteriza por su planitud, lo que, unido al desnivel que salva desde su nacimiento, constituira el fundamento material del temprano e intensivo aprovechamiento economico de sus recursos hidniulicos mediante la derivacion del mismo de cauces de agua artificiales. Asi, desde que la documentacion medieval comience a hacerse explicita hallaremos multitud de noticias relativas a explotaciones horticolas, herrenes, plantaciones de lino, arboles propios de terrenos hUmedos como chopos, sauces y frutales y, en fin, ingenios molineros 2• El terreno sobre el que estas realidades se asentaran sera elllamado "de Vega", configurado por arenas y materiales aluviales arrastrados por el rio y adecuado, por ende, para su explotacion agricola. En estos terminos, conviene aclarar que nuestro estudio se limita a las vegas de Saldaiia y Carrion, situadas en ese "curso medio" del rio a] que se refiere su titulo, quedando excluido, por su contextura historica diferenciada, el tramo bajo del mismo, en el que se asienta la ciudad de Palencia - capital de la provincia -, a muy pocos kilometros de su afluencia a! Pisuerga. Por lo demas, e insistiendo en lo dicho mas arriba sobre la indole de las riberas del rio, podremos constatar como el aprovechamiento hidraulico se realiza de manera casi exclusiva en la vega accidental, a excepcion de Ios tramos correspondientes a Nogal de las Huertas y Poblacion de Soto y, mas al Sur, a Cestillos y Villanueva del Rio. Por lo que se refiere alas determinaciones climaticas que han operado historicamente sobre el ambito geografico en que se enmarca el objeto de este estudio, debe seiialarse que, ademas de !as implicaciones de su continentalidad - frecuentes heladas desde octubre-noviembre hasta abril-mayo, veranos calurosos - sobre el logro de Ios cultivos a que nos acabamos de referir, el regimen de lluvias se caracteriza - a causa, en buena medida, del "efecto pantalla" ejercido por la Cordillera Cantabrica- por su escasez, lo que da lugar a un caracteristico paisaje arido. Las pocas precipitaciones tienen lugar especialmente durante la primavera y el otoiio, presentando una acusada irregularidad y produciendose con relativa frecuencia de forma torrencial. Estas circunstancias nos ponen ante la realidad cotidiana que hubieron de afrontar los colectivos humanos que poblaron las vegas de Saldaiia y Carrion; una realidad Sobre las caracteristicas hist6ricas y geognificas de este espacio son de obligada consulta Ios trabajos de VACA LORENZO A., La estructura socioecon6mica de la Tierra de Campos a mediados del siglo XIV, Pub/icaciones de la !nslilucion Te//o Tellez de Meneses 39, 1977, 229-398, y 42, 1979, 203-387, y: La Tierra de Campos y sus bases ecol6gicas en el siglo XIV, Studia Historica. Historia Medieval X, 1992,149-185. 271 J.A. Perez Celada consistente en intensos estiajes y ocasionales inundaciones que, como veremos con mas detenimiento, llevaron a aquellos a extremar el control sobre Ios cursos artificiales de agua y a sufrir frecuentes enfrentamientos en su seno relacionados con el aprovechamiento de un bien tan escaso. En efecto, y principalmente cuando llegaba el verano, cualquier presa o boquera construida recientemente en el rio Carrion o en cualquiera de Ios canales de el derivados, mermaba el caudal que Ios beneficiaries tradicionales del agua recibian segun un riguroso reparto de porciones aplicado a! sistema de irrigacion de !as riberas desde su inicio en el termino de Poza de la Vega hasta su conclusion en el de Villoldo 3 • Antes de pasar a otras consideraciones conviene sefialar que el nombre de Carrion aplicado a! rio proviene del de la villa homonima - hoy Carrion de Ios Condes -, circunstancia que se explica porque en esta, que primero se llamo Santa Maria, existia un pequefio camino o carria que permitia salvar el escarpe existente en la orilla oriental del rio y cruzarlo4. Cuando Ios condes saldafieses decidan trasladar la capital de la circunscripcion que gobieman a Santa Maria a fines del siglo X y, media centuria despues, se construya, por voluntad de uno de Ios ultimos representantes del linaje condal, un puente sobre el rio en dicho punto, quedara consolidado el caracter de etapa significada de la villa en el Camino de Santiago, a! que se sumara, en razon de la importancia de Ios recursos agricolas en general de la zona en que se asienta, el de principal nucleo de poblacion del ambito delimitado por Ios rios Cea y Pisuerga durante la Edad Media. Por otra parte, el establecimiento a partir de 1076 de monjes cluniacenses en el monasterio de San Zoilo, situado, con el rio de por medio, frente a la villa de Carrion, convertira a esta en un centro de irradiacion ideologico-religiosa. El "riuulo maiore" de que hablan !as fuentes del siglo XI se ira identificando paulatinamente con el nombre de la villa situada a orillas de su curso central. A titulo de ejemplo, citaremos un testimonio documental significativo en la zona: el rey Alfonso VII reintegra en 1127 el monasterio de Nogal de !as Huertas a la jurisdiccion del de Sahagun y lo hace, entre otras cosas, "cum tern's et uineis, pratis, pascuis, paludibus, are is, montibus, fontibus, molendinis, sesicis molendinorum, aquis aquarum, cum aqueductibus earum, arboribus fructuosis et infructuosis ... " (FEIU'lANDEZ FLOREZ J.A., Colecci6n diplomatica del monasterio de Sahagzin, IV (1!/0-1199), Lean, 1991,docno 1230). 3 En 1719 las Ordenanzas relativas a! regimen de explotacion del agua en la Ribera de San Zoles, que se basan en costumbres existentes cuando menos 500 afios antes, sefialan que "suele acaecer Ios mas de Ios aiios traer poca agua el rio de esta villa yen razon de e//o haber algunas diferencias en la lama del agua de el para dicha ribera de San Zoles y la de Nogal" (LALANDA 272 La problematica historica del uso del aqua En una sociedad como la feudal, en la que se encuadran Ios elementos humanos que ocupan !as vegas objeto de nuestra consideracion, la disponibilidad del agua del rio, de Ios terrenos que la misma riega y de Ios ingenios molineros que mueve, se halla sometida a determinaciones que se manifiestan en la capacidad diferenciada de !as distintas clases sociales para poseer en propiedad y/o usufructuar dichos medios de produccion. Sin animo de abordar la problematica del feudalismo iberico - porque ello trasciende ahora nuestras pretensiones - estimamos imprescindible, en todo caso, referimos someramente a esta cuestion en relacion con el tema que tratamos. En primer lugar, cabe preguntarse sobre la instancia que tiene la capacidad de decision sobre la atribucion de aprovechamiento de Ios recursos hidraulicos. Si se parte de la consideracion del agua como "res communis", como "bien comunal" 5, habra que convenir, en principio, que son Ios organismos concejiles- y, en ultima instancia, Ios reyes- quienes se hallan facultados para regular, en tanto que representantes de Ios distintos nucleos de poblacion, el uso del agua. Pero para fijar esta cuestion en Ios terminos adecuados conviene remontarse a! proceso de "repoblacion" de este territorio. Tras la desarticulacion politica y demografica del mismo causada por la invasion musulmana, desde el siglo IX, y procedente de la comarca montafiesa de La Liebana y su entorno, descendera de modo continuado una corriente de repobladores a traves, especialmente, de Ios pasos de Cervera de Pisuerga y de Ios situado a! Este de Ios mismos que se iran instalando en la zona bajo el regimen de "presuras" de tierra individualizadas 0 colectivas, dirigidas o no por nobles o monarcas. Paralelamente, a lo largo del valle del Carrion- como en otros ambitos de la vertiente septentrional de la Cuenca - se configura una sucesion de comissa o demarcaciones militares articuladas por fortalezas CARROBLES P., Las vegas de Saldafia y Carrion, antecedentes historicos de sus regad[os, PITTM 36, 1972, 193). Por lo que se refiere a las avenidas del rio, ya en 1349 se hallan atestiguadas inundaciones en el monasterio de San Zoilo de Carrion, situado en la rib era derecha del rio: "...jluvius intra/ sepia prioratus, nisi . remedium apponatur, ecclesia et claustrum minantur ruinam" (ROBERT U., Etat des monasteres espagnols de l'Ordre de Cluny aux XIIIe.-XVe. s1ecles, d'apres les actes des Chapitres Generaux, Boletin de la Real Academia de la Historia XX, 1892, 398. 4 DIEZ ASENSIO J., Substrato prerromano en la toponimia palentina: Calahorra, Carrion, Saldafia, Tamara, Aetas del !I Congreso de Historia de Palencia I. I, Prehistoria, Arqueologia e Historia Antigua, Palencia, 1990, 738; GONZALEZ J., Cuestiones de repoblacion en tierras palentinas, en: Palencia en la Historia, Palencia, 1982, 56. Horticultura y molinos de agua en el curso media del rio Carrion en la edad media Villosilla o Poza de la Vega o Csna/ o Rio Nuevo (s.XVI) • Canal de la Parionda Pago del Nido del Cuervo Villaturd~e Villanueva de loa Nabos Villotilla \ • Villamez 0 Ctwal del Monasterio de San Zoilo Canal da lztln o Abanzilas 8 Carri6n da Ios Condes ........ .... .... • Cestillos ...... ........ Cane/ de Cest.i/Jos Torre de Ios Molinos El .... "' .. 0 ....~~ .... Ob,~, ....~l' .... ...... ....'{~go 'P t2 •• 11 Reconstrucci6n hypotetica del sistema principal de canales en la Edad Media. 273 J.A. Perez Celada que servinin para proteger Ios asentamientos humanos. Desde principios del siglo X cuando menos, un comes - o conde -, teorico delegado del monarca astur-leones, gobernani estos distritos. Pese a la condicion de removible del cargo por el rey que tiene el titular del condado, un linaje, el de Ios Banu-Gomez o "descendientes de Gomez", que no por casualidad coincide con la familia mejor dotada patrimonialmente en la zona, monopolizani la dignidad condal hasta el siglo XII. La escasa capacidad de la Corona para hacer efectivo su dominio sobre el conjunto del reino dotani a los condes Banu-Gomez de una gran capacidad de intervencion sobre el espacio de su demarcacion, que abarcani !os distritos de Saldafia, Entrepefias y Carrion. La importante ocupacion de tierras por campesinos ha dado lugar a comunidades aldeanas que tienen un gran protagonismo en la puesta en explotacion de Ios recursos. De este modo, estos colectivos - pero tambien los situados bajo la ferula de Ios poderosos - comienzan a aprovechar con regularidad las virtualidades productivas que las vegas de Saldafia y Carrion brindan, labrando canales, construyendo molinos o plantando huertos, y ello bajo la atenta mirada- si no de la induccion- de los condes, de sus delegados o de Ios componentes de una incipiente clase nobiliaria. Las obras de apertura de acequias y de construccion de molinos, asi como su mantenimiento, habnin de implicar severas disciplinas colectivas que las comunidades campesinas se impondran a si mismas o sufrinin como compulsion externa, y supondnin con frecuencia la coordinacion de los habitantes de diferentes nucleos de poblacion6 . Por lo demas, ya desde los inicios de este proceso estaba teniendo lugar otro que se orientaba hacia la captacion de presuras territoriales de los campesinos y el control de !as infraestructuras construidas por estos para su explotacion. Nos referimos al proceso de configuracion de las grandes propiedades feudales que ira menoscabando, desde el mismo seno de los grupos aldeanos o desde instancias ajenas a !os mismos, la pequefia propiedad campesina 7 • Conocemos algunos casos, correspondientes a Ios siglos XI, XII y XIII, que ilustran el movimiento en lo que se refiere al objeto de este articulo. En 1057 Ios campesinos de varias villas o aldeas situadas a algunos Km al SSE de Carrion poseian un amplio conjunto de tierras de regadio y tres grupos de molinos a lo largo de un canal - "rego ", "cornago"- que eran obra de sus antepasados; el canal constituia una derivacion del "riuulo maiore ", y las tierras y este mismo se nos presentan con toda probabilidad como resultado de presuras,· el aprovechamiento del agua, por su parte, constituia el ejercicio de un derecho derivado de su caracter de bien comunal. Pues bien, en el afio mencionado mas arriba, el conde G6mez Diaz - uno de Ios ultimos Banu-Gomez- se hani con la totalidad del complejo mediante una operacion masiva de compra de tierras y "raciones" o partes de Ios molinos que privara a este amplio grupo de villanos de su condici6n de propietario del mismo. En 1194, el monasterio de Santa Maria de Benevivere realiza una compra, de analogo corte a! de la anterior, mediante la cual incorpora a su patrimonio una importante numero de participaciones- "raciones" y "partes"pertenecientes a Ios campesinos de Poblacion de Yuso - Poblaci6n de Soto - en sus molinos de esta villa, llamados Redondillo, Hostio, Gordo y Afiar. En fin, en 1221 el monasterio burgales de Las Huelgas compra !as veces o participaciones de los propietarios del molino de Sopena, asimismo en Poblaci6n de Soto 8 . Se ha incidido hasta hace no mucho tiempo en el caracter de monopolio sefiorial de !os molinos, que habrian constituido uno de los fundamentos de la dominacion feudal 9 . En nuestro ambito de estudio cabe decir que tal cosa no es cierta. Existe hoy, en efecto, la coincidencia generalizada en que la documentaci6n revela que con frecuencia fueron !as Vease a este respeeto GARCIA DE CORTAZAR J.A., El equipamiento molinar en la Rioja Alta en Ios siglos X al XII, en: Homenaje a Fray Justo Perez de Urbel, t. I, Studia Silensia, Abadia de Silos, 1976, 387-405; GAUTIER DALCHE J., Moulin a eau, seigneurie, eommunaute rurale clans le nord de I' Espagne (!Xe-XI!e sieeles), en: Etudes de Civilisation medilfvale (IXeX!Je siecles). Melanges offerts a Edmond-Rene Labande, Poitiers, 1974, 337-349. PEREZ CELADA J A., El monasterio de San Zoilo de Carrion. 428. Sobre Ios derechos de uso del agua fundados en las presuras de la Repoblaci6n, GL!CK T.F., Islamic and Christian Spain in the Early Middle Ages, Princeton, New Jersey, 1979, 96 y 97. Formacion, estructura y decurso historico de tm seiiorio castellano-leontis (siglos XI-XVI), Tesis de Doetorado, Universidad de Valladolid, 1994, en prensa, Capitulos !I y Ill, y, del mismo: Notas sobre la ampliaei6n del patrimonio eondal en el siglo XI: una operaei6n multiple de eompraventa y permuta protagonizada por don G6mez Diaz y doiia Teres a en I 057, en: Aetas del JJ Congreso de Historia de Palencia, t. I!, Palencia, 1990, 415274 7 El proeeso, por lo que respecta a una zona no lejana de la que trata este estudio, ha sido magistralmente analizado por MARTINEZ SOPENA P., La Tierra de Campos accidental. Poblamiento, poder y comunidad del siglo X al XJJJ, Valladolid, 1985, 207-321. PEREZ CELADA J.A., Documentacion del monasterio de San Zoilo de Carrion (/047-1300), Burgos, 1986, doe. 3; FERNANDEZ L., Coleccion diplomcitica de la abadia de Santa Maria de Benevivere (Palencia) 1020-/561, Madrid, 1967, doe. 31, y LIZOArN GARRIDO J.M., Documentacion del monasterio de Las Hue/gas de Burgos (1!16-1230), Burgos, 1985, doe. 158. 9 BLOCH M., La historia rural francesa: caracteres originates, Barcelona, 1978, 229. Horticultura y molinos de agua en el curso media del rio Carrion en la edad media comunidades campesinas- e incluso Ios particulares - quienes edificaron Ios molinos y conservaron su propiedad- dividida en participaciones, veces o vices -durante largo tiempo en ocasiones 10 • El hecho de que la clase senorial se presente cada vez mas como la mayor propietaria de molinos- a! igual que sucede con Ios demas medios de produccion - es una circunstancia imputable a! proceso de configuracion de la gran propiedad feudal, verdadero pilar de la sociedad medieval, pero ello no puede considerarse un monopolio formal sobre Ios ingenios molineros 11 • A traves de mecanismos como la compra, la profiliacion, la maneria, la recepcion de donaciones o, en fin, de otros, coercitivos ono, Ios senores se iran haciendo con las partes o vices que Ios campesinos poseen en Ios ingenios construidos por sus antepasados o por ellos mismos y, en muchas ocasiones, a adquirirlos en su totalidad 12 • Por lo de m as, cuando Ios senores reciban atribuciones jurisdiccionales sobre la totalidad de Ios terminos de las villas de la zona, disfrutaran de una excelente plataforma suplementaria para convertirse en propietarios de Ios molinos como parece que sucedi6 en la villa de Lobera, donde Ios monjes de San Zoilo de Carrion, que recibieron la jurisdicci6n singular sobre la misma en 1129, sin que conste que por ello accedieran a la propiedad de Ios molinos, se presentan en 1213 como duenos de Ios mismos, que explota ahora el concejo mediante arrendamiento 13 • Sin embargo, vemos que, en el caso de Poblacion de Soto, situada bajo la jurisdiccion de Ios monasterios de Sahagun y Nogal de !as Huertas desde algun momento del siglo XI, existian herederos que poseian participaciones en molinos hasta que se desprendieron de las mismas mediante dos ventas, ya glosadas mas arriba, que tuvieron lugar en 1194 y 1221 y pusieron !as mismas en manos, respectivamente, de Ios monasterios de Santa Maria de Benevivere y Las Huelgas de Burgos 14 • En fin, las menciones de molinos se presentan ya desde el siglo XI en manos de campesinos que estan perdiendo su propiedad en beneficia de Ios senores. Por lo demas, vemos como Ios institutos religiosos adquieren, con frecuencia por donacion, molinos hasta el momento en manos de senores laicos sobre todo durante Ios siglos XII y XIII. Por lo que respecta a Ios huertos y plantaciones de regadio en general, senin tambien Ios senores quienes se presenten en la documentacion como Ios mayores propietarios, aunque podremos contemplar la persistencia de pequenas propiedades horticolas a lo largo de todo el periodo considerado 15 • El caracter comunal de !as aguas, y su consecuente administracion coordinada por Ios respectivos concejos - y no entraremos aqui en el proceso diferencial de oligarquizacion de Ios mismos -, constituira una realidad a lo largo de todo el medievo, aunque sea necesario referirse a Ios importantes condicionamientos a que esta sometida tal capacidad. Por una parte, son Ios monarcas quienes tienen la ultima palabra, por la via de la emision de privilegios o el dictado de 6rdenes, en el reparto de !as aguas. Por otra, el mencionado incremento de la propiedad feudal a costa de la campesina convertira con frecuencia a Ios agricultores en usufructuarios de propiedades senoriales, por lo que Ios concejos son sustituidos en ocasiones por Ios senores o por Ios concesionarios de molinos o huertos a titulo particular en la regulacion de Ios derechos de uso del agua 16 • El derecho individualizado de aprovechamiento del agua, en todo caso, se deriva de la condici6n de "heredado" en un lugar, esto es, de propietario o usufructuario de tierras. Esta circunstancia generani solidaridades, normalmente expresadas a traves de Ios concejos, que agrupan a veces a personas fisicas o juridicas con intereses objetivamente diferenciados. Por otra parte, son Ios monarcas quienes, en la practica, determinan la resolucion de !as diferencias importantes relativas al uso del agua 17 • Siesta habia experimentado una explotacion que se basaba en 10 PASTOR R., Resistencias y luchas campesinas en /a epoca del crecimiento y consolidacion de laformacionfeudal. Castilla y Lean, siglos X-XII!, Madrid, 1980, 56-60. 11 MARTINEZ SOPENA P., op. cit., 316 y 317. 12 PASTOR R., ibidem. 13 PEREZ CELADA J.A., Documentacion ... , does. 27 y 80. 14 Vease supra, nota 8. 15 La enumeracion de ejemplos resultaria farragosa. Baste citar el registro documental del monasterio de San Zoilo de Carrion- lndice de San Zoilo- depositado en el Archivo Historico Provincial de Palencia, que principalmente a partir de su fol. 144 ofrece casos, correspondientes principalmente al siglo XV, de pequefios propietarios que donan, compran o venden tierras en el entomo de Carrion; incluso disponemos del dato de un "moro"- un mudejar-, de nombre Hamete, vecino del barrio de Ios monjes, que vende a estos dos pedazos de tierra de vega - fol. 148 V 16 Asf, por ejemplo, el monasterio de San Zoilo de Carrion siempre actuo directamente en lo relativo a la defensa de sus privilegios relativos al agua, quedando e1 concejo que existfa en su barrio de San Zoles sistematicamente en la so m bra a este respecto. Hay ocasiones en que un concejo se inhibe por no considerarse concernido por un problema: el concejo de Poblacion de Soto, en efecto, se desintereso de un enfrentamiento entre el monasterio de San Zoilo y Ios arrendatarios de Ios molinos de Las Huelgas en dicho lugar surgido en 1398 (PEREZ CELADA J.A., Documentacion del monasterio de San Zoilo de Carrion (1301-1400), Burgos, 1987, doe. 335). 17 No obstante, Ios senores, en virtud de sus capacidades propietarias y/o jurisdiccionales pod fan deterrninar Ios derechos de uso del agua (GAUTIER DALCHE J., art. cit., 348), aunque en el ambito que tratamos no hemos verificado tal circunstancia. 0 • 275 J.A. Perez Celada sucesivos pactos explicitos o tacitos entre Ios usuarios que se iban agregando a lo largo del proceso de compactacion demografica de la zona, lo cierto es que a partir de fines del siglo XII comienzan a manifestarse tensiones por causa de la construccion de nuevas presas para captar agua del rio. Asi, en un momento no determinado de fines del siglo XII, una tal Maria Boiso, posiblemente una importante hacendada territorial en este espacio, ordeno abrir un "cornagus o canal entre San Martin del Obispo y Los Barrios, en la vega saldafiesa. El rey Alfonso VIII, requerido por algunos concejos, ordeno efectuar una pesquisa a! respecto al abad de Benevivere y a dos caballeros de la zona y, tras ello, a traves de su merino en Saldafia, clausuro dicho curso de agua. La razon que dio lugar a tal determinacion fue que la mencionada sefiora habia abierto el canal "non de· fuero neque de consuetudine, et nunquam ibi fuit apertus es decir, transgrediendo un uso consolidado - principalmente por parte de Ios vecinos de Gafiinas y Lobera - que la Corona se encargaria de restaurar. Muy poco tiempo despues, en 1203, veremos al abad de Sahagun ejecutando una pesquisa tambien a instancias de Alfonso VIII algunos Km rio abajo, concretamente en el entomo de La Sema, en relacion con una disputa entre el monasterio de San Zoilo de Carrion y Ios "homines de No gal de Las Huertas por causa de la apertura de una presa que estos habian realizado a la altura de dicho nucleo, lo que privaba al monasterio de sus derechos - que se remontaban cuando menos a tiempos de Alfonso VII - a la tercera parte del agua del rio que pasara allado de Nogal - las otras dos debian ir hacia la villa de Carrion - y que captaban mediante una presa colocada unos tres Km rio abajo de Noga! 19 • En realidad, estos "homines" parecian haber actuado por propia iniciativa, al margen de sus senores, Ios monasterios benedictinos de Sahagun y San Salvador de Nogal, que lo eran desde 1131, asi como propietarios del pago de Nido del Cuervo y de terrenos en La Sema, unos tres Km a! norte de Nogal. En Nido del Cuervo - situado en la orilla derecha del rio- nacia un cauce 11 1118 , 11 18 19 20 PEREZ CELADA J.A., Documentacion ... (/047-1300), doe. 68. Ibidem, doe. 72. AHN. Clero. Legajo 5328. Relacion de Ios fundamentos que liene este monasterio de San Zoil de Carrion para gozar la tercera parte de todo el agua que viniese por el Rio Mayor desde la presa del Nido del Cuervo ... (siglo XVII). (V er croquis del re- parto del agua desde la toma de Iz:in). Pero ya en 1395, en efecto, la situaci6n era identiea, eomo lo demuestra el concierto entre el Coneejo de Carrion y el monasterio de Benevivere (LALANDA CARROBLES P., art. cit., 196 y 197: articulo 226 de Ias Ordenanzas de la Ribera de Torre y Calzada -Izan). 21 Vernota19. 276 artificial, seguramente el de Izan o Abanziles, en el que en calidad de propietarios tenian intereses territoriales Ios monasterios de Sahagun y Nogal, que aprovechaban dos boqueras del curso alto del mismo en el siglo XVII, una circunstancia que se puede remontar en principio a este momento 20 . Lo cierto es que Ios "homines de Nogal, aparte de haber cegado la toma de agua de este canal, habian construido a! otro lado del Carrion, a la altura de la villa, una presa que menoscababa el derecho de San Zoilo de Carrion a recibir su cupo de agua algunos Km mas abajo, pasada Poblacion de Soto y en la ribera derecha del rio. Tras la pesquisa, efectuada significativamente por el abad de Sahagun- otro de Ios damnificados por la actuacion de sus dependientes -, el rey prohibira que se hagan en adelante presas nuevas entre Nido del Cuervo y la villa de Carrion, con lo que Ios derechos de Ios usuarios situados bajo dicho pago quedaban por el momento salvaguardados 21 • Conviene recalcar esta nueva actuacion regia, que se concretara en la confirmacion del reparto antedicho y que sera reiteradamente aducida por Ios interesados y confirmada a lo largo de toda la Edad Media y de la propia Edad Modema por Ios sucesivos monarcas. Nos hallamos, pues, ante la evidencia de que, si no era posible un acuerdo directo entre !as instituciones y/o concejos, la monarquia tiene la capacidad ultima de decision 22 • Una vez que el rey adjudica un cupo en el uso de agua, su beneficiario se convierte de facto en propietario del mismo, aunque ello, como veremos, no evite en absoluto transgresiones de tal concesion en el contexto de escasez de agua que, segun indicabamos, caracteriza a este ambito. 11 Los canicteres de la red hidraulica y el regimen de explotaci6n de Ios recursos Las primeras menciones de estas infraestructuras hidraulicas tienen un caracter relativamente temprano y nos presentan por lo demas a !as mismas y a 22 Sobre esta euesti6n en el entomo de Burgos, ver CASADO ALONSO H., Senores, mercaderes y campesinos. La comarca de Burgos a fines de la Edad Media, Valladolid, 1987, 183-186. 23 PEREZ CELADA J.A., Notas sabre la ampliaci6n ... 24 BERNARD A. & BRUEL A., Recueil des chartes de l'abbaye de Cluny, t. IV, Paris, 1900, doe. 3507, donde se mencionan, entre otros bienes, "or/os, piscarias, molendinos, aquas,fontes, rivulos cum suis reductilibus, aquis aquarumque introitus et exitus". Sob re la donaeion a Sahagun del monasterio de No gal: HERRERO DE LA FUENTE M., Coleccion diplomatica del monasterio de Sahagun (I 07 3-11 09), Lean, 1988, doe. 912. Horticultura y molinos de agua en el curso medio del rio Carrion en la edad media los huertos y molinos en una situacion consolidada. En efecto, al testimonio citado mas arriba correspondiente al afio l 057, en el que observamos la presencia de molinos y canales al sur de Carrion en pleno funcionamiento y en dependencia de un curso artificial de agua y que deja claro que dicho funcionamiento se estaba produciendo al menos desde los alrededores del afio 1000 23 , cabe afiadir la mencion de molinos, canales y huertos que hace la carta de donacion del monasterio de San Zoilo a Cluny realizada por Teresa Diaz, viuda del conde BanuG6mez Gomez Diaz, correspondiente al afio 10761077, en la que se pone de manifiesto que los monjes disponian ya de un complejo de irrigacion y molienda. Por lo demas, ya en 1093 el monasterio de San Salvador de Nogal recibe, entre otras cosas, cuatro molinos a orillas del rio Carrion en ubicaciones no precisadas 24 . En estos terminos, y sobre todo por lo que se refiere al primero de Ios testimonios citados, estamos en condiciones de suponer con fundamento que, como enunciabamos mas arriba, el aprovechamiento del agua en Ios terminos referidos se venia produciendo seguramente ya en el siglo X y que los sucesivos aportes demograficos a la zona no habrian hecho sino colaborar en la ampliacion de la "red" hidraulica que habria comenzado a configurarse durante esta centuria, si no en la anterior. En fin, las colecciones documentales monasticas de la zona recogen durante Ios siglos XII y XIII bastantes testimonios de donaciones de molinos sefioriales y de sus derechos anejos sobre el agua 25 . Conviene ahora que presentemos a! lector una imagen mas perfilada del sistema de canales sobre la base de los datos concretos que nos ha brindado la encuesta documental que hemos realizado. Y, para ello, efectuaremos un recorrido en sentido norte-sur, es decir, comenzando por las primeras tomas de agua en la vega saldafiesa, situadas en las proximidades de Poza de la Vega, y concluyendo en el ultimo desagi.ie de los canales, a la altura de Villoldo. Las primeras noticias explicitas de que disponemos sobre las presas mas septentrionales, las de la vega saldafiesa, donde la inclinacion continuada del terreno resulta idonea a tal efecto 26 , corresponden a! afio 1285, cuando dofia Berenguela, abadesa de Las Huelgas, hace valer ante el monarca Sancho IV una pesquisa realizada tiempo antes por orden de Alfonso X 27 • En ella se consignan !as presas para tomar agua existentes desde Poza de la Vega- "la presa de Sob re Por;uela "- hasta la de La Serna, es decir, las correspondientes a la vega de Saldafia y a los inicios de la de Carrion. Las tomas de agua relacionadas en la pesquisa son !as siguientes en orden descendente: de "Sabre Por;uela ", de "Domingo Abbat ", de "!os Sauariegos ", de "Sarannan ", de Lobera y Gafiinas, de Moslares, de Renedo, de Nido del Cuervo y de La Serna. Esta enumeracion de obras hidraulicas creemos que ofrece una informacion fundamental respecto a la estructura del sistema de cauces de agua, que no experimentara, si exceptuamos la apertura del Rio Nuevo o de Los Molledos a la altura de Villosilla a principios del siglo XVI, modificaciones sustanciales con posterioridad 28 . En efecto, podemos hacer coincidir la primera presa - "Sabre Por;uela " - con la saca de agua del canal o "Puerto" de Matazorita mencionado en las Ordenanzas relativas al agua de esta acequia del afio 1630 y que se situa en Poza de la Vega, desaguando con posterioridad el caudal que toma en el Rio Nuevo 29 . La presa de "Domingo Ab bat" pudiera ser la que, sin ten er unas Ordenanzas especificas en la Edad Moderna, se ha conocido tradicionalmente como Rozas Viejas y que nace tambien en las cercanias de Poza 30 . La de "Sarannan" creemos que puede identificarse con la de Saldafia y constituir el canal con mayor desarrollo espacial en la vega de Carrion: la Perihonda, que, naciendo en el pago que sus Ordenanzas de l5671laman El Sedenal, se situaria en la ribera opuesta a la de la villa de Saldafia31 . Las de Lobera y Gafiinas, Moslares, Renedo y Lerones serian tomas de caracter menor destinadas a la irrigacion y molienda de trarnos de ribera restringidos32, y la de Nido del Cuervo, situada en termino de La Serna33 , es la que iniciaba el canal conocido como de Izan o Abanziles 34 . 25 tar que en el documento de 1285 citado supra se consignan ocho molinos escalonados entre Saldaiia- donde hay un molino del rey- y Lerones- que alberga dos-, concluyendo la enumeraci6n con una menci6n a Ios molinos de Torre, ya en el cuemago de Izan. 29 Ibidem, 175. 30 Ibidem, 143. 31 Ibidem, 165-168. 32 CASTRO GARRJDO A. & LIZOAIN GARRIDO J.M., op. cif., doc.39. 33 Sobre el pago de Nido del Cuervo, vid. FERNANDEZ FLOREZ J.A., op. cif., doe. 1389, y AHN. Clero. Legajo 5328. Relacion de Ios fundamentos. 34 AHN. Clero. Ibidem. Ver, por ejemplo, PEREZ CELADA J.A., Documentacion ... ( 1047-1300), does. 15, 27, 52 y 76; FERNANDEZ L., op. cit., doe. 16, y FERNANDEZ FLOREZ J.A., op. cit., does. 1204, 1248, 1365 y 1438. 26 Asi consideran las Ordenanzas de la acequia de Matazorita en 1630 a Poza y su entomo - "donde el rio se inclina y pone mejor disposicion" (LALANDA CARROBLES P., art. cit., 175) -, y asi puede verificarse sabre el terreno o mediante la contemplaci6n de un mapa topografico. 27 CASTRO GARRIDO A. & LIZOAIN GARRJDO J.M., Documentacion del monasterio de Las Hue/gas de Burgos (1284-1306), Burgos, 1987, doe. 39. LALANDA CARROBLES P., art. cit., 149 y ss. Conviene seiia- 28 277 J.A. Perez Celada Fuera de la enumeracion que hemos comentado se encuentra la toma o boquera que riega la ribera de Nogal y Poblacion de Soto y que se encuentra a la altura del primer lugar citado. En tomo a ella tuvieron lugar Ios enfrentamientos mas intensos y constantes por el uso del agua entre algunas de !as instancias sefioriales con mas peso en la zona y con un interes mas acusado en el uso masivo del agua: el monasterio de San Salvador de Nogal, a! que ya nos hemos referido, y, sobre todo, el de Las Huelgas de Burgos, con intereses ambos en la misma ribera, y el de San Zoilo de Carrion. En efecto, ya en 1203 nos hallamos ante el contencioso suscitado entre Ios "homines de Nogal y el monasterio de San Zoilo, que, sin embargo, no se reavivani hasta el siglo XVII 35 . No asi sucedera con Poblacion de Soto. En dicho lugar poseian !as monjas de Las Huelgas de Burgos una "domus o explotacion agraria destinada a! aprovechamiento del terrazgo en regimen de gestion directa, pero tambien a la recaudacion de rentas de !as explotaciones campesinas, y a! aprovechamiento de un conjunto de molinos. La "domus de Poblacion formaba parte de un conjunto de explotaciones de la misma indole dispersas por Ios amplios dominios de la institucion burgalesa. Aunque la primera mencion en la documentacion corresponde a! afio 1263, cabe suponer que la integracion en el dominio se produjo con anterioridad. Regida por "freyres cistercienses a !as ordenes de !as monjas de Burgos, la "domus o "palacio de Poblacion parece que orientaba sus excedentes a! mercado de Carrion, pues sabemos, por un documento de 1336, que a Ios vasallos dellugar se les obligaba a transportar !as rentas del mismo hasta dicha villa 36 . Pues bien, la institucion burgalesa mantuvo un autentico "tira y af!oja" en lo relativo a! uso del agua con el monasterio de San Zoilo a lo largo de mucho tiempo. Ya en 1285 pidieron !as monjas que se ratificara su derecho a disponer de su boquera, con el caudal de agua correspondiente, en la ribera izquierda del rio. Pero muy pronto, en 1305, la abadesa se quejara a! rey Femando IV de que "Ios monjes de Sanct Zoyl de Carrion e otros ommes que les embargan el agua a monte la presa de donna Sancha, sabre! de Nogar, e en otros logares (...), en manera que sus molinos non muelen nin se aprovechan del agua segund deuen una queja que cabe relacionar tanto con el interes del monasterio de San Zoilo por que su porcion de un tercio del agua reconocida en 1203 llegara integra a su presa - lo que quiza les obligara a menoscabar la de doiia Sancha- como con Ios de !as poblaciones de !as riberas del norte, en lo que parece que se trataba de un afio particularmente seco. Por lo que respecta a la presa de dofia Sancha que menciona el documento citado, hay que seiialar que debio de ser esta abadesa de Las Huelgas (12051230) quien la mando construir, una actuacion que transgredia a todas luces el ordenamiento del agua realizado por Alfonso VIII en 1203. Ahora bien, este hecho se explica si tomamos en consideracion que Alfonso VIII fue un significado protector del monasterio burgales, donde en union con su esposa se halla enterrado, y que no hubo de resultar dificil a !as monjes arrancarle a! monarca una autorizacion para abrir su propia boquera, precisamente mas abajo de Nido del Cuervo, a la altura de Nogal - "sabre! de Nogar ~~, como afirma el documento citado 37 . En 1398 sera el monasterio de San Zoilo el que reivindique sus derechos contra Ios de Poblacion de Soto. En efecto, parece que un clerigo y otras personas, arrendatarios del complejo productivo de !as monjas en Poblacion- circunstancia que indica que Ios "freyres" a! servicio de !as mismas habian abandonado ya la "domus habian realizado una toma de agua a la altura de Poblacion- es decir, otra mas abajo de la de doiia Sancha en Nogal- para alimentar el molino de Las Casillas, lo cual menoscababa el tercio de agua que le correspondia a San Zoilo. Ante la inhibicion del concejo de Poblacion, que no se sentia afectado por la actuacion de dos particulares, estos llegaron a! acuerdo con Ios monjes de que no se tocara la presa que habian hecho y que se rebajara, en cambio, la de dofia Sancha, situada rio arriba, una solucion que ha de responder a la ausencia de cualquier gestor del monasterio en Poblacion y a! desinteres que dicha institucion mostraba a estas alturas por sus propiedades en ellugar38 . En cualquier caso, en 1412 el monasterio de San Zoilo lograra que la presa del molino de Soto, que el nuevo arrendatario del complejo molinero de !as monjas, precisamente el merino de Poblacion, Juan Garcia, utilizaba a la altura de dicho lugar para alimentar el mencionado ingenio, fuera rebajada porque contravenia su derecho a un tercio del agua obtenido en 1203 39 . En 1416 Ios monjes carrioneses volveran a efectuar requerimientos, esta vez a Ios alcaldes de Carrion, porque rio abajo de Nogal se han erigido presas con ramas y cespedes que impiden que el agua llegue a su boquera y funcionen adecuadamente sus molinos 40 . 35 37 11 11 11 11 11 11 11 ; j 6 AHPP. In dice de San Zoilo, fol I 0 I. J.M. & GARCIA GONZALEZ J.J., El monasterio de LIZOAIN Las Hue/gas. Historia de un seiiorio cisterciense burgales (siglos Xff y X[[[), Burgos, 1988, 222-225. 11 -, 38 39 40 278 CASTRO GARRIDO A. & LIZOAIN GARRIDO J.M., op. cit., does. 39 y 181. PEREZ CELADA J.A., Documentacion ... (1301-1400), doe. 335. AHN. Clero. Legajo 5328, 31-VII-1412. AHN. C1ero. Legajo 5328, VIII/IX-1416. Horticultura y molinos de agua en el curso media del rio Carrion en la edad media Conviene ahora referirse a! canal de bin o Abanziles. Este curso, situado en la margen derecha del rio Carrion, se presenta como el principal de la vega de Carrion en la Edad Media. En un orden de sucesion de canales considerado desde el oeste, aparece en tercer lugar, tras el Rio Nuevo- que, como ya dijimos, se abrio a principios del siglo XVI - y la Perihonda. El Iz:in nace en termino de La Serna, en la orilla derecha del rio y fue concebido con el doble fin de aplicarlo primordialmente a! riego en su tramo alto y a la alimentacion de ingenios molineros en el bajo, concretamente en !os terminos de Torre de !os Molinos y Calzada de Ios Molinos. En 1285, la confirmacion de la pesquisa regia solicitada por !as monjas de Las Huelgas nos presenta a! Izan en estos terminos: 'et la presa del Nido del Cuervo faganla !os de Poblap'on e de Nogar e !os de Poblar,:ion de Suso, e !os de La Serna den la terr,:era parte del agua que vaya por Castelanos ea Torre ea !os molinos del rey', lo que parece indicar que !os monjes de Nogal y Sahagun estaban en condiciones de exigir a sus dependientes en !os lugares mencionados prestaciones personales aplicadas a la construccion y el mantenimiento de la presa que se ubicaba en su pago de Nido del Cuervo 41 • Un compromiso establecido en 1395 entre el monasterio de Santa Maria de Benevivere, ocupado por canonigos regu1ares que seguian la Regia de San Agustfn y dependiente, en lo que se refiere al aprovechamiento del agua, del caudal que proporcionaba esta acequia de Izan, y el concejo de la villa de Carrion en relacion con los derechos de ambas instituciones sobre el mismo, nos informa de los cursos secundarios que se derivaban del canal por su ribera izquierda; una informacion que, por lo demas, es plenamente aplicable mas de tres siglos despues, cuando se aprueban, fundandose en la misma, las ordenanzas42. Asi, del Izan se derivanin una serie de canales o regatos secundarios que son, enumerados de Sur a Norte, los siguientes: el primer ojo se situaba inmediatamente el norte del molino llamado de La Duefia y discurria hacia el sur regando en primer termino !as huertas de la abadia de Benevfvere, que le daba su nombre - acequia Canoniga o de Benevivere -; la ubicacion de esta toma de agua se hallaba allado mismo del Camino de Santiago, lindando con tierras del Hospital de San Torcuato, un priorato dependiente de la abadia mencionada que se hallaba a un centenar de metros de la misma 43 . El segundo ojo se encontraba cerea del anterior y tambien discurria hacia el Sur, avenando el pago de Hingidro y otras tierras cercanas propiedad de habitantes de la villa de Carrion44 • El tercer ojo estaba entre el molino de Abanziles- que daba al canal principal uno de sus nombres- y el de Casasola, y se situaba en el actual pago de Ojo Muela, allado del camino que llevaba de Benevivere a Villanueva de Ios Nabos. La cuarta toma estaba entre Ios molinos de Casasola y el de Vado- perteneciente este ultimo al Hospital de la Herrada, que dependia del obispado palentino y se dedicaba a la atencion de los peregrinos que iban a Compostela - 45 ; como este curso secundario regaba principalmente tierras de dicho hospital, ha conservado hasta nuestros dias el nombre de Arroyo de La Herrada. El quinto ojo salia a la altura de Villanueva de los Nabos. El sexto, en fin, lo hacia a la altura de Nogal de las Huertas - situada al otro !ado del rio Carrion- y se aplicaba fundamentalmente al riego de los terminos de Villanueva de los Nabos y Villamez, en Ios que los monjes de Nogal tenian una importante dotacion territorial. Aguas abajo, y sin excluir en absoluto el riego de huertos, linares y frutales, el agua servia para mover el mayor conjunto de molinos de este espacio en un terreno idoneo para dicha aplicacion: el de Ios terminos de Torre y Calzada46 . Por lo que se refiere a! final de estos cauces principales, sefialaremos que elllamado Rio Nuevo desembocara en la Perihonda, entre Villaturde y Villotilla; esta confluira unos Km mas abajo, recien pasada Calzada de los Molinos, con el Izan, y, finalmente, este curso comun desembocani en Villoldo, cerrando el ciclo hidrologico de !as vegas de Saldafia y Carrion47 . 41 CASTRO GARRJDO A. & LIZOAIN GARRIDO J.M., op. cit., doe. 39. 42 LALANDA CARROBLES P., art. cit., 196 y 197. 43 FERNANDEZ L., op. cit., 3; VAZQUEZ DE PARGA L., LACARRA J.M. & URJA RIU J., Las peregrinaciones a Santiago de Camposte/a, t. !!, Madrid, 1948, 213-218. 44 LALANDA CARROBLES P., ibidem. 45 Ibidem. 46 Ibidem, y AHN. Clero. Legajo 5328, Relacion de losfimdamentos... En Calzada se doeumentan en el siglo XV Ios grupos molineros de Palaeio y de la Sema, perteneeientes al monasterio de San Zoilo, y el de la Duei\a (AHPP. Libra 2" de Apeos del monasterio de San Zoilo, fols. 4! -5 [, e Indice de San Zoilo, fol 132, respeetivamente). Asimismo, sabemos que en 1414 el monasterio de Benevfvere reeibe en donaei6n una parte en el molino de Sotillo en dieho lugar (FERNANDEZ L., op. cit., doe. 127). En Torre de Ios molinos hay meneiones a diehos ingenios -en plural y sei\alando su propiedad regia- en 1285 (LIZOAIN GARRIDO J.M. & CASTRO GARRIDO A., op. cit., doe. 39). 47 Pueden servir de adeeuada orientaei6n a este respeeto las hojas no 164, 197 y 235 del Mapa Topograflco Nacional de Espatia- eseala I :50.000- y el Mapa Provincial de Palenciaeseala l :200.000- elaborados por el Instituto Geogratieo Naeional, asf eomo el Diccionario geograflco, estadistico e historico de £spatia y sus provincias de ultramar, de MADOZ P., Madrid, 1850, en el eual habn1n de busearse Ios top6nimos que apareeen en el Mapa que nosotros adjuntamos a este trabajo. 279 J.A. Perez Celada Para concluir con la descripcion del sistema de canales de estas vegas, conviene que nos refiramos a Ios dos restantes cursos artificiales del sistema. Uno de ellos, el canal de Cestillos, se sirua, como el de Nogal-Poblacion, en la orilla derecha del rio, y aprovecha, como Villanueva del Rio tres km mas abajo, una pequefia red de canales adaptados a la planitud de este tramo de la ribera. En Cestillos poseia el monasterio de San Zoilo de Carrion, ademas de la parroquia de San Juan, importantes propiedades territoriales y a! menos un ingenio molinero 48 . Finalmente, el monasterio de San Zoilo poseia su propio canal, que nacia, cuando menos hasta el siglo XVIII, algo a! Sur de Poblacion de Soto, en la ribera derecha49 y tras recorrer unos seis km se reintegraba a! rio Carrion inmediatamente despues de la villa del mismo nombre, discurriendo a unos 250 metros del curso del rio y dejando a su izquierda el monasterio de San Zoilo y su barrio anejo, el de San Zoles. A la altura de este, Ios monjes habian derivado un canal que desembocaba tambien en el rio Carrion, atravesando Ios huertos inmediatos a! monasterio: era elllamado Arroyo Forero. El canal de Ios monjes regaba !as plantaciones de Ios mismos y de sus concesionarios, a la vez que daba movimiento a un complejo molinar50. Conviene efectuar, finalmente, algunas consideraciones sobre el regimen de aprovechamiento de Ios ingenios y Ios cultivos mencionados. Por lo que respecta a los molinos, cabe decir que nos hallamos en presencia de ingenios que responden de modo generalizado a! prototipo de rueda horizontal que transmite en tal disposicion su movimiento directamente a !as muelas, quedando en general excluidos, a! men os para estos tiempos medievales, Ios mol in os de rueda vertical que llevan a cabo su transmision a !as muelas a traves de un engranaje. La modestia del volumen hidrico del Carrion y, por ende, la de Ios canales derivados del mismo, nos situa ante la practica inevitabilidad de recurrir de modo sistematico al tipo de ingenio mencionado en primer lugar, ya que el molino de rueda vertical precisa de caudales de agua importantes y su coste resulta mas elevado 51 . La documentacion medieval castellana distingue con frecuencia entre el molino simple - molendino, molendinum, molino- y el complejo de rueda vertical - acenia, aceiia 52 , del arabe saniya -, pero lo cierto es que la simple mencion de la palabra "molino" encubre en ocasiones la existencia de una acefia, aunque tal cosa no sucede en sentido inverso 53 . Pero en el caso que tratamos, a !as razones antedichas en favor de un uso generalizado del molino de rueda horizontal en este espacio, se une la circunstancia de que la documentacion en el generada parece poner cuidado en la distincion entre molinos y acefias. Asi, por ejemplo, el monasterio de San Zoilo de Carrion, uno de Ios principales propietarios de molinos en esta zona, extendia sus dominios tambien por la vega del Duero, mas concretamente en Toro. Alii era poseedor de varias acefias que se mencionan como tales en la documentacion ya desde mediados del siglo XIV y asi se seguira hacienda en adelante 54 . No es preciso aclarar que el caudal del Duero resulta mucho mas apropiado que el de sus afluentes - y que el de !as acequias derivadas de estos- para la implantacion de tales ingenios; y, asi, tanto en Toro como en Zamora la presencia de acefias era un hecho comun ya en la Edad Media. Lo cierto es que el rio Carrion conocera la implantacion de estas maquinas, pero estimamos, en general, que en tiempos posteriores y con un caracter selectivo: en Palencia, situada inmediatamente antes de la desembocadura del Carrion en el 48 En !2!3 consta que el monasterio tenia dos molinos en Cestillos (PEREZ CELADA J.A., Documentacion ... (l 047-1 300), doe. 80, pero posteriormente hemos haladfo la menci6n de tres a principios del siglo XV (AHPP. Libra 2" de Apeos, fol. 5). Asimismo, existe una referencia al molino del Campo en !448 (AHPP. lndice ... , fol !75). 49 En AHN. Clero. Legajo 5328. Relacion de Ios fundamen/os ... , se hacen una descripci6n inequivoca en estos tenninos y una remisi6n a la pesquisa realizada en !203 que hemos citado reiteradamente. · 50 El Arroyo Forero, mencionado reiteradamente en la documentaci6n, es descrito en !527 en estos terminos: " ... el arroyo de !482 -,La Cespedera- Documen/acion ... (f 30 I -/400), referencia de 1338, doe. 2!5, y Libra 2"... , afio !482 -, TapiaLibra 2" ... , fol. 41, refs. de !48! y !482- y del Huerto- Libra 2"... , fols. 4! y 49 v•, refs. de !48! y !484; se indica en ambas que este molino tiene dos paradas o ruedas -. Hay asimismo menciones del Molino de Mediavilla, que no era de Ios monjes - AHN. Clero, Legajo 5328, ref. de !530 -.Se habla tambien de un molino de La Salceda en el barrio propiedad de Ios monjesAHN. Clero. Legajo 5328, afio !546, en que se arrienda, haciendose menci6n de que esto se venia realizando con anterioridad. 51 CARO BAROJA J., Tecno/ogia popular espaiiola, Madrid, !983, 263-265; VILLAR GARCIA L.M., La Extremadura castellano-leonesa, Valladolid, !986, 335. 52 CAS ADO ALONSO H., op. cil., !92 y ss. forero que sale del cuernago de nues/ro monasterio de San Zoyl e pasa por nuestras lwertas" (AHN. Clero. Legajo 5328, 25-IX- !527). Los molinos de Ios monjes que alimentaba su propio cuemago- y que salvo el que mencionamos en primer lugar se hallaban en el barrio de San Zoles- son Ios siguientes en el siglo XV: Yuanades- en el que Ios monjes ccdieron su participaci6n en 1399 (PEREZ CELADA J.A., Documenlacion ... (J 301-1 400), doe. 338) -,El Pison- Libra 2" de Apeos, fol. 42 v•; referencia 280 53 54 Ibidem. PEREZ CELADA J.A., La "casa" de San Pelayo de Toro y sus depencencias entre Ios siglos XI y XV. Una aproximaci6n al sefiorio cluniacense en la provincia de Zamora, en: Aetas del I Congreso de His/aria de Zamora, T. !!!, Medieval y Moderna, Zamora, 1988, 223-23!. Horticultura y molinos de agua en el curso medio del rio Carrion en la edad media Pisuerga, la expansion de !as manufacturas textiles determino la presencia de acefias en la Edad Modema, si no antes 55 . Por lo demas, !as descripciones disponibles del equipamiento de Ios ingenios hidraulicos que menciona la documentacion nos ponen ante molinos simples 56 . Otra cuestion es la mencion en la documentacion de Ios molinos en plural o en singular y, en conexion con ello, la del vocablo "parada H. Casado ha sefialado la identidad existente entre la "parada" y el "sitio para hacer un molino", de tal suerte que cuando la documentacion indica la existencia de x "paradas" en un molino, nos esta hablando de x ruedas con sus correspondientes muelas, rodeznos, etc. Ello explica tambien el hecho de que un mismo molino se mencione en unas ocasiones en singular y otras en plural, porque, efectivamente, lo usual era que un molino constara de varias paradas a la vez. Asi, el monasterio de San Zoilo era propietario de "dos casas de molinos en Calzada de Ios Molinos, !as de Palacio y La Sema, que contaban respectivamente, con cuatro y tres paradas. En el caso de Ios de Palacio, la documentacion desglosa cada una de !as paradas por el nombre que se les daba - "Cabe el postigo "El Gall ego "Chamorro y "El Cabo y describe el equipamiento de cada una de las mismas, lo que deja claro que, en efecto, cada "parada" era un ingenio molinero 57 • En otro orden de cosas, cuando la documentacion se hace explicita, Ios molinos se presentan en manos de Ios senores o en trance de estarlo, sean estos eclesiasticos o laicos - incluido el propio monarca. No cabe, sin embargo, descartar la posibilidad de que en !as pequefias a ideas que jalonan el curso del rio y que se hallan alejadas de ambitos economicamente dinamicos como Saldafia o Carrion, algunas comunidades hubieran podido mantener en su seno, total o parcialmente, !as "veces que correspondieran a sus vecinos hasta tiempos avanzados. Es importante hacer hincapie en la generalizada resistencia de Ios senores, una vez que Ios han adquirido, a desprenderse de la propiedad de sus molinos y aun a perder el control efectivo de Ios mismos mediante figuras de concesion hereditarias 58 . Lo que predominara de modo claro cuando menos en Ios siglos XIV y XV - que es cuando !as fuentes son explicitas al respecto- sera la cesion de Ios mismos en arrendamiento a corto plazo - entre tres y seis afios - a concesionarios a Ios que, en ocasiones, se atribuye explicitamente en Ios documentos la condicion socio-laboral de "molineros" 59 . Las condiciones de !as cesiones establecen que la contraprestacion que ha de hacer efectiva el arrendatario sea en especie - en cereal - a lo largo del siglo XV, aunque en la centuria anterior conocemos rentas dinerarias 60 . Por lo demas, el concesionario corren1 con Ios gastos de mantenimiento del canal y del equipamiento del molino 61 . Respecto a la cuestion del "valor"de Ios molinos, lo costoso de su construccion y su mantenimiento, cabe decir,en conexion con lo apuntado mas arriba, que la idea de que Ios mismos resultaban muy elevados, conviene que sea matizada por cuanto, siendo cierto que la aten cion a un ingenio de estas caracteristicas habia de ser intensiva y que Ios gastos de reposicion de sus componentes particularmente el canalillo que proyectaba el agua sobre la rueda, el rodezno, la turbina y !as mueles 55 tenia "el martillo y rodeznero, ... el rodezno ... "; el de El Gal/ego, "rodezno sin hierro, la rodeznera ... "; el de Chamorro, "rodezno sin hierro, martillo y rodeznera ... "; el del Cabo, "rodezno sin hierro, martillo y rodeznera ... "(AHPP. Libra 2" de Apeas, fols. 11 • 11 11 , 11 11 , 11 - A principios del siglo XVIII tuvo lugar un agrio enfrentamiento entre la ciudad de Palencia y !as poblaciones de las vegas de Saldana y Carrion, pues la Chancillerfa de Valladolid privo a estas de cualquier forma de aprovechamiento de Ias aguas del rio Carrion en favor de aquella, que afirmaba precisarlas para sus molinos, acenas y pisones aplicados en gran medida a la produccion textil. No obstante, el recurso de las vegas prosperani poco despues y senin restituidas en sus derechos (LALANDA CARROBLES P., art. cif., 160-161). RUIZ T.F., ha detectado una relacion entre el desenvolvimiento de las manufacturas y la conversion de Ios molinos en acenas: asi, en Burgos, las acenas comienzan a aparecer con frecuencia en la documentacion en !os siglos XIII y XIV (Tecnologia y division de la propiedad. Los molinos de Burgos en la Baja Edad Media, en Sociedad y poder real en Cast ilia, Barcelona, 1981, 77 -80). 56 Compartimos con CASADO H.(op. cif., 194) la afirmacion de que "la existencia de un rodezno indica claramente que nos encontramos ante un molino horizontal". En tal sentido, las descripciones de Ios molinos del monasterio de San Zoilo en Ios afios 80 del siglo XV, incluyen siempre menciones al rodezno, la rodeznera, etc (AHPP. Libra 2" de Apeos, fols. 47-49 v"). 57 CASADO H., op. cit., pag 195. Entre Ios molinos de San Zoilo, vemos, en efecto, a fines del siglo XV, que el de Cabe el Postigo 11 47 y 48). 58 Hasta 1500 solo conocemos, explicitamente, el caso del molino de Yuanades (PEREZ CELADA J.A., Documentacion ... (/301-1400), doe. 338. 59 Asi, y aparte de Ios reiterados arrendamientos de Ios molinos de Las Huelgas en Poblacion de Soto citados supra, ver Ios que, de modo sistematico hacia, en !as condiciones de duraci6n antedichas,el monasterio de San Zoilo en la segunda mitad del siglo XV de Ios que poseia en su barrio y en el entorno de Calzada en AHPP. Libra 2" de Apeas, fols 40-76 v•. En dicha fuente aparece- fol. 51 - un concesionario a qui en se atribuye la condici6nde "molinero"; una condicion que debia de ser por lo demas comun en quienes arrendaban Ios molinos monasticos, aunque no debe descartarse la posibilidad de subarrendamientos. Una cesion vitalicia en ibidem, fol. 43 v•. 60 En el siglo XIV, concretamente en 1338, si que observamos, en cambio entregas en dinero pore! disfrute de cuatro molinos cercanos al monasterio (PEREZ CELADA J.A., Documentacion ... (/301-1400), doe. 215). 281 J.A. Perez Celada estriadas - habfan de ser frecuentes, no parece en cualquier caso que puedan calificarse de excesivos, a tenor de !as informaciones de que disponemos correspondientes a fines del siglo XV, en que el equipamiento completo de un ingenio superaba apenas en precio al de una mula 62 . Para concluir, y por lo que respecta a Ios cultivos de regadfo, la documentaci6n reitera durante todo el periodo !as menciones de linares, herrenes, arboles frutales y otros de vega como el sauce o el olmo y, sobre todo, de huertos, y nos pone ante la circunstancia, habida cuenta de que el mayor volumen documental disponible corresponde a! entomo de Carrion de Ios Condes, de una gran proliferaci6n de estas explotaciones, lo que se corresponde, por lo demas, con el caracter de activo centro de intercambios comerciales de dicha villa. Pero Ios regadios tienen tambien una importante presencia rio arriba, consecuentemente con la adecuada disposici6n del terreno. Los sefiores se presentan como Ios mayores propietarios de estos bienes, y a traves de !as cartas de cesi6n de Ios mismos- y muy especialmente de Ios huertos- comprobamos que su puesta en valor de se realizaba a traves de concesionarios, aunque, a diferencia de Ios molinos, !as concesiones tenfan casi siempre el caracter de vitalicias 0 hereditarias 63 . 61 AHPP. Libra 2" de Apeos. Ibidem. Asf, en 1484, Ios equipamientos productivos de Ios ya citados molinos del Postigo, el Gal/ego, de Chamorro y el Cabo son "apreciados ", respectivamente, en 834, 1776, 2354 y 2193 maravedfs (ibidem, fol 47), mientras que una "bestia parda"una mula- lo es, en el mismo momento, en 1300 maravedfs (ibidem, fol 49 v•). 63 Las citas se multipiicarfan hasta la saciedad. Basten solo algunas: AHN. Clero. Carpeta 1709, n• 6 (3-IV-1410); AHPP. lndice de San Zoilo, fol. 163 ( 1421 ), fol. 164 v• ( 1448), fol. 158 (1470); AHN. Clero. Legajo 5329 (5-IV-1472). 62 282 Julio A. Perez Celada Universidad de Burgos Cl. Zaragoza N° l-6°A 0900 I Burgos Espaiia Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Auto1lio Fermindez Ugalde El almacenamiento subterraneo y la conquista feudal en la peninsula iberica: aportaciones de la arqueologia Introducci6n Este trabajo avanza sobre Ios planteamientos expuestos por nosotros con anterioridad basados, como principal hip6tesis, en que Ios silos subtern1neos de epoca medieval localizados en diversas regiones de la Peninsula Iberica (Reino de Toledo, especialmente Madrid, Catalufia, Arag6n, Valencia, Baleares, Castilla y Le6n, Andalucia, Sur de Portugal) son estructuras de almacenamiento a largo plazo caracteristicas de formaciones sociales en general pre-feudales, a menudo de raiz tribal, re!acionadas con formas de propiedad familiares, privadas o comunales y, en general, con sistemas de distribuci6n que permiten una cierta autonomia en la acumulaci6n privada y domestica de excedentes (Femandez Ugalde 1994). Presentamos un primer mapa de distribuci6n de silos medievales en la Peninsula Iberica, elaborado a partir de !as mas de 200 referencias publicadas en la bibliografia arqueol6gica. Consideramos que el declive del almacenamiento a largo plazo en silos, privado o colectivo, puede ser un buen indicador de la extension de relaciones de producci6n feudales, y coincidir con cam bios estructurales relacionados con transformaciones en la distribuci6n del excedente y con su disponibilidad por parte de !as unidades familiares. Desde tales planteamientos, la oclusi6n generalizada de silos subterraneos es un e!emento de especial valor a la hora de determinar el avance de la sociedad feudal. Sorprendentemente, Ios estudios arqueol6gicos de silos son escasos e incompletos: la simple identificaci6n funcional suele rehuirse, y en !as publica- Un caso excepcional en ambito islamico es el de Mesas de Castelinho (Almodovar, Algarve, Portugal), que contenia varios centenares de granos de Triticum diccocum Schrank, especie tetraploide hoy desaparecida de la Peninsula, y de trigo comun (Triticum aestivum L.) (Pais 1993). En cuanto al Norte peninsular, se distinguieron restos de cereales en Ios silos de Monte Cantabria (La Rioja) (Perez Arrondo y Andres 1986) y Burad6n (Rioja alavesa) (Martinez Salcedo & Cepeda 1994). ciones raramente se aportan secciones o estimaciones de capacidad. Efectuamos aqui un primer intento de aproximaci6n a la posibi!idad de una secuencia de abandono de silos vinculable a! avance de la conquista y la imposici6n de estructuras de poder feudales en la Peninsula Iberica. El almacenamiento subternineo en la peninsula iberica: caracteristicas y antecedentes premedievales Los silos subterraneos constituyen una soluci6n simple y econ6mica para conseguir un almacenamiento a largo plazo de cereales y otros productos agricolas que contribuye a la minimizaci6n de !as crisis de subsistencia peri6dicas. El silo suele llegar a nosotros con una utilizaci6n secundaria, reempleado como contenedor de desperdicios en un momento posterior a su abandono como almacen. Por esta raz6n, son muy contadas !as ocasiones en !as que se han preservado silos junto a restos de su contenido1. Frente a Ios t6picos frecuentes en la bibliografia arqueol6gica, deben tenerse en cuenta dos nociones basicas: - un silo no requiere habitualmente enlucidos de protecci6n de !as paredes (lo habitual es aislarlas con . paja); Ios materiales que se encuentran en el interior de un silo fechan el momento de su abandono, y nunca el de su utilizaci6n. El sistema de conservaci6n en silos es conocido en la Peninsula Iberica y, en general, en el continente europeo, a partir del Neolitico. Segun algunos investigadores, el hecho mismo de la "neolitizaci6n" (esto es, el paso desde la economia de caza y recolecci6n a una economia productora) pudo estar precedido de una intensificaci6n del almacenamiento como practica econ6mica (Testart 1985). En la Peninsula Iberica, el procedimiento de conservaci6n subterranea del excedente se hace relativamente frecuente en epoca romana imperial y 283 A. Fernimdez Ugalde tardia. Sin embargo, es obvio que Ios silos coexistieron con otros procedimientos de conservacion no subternineos: Ios graneros construidos (horrea o granaria) y Ios h6rreos o graneros sobreelevados (granaria sublimata o sublimia, horrea pensilia). Los silos siguieron siendo empleados habitualmente durante el reino visigodo, tanto en contextos eclesiasticos o monasticos como domesticos, urbanos y rurales. En lo que se refiere a Ios oscuros momentos entre !as postrimerias del reino visigodo y la aparicion de la primera documentacion sefiorial o eclesiastica, en zonas de Catalufia como el Maresme, se ha podido comprobar una perduracion del uso de silos entre !as epocas tardorromana y altomedieval. Asimismo, en la Submeseta Norte, en ciertos asentamientos tardorromanos y altomedievales que algunos investigadores interpretan como pertenecientes a comunidades de aldea se han localizado silos subterraneos considerados como de uso comunal (Reyes 1986). Con todo, la evidencia de estos momentos entre Ios siglos VIINIII y XI parece todavia demasiado tenue y escasa coma para sentar argumentaciones de tipo diacronico sabre continuidad o discontinuidad de Ios sistemas de almacenamiento. Los silos medievales en la peninsula iberica Coma precision inicial, hay que dejar sentado que es probable que el cereal no constituyera la base de la explotacion agricola hasta la implantacion de la renta feudal: en Ios silos se almacenaria una parte del excedente agricola, en sociedades que, generalmente, tendieron a desarrollar una agricultura diversificada que no era prioritaria o basicamente cerealista, oleicola o viticultora. De hecho, paradojicamente, el declive del uso de silos medievales coincidira con la introduccion de una agricultura eminentemente cerealista como es la feudal, derivada de !as exigencias de renta. El mapa adjunto permite comprobar la localizacion de silos medievales en la Peninsula Iberica. Sin pretensiones de exhaustividad, pues no se ha hecho una revision sistematica de publicaciones locales ni de "cartas arqueologicas", en else recoge la mayoria de Ios hallazgos publicados de silos fechados entre Ios siglos VII y XVI. Los ejemplos de silos son casi inexistentes en el cuadrante Noroccidental de la Peninsula, debido a su litologia paleozoica, poco favo- Aparte de su localizaci6n en ambientes rurales y urbanos, consta tambien el hallazgo de silos en areas fortificadas (castillos o alcazabas), que en este caso deben vincularse al abastecimiento de !as guamiciones. 284 rable. Conviene destacar que Ios aparentes vacios que presentan ciertas zonas, especialmente La Mancha, pueden deberse esencialmente a la ausencia de investigacion arqueologica de epoca medieval, mas que a una carencia real de silos medievales. Es obvio que la consulta exhaustiva de "cartas arqueologicas" puede densificar la presencia de silos, aunque - creemos no variaria en demasia !as areas de extension de este procedimiento. Los silos en al-Andalus Paralelamente a la perduracion de silos desde epoca tardorromana, la conservacion subterranea de Ios excedentes parece adquirir un renovado desarrollo a partir de Ios momentos de dominacion islamica, cuando Ios vemos extendidos por buena parte de al-Andalus y del Magreb, tanto en espacios urbanos como rurales. En nuestra opinion, estos almacenes subtemineos fueron construidos y utilizados como medias de conservacion de reservas alimenticias por productores insertos en formaciones sociales de raiz comunal o tribal o por grupos familiares extensos, y su gran difusion se debe probablemente a la conjuncion de al menos otros dos factores: la existencia de un subsuelo impermeable y facilmente excavable y la especial adaptabilidad de Ios silos a su insercion en o junta a unidades de habitacion 2• Los silos andalusies deben entenderse como elementos de almacenamiento de una parte de Ios excedentes agricolas (cere ales y quiza algunos tipos de leguminosas: es decir, Ios productos susceptibles de almacenamiento a largo plaza) para constituir una reserva esencialmente destinada a la seguridad alimenticia en afios de penuria. Ciertos objetos frecuentemente hallados en silos, que interpretamos como amuletos, apuntan a la importancia de la preservacion de !as reservas de grana en !as sociedades andalusies (Femandez Ugalde e.p.). Con todo, !as referencias en Ios tratados de agronomia hispanomusulmanes son casi anecdoticas - sabre todo, Ibn al-Awwan 1988, I, XVI) y, para el reino de Toledo, Yaqut (Bolens, 1979) -, y !as noticias escritas son asimismo escasas, aunque constan algunas consultas juridicas sabre el particular fechadas a partir del siglo IX (Lagardere 1995). Los silos de epoca andalusi que conocemos gracias a !as excavaciones arqueologicas son generalmente fosas de forma cilindrica, ultrahemiesferica o piriforme, excavadas en el firme geologico (a menudo de arenas y arcillas terciarias, pero tambien en sustratos rocosos). Su boca suele ser circular, con un ditimetro de alga mas de 1 m., y poseen por lo general una profundidad de alrededor de 1,40 m., que puede alcanzar (y superar) Ios 3 m. En ocasiones, se aprecia El almacenamiento subterraneo y la conquista feudal en la peninsula iberica: aportaciones de la arqueologia I .. • 6 I .. I . ·. ..· ~. .. . ..... ~ .... ••••• 'c;:~~: .. ... ..· .. .. .. • ... .. .~ .. ... i• ·.· I Fig. 1.- Distribuci6n de silos medievales en la Peninsula Jberica segun !as referencias en bibliografia arqueol6gica. en la boca un rebaje destinado a encajar una tapadera de madera o de piedra. Las paredes de los silos estan cuidadosamente redondeadas y alisadas y, en algunos de los ejemplares cilindricos de mayor profundidad, poseen agujeros escalonados para facilitar el ascenso desde el interior. La capacidad de estos silos subterraneos no es regular: en los excavados por nosotros en Madrid suele encontrarse entre 13 y 34 Hectolitros; en la ciudad de Lerida, su capacidad suele oscilar entre 6 y 14 HI. En general, puede afirmarse que en todo el territorio de al-Andalus se recurre a la construcci6n de silos subterraneos siempre que el subsuelo lo permite, bien relacionados con viviendas urbanas (Merida, Vascos, Calatrava la Vieja, Lerida), dentro de las mismas o en corrales o estructuras anejas, o bien agrupados para constituir areas de almacenamiento ~Madrid?). La distribuci6n espacial de los silos que aparecen en agrupaciones urbanas no parece obedecer a una planificaci6n organica. Aunque a menudo han desaparecido los restos de las viviendas coetaneas, es probable que cada silo o conjunto de silos se relacionara con estructuras de habitaci6n familiares, bien en el interior de las viviendas, bien en espacios o construcciones anexas como corrales. Consideramos que la iniciativa de la excavaci6n de silos como los descritos debi6 de corresponder a unidades familiares que desarrollaran una actividad de producci6n agricola y que, en cualquier caso, poseerian una cierta capacidad de almacenamiento domestico del excedente. En ocasiones, los grupos familiares pueden asociarse para constituir espacios de almacenamiento colectivos, si bien la propiedad de cada uno de los silos seguia probablemente siendo privada o familiar, no comunal. Sin duda, no los silos no son el unico sistema de almacenamiento empleado por las sociedades hispano-musulmanas, aunque sean el mas reconocible: por ejemplo, las algorfas o sobrados de las viviendas y las bodegas cumplieron sin duda esta funci6n, como lo hacian en la Sevilla del XIII y, sobre todo, los alhories, trojes o graneros ventilados, las estructuras de almacenamiento (hri) que se encontraban en los corrales o dentro de las viviendas, y el almacenamiento en tinajas de ceramica. Hay que citar, finalmente, la extension de los graneros en cuevas artificiales sobre acantilados por amplias zonas de Levante y del Sudeste, y la reciente excavaci6n de un granero fortifi285 A. Fernandez Ugalde cado similar a Ios ksar(s) marroquies, en Murcia (De Meulemeester, Matthys & Amigues 1995). Los silos en territorios "cristianos" del tercio norte peninsular: Cataluiia, Cantabria, Castilla y Leon, Alava La mayoria de Ios ejemplos de silos en territories cristianos se encuentran en ubicaciones rurales, a diferencia de Ios islamicos, que son tanto urbanos como rurales. A la luz de Ios hallazgos arqueologicos, Castilla se revela como uno de Ios nucleos mas densos de extension de Ios silos. En La Rioja, un buen ejemplo de silos relacionados con viviendas es el de la primera ocupacion medieval del asentamiento de Monte Cantabria, en Logrofio, datada entre el siglo X y principios del XII (Perez Arrondo & Andres 1986). La Catalufia feudal es otra de !as zonas con mayor densidad de silos medievales y para la que contamos con informacion arqueologica sistematica. Varias excavaciones proporcionan argumentos para considerar la continuidad del uso de silos entre la epoca romana y la altomedieval: destacan secuencias como la observada en L'Aiguacuit (Tarrasa), villa romana a la que se superponen un poblado de cabafias del siglo VII, asociado a un "campo de silos", y, despues, un nuevo poblado de viviendas con zocalo, en Ios siglos VIII-IX asociado a otro "campo de silos" (Coil, Molina & Roig 1993). Una modalidad de localizacion de silos bien documentada en el area catalana es la agrupacion en sagreras en torno a las iglesias. Se conocen ejemplares que funcionaban en !os siglos X y XI, aparecidos durante !as intervenciones arqueologicas en iglesias romanicas. En ocasiones se ha enunciado su asociacion a ocupaciones de epoca "de repoblacion", como en el caso de Santa C<'mdia d'Orpi en Barcelona (Lopez Mullor, Caixal & Fierro 1986), aunque se destaquen Ios precedentes del ensagrerament entre Ios siglos VII y X - caso, por ejemplo, de Ios silos de Santa Margarida (Navarro & Mauri 1986, 1994) o de Ios de Sant Mar9al de Terrassola (Fierro-Macia & Domingo 1987) -; estos autores han hipotetizado que la constitucion de sagreras podria relacionarse con una busqueda campesina de la proteccion sacra en momentos de feudalizacion y bandolerismo sefiorial. Sin embargo, coincidimos con Marti en considerarlas mas bien como parte del proceso de reorganizacion del espacio rural por la Iglesia y !os sefiores feudales que entre el siglo XI y mediados del XII conduce a la sustitucion de la estructura eclesiastica de villae por la parroquial, unido a la concentracion del poblamiento en torno a fortificaciones, a modo de incastellamento (Marti 1988). Ciertos silos aparecidos en Tarragona se inter- 286 pretan como pertenecientes a la fase condal de la ciudad (fines del siglo XI-siglo XII) (Menchon, Macfas & Mufioz 1994), y fueron amortizados en el siglo XIV. El almacenamiento subterraneo a largo plazo en epoca medieval Es evidente que !os silos subterraneos pueden funcionar dentro de regimenes de almacenamiento comunales, domesticos o sefioriales e, incluso, de acumulacion capitalista. Por tanto, si en determinados momentos o zonas su aparicion puede coincidir con un proceso de feudalizacion (por ejemplo, en las sagreras de la Catalunya Vella entre Ios siglos XI y mediados del XII) (Marti 1988), en otros es su cancelacion la que indicani la implantacion de estructuras feudales (por ejemplo, con posterioridad a la conquista castellana del reino de Toledo). En general, parece observarse que, historicamente, la construccion y mantenimiento de silos subtemineos suele responder a iniciativas particulares o comunales, coincidir con un almacenamiento domestico (suele mantenerse la propiedad privada de los silos, aunque esten agrupados o formen parte de graneros colectivos) y con modos de produccion antiguos o no feudales. Cuando Ios silos perviven en epoca feudal, lo hacen generalmente como sistemas de almacenamiento de excedente organizados y, a menudo, centralizados, establecidos por iniciativa de entidades nobiliarias, momirquicas, eclesiasticas o concejiles. En nuestra opinion, la evidencia historica y etnografica parece inequivocamente demostrar que !os silos subterraneos son una solucion preferente para la conservacion del excedente a largo plazo por parte de Ios productores insertos en formaciones sociales de raiz tribal o pre-feudal. Aun mas, como demuestran estudios historicos relativos a! Magreb moderno, la capacidad de almacenamiento de alimentos es clave para el mantenimiento de la autonomia familiar (Bergeret 1985, 152; Rosenberger 1985), y el simbolismo tradicional asocia Ios silos a la vitalidad del grupo (Lefebure 1985). Viceversa, como demuestra la experiencia colonial en el siglo XX, la destruccion del autoabastecimiento campesino es la clave de la sumision y de la dependencia, y la estrategia prioritaria de los conquistadores. El fenomeno del abandono de silos en la peninsula iberica: algunas evidencias arqueologicas Un problema fundamental a la hora de profundizar en un estudio diacronico es el de las dataciones El almacenamiento subterraneo y la conquista feudal en lapeninsula iberica: aportaciones de la arqueologia de Ios estratos de relleno de silos: solo una cuarta parte de !as referencias bibliognificas aportan una cronologia mas o menos precisa y/o fiable. No entraremos, en general, a discutir !as cronologias adjudicadas por sus excavadores. No obstante, consideramos que a menudo se tiende a elevar !as dataciones, en buena medida por planteamientos reduccionistas que contraponen lo "islamico" y lo "cristiano". Con todo, mencionemos que la gran mayoria de Ios rellenos de silos no han sido excavados estratigraficamente ni insertos en cronologias relativas junto a secuencias extemas a! propio silo. Una de !as zonas en !as que la hipotesis de un abandono de silos coincidente con la creciente feudalizacion parece verificarse es en Barcelona, donde se observa una sistematica oclusion de silos en tomo a! siglo XII. La secuencia de abandono es bien conocida en el Maresme y en algunas otras comarcas: se produce una reutilizacion de silos como contenedores de basuras fechable hacia fines del siglo XI o principios del XII (Clariana et al. 1986, 1987). Los silos agrupados en sagreras eclesiasticas parecen asimismo sufrir una oclusion general fechable entre fines del siglo XI y principios del XIII (Lopez Mullor et al. 1986; Moro & Roig 1994; Navarro & Mauri 1994). Algunos autores ya han relacionado la oclusion de silos y el abandono de algunos poblados con la reorganizacion territorial derivada del establecimiento del poder condal (Coil, Molina & Roig 1994 para el Valles). En Castilla y Leon, un caso de oclusion generalizada de un conjunto de silos, producida hacia el siglo XI o quiza todavia en el X, es la observada recientemente en la Tierra de Campos palentina (De la Cruz & Lamalfa 1994); tambien se ha querido relacionar con la reorganizacion del poblamiento impuesta por el poder condal. En general, abundan en Castilla y Leon los casos de cegados producidos genericamente entre Ios siglos XI y XIII, aunque hay que hacer mencion a la particular imprecision de !as cronologias en la zona. Inutilizaciones generales de silos se han observado en las viviendas de la fase medieval temprana del asentamiento de Monte Cantabria (Logrofio ), a principios del siglo XII (Perez Arrondo & Andres 1986) o en Ios 36 silos del yacimiento medieval de Los Paletones (Cenicero, La Rioja) (Gomez Martinez 1989). Casos de oclusiones discordantes, por tardias, se constatan en A.Iava, Burgos y, sobre todo, en la ciudad de Valladolid, donde se extienden, e incluso superan el siglo XV. El reino de Toledo, objeto de estudios pormenorizados por nuestra parte (Femandez Ugalde 1994), parece ser una de !as principales zonas de confirmacion de la hipotesis de una oclusion subsiguiente a la conquista cristiana de 1085, y no existen evi- dencias de perduracion del procedimiento de conservacion en silos a partir de epoca bajomedieval o en epoca modema. En cuanto a la Marca Superior, y especialmente a! territorio leridano, el momento supuesto de amortizacion de la mayoria de Ios silos del Antic Portal de Magdalena es poco anterior a la conquista (1150) (Loriente 1990), aunque podria quiza retrasarse algunos decenios, y se constatan rellenos de silos ya en epoca de taifas. En todo caso, las viviendas de nueva planta edificadas en Lerida en el siglo XIII carecen de silos. De particular interes resulta el analisis de la evolucion de Ios silos en !as regiones conquistadas entre mediados del siglo XII y mediados del siglo XIII: Valencia, Baleares, Sevilla, el Algarve. En cuanto a Andalucia, es sintomatico que la mayor parte de !as posibles estructuras de tipo silo conocidas posean rellenos datables "en epoca almohade", que bien podrian remontarse al final de la dominacion musulmana de la zona. Un ejemplo que se adapta perfectamente a la hipotesis es el de Villa Julia (Carcabuey, Cordoba), presunta alqueria en la que el momento de cegado intencionado de Ios ocho silos localizados podria coincidir con la conquista por la orden de Calatrava (1240-45), constatandose una subsiguiente continuidad del asentamiento (Carmona 1995). En Levante y Baleares, la mayoria de Ios rellenos de silos pueden enmarcarse en un periodo proximo a la conquista o inmediatamente posterior a ella. Asi, Ios dos silos hallados en Paterna, empleados como tales hasta fines del siglo XII (Mesquida 1990), o el conjunta de cinco silos junto a la Torre de Silla, que parece haber sido inutilizado a fines del siglo XIII. Otros posibles casos similares son Ios de un silo domestico en Alhama de Murcia (Bafios & Bemabe 1994), Ios diversos silos localizados en Javea y la comarca de la Marina Alta alicantina o el de ciertos "pozos" descubiertos en la ciudad de Elda (Alicante), rellenos a fines del XII o principios del XIII. Conclusiones En nuestra opinion, puede afirmarse que existen zonas de la Peninsula Iberica en !as que parece verificarse la hipotesis esbozada y otras que presentan casos discordantes. Por supuesto, existen en todas !as zonas ejemplos de silos cuya oclusion se data en momentos demasiado tardios para lo planteado en este articulo - por cierto, buena parte de ellos en el entomo de fortificaciones militares, residencias sefioriales o edificios religiosos -, y es obvio que existen abandonos de silos antes de la conquista feudal por diversas razones; ademas, el relleno de un silo no implica necesariamente que haya seguido en uso 287 A. Femandez Ugalde hasta el momento de su oclusion, y el abandono de una parte de Ios silos puede ser producto de procesos particulares o de la expulsion de poblacion musulmana. Todo lo anterior no invalida la hipotesis de la contradiccion de Ios sistemas de almacenamiento a largo plazo, privados, familiares o chinicos, con el feudalismo. Insistimos, por ultimo, en que el modelo propuesto se aplica esencialmente a Ios silos domesticos o a Ios silos comunales de comunidades de aldea o tribales, y se verifica esencialmente en yacimientos con continuidad de ocupacion. Lo cierto es que desde la epoca bajomedieval, cuando las fuentes documentales comienzan a ser relativamente abundantes y ha desaparecido del registro arqueologico toda evidencia de almacenamiento domestico y subterraneo, Ios t1nicos sistemas de almacenamiento mencionados son !as alhondigas, de canicter comunitario, concentrado u organizado, es decir, no relacionadas directa y fisicamente con las viviendas de Ios productores sino vinculadas a un cierto grado de centralizacion, y Ios graneros o conjuntos de silos vinculados a castillos, edificios palatinos, nobiliarios, monasticos o eclesiasticos: en la ciudad feudal se tienden a concentrar el almacenamiento y la distribucion de Ios productos, a menudo mediante la concesion de monopolios. Asf pues, la cuestion clave no es que desaparezca el procedimiento de conservacion en silos en si, sino el fin de su utilizacion generalizada en el entomo domestico, y la evidencia de una concentracion fisica de Ios excedentes. Es sintomatico que Ios moriscos del reino de Granada conservaran en uso Ios silos todavia a mediados del siglo XVI, o que en Ceuta, donde Ios cronistas portugueses de la conquista destacan la amplia extension de Ios silos por el espacio urbano, se distingan oclusiones a partir de la conquista de 1415. La ruptura de la autonomfa campesina por el feudalismo es un fenomeno parangonable a otros procesos mejor conocidos: la disgregacion de las familias extensas, fenomeno observado historicamente en Galicia y el valle del Duero o, arqueologicamente, a traves de la transformaci6n de !as viviendas moriscas de las sierras alicantinas (Torr6 & Ivars 1990); la extension de Ios derechos banales sobre molinos o la prohibicion de uso de tanores u homos de pan porta tiles (del arabe tannur) impuesta por Ios conquistadores catalanes de la ciudad de Valencia (Gutierrez 1996, cit. trabajos de Torro): el epflogo de todos estos procesos es la dependencia sefiorial de la poblacion rural y urbana. En buena medida, el almacenamiento a largo plazo en silos, que dificulta la "visibilidad" y, por tanto, la cuantificacion y fiscalizacion de !as cosechas3 , supone una contradicci6n con la imposicion de renta feudal. 288 Bibliografia BANOS J. & BERNABE M. 1994: Excavaciones arqueologicas en el casco antiguo de Alhama de Murcia, Revista de arqueologia 157, 60-61. BERGERET A. 1985: Les reserves de grains des societes africaines au Proche-Orient, au Maghreb et dans les pays du Sahel, en: Gast, Sigaut & Beutler (eds.), fasc. 1, 151-153. BOLENS L. 1979: La conservation des grains en Andalousie medievale d'apres les traites d'agronomie hispano-arabes, en: Gast & Sigaut F. (eds.), vol. 1, 105-112. CARMONA R. 1995: Los silos hispanomusulmanes de Villa Julia (Carcabuey, Cordoba), Antiquitas 6, 133-140. CLARIANA J.F., 0ROBITG M.J., PORTILLO T. & PREVOSTI M. 1986: Datos para el estudio del poblamiento rural altomedieval del Maresme (Barcelona), Aetas del I Congreso de arqueologia medieval espafiola, Zaragoza, vol. IV, 569-585. COLL J.M., MOLfNA J.A. & Roro J. 1993: L'Aiguacuit (Terrassa, Valles occidental). La transicio d'una vila romana en un vilatge medieval, IV Congreso de arqueologia medieval espafiola, Resumenes, Alicante. DE LA CRUZ A. & LAMALFA C. 1994: Monzon de Campos: la transformacion del sistema de almacenamiento como consecuencia del cambio en !as estructuras agricolas, IV Congreso de arqueologia medieval espafiola, Alicante, vol. III, 605-610. DE MEULEMEESTER J. MATTHYS A. & AMIGUES F. 1995: Un grenier collectif fortifie hispano-musulman? Le Cabezo de la Cobertera (Vallee du rio Segura/Murcie), Bilan provisoire d'une approche ethnoarcheologique, en: BAZZANA A. & DELAIGUE M.C. (eds.), Ethno-archeologie mediterraneenne. Finalites, demarches et resultats, Madrid, 181-196. DEMERSON J. & ZOZA YA, J. (1983): Ceramicas islamicas de C'an Portmany (Ibiza), Boletin de la Sociedad Espafiola de Orientalistas XIX, 163-184. FERNANDEZ UGALDE A. 1994: El fenomeno del relleno de silos y la implantacion del feudalismo en Madrid y en el reino de Toledo, IV Congreso Es notorio que el almacenamiento en silos dificulta per se la extracci6n de renta, al esconder el total de la cosecha (aspecto mucho mas dificil de conseguir mientras se empleen estructuras al aire libre). En epoca carolingia, el Cartulario carolingio de Villis prohfbe la ocultaci6n subterranea de cosechas (Mauny 1979). Noticias sobre una tardfa ocultaci6n subterranea de cosechas en la Ibiza del siglo XVIII, con la consiguiente indignaci6n del obispo por el impago de diezmos y tributos, pueden consultarse en Demerson & Zozaya 1983. El almacenamiento subterraneo y la conquista feudal en la peninsula iberica: aportaciones de la arqueologia de arqueologia medieval espafiola, Alicante, vol. Ill, 611-61 7. FERNANDEZ UGALDE A. e.p.: jQue Dios nos conserve el grano! Una interpretacion de !os omoplatos con inscripcion arabe procedentes de yacimientos medievales, Al-Qantara. FIERRO-MACIA J. & DOMINGO R. 1987: Excavaciones en la iglesia de Sant Maryal de Terrassola (Torrelavit, Barcelona), If Congreso de arqueologia medieval espafiola, Madrid, Ill, 421-428. GAST M. & SIGAUT F. (eds.) 1979: Les techniques de a conservation des grains long terme. Leur role dans la dynamique des systemes de cultures et des societes, Vol. 1, Paris. GAST M. & Sigaut F. (eds.) 1981: Les techniques de conservation des grains long terme, Leur role dans la dynamique des systhnes de cultures et des societes, Vol. 2, Paris. GAST M., S!GAUT F. & BEUTLER C. (eds.) 1985: Les techniques de conservation des grains long terme, Leur role dans la dynamique des systemes de cultures et des societes, Vol. 3, fasc. 1 y 2, Paris. G6MEZ MARTINEZ J.R. 1989: Carreteras y arqueologia, Revista de arqueologia 101, 6-7. GUTIERREZ S. 1996: La producci6n de pan y aceite en ambientes domesticos. Limites y posibilidades de una aproximaci6n etnoarqueol6gica, Arqueologia medieval 4 (Formas de habitar e alimentar;:iio na Idade Media), 237-254. IBN AL AWWAN 1988: Libra de agricultura. Su autor, el Doctor excelente Abu Zacaria Iahia Aben Mohamed ben Ahmed Ebn el Awam, sevillano, Ed. facsimil de la de J.A. Banqueri ( 1802), 2 vols, Madrid, Ministerio de Agricultura. LAGARDERE V. 1995: Histoire et societe en accident musulman au Moyen Age. Analyse du "Mi 'yar" d'Al Wansarisi, Madrid. LEFEBURE C. 1985: Reserves cerealieres et societe: 1'ensilage chez les marocains, en: Gast, Sigaut & Beutler (eds.), vol. 3, fasc. 1, 210-235. LOPEZ MULLOR A., CAIXAL A., FIERRO J., DOMINGO R. & JUAN M. 1986: Excavaciones en la iglesia de Santa O'mdia d'Orpi (Barcelona), Aetas del I congreso de arqueologia medieval espafiola, vol. V, 95-109. LORIENTE A. 1990: L 'horitz6 andalusi de I 'antic Portal de Magdalena, Monografies d'arqueologia urbana 2, Lerida. MARTi R. 1988: L'ensagrerament: l'adveniment de les sagreres feudals, Faventia 10, 153-182. MARTINEZ SALCEDO A. & CEPEDA J.J. 1994: Conjunto arqueol6gico de Buradon (Salinillas de Buradon), Arkeoikuska 93, 257-270. MAUNY R. 1979: Contribution a !'etude des fosses a a ovoides et silos, en: Gast & Sigaut F. (eds.), vol. 1, 48-53 .. MENCHON J., MAciAS J.M. & MuNOZ A. 1994: Aproximaci6 a! proces transformador de la ciutat de Tarraco. Del Baix Imperi a I 'Edat Mitjana, Pyrenae 25, 225-258. MESQUIDA M. 1990: Plaza del Pueblo. Paterna, 1'Horta, Excavacions arqueologiques de salvament a la Comunitat Valenciana 1984-1988. I. Intervencions urbanes, Valencia, 112-114. MORO A. & ROIG J. 1994: El conjunt de sitges altmedievals de Sta. Maria d'Egara per a l'emmagatzematge de cereal, IV Congreso de arqueologia medieval espafiola, Alicante, vol. Ill, 619-624. NA v ARRO R. & MAURI A. 1986: La excavacion de un silo medieval en Santa Margarida (Martorell, Barcelona), Aetas del I Congreso de arqueologia medieval espafiola, vol. V, 435-452. NA VARRO R. & MAURI A. 1994: Santa Margarida de Martorell: la transici6 de I' antiguitat tardana a! m6n medieval, IV Congreso de arqueologia medieval espafiola, vol. Ill, 341-344. PAIS J. 1993: Sementes de urn silo omiada (UE 67) de Mesas de Castelinho (Almodovar), Arqueologia medieval2, 109-110. PEREZ ARRONDO C.L. & ANDRES VALERO S. 1986: El poblamiento medieval en el yacimiento arqueol6gico de Monte Cantabria (Logrofio, La Rioja), Aetas del I congreso de arqueologia medieval espafiola, Zaragoza, vol. IV, 485-505. REYES F. 1986: Excavaciones en la ermita de Santa Cruz (Valdezate, Burgos), Aetas del I congreso de arqueologia medieval espafiola, Zaragoza, vol. V, 7-27. RosENBERGER B. 1985: Reserves de grains et pouvoir dans le Maroc precolonial, en: Gast, Sigaut & Beutler (eds.), vol. 3, fasc. 1, 237-268. TEST ART A. 1985: Les reserves alimentaires dans les societes de chasse-cueillette. Presentation, en: Gast, Sigaut & Beutler (eds.), vol. 3, fasc. 1, 9-13. ToRR6 J. & IV ARS J. 1990: La vivienda rural mudejar y morisca en el Sur del pais valenciano, en: BERMUDEZ J. & BAZZANA A. (coords.): La casa hispano-musulmana. Aportaciones de la arqueologia. La maison hispano-musulmane. Apports de I 'archeologie, Granada, 73-97. Antonio Femandez Ugalde 'AREA Sociedad Cooperativa Arqueologica cl Villa, 3 - 2° D 28005 Madrid Espaiia 289 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference - Volume 6 Elena Serrano, Antonio Fermindez Ugalde & Leonor Peiia-Chocarro Los Silos Medievales en el Reino de Toledo 1 Introducci6n Comentarios de geografos musulmanes de Ios siglos XI y XII ilustran sobre la singular difusion que alcanzo este sistema de conservacion de reservas de cereal en la region de Toledo. Los silos son Ios vestigios arqueologicos mas caracteristicos de la Meseta iberica desde la prehistoria reciente, a partir de mediados del tercer milenio antes de nuestra era, y su uso perdura durante !as epocas prerromana, romana y visigoda, habiendo alcanzado incluso hasta el siglo XIX. En este trabajo se trata la informacion arqueologica relativa a Ios silos medievales en el reino musulman de Toledo y a su evolucion posterior. El reino de Toledo fue uno de Ios pequefios Estados que surgieron como resultado de la descomposicion del Califato de Cordoba a partir de principios del siglo XI (Fig. 1). La conquista castellano-leonesa de la capital y de gran parte del reino en 1085 supone el primer gran avance de la expansion feudal en la Peninsula Iberica; sin embargo, una parte del antiguo territorio de Toledo perdura bajo el poder del imperio almoravide durante el siglo XII. Los hallazgos de silos se multiplican en epoca islamica (711-1 085) en ambientes urban os y rurales: ciudades, castillos, fortificaciones, alquerias. No obstante, nuestro interes nos lleva a centramos en Ios ejemplares documentados estratigraficamente en contextos urbanos con continuidad temporal durante la Baja Edad Media (lo cual, por desgracia, nos limita casi exclusivamente a la ciudad de Madrid). En el entomo de Madrid son abundantisimos Ios silos localizados en !as excavaciones recientes, en general rela- cionadas con Ios vigentes programas de intervencion arqueologica vinculada a proyectos constructivos y, en buena medida, todavia ineditas 1• 2 La identificaci6n de las estructuras como silos No ha sido hasta epoca relativamente reciente cuando, en contra de la opinion generalizada, se han empezado a interpretar Ios numerosos "agujeros" excavados directamente en el terreno natural arcilloso 2 y generalmente carentes de recubrimiento, como silos subternineos destinados a! almacenamiento de cereal u otro tipo de reservas alimentarias. La interpretacion predominante era la de basureros, ya que su excavacion proporcionaba evidencias de la deposicion de vertidos de diversa naturaleza, y el interes de la investigacion se limitaba a la informacion aportada por Ios objetos contenidos en Ios estratos de relleno de estos "agujeros". Los estudios se han centrado en Ios artefactos como unidad de amllisis, destinados a la elaboracion de tipologias (generalmente ceramicas) que, aunque necesarias, han ocultado informacion sobre la funcionalidad original de !as estructuras excavadas. Efectivamente, la identificacion de !as estructuras es problematica. Pozos de captacion, pozos de noria, pozos negros, etc. podrian ser considerados como interpretaciones plausibles que, de no ser convenientemente refutadas para cada caso, serian tan validas como la aqui esbozada. Carecemos, en la mayoria de Ios casos, de informacion relativa a! uso original de !as estructuras, ya que no nos ha llegado ni el contenido original ni elementos susceptibles de relacio- Entre las excavaciones madrilefias publicadas sinteticamente por Ios firmantes, mencionemos Fem{mdez Ugalde 1996; Femandez Ugalde & Serrano 1997 (e.p.); Serrano 1996; Serrano & Yanez 1996; Vallespin et al. 1990 y Yafiez et al. 1992. 2 En el caso peninsular si parece observarse una asociaci6n entre el sustrato geo16gico arcilloso (terrenos terciarios) y el empleo de tecnicas de almacenamiento subterraneo. No obstante, y a modo de ejemplo, vease el caso de Ios asentamientos de la Edad del Hierro inglesa donde Ios vestigios arqueo16gicos mayoritariamente representados son tambien "agujeros" excavados en varios tipos de rocas, para Ios que se ha demostrado un uso original como silos subterraneos (Reynolds 1979). En cualquier caso, clima y condiciones naturales del subsuelo condicionan, que no deterrninan, la elecci6n de una determinada tecnica de almacenamiento. 291 E. Serrano, A. Femandez Ugalde & L. Peiia-Chocarro Fig. 1. - a) Distribuci6n general de silos medievales en la Peninsula Iberica y limites del Reino de Toledo. b) Lugares de hallazgo de silos medievales en el reino de Toledo. ,·,~··· . . . . ~:: •.' 1 u 2 u u uUU 'D2 5 11 9 19 u .24 13 14 uu []25 10 026 narse con esta primera funcion. La presencia de enlucidos no constituye por si misma un argumento, por cuanto Ios silos medievales del reino de Toledo, y en general Ios del resto de la Peninsula Iberica, no presentan revestimientos permanentes, siendo lo habitual el empleo de capas de paja protectoras. En conclusion, en la mayoria de Ios ejemplos analizados se observa una ausencia de elementos aportados por el registro arqueologico que apoyen la hipotesis del uso como silos. Como problema afiadido, a menudo se echa en falta en !as publicaciones de estructuras como !as aqui analizadas la aportacion de la necesaria documentacion planimetrica y de secciones arqueologicas. 292 I. 2. 3. 4. 5. 6. TALAMANCA 7. HUETE 8. 9. 10. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. GUADALAJARA LAS CHORRERAS FRJAS DE ALBARRACIN ALCALA DE HENARES RECOPOLIS PAJARONCILLO MADRJD PINTO ESCALONA 0LMOS lLLESCAS MELQUE BEL VIS VASCOS CALATRAVA LA Y!EJA LAS FUENTECILLAS MAQUEDA ALCOHUJATE PERALES DEL RIO CANAVERUELAS V!LLAREJO DE SAL VANES PANTOJA MAGAN La falta de evidencias en el propio contexto arqueologico ha dificultado la identificacion, clasificacion e interpretacion de !as fosas. Una vez vaciado del contenido final, Ios principales elementos susceptibles de amilisis y cuantificacion son Ios relacionados con aspectos morfologicos y volumetricos, a partir de Ios cuales se pueden efectuar estimaciones sobre capacidad. Inevitablemente tenemos que acudir a! apoyo de fuentes documentales, que para el caso que nos ocupa resultarian suficientemente explfcitas (Ibn al-A wwan 1988, I, cap XVI; Bolens 1979), a ejemplos constatados arqueologicamente fuera de la region y, con las debidas precauciones, a paralelos etnograficos en el mundo ishimico meditemineo (Lefebure 1985). Los Silos Medievales en el Reino de Toledo Para Ios ejemplos analizados del reino de Toledo es notoria la ausencia de elementos que corroboren la hip6tesis esbozada sobre la vinculaci6n de las estructuras excavadas a la conservaci6n de alimentos. Lamentablemente, carecemos en general de asociaciones entre viviendas y silos. Un solo caso documentado en Madrid (Cava Baja 30), cuya fiabilidad estratigrafica no es total, plantea la posible relaci6n entre una estructura construida y uno de Ios silos documentados; salvo en el caso de la ciudad de Vascos, donde su excavador documenta la presencia de un pozo o silo en el interior de una estancia de una de las viviendas excavadas, el resto de Ios ejemplos publicados son hallazgos de silos aislados. Excavaciones en curso en Madrid y otras comarcas del centro de la Peninsula pueden, sin embargo, proporcionar nuevos elementos sobre la asociaci6n entre silos y viviendas. lgualmente interesante es el analisis del ultimo contenido de los silos. Como han demostrado trabajos de investigaci6n encaminados a la resoluci6n de problemas similares, los vertidos que han sido analizados presentan un elevado componente de materia organica susceptible de ser destinada a fines practicos como el abonado de los campos mas pr6ximos. La explicaci6n de su deposici6n final en fosas debe vincularse al abandono forzado de una estructura subterranea preexistente o a imposiciones de distinto caracter (profilacticas, higienicas, etc.). Por ultimo, mencionemos que deterrninados objetos hallados en el interior de silos (amuletos, omoplatos inscritos en arabe, ceramicas romanas recortadas, cuemos de ciervo) remiten posiblemente a practicas magicas vinculadas a la conservaci6n de las reservas de grano (Femandez Ugalde e.p.). 3 Capacidades, morfologia, distribuci6n, estratificaci6n interna Parad6jicamente, el momento de la utilizaci6n de los silos como tales es aquel de que poseemos menor inforrnaci6n. Raramente se han preservado Ios estratos de ocupaci6n coetaneos a los silos. Esto hace imprescindible proceder a analisis detallados de la capacidad y distribuci6n espacial de cara a la aproximaci6n a su funcionalidad original. Los silos documentados en Madrid consisten en fosas piriforrnes, troncoc6nicas o cilindricas, excavadas en el substrato terciario de arenas o arcillas (Fig. 2). Las paredes estan cuidadosamente alisadas y, en ocasiones, se observan rebajes para encajar tapaderas u oquedades a ambos lados de la boca. Las profundidades Ios diferencian netamente de Ios pozos de captaci6n de agua o de los pozos de noria, mucho V u u u V 2m. Fig. 2. - Secciones de silos excavados en la ciudad de Madrid. mas profundos, y la ausencia de elementos estructurales anexos impide plantear su interpretaci6n como letrinas o pozos negros de viviendas. Las capacidades de los silos oscilan generalmente entre Ios 13 y 34 Hect6litros. Si atendemos a la documentaci6n medieval de la Geniza de El Cairo estudiada por Goitein, alrededor de 32 HI. de trigo constituian el abastecimiento medio durante un afio de una familia hebrea de 8 a 10 personas de Fustat (Goitein 1983, 129). Otros silos mas pequefios pueden haber sido destinados al almacenamiento de la sementera o de frutos secos. En cuanto a su distribuci6n en entomos urbanos, hacia el siglo XI, en una ciudad del reino de Toledo de importancia secundaria, como es el caso de Madrid, Ios silos poseian una amplia dispersion: hasta la fecha aparecen forrnando agrupaciones en la practica totalidad de !as numerosas excavaciones efectuadas intramuros del denominado "segundo recinto" (area amurallada a fines del siglo XI o principios del XII, que se considera habitada desde epocas algo anteriores) (Fig. 3; Lam. 1). Faltan por el momento elementos que perrnitan asegurar si nos encontramos ante areas de almacenamiento en espacio abierto, como podrian indicar el emplazamiento periferico de la mayor parte de !as excavaciones efectuadas y la general ausencia 293 E. Serrano, A. Femandez Ugalde & L. Peiia-Chocarro 0 0 0 0 0 0 Cava Baja 30 0 'Casa de San lsidro' Fig. 3. -Madrid medieval: limites de la madina emiral y del "segundo recinto "amurallado. En Ios recuadros, planta de Ios silos de !as excavaciones urbanas en a "Casa de San Isidro "yen Cava Baja, 30. de estructuras, o bien si Ios silos se relacionaron fisica y economicamente con unidades de habitacion. En general, Ios silos excavados en la region toledana presentan un relleno pluriestratificado. Los estratos de relleno, compuestos esencialmente de tierras con elevada proporcion en materia organica carbonizada que contiene numerosos fragmentos ceramicos y restos oseos, son de superficies sensiblemente horizontales y casi siempre netamente diferenciables por su composicion (mayor o menor propercion de cenizas y materia organica carbonizada) y su color. En Ios excavados por nosotros, el numero de 294 estratos diferenciables dentro de un mismo silo oscila entre 2 y 13. Es frecuente que fragmentos ceramicos de un mismo recipiente se encuentren repartidos en distintos estratos, y no solo en Ios inmediatamente contiguos, lo cual indica que, en general, el relleno de Ios silos se produjo rapidamente y de una sola vez. En nuestra opinion, la estratificacion que se observa en el interior de la mayoria de Ios silos esta generada por la limpieza de ambientes domesticos y el ulterior transporte de Ios desperdicios en contenedores de escasa capacidad (a cuestas en espuertas o en alforjas, a lomos de bestias de carga), procedimiento que pue- Los Silos Medievales en el Reino de Toledo Lam. 1. - Vista de !os silos aparecidos en la excavaci6n de Cava Baja, 30 (ivfadrid),junto a la muralla del "segundo recinto ". de llegar a producir estratos de escasa potencia, como es el caso de los silos rellenos por numerosos estratos sutiles. 4 Cronologia La informaci6n que habitualmente ha generado la excavaci6n de silos se refiere a dataciones o aproximaciones cronol6gicas procedentes del amilisis de los artefactos contenidos en los estratos de relleno, y a estudios tipol6gicos centrados en el material ceramice. Al emplearse referentes tipol6gicos, se vienen mayoritariamente encuadrando Ios silos entre los siglos IX-XI y XII-XIII, a partir de la presencia/ ausencia de determinados f6siles directores, como las decoraciones en "verde y manganese" y "cuerda seca", 0 de determinadas formas ceramicas empleadas para obtener precisiones mas certeras. Las dataciones evidencian un rigido panorama reduccionista que contrapone lo islamico a lo cristiano, a partir del cual resulta dificil establecer evoluciones diacr6nicas, y que tiende a sobrevalorar Ios cam bios esteticos reflejados en las diferentes producciones materiales y en ningun caso a cuestionar hip6tesis relativas a! proceso hist6rico que ha generado el registro. Debemos hacer una vez mas hincapie en la noci6n de que Ios materiales de relleno de un silo fechan el momento de su amortizaci6n, y nunca el de su uso, salvo en Ios casos excepcionales en Ios que se ha preservado el silo junto a !as reservas que contenia. Asi pues, si parece clara la amplia difusi6n del sistema de silos en Ios ambientes urbanos y rurales toledanos de epoca islamica, a pesar de la escasa evidencia de asociaci6n a viviendas y estratos andalusies, recientes trabajos plantean la existencia de indicios suficientes para hipotetizar su abandono generalizado en el reino a partir de la conquista feudal, desde fines del siglo XI y durante Ios siglos XII y XIII. El proceso de oclusi6n pudo comenzar ya en las postrimerias del reino de taifas, pero puede seguirse sin dificultades en el siglo posterior a la conquista, gracias al reestudio de materiales cenimicos y a la documentaci6n de objetos de preciso valor datante, entre ellos monedas. Este proceso de amortizaci6n subsiguiente a la conquista acaso puede reconocerse en otras regiones de la Peninsula (Femandez Ugalde 1994). 5 Analisis del material organico Habitualmente, la ultima funci6n de Ios silos medievales es la de contenedores de desperdicios domesticos. Por ello, una de las principales ventajas de su excavaci6n es la de proporcionar abundantes contenidos de materia organica, generalmente en buen estado de conservaci6n. Sin embargo, no se conocen en el reino de Toledo ejemplos de silos medievales que conserven su contenido de cereal u otras reservas alimenticias. Con todo, la f1otaci6n sistematica de los estratos de relleno permite la recuperaci6n de restos vegetales conservados por carbonizaci6n y mineralizaci6n. En particular, la excelente conservaci6n de los restos mineralizados ha permitido recuperar frutos enteros y numerosas semillas de 295 E. Serrano, A. Fermindez Ugalde & L. Pena-Chocarro diferentes especies. Este tipo de restos aporta una informaci6n raramente asequible a traves de restos conservados a partir de otro tipo de procesos. Un estudio preliminar de Ios restos vegetates de varios silos reutilizados como basureros del Madrid medieval sefiala el potencial de Ios estudios arqueobotanicos. Los restos vegetales recuperados en este tipo de contextos proporcionan una valiosa informaci6n a la hora de reconstruir no solo la dieta, sino tambien algunas de !as actividades relacionadas con la preparaci6n de alimentos, procesado de cereales y utilizaci6n de plantas con otros fines diferentes a la alimentaci6n. Los analisis faunisticos de restos contenidos en estratos de relleno estan asimismo proporcionando interesantes resultados, especialmente en lo referente a la avifauna, con indicios de cetreria y observaci6n de especies ausentes hasta la fecha del registro arqueol6gico peninsular (Hem{mdez Carrasquilla 1991, 1993). Bibliografia BoLENS L. 1979: La conservation des grains en Andalousie medievale d'apres les traites d'agronomie hispano-arabes, en: GAST M. & SIGAUT F. (eds.), Les techniques de conservation des grains along terme. Leur role dans la dynamique des systemes de cultures et des societes 1, Paris, 105-112. FERNANDEZ UGALDE A. 1994: El fen6meno del relleno de silos y la implantaci6n del feudalismo en Madrid y en el reino de Toledo, IV Congreso de arqueologia medieval espafiola, vol. Ill, Alicante, 611-617. FERNANDEZ UGALDE A 1996: Excavaciones en la Cava Baja, 38, 32 y 30 de Madrid: muralla del 'se gun do recinto' y vestigios del poblamiento medieval, Reunion de arqueologia madrilefia, Madrid, 21-29. FERNANDEZ UGALDE A. e.p.: jQue Dios nos conserve el grano! Una interpretaci6n de Ios omoplatos con inscripci6n arabe procedentes de yacimientos medievales, Al-Qantara. FERNANDEZ UGALDE A. & SERRANO E. 1997 e.p.: Las murallas de Madrid: excavaciones recientes y apuntes para su evoluci6n, Estudios de prehistoria y arqueologia madrilefias 10. GorTEIN S. 1983: A Mediterranean Society. The Jewish Communities of the Arab World as Portrayed in the Documents of the Cairo Geniza, vol. 4, Berkeley-Los Angeles-Londres. HERNANDEZ CARRASQUILLA F. 1991: Las aves del yacimiento de Angosta de Ios Mancebos (Madrid), Boletin de arqueologia medieval 5, 181-191. 296 HERNANDEZ CARRASQUILLA F. 1993: Una fauna medieval inusual: !as aves de Cava Baja (prov. de Madrid, Espafia), Archaeofauna 2, 169-174. IBN AL-AWWAN 1988: Libra de agricultura. Su autor, el Doctor excelente Abu Zacaria Iahia Aben Mohamed ben Alzmed Ebn el Awam, sevillano, Ed. facsimil de la de J.A. Banqueri (1802), 2 vols, Ministerio de Agricultura, Madrid. LEFEBURE C. 1985: Reserves cerealieres et societe: !'ensilage chez les marocains, en: GAST M., S!GAUT F. & BEUTLER C. (eds.): Les techniques de conservation des grains along terme 3' fasc. 1' Paris, 210-235. REYNOLDS P.J. 1979: A general report of underground grain storage experiments at the Butser Ancient Farm Research Project, en: GAST M. & SIGAUT F. (eds.): Les techniques de conservation des grains a long terme. Leur role dans la dynamique des systemes de cultures et des soci13tes 1, Paris, 70-78. SERRANO E. 1996: Intervenci6n arqueol6gica realizada para la canalizaci6n de GASNATURAL en cl Hileras-Pza. de Ramales, Reunion de arqueologia madrilefia, Madrid, 71-73. SERRANO E. & Y Mmz G.I. 1996: Intervenci6n arqueol6gica en el inmueble de c/ del Rollo n° 7 (Madrid), Reunion de arqueologia madrilefia, Madrid, 74-76. VALLESPiN 0., SERRANO E., L6PEZ MARCOS M.A. & MARiN F. 1990: Excavaciones en el solar 'Casa de San Isidro', Madrid del siglo IX al XI, Madrid, 287-296. Y ANEZ G.l., SERRANO E. & LOPEZ MARCOS M.A. 1992: La Capilla del Obispo, Arqueologia, paleontologia y etnografia 3, 277-318. Elena Serrano T.A.R., Madrid Antonio Femandez Ugalde AREA Sociedad Cooperativa Arqueologica cl Villa, 3 - 2° D 28005 Madrid Leonor Pefia-Chocarro Laboratorio de Arqueozoologia, Universidad Aut6noma de Madrid Espana Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Jiirg Tauber Landliche Siedlu.ngen in der Nordwestschweiz von der Merowingerzeit bis zu.m Mittelalter Archaologische Qu.ellen u.nd historische Interpretation Die Landschaft, mit der ich mich im folgenden befasse, ist das linksrheinische Hinterland von Base!, d. h. jene Gegend, die si.idlich an den Rhein angrenzt und sich bis zum Kamm des Juragebirges erstreckt 1• In diesen Raum greift von Norden her die Oberrheinische Tiefebene ein, zu welcher der Unterlauf der Birs von Aesch bis zur Mi.indung in den Rhein geh6rt. Westlich schliessen die si.idostlichsten Auslaufer des Sundgauer Hi.igellandes an, ostlich das Tal des hier von Osten nach Westen fliessenden Hochrheins. Der grosste Teil des Untersuchungsgebietes ist jedoch dem Juragebirge zuzurechnen. Der Tafeljura schliesst an das Rheintal an und ist gekennzeichnet durch fruchtbare Hochflachen und tief eingeschnittene Tal er. Gegen Si.iden grenzt er an den Ketten- oder Fattenjura, der aus ost-west-verlaufenden steilen Gebirgszi.igen besteht. Die zum Gegenstand meiner Betrachtungen gewahlte Region ist somit nicht ein vollig willki.irlich ausgewahltes Gebiet, sondern kann als in sich geschlossenes Siedlungsgebiet betrachtet werden. Die Zeitspanne, welche die Ausfi.ihrungen umfassen sollen, reicht vom Ende der Antike bis ins Mittelalter. Das Einsetzen in oder am Ende der Romerzeit ist nicht einfach eine wissenschaftliche Konvention, sondern ist durch tatsachlich existierende Traditionen vorgegeben: wir befinden uns im Hinterland der Romerstadt Augusta Raurica bzw. des Castrum Rauracense 2 . Dieses Hinterland war erschlossen durch ein dichtes Netz romischer Gutsh6fe, sowie Infrastrukturbauten wie Strassen und die romische Wasserleitung von Lausen nach Augst. Nach einem Erdbeben urn die Mitte des 3. Jahrhunderts und den wohl auch hier nicht folgenlosen Alamanneneinfallen gingen die Bevolkerungszahlen und damit auch die Bedeutung der Stadt massiv zuri.ick. Das Gebiet wurde aber nicht aufgegeben, sondern in die Anstrengungen zur Sicherung des Reichsgebietes im 4. Jahrhundert einbezogen; verwaltungstechnisch wurde es der Maxima Sequanorum zugeschlagen. Von dieser Konstellation haben wir auszugehen, wenn wir uns mit der Siedlungsgeschichte des fri.ihen und hohen Mittelalters befassen. Noch eine Vorbemerkung zum Inhalt meiner Ausfi.ihrungen: es handelt sich nicht urn die Resultate eines einzelnen Forschungsprojektes, sondern urn einen Bericht zum Stand der Arbeiten in einer ganzen Reihe verschiedener parallellaufender Untersuchungen. Aufhanger ist die in den spaten 80er und fri.ihen 90er Jahren durchgefi.ihrte Grabung in Lausen-Bettenach bzw. die laufende Detailauswertung dieser Grabung (Schmaedecke & Tauber 1992; Schmaedecke 199 5) 3• Ein historisch ausgerichtetes Projekt, das die Einbindung der in Lausen zum Vorschein gekommenen Siedlung im historischen Umfeld zum Ziel hatte, ist eben zu Ende gegangen 4 . Ausserdem werden Ergebnisse einfliessen, die im Rahmen einer Dissertation zur fri.ihmittelalterlichen Siedlungsgeschichte erarbeitet werden5 . Schliesslich werden auch Aspekte der Burgenforschung in der Nordwestschweiz beri.icksichtigt. Das Ziel ist nicht eine Zusammenstellung der archaologisch erforschten landlichen Siedlungen 6 ; !m wesentlichen beschaftige ich mich mit dem Gebiet des Kantons Basei-Landschaft, das gegen 520 Quadratkilometer umfasst, sowie einzelnen Ausblicken auf die westlich (Kanton Jura) und ostlich (Fricktal, Kanton Aargau) anschliessenden Gebiete. 2 D;e Oberreste der romischen Stadt verteilen sich auf die Gemeinden Augst (Kt. Basel-Landschaft) und Kaiseraugst (Kt. Aargau), jene des Kastelles liegen in Kaiseraugst. 3 Die Auswertungsarbeiten werden von Michael Schmaedecke koordiniert. Da sie jedoch nicht durch ein eigenes Projektteam durchgefiihrt werden konnen, sondem neben dem "courant normal" vonstatten gehen milssen, kommen sie !eider nicht in der erwilnschten Geschwindigkeit voran. Der Titel des Projektes lautet "Lausen-Bettenach- ein frilhund hochmittelalterlicher Herrenhof und seine historische Einbindung in das Hinterland einer ehemaligen romischen Grossstadt. Studien zur genetischen Siedlungsforschung". Es wird ftnanziert vom Schweizerischen Nationalfonds zur Forderung der wissenschaftlichen Forschung. 5 Dissertation von Reto Marti unter der Leitung von Max Martin. Allgemeine Aspekte landlicher Siedlungen in der Schweiz behandelt Michael Schmaedecke in der schriftlichen Fassung eines Referates, das er in Prag (Ruralia 1996) gehalten hat (Schmaedecke im Druck). 297 J. Tauber vielmehr mochte ich zu zeigen versuchen, wie aufgrund entsprechender Grabungen und ihrer Interpretation neue Impulse fUr die Geschichte gegeben werden konnen. tion in einem grossen Teil unserer Region zur Beseitigung von Traditionen geftlhrt, die unter Umstanden altere Patrozinien noch batten i.iberliefem konnen. Archaologische Quell en Zur Quellenlage Schriftliche Quellen Traditionellerweise werden bei der Erorterung historischer Themen die schriftlichen Quellen als erste behandelt. Der Bestand an Urkunden und anderen Schriftsti.icken, die sich konkret auf die hier zur Diskussion stehende Region beziehen, ist jedoch so liickenhaft und zuHi.llig, dass ein darauf aufbauender Entwurf zu einer Siedlungsgeschichte als spekulativ oder gar unmoglich erscheinen muss. Die Quellenarmut hat dazu gefllhrt, dass die Gegend am Rheinknie als weitgehend unbedeutend, als Machtvakuum weitab der wichtigen Schaupliitze und Heerstrassen der Geschichte betrachtet wurde. Diese Einschatzung betrifft auch das Bistum, das zunachst in Augst, dann in Base! angesiedelt war und bis urn die Jahrtausendwende merkwOrdig schemenhaft erscheint?. Ortsnamen Der Bestand an Orts-(und Flur-)Namen wird geme zur Nachzeichnung von Siedlungsentwicklungen herangezogen. In unserer Region ist schon mehrfach ein starkes Substrat vorgermanischer Ortsnamen konstatiert worden, das ein Uberleben romanischer Traditionen erschliessen lasst. Daneben kann ein reicher Namensbestand von "klassischen" germanischen Ortsnamen auf -ingen, -inghofen (=-ikon), -wil, -dorfund ausnahmsweise -heim nachgewiesen werden. Frtiher wurden diese Namen stark chronologisch (und ideologisch) beladen und klar trennbaren Stufen des Landesausbaus seit der "alamannischen Landnahme im 5. Jahrhundert" zugewiesen. Heute ist man in dieser Hinsicht vorsichtiger geworden 8. Kirchenpatrozinien Ein weiteres Indiz fur die (Re)Konstruktion von Siedlungsvorgangen konnen Kirchenpatrozinien sein (Wittmer-Butsch 1995). Ich erinnere an die MartinsPatrozinien, die oft mit frankischem Einfluss in Verbindung gebracht werden. Allerdings ist diese Gruppe historischer Indizien nicht unproblematisch. Zum einen sind !angst nicht alle Kirchen schon im Hochmittelalter erwahnt, zum anderen hat die Reforma- 298 Dass die bei Ausgrabungen zutage geforderten Befunde und Funde in der historischen Forschung einen hohen Stellenwert einnehmen, wird heute auch von eingefleischten Schriftquellen-Enthusiasten kaum mehr bezweifelt. Vor all em die Tatsache, dass immer neue Quellen erschlossen werden, lasst die Kolleginnen und Kollegen der Nachbarwissenschaften vor Neid erblassen. Ein Blick auf die Verhaltnisse in der Nordwestschweiz Hisst erkennen, dass der Zuwachs an archaologischen Quellen in den letzten 15 Jahren enorm war. Insbesondere filr die Zeit zwischen dem Ende der Romerzeit und der Jahrtausendwende hat die Zahl neuer Befunde und Funde derart zugenommen, dass verschiedene altere Thesen widerlegt oder aber durch Fakten massiv untermauert werden konnten. Ich mochte im folgenden versuchen, die verschiedenen Quellenarten zu kombinieren und aufzuzeigen, wie auf diese W eise neue Erkenntnisse oder Thesen erarbeitet werden konnen, welche das Geschichtsbild erweitem oder verandem, und zwar nicht nur, was die blosse Feststellung der Existenz landlicher Sied1ungen und der darin vorherrschenden Lebensweise angeht, sondem auch in herrschafts- oder besitzgeschichtlicher Hinsicht. Man darfnicht vergessen, dass jede landliche Siedlung in ein Netz von Beziehungen und Abhangigkeiten eingebettet ist, die nicht unterschlagen werden di.irfen. Ausserdem muss man sich immer wieder vor Augen halten, dass abgesehen von wenigen punktuell als eindeutig stadtisch definierbaren Siedlungen grundsatzlich jede Siedlung zunachst einmal als landlich zu gelten hat. Kontinuitiit und Wandel zwischen Romerzeit und Mittelalter Dass die Romerzeit nicht mit einem Schlag zu Ende ging, ist seit langerem bekannt9 . In den Kastell- Die Quell en zum frtihen Bistum und das historische Umfeld beleuchtet Btittner 1939. Ausgewahlte Beispiele fOr die Aussagemi:iglichkeiten von Ortsnamen bei Marti 1995. Dort auch Verweise aufweiterfOhrende Literatur. Ich stiitze mich bei den folgenden Ausfiihrungen v.a. auf die Arbeiten von und Diskussionen mit Reto Marti, der dieses Uindliche Siedlungen in der Nordwestschweiz von der Merowingerzeit bis zum Mittelalter nekropolen von Kaiseraugst und Basel-Aeschenvorstadt liess sich eine kontinuierliche Belegung bis ins 7. Jahrhundert seit langern nachweisen (Marti 1995, 11 ). Auch der vorgermanische Ortsnamensbestand wurde schon rnehrfach als Indiz fUr ein Oberleben romanischer Inseln irn Hinterland von Augst herangezogen (Marti 1995, 9). Unklar war bis vor wenigen Jahren, ob si eh dieses Oberleben auf die unrnittelbare Urngebung der Kastelle beschrankte oder noch etwas weiter ins Hinterland hinein fassbar sein konnte. Funde der letzten Jahre haben nun gezeigt, class der rornanische Siedlungsraurn doch etwas grosser gewesen sein muss als frtiher angenornrnen, auch wenn nicht zu bestreiten ist, class zahlreiche rornische GutshOfe aufgegeben oder nur noch in reduzierter Weise weiterbenutzt wurden. Die gesicherten Funde des 4., 5. oder gar 6. Jahrhunderts sind in der Umgebung der Kastelle und an den Hauptverkehrsachsen nachzuweisen, wahrend aufgrund der heutigen Kenntnisse die Siedlungen in den abgelegeneren Talem zeitweise abzubrechen scheinen. Rornische Traditionen tiber das 4./5. Jahrhundert hinaus sind sehr schwierig nachzuweisen, da die in frtiheren Zeiten so zuverlassigen Datierungssttitzen wie irnportiertes Luxusgeschirr oder Mtinzen nach 400 zunehmend ausfallen. Deshalb brauchte es einige Modellbefunde, welche die Lticken ftillen halfen. Einer dieser Befunde ist ein Grubenhaus der Zeit urn 600 in Reinach (Marti 1990). Die in der Verftillung gefundene Kerarnik zeigt, class bis urn diese Zeit "rornische" Keramik in Form rauhwandiger Drehscheibenware produziert wurde. Besonders eindrticklich lasst sich eine Kontinuitat antiker Traditionen in Lausen-Bettenach nachweisen. Ich werde auf diesen Fundort zurtickkornmen. Kontinuitat schliesst aber einen Wandel nicht aus, und so lasst sich- vorausgesetzt, es steht genug Material zur Verftigung- auch bei dieser rauhwandigen Drehscheibenware eine Entwicklung beobachten (Marti 1994). Eine Zasur ist erst im 7. Jahrhundert fassbar: jetzt setzt rnit einer sandigen Drehscheibenware eine deutlich unterschiedliche Keramik ein, die auf andere Traditionen zurtickgeht. Man ist versucht, sie rnit gesellschaftlichen Veranderungen in Beziehung zu bringen. Dass diese Kerarnik eindeutig eine einheirnische Ware ist, zeigen Funde von TopferOfen in Oberwil (Steinle/Tauber 1974), Therwil und Reinach (Tauber 1988; Marti 1990). Auch in Thema im Rahmen seiner Dissertation vertieft behandelt. Bisher greitbare einschlagige Artikel sind etwa Marti 1994 (zur Entwicklung der Keramik) und Marti 1995 (methodisch orientierter Oberblick mit einer knappen Zusammenfassung des Forschungsstandes; dart auch weiterfilhrende Literaturangeben). den Befunden zeitlich entsprechender Siedlungen tiberwiegt diese Ware; daneben findet sich haufig auch die altere gelbtonige Drehscheibenware, die wohl aus dern Elsass starnrnt. Der in den vergangenen Jahren erfreulich forgeschrittene Forschungsstand bei der Kerarnik hat erlaubt, die Siedlungsbefunde chronologisch richtig einzuordnen. Diese fruh- und hochrnittelalterlichen Spuren sind jedoch nicht sehr zahlreich. Haufig handelt es sich urn Einzelfunde und -befunde, die in einern heute noch bestehenden Dorf oder seiner unrnittelbaren Urngebung entdeckt werden. Siedlungsbefunde rnit Resten von Gebauden wie in Reinach (Marti 1990) und Aesch (unpubliziert) sind eher selten. Erstaunlich ist femer die Tatsache, class die Belege rnit fortschreitender Zeit abnehmen, d.h. Befunde und Funde des Hoch- und Spatrnittelalters sind wesentlich seltener als jene der rneorwingisch/ karolingischen Epoche! An die Stelle der Reste von landlichen Siedlungen treten von der Jahrtausendwende an vermehrt die Kirchen und die Burgen. Nicht zu vergessen sind femer die frtih- und hochrnittelalterlichen Siedlungsspuren in den Kastellen von Base! (Helmig 1982) und Kaiseraugst (Prey 1992; Marti 1997), auf die ich jedoch in unserern Zusamrnenhang nicht naher eingehen werde. Besonders hervorheben rnochte ich jedoch eine Kleinregion urn Liestal, die rnit Lausen-Bettenach und Liestal-Rosemtal zwei Fundstellen geliefert hat, die beide fUr ganze Problembtinde1 als Schlusselgrabungen gelten dtirfen. Lausen-Bettenach ist eine Siedlung, die eine Kontinuitat von der mittleren rornischen Kaiserzeit bis zum Beginn des 13. Jahrhunderts aufweist. Sie zeichnet sich durch eine ausserordentliche Befunddichte aus (Schrnaedecke/Tauber 1992; Schrnaedecke 1995). Zahlreiche Gruben und Grubenhauser sowie Reste von Pfostenbauten werden erganzt durch Steinbauten einmal des 5. (ev. 6.), aber auch des 9. oder 10. Jahrhunderts. Eine frtihe Kirche wird im 11. Jahrhundert unter radikaler Anderung der Orientierung durch eine dreischiffige Basilika ersetzt. Wohl zur selben Zeit wird auch das Siedlungsareal urngestaltet, indern die Steinbauten niedergelegt werden; an ihrer Stelle entstehen Grubenhauser, die gegen Ende des 11. Jahrhunderts abgebrochen und wieder zugeschtittet werden. Zu Beg inn des 13. Jahrhunderts wird die Siedlung allmahlich verlassen; nur die Kirche bleibt stehen. Westlich der Kirche wird ein Bau errichtet, der in der ersten Halfte des 15. Jahrhunderts bereits wieder aufgegeben wird, wohl irn Zusamrnenhang rnit dern Anbau des heutigen "Sigristenhauses" an die Westfassade der Kirche. Die Kirche- es handelte si eh bis zur Reformation urn eine Nikolauskirche- blieb Pfarrkirche des jenseits 299 J. Tauber der Ergolz auf der anderen Talseite liegenden Dorfes Lausen. Neben den Steinbauten weisen auch Funde auf eine Bewohnerschaft mit gehobenem Standard hin: Hufnagel und Hufeisen dokumentieren die haufige Anwesenheit von Berittenen, Funde von kostbaren Kreuz- und Emailscheibenfibeln fehlen ebensowenig wie ein Stilus. Die Liste aussergewohnlicher Funde liesse sich verlangem und entspricht jedenfalls nicht dem Spektrum, das in einer rein Hindlichen Siedlung erwartet werden wilrde. Der Fundort gibt also Gelegenheit, im Vergleich mit anderen Siedlungen Fragen zur sozialen Gliederung der Gesellschaft und zur Funktion der einzelnen Fundstellen zu diskutieren. Etwas anders gelagert sind Liestal-Munzach und Liestal-Rosemtal: Die Siedlung Munzach liegt im Area! des romischen Gutshofes. Sie zeigt eine Koninuitat wie Lausen-Bettenach. Allerdings liegen die Grabungen einige Jahrzehnte zuri.ick und wurden auch ftir damalige Verhaltnisse nicht sehr fachmannisch durchgefi.ihrt, sodass die Dokumentation sehr zu wilnschen i.ibrig Hisst. Auch auf dem Area! der Villa Munzach entsteht aufantiken Wurzeln eine fri.ih- und hochmittelalterliche Siedlung, die zu Beginn des 13. Jahrhunderts aufgegeben wurde. Hier stand die in Spuren nachgewiesene Kirche St. Laurentius, die erst im 19. Jahrhundert abgebrochen wurde. Die Funde weisen aber ein deutlich bescheideneres Spektrum auf als jene von Lausen. Die Siedlung im Rosemtal hingegen ist eine Neugri.indung der Zeit urn 800 (Tauber 1993; Lavicka 1995): Die alteste Keramik entspricht einer spaten Variante der sandigen Drehscheibenware. Noch im Laufe des 12. Jahrunderts wird der Platz wieder aufgegeben. Was ihm zu einer Bedeutung i.iber das Ubliche hinaus verhilft, sind die nachgewiesenen handwerklichen Tatigkeiten. Neben Spuren von Glasund Buntmetallverarbeitung sind es vor allem Dberreste der Verhi.ittung von Eisenerz und der Verarbeitung des gewonnenen Eisens (Semeels 1995). Mehrere Tonnen Schlacken wurden wahrend der Ausgrabung geborgen und analysiert; sie belegen die Verhi.ittung verschiedener Erze in RennOfen sowie mehrere darauf folgende Abeitsschritte bis hin zur Fertigung von Geraten. Auffallend ist, dass diese Siedlung deutlich von der gleichzeitigen Siedlung auf dem Area! der ehemaligen Villa getrennt ist. Hi er lassen sich Dberlegungen zu wirtschaftlichen Aspekten ausserhalb der Landwirtschaft im engeren Sinn sowie Fragen zur Arbeitsteilung ankni.ipfen. In diesem Zusamenhang zu erwahnen ist, dass auch in Lausen Eisenverhi.ittung durch Schlacken nachgewiesen ist. Eine weitere fri.ihmittelalterliche Sied300 lung, in der neben Landwirtschaft auch Eisenverarbeitung eine Rolle spielt, liegt in den Gemeinden Develier und Courtetelle (Kt. Jura); dieser Fundort weist einen vollig anderen Charakter auf, mit kleinen, ilber grossere Distanzen gestreuten Siedlungskernen (Fellner et al. 1995; Schenardi 1995; Fellner & Schenardi 1995). Zwischen Lausen-Bettenach und Liestal-Munzach bzw. Roserntal liegt Liestal, auf das ich ebenfalls noch kurz eingehen mochte. Der Ort wurde im fri.ihen 13. Jahrhundert durch die Grafen von Frohburg zur Stadt erhoben, geht aber ebenfalls aufaltere Wurzeln zuri.ick. Der Kern der Altstadt, d.h. das Geviert urn die Kirche, die St. Martin und Brida geweiht war, di.irfte seinen Ursprung in einem romischen Kastell haben, wie nicht nur der Grundriss, sondern auch Funde aus der mehrere Jahrzehnte zuri.ickliegenden Grabung im Kircheninnern erschliessen lassen (Marti 1988) 10 • Weitere Spuren des romischen oder fri.ihbzw. hochmittelalterlichen Liestal sind ausserhalb der Kirche allerdings ausgeblieben, da das Gelande in der Fri.ihen Neuzeit offenbar auf dem ganzen Stadtareal abgetragen worden ist. Immerhin sind seit kurzem Funde des 10. und 11. Jahrhunderts bekannt; ihre Fundlage in einem Hauseingang sechzig Zentimeter i.iber dem heutigen Strasseniveau bestatigt die These der Gelandeabsenkungen. Die Konstellation in Liestal bringt uns zur Problematik von Besitzgeschichte und Herrschaftsbildung, ausserdem weist sie auf eine weitere Quellengattung hin, die siedlungsgeschichtlich von hervorragender Bedeutung ist: die Kirchen. Auch wenn es mitunter schwierig ist, allein aufgrund von Grundrissen eine einigermassen prazise Datierung vorzuschlagen, so zeigen mehrere Beispiele mit beigabenfi.ihrenden Grabern, dass der Bau von Kirchen schon im 7. Jahrhundert einsetzt und die auf spatantike Wurzeln zuri.ickgehende Anlage in Kaiseraugst erganzen. Mit der einzigen bisher nachgewiesenen holzernen Ausnahme von Buus (Ewald 1996) bestehen in der Nordwestschweiz alle fri.ihmittelalterlichen Kirchen aus gemorteltem Mauerwerk (Ewald 1991). Wenn wir die fri.ihmittelalterlichen Siedlungsspuren und Kirchen kombinieren, erhalten wir zwar eine erfreuliche Anzahl von Einzelbelegen, aber kein geschlossenes Bild. Eine wichtige archaologische Quellengattung fehlt namlich: die Graber und Gra- 10 Neben der rauhwandigen Drehscheibenware vor allem Argonnensigillata sowie ein Fragment der Sigillee Paleochretienne. Ltindliche Siedlungen in der Nordwestschweiz von der Merowingerzeit bis zum Mittelalter Abb. 1.- Frilhmittelalter!iche Siedlungsfunde (5.-9. Jh.). 301 J. Tauber berfelder. Die Verbreitungskarten der Siedlungsbefunde und der Grabfunde machen deutlich, wie unverzichtbar die Betrachtung der Grabfunde ist (Abb. 1 und 2). Fur die Siedlungsgeschichte (Marti 1995, mit weiterflihrender Literatur) von Bedeutung ist die Moglichkeit einer zeitlichen Differenzierung: Neben den Grabem des 5. Jahrhunderts in den rechtsrheinischen alamannischen Graberfeldem von BaselKleinhtiningen, Base1-Gotterbarmweg und Herten sowie zeitlich entsprechenden Grabem in den Kastellnekropo1en von Kaiseraugst und Base! zeigen sich links des Rheins erst im Laufe des 6. Jahrhunderts nach germanischer Sitte mit Beigaben versehene Graber, etwa in Basel-Bemerring, Therwil und Liestal-Radacker - wohl nicht zufallig in der Nahe spatankiker Zentren oder wichtiger Verkehrswege. Die Funde in diesen Grabem zeigen tiberdies, class die frtihesten im linksrheinischen Gebiet fassbaren Germanen Franken oder doch stark frankisch beeinflusst waren. Erst im 7. Jahrhundert intensiviert sich der Landesausbau, die Dichte der Graberfelder wachst. An diesem Bild konnten durch Neufunde allenfalls noch Retuschen angebracht werden; angesichts der bereits heute mehr oder weniger flachendeckenden Erfassung unseres Arbeitsgebietes dtirfen wir jedoch davon ausgehen, class das heute greifbare Material als reprasentativ gelten kann. Interessant ist die Beobachtung, class tiber die Grabbeigaben Beziehungen Richtung Norden ins Elsass bzw. ins Gebiet des stidlichen Oberrheins sowie Richtung Westen und Stidwesten (also in den Bereich der spatantiken Maxima Sequanorum) erschlossen werden konnen, wahrend Kontakte ins rechtsrheinische Kemgebiet der Alamannen wesentlich seltener nachzuweisen sind. Allein aufgrund dieser archaologischen Befunde zeigt unser Arbeitsgebiet bereits ftir das 7. bis 9. Jahrhundert eine erstaunlich dichte Besiedlung. Sie umfasst zunachst zwar noch nicht die maximale Ausdehnung der in romischer Zeit besiedelten Flache, sondem nur die besseren Lagen an Verkehrswegen, auf fruchtbaren Talboden und in der Nahe der Zentren (z.B. der Kastelle), greift aber bald auch auf eher marginale Gebiete aus, die in romischer Zeit nicht besiedelt waren. Beispiele hieftir sind etwa Eptingen und Reigoldswil, zwei Dorfer ganz am oberen Ende von tief in die Berge des Faltenjura eingeschnittenen Talem. 11 Nur aufmerksamsten Beobachtungen ortsansassiger Interessierter ist es zu verdanken, dass hin und wieder entsprechende Funde aufgelesen werden wie jungst in Rothenfluh, wo beim in Quellen der frilhen Neuzeit genannten Hof "Wcrdlingcn" cinzelne Scherben des 6.17. und des 8. Jahrhunderts zum Vorschein kamen (Rippmann 1996). 302 Dieses Bild kann durch die Betrachtung der Ortsnamen noch verdichtet werden; es zeigt sich, class diese drei Jahrhunderte als Zeit eines intensiven Landesausbaus verstanden werden mtissen (Marti 1995). Bis zur Jahrtausendwende scheint sich an diesem Bild wenig geandert zu haben. Es kann vermutet werden, class der Landesausbau in etwas verlangsamter Form fortgeftihrt wurde, doch fehlen uns mangels Beigaben in den Grabem und vor allem mangels Funden aus den Dorfem selbst entsprechende archaologische Hinweise 11 • Soweit das aufgrund der archaologischen Quellen und der Ortsnamen gewonnene Siedlungsbild, das zeigt, class die Anzahl der landlichen Siedlungen schon vor der Jahrtausendwende im wesentlichen bereits eine Zahl erreicht haben muss, die von der heutigen nicht sehr stark abweicht. Dies heisst aber nicht, class eine gewaltige Bevolkerungsvermehrung postuliert werden soil, denn man muss davon ausgehen, class die Siedlungen sehr klein waren. Die Anbauflachen in den zum Teil engen T~ilem batten einer grossen Bevolkerung keine gentigende Lebensgrundlage liefern konnen, und man ist versucht zu fragen, was die Menschen bewogen hat, sich in den unwirtlichen Talkesseln niederzulassen. Ein Grund (selbstverstandlich unter anderen) konnten die reichen Vorkommen an Eisenerz gewesen sein, dessen Ausbeutung im Jura nachweislich im Frilhmittelalter auf breiter Front einsetzt (Tauber 1992 und 1993; Pelet 1993; Eschenlohr/Semeels 1991; Semeels 1993 ). Versuch einer historischen Einbindung Urn die archaologischen Ergebnisse historisch fruchtbar zu machen - und dies ist letztlich ja das Ziel unserer Arbeit - sollten sie in einen gesamtgeschichtlichen Zusammenhang gestellt werden, der i.iber den Versuch hinausgeht, die Lebensbedingungen der in den nachgewiesenen Pfostenhausem lebenden Menschen zu (re)konstruieren. Dieser Ansatz ist zwar legitim, klammert aber aus, class jede landliche Siedlung eingebunden ist in ein Geflecht aus Besitz, Macht und Herrschaft und deshalb nicht isoliert betrachtet werden kann. Die folgenden Ausftihrungen versuchen, einige Aspekte dieser Thematik zu skizzieren. Gewiss ist dies eine Arbeit mit Hypothesen, was sie auf den ersten Blick als zweifelhafter erscheinen lasst als die aufgrund der mit "handfesten" archaologischen Quellen erarbeiteten Ergebnisse; auch die Archaologie liefert aber niemals "Beweise", sondem lediglich Indizien, die von uns interpretiert und zu einem sinnvollen Ganzen zusammengefligt, im schlimmsten Fall auch zurechtgebogen werden. Uindliche Siedlungen in der Nordwestschweiz von der Merowingerzeit bis zum Mittelalter Abb. 2. - Friihmittelalterliche Grabfunde m it Beigaben. 303 J. Tauber Angesichts der sparlichen schriftlichen Oberlieferung stellen sich beim Versuch einer historischen Interpretation allerdings erhebliche Probleme. Wir konnen nicht einfach von den wenigen Urkunden ausgehen, sondem milssen auch die allgemeinhistorischen Gegebenheiten berilcksichtigen. Das Fehlen von Urkunden ist namlich nicht zwingend mit einem siedlungsgeschichtlichen "Vakuum" gleichzusetzen, da sie in den meisten Fallen Handanderungen betreffen. Ihr Ausbleiben kann deshalb auch anzeigen, class die Besitzverhaltnisse in der entsprechenden Region stabil blieben. Angesichts des dichten Netzes von Siedlungen, die archaologisch nachgewiesen werden konnen, sowie der Bedeutung der Jurapasse, insbesondere des Oberen Hauensteins, mochten wir eher dieser Interpretation den Vorzug geben. Betrachten wir zunachst die "politische Grosswetterlage": Die zur Diskussion stehende Region fie! im Vertrag von Verdun 843 ans Mittelreich, injenem von Meersen 870 ans ostfrankische Reich. Noch im 9. Jahrhundert aber fie! sie dem Konigreich Hochburgund zu. Erst 1033 gelangte sie mit diesem ans Reich. Dass dies aber nicht mit einer Losung der burgundischen Bindungen einherging, zeigen eindrticklich die Verhii.ltnisse im Investiturstreit: Sowohl der Basler BischofBurkard von Fenis (Stammburg zwischen Bieler- und Neuenburgersee) als auch sein Widersacher Rudolfvon Rheinfelden (bzw. von Schwaben), also der Gegenkonig hOchstpersonlich, stammen beide aus burgundischen Hochadelsge. schlechtem. Neben Beziehungen zu Burgund bestanden auch starke Bindungen zum Elsass, die in verschiedenen Aspekten greifbar sind: Ottilia, die Tochter Etichos, vermacht in ihrem Testament 708 die Curia von Arlesheim mit zugehOrigen Gtitem in elsassischen Dorfem dem Kloster Hohenburg. Das Dorf Muttenz mit seiner St. ArbogastKirche und wohl auch der vorderen Burg auf dem Wartenberg (die in karolingische Zeit zurilckgeht) ist nicht nur durch dieses Patrozinium mit Strassburg verbunden; es befindet sich auch im Besitz des dortigen Bistums, wie aus der Lehenshoheit in spateren Zeiten ersichtlich ist. Unklar ist lediglich, wann und wie es in den Besitz dieses Bistums gelangte. Besitz an wichtigen Stellen, insbesondere an der Passstrasse zum Oberen Hauenstein, ist auch filr das Kloster Murbach belegt. 12 Heinrich Boos, Urkundenbuch der Landschaft Base/ I, Base!, 1881, Nr. 8. 13 Heinrich Boos, Urkundenbuch der Landschaft Base/ I, Base!, 1881, Nr. 12. 14 WARTMANN, Urkundenbuch der Abtei St. Gal/en I, Nr. 191. 304 Dass das Bistum Base! bis zur Jahrtausendwende wenig bedeutend war und offensichtlich nicht vermochte, sich eine Machtbasis zu schaffen (Btittner 1939), konnte mit der starken Prasenz anderer Machte zusammenhangen. Neben dem Bistum Strassburg und allenfalls den elsassischen Herzogen kamen auch das merowingisch/karolingische Konigtum und seine Rechtsnachfolger in Frage. Auffallend ist die Existenz gleich dreier Fundorte (Lausen-Bettenach, Liestal, Munzach) mit einer archaologisch nachgewiesenen Kontinuitat von der Spatantike bis ins Hochmittelalter auf engstem Raum urn Liestal, das mit einiger Wahrscheinlichkeit aus einem spatromischen Kastell hervorging. LausenBettenach liegt am oberen Ende der romischen Wasserleitung, die an der rechten Talflanke an Liestal und Munzach vorbei nach Augst filhrt. Alle drei Orte saumen wichtige Verkehrswege vom Rhein in Richtung der Jurapasse. Die starken antiken Traditionen im Hinterland von Augst und die Tatsache, class die frilhesten linksrheinisch nachweisbaren Germanen Franken waren, lasst die Hypothese zu, class reichlich vorhandenes romisches Fiskalgut zu merowingischem Konigsgut geworden ware (zu diesem Problem siehe Kaiser 1993). Dass Konigsgut vorhanden gewesen sein muss, zeigt etwa die Schenkung 12 Kaiser Amulfs von 891 an Anno: sie betrifft die Kirche und sieben Hufen in Augst (=Kaiseraugst, dem Standort des spatromischen Castrum Rauracense und frilhen Bischofssitzes). In dieselbe Richtung weist auch die Schenkung Heinrichs Ill. an das B is turn Base! von 1041: Heinrich schenkt dem Bistum die Grafschaft Augst im Augst- und im Sisgau (comitatum Augusta vocatum in pago Ougestgowe et Sisgowe situm) 13 . Die Konigsgut-Hypothese mochten wir mindestens flir Liestal und Lausen-Bettenach in Anpruch nehmen, was aufgrund der aussergewohnlichen Lausener Funde nicht abwegig erscheint. In Munzach konnten die Dinge etwas anders liegen: in einer zwar undatierten, aber aufgrund verschiedener Indizien gegen 790 angefertigten Zeugenliste wird es als Gerichtsort erwahnt (Wittmer-Butsch 1995). Dies weist auch Munzach eine ilber das Mass des Durchschnitts hinausgehende zentralortliche Bedeutung zu. In einer weiteren Munzach betreffenden Urkunder4 von 824/25 zeigt sich, class das postulierte Konigsgut im Tal der Ergolz nicht flachendeckend gewesen sein kann: Dort schenkt namlich ein sonst nicht nachgewies~ner Uppert ein Drittel seiner Gilter in Filllinsdorf und Munzach, die er vom Vater und der Mutter geerbt hat, an das Kloster St. Gallen. Wir mochten in unserem Fundort Rosemtal, dem "Schmiededorf', das urn diese Zeit gleich neben dem Dorf Munzach gegrilndet wird, eine Einrichtung dieses Klosters sehen (Tauber 1995, 61). Ae Ar Au Ba Bu CO De Aesch Arlesheim Augst Base! Buus Ep Fii He Ka La Li Mu Courtt~telle Develier Eptingen Fiillinsdorf Herten Kaiseraugst Lausen Lies tal Muttenz Mz Ob Re Rw Rh Th Munzach Oberwil Reinach Reigoldswil Rheinfelden Therwil r "'' e: c;· ;:l ::::r(1) ;... 0" ?" V. ~ ..., ~ ~ ~- ~ ~ ~ ~ ;:: I:> ;:: ;:: ~ ;:: a ~ . _., ...------..---------, 10 2Qkm A N. C/J o;· p.. 2" ;:l (IQ (!) ;:l s· ~ ,..., z0 a. ~ (!) ~ en () ::::r~ (1) N. < 0 ;:l p.. (1) ,..., <" (;"' 0 ~ s· [JQ ,.., (l) N g.. 0"' c;;;· N c 3 3:.: ~- (1) w 0 V\ ;;;(p .... J. Tauber Ob solche Beziehungen sich im Fundgut Hindlicher Siedlungen niederschlagen, ist eine offene Frage. Anhand der drei Beispiele Lausen-Bettenach, Liestal-Munzach und Liestal -Rosemtal ist aber auffallend, dass sehr wohl Unterschiede auszumachen sind. Gewiss geht ein Teil davon auf die unterschiedlichen Grabungsbedingungen zun1ck, doch lassen sich damit nicht alle Unterschiede erkHiren. Dass In Lausen deutlich mehr Funde von hoher QualiUit i.iberliefert sind als an den anderen beiden Fundorten, ist kaum ein Zufall. Auch das haufige Vorkommen von Lavezgeffissen in der Siedlung im Rosemtal konnte auf Beziehungen zur Ostschweiz und zum Alpenraum hindeuten - vermittelt eben durch das Kloster St. Gallen. Ob sich schliesslich im haufigen Vorkommen der alteren gelben Drehscheibenware in allen drei Fundorten, schwergewichtig aber doch in Lausen, Beziehungen zum Elsass ablesen lassen, ware nur durch aufwendige Analysen des entsprechenden Materials zu eruieren. Nach der Jahrtausendwende wird der Informationsstand zu den landlichen Siedlungen nicht wesentlich besser; es sind immer noch die erwahnten Ausgrabungen im Tal der Ergolz, die am meisten Material liefem. Aus schriftlichen Quell en ist vor der Zeit urn 1100 ebenfalls nicht mehr zu erfahren 15 , und so mi.issen wir einmal mehr versuchen, die archaologischen Quellen mit den historischen Gegebenheiten zu kombinieren, urn unsere Schli.isse daraus zu ziehen. Bei der Betrachtung der fri.ihen Burgen haben wir bereits fri.iher festgestellt, dass sich mit dem Bau der drei grossen Burgen Alt-Tierstein, Alt-Homberg und Odenburg eine fri.ihe Herrschaftsbildung fassen lasst (Tauber 1995). Der Erbauer dieser Burgen ist zweifellos jener "comes Rudolfus", der 1045 in einer Urkunde erwahnt wird 16 . Wir halten ihn sowohl fi.ir einen Vorfahren der Grafen von Alt-Homberg und Alt-Tierstein, die urn 1100 a us dem Nichts auftauchen, zum hohen Reichsadel zahlen und Rudolf als dominanten Leitnamen fi.ihren, als auch fUr identisch mit Rudolf von Rheinfelden (bzw. von Schwa ben). Dies wi.irde bedeuten, dass die Grafen von Rheinfelden schon urn die Mitte des 11. Jahrhunderts fest in unserer Gegend etabliert waren. Diese Hypothese konnte gesti.itzt werden durch die Erwahnung eines Grafen im Elsass Kuno im Jahre 1052, in dem die neuere Forschung (Hlawitschka 1991, S. 202) den Vater Rudolfs annimmt. Betrachten wir unter diesen Aspekten die archiiologischen Befunde in Lausen, wo sichjust in dieser 15 Zusammenstellung der Erstcrwahnungen bei Rippmann 1995. 16 MGH 0 H m., 219. 306 Zeit markante Veranderungen beobachten lassen. Am augenfalligsten ist der Neubau einer Kirche, die nicht nur anders orientiert ist als der Vorgangerbau, sondem zu einer fi.ir eine kleine landliche Siedlung i.iberdimensionierten dreischiffigen Basilika ausgebaut wurde. Etwa gleichzeitig mit dem Kirchenneubau wurde auch die Siedlung umgestaltet (s. oben). Ein solcher Kirchenbau kann nicht einfach mit einem Wachstum der Gemeinde begri.indet werden; er ist vielmehr ein Programm. Dass damit auch die Einfi.ihrung eines neuen Patroziniums verbunden sein di.irfte, ist wahrscheinlich. Das Nikolaus-Patrozinium der Kirche von Lausen erOffnet Perspektiven, die geradezu ideal zum ganzen restlichen Umfeld passen (Wittmer-Butsch 1995, 50ff.). Demnach ware Lausen einer jener nicht seltenen Orte in der Nordwestschweiz, wo ein Nikolauskult schon vor 1087 (der Translation der Gebeine nach Bari) zu fassen ist. Unbestritten ist, class Rudolf von Rheinfelden ein grosser Forderer dieses Kultes war. Folgerichtig ware es ebendieser Rudolf von Rheinfelden gewesen, der in Lausen als grosser Umgestalter gewirkt hat. Bleibt die Frage, wie er in den Besitz dieses Hofes gelangt ist, der von uns ja als Konigshof interpretiert wird. Akzeptiert man diese These, so bietet sich eine verhaltnismassig einfache Losung an: Lausen ware 1059 als Mitgift der Kaiserstochter Mathilde, der ersten, nach einem Jahr Ehe bereits verstorbenen Frau Rudolfs in seinen Besitz gelangt. Die landliche Siedlung Lausen ist somit nicht nur als archaologisch weiterflihrende Ausgrabung zu verstehen, sondemin Verbindung mit anderen Grabungen und historischen Forschungen- als eine Schli.isselstelle fi.ir das Verstandnis der Geschichte in jener Zeit, in der wir nach wie vor mehr auf archiiologische Quell en als auf Schiriftsti.icke bauen mi.issen. Fazit Die Erforschung der landlichen Siedlungen hat in der Schweiz in den letzten Jahren einen erfreulichen Verlauf genommen (Schmaedecke im Druck). Wo fri.iher entsprechende Befunde unbeachtet zerstort wurden, erlangen sie heute die Aufmerksamkeit, die sie verdienen. Fi.ir die Forschungen der Nordwestschweiz lassen sich folgende wichtige Punkte herausschalen: 1. Die landliche Siedlung an sich gibt es nicht. Betrachtet man die in den letzten Jahren erforschten Siedlungen wie Lausen-Bettenach, Liestal-Rosemtal, Reinach-Alte Brauerei, aber auch Develier/Courtetelle, so ergibt sich eine reichhaltige Palette verschiedenartiger Auspragungen. 2. Eine wichtige Rolle bei der wirtschaftlichen Ausrichtung scheint die Eisenverhi.ittung gespielt zu Uindliche Siedlungen in der Nordwestschweiz von der Merowingerzeit bis zum Mittelalter haben, wie nicht nur die Untersuchungen entsprechender Gewerbeanlagen durch Ludwig Eschenlohr in den kantonen Jura und Bern zeigen, sondern auch die Reste von Verhi.ittungsschlacken in den fri.ih- und hochmittelalterlichen Siedlungen von Lausen, Roserntal und Develier/Courtetelle. 3. Die reine Feststellung der in den einzelnen Siedlungen feststellbaren Unterschiede bringt uns nicht weiter; wir mi.issen versuchen, diese Unterschiede auch in Richtung einer gesellschaftsgeschichtlichen Interpretation zu deuten. 4. Schriftliche und andere nichtarchaologische Quellen sind konsequent einzubeziehen und mit den archaologischen zu konfrontieren. Dies ermoglicht eine standige gegenseitige Befruchtung der Denkansatze und Hypothesen. Wie unsere Beispiele zeigen, werden damit weitgehende historische Perspektiven erOffnet. Literatur BOTTNER H. 1939: Die Landschaft urn Base! von der Einwanderung der Alamannen his zur Mitte des 8. Jahrhunderts, Vom Jura zum Schwarzwald NF, 59-82. ESCHENLOHR L. & SERNEELS V. 1991: Les bas fourneaux merovingiens de Boecourt, Les Boulies (JU, Suisse), Cahier d'archeologiejurassienne 3. EWALD J. 1991: Kitchen und Kirchengrabungen im Baselbiet. Ein Beitrag zur Geschichte der Kirchen-Landschaft der Nordwestschweiz, in: Tauber (Hsg.) 1991, 57-84. EWALD J.1996: Zur Baugeschichte der Kirche St. Michael zu Buus, Baselbieter Heimatblatter 6114, 121-136. FELLNER R., POUSAZ N. & TAILLARD P. 1995: Le site de Courtetelle, Tivila (JU, Suisse), Haut moyen age et age dufer, Fouilles 1994, (avec les contributions de M. Guelat, geologue, L. Eschenlohr, V. Friedli, C. Peter, M-H. Paratte-Rana, archeologues), Republique et Canton du Jura, Section d'archeologie de !'office du patrimoine historique et N16 39. FELLNER R. & SCHENARDI M. 1995: Le haut moyen age Develier, La Pran et Courtetelle, Tivila (JU, Suisse). Synthese- premier bilan, (avec les contributions de L. Eschenlohr, archeologue, M. Guelat, geologue, C. Peter, archeologue, S. Stekoffer, historienne de 1'art), Republique et Canton du Jura, Section d'archeologie de !'office du patrimoine historique et N16 40. FREY P. 1992: Mittelalterliche Siedlungsbefunde in Kaiseraugst, Jahresberichte aus Augst und Kaiseraugst 13, 231-262. a a HELMIG G. 1982: Fri.ihmittelalterliche Grubenhauser auf dem Mi.insterhi.igel. Ein Kapitel Basler Stadtgeschichte, Archaologie der Schweiz 5/Heft 2, 153-157. HLA WITSCHKA E. 1991: Zur Herkunft und zu den Seitenverwandten des Gegenkonigs Rudolf von Rheinfelden- Genealogische und politisch-historische Untersuchungen, in: WEINFURTER S. (Hsg.), Die Salier und das Reich 1. Salier, Adel und Reichsverfassung, Sigmaringen, 175-220. KAISER R. 1993: Das romische Erbe und das Merowingerreich, Enzyklopadie deutscher Geschichte 26,Mi.inchen. LA VICKA P. 1995: Eine Eisengewerbesiedlung des 9. his 12. Jahrhunderts in Liestal-Roserntal, in: Schrnaedecke (Bearb.) 1995, 27-34. MART! R. 1990: Bedeutende fri.ihmittelalterliche Siedlungsreste in Reinach BL, Archaologie der Schweiz 13, 136-153. MARTI R. 1994: Kontinuitat und Wandel am Beispiel fri.ihmittelalterlicher Keramik aus der Nordwestschweiz, in: Keramik zwischen den Epochen. Funktion-Formenwandel- Technik. Wissenschaftliche Fachtagung vom 19.120. August 1994 auf Schloss Miinchenwiler BE, Bern, 49-55. MARTI R. 1995: Ansatze zu einer fri.ihmittelalterlichen Siedlungsgeschichte der Nordwestschweiz, in: Schmaedecke (Bearb.) 1995,9-16 MARTI R. 1996: Fri.ihrnittelalterliche Siedlungsbefunde aus dem Castrum Rauracense (Grabung Kaiseraugst, "Jakobli-Haus" 1994.02), mit Beitragen von S. Fi.infschilling & Markus P., Jahresberichte aus Augst und Kaiseraugst 17, 149-195 PELET P-L. 1993: Une industrie reconnue: Fer Charbon Acier dans le Pays de Vaud, Cahiers d'archeologie romande 60. RIPPMANN D. 1991: Zur Geschichte des Dorfes im Mittelalter am Beispiel des Kantons Baselland, in: Tauber (Hsg.) 1991, 31-55. RrPPMANN D. 1996: Bauern und Herren. Rothenjluh im Mittelalter. Ein Beitrag zur Geschichte der landlichen Gesellschaft im Mittelalter, mit einem Beitrag von J. Tauber , Liestal. SCHENARDI M. 1995: Le site du haut moyen age de Develier, La Pran (JU, Suisse). Fouilles 1994, (avec les contributions de M. Guelat, geologue, L .. Eschenlohr, F. Lorenz, C. Peter, archeologues), Republique et Canton du Jura, Section d'archeologie de !'office du patrimoine historique et N16 38. SCHMAEDECKE M. (Bearb.) 1995: Landliche Sied!ungen zwischen Spatantike und Mittelalter. Beitrage zum Kolloquium in Liestal (Schweiz) vom 13. bis 15. Marz 1995, Archaologie und Museum 33. 307 J. Tauber SCHMAEDECKE M. 1996: Liindliche Siedlungen des frtihen und hohen Mittelalters in der Schweiz, in: Ruralia 1996, Pamatky archeologicke Supplementum 5, im Druck. SCHMAEDECKE M. & TAUBER J. 1992: Ausgrabungen in Lausen-Bettenach. Vorbericht iiber die archaologischen Untersuchungen 1985-1992, (unter Mitarbeit von J. Obrecht und mit Beitragen von H. Hi.ister und S. Jacomet), Archaologie und Museum 25. SERNEELS V. 1993: Archeometrie des scories de fer. Recherches sur la siderurgie ancienne en Suisse occidentale, Cahiers d'archeologie romande 61. SERNEELS V. 1995: Du minerai a l'objet: un village de siderurgistes du IXe au XIIe siecle a LiestalRosemtal, in: Schmaedecke (Bearb.) 1995,35-43. STEINLE S. & TAUBER J. 1974: Ein karolingischer Topferbezirk in Oberwil, Kanton Basel-Landschaft (Schweiz), Archaologisches Korrespondenzblatt 4, 181-188. TAUBER J. 1988: Ein karolingischer Topferofen in Reinach, Archiiologie und Museum 11, 69-83. TAUBER J. (Hsg.) 1991: Methoden und Perspektiven der Archaologie des Mittelalters. Tagungsberichte zum Kolloquium vom 27.-30. September 1989 in Liestal (Schweiz), Archaologie und Museum 20. TAUBER J. 1992: Zum Stand der Eisenarchaologie im Kanton Basei-Landschaft, Minaria Helvetica 12 a, 22-30. TAUBER J. 1993: Zum Stand der Eisenarchaologie im Kanton Basel-Landschaft - das Beispiel einer Handwerkersiedlung im Rosemtal bei Liestal, Archiiologische Informationen 16/2, 243-251. TAUBER J. 1995: Archaologie und Geschichte. Zur Frage der Rolle von Konigtum und Hochadel in der mittelalterlichen Siedlungsgeschichte der Nordwestschweiz, in: Schmaedecke (Bearb.) 1995, 57-67. WITTMER-BUTSCH M. 1995: Hypothesen zur Deutung ausgewahlter Quellenfragmente, in: Schmaedecke (Bearb.) 1995, 45-56. Ji.irg Tauber Kanton Basei-Landschaft Archaologie und Kantonsmuseum Regierunsgebaude CH -441 0 Liestal 308 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 J.-L. Abbe Le parcellaire rural des bastides du sud-ouest de la France: l'apport des sources ecrites et planimetriques Depuis le milieu du XIXe siecle, les bastides ont attire 1'attention des chercheurs. La regularite geometrique des plans de village ou de ville, les privileges accordes par les chartes de franchises au moment des fondations (aux XIIIe et XIVe siecles), ont focalise 1'attention des historiens, des geographes, des architectes et des urbanistes. A tel point que ces villeneuves ont ete considerees comme le symbole de l'urbanisme et des libertes au Moyen Age 1• Les travaux de C. Higounet ont perrnis de rectifier certaines exagerations et constituent le fondement du sa voir actuel sur ce phenomene exceptionnel par son ampleur 2 • Pourtant, cela vient d'etre dit, les etudes ont surtout porte sur deux aspects: les conditions historiques des fondations et la morphogenese des agglomerations. A !'inverse, le parcellaire agraire des nouveaux villages, creations rurales par excellence, a ete pendant longtemps le grand absent de I 'interet des chercheurs. Il a fallu attend re 1' essor recent des etudes de parcellaires conduites par les antiquisants pour que I 'attention se porte sur le paysage que peuvent engendrer les villeneuves medie- vales. Les bastides offrent un terrain d'investigation privilegie: les chartes de fondation et de franchises datent les creations, les plans cadastraux et les photographies aeriennes restituent I' empreinte d' amenagements ruraux geometriques. Quelques etudes monographiques menees depuis le debut des annees 1990 et d'autres avenir laissent entrevoir la richesse de ce chantier apeine ouvert3 • Il reste que la genese des parcellaires agraires des bastides n 'est pas si aisee que cela a men er. Certes, les textes sont relativement explicites sur I' organisation du futur village. La dimension des parcelles a batir figure souvent, avec des mesures de longueur et de largeur precises. La confrontation avec les plans cadastraux perrnet alors d'identifier les llots lies a 1' operation, et meme de calculer la valeur des unites de mesure utilisees 4 • Les textes sont par contre moins precis au sujet des parcelles amettre en valeur. Leur mention, leurs mesures, sont tres lacunaires et les modalites du parcellement ne figurent pas non plus dans les chartes. Pourtant des systemes parcellaires reguliers, prenant la forme d'un quadrillage tres F. PUJOL, L'eiaboration de !'image symbolique de la bastide, Annales du Midi I 03, 1991, 345-367. Les principaux articles de C. HIGOUNET sont rassembles clans Paysages et villages neufs du Moyen Age, Federation historique du Sud-Ouest, Bordeaux, I975 et Villes, societes et economies medievales, Federation historique du Sud-Ouest, Bordeaux, I992. ll n' est pas possible de fa ire etat ici de toute la litterature relative aux bastides depuis M. A. CURIE-SEIMBRES, offre un cadre de travail primordial pour I'etude des parcellaires medievaux. Le colloque d'Orleans de mars 1996 sur "I' Archeologie des parcellaires" et les deux volumes sur Les formes du paysage, G. CHOUQUER (dir.), t. I, Etudes sur les parcellaires, et t. 2, Archeologie des parcellaires (Actes du colloque d 'Orleans), collection Archeologie Aujourd'hui, Errance/ARCHEA, Paris, 1996, traduisent les progres de cette recherche et font le point sur les connaissances actuelles. Voir en particulier le biian des travaux sur le Moyen Age: J.-L. ABBE, Perrnanences et mutations des parcellaires medievaux, t. 2, 223-233. Deux monographies sur les parcellaires agraires des bastides sont a re lever: J.-L. ABBE, L 'amenagement de l'espace: le parcellaire rural de la bastide de Saint-Denis (Aude), clans: Travaux reunis par E. MORNET, Campagnes medievales: I 'homme et son espace. Etudes ojfertes aRobert Fossier, Publications de la Sorbonne, Paris, 1995, I 03-119 et C. LA VIGNE, Recherches sur les systemes parcellaires de fondation en Gascogne au Moyen Age, clans: G. CHOUQUER (dir.), op. cit. t. I, I82-I98 (sur les bastides de Barcelonne-du-Gers, de Mielan, clans le Gers, et de Grenade-sur-Garonne, en Haute-Garonne). C. Lavigne prepare une these de Doctorat sur Fondations d'habitats et restruc- Essai sur les villes fondees dans le Sud-Ouest sous le nom generique de bastides, Toulouse, 1880. Parmi les travaux d'ensemble recents, il faut relever: J. HUGUENEY & P. LAVEDAN, L 'urbanisme au Moyen Age, Droz, Geneve, I974; G. BERNARD, Les bastides du sud-ouest de la France. Morphologie et fonction. Etude de geographie historique, these de doctorat de Illeme cycle, Universite de Paris IV, I983; F. DIYORNE, B. GENDRE, 8. LAVERGNE & P. PANERAI, Essai sur la regularite. Les bastides d 'Aquitaine, du Bas-Languedoc et du Beam, Archives d 'Architecture Modeme, Bruxelles, I985; A. LAURENT, R. MALEBRANCHE & G. SERAPHIN, Bastides, vi lies nouvel/es de Moyen Age, Toulouse, 1988. 3 Le projet collectif interregional "Cartographie des anciens parcellaires de la France" (UMR 9966 "Archeologie et territoires", CNRS!Universite de Tours), coordonne par G. Chouquer, turations des paysages au Moyen Age dans le bassin de la Garonne (dir. J.-8. Marquette, Universite de Bordeaux Ill). 309 J.-L. Abbe homogene, ont ete observes a plusieurs reprises, soit regroupes a la peripherie immediate du nouveau aux environs immediats de la bastide (a Saint-Denis, village. C'est pourquoije delaisserai les textes ne faipar exemple) ou sur la majorite de son territoire (a sant reference qu 'aces deux categories de parcelles Barcelonne-du-Gers, Mielan et Grenade-sur-Garonne, urbaines ou peri-urbaines pour ne retenir que les voir n. 3). Les photographies aeriennes revelent documents mentionnant de maniere explicite des d 'autres cas evidents et ont deja ete interpretees clans parcelles rurales avec des donnees chiffrees. L'objecce sens 5 • I1 reste que photographies et cadastres sont . tif est d' observer et d' interpreter le contenu des chardes documents tres posterieurs aux operations d'ametes a propos du parcellaire rural. nagement et qu, il fa ut se mefier des interpretions Le tableau elabore a partir de ces criteres (fig. 1) hatives, surtout si elles ne peuvent etre soutenues par ne se veut pas exhaustif. Tous les textes conserves des sources ecrites. n'ont pas ete vus. Neanmoins, les travaux de C. Aussi, il semble utile de reflechir sur les methodes Higounet et le repertoire contenu dans I' ouvrage Basqui peuvent etre mises en oeuvre pour mieux detecter tides, villes nouvelles du Moyen Age permettent de et caracteriser les paysages issus des bastides meridresser une liste significative completee par les dionales et, au-dela, des diverses formes d'habitats editions de textes ou la consultation des documents fondes ou developpes pendant la periode de croiseux-memes 7 • Le petit nombre de chartes recensees, sance economique des XIe-debut du XN e siecles. En dix -sept, retient d' abord I' attention. La mention de particulier, quels peuvent etre les apports respectifs superficie en vignes, terres labourables, pres ou bois des sources ecrites et des sources planimetriques qui se justifie par les avantages offerts, a sa voir de faibles viennent d 'etre evoquees ? Que fa ire pour les rend re redevances ( cens ou agrier pour les cereales). A complementaires ? La reponse a ces interrogations l'instar des modalites equivalentes sur les lots a batir necessite des etudes monographiques nombreuses, ce et sur les jardins, I' objectif est de dynamiser le peuqui n' est pas encore le cas. I1 ne sera done question plement et l'activite agricole en rendant les conici que de proposer des pistes de reflexion a partir de ditions de residence et d'exploitation attrayantes. quelques exemples. Comment interpreter alors le peu de textes rencontres? La rarete des mentions de terres a mettre en valeur la is se entendre, c 'est en tout cas une hypo these Les chartes seigneuriales et la restructuration des envisageable, que ce n 'est pas une priorite pour les terroirs seigneurs. L' essentiel est de favoriser I' installation de nouveaux venus au village neufpar un faible cens Les chartes seigneuriales, cela vient d'etre dit, sur la maison et sou vent aussi le jardin. En accordant sont plutot avares de renseignements sur les lots de des "remises" sur les redevances agraires, le seigneur terres alloues aux tenanciers clans le cadre de ces risque d'oberer exagerement ses revenus fanciers. I! fondations. I1 arrive assez souvent qu'en plus de la s'en dispense done la plupart du temps. parcelle a construire - local ou ayral - i1 soit fait A I 'inverse, il est possible de comprendre la premention du jardin- art, casal ou casalere. Mais au sence des biens ruraux comme une volonte accrue de lieu de donner les mesures de longueur prevues pour reussir !'operation et de lui donner une plus forte ces demiers, le texte indique de preference la superdimension agraire. Le cas des fondations de Sicard ficie, ce qui rend I' identification de ces parcelles sur Alaman est tout a fait revelateur. Homme de conun plan un peu peu moins simple que pour les fiance des comtes de Toulouse Raymond VII et habitations 6 . Les jardins sont cependant toujours Alphonse de Poitiers, il beneficie de nombreuses situes clans l'aire de !'agglomeration, soit attenant a infeodations qu' il complete par des achats fanciers I 'habitation, so it - le plus souvent, semble-t-il en Albigeois et en Toulousain 8 • Sur les cinq fonda- Sur !'etude mathematique de la metrologie des unites de longueur apartir d 'un groupe de bastides: H. LEBLOND, Recherches metrologiques sur les plans de bastides medievales, Histoire et Mesure, 1987, 11-3/4, 55-87. 5 Des les annees 1960, J. SoYER analysait plusieurs cliches verticaux de I'IGN et attirait !'attention sur le parcellaire des bastides dans: La vie urbaine, nouvelle serie, 2, avril-juin 1960, 81-88; Photo-interpretation, no 2, 1962, fascicule I 0 et no 4, 1966, 47-50. G. Bemard (voir n. 2) presente de nombreux schemas d 'interpretation de cliches, mais sans veritablement les analyser (entre les p. 87 et 88). 6 Quelques exceptions a noter, cependant, comme a Gargas 310 ou le jardin do it mesurer 8 perches sur 24 (voir fig. I). 7 References des ouvrages n. 2. Les auteurs de Bastides, villes nouvelles du Moyen Age font un repertoire (279-305) de toutes les agglomerations identifiees par eux comme bastides, soit 600. Pour chaque notice, les chartes de fondation et de coutumes sont indiquees lorsqu' elles sont connues (date, seigneur(s) a I' origine de I' acte ). Les informations originales, comme les mentions de lots attribues, sont parfois citees. Les sources utilisees pour chaque document sont repertoriees ala fin de !'article. Voir a son sujet: C. HIGOUNET, Les Alaman, seigneurs bastidors et peagers du XIIIe siecle, op. cit. n. 2, 1975, 305-323. Le parcellaire rural des bastides du sud-ouest de la France: l'apport des sources ecrites et planimetriques communedepartement date seigneurs maison jardin vigne terre pre bois AIGNES 31 1242 Hospitaliers BARRAN 32 1270 1 1 1 114 arpent 4*12 perches 1/4 arpent BEAUMONT-DELOMAGNE 82 BOULOC 31 BUZET-SUR-TARN 31 CASTELNAU-DELEVIS 81 0ARGAS 32 GIMONT 32 1278 Roi-Cisterciens 1 I 0 0 0 1 arpent 32*64 perches 1 arpent 32*64 perches 0 0 0 0 0 1295 1 seteree 1aboureurs 1/2 seterree brassiers 0 seteree femmes 1 arpent 32*64perches I arpent 32*64 perches 1 arpent 0 0 0 0 8ASSOUES 32 Cte Armagnac - Eveque Auch Eveque Auch 1 1 1 4*12 perches 1242 1241 1256 S. Alaman Cte Toulouse S. Alaman 1 4*6 brasses 4*6 cannes I pugnen!e 1 pugneree 112 carteree I eminee I carteree I eminee I cartonnee I seteree 1240 12661274 1291 Seigneur Cte Toulouse Cisterciens Roi-Cisterciens I 8*24 perches 1 arpent 5*15 brasses 1265 1242 1322 1303 Roi Angleterre S. Alaman Seigneur Templiers PUYBEGON 81 SAINTE-GEMME 33 1246 1275 SAINT-PAPOUL 11 1255 S. A1aman Cte Armagnac2 coseigneurs Benedictins GRENADE-SURGARONNE 31 MONSEGUR 33 MONTASTRUC 31 0RNEZAN 32 PLAGNE 31 0 0 0 0 0 0 0 0 0 0 0 1 arpent 1 arpent 0 0 1 arpent 1 arpent 0 0 1/4 arpent 0 1 arpent 0 0 24*72 pieds 1 12*35 coudees 5*12 periones 1 esturon 0 1 concade 1 eminee 1 concade 0 0 0 0 0 0 0 0 1/2arpent 4*6 brasses 4*12 stades 1 carteree 2/5 arpent 1 paire boeuf/jour 2 cartonnees 1 seteree 1 arpent 32*64 periones 1seteree 1 arpent 0 0 l l eminee 1 seten!e 1 arpent 32*64 perches 0 I 1 1 eminee 1arpent 32*64 perches l seteree 1 modiatee 0 a Fig. 1.- Tableau des mesures relatives aux parcelles rurales concedees par chartes defondation ou de franchises des localites du sud-ouest de la France aux XII!e et X!Ve siecles. Legende: Seigneur: seigneur de rang secondaire; 0: absence de mention de cette categorie de parcelle; 1: la parcelle est mentionmie, mais sans mesure. References des chartes en fin d 'article. tions rurales qu' i1 realise entre 1242 et 125 6, quatre (Bouloc, Castelnau-de-Levis, Montastruc, Puybegon) precisent les superficies des terres privilegiees. A Puybegon, elles concement les jardins, les vignes, les terres labourables et les pres: c 'est 1' exemple le plus complet, avec Sainte-Gemme, des dix-sept chartes retenues. I! s'agit d'une veritable politique de peuplement et de mise en valeur des terroirs. M ais c 'est le seul cas ou elle se manifeste aussi bien. Pourtant il ne faut pas negliger les trois fondations en pareage des cisterciens avec les Capetiens. Beaumont-deLomagne, Gimont et Grenade-sur-Garonne appartiennent au groupe des plus eminentes bastides cisterciennes. Elles relevent d'une autre generation que celles de Si card Alaman puisqu' elles datent du dernier tiers du XIIIe siecle. Et les avantages octroyees sont plus limites: les vignes seulement aBeaumont et a Gimont, les terres, sans precision, a Grenade. I! reste a savoir si ces dispositions incitatives debouchent sur un amenagement du paysage agraire, si elles en sont le revelateur. Les textes ne font pas etat de mesures de longueur pour les parcelles rurales, a quelques exceptions pres. A Barran, Bassoues et Sainte-Gemme, les arpents de vigne et de pre font 32 perches sur 64 et a Plagne I' arpent de terre labourable, 32 periones sur 64 9 • Ces dimensions peuvent correspondre a un format de parcelle trapue (la longueur double de la largeur) effectivement repandu dans les regions meridionales. Pourtant a Plagne, C. Higounet n 'a pu identifier des arpents de terre reguliers qui correspondraient a ceux des franchises alors qu' i1 a restitue les casaleres et les parcelles habitees d' origine 10 • I1 convient plutot d' envisager que ces mesures de longueur servent a rappeler la valeur de I' arpent pour determiner I' assiette de la redevance. Le plus sou vent, d 'ailleurs, les parcelles de vigne, de pre et de terre sont indiquees par des unites de mesures agraires de superficie (un arpent, une seteree, une eminee ... ) 11 • La surface importe plus que la forme en la matiere. Sur cette mesure "conventionnelle", voir C. HIGOUNET, Une bastide de colonisation des Templiers clans les prePyrenees: Plagne, op. cit. n. 2, 1975, 295. 10 Ibid., 300-303. 11 L'exemple des jardins est significatif. A Sainte-Gemme, l'arpent servant faire un casal doit 4 deniers, tout comme celui pour le pre ou la vigne do it l 0 deniers. M a is les droits de retroacapte sont de 5 deniers par arpent de casal et de 2 deniers a 311 J.-L. Abbe N 1 0 100 200 m Azas Fig. 2. - Le parcellaire de Buzet-sur- Tarn (Haute-Garonne) d 'apres le plan cadastral de 1836. Dessin J. -L. Abbl Le relief n 'est pas represente car il est tres plat cet endroit de la rive gauche de la vallee du Tarn, part quelques faibles escarpements en bordure de la riviere. L 'altitude varie entre 109 et 112 m. a Il faut done confronter les chartes aux plans cadastraux et aux photographies aeriennes pour evaluer 1'impact des fondations sur le paysage rural et le lien avec les franchises. Pour les trois bastides cisterciennes, cet amenagement est tout fait sensible: des zones quadrillees entourent les nouveaux villages sur a 312 a une grande partie du territoire communal actuel. Ainsi, a Grenade, deux trames agraires paraissent liees a la fondation 12 • I! est par consequent possible d'affirmer que ces entreprises mettent en oeuvre une gestion rationalisee par 1'uniformisation des redevances et un nouveau parcellaire regulier. Il reste nean- Le parcellaire rural des bastides du sud-ouest de la France: l'apport des sources ecrites et planimetriques anc. chAteau / '' '',, '~ '1--.s ,, '',, ' '' ' ' '-. I '-. I '- ' ' '" ' ' ' '- ' I -< I ~ ', ',;''', ) I ", I I I I I I I ' I ', ,j I I I N 0 l 100 I 1 I I I 1 1'- I I I I '-..._ I 200 m Fig. 3.- Schema d 'interpretation du parcellaire de Buzet-sur- Tarn (Haute-Garonne) d 'apres le plan cadastral de 1836. Dessin J.-L. Abbe. Legende: 1. !lots urbains. 2. Parcellaire N- 23°£. 3. Parcellaire N- l9°E. 4. Trace suppose de la premiere enceinte villageoise. par casal. Ainsi le casal fait en principe 2/5 d'arpent et non I arpent comme le cens pourrait le Iaisser croire. L'arpent sert done d'etalon pour calculer les redevances et non pas d'indicateur de la superficie reelle. 12 Sur ces trois bastides, voir Ies schemas clans G. BERNARD, op. cit. n. 2. Pour Grenade, etudes des parcellaires clans: C. LAVIGNE, art. cil. n. 3, I 92- I95 avec la bibliographie des etudes relatives au fin age de cette bastide. A Beaumont-de-Lomagne, des operations de parcellement ont lieu en 1280 et I282 (0. DE SAINT-BLANCAT, Lafondation des bastides royales dans la semichaussee de Toulouse aux XII!e et X!Ve siecles, these de I'Ecole des Chartes, I 94 I, CRDP, Toulouse, 1985, 62). 313 J.-L. Abbe TF. du SIGNAL ACCUMULATION 1 0.96 0.00 O.?e 2 0.60 3 o.se 0.40 0.3e 0.20 0.10 20 4(:) 6(:) 80 1(:)(:) 12(:) 140 160 180 2(:)(:) Fig. 4.- Saint-Denis (Aude). Traitement numerique d'un agrandissement du clicle IGN 1976 FR 2810/200, no/869. Transformation de Fourier du signal d 'accumulation des !ignes verticales (d 'orientation ouest-est). D. Charraut, laboratoire d'Optique P.-M. Duffieux de l'Universite de Besanr;on. Legende: 1, 2 et 3: pies remarquables. Abscisses: frequences spatiales, ordonnees: importance des pies. moins a determiner si les arpents de terre evoques dans les coutumes se retrouvent comme unites de parcellement, a l'instar des jardins et des ayrals. Pourtant, cette adequation entre la charte et le paysage rural de la bastide n'est pas une constante. Si les fondations de Sicard Alaman sont prises en consideration, le resultat est tres different. Les plans cadastraux du XIXe siecle montrent bien la regularite des !lots des villages, mais rien dans le parcellaire rural proche ne correspond a une restructuration organisee et geometrique 13 • C'est le cas aussi pour Gargas 14 et apropos de Monsegur, C. Higounet arrive aux memes conclusions. Neanmoins, le parcellaire proche de cette bastide met en valeur des orientations et un maillage qui pourraient etre le fruit d 'operations concertees 15 • Par consequent, i1 semble difficile d'etablir une correspondance etroite entre les mesures de longueur et de superficie consignees dans les chartes et 1' organisation du parcellaire rural, a la difference de 1, espace urbanise ou maisons et jardins sont disposees selon les mesures figurant dans les textes. Autrement dit, les franchises sur les terres revelent la preoccupation des seigneurs d'attirer ceux qui !es mettent en valeur, mais n' engendrent pas necessairement un nouveau parcellement. Les situations sont en fait tres diverses. Ainsi a Aignes, les Hospitaliers ne donnent des mesures et des redevances que pour les vignes, en distinguant une seteree pour les laboureurs, une demi-seteree pour les brassiers, alors que les femmes, faisant 1'objet d 'un article specifique, n'ont qu'une maison et un jardin. Cette hierarchie, 1l Plans parcellaires: Bouloc: A. D. Haute-Garonne, 3P2 I 66, Castelnau-de-Levis: A.D. Tarn, 3P 432 I 5, Montastruc: A.D. HauteGaronne, 3P2 I 326, Puybegon: A.D. Tarn, 3P I 2438. 4 ' A.D. Haute-Garonne, 3P2 I 189. 1s Pour C. Higounet, "La concession achaque habitant de Monsegur en 1267 (sic, 1265 en fait) d'autant de terrain qu'une paire de boeufs pouvait en labourer en un jour n 'a laisse aucune trace clans le plan cadastral" (Bastides et frontieres, op. cit. n. 2, 1975, 249). Schema clans G. BERNARD, op. cit. n. 2. 314 Le parcellaire rural des bastides du sud-ouest de la France: l'apport des sources ecrites et planimetriques AFFICHAGE DES PERIODICITES DETECTEES Critere (0-1 ): .40 512 pixels en x = 1445 m FnSquence PIC Periode (m) 27 26 (1.00) ( .90) 27 27 54 56 2 ( .79) 2 723 99 ( .73) 99 15 7 94 5 3 ( .70) ( .62) ( .61) ( .58) 7 94 5 2 206 15 289 482 51 (.57) 51 28 11 131 11 ( .56) 25 (.51) 27 58 41 97 58 13 (.50) ( .49) ( .49) ( .47) 41 97 58 13 35 15 25 111 52 98 ( .46) ( .46) 51 99 28 15 30 144 49 34 92 110 76 160 121 130 12 ( .45) ( .44) ( .43) ( .43) ( .43) ( .42) ( .42) ( .42) ( .41) ( .40) ( .40) 30 144 49 34 92 110 76 160 121 130 11 48 10 29 43 16 13 19 9 12 11 120 Fig. 5. Saint-Denis (Aude). Affichage des periodicites detectees a partir de la transformation de Fourier precedente (jig. 4). qui n 'est pas sans rappeler la dichotomie entre les manses libres et les manses serviles du haut Moyen Age, est de nature sociale et favorise les paysans aises. Mais il ne s'agit pas de reorganiser le vignoble local. L' exemple de Buzet-sur-Tarn Il m'a paru interessant de s'arreter sur le cas de Buzet-sur-Tarn (Haute-Garonne, canton de Montastruc, fig. 2 et 3) pour illustrer les apports des sources ecrites et planimetriques et leurs eventuelles correlations en matiere de parcellaire agraire. Le comte de Toulouse Raymond VII accorde le 15 aout 1241 une charte de coutumes aux habitants presents et a venir de Buzet 16 . I1 vient d'acheter cette seigneurie a divers seigneurs entre 1237 et 1239 et en a fait delimiter le territoire 17 . L 'agglomeration existe deja puisque le comte s 'adresse a ceux qui habitent in castro nostro de Buzeto. Le castrum designe a la fois une fortification - le senechal de Carcassonne s 'y rend en 1271 18 - et 1'habitat villageois. En effet, la charte de 1241 distingue, parmi les habitants, ceux habentes domum vel locale infra clausuram sive parietes dicti castri de ceux qui possedent domum vel locale ... extra castrum, in villa seu in boria. Un rempart entoure done le "chateau" et des maisons. Le premier, aujourd'hui disparu, etait situe sur un promontoire rocheux au confluent du Tarn et du ruisseau de Marignol. Le plan cadastral du XIXe siecle (fig. 2) montre la plusieurs grandes parcelles formant presque un anneau rappelant !'emplacement des fosses 19 . Une rue plus large que les autres, en arc de cercle a 1'est de 1'eglise (XIIIe siecle), pourrait correspondre au premier rempart villageois mentionne clans la charte, mais cela reste a verifier. Les maisons situees "hors du castrum " se trouvent in villa, qui serait la partie ouverte du village et in boria, dans un domaine ruraF0 • Buzet, a une trentaine de kilometres au nord-est de Toulouse, est bien situe sur le Tarn, voie navigeable. Le chateau controle un pont qui franchit la riviere, permettant d'aller de Toulousain en Albigeois21. Son role important pour les communications et le commerce regionaux permet de comprendre l'interet du comte qui rejoint celui de Sicard Alaman dont les fondations sont toutes proches 22 • C'est un 16 La charte est connue par la copie faite au XVe siecle d'une copie d'un vidimus de Louis XI date de mars 1462. 17 0. DE SA!NT-BLANQUAT, op. cit. n. 12, 35 et Y. OOSSAT, Saisimentum comitatus tholosani, Collection de documents inedits sur I'histoire de France, serie in-8°, vol. I, CNRS, Paris, 1966, p. 189, n. I. 18 Y. OOSSAT, op. cit. n. 17, 187-193. 19 Plan parcellaire de 1836, A.D. Haute-Garonne, 3P2 I 79. 2 ° Cette interpretation du vocabulaire lie a I'habitat et au terroir est proposee avec prudence. 0. de Saint-Bianquat considere que villa et boria sont synonymes dans le texte. Le deuxieme terme sera it une latinisation de I'occitan barri design ant le faubourg (op. cit. n.l2, p. 132 et 137, n. 1). 11 est difficile de trancher. 21 Le pont medieval de Buzet a ete emporte par les crues au XVIlle siecle. Le plan cadastral represente les deux piles subsistantes au X!Xe siecle. Son emplacement, aux abords immediats du chateau, est marque sur les fig. 2 et 3 par des tirets. 22 Sur les objectifs economiques, comrnerciaux et agricoles de ces fondations, voir: C. Higounet, art. cit. n. 8, et L'occupation du sol du pays entre Tarn et Garonne au Moyen Age, op. cit. n. 2, 1975, 129-150. 315 J.-L. Abbe chef-lieu de bailie dont dependent cinq bastides et six castra en 1271. Le succes de la charte de franchises se lit sur le plan du village. L'agglomeration comprend une serie de rues et d 'ilots dont le trace n 'est pas d'une geometrie parfaite - le raccord avec le castrum en est probablement la raison - mais 1'homogeneite de !'ensemble est significative. La charte de franchises prevoyait d'ailleurs des localia pour les maisons de quatre brasses sur six. L' amenagement urbain est-il developpe par celui de 1'espace rural ? Raymond VII concede a chaque habitant, present'ou futur, unjardin d'une superficie d'une pugneree avec un cens d'un denier toulousain, une carteree pour faire une vigne avec un cens de trois deniers et une seteree de terre a labourer sans aucun cens (fig. 1). Chacun peut defricher, s'ille veut, en acquittant un agrier correspondant au neuvieme de la recolte. Ces dispositions visent bien sur a accorder des avantages pour les redevances. Debouchent-elle sur un nouveau paysage agraire ? La fig. 3 propose une interpretation du parcellaire autour de Buzet23 • Deux ensembles ont retenu !'attention. Aux abords meridionaux immediats du village, plusieurs limites de parcelles, parfois des tronc;:ons de chemins, esquissent une trame lanieree reguliere. Elle est reperable jusqu' a environ 700 metres du village et structure nettement 1'espace entre le ruisseau de Marignol et le grand chemin allant a Lavaur et a Castres. Les secteurs de voies v1 et v2, perpendiculaires entre eux, semblent lies a une reorganisation du reseau viaire que trahit le trace en baYonnette du chemin de Bels et d' Azas ainsi que !'axe v3, limite parcellaire pouvant reveler une ancienne voie. Le developpement du castrum du Buzet a engendre, c' est ain si qu 'il est possible de 1'interpreter, la transformation partielle de la trame via ire et du parcellaire rural. Cependant, la superficie concernee reste assez reduite. L'examen du parcellaire a aussi mis en valeur un deuxieme ensemble original. Situe plus a l'ecart de Buzet, il s'etend par contre sur un espace relativement important. Caracterise par un fort lanierage, sa densite est remarquable clans les quartiers de la Mouline, du Gres, jusqu'au moulin de Laberano, sur environ 1, 7 km. Il se prolonge, semble-t-il, au-dela du ruisseau de Marignol, et surtout quelques grands axes du parcellaire lui sont isoclines a I' est (v6 - 1 km de long- et v7), bien que relativement eloignes. Comment comprendre cet ensemble parcellaire? Son etendu est telle qu'il faudrait l'etudier sur une plus grande superficie, afin de juger s'il tient compte ou non des limites communales, et done s'il correspond a une mise en oeuvre locale, clans le cadre d'un finage, ou a une organisation a plus grande echelle. I1 faut cependant observer qu'aucun chemin du cadastre ne s 'integre ases orientations, sauf le tronc;:on v5. Par contre, la limite v3 deja evoquee joue le role d'axe sur lequel butent les autres parcelles. Si c'est effectivement un ancienne voie, cela attesterait la relative anciennete de la formation de ce parcellaire, son anteriorite a celui identifie en premier. La meme observation peut etre faite apropos de la limite v4. I1 est difficile de proposer une chronologie absolue des formes parcellaires identifiees. La trame proche de Buzet parte le temoignage d'un amenagement modeste mais ordonne qui doit etre consecutif a l'essor de !'habitat au XIIIe siecle. Cependant !'orientation (23° Est) ne correspond a rien clans le village: il ne s' agit pas d 'une operation integrant les parcellaires urbain et agraire. La deuxieme trame, dont I' empreinte est plus vive, correspond a une autre phase: !'orientation differe (19° Est) avec une predominance des limites est-ouest, contrairement au premier parcellaire. Elle pourrait correspondre a une phase ulterieure de 1'amenagement medieval, mais son extension et son agencement interne (role de v3 et v4) font plutot envisager son anteriorite. L 'exemple de Buzet-sur-Tarn con vie a prendre en compte la complexite des amenagements agraires medievaux lies aux villages neufs. L' octroi en 1241 de coutumes semble trouver un prolongement sur le sol autour de village. Mais c'est une operation visiblement modeste, qui ne bouleverse pas le paysage, a la difference des bastides cisterciennes apparaissant clans le tableau. Au lieu de faire table rase de ce qui existe, le parcellaire anterieur a ete globalement maintenu. Sur ce point, Buzet est en definitive amichemin des fondations de Sicard Alaman et de celles de C1teaux, pour rester clans le cadre comparatif dresse clans cette etude. Les operations du comte de Toulouse et de son conseiller ont en commun des mesures de superficie identiques ou proches (pour les maisons, les jardins, les terres labourables) et !'impact des dispositions dans 1'amenagement transparait surtout dans le lotissement villageois. Il faut attendre les annees 1260 et plus encore 1270 pour que des operation d' envergure touchent des fractions importantes de finages lies a des fondations: la regularisation des champs proches de Buzet n'en serait qu'une ebauche. Les etudes deja menees sur 23 Le schema a ete realise a !'aide d'une grille orthonormee et en reportant les limites parcellaires qui s'integraient aux orientations selectionnees. Le territoire a l'ouest du ruisseau de Marignol n'est pas represente: le parcellaire n'adopte pas du tout les memes dispositions qu'a !'est et ne paralt pas lie morphologiquement aux transformations du XII!e siecle. 316 Le parcellaire rural des bastides du sud-ouest de la France: I' apport des sources ecrites et planimetriques Grenade-sur-Garonne, Mielan, Barcelonne-du-Gers et Saint-Denis prouvent nettement la radicalisation des changements ulterieurs" 4 • Traitement statistique et numerique des sources Jusqu'a maintenant, i1 a ete question de l'apport des chartes de fondation et de coutumes a la connaissance du paysage rural des bastides et a a leur mise en correlation, delicate parfois, cela vient d'etre vu, avec les plans cadastraux. Or, dans de nombreux cas, les chartes ont disparu et les conditions d'elaboration des paysages neufs ne peuvent etre apprehendees que par d 'autres sources. Ainsi 1'etude de la bastide audoise de Saint-Denis m'a amene a rechercher la genese de son parcellement a partir d 'un censier realise en 1490 et des documents planimetriques. Les chartes d 'origine de cette fondation royale des annees 1290-1293 ne sont pas connues. Or, a la suite d'un traitement statistique, le censier a revele que les 2/3 (31/50) des jardins et des vergers avaient une superficie d'une carteree ou de deux pugnerees (la moitie d'une carteree). Or, la carteree, ou quart d' arpent, est une mesure frequente des jardins (voir fig. 1) dans les chartes. I! a a us si indique que la superficie d 'un journal correspond a celles de treize grandes parcelles carrees. Chacune est limitee par quatre chemins et occupee par des pres. Ces parcelles constituent en fait le module de base de la bastide, celui aussi des llots batis et des jardins. Le journal, 36 ares, est I' equivalent de la seteree ou de l'arpent local. Ceux-ci sont effectivement des superficies recurrentes des parcelles rurales dans les chartes. Verifies sur le plan cadastral, les resultats ainsi obtenus par le traitement des donnees du censier permettent de proposer des hypotheses fondees sur le parcellement de Saint-Denis (activites et superficies, unites de mesure). Ils vont aussi dans le sens d'une prise en compte des mesures inscrites dans les chartes pour I' amenagement des terroirs puisqu' elles se retrouvent mises en pratique. Par contre il n'a pas ete possible d'etablir de telles correlations pour les vignes et les terres labourables. Le traitement numerique des photographies aeriennes offre depuis peu de nouvelles possibilites d'ana- 24 25 C. LAVIGNE et J.-L. ABBE, arts. cits. n. 3. D. Charraut, F. Favory, C. Raynaud, avec le concours de P. JOURDAIN, M.-J. OURIACHI et H. PERRIN, Paysages rythmes: recherches sur l'empreinte des mesures antiques dans le parcellaire agraire languedocien, Mappemonde, 311992, 28-33; D. CHARRAUT, G. CHOUQUER & F. FAVORY, Traitementnumerique de l' image, Archeologia, n° 307, decembre 1994, 24-32, et C. LAYIGNE, art. cit. n. 3, 183-184. Traitement de !'image realise a lyse des paysages. La recherche des orientations et des periodicites metrologiques contenues dans un cliche de I 'IGN perm et aujourd 'hui de detecter les traits dominants d 'une organisation parcellaire. Les principes du traitement numerique ayant ete exposes dans d'autres publications, je n'y reviendrai pas 25 • I! s 'agit de de gager les orientations majeures du parcellaire en eliminant progressivement celles qui sont secondaires: c'est le "filtrage" de !'image numerique obtenue a partir du cliche. De la, le calcul des periodicites principales est realise par la recherche des ecarts types entre les !ignes selectionnees. Comme dans le cas des cadastres antiques, cette methode convient parfaitement a I 'etude des villeneuves ayant engendre leur propre parcellaire. Lorsque celui-ci est organise selon un quadrillage regulier, le traitement d'image permet de mettre en valeur son etendu et ses rythmes dominants. C. Lavigne pour Barcelonne-duGers et moi-meme pour Saint-Denis avons mis en oeuvre cette methode d'analyse" 6 • A Saint-Denis, le filtrage numerique d'une photographie aerienne de 1'IGN de 197 6 met en valeur les grands axes du nouveau parcellaire organise selon un quadrillage orthonorme rigoureux autour de la bastide27. Les fig. 4 et 5 temoignent des etapes du traitement du ca!cul des periodicites. La transformation de Fourier du signal d'accumulation (fig. 4) represente par une courbe les ecarts types entre les !ignes de pixels selectionnees par le filtrage numerique, autrement dit les intervalles entre les !ignes du parcellaire. Trois "pies" emergent assez nettement (1, 2 et 3). Le tableau d'affichage des periodes detectees (fig. 5) convertit en metres les pixels suivant un ordre decroissant en fonction de la frequence (la hauteur des pies sur le graphique) et degage ainsi les ecarts dominants: le pie 27 (le 1, de loin le plus eleve) correspond a une periode metrique de 54 a 58 m, - le pie 99 ( le 2), a une periode de 15 m, - le pie 51 (le 3), a une periode de 28 m. La relation est etroite entre les trois periodes: la premiere est le double de troisieme et pratiquement le quadruple de la deuxieme. I1 est aise d'identifier la periode de 54-58 m au cote des !lots cam~s de la bastide (voir supra ). Calculee a partir du plan parcellaire, leur superficie moyenne est de 36 ares (le journal du censier de 1490), soit 60 m de cote. Le partir d'un logiciel mis au point au laboratoire d'Optique P.-M. Duffieux de l 'Universite de Besanc;:on (D. Charraut). 26 C. LAYIGNE, art. cit. n. 3. J'ai effectue le traitement d'image sur Saint-Den is au laboratoire d 'Optique de Besanc;:on avec D. Charraut et G. Chouquer le 13 novembre 1992. 27 Le cliche IGN ( 1976 FR 2810/200, n° 1869) est reproduit avec une image du filtrage obtenu dans D. CHARRAUT, G. CHOUQUER & F. FAVORY, art. cit. n. 25, 30. 317 1.-L. Abbe traitement numerique arrive a un resultat tres voisin. Les subdivisions de la periode de 54-58 m representent les divisions intemes de ces !lots, en demi ou en tiers, en particulier pres du village. La presence sur le tableau des periodicites d'un pie de frequence 11, so it 120-131 m, c 'esH't-dire le double du module de 60 m, confirme la validite du traitement. Le paysage de la bastide de Saint-Denis apparait par consequent fortement marque par le parcellement issu de sa fondation a la fin du XIIIe siecle. La bastide de Barcelonne-du-Gers resulte d'un pare age entre le comte Bemard VI d' Armagnac et les Hospitaliers en 1316. Le texte est perdu, de meme que les coutumes, mais la bastide est erigee peu apres 1336. Dans une etude exemp1aire menee a partir de photographies aeriennes, C. Lavigne individualise quatre trames parcellaires quadrillees sur le finage de Barcelonne. Le traitement numerique des cliches confirme une periode majeure, de 150 m, qui correspond a 1'ecart entre les chemins servant de lineaments a ces trames. Surtout il revele six autres periodes sous-multiples de la premiere. A partir de ces donnees, la recherche des parcelles dont les cotes equivalent a ces periodes debouche sur 1'identification de 533 unites. Les superficies de ces parcelles apparaissent reposer sur un module de 56 ares (23% des unites retenues) et sur ses multiples (de 1 a 4). Le nombre et 1'ubiquite de ces modules parcellaires en font evidemment le fondement des divers parcellements agraires de la bastide. Les exemples de traitement numerique des photographies aeriennes viennent de demontrer, me semblet-it, les vastes perspectives proposees a !'etude des paysages ruraux des villeneuves. Cette etude est par ailleurs complexe, comme l'examen des chartes et du plan cadastral de Buzet l'ont fait comprendre. Que faut-il en retenir ? Les actes de fondation et de coutumes contenant des mesures de surface, parfois de longueur, relatives aux terres du finage sont minoritaires. Ces mentions ont une finalite certaine: preciser des redevances dont la nature et le montant doivent etre attractifs. Le choix des mesures est plus equivoque. Il est bien sur destine afa ire sa voir la superficie de la parcelle sur laquelle doit peser le cens. Est-il aussi le signed 'une division des terroirs en lots egaux afin de faciliter 1'assignation et la gestion des terres et de leurs revenus? La reponse est loin d'etre univoque. Dans certains cas, le parcellaire des fondations ne presente pas de regularite averee. La reference a l'arpent, a la seteree, a 1'emine, est alors a comprendre uniquement comme celle d'une surface a mesurer pour determiner les parcelles beneficiant de la redevance fixee par la charte. Mais le parcellaire n'est pas transforme. Cette situation parait prevaloir pendant la premiere moitie du XIIIe siecle. 318 Apres 1260 la mise en valeur des terres s'accompagne souvent d'une reorganisation etendue du parcellaire dont les sources planimetriques, et meme des documents ecrits comme les censiers ou les compoix, gardent 1' empreinte, les unes a travers les formes et les orientations, les autres a travers les mesures de superficie. Les chartes seigneuriales concedees dans ces cas-la peuvent etre revelatrices de ces operations concertees. Mais elles n 'en donnent pas les modalites pratiques: ce n'est pas leur finalite. La difference est nette avec les parcelles a batir, voire les jardins, pour lesquelles les dimensions sont fixees precisement et dont les plans attestent la realite des dispositions. Les textes retenus dans cette etude attirent encore deux remarques. En premier lieu, la mention des vignes est tout aussi systematique que celle des maisons et des jardins. Ce n' est pas le cas pour les terres labourables, les pres et plus encore les bois. Cela demontre la presence generalisee de la culture de la vigne, mais aussi de son interet pour les villageois: proposer une redevance avantageuse parait necessaire. A 1'inverse, la rarete des pres est sou vent reelle dans les surfaces cultives, mais leur grande valeur n'incite pas non plus les seigneurs a etre trop larges en ce domaine. C'est encore plus vrai pour les bois que l'aristocratie veut conserver. Seule Plagne, bastide de colonisation au milieu d'un massif forestier pre-pyreneen, pouvait faire exception. L'autre observation porte sur les villages obtenant les franchises. Ils presentent une grande heterogeneite quant a leur origine. Des bastides appelees ainsi dans les textes, correspondant a une agglomeration tout a fait nouvelle, sont presentes. C' est le cas des fondations cisterciennes, de Plagne, de Barran et de Bassoues, mais aussi de Saint-Denis et de Barcelonne-du-Gers. Souvent, ces fondations tardives (pas avant le demier tiers du XIIIe siecle) debouchent sur de profondes reorganisations du parcellaire rural. D'autres habitats sont designes dans les chartes par le terme de castrom, comme Buzet, Castelnau-de-Levis, Omezan ou Puybegon. L' ampleur du remembrement est plus faible dans ces villages, voire absent. Enfin, il faut re lever la presence du village abbatial de SaintPapoul. Les benedictins concedent des droits identiques aux bastides et sur des superficies plus grandes. Cette diversite semble indiquer que l'interet pour la valorisation des terres ne conceme pas que les hastides stricto sensu mais aus si 1'ensemble des localites favorisees par 1'essor economique et demographique. Les bastides ne constituent que la face la plus spectaculaire, la plus emblematique d'un mouvement de fond qui innerve !'ensemble du Sud-Ouest. Les progres dans I' etude des paysages ruraux lies a une fondation passe par consequent par la conjonction des sources ecrites et planimetriques. Les Le parcellaire rural des bastides du sud-ouest de la France: l'apport des sources ecrites et planimetriques chartes servent ajuger I' interet porte aux ressources des terroirs agraires et par consequent a deceler une eventuelle volonte de modeler ces terroirs. Les unites et les mesures sont aprendre en compte lorsque sont examines les documents planimetriques afin de deteeter les modalites de I' assignation des parcelles. Cela dit, beaucoup de chartes n'evoquent pas les terres, et parfois elles ont disparu. I! fa ut alors utiliser les sources economiques et fiscales ulterieures. De toute fayon, les plans et les photographies representent le plus sou vent I' element essentiel car etant les seuls a restituer la realite, meme a plusieurs siecles d'ecart. La genese des paysages agraires des bastides, le lien avec les autres modes de fondation, le comportement des seigneurs face a !'evolution economique: autant de domaines aapprofondir en donnant toute leur place ades sources complementaires 28 , Sources Sont indiquees ici les sources utilisees pour les chartes de fondation et de franchises des localites du tableau de la fig. 1. Aignes (31): J. RAMIERE DE FORTANIER, Charles de franchises du Lauragais, these de Doctorat, Universite de Toulouse, Toul, 1939, 120-121. Barran (32): J.J. MONLEZUN, Histoire de la Gascogne depuis les temps les plus recules jusqu 'a nos }ours , Auch, 1849, t. VI, 110-116. Bassoues (32): Idem (en fait celle de Barran, les textes sont identiques ). Beaumont-de-Lomagne (82): C. BABINET DE RENCOGNE, Le livre juratoire de Beaumont-deLomagne, Montauban, 1888, 257. Bouloc (31): A. TEULET, Layettes du Tresor des Charles, II, Paris, 1866,474-475. Buzet-sur-Tam (31): 0. DE SAINT-BLANQUAT, La Jondation des bastides royales dans la senechaussee de Toulouse aux XII!e et XIVe siecles, these de l'Ecole des Chartes, 1941, CRDP, Toulouse, 1985, 131-137. Castelnau-de-Levis (81): C. COMPAYRE, Etudes historiques et documents inedits sur I 'Albigeois, le Castrais et !'ancien diocese de Lavaur, Albi, 1841,313-320. Gargas (31): J. LESTRADE, La coutume de Gargas, Revue historique de Toulouse X, n° 1, 1923, 297. Gimont (32): J. J. MONLEZUN, op. cit., 201-205 (charte de fondation) et A. THOMAS, Charte de coutumes de Gimont, Annates du Midi VIII, 1896, 5-14. Grenade-sur-Garonne (31): R. RUMEAU, Monographie de la vi!le de Grenade-sur-Garonne, Toulouse, 1879, 64. Monsegur (33): F. MICHEL, L 'Esclapot ou Cartulaire de Monsegur, Bordeaux, 1963 (les mesures ont ete relevees clans J. HUGUENEY & P. LAVEDAN, L 'urban is me au Moyen Age, Droz, Geneve, 197 4, 73-74). Montastruc (31): C. DE Vrc & J. VAISSETE, Histoire generate de Languedoc, t. VIII, Privat, Toulouse, 1876-1892, col. 1081-1084. Omezan (32): E. DE VACQUIE, Coutumes d'Orbessan et d'Omezan, 1320, 1322, Memoires de la Societe Archeologique du Midi de la France IV, 18401841' 329-342. Plagne (31): A.D. Haute-Garonne, H Malte, Montsaunes, liasse 15, piece 1 et C. HIGOUNET, Une bastide de colonisation des Templiers clans les pre-Pyrenees: Plagne, Paysages et villages neufs du Moyen Age, Federation historique du SudOuest, Bordeaux, 1975, 293-303. Puybegon (81 ): E. CABlE & L. MAZENS, Un cartulaire et divers actes des Alaman, des de Lautrec et des de Levis, XI!Ie et XIVe siecles, Toulouse, 1882, 67. Sainte-Gemme (33): J. J. MONLEZUN, op. cit., 269284. Saint-Papoul (11): J. RAMIERE DE FORTANIER, op. cit., 661-666. (A.D. =Archives Departementales). 28 C. Higounet a parfaitement per9u les enjeux des fondations (sauvetes, castelnaux, bastides): "les grandes I ignes du paysage agraire et de I 'habitat de la region ont ete mises en place par ces grandes entreprises des hommes des trois siecles medievaux. Mais un tel phenomene n 'a pas ete particulier aI' Aquitaine, et nous savons qu'il a ete, a diverses variantes pres, a l'echelle europeenne" (Pour l'histoire de !'occupation du sol et du peuplement de la France du Sud-Ouest, du Xle au X!Ve siecle, op. cif. n. 2, 1975, 397). J.-L. Abbe Universite Paul Valery, Montpellier III 19A, rue du Bois-Rolland 11 100 Narbonne France 319 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 Dirk Meier Der Wandel der Landschaft und Besiedlung von der romischen Kaiserzeit bis in das Mittelalter im Dithmarscher Kiistengebiet (Deutschland) 1 Einleitung Seit priihistorischer Zeit sind in den festliindischen Nordseemarschen die Genese der Landschaft und ihrer Besiedlung eng miteinander verbunden. Im Bereich der schleswig-holsteinischen Nordseeki.iste untersucht die Arbeitsgruppe Kustenarchiiologie am Forschungsund Technologiezentrum Westkuste (FTZ) der Christian-Albrechts-Universitiit Kiel die Besiedlungsgeschichte wie die Lebensbedingungen der Menschen an der Ki.iste unter dem Einflu/3 einer sich stetig wandelnden Umwelt, die damit eng verbundene Landschaftsgeschichte mit der Veriinderung von Ki.istenlinien sowie den Verlaufund die Auswirkungen nacheiszeitlicher Meeresspiegelschwankungen. Nach der Betrachtung des mittelalterlichen Landesausaus und der fiiihen Bedeichung in Eiderstedt (u.a. Meier u.a. 1989; Meier 1992) erfolgten Untersuchungen zur Abb. 1. - Lage des Untersuchungsgebietes an der Nordseekuste SchleswigHolsteins mit Wurtengrabungen (Altuntersuchungen und DFG Projekt Dithmarschen). I. Tofting; 2. Elisenhof; 3. Tiebensee; 4. Haferwisch; 5. Jarrenwisch; 6. Wellinghusen; 7. Hassenbattel; 8. Busum; 9. Nord erbusenwurth; I 0. S ii d e r b u s e n wurth; 11. Fahrstedt. "Landschaftsgeschichte und zu den Siedlungsmustem von der romischen Kaiserzeit bis in das Mittelalter in Norderdithmarschen", die in Jahren 1993 bis 1995 durch die Deutsche Forschungsgemeinschaft finanziert wurden (Meier 1995; Meier u.a. 1997). Die Auswahl der einzelnen Grabungsorte erfolgt auf der Basis einer Kartierung der historischen Kulturlandschaft des Eiderstedter und Dithmarscher Ki.istengebietes mit ihren Warften (Wurten), Deichen, wasserbautechnischen Einrichtungen, archiiologischen Fundstellen und a! ten Meereseinbruchen, in die altere Vorarbeiten einbezogen sind (Abb. 1). Durch ihre interdiszipliniiren Untersuchungen triigt die historische Ki.isten- und Umweltforschung auch zum Verstiindnis der heutigen okologischen Problematik der Nordseeki.iste bei, einem Raum, in dem Landschaftsentwicklung und Siedlungsgeschichte eng miteinander verzahnt sind und sich gegenseitig tiefbeeinflu/3t haben. NoJ:dfT!es!dho ml Geest D Ill Marsch Dune Ne 0 km 321 D. Meier 2 Naturraumliche Verhaltnisse und Siedlungsmuster in der romischen Kaiserzeit Von der Elbe im Siiden bis zur Eider im Norden erstreckt sich an der Nordseekiiste Schleswig-Holsteins die Landschaft Dithmarschen, einer der drei nordalbingischen Sachsengaue, der seine Grenzen vom friihen Mittelalter bis zum heutigen Kreis fast unverandert bewahrt hat (Abb. 1). Das Landesinnere nehmen hOhere, in der Saale-Kaltzeit geformte Geestkeme ein, die mehrere, ehemals vermoorte Taler inselartig zergliedern. Westlich des teilweise kliffartig abfallenden Geestrandes erstrecken sich stabile Seemarschen, die in ein kiistennah hoch aufgelandeten Marschstreifen und ein tieferes Sietland gegliedert sind. Das Kiistenvorfeld bestimmt das Wattenmeer mit mehreren Vorsanden und Gezeitenstromen. Im Verlauf des nacheiszeitlichen Meeresspiegelanstiegs hatte die Nordsee vor etwa 6500 Jahren BP den Dithmarscher Geestrand erreicht (Menke 1988). Dabei erfolgte die Uberflutung der im Kiistenbereich bis NN -20 m abfallenden holozanen Oberflache so rasch, das es nicht wie in anderen Bereichen des Nordseekiistengebietes zur Ausbildung von Torfen kam, die als datierbare Fixpunkte des nacheiszeitlichen Meeresspiegelanstiegs herangezogen werden konnen (Meier u.a. 1997; Behre 1987). Lediglich palynologisch datierte Aufschliisse (Menke 1988, 21) a us den vermoorten Niederungen zwischen den inselartigen Geestkernen vermitteln in Dithmarschen ein Bild des alteren und jiingeren Transgressionsabschnittes. Die Ablagerung der sandigen und schluffigen Sedimente im Kiistenraum ebenso wie die Erosion der vorspringenden Geestkerne durch das Meer fuhrte im Dithmarscher Kiistengebiet vor etwa 3800 Jahren BP zur Entstehung von Nehrungen, auf denen Diinen aufwuchsen (Ab b. 1: 5). Diese schlossen sich in nordsiidlicher Richtung an die Geestkeme an, so daf3 eine Ausgleichskiiste entstand. Da sich ab dem Ubergang zum jiingeren Holozan durch verstarkte Sedimentation die Wassertiefen verringerten bildete sich ein ausgedehntes Wattenmeer mit Gezeitenrinnen und Sandplaten. In einem langen Zeitraum hatte das Meer dann soviel Material herangefuhrt, daf3 ab der Mitte des letzten vorchristlichen Jahrtausends vor der Ausgleichskiiste ein schmaler Marschstreifen aufgelandet war. Urn Chr. Geburt erstreckte sich entlang des Geestrandes die "alte Marsch", deren Landnahme im friihen 1. Jahrhundert n.Chr. einsetzte. Eine grof3flachige Bodenbildung, auf der die kaiserzeitlichen Wohnplatze in der Marsch lagen, istjedoch heute kaum mehr nachweisbar, wie Beobachtungen entlang von Rohrleitungsgraben einer Leitung des Wasserbeschaffungsverbandes durch die Dith322 marscher Nordermarsch ergaben (Meier u.a. 1997). Reste humoser Lagen traten vor allem dort auf, wo sich Priele in die Marschoberflachen eingeschnitten hatten, deren Rinnen spater zuschlickten und deren Fiillung nachsackte. In spaterer Zeit wurde die dicht unter der heutigen Oberflache liegende kaiserzeitliche Marsch nur noch in geringem Maf3e von Sedimenten iiberdeckt. Die intensive landwirtschaftliche Bearbeitung und die dicht unter oder im Bereich der heutigen Oberflache liegenden Boden der romischen Kaiserzeit erschweren die Rekonstruktion des Naturraums der ersten nachchristlichen Jahrhunderte. Doch darfals gesichert gelten, daf3 im siidlichen schleswigholsteinischen Nordseekiistengebiet erstrnals urn Chr. Geburt gute Voraussetzungen fUr die Anlage bauerlicher Siedlungen in den Seemarschen vorhanden waren. Dabei begiinstige eine im Nordseekiistengebiet weit verbreitete Meeresregression eine flachenhafte Landnahme der Seemarsch. Geht man von der Verteilung der archaologisch bekannten Siedlungen (Bantelmann 1949; Bokelmann 1988) aus, erstreckte sich entlang des Dithmarscher Geestrandes ein wohl maximal bis 5000 m breiter Marschstreifen, den die Meldorfer Bucht in die Norder- und Siidermarsch trennte (Abb. 1: 2). In der Siidermarsch lagen einzelne kaiserzeitliche Siedlungen nahe der Kiiste westlich eines vermoorten Sietlandes wie kaiserzeitliche Funde aus Krumwehl und der Dorfwurt Siiderbusenwurth (Bantelmann 1949) andeuten. Weitere kaiserzeitliche Siedlungen bestanden nach Aussage archaologischer Funde auf der Nehrung des Elpersbiitteler Donns (Buchholz 1963 ), am Gee strand bei Hemmingstedt und Heide und in der Randlage zur Marsch bei Eddelack. Auf den bewaldeten, von vermoorten Talern getrennten Geestinseln sind zudem zahlreiche Graberfelder belegt (Hingst 1983). Beide Siedlungsgebiete der romischen Kaiserzeit, die der Seemarsch und der Geestkammem trennte ein tei1weise vermoortes Sietland voneinander. Reste von bislang nicht datierten Torfen traten im Sietland bei Baustellenbeobachtungen unter einer mittelalterlichen Hofwurt in Barlt zutage, einer reihenformigen Marschenhufensiedlung des hohen bis spaten Mittelalters. Etwas anders waren die Verhaltnisse in der Nordermarsch zwischen der Meldorfer Bucht im Siiden und der Eider im Norden (Abb. 2). Archaologische Sammelfunde (Bokelmann 1988) und Grabungen auf den Wurten Wennemannswisch (Bantelmann 1949), Tiebensee und Haferwisch (Meier 1994; Meier u.a. 1997) vermitteln dort das B ild zweier langgestreckter Siedlungsreihen kleiner bis mittelgrof3er, auf Viehhaltung ausgerichteter Gruppensiedlungen mehrerer bauerlicher Betriebe. Moglicherweise sind diese linienformigen Siedlungsmuster vermutlich ein Der Wandel der Landschaft und Besiedlung im Dithmarscher Ktistengebiet Abb. 2. - Paliiographische Karte des Eiderstedter und Dithmarscher Kustengebietes in der romischen Kaiserzeit. EIDERSTEDT -ffi~ L3:> @ tl Dorfwurt/Siedlung Hortfund Graberfeld Marsch der ram. Kalserzelt Moor D Geest Ill! SandWall DITHMARSC 0 Abbild der kaiserzeitlichen Topographie, die sich aber infolge ji.ingerer, allerdigns nur geringmachtiger Sedimentablagerungen und jahrhundertelanger landwirtschaftlicher Bearbeitung heute nicht mehr deutlich erkennen la/3t. Moglicherweise lagen zwar einzelne Wohnplatze auf geringftigig hOher aufsedimentierten Marschri.icken, fUr einen Teil gilt dies jedoch nicht. Zwar deuten bodenkundliche Karten an, da/3 bei der Anlage der Siedlungen teilweise etwas sandigere Boden bevorzugt wurden, doch konnten dies geologische Untersuchungen nicht belegen. Wahrend der gesamten romischen Kaiserzeit blieben die auf Viehhaltung ausgerichteten Siedlungen jedoch von der naturraumlichen Entwicklung abhangig, wie die unterschiedliche lange Nutzungszeit der . Wohnplatze verdeutlicht. So entstanden nach Aussage verdickt facettierter eisenzeitlicher Keramik die Siedlungen der geestnahen Reihe von Tiebensee, bverwisch und Wennemannswisch (Bantelmann 1949) im fri.ihen I. Jahrhundert n.Chr., wahrend in der westlicheren Reihe in Haferwisch mit dem Bau von Wurten nicht vor der Mitte des 2. Jahrhunderts begonnen wurde (Abb. 1; 2). Wie die auf einer im Durchmesser etwa lOO m gro/3en, bis NN +3,00 m hohen Wurt der geestnahen Reihe durchgeftihrten archaologischen Untersuchungen ergaben, bot si eh in Tiebensee ein trockenes zwischen NN +1,00 und NN + 1,30 m hohes Area! zur Anlage einer von Hofstellen zur ebenen Erde an. Nach der Auswertung palaobotanischer Proben aus einem Brunnen des 1./2. Jahrhunderts zu schlie/3en, Jag die Siedlung selbst recht hoch und trocken, da si.i/3wassergepragte Arten i.iberwogen 1. Reste der Krotenbinse (Juncus bufonius ), des Wei/3en Gansefu/3es (Chenepodium album) oder der kleinen Brennessel ( Urtica urensis) deuten auf Ackerbau hin, wenn auch keine Kulturpflanzenreste nachgewiesen werden konnten. Die archaologischen Grabungen legten einen W ohnplatz der vermutlich a us vier bis sechs Wirtschaftsbetrieben bestehenden Flachsiedlung frei. Nach dem Abbruch des auf dem Wohnplatz I einmal emeuerten Gebaudes erfolgte im 2. Jahrhundert eine geringfi.igige Erhohung des Siedlungsareals von NN + 1,50 auf i.iber NN +2,00 m bevor die Siedlung am Die botanischen Untersuchungen fOhrten H. Kroll und S. Medovic, lnstitut fOr Ur- und Friihgeschichte der Universitat Kiel, durch. lhnen sei fOr die Oberlassung erster Ergebnisse herzlich gedankt. 323 D. Meier Obergang zur jilngeren romischen Kaiserzeit verlassen wurde. Die Aufgabe der Siedlung war vermutlich ebenso wie die weiterer eine Folge des Landschaftswandels, da moglicherweise der seit der Mitte des ersten nachchristlichen Jahrtausends nachgewiesenen Vermoorung der geestnahen Marsch (Wiermann 1962) Stauniissebildungen vorausgingen, die zu einer Einengung der kaiserzeitlichen Wirtschaftstliichen filhrten. Weiter nordlich dehnten sich in dieser Zeit bereits Moore aus, wie Radiokarbondaten eines Torfhorizontes unter einer Hofwurt in Hemme andeuten, der ein Alter urn 1790 BP aufwies, was einem Kalenderalter 40-380 AD entspricht (Hoffmann 1986, 255). Die weitere Vermoorung der kilstenfernen Marsch wurde durch die Auflandungjunger Marschtliichen im Westen noch begilnstigt, so dal3 im fri.ihen Mittelalter bereits weite Teile der noch in der romischen Kaiserzeit dicht besiedelten alten Marsch vermoort waren. Si eh wandelnde Umweltbedingungen mit der Einschriinkung von Wirtschaftsfliichen dilrften aber schon in der romischen Kaiserzeit Ursache fUr die Verlagerung von Siedlungen gewesen sein. Dabei wurden auch niedrig aufgelandete Marschflachen in den Siedelraum miteinbezogen, die bis dahin noch gemieden warden waren. So waren auf den hiiufiger von Salzwasser erfal3ten niedrigen Marschfliichen bei Haferwisch die Neusiedler anders als in Tiebensee vermutlich von Anfang an zum Bau absichtlich aufgehohter Wurten gezwungen, wie archaologische und paliiobotanische Untersuchungen erkennen lassen (Abb. 2; 5). Nach den hohen Werten der Salzbinse (Juncus gerardi) und anderer salzliebender Pflanzen kann die nur NN +0,50 m hohe Marsch nicht einmal von sommerlichen Sturmfluten verschont geblieben sein. Die Datierung einer der auf dieser Obertliiche errichteten Wurten ergab sich aus einer Reihe, beim weiteren Wurtenausbau mit Klei ilberdeckter Eichenspaltbohlen, deren dendrochronologisch ermittelte Fiilldaten in die Zeit "urn oder nach 140" und "urn oder nach 168" wiesen 2• V on der Obertliiche der Marsch aus in den Untergrund eingelassene Brunnen und andere Formen einfacher Wasserstellen deuten ebenso wie Pflugspuren beetartiger, von Griiben eingefa13ter Ackerparzellen aber noch auf kurzzeitige, i.iberflutungsfreie Perioden wiihrend des Siedlungsbeginns hin. Wie eine Auswertung der paliiobotanischen Pro- ben anzeigt, ist zwar der Anbau von Vierzeil-Spelzgerste (Hordeuum vulgare vulgare), Hafer (Avena) und Leinen (Limon usitassimum) wahrscheinlich, jedoch wurden Rispenhirse (Panicum miliaceum) und Roggen (Secale cera/e) vermutlich von der Geest her importiert, da diese Kulturpflanzen nicht zeittypisch filr die Marsch sind, in Haferwisch aber aus kaiserzeitlichen Fundzusammenhiingen stammen. Durch den Ausbau der Kernwurten im 3. Jahrhundert wurde das Ackerfeld ebenso wie die iiltesten Wasserstellen von Kleiauftriigen i.iberdeckt. Im 4. Jahrhundert war schliel3lich eine nicht nur grol3ere, sondern auch hohere Gesamtwurt entstanden, deren Siedelniveau hOher als NN +2,00 m lag. Die einzelnen Ausbaustadien der bereits stark planierten Gesamtwurt zeichneten sich dabei deutlicher im Randbereich als im Kern ab. Nach Aussage der archiiologischen Sammelfunde i.iberdauerten die meisten der in der Dithmarscher Nordermarsch angelegten Siedlungen das 3. bis 4. Jahrhundert nicht. Viele der Wohnpliitze bestanden nur i.iber einen Zeitraum von ein bis zwei Jahrhunderten und wuchsen nicht zu Dorfwurten heran, wie dies vor allem entlang des iiul3eren Milndungsgebietes der Eider der Fall war. Dort ermoglichten hOher aufgelandete Uferwiille eine platzkonstant~ Besiedlung i.iber mehrere Jahrhunderte, wie die Grabungen von Bantelmann (1955) auf der Dorfwurt Tofting ergaben (Abb. 1: 2). Nach den paliiobotanischen Untersuchungen (Behre 1976) war dort wiihrend der Flachsiedlungsperiode im 2. Jahrhundert der Meereseinf1ul3 noch gering, doch ilberschwemmten bald hiiufiger Sturmfluten die Salzmarschen und zwangen die Bewohner zur Aufhohung ihrer WohnpHitze. Haufigere Sturmflutperioden mogen se it dem 3 .I 4. Jahrhundert im Nordseekilstenraum ortlich zum Verlassen der Siedlungsgebiete gefilhrt haben, deren Auswirkungen waren jedoch lokal verschieden. Wiihrend beispielsweise im Dithmarscher Kilstengebiet nur geringe Sedimentrnengen die alte Marsch bedeckten, zerstorten Sturmfluten sildlich der Elbe, im Land Wursten, nordlich der Wurtenkette der Feddersen Wierde (Haarnagel 1979; Schmid 1988; 1991; 1995), Wirtschaftsflachen. Die Reduzierung der Nutzflachen entzog den Wurtbewohnern schliel3lich die Erniiherungsgrundlage und zwang sie im 5. Jahrhundert zur Aufgabe der Dorfwurten. 3 Naturraumliche Verhaltnisse und Siedlungsmuster im friihen Mittelalter Die dendrochronologischen Altersbestimmungen filhrte dankenswerterweise Dipi.-Holzwirten S. Wrobel, Ordinaria! fiir Holzbiologie der Universitat Hamburg, durch. Die Radiokarbondatierungen ermittelte Dr. H. Erlenkeuser, Institut fUr Kernphysik der Universitat Kiel. 324 Haufige Sturmflutperioden in dem Zeitraum zwischen der jilngeren romischen Kaiserzeit und dem frilhen Mittelalter hatten teilweise erhebliche Aus- Der Wandel der Landschaft und Besiedlung im Dithmarscher Ktistengebiet Abb. 3. - Paliiographische Karte des Eiderstedter und Dithmarscher Kiistengebietes im friihen Mittelalter. ffi ~ Wurt/Siedlung L3> Hortfund © Graberfeld tl Marsch Moor und Moormarsch D • Geest Sandwall 0 wirkungen auf das Ki.istengebiet der festlandischen Nordseemarschen. Sie fi.ihrten zum Einbruch wie zur Verlandung alter Meereseinbruche, zu Landverlusten und zum Aufwuchs jungen Marschlandes entlang der Ki.iste. So hatte sich mit dem Nachlassen der Sturmflutaktivitaten in einzelnen Regionen zwischen Weser und Elbe die Ki.istenlinie seewarts verlagert. Auch fi.ir die Dithmarscher Nordermarsch mag man dies annehmen, wenn man wiederum die Verteilung der archaologisch untersuchten W ohnplatze zugrundelegt. Dort erfaJ3te die fri.ihmittelalterliche Landnahme die auJ3ere Seemarsch jenseits des kaiserzeitlichen Altsiedellandes, wahrend in der Si.idermarsch von der Elbemi.indungung bis zur Meldorfer Bucht in Ki.istennahe noch hohe, schon in den ersten nachchristlichen Jahrhunderten bewohnte Flachen zur Verfi.igung standen, WO die Kolonisation moglicherweise zunachst Wurten der romischen Kaiserzeit, wie Norder- und Si.idbusenwurth (Bantelmann 1949), erneut besiedel te, bevor auch dort Neugrtindungen, wie das 1883 erkundete Fahrstedt (Hartmann 1883), entstanden. Wahrend in der Si.idermarsch mit Ausnahme einer Notgrabung 1996 durch die Arbeitsgruppe Ki.istenarchaologie in Norderbusenwurth neuere Untersu- chungen fehlen, sind wir i.iber den Verlauf der fri.ihmittelalterlichen Landnahme in der Nordermarsch durch die Untersuchungen der Arbeitsgruppe Ki.istenarchao1ogie besser unterrichtet (Abb. 1: Nr. 3-8). Im fri.ihen Mittelalter wies dort die Marsch noch starker als in der romischen Kaiserzeit die typische Gliederung einer ki.istennahen hoch aufgelandeten Marschstreifens und eines binnenwartigen, niedrigeren, durch Stauwasserprobleme gekennzeichneten Sietlandes der alten kaiserzeitlichen Marsch auf, in dem sich teilweise Moore erstreckten (Abb. 3). Reste eines Schilftorfes wurden unter in den im hohen Mittelalter angelegten Hofwurten im Sietland bei Jarrenwisch urn 1020±50 BP (KI-3797) datiert (Abb. 3: Nr. 5). Gi.instige Siedel- und Wirtschaftsflachen erstreckten sich somit im fruhen Mittelalter nur westlich dieses Sietlandes nahe der Ki.iste. In der Dithmarscher Nordermarsch, wo der Ki.istenverlauf zwischen der Meldorfer Bucht im Si.iden und der Eider im Norden weiter nach Westen reichte, erstreckten sich mit Wohrden, Wellinghusen, Hassenbi.ittel und Wesselburen eine innere sowie mit GroJ3bi.ittel, Wesselburener Deichhausen, Westerbi.ittel und moglicherweise Si.iderdeich eine auJ3ere Reihe groJ3er Wurtendorfer westlich der alten, wahrend der 325 D. Meier 30 m -r----------------------------------~- ISH 4 um 1000 n.Chr. ISH 2 um 800 n.Chr. SH 1 um 700 n.Chr. Abb. 4. - Wellinghusen, Dithrnarschen. Archiiologische Fliichenbefunde (schernatisiert). SH = Siedlungshorizont: SH 1 Flachsiedlung urn 700 n.Chr.; SH 2 Hofwurten urn 800 n.Chr.; SH 4 Dorfwurt um 1000 n.Chr. romischen Kaiserzeit noch dicht besiedelten alten Marsch, in der sich nun ausgedehnte Moore erstreckten, wie Reste eines Schilftorfes unter der hochmittelalterlichen Hofwurtenreihe in Jarrenwisch und geologische Beobachtungen entlang der Wasser!eitungstrasse erkennen lassen (Meier u.a. 1997). Den Verlauf der friihmittelalterlichen Siedlungsentwicklung lassen die im Rahmen des DFG-Projektes Dithmarschen durchgeftihrten Grabungen aufzwei Dorfwurten der inneren Siedlungsreihe in der Dithmarscher Nordermarsch erkennen (Meier 1995; Meier u.a. 1997). Danach erschlof3 eine erste Landnahme im 7 ./8. Jahrhundert vor all em die entlang von Prielen besonders hoch aufsedimentierten Marschriicken. Ein giinstiges Siedelareal mit einem bis NN +1,80 m hoch aufgelandetem, halbinselartig von Prielen eingefaf3ten Marschriicken bot sich in Wellinghusen an (Abb. 3). Teile des Marschri.ickens waren mit Phragmites bewachsen, der ein Radiokarbonalter (KI 3797) von 1440±40 BP aufwies, was einem Kalenderalter 555660 AD entspricht (Abb. 4; 5). Im Umland des zur Landnahmezeit im 7./8. Jahrhundertnoch weitgehend sturmflutfreien Prieluferriickens erstreckten sich nach 326 Ausweis der paHiobotanischen Untersuchungen niedere Salzmarschen mit Milchlaaut (Glau;r: maritima), Strandsalzschwaden (Puccinella maritima), Salzbinsen (Juncus geradii), Schuppenmieren (Spergularien), Salz Dreizack (Triglochin maritimum) und h6here Salzmarschen mit Seggen (Carex) und Strauf3gras (Agrostis), Schilfrohr (Phragmites australis), die das Weideland der auf Viehhaltung ausgerichteten Marschensiedlung bildeten. Auf den hochstgelegenen FHichen war ein Anbau von Hafer (Avena), Gerste (Hordeum vulgare vulgare), Roggen (Secale cera/e) und Leinen (Linum usitatissimum) moglich. Wo sich der frilhmittelalterliche Kiistenverlauf befand, ist nicht genau bekannt, doch diirfte dieser kaum westlicher als der 1500 m entfemte Deich des Mittelalters gelegen haben. Wie der auf der im Durchmesser bis 250 m groBen und bis NN +6,20 m hohen Dorfwurt angelegte 30 m lange, 12 m breite und 4-5 m tiefe Grabungsschnitt erkennen Iief3, durchzog das von gr6f3eren Prielen umgebene friihmittelalterliche Siedlungsareal auf den hochsten Stellen des Uferriickens der Seitenarm eines etwa 6 m brei ten Prieles, an dessem Verlauf die altesten Wohnplatze errichtet wurden. Soweit anhand der dendrochronologisch ermittelten Falldaten nachgewiesen, wurde das auf dem Hofplatz II errichtete Gebaude der aus mehreren westostlich orientierten Wohnstallhausem bestehenden Reihensiedlung "urn 691 n.Chr." erbaut. Der Wohnplatz II war ebenso wie ein weiterer, ebenfalls randlich erfaf3ter Hofplatz (I) auf einem 0,20 m hohen Sodenpodest iiber der NN +1,80 m hohen Marsch errichtet. Da der Hofplatz I einen alteren Graben iiberlagerte, mag man einen alteren Wohnplatz weiter im Westen vermuten. Im Bereich der Kemsiedlung (Siedlungshorizont 1) fiihrte moglicherweise eine Brilcke iiber den kleinen Priel, die noch urn 785 n.Chr. in Benutzung war, einer Zeit, in der die Wohnplatze der Flachsiedlung schon zu einen Meter hohen Hofwurten (Siedlungshorizont 2) aufgehoht waren (Abb. 4: 5). Im Zuge der Ausweitung der Hofwurten wurde der Priel mit Mist verftillt und in das Siedelareal miteinbezogen. Auf der iiber dem Wohnplatz I a us Mist aufgeworfenen Hofwurt lagen zwei kurz nacheinander errichtete Wohnstallhauser (Siedlungshorizont 2). Die in W est-Ost Richtung erbauten Wohnstallhauser waren stallwarts geneigt, wobei der Wohnteil eine Hohe von etwa NN +3,00 m, der Stallteil von NN +2,50 m aufwies. Nach den Datierungen zweier gleichfalls nur noch in Resten erhaltener Nebenbauten zu schlief3en, diirfte das jiingere Wohnstallhaus "urn oder nach 785" oder/und "820 n.Chr." in Benutzung gewesen sein. Demnach erfolgte die Aufhohung von Hofplatzen der Flachsiedlung urn etwa Der Wandel der Landschaft und Besiedlung im Dithmarscher Klistengebiet NN mNN HassenbOttel +5 Welllnghusen Jarrenwlsch Haferw1sch Tlebensee heutlge H6he +5 N.~--------~~----~~--~----~~~--------~r-----~~--~~--------~ NN Anwachs (UberflutungsSedlrnente) -20 Westen Osten Abb. 5.- Siedlungsabfolge in der Dithmarscher Nordermarsch von der romischen Kaiserzeit bis in dasfriihe Mittelalter anhand geologischer und archiiologischer Untersuchungen. einen Meter friihestens amEnde des 8. Jahrhunderts, sicherlich aber zu Beginn des 9. Jahrhunderts, wobei eine sekundare Verwendung der verbauten Eichenspaltbohlen allerdings nicht ausgeschlossen werden kann. Im Laufe der Siedlungszeit verschoben si eh die grof3en Wirtschaftsbetriebe mit ihren hOheren Wohnteilen im Westen und niedrigeren Stallteilen mehrfach in Richtung ihrer Langsachse nach Osten. Weitere Auftrage, nun iiberwiegend aus Klei, erfolgten im 9. und 10. Jahrhundert, wobei im Siedlungshorizont 4 auf einer Hohe von NN +3,80 bis etwa +4,00 zwei dicht iibereinander liegende Kleinbauten des 10. Jahrhunderts erfaf3t wurden. In dieser Zeit wies die Dorfwurt vermutlich schon ihre heutige Grof3e von etwa 250 m im Durchmesser auf. Die friihmittelalterlichen Siedlungsschichten bedeckte ein bis 1,5 m machtiger Kleiauftrag des Hoch- und Spatmittelalters bis zu einer Hohe von NN +6,20 m (Abb. 5). Offensichtlich im spiHen Mittelalter endete die dichte Bebauung der Dorfwurt, in der friihen Neuzeit verblieben nur noch zwei HOfe auf der Wurt (Hausigk 1995, 56 ff.). Anders als auf der nordlich der Eider gelegenen Marschensiedlung am Elisenhof (Bantelmann 1975) konnten in Wellinghusen keine dicht iibereinander errichteten Wohnstallhauser nachgewiesen werden, sondem die Bauten der verschiedenen Siedlungsphasen waren stets durch deutliche, etwa einen Meter hohe Auftrage getrennt, ein deutliches Indiz dafiir, daf3 der Wurtenbau in der kiisten- nahen Dithmarscher Nordermarsch eine Folge hOher auflaufender Sturmfluten war. Vermutlich im 9./10. Jahrhundert erfolgte eine Ausweitung der Besiedlung in der Dithmarscher Nordermarsch. Diese beschrankte sichjedoch aufdie kiistennahe Marsch, da das vermoorte Sietland keine Entfaltungsmoglichkeiten bot. Da nicht mehr in ausreichendem Maf3e hoher aufgelandete Prieluferrander zur Verfiigung standen, erfolgte die Anlage neuer Wohnplatze auch auf den haufiger von Salzwasser iiberschwemmten, niedrigeren Marschflachen. Dort waren die Neusiedler, wie eine grof3ere Schnittgrabung auf der 2000 m nordlich von Wellinghusen gelegenen, im Durchmesser 300 m grof3en und bis NN +5 ,20 m hohen Dorfwurt Hassenbiittel andeutet, von Anfang an zum Bau von Wurten gezwungen, die im 9./10. Jahrhundert bereits eine Hohe von NN +3,00 m aufwiesen (Abb. 3: 5). Die alteste Marschoberflache lag dort einen Meter tiefer als in Wellinghusen auf einem Niveau von NN +0,80 m. Zwar deutet ein grabenumgebenes Feld mit Spuren eines Streichbrettpfluges auf iiberflutungsfreie Perioden hin, was u.a. den Anbau von Vierzeil-Spelzgerste (Hordeum vulgare vulgare) und Hafer (Avena) erlaubte, doch iiberdeckten bald 0,40 m machtige Sedimente das Ackerland (Abb. 5). Auf diesem Anwachs wurden die altesten, im Grabungsschnitt erfaf3ten Wurten errichtet. Das Siedlungsniveau einer der nur randlich erfaf3ten, aus Klei auf327 D. Meier geworfenen und randlich mit Sodenwiillen gesicherten Primarwurten (Siedlungshorizont 1) lag mindestens bei NN +2,00 m. Stellenweise uber dieser erfolgte im 9./10. Jahrhundert die Errichtung einer bis NN +3,00 m hohen Hofwurt (Siedlungshorizont 2), deren Auftrag im unteren Teil aus Mist bestand, den mehrere Kleisodenlagen bedeckten. Die Wurt bot Platz fur ein grof3es Wohnstallhaus, dessen nach Westen abfallender Stallteil angeschnitten wurde. Nordlich dieser Wurt erstreckte sich eine weitere, ebenfalls aus Mist und Klei aufgehohte Hofwurt, die ebenfalls eine Ho he von etwa 2 m iiber der Marsch aufwies und in einem Arbeitsgang errichtet wurde. Unmittelbar uber den friihmittelalterlichen Wohnschichten lagen Auftrage des Hoch- und Spatrnittelalters bis zur heutigen Hohe der teilweise noch bebauten grof3en Dorfwurt. Zwar uberwog auf beiden Dorfwurten die agra,risch ausgerichtete Wirtschaftsform, doch deuten gegen landwirtschaftliche Produkte eingetauschte Importe wie Wetzsteine aus norwegischen Tonschiefer, Dreilagenkamme, Mahlsteine aus dem Rheinland und Fibeln aus dem karolingisch-ottonischen Reich auf eine partielle Teilnahme am Nah- und Fernhandel hin (Meier u.a. 1997). So fallt die Landnahme wie die Siedlungsverdichtung in den Dithmarscher und Eiderstedter Kustengebieten in eine Zeit des Aufbluhens des Frankisch-Friesischen Handels. Anders als entlang der niedersachsischen und niederlandlischen Kuste entstanden jedoch keine Langwurten als Zentren des Nah- und Femhandels wie des Gewerbes. Speziell auf den Handel ausgerichtete Siedlungen gab es im fruhen Mittelalter weder in den Marschgebieten Eiderstedts noch Dithmarschens. Die Bevolkerung auf den Wurten bewahrte im fruhen Mittelalter in Dithmarschen weitgehend ihre Unabhangigkeit von jeglicher auswartiger Oberhoheit. Dabei sicherten die Stellerburg und die Bokelnburg in Dithmarschen als Ringwalle den sachsischen Gau vor Angriffen von auf3en. Ob und in welchem Maf3e sich die Machtverhaltnisse nach der karolingischen Eroberung des nordalbingischen Sachsens anderten, bleibt unklar. Lediglich einer Auf3erung Helmold von Bosaus laf3t sich entnehmen, daf3 in Dithmarschen vor der Stader Herrschaft eigene Grafen regierten, deren gesellschaftlicher Status aber unbekannt bleibt. Grof3 dilrfte ihr Einf1uf3 in den Kustengebieten nicht gewesen sein, da die dicht besiedelte Zone der Seemarschen durch weite, vermoorte Sietlander von der Geest her nahezu unzuganglich war. 4 Hohes und spates Mittelalter Seit dem hohen Mittelalter erfolgte in Dithmarschen ebenso wie in anderen Bereichen des Nordsee328 kiistengebietes (Schmid 1988; Schmid 1991) ein mit der Bedeichung einhergehender Landesausbau, der neben der Seemarsch vor allem das bis dahin nicht nutzbare Sietland erfaf3te. Diese wasserreichen Sietland- und Moorgebiete boten im Mittelalter nur schwer zu kultivierende, aber notwendige FHichen fur die Landnutzung und Ansiedlung einer zunehmenden Bevolkerung (Nitz 1994, 248 ff.). Der aus den bereits vorhandenen Siedlungsraumen in den Seemarschen heraus erweiterte Landesausbau bildet eine Binnenkolonisation kolonisierter Naturraume am Rande oder zwischen alteren Siedlungen und ihren Kulturflachen. In diesen Gebieten kam es zur Ausformung bestimmter Siedlungsstrukturen in Form von Marschenhufensiedlungen. Die Anlage dieser Siedlungen ist eine unmittelbare Folge des genossenwirtschaftlichen Kolonisationsvorganges mehrerer Familien- oder Personalverbande. In Dithmarschen waren sozial und wirtschaftlich fuhrende Schichten auf den altbesiedelten Dorfwurten entstanden, die zu Tragem des Landesausbaus wurden. So bildete ne ben W esselburen vor all em Wohrden mit urkundlich im Jahre 1281 uberlieferter Kirche, Markt und Hafen ein wichtiges regionales Zentrum, das im spa ten Mittelalter sogar Handelsverbindungen mit Lubeck besaf3 (Abb. 6). Wahrend Wohrden und Wesselburen seit dem hohen Mittelalter noch an Bedeutung gewannen, fielen andere Wurtendorfer teilweise oder ganz wtist. Dort wo Dorfwurten wie Wesselburen moglicherweise ihren Wasseranschluf3 verloren entstanden an der Kuste mit Norddeich und Reinsbuttel Langwurten, von deren Kleinhiifen aus eine Verteilung der Waren erfolgte. Ebenso wie diese Wurten wurden mit Schiilp und Busumer Deichhausen langrechteckige Wurten mit umlaufender Ringstraf3e und schachbrettformiger Parzellenstruktur erst seit dem 12. Jahrhundert (Meier 1995, 102 ff.) errichtet. Seit dieser Zeit schutzte ein kustenparalleler Dei eh die Dithmarscher Nordermarsch, der im Westen die Insel Busum vorgelagert war, deren Reste seit der zweiten Halfte des 16. Jahrhunderts an das Festland angedeicht wurden. Der mit der Bedeichung der Seemarsch einhergehende hoch- und spatmittelalterliche Landesausbau erfaf3te jedoch vor all em das his dahin nicht nutzbare durch Moore und Stauwasserprobleme gekennzeichnete Sietland, wo reihenf6rmige, von Geschlechtem als Siedlungsgenossenschaften angelegte Marschhufensiedlungen en~standen (Stoob 1951; 1953). Wie archaologische Grabungen auf einer Hofwurt in Jarrenwisch zeigten, wurden im hohen Mittelalter die ersten Hofwurten in der westlichsten der drei Siedlungsreihen noch auf dem nicht oder nur teilweise abgetorften Moor angelegt, das ahnlich wie in Zentralholland durch Oxidation und Entwasserung ver- Der Wandel der Landschaft und Besiedlung im Dithmarscher Ki.istengebiet schwand und deren Relikte nur unter einzelnen Hofwurten ("Restheem") erhalten blieben (Borger I984). Die Kolonisation des Sietlandes schuf die V oraussetzung fUr den Aufstieg der Geschlechter als Personalverbande, die bis I300 den Adel aus dem Lande drangten (Stoob 195I; 1953). Ihre Macht dokumentierte sich am auffalligsten in der Griindung eigener Kirchen wie Neuenkirchen (1323) in der Nordermarsch und Barlt (I428) in der Si.idermarsch. Zur 1323 urkundlich erwiihnten Kirche in Neuenkirchen gehorte ein umfangreicher Landbesitz, der im I8. Jahrhundert etwa 54 ha betrug (Hausigk I995, 85). Reste des an das Kirchspiel angrenzenden W eif3en Moo res wurden bis I5 80 unter den Kirchspielen Hemme, Weddingstedt und Neuenkirchen aufgetei1t. Im Unterschied zu den Reihensiedlungen mit ihren vielen Hofstellen konzentrierte sich in dem zentral zwischen den Marschenhufensiedlungen ge1egenen Kirchort die Verwaltung wie das Gewerbe und Handwerk. AmEnde des 13. Jahrhunderts entwickelten sich die 15 Kirchspiele in Dithmarschen zu den eigentlichen Tragern der politischen Gewalt (Stoob 1953). In ihnen bildeten seit der zweiten Halfte des I2. J ahrhunderts die fi.ihrenden Grof3bauern der Personalverbande als "clavigeri" und "iurati" Kollegien, die den Ratskollegien von Stiidten vergieichbar waren. Diese Kollegien steuerten Deich-, Entwiisserungs- sowie Wegeanlegenheiten und sprachen Recht. Im spaten Mittelalter galt der "Norderstrand" als der reichste Teil des gesamten Dithmarscher Landes. Literatur BANTELMANN A. 1949: Ergebnisse der Marschenarchaologie in Schleswig-Holstein, Offa 8, 75-88. BANTELMANN A. 1955: Tofting, eine vorgeschichtliche Warft an der Eidermiindung, Offa-Bi.icher 12, Neumi.inster. BANTELMANN A. 1975: Die friihgeschichtliche lvfarschensiedlung beim Elisenhof in Eiderstedt. Landschaftsgeschichte und Baubefunde, Stud. Ki.istenarch. Schleswig-Holstein, Ser. A, Elisenhof 1, Bern-Frankfurt. BEHRE K.E. I976: Die Pjlanzenreste aus der friihgesch:'chtlichen Warft Elisenhof, Stud. Ki.istenarch. Schleswig-Holstein, Ser. A, Elisenhof 2, Bern-Frankfurt. BEHRE K.E. I987: Meeresspiegelschwankungen und Siedlungsgeschichte in den Nordseemarschen, Vortrage Oldenburgische Landschaft I7, Oldenburg. /~l \,J.- BOS!Jm I ' I 0 • 0 Geest Moor Marsch a Dorfwurt A• R Hotwurt Klrche M Markt 0 2 km ~ Hafen """-' Delch Abb. 6. - Hoch- und spatmittelalterlicher Landesausbau, fruhe Bedeichung und Kirchengriindungen in der Dithmarscher Nordermarsch. BOKELMANN K. I988: Warften und Flachsiedlungen der romischen Kaiserzeit. Ergebnisse einer Prospektion in Norderdithmarschen und Eiderstedt. In: MOLLER-WILLE M. u.a., Norderhever-Projekt. 1 Landschaftsentwic!dung und Siedlungsgeschichte im Einzugsgebiet der Norderhever (Nordfriesland), Offa-Bi.icher 66 (=Stud. Ki.istenarch. Schleswig-Holstein, Ser. C, I Norderhever-Projekt), Neumi.inster, I49-I62. BORGER G.J. 1984: Die mittelalterliche und friihneuzeitliche Marschen- und Moorbesiedlung in den Niederlanden. Einige Bemerkungen zum Forschungsstand, Siedlungsforschung 2, I 0 I-I1 0. BUCHHOLZ H.G. I963: Eine eisenzeitliche Siedlung auf dem Elpesbi.itteler Donn in Dithmarschen, Offa 20, 1I6-124. HAARNAGEL W. 1979: Die Grabung Feddersen Wierde. ivfethode, Hausbau, Siedlungs- und Wirtschaftsformen sowie Sozialstruktur, Feddersen Wierde 2, Wiesbaden. HARTMANN R. 1883: Die a/ten Dithmarscher Wurten und ihr Packwerkbau, Marne. HAUSIGK O.F. 1995: lvfittelalterlicher und friihneuzeitlicher Landesausbau in Norderdithmarschen am Beispiel zweier Siedlungen in der Marsch. Historisch-geographische Untersuchungen, Ungedr. Diplom-Arbeit, Kiel. 329 D. Meier HINGST H. 1983: Die vorromische Eisenzeit Westha/steins, Offa-Bi.icher 49 (UmenfriedhOfe Schleswig-Holsteins 8), Neumi.inster. HOFFMANN D. 1986: Beobachtungen und Daten zur ji.ingeren Entwicklung des Ki.istengebietes von Dithmarschen und Nordfriesland, Offa 43, 251264. MEIER D. 1992: Fri.ihe Deiche in Eiderstedt. In: STEENSEN Th. (Hrsg.)., Deichbau und Sturmfluten in den Frieslanden, 2. Historikertreffen Nordfriisk lnstituut 1991, Husum, 20-31. MEIER D. 1994:, Landschaftsentwicklung und Siedlungsmuster von der romischen Kaiserzeit bis in das Mittelalter in den sch!eswig-holsteinischen Marschen, Schr. Naturwiss. Verein SchleswigHolstein 63, 117-144. ME !ER D. 1995: Die Neubesiedlung der Dithmarscher Seemarsch und der Wandel der Kulturlandschaft. In: RADTKE U. (Hrsg.), Vom Sudatlantik bis zur Ostsee- neue Ergebnisse der Meeres- und Kustenforschung. Beitrage des 13. Jahrestagung der Arbeitskreis Geographie der Meere und Kusten, Kolner Geographische Arbeiten 66, 1995, 97-108. MEIER D., HOFFMANN D. & MOLLER-WILLE M. 1989: Zum mittelalterlichen Landesausbau Eiderstedts. Ein Forschungsprojekt der Arbeitsgruppe Ki.istenarchaologie am Forschungs- und Technologiezentrum Bi.isum, Offa 46, 285-300. MEIER D., HOFFMANN D. & MOLLER-WILLE M. 1997: Geologische und archaologische Untersuchungen zur Landschafts- und Siedlungsgeschichte der Dithmarscher Nordermarsch, Germania, im Druck. MENKE B. 1988: Die holozane Nordseeki.istentransgression im Kiistenbereich der si.idostlichen Deutschen Bucht. In: MOLLER-WILLE M. u.a., Norderhever-Projekt 1. Landschaftsentwicklung und Siedlungsgeschichte im Einzugsgebiet der Norderhever (Nordfriesland), Offa-Bi.icher 66 (Studien Ki.istenarch. Schleswig-Holstein, Ser. C, 1 Norderhever-Projekt), Neumi.inster, 117-137. NITZ H.J 1994: Historische Kolonisation und Plansiedlung in Deutschland. Ausgewahlte Arbeiten Bd. I, Kleine Geographische Schriften Gottingen 8, Berlin, 1994. SCHMID P. 1988: Die mittelalterliche Neubesiedlung der niedersachsischen Marsch. In: W ATERBOLK H.T. (Hrsg.), Archeologie en landschap. Bijdragen aan het gelijknamige symposium, Groningen, 133-164. SCHMID P. 1991: Mittelalterliche Besiedlung, Deichund Landesausbau im niedersachsischen Ki.istengebiet. In: B6HME H. (Hrsg.), Siedlungen und Landesausbau zur Salierzeit. 1 In den nordlichen Landschafien des Reiches, Sigmaringen, 9-36. ScHMID P. 1995: Archaologische Ergebnisse zur Wirtschaftsweise in der Marsch. In: H.-E. DANNENBERG & J.-H. SCHULZE, Geschichte des Landes zwischen Elbe und Weser. Bd. l Vor- und Friihgeschichte, Stade, 221-250. STOOB H. 1951: Die dithmarsischen Geschlechterverbande. Grnndfragen der Siedlungs- und Rechtsgeschichte in den Nordseemarschen, Heide. STOOB H. 1953: Dithmarschens Kirchspiele im Mittelalter, Zeitschr. d. Ges.f Schlesw. Hoist. Gesch. 77, 97-140. WIERMANN R. 1962: Botanisch-moorkundliche Untersuchungen in Nordfriesland. Ein Beitrag zur Frage nach dem zeitlichen Ablauf der Meeresspiegelschwankungen, Meyniana 12, 1962, 97-146. Dr. Dirk Meier Forschungs- und Technologiezentrum Westktiste Zentrale Einrichtung der Christian-Albrechts-Universitat Kiel, Arbeitsgruppe Ktistenarchaologie Hafentom 25761 Btisum Deutschland 330 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 DellaHooke The effect of English settlement in medieval North Wales Abstract After the conquest of Wales by Edward I in 1284 native farming was to be completely reorganised under the influence of settlers from the English burhs. The changes which occurred are reflected in the present-day landscape: the resulting palimpsest is investigated in a region of the Conwy valley of North Wales. The evidence of landscape archaeology is set alongside other sources of evidence. Introduction The landscape of upland North Wales has been as affected by the intervention of man as any lowland landscape. Prehistoric occupation reached relatively high levels on the slopes of the mountain ranges and into the most remote valleys of Snowdonia. It is considered that the removal of tree cover in such marginal areas of low fertility and high rainfall contributed towards the soil deterioration which produced the impoverished soils found today and to the formation of the barren peat-covered moorlands which are so characteristic of parts of the region (Walker & Taylor 1976). Similar landscapes can be found throughout western Britain, from the Western Isles off the Scottish coast to the peat moors of Connemara in western Ireland. By the first millennium AD, however, the focus of settlement had become firmly established in the main valley regions. Even in prehistoric times the greatest concentration of settlement has been observed in such regions and in the Roman period improved access may have been a contributory factor (RCH MW 1964; Casey 1969). Jones (1972, 295) has noted how the distribution of sculptured stone monuments of the Early Christian period reinforces the suggestion that economic activity was focused on the lowlands in the fifth and sixth centuries. In Gwynedd, the uplands of Snowdonia were primarily a place of refuge, offering security and a place to regroup when outside threat became too dangerous. These and other upland areas, however, also offered valuable resources: the earliest documentary evidence suggests that the marginal uplands served as areas of seasonal grazing, their resources complementing those of the more intensively developed lowlands. Unfortunately, the earliest Welsh documentation dates in its present form from no earlier than the later twelfth century; this includes the law books such as Liber Landavensis, etc. Although these sources contain much that is of genuinely early medieval derivation, this is difficult to disentangle from the surviving texts (Jones 1972, 284-287). The most useful detail for the reconstruction of local communities is found in the medieval extents, compiled after the English conquest. Nevertheless, just as Domesday Book, compiled soon after William's victory at Hastings, portrays the situation prevailing in late AngloSaxon England, so the medieval extents reveal the agrarian communities and systems of the native Welsh. It remains true, however, that 'the contemporary archaeological evidence remains our surest guide since it enables us not only to test the written evidence but also to view it clearly in its territorial setting. More of the history of Wales lies in the earth than is recorded in our libraries ... ' (ibid., 286). To this might be added the evidence on the earth, for landscape archaeology has shown that much field evidence survives, especially in marginal zones. Ideally, investigation adopts an inter-disciplinary approach reviewing the evidence from all available sources including, in addition to the documentary and cartographic sources, that of place-names (Hooke forthcoming) and literature, and among the wide range of archaeological techniques, air photography (Musson n.d.; Aris 1996) and field survey, complemented where possible by judicious excavation and environmental analysis. Medieval communities The medieval extents reveal valley-based communities utilising the hill pastures, probably initially 331 D. Hooke upon a seasonal basis. Thus the extent of 1284 (Extent of Merioneth 1284) names the king's pastures and vaccaries of Ardudwy, a division ofthe cantrefof Dunoding in Gwynedd, as the pastures ofBryn Coch, Y Feidiog and Prysor, lying to the east of the Rhinogau range; these could have sustained over 200 stock had they then not been devastated by war (Fig. 1). In the Record of Caernarvon (1420 for Merioneth; Ellis 1838), these havotref!rith provided pasture (herbag f!oreste) for the stallions of Uwch Artro. The royal vill or maerdrefofthe commote of Ardudwy Is Artro lay at Ystumgwem, upon the shores of the Cardigan Bay, and its bondmen held appurtenant pastures in the rocky valley ofCwm Nantcol between the Rhinogau and the sea (Extent of Merioneth 1284). Studies by Jones and Gresham have shown how in the most servile tenure of the maerdref a regular system of land allotment ensured that the community was responsible as a whole for provisioning the royal vill and both here and on the holdings of free tribesmen the arable strips formed intermingled holdings of the type reconstructed for Castell (below). Outside the maerdref, rights to land usually operated through kin groups, with systems in place for the reallocation of land to prevent excessive morcellation of holdings once boundaries had become fixed. However, the process of territorial fragmentation was at work here in Wales before the Edwardian conquest. As in England, tribal society was becoming increasingly landoriented as the boundaries of both township and native land units (the gafael and the gwely, which might be scattered through several townships) were laid down, possibly in the twelfth century (Gresham 1987, 137-138). The township was closely related to the ecclesiastical parish, an area served by a church, which was a system of organisation already long used by the church throughout England. By 1326 it was a clansman of the township of Llanaber, rather than those of any other land unit, who was 'fined for keeping his animals "in the common pastures of the old settlement'' (in communi pastura del hendreve) after the community of the township (communitas villate) had moved, early in May of that year, with its animals to the mountains' (Jones 1972, 298). In the fifteenth-century extent, the bond and free holdings of tribal Wales are grouped into such townships but it may have been a genuine fact that the boundaries of eight native bond gweliau in Llanaber, Llanddwywe and Llanenddwyn were unknown. Their lands were probably intermingled, making use of the narrow coastal strip which lies here between the Rhinogau and the sea (Hooke 1975) (Fig. 2). Jones Pierce (1939, 1972) has shown how the holdings of the native Welsh were similarly intermingled in the township of Castell, now part of the 332 parish of Caerhun. This parish lies on the west bank of the River Conwy on the eastern side of the Carneddau range, also within the kingdom of Gwynedd. The locations ofCastell and its appendant hamlets of Penfro, Bodidda, Cymryd and Merchlyn can be roughly identified, spread across the present-day parishes ofGyffin, Caerhun and Llanbedrycennin. Castelllay in the commote of Arllechwedd Isaph but the site of the maerdrefhas not been identified. A likely location is close to the matte ofBryncastell in the hamlet of Gronant, close to an important crossing point of the Conwy, rather than at the Roman fort of Kanovium in Caerhun (Gresham 1979) (Fig. 1). Inland, in the parish of Llangelynnin, the place-name 'Ffriddlys' may indicate the summer pastures of the llys (court) or maerdref By c. 1350, however, a further bond gafael in the remote inland valley of Cwm Eigiau may represent other pastures which had originally belonged to Castell (Ellis 1838, Record of Caernarvon). It is not clear whether these were occupied throughout the year. A period of more clement climatic conditions in the late medieval period apparently permitted intermittent cultivation to take place at relatively high altitudes (Jones 1964, 24-26). The land in Cwm Eigiau may represent an assart on the mountain land and, unlike many other native gafaelion, it was demarcated by precise metes and bounds, recorded in the sixteenth century (Univ Wales Bangor, Baron Hill 2453), but was escheat (taken from native tenure to the king) by 1352. It is the period of the demise of the native holdings which provides the closest insight into their structure. Some were entirely denuded of their occupants: Glyn and Gronant in Castell, together with the Cwm Eigiau gafael, were emptied at this time. Before any holdings subject to the particularly restrictive tenure of tir cyfrif('reckoned land'- usually associated with the maerdref or royal demesne), could be taken over by the emerging landholding classes after the Edwardian conquest the land had to be freed from this associated restrictive bond tenure and other bond lands would similarly be avoided by prospective tenants so long as freehold tenements were available (Jones Pierce 1972, 47-48). The lands of the bondsmen ofYstumgwem in Ardudwy is Artro were given to villein families displaced by the building of Edward I's borough at Harlech (Gresham 1988). Twenty-two free clans were recorded in Castell in 1352, occupying that number of gafaelion (Ellis 1838, Record of Caernarvon). Their lands had been fragmented but with recognisable focal areas: 'none of the gafaelion of Castell can be regarded as a distinct territorial unit demarcated from its neighbours by definite metes and bounds like the nearby bond gafael ofCwm Eigiau, unless an organisation of that The effect of English settlement in medieval North Wales 10 miles 10 20 km . '' Upland range V • Royal vill Ystumgwern + Early religious centre 0 Borough IE!iiJiiliSEIIIi!!lllll!illl 1 NAME 1 Cantref boundary ~---- ~ .................... i Name i ..................... Commote boundary Transhumance Link ~------, 1 l ______ 1 I Area of figures 2 & 3 Fig. 1. -Administration units in North Wales. kind be conceded to the eight gafaelion limited to Castell proper. The remainder must have been divided into at least two, and in some cases into as many as five, divisions' (Jones Pierce 1972, 201 ). Most were named after the ancestors of the tribal groups holding them in the fifteenth century and some can be traced through into later documents as their lands passed to new owners (Gresham 1965, 33). These 333 D. Hooke help to cast light upon the pattern of native landholding and also show how the early systems was to be almost entirely eradicated, giving rise to new settlement and field patterns. First, to look backwards to what can be deducted about the native patterns of land tenure: Castell is unusual in the amount of documentation that has survived, for a fifteenth-century rental (Univ Wales Bangor, Baron Hill1939; Gresham 1965) shows how one Bartholomew Bolde, a descendent of a Lancashire family who had presumably moved into North Wales as settlers in the English plantation boroughs after the Edwardian conquest, was acquiring the lands of Welsh peasant farmers piecemeal in order to build up his estates. The rental identifies many of the former native holdings: some sixty cottages, 600 acres of arable, 200 acres of meadow and 1000 acres of pasture and permits their scattered locations to be identified. Subsequent deeds illustrate their subsequent amalgamation into farming units. Jones Pierce was able to reconstruct the layout of the arable and meadow of the eight Castell gafaelion which lay on the valley land above the marsh and the river: 'On the site occupied a hundred years ago by nine holdings there stood at one stage in the past holdings belonging to as many as sixty-nine tenants', their tyddynnod (homesteads) situated just within the boundaries of the inner hamlet ofLlwydfaen or between the arable and the lower lying meadowland of Morfa Llwydfaen. Jones Pierce pictures the scene which would have met a traveller landing at Cafn Gronant and making his way towards Bwlch y Deufaen: he would see the tyddynnod, erwau, and drylliau of the vill of Gronant, once cultivated by the prince's bondmen and now held in villeinage by free tenants ofCastell, covering the rises on either hand and merging imperceptibly, unbroken by wall or hedgerow, into the fields of Llwydfaen ... (ibid., 209). Elsewhere he describes 'a patchwork of quilleted fields' and notes that the holdings purchased by Bolde were dispersed in scattered and widely separated parcels- a holding of 309 acres might be held in as many a ten parcels. This fragmentation he interpreted as arising from the subdivision of former clan holdings, subdivided by the later Welsh law offamily inheritance. The average holding of a native tribesman may have been as little as eight or ten acres by the beginning of the fifteenth century, widely scattered throughout the township and its hamlets and with parcels of arable and meadow thoroughly intermingled with each other and separated by the strips belonging to others - even the lands of individual gafaelion were intermingled (Gresham 1965). 334 Since cultivation continued over most of the lowlands in the same places throughout the centuries, the archaeological field evidence of medieval fields and settlements is generally very poor. Only on the borders of marginal zones have field relics relating to the ebb and flow of settlement survived to any appreciable extent. These are usually located on the margins of the upland and only rarely have any relict systems been identified in the main areas of settlement below; occasional patches of poor terrain at low levels which have subsequently been ignored by development are unfortunately unrepresentative of the areas usually sought out by medieval cultivators. Sometimes the field remains are sufficiently clear for detailed survey but less obvious remains are now also becoming apparent on air photographs. Rarely has any of this evidence been dated by excavation or environmental analysis. Most prolific are the remains of long huts of various sizes, often surviving above the present level of cultivation. They either stand alone or are related to simple paddocks and irregular field systems; only occasionally are there signs of arable cultivation (RCHMW 1956). These are usually unrelated to the later farm complexes and are likely to pre-date them in use. Some huts may represent continually occupied settlements, either representing a period of land pressure or the period of climatic optimum noted above, but many others are at heights which suggest they were associated with the seasonal pasturing of the upland commons. It is rare for them to link with any surviving field name and although a long hut stands in a hafod-named field in Llanaber the boundaries of the field itself clearly cut through other adjacent long huts (this may represent two periods of use). Field systems have also been identified on air photographs. Jones (1972, 344-345, fig. 45, pi. Ill) showed how strip systems photographed by the Air Force Department in Llanynys, Denbighshire, appeared to represent a survival of nucleal land (tir corddlan) and hereditary land (tir gwelyog) associated with an early clas (clerical community). On the Great Orme, near Llandudno, overlying systems of cultivation have recently been identified, arising from two distinct periods of cultivation (Aris 1996). Linear systems of ridge and furrow may overlie .an earlier complex of regular rectangular fields which give rise to pronounced regularly spaced lynchets influencing the size of the ridges themselves. Although no dates for either of these systems have been established they lie close to the township foci of the medieval extents. Similar strip field complexes identified on Ffridd Cam en along the western edge of the Berwyn in Clwyd at heights of between 335 and 395 metres O.D. have tentatively been interpreted as medieval in date, associated with a long hut which may The effect of English settlement.in medieval North Wales • Harlech LLANDDWYWE Upland range \ 305m (1 OOOft) contour Area of Long hut clusters (Ardudwy Is Artro only; individual sites not shown) Commote division boundary Land of Ystumgwern maerdref + Gwern y cape! Trackway 0 Free land of bond gafael Bond holding >< Field-names indicative of arable strips (Lianaber only) Commons enclosed C1 9 Fig. 2. · Medieval/and holdings in Ardudwy Is Artro. have been a permanently occupied farmstead. Smaller huts in the vicinity may have been ?earlier hafod settlements (Silvester 1991). These recently recognised systems of regular strip fields are not unlike those found at higher levels on Bodmin Moor (Johnson & Rose 1994,97, 107-115). 335 D. Hooke At 274-305 metres O.D. on Brown Willy a cluster of medieval long huts were associated with regular strip fields; these were later overlain by a later system of larger fields which incorporated some of the earlier boundary walls. It is thought that, on Bodmin Moor, cultivation of the linear strips may have begun in the late fourteenth century after earlier experimental sowings had been tried and presumably found successful. At Roughtor South, between 289 and 305 metres O.D., irregular ridges within the fields may have been produced by either ploughing or spade cultivation (ibid., 67) but at numerous sites more regular broader ridging appeared to be of postmedieval origin. The linear strips were thought to represent a system of subdivided arable within an in-field, out-field context. Snatch crops could be taken from the upland commons, if necessary, in most farming systems of upland Britain and such fields have been identified by aerial photography on some of the other Cornish moors. This would also be applicable to what is known for Wales but the strip fields of the Great Orme and the Berwyn seem to be far too pronounced to represent temporary cultivation. Even when medieval field systems have been eradicated by later cultivation, as seems to be the case in the Conwy valley area of Caerhun, the area of medieval arable can often be identified from fieldnames recorded in post-medieval documents. Terms such as dry!!, 'a strip or ridge', talar, 'a headland', lleiniau, 'quillets' and erw and cyfar, measures of land, seem to refer to the sub-divided arable strips of the medieval fields (Hooke 1975; Thomas 1980, 1992). In Llanaber these were strongly concentrated along a narrow band on the coastal strip below the 152-metre contour (Fig. 2) but in Caerhun, despite a predominantly lowland location, they also spread along access routes (a Roman road) onto the lower slopes of the hills (Hooke 1997 forthcoming) (Fig. 3). In the coastal area of Ardudwy is Artro, the names of some of the native holdings became attached to the farms which replaced them (Thomas 1970; Davies 1988; Gresham 1988) (Fig. 2). A parcel of free land called Egrin became a substantial holding (Egryn) by the fifteenth century, supporting a noble open-hearth hall-house (Smith 1975, 114, fig. 53), and other free parcels identified include Taltrefthyn (Taltreuddyn) and Taluru (Talwem). Taltreuddyn was, however, still a hamlet of thirteen households in an extent of 1293 (PRO, E 179/242/53; Thomas 1970, 130, n. 5) and only later was it to be replaced by a consolidated holding with a substantial ?post-Tudor house. The gwelyau (hereditary land) of Goleedd, Llanvechwy, Llecheithior, y Ben Erwe, Llanddwywe, and Bryn y voel (often mis-spelt by the English commissioners), 336 which lay intermingled through the three coastal townships of Llanaber, Llanddwywe and Llanenddwyn, are found giving their names to the later farmsteads or hamlets of Golodd, Llanyfachwenisaf, Llecheiddior, Bennar, Llanddwywe and Bron y foe!; Ythaildreffis remembered in the modem farm-name Faildref but was not in the same location; the bond land of Keirtrejj' is remembered in a house name outside Barmouth. Only one of the names on the coastal side of the Rhinogau range referred to a holding above 150 metres (500 ft) (Thomas 1970) and only Golodd lay beyond the coast range, the later farm occupying a site at the head of Cwm Sylfaen at a height of245 metres (800ft) just over a col, Bwlch y Rhiwgyr, which leads up from the coast and the valley of the Y sgethin. The demesne holdings of the royal vill of Ystumgwern are represented in the Edwardian extents by bond gafaelion, some of which have been identified. One of the bond gafaelion (a non-hereditary holding), y Migliwe, of Llanenddwyn reappears as a farm Migliw on the banks of the Artro, others lay in the valley of the Afon y Gornant, including Ygornant, represented today by a farm-name Gornant. Other bond holdings identified include Llwyngwian, andy Lloynon (LLwyn-yn), all situated on the coastal lowland strip. By 1420, seven of the twenty bond gafaelion of Llanenddwyn had been converted into fir mal, or holdings which were available for renting out to freemen. Gresham (1988) traces these holdings back to an extent of 1284 and by identifying these he has been able to locate the site of the maerdrefland. Four of the gafaelion noted as 'the prince's land' in 1420 were called Y Faerdref(cf maerdref) in 1284, individually named perhaps after the men who 'may well have been the heads of the displaced families settled there': Gwyn y moch, Ieuan ap Grono y lloc, his brother Einion, and Cocholyn. In the sixteenth century, eight cottages on 109\/z acres of these four gafaelion were listed in a lease of crown land but a further sixteen acres were unaccounted for. At some time a chapel had been built upon this land (possibly associated with the original prince's hall, which had been moved to the castle at Harlech after the conquest), and it was the official search for this site that had led to the documentation which pin-pointed the land of the gafaelion. The chapel, Cappell Teulwyd, was found to be then in existence but unused, its land encroached illegally !:Jy Griffith ap John ap Ieuan ap Einion. Allocated with sixteen acres to a nearby estate, some of this land bears the present-day fieldname Gwern y Cape!, immediately to the west of the modern farm of Faildref between it and Morfa Meirion. The modern farm ofYstumgwem lies adjacent on the southern side. The effect of English settlement in medieval North Wales v.::::·tl 305m Contour (1000 ft) Parish boundary Commons enclosed C19 Early arable names Other early arable names not precisely located 1 km Upland Fringe MB Maen-y-bardd G TA Garth-mor Tyddynrobin TE Tyddyneithiniog R Rowlyn Roman road ICl Roman fort Valley Farms LL Llwydlaen s Soglog MC F HG T G Maesycastell Farchwel Garthmor Upland Farms M Maenelra D Halodygarreg Tal-llyn Eigiau Area of Figure 4 Fig. 3.- Farm holdings in Caerhun and Llanbedrycennin. Close to Ystumgwem, a small area of relict fields associated with a possible long hut survives in a particularly rocky location beside the Barmouth-Harlech road. This may merely represent late encrqachrnent onto an area left as common grazing but there is a hint of linear strips some fifteen to twenty metres in width associated with clearance mounds indicating a first stage in land clearance (Hooke 1983, 251-252, fig. 3). It is doubtful whether this land would have been cleared to this extent unless it was going to be used for cultivation, even ploughing. In general, wellpreserved long hut groups and associated irregular paddocks are best preserved along this coastal strip on the hill slopes above the coastal arable (Hooke 1975, 1983) (Fig. 2). Clusters of long huts on benches ofland at heights of 180 and 214 metres seem mostly to have served as seasonal settlements, although a small number oflower buildings are more substantial and associated with ancillary buildings, suggesting that they became at some stage permanent farmsteads. This area has recently been rephotographed from the air and fresh computer-drawn surveys produced, resulting in the recognition of additional sites (RCHMW, SMR). These clusters are clearly earlier than the largeffriddoedd enclosures which cut across and through them and a number lie above the mountain wall. Most lie on the seaward side of the coastal range but a previously unrecorded site has now been located above Golodd (the one named gafael which lay beyond the coastal ridge). A scatter of ill-pre337 D. Hooke served long huts, including a platform house cut into the slope, lie at a height of275-306 metres (900-1000 ft) beside the lower part of the trackway which crosses the range by the Bwlch y Rhiwgyr. In Caerhun and Llanbedrycennin long hut groups are more widely dispersed across the upper moorlands but with the main clusters lying along the margin of the hill commons, some on land known to have been enclosed only in the sixteenth century (Hooke 1997 forthcoming a). A group lie within the bounds of the Cwm Eigiau gafael, located between 3 97 and 428 metres on land sloping down to the upper waters of the Afon Eigiau, but are in a poorly preserved state. There is considerable variation in the size of individual huts in both Llanaber parish and the parishes of Caerhun and Llanbedrycennin, although in both areas they are generally rectangular shape and set at right angles to the slope. Some are slightly set into the slope, giving a curving hood around a levelled platform, the lower part of which stands on an 'apron' of earth: all the Cwm Eigiau huts were of this type. In Llanaber, many long huts were clearly divided into two compartments by an internal dividing wall. Size varies: on the holding of Llwynwcws in Llanaber the best-preserved long huts vary from 7.9 m to 9.8m in length, with widths of between 4.9 m and 5.5 m. In Caerhun and Llanbedrycennin size is more variable with some huts over 13 m in length and with widths of up to 5.5 m, although some are also much smaller, no longer than 6.1 m and only 2.4 m in width. It is unusual for anything other than foundation walls to survive and sods may have been used for upper courses on many occasions, with straw or heather thatch used for roofing. The classification by size currently being carried out by Cadw: Welsh Historic Monuments may well provide a clearer understanding of the nature of these dwellings. In the interim, it may be suggested that the rectangular long hut was the normal form of medieval native dwelling throughout upland North Wales, this form being used for both permanently occupied tyddynnod and seasonal hafodydd, although few of the former have survived. The evidence points to an early pattern of clustered but by no means nucleated settlements, with simple tyddynnod scattered between or around the arable strips of the more intensively cultivated patches of infield. The dwellings were probably similar to surviving long huts. Beyond, isolated or clustered long hut groups represented either permanent (in some cases) or seasonal settlement, with most of the upland common being used as stock pasture although parts could be treated as outfield with intermittent cropping should the need arise or should economic or climatic factors encourage such ventures. This sys338 tern fell apart after the Edwardian conquest in the late thirteenth century. Some areas of native farming became abandoned but the more promising areas were acquired by both English and Welsh owners. These holdings were at first widely scattered as the more enterprising, such as Bartholomew Bolde, purchased any land that became available, eventually holding sufficient to reorganise their estates. Not all the land became part of such estates but several families in Caerhun and Llanbedrycennin acquired enough to dominate the landholding pattern in these parishes. Even when such large estate owners were not present, scattered native holdings gave way to individual farms by the fifteenth and sixteenth centuries, their lands increasingly amalgamated into composite holdings. The establishment of post-medieval farms Beyond the survival of names, there is little documentation in Llanaber to link the native holdings with subsequent farms but in the Conwy valley the rich documentation of several estate groupings allows a clearer picture to be built up. Many names can be recognised in documents like the Bolde rental of c. 1450 (UWB, Bangor 1939): the native holdings ofL!wydfaen appear regularly and Bryn y Castell in Gronant is named. There are several entries to names which are later known to be farms such as 'a place called Mayn y barth' (Maen-y-bardd), 'a place called Garthmore'. Although this seems to imply the existence of a dwelling it is rarely possible to confirm this: an actual house at Maen-y-bardd is only referred to in 1546 (UWB, Bangor 1920), its ruins still standing among today's farm buildings (RCHMW 1956, 26, Site No 96). Bronygadair is recorded in 1448/9 but only referred to as a tyddyn in 1600 (UWB, Baron Hill 2324, 2535). Surprisingly one intake in the upper part of the Afon Dulyn valley at 367 metres is also recorded at about this time: Maeneira, which was clearly a tenement in 1468 (UWB, Baron Hill 2654). Summer dwellings with hafod or llety names also begin to be recorded at this date such as the six acres of land called Kay llette Tudor noted in the Bolde rental and later said to lie in Bryn Gwenith (UWB, Baron Hill 2683), hill pastures in the south of Caerhun parish. It was suggested in my 197 5 paper that many of the cae names may have originated in the fifteenth and sixteenth centuries as new farms were being established and Caeithel in Llanbedrycennin is recorded as early as 1438 (UWB, Baron Hill 2290) when it was already divided into two tenements. Cae-y-pin appears in the records in 1451 with a new house built in 1555 (UWB, Baron Hill The effect of English settlementin medieval North Wales Commons enclosed C19 LB Llanbedrycennin village Hafodydd Enclosed ffriddoedd BGd Bronygadair Hg Hafod-y-gwyn Tr Tyddynrobin HyC Hafod-y-clawdd Bg Bwlch-y-gaer He Hafod-y-cae Limits of C16 enclosure Te Tyddyneithiniog H Hafod Field boundaries C 19 Ci Caeithel Hgw Hafodygorswen 2/3 Wall types Hafod holdings ~ ~ 11111 [J • Tb Tan-y-bwlch Underlying walls s Soglog Approximate boundary between stone walling and hedges Fb Ffridd-y-bont R Rowlynuchaf Former mill Pandyuchaf Lo Llwyn-on Occupied settlement Cl Cae'r·llin Abandoned settlement Long hut site Fig. 4. - Field patterns in Caerhun and Llanbedrycennin. 339 D. Hooke 2335, 2423) and Caemalach in 1464 (UWB, Baron Hill2358). There are many named tyddynnod which might be further located by detailed local study but most have not become modem farms. It is easier to locate those documented from the sixteenth century, although they may well have existed earlier: Tyddynrobinisafis referred to as 'a place' in 1501 (UWB, Baron Hill 2262), probably again implying the existence of a house or tyddyn. This lies on the upper edge of the valley land where the land is rising steeply to the upland plateau. On the plateau edge, Tyddyneithiniog occupies a heart-shaped shelf ofland below the hillfort of Pen y gaer in an area previously occupied by clusters of long huts. The long huts seem to represent an under-layer of settlement abandoned before the establishment of individual farms. The name ofTyddyneithiniog, recorded by the sixteenth-century, means 'house in the gorse', suggestive of such abandoned land. It appears in the documents in 1549 (UWB, Baron Hill 2417) and a ruined house site on the eastern edge of the holding may date from this period; it was to be replaced by another dwelling at a slightly higher level by the eighteenth century (Fig. 4). There are other references to farmsteads which lie at the outer edge of the cultivated land and which incorporated new intakes along the moorland fringe, representing this further wave of settlement encroachment onto the upland commons. A place called Tyddynrowlyn referred to one of the Rowlyn farms, located again on the upper edge of cultivation in the Dulyn valley, and was in existence by 1501 (UWB, Baron Hill 2662; Emery 1967) (Fig. 4). Higher up the Dulyn valley there were intakes in Gwern Jays y glog recorded c. 1600, which were to give rise to the farm ofSoglog (Hughes 1940, 3; mapped in Emery 1967, 148-8, fig. 7; see too Withers 1995, fig. 2; UWB, Baron Hill 2953). In Llanaber expansion onto the upland grazings was also occurring in this period; it is possible to identify several intakes, usually added to pre-existing farms, which it was claimed in 1575 had been illegal encroachments from the 'Forest of Snowdon' (Hooke 1975). In Caerhun and Llanbedrycennin several intakes formed islands within the upland commons as these hitherto common pastures were gradually eroded at this time. Deeds show estate owners buying, leasing and enclosing sections of the mountain pastures, usually referred to as the ffriddoedd: Fridd-ddu is documented in 1564 (UWB, Baron Hill 2683), taking in rocky land on the banks of the Afon Porthllwyd. Other intakes formed islands in the open upland grazings near Maeneira but most seem from their names to have started as seasonal hafodydd: Hafodygarreg is mentioned in 1544, its tyddyn in 1572-3 340 (UWB, Baron Hill 2413, 2953), and other seasonal upland holdings include Lletty yr Ddeufaen recorded in 1598 (UWB, Baron Hill 2530). Llettytudor recorded c. 1450 (noted above) was not certainly enclosed from the commons until after 1574/5 (UWB, Baron Hill 2685). To the north-east of Tyddyneithiniog an area of enclosed but steeply sloping rocky land still bore a hafod name in the eighteenth century (GA, Caerhun MS XM437 3) but was farmed from Cae'r llin, a holding beside the Afon Dulyn, itself not recorded before 1612 (UWB, Baron Hill 2556). Some of the hafodyddbecame permanently occupied at this time: farms were established, for instance, in the remote valley of Cwm Eigaiu. Tal-llyn Eigiau, along with other cottages close by, had originally been built as a dairy-house (a summer hafod) in the earlier part of the sixteenth century (UWB, Baron Hill2718; Davies 1979, 31), the land used for grazing stock which were brought down to the farms of Dolymarchog and Gronant in winter (UCW, Baron Hill 2637). Enclosures made around the margins of the upland commons from the sixteenth century onwards began to undermine the native tradition of the community use of upland pasture. Several hafod sites also date from this period but by this date are closely associated with specific lowland farms. Interestingly, fine specimans of sycamore, a tree widely introduced into Britain in the sixteenth century, and often planted to provide shelter, are found close beside several of these hafodydd. The associated paddocks often form islands within the commons (Fig. 4). In Caerhun encroachment was also beginning to take place on the lowland common of Allt Wyllt (Hooke, in preparation a) although most of the cottages which were built on this rocky cliff are likely to be much later in date. This tiny common supported a substantial community oflandless labourers and miners at the end of the nineteenth century but is now practically deserted. The tradition of constructing what were known as ty unnos houses persisted long in upland Wales although it had no recognised legal basis. It was thought that if a cabin could be erected on the waste during the night, with smoke issuing from its chimney by dawn, then its owner might legitimately claim the site as his own (RCL WM 1896, 576, No. 490). By this time, some substantial farms had developed in the valley itself: Farchwel is documented from the early sixteenth century and some of the fabric of the present house dates from the middle of that century; the present house of Maesycastell contains some material from the original house built in 15 82 and these reflect the growing prosperity of some local estate owners (Captain Edward Williams of Maesycastell was Sheriff of Caemarfonshire in 1570) The effect of English settlement in medieval North Wales (RCHMW 23-4, Nos 92 and 93). A large proportion of the parish, however, formed part of three large estates: the Baron Hill, Vaynol and Davies-Griffith estates. The former belonged to the Bulkeleys, a Beaumaris family, originally from Cheshire, who had their family seat in Lincolnshire but managed their estates in the seventeenth century through a steward based in Conwy. The core of their lands in the Conwy valley had been acquired by the marriage of a son to the daughter and heiress of Bartholomew Bolde in 1448; they had added to this into the seventeenth century (Williams 1979). Vaynollay near Bangor and had belonged to the Williams family (originally of Cochwillan) but had passed to the Smiths at the end of the seventeenth century, Thomas Assheton Smith owning these estates in the mid nineteenth century. The Davies-Griffith family of Caerhun were, however, based locally. The lands of both estates were subdivided into numerous small farms and it is the records ofland transfers etc which supply so much of the historical detail. Expansion onto the upland commons continued and the date at which many farms were established remains undocumented. Bwlch-y-gaer in Llanbedrycennin and Fridd-y-bont in Caerhun, for instance, appear as islands of enclosed land surrounded by land that was open until after 1858. Tan-y-bwlch was only enclosed after parliamentary enclosure. The enclosure movement swept across England and Wales in the eighteenth and nineteenth century and in this part of Wales mainly concerned the upland commons. In Caerhun and Llanbedrycennin, 6,242 acres (2,526 hectares), mostly upland common, was affected by this act, in each case over a quarter of the total parish area. Such enclosure was fiercely resisted and although an Enclosure order was made in 1850, it was not carried through until 1858. Fences were pulled down as fast as they were erected, ricks were burnt and extra police had to be brought in but it was ten years before resentment simmered down (Dodd 1990, 240; Aris 1987). By enclosures on this scale the last vestiges of native farming were thereby eradicated; two landowners - those of the estates noted above: the Bulkeley family of Beaumaris and the Davies-Griffith family of Caerhun - each obtained over ten per cent of the land enclosed in both parishes (Chapman 1992, 42, 45). Field shapes and boundaries Unfortunately the dates at which names are first recorded do no not necessarily indicate the date of a settlement and can only provide a terminus post quem date in assisting landscape reconstruction. Another type of evidence which is being investigated here and elsewhere is field boundary evidence. Since most of the present field pattern reflects the changed farming system which replaced the earlier native pattern this is of considerable importance if subsequent adaptations are to be understood. A glance at the modem map shows how small irregular fields still characterise much of the valley land although larger fields are found on the flood plain itself. Some of these represent the amalgamation of older fields, especially around major farms and estate centres, but much of the low-lying meadowland was subject to flooding and was only drained and enclosed over the last few centuries. Along the lower slopes between the 15m and 180m contours, where many of the small fifteenth and sixteenth-century farms developed, fields continue to be irregular and estate surveys show that most of them have remained unchanged in shape and areas since the seventeenth and eighteenth centuries (Fig. 4). Jones Pierce notes how hedging is first mentioned in the Baron Hill leases of the sixteenth century (Jones Pierce 1972, 209, n. 29) and living hedges form the main type of boundary used on the valley floor. This probably reflects an absence of stone on the flood plain but also confirms that the land had been cleared of readily available stone as far as the present margin between hedge and wall. Detailed examination of the area in which both types of boundary overlap might cast more light upon their sequence of use because ring fences of farms and their internal boundaries can be ascertained from the surviving cartographic evidence. On occasions, undergrowth has grown up alongside or even over stone walling. Sample hedge counts carried out in Llanbedrycennin gave a consistent count of four species being present in most hedges forming field boundaries. This, if reliable, would confirm a sixteenth-century origin (Hooper 1970). A close chronological sequence for the stone walling has not yet been satisfactorily established but broad chronological trends are clear. The commonest type of stone walling throughout the area of the farmed zone, and especially around the hedgerow border zone, consists of random stone construction using stones of variable size - in other words, walls which were constructed utilising any stone that was readily at hand (Type 2). The walls are often sinuous and look relatively precarious but they have obviously functioned effectively for a very long time. These may well represent the walls associated with the tenanted small farms of the sixteenth and seventeenth centuries, constructed and maintained by the tenants themselves; they are also found around some upper hafod sites. Another common type of wall 341 D. Hooke (Type 3) represents a further stage in the rebuilding of walls to produce a less random style, often with a horizontally laid top layer. Larger through-stones have been used to give a firmer construction. These walls probably represent the improved techniques being carried out on well-managed estates after the seventeenth century. Field survey of boundary walling may help to investigate the complicated pattern of development that can be seen on some holdings and thereby help to identify the trends of settlement and land use change across wider areas. Study of walling has so far only been incidental to wider research in the Conwy study but it is clear that several stages of development may be represented on the holding of Tyydyneithiniog (Fig. 4). As noted above, this tenement is first recorded by name in the sixteenth century but its name, 'house amongst the gorse', suggests that it was established in an area of previous clearance that had been at least temporarily abandoned. The long hut sites may stem from this earlier period of occupation together with, perhaps, some of the remnant walls, but a house site near the present !eat probably represents the sixteenth-century farm; at least one standing wall (Type 2) appears to date from this period. However, much rebuilding of walls was carried out later in association with a new house erected to the north of the old site; both are mapped in eighteenth-century surveys (GA, Vaynol 4056). These reveal a pattern of medium-large fields separated by straight walls (several Type 3 walls are indicated on Fig. 4), apart from the small enclosures around the farmstead (most abandoned by 1832- GA, Vaynol4071). The holding formed part of the Davies-Griffith Caerhun estate. Most of the newly-built walls survive but the plundered walls of earlier layouts can still be seen, together with an area of undated ridge and furrow. To the east of the hillfort of Pen-y-gaer, Type 3 walls also enclosed the rough land or ffridd of Ochr Gaer and separated this from a similar area of land which had become attached to the holding of Cae'r llin in the Dulyn valley. This had earlier been associated with a hafod in the adjacent field, also part of the Caerhun estate (1774: GA XM 437 3). On Rowlyn, too, an earlier field layout can be detected in distant views but the a new farm and field layout was made in the nineteenth century. Although any earlier walls would have been readily plundered for stone the bottom courses of earlier, probably medieval, walls survive (Type 1) in association with several long hut groups. They are rarely now above one course in height- robbed for the construction of sheep folds as well as walls - but the lower stones were often too large or too buried to be easily removed. Such walls survive best in Llanaber 342 in association with the long huts and can still be traced over considerable distances. They form small irregular paddocks on the holding of Egryn where they are associated with the long hut complexes which occur on the hillsides at heights of 180 and 214 metres. In this parish the sequence of walling is remarkably similar to that noted in the parishes of Caerhun and Llanbedrycennin despite the different types of rocks and stones available (Hooke, in preparation b). One type of boundary walling can be easily detected- that associated with nineteenth-century parliamentary enclosure (Type 4). Not only are the intakes on the hills large, straight-sided and geometric in shape but the walling is particularly strong and of a consistent style: through-stones are consistently used, the walls are high and stable, and frequently have a top of vertical or angled stones. In the Conwy parishes, a variant utilised short and long stones to produce a castellated appearance, a style restricted in Llanaber to a late (nineteenth-century) holding. Experienced gangs were often employed by large estate owners at this time, helping to produce consistency across a wide area. Although, on Fig. 4, the farms of Ffridd-y-bont and Hafodygorswen appear as enclosed islands within the commons, Tan-y-bwlch was a new farm established after enclosure. Exceptions to the above styles are always of interest, revealing local fashion, idiosyncrasy or adaptation to a special kind of terrain, and several instances of localised styles have been found in the two Conwy parishes. Near the village of Llanbedrycennin, slate, possibly old roof tiles, has been incorporated into the walls (there were slate workings at the head of the Dulyn valley in the nineteenth century). An exceptional area in the lower zone is the small area of common of Allt Wyllt which is characterised by stone walling: this is the steep rocky cliff which was encroached and settled from the sixteenth century. Conclusions Today, in Caerhun and Llanbedrycennin, all the upland farms have been abandoned as self-contained units and their dwelling houses are falling into ruin. There is no permanent occupation in the Dulyn valley above Rowlynuchaf and the land is grazed, chiefly by sheep, belonging to the valley farms. Once again, the remaining open uplands serve only as pasture for sheep and cattle, but although the appearance of much of the higher land may have changed little from medieval times, farming systems and economic arrangements are very different. In Llanaber, too, The effect of English settlement in medieval North Wales higher farms have in places given way to forestry. The landscape could be on the brink of further change as the lowland villages expand and change the traditional pattern of dispersed settlement to one closer to the English model. In this case the change is less a response to methods of land management than to the preferred requirements of planners, administrators, those providing emergency services and, it must be said, of the people themselves. It is to be hoped that the essential 'Welshness' of the uplands will not be eroded yet again. Acknowledgements I should like to thank the staff of the Dolgellau and Caernarfon Record Offices, and especially Gareth Haulfryn Williams, for their continued help and support over many years; Tomos Roberts for providing access to the Bangor archives and to Melville Richard's invaluable place-name index; Nick Eyles for his preliminary survey of the field boundary walling in the Conwy valley; and Kathryn Sharp for her cartographic assistance. In particular, I should like to thank Cheltenham and Gloucester College for the funding that permitted some of this research to by undertaken. Abbreviations Gwynedd Archives, County Record Office, Caernarfon. Public Record Office, London. PRO RCLWM Royal Commission on Land in Wales and Monmouthshire (see Bibliography). RCHMW Royal Commission on Ancient and Historical Monuments in Wales and Monmouthshire; SMR Sites & Monuments Record. ucw University ofWales, Bangor, Collection of manuscripts. GA Bibliography ARIS M. 1987: Crime and Punishment: a Welsh Perspective, Gwynedd Archives Service, Caernarfon. ARIS M. 1996: Historic Landscapes of the Great Orme, Llanrwst. CASEY P.J. 1969: Caerhun, in: V.E. NASH-WILLIAMS (ed.), The Roman Frontier in Wales, 2nd edn revised M.G. Jarret, Cardiff, 56-59. CHAPMAN J. 1992: A Guide to Parliamentary Enclosures in Wales, Cardiff. DA VIES E. 1979: Hendred and hafod in Caernarvonshire, Transactions of the Caernarvonshire Historical Society 40, 17-46. DAVIES T. 1988: Ystumgwern yn Ardudwy, Journal of the Merioneth Historical and Record Society 10, 205-220. Dooo A.H. 1990: A History of Caernarvonshire 1284-1900, (1968), 2nd edn., Wrexham, 1990. ELLIS H. 1838: Record of Caernarvon. EMERY F. 1967: The farming regions of Wales, in: J. THIRSK (ed.), The Agrarian History of England and Wales IV. 1500-1640, Cambridge, 113-160. Extent of Merioneth 1284, Archaeologia Cambrensis, 3rd ser. 13, 1867. GRESHAM C.A. 1965: The Bolde rental (Bangor MS. 1939), Transactions of the Caernarvonshire Historical Society 26, 31-49. GRESHAM C.A. 1979: The commotal centre of Arllechwedd Isaf, Transactions of the Caernarvonshire Historical Society 40, 11-16. GRESHAM C.A. 1987: Medieval parish and township boundaries in Gwynedd, Bulletin of the Board of Celtic Studies 34, 137-149. GRESHAM C.A. 1988: Addendum, 'Vairdre alias Vaildre', Journal of the Merioneth Historical and Record Society 10, 221-6. HooKED. 1975: Llanaber; a study in landscape development, Journal of the Merioneth Historical and Record Society 7, 221-230. HooKED. 1983: The Ardudwy landscape, Journal of the Merioneth Historical and Record Society 9, 245-260. HOOKED. 1997 forthcoming: Place-names and vegetation history as a key to understanding settlement in the Conwy valley, in: N. EDWARDS (ed.), Landscape and Settlement in Medieval Wales, Oxbow Books, Oxford. HOOKED., in preparation (a): The settlement on Allt Wyllt, Caerhun. HooKE D. in preparation (b): A chronology of boundary walling. HOOPER M. 1970: Dating hedges, Area 4, 63-65. HUGHES R.E. 1940: Environment and human settlement in the commote of Arllechwedd Isaf, Transactions of the Caernarvonshire Historical Society 2, 1-25. JOHNSON N. & ROSE P. 1994: Bodmin Moor, an Archaeological Survey, Vol1: The human landscape to c 1800, English Heritage, London. JONES G.R.J. 1964: The distribution of bond settlements in north-west Wales, Welsh History Review 2, 19-36. JoNES G.R.J. 1972: Post-Roman Wales, in: H.P.R. FINBERG (ed.), The Agrarian History of England and Wales !If. A.D. 43-1042, Cambridge, 283-382. JONES PIERCE T. 1939: Some tendencies in the agrarian history of Caernarvonshire during the later Middle Ages, Transactions of the Caernarvon343 D. Hooke shire Historical Society 1, 18-36, repr in: J. BEVERLEY SMITH (ed.), Medieval Welsh Society, Selected Esssays by T Jones Pierce, Cardiff 1972, 39-60. JONES-PIERCE T. 1942: The gafael in Bangor Manuscript 1939, Transactions of the Honourable Society of Cymmrodorium, 158-188, repr. in: J. BEVERLEY SMITH (ed.), Medieval Welsh Society, Selected Esssays by T Jones Pierce, Cardiff, 1972, 195-228. MussoN C. n.d.: Wales from the Air, Patterns ofPast and Present, RCHMW, Aberystwyth. RCLWM 1896: Royal Commission on land in Wales and Monmouthshire, Report, London. RCHMW 1956: Royal Commission on Ancient and Historical Monuments in Wales and Monmouthshire, An Inventory of the Ancient Monuments in Caernarvonshire, Vol. I: East, the cantref of Arllechwedd and the commote ofCreuddyn. RCHMW 1964: Royal Commission on Ancient and Historical Monuments in Wales and Monmouthshire, An Inventory of the Ancient Monuments in Caernarvonshire, Vol. Ill: West, the cantref of Lleyn. SILVESTER R. 1991: Medieval farming on the Berwyn, North Wales, Medieval Settlement Research Group Annual Report 6, 12-14. SMITH P. 1975: Houses of the Welsh Countryside, London. THOMAS C. 1970: Social organisation and rural settlement in medieval North Wales, Journal of the Merioneth Historical and Record Society 6, 121131. THOMAS C. 1980: Field-name evidence in the reconstruction of medieval settlement nuclei in North Wales, National Library of Wales Journal 21, 340-356. THOMAS C. 1992: A cultural-biological model of agrarian colonisation in upland Wales, Landscape History 14, 37-50. WALKER M.F. & Taylor. J.A. 1976: Post-Neolithic vegetation changes in the western Rhinogau, Gwynedd, north-west Wales, Transactions of the Institute of British Geographers, new ser 1, 323345. WILLIAMS G .H. 1979: Estate management in Dyffryn Conwy, c. 1685: the Caerhun, Baron Hill and Gwydir estates, Transactions of the Honourable Society ofCymmrodorion, 31-74. WITHERS C.W.J. 1995: Conceptions of cultural landscape change in upland North Wales: a case study of Llanbedr-y-cennin and Caerhun parishes, c. 1560-c.1891, Landscape History 17, 35-47. Dr Della Hooke Dept Geography & Geology Cheltenham & Gloucester College Francis Close Hall Swindon Road Cheltenham GL50 4AZ UK 344 Rural Settlements in Medieval Europe- Papers of the 'Medieval Europe Brugge 1997' Conference- Volume 6 P.J. Dixon Settlement in the the Hunting Forests of Southern Scotland in the Medieval and Later Periods Introduction Recent archaeological field survey in the lowlands of southern Scotland by the Royal Commission on the Ancient and Historical Monuemnts of Scotland (RCAHMS) has identified extensive areas of abandoned medieval and post-medieval settlement. These settlements are littered across the rough pastures, taking the form of buildings reduced to grasscovered banks, earthen-banked field-systems, and plough-rig. In some areas, the settlement remains are enclosed within earthen boundary-dykes, which comprise a bank and an external ditch, known as a "deerdyke". Two of the areas surveyed, Southdean and Liddesdale (Fig. 1) lie within the medieval hunting forests of Jedburgh and Liddesdale respectively. Here, this type of head-dyke is attributed to the strictures of Forest Law that were applied in the hunting forests of Scotland from the mid-12th century until the end of the medieval period (c.1600), and should be distinguished from the earthen head-dykes that surround so many post-medieval rural settlements elsewhere in Scotland. largely confined to southern Scotland and included both royal demesne, i.e., the forest of Ettrick, and forest grants to his Norman barons, such as that of the Bruce lordship of Annandale. Indeed, many of the new baronial forests were granted to Norman followers of King David who had been feudal dependents of the earldom ofHuntingdon, where they could not hunt without the permission of King Henry. In contrast with England, this was a beneficial arrangement, whereby the king instituted his control of hunting reserves by granting a subject the right to create a private forest. Although this reserved the hunting to the baron, it did not preclude the king from hunting there, for example, Robert II hunted in the Earl of The establishment of Hunting Forests in Scotland: Historical Introduction The concept of the hunting forest was introduced to Scotland by King David during the 1130s, thus altering and cutting across the traditional view of hunting and access to game. Up till that time the law in Scotland with respect to game depended on the principle of Roman Law known as 'res nullius ', which stated that game belonged to the person who killed it. Whilst this was limited by the rights of the possessor of the lad, the creation of hunting reserves was an outright denial of this traditional prerogative. However, the creation of royal forests was limited to the royal demesne and did not include baronial holdings, as it did in England, so that the rights of the barons as possessors of land were not impugned. In the first instance the new forests created by King David were SCOTLAND Fig. 1. - Location Map. 345 P.J. Dixon Mar's forest of Mar in the 14th century (Gilbert 1979). Scots forest law was, in general, more humane than the English system, and appears to have been tempered under the influence of native custom of 'res nullius'. As in England, however, deer hunting, in particular, was reserved, and actions which affected the livelihood of the deer such as woodland clearance, grazing, muir-burn, enclosure and settlement, were not permitted. However, the penalties for poaching not so harsh, for example, the maximum fines were not usually exacted until the 4th offence, and the mutilation of dogs and the banning of hunting weapons were not enforced in Scotland. On the other hand there were fines for anyone travelling through a forest without permission or deviating from the designated route. Hunting Forests and settlement Until recently, there has been little archaeological interest in the remains of medieval settlement in the hunting forests of Scotland. The model for settlement in the forest was established by the RCAHMS in their work on Selkirkshire (1957). In this Inventory, peel towers were placed in the context of a system of "forest steads", which the Crown inherited from the Douglas lordship of the Forest ofEttrick in 1455. The concentration on the ruins of these defensible structures to the relative exclusion of the earthworks of the rest of the rural settlement landscape has created a bias in our understanding of the Forest of Ettrick towards its defensible structures of later 15th and 16th century date. Thus the parks at Newark Castle and Aikwood, mapped by Pont (Blaeu 1654), are not located or recorded, and only a handful of the other settlements, such as Douglas Craig or Over Kirkhope, were deemed worthy of record (RCAHMS 1957). The symbiosis of park and hunting-lodge or castle (e.g. Newark Castle) was not examined, yet this is a common aspect of the medieval estate-centre, since hunting was such an important facet of the life of the king and his barons. Thus the royal castles of Stirling had two nearby deer parks by the late 13th century, and Kincardine and Jedburgh one each, whilst baronial parks proliferated in the 14th and 15th centuries (Gilbert 1979). However, not all steads sported peel towers, which are in any case predominantly 16th century in date. The antiquity of the system of forest steads in Ettrick Forest and the pattern of settlement indicated by it, is difficult to establish, either by documentary evidence, in the absence of any suitable documentation of the forest for the period of Douglas ownership prior to 1455, or on archaeological grounds, without fur- 346 ther fieldwork. What may be said is that the steads were in part a military system in that the holders of each stead had to provide the royal army in time of war with two bows, a spear, a horse and gear, in addition to the usual annual payment of rent in kind. The system is most likely to have its origins in the later 14th or early 15th centuries as a reaction to the occupation of southern Scotland by the English, but it is not clear that it had any effect on the pre-war pattern of settlement or what form this may have taken. In the conditions of warfare and pestilence that obtained in southern Scotland in the 14th and 15th centuries, the crown and other landlords were obliged to find ways of maintaining their tenantry in their farms. With a dislocated economy and a declining population, unfree serfs were replaced by husbandmen, demesnes were leased, and rents fell or were held static. There was also a preference for the substantial husbandman, who would be better able to support military service. In Ettrick, as elsewhere in southern Scotland, the tenants of the forest steads, apart from the forest officials, held their lands at the will of the lord. Their lease was by a tack of three or five years, with annual payments of money, cattle and lambs in the later 15th century with no right to sub-let. James IV converted the tacks to feu-farms in 1506, giving the tenants permanent tenures and the right to alienate the farm, whilst retaining the feu payment. Whilst this did little to raise the income of the crown, it provided the stability that enabled some of the tenants to build peel towers. However, this process of alienation of forest lands did not apply to forests that were not in royal hands, indeed the evidence is to the contrary. In the Forest of Jedburgh in 1541, although the tenants held their steads on an annual basis, they appear to have benefited from a "kindly tenure", which allowed the tenant to continue in possession at the end of the year in the interest of stability. This was important if the steads were to provide military service and may also have created the beneficial conditions that led to the construction of peel houses (RCAHMS 1994). In Liddesdale, another Douglas possession, which passed to a different branch of the family in the 15th century, the farms were held at will in 1376, but the length of tenure of the tenants-at-will was not defined. However, in 1541 the tenants paid by the year, presumably at the will of the lord, but a significant proportion of the steads or holdings were either vacant or worth nothing (RHM 1853). This suggests that the landowner was not very effective in maintaining the tenantry in occupation. Whilst there may have been other factors at play, the small size of many of the holdings of 1376 may have limited their viability. Settlement in the the Hunting Forests of Southern Scotland in the Medieval and Later Periods ~ ~ 079 -t +. +. A -+.- +. +. 075 N ~ + ~._:_ .+.. ;t, -----·~ .t. -t .+.. ;t, .+.. +. . ~: : : /' ' ;t, +. 100 --_ 200 · 300m /l ~It-'t-1t-1w1w1w'...J1....!1...J'...J'------'-'------'' / '/ .+.. + +. +. ... .+.. .t. .:t--·----- .+.. ~ft .+.. +. // NT 652 653 +. 654 • 655 i/ 656 ~ 657 658 Fig. 2. -Ass arts at Martinlee Sike, Southdean in the forest of Jedburgh. Ass arts Settlement within the forest was limited or controlled because it was at variance with the maintenance of the habitat of the deer. Thus any clearance of woodland within the forest for farming was only permitted under licence from the forester or the landowner. The clearance of land in this way was called an assart (derived from the old French essarter, 'to grub up trees'). In Scotland the majority of assarts are recorded in charters of the mid-12th to the early-14th century. Although there are relatively few documented assarts in Scotland, this does not mean that there are few assarts, since most holdings were held by unfree tenures until the 14th century and thereafter at the will of the lord. The assarts of this type of tenant would be recorded not by charter but by a licence recorded in the administrative 'rolls' of the forest, none of which survive. In England there are forest rolls with this type of record for the Royal Forests ofRockingham and Pamber (Stamper 1983). Thus in an estate like Liddesdale, where in 13 76 only twelve holdings out of 173 were free tenures or church lands, the assarts of any tenants-at-will, who formed by far the majority, would not be recorded by charter but in the forest rolls. After the early 14th century charters of assarts cease, either because there was no pressure on land, which might be true in southern 347 P.J. Dixon Scotland but is demonstrably not the case in other parts of Scotland, e.g. Perthshire (Gilbert 1979 and RCAHMS 1990), or because they were administered in other ways such as purprestures, a less specific term than assart, used as a method of fining those make unlicenced encroachments in the forest, but in the absence of forest court rolls they are hardly likely to be well recorded in Scotland. Some idea of the form of an assart may be gained from late 13th-century charter ofRobert de Brus, lord of Annandale, to William de Carlyle in which the grantee was given the right to increase his lands of Kinmount by taking a portion of the adjacent common of the tenement ofNewby. The grant specified that William was empowered to enclose by "hedges and ditches, to cultivate, build, make meadow and arable" (Fraser 1894, I, n°. 11). Unfortunately there are no surviving physical remains of this assart or of any other documented assart in Scotland. The form which the forest clearance or assart might take on the ground was described in an analysis of the field-walls of a forest edge settlement at Holne Moor on Dartmoor. It argued that there were particular characteristics associated with the enclosures of this type of settlement. A type of field-wall with an assymetrical profile, known locally as a 'corn-ditch', was identified that comprised an externally-reverted bank and external ditch. Such a fieldwall was ideally designed to prevent the ingress, but not egress, of deer. The 'corn-ditch' formed an early stage in the development of the so-called 'lobes' or enclosures that surrounded the fields of the settlement. This phase was attributed to the period from the late 11th century to 123 9 when Dartmoor was part of the English king's Royal Forest (Ralph & Fleming 1982). In Scotland assarting has been recorded in recent surveys carried out in three of the forests of southern Scotland, Annandale, Southdean and Liddesdale, and in all three areas the Dartmoor type of boundary-dyke has been identified. Of these forest areas, Annandale has been so heavily improved by modem farming that there are few surviving medieval or post-medieval settlement landscapes and, apart from one possible assart at Cowbum on the Corrie Water, there is little field evidence left. The Corrie Water example comprises a D-shaped enclosure of about 60ha bounded by an earthen bank with an external ditch for most of its course with an adjacent intake of 18ha to the south. Another example at Carterton, near Cowbum which has been much reduced by modem ploughing, comprises aD-shaped enclosure of about 52ha with the chord of the Don a bum (RCAHMS 1997). The main body of data from southern Scotland comes from Southdean and Liddesdale (Fig. 1). Here, 348 there are extensive settlement landscapes of the medieval and post-medieval periods. Field-analysis of which has identified a characteristic landscape of assart banks. The assart banks are built of earth and stone, occasionally with a visible external revetment (for example at Greenshiels) and an external ditch. It is probable that a hedge surmounted the earthen bank, as referred to in the assart of William de Carlyle referred to above, but there is no archaeological evidence for this at present. This type of boundary is known in Scotland as a deer-dyke, a type of dyke which also surrounds parks such as Kincardine, but with the ditch on the inside instead of on the outside. In other words the greatest barrier is on the side of the ditch where the combined height of ditch and bank form a boundary of sufficient height to deter a leeping deer, whilst an approach from the other side presents a relatively easy exit. This type of bank typifies the head-dykes of both Southdean and Liddesdale, where they enclose blocks of ground of variable extent, the banks running in a curvilinear fashion from bum to bum. Thus they may define roughly D-shaped areas, the chord of the D being usually defined by a river, and may build one on another as further land is taken in. The Southdean Assarts Southdean parish lies at the head-waters of the Jed Water, the far extremity of the Royal Forest of Jedburgh, and marches with border of England to the south. A small strip of ground (c.25 km 2) between the Jed Water River and the conifer plantations that cover much of the parish was surveyed. In this small area six settlements with peel houses or towers were identified, all of which were occupied in the 16th century, and abandoned by the late 18th century. In addition an extensive system of assart-banks was recognised, which encompassed not only the peel tower settlements and their cultivated lands, but also some undocumented farmsteads. Since no vacant steads are recorded in the forest in the Crown rental of 1541, and Pont (Blaeu 1654) does not depict them, it is a reasonable inference that these farmsteads were abandoned before the forest steads listed in 1541 were established. The best-preserved of the assarts containing farmsteads in Southdean (Martinlee Sike) comprised a D-shaped enclosure of about 7ha in its primary form with an extension abutting it to the north-west of 14ha, of which only a small part was cultivated (Fig 2). The primary enclosure contains two farmsteads, each comprising a large and a small building and an enclosure, and an expanse of plough rig In Settlement in the the Hunting Forests of Southern Scotland in the Medieval and Later Periods Fig. 3. - Assarts on Sufficient Hill, Kirk Hill and Park Hill near Newcastleton in the forest ofLiddesdale. 349 P.J. Dixon contrast the assart that contains the peel tower, called North bank Tower, encompasses some 30ha, much of it covered with rig (RCAHMS 1994). However, there is not enough evidence to make any certain equation of the size of assart with the type of settlement enclosed. In addition to the expansion of settlement and ass arts, there was a phase of contraction. This is partly illustrated by the complete abandonement of the assart at Martinlee Sike, but also by the Northbank assart, which expanded in two stages up the Jordan Sike, each assart containing a farmstead, and enclosing about 8km 2 and 12km2 respectively, but both were subsequently abandoned as a new assart-bank was established just above the original boundary. Liddesdale: the documentary history of settlement Liddesdale lies adjacent to the English border and comprises the valley of the Liddel River. Prior to the Wars of Independence, it was held by the de Soulis family and run as a private forest, but after Bannockbum it was granted by Robert Bruce to the Black Douglas with forest rights. A rental of 1376 (RHM 1853) indicates that by this date, the greater part of the valley had been deforested, since only a limited body ofland is listed under the title "foresta ",much of it disposed on the north and west extremities of the valley, and including what must be the grazing of the parks of Hermitage and Castletoun. By 1541 when the estate was temporarily in Crown hands, the distinction between the forest and the rest had been abandoned. Significantly, the value of the estate had declined by nearly two-thirds since 1376, one in four farms was vacant and another one in six was of no value (ERS 1897). The reasons for this degree of deterioration are not immediately evident, since the war with England did not start until 1542, but it may have been related to James Vs attempt to bring a measure of law and order to the valley. Be that as it may, this trend of a decline in the numbers of farms appears to have been accepted as reality by the late 17th century when the Hearth Tax returns 1691 indicate a reduction of more than 50% in the number of settlements from that of 1376, and, by 1718, when the estate was surveyed for the Duke ofBuccleuch, only fifty farms were mapped, apart from nine freeholds (SRO RHP). It is a reasonable inference that there was a drastic reorganisation and amalgamation of farms prior to the late 17th century, the most likely moment being at the union of the crowns (1603) when James VI turned out the freebooters ofLiddesdale, destroying their peel towers to bring peace to the border. 350 The archaeology of medieval settlement in Liddesdale In Liddesdale the greater part of the west side of the valley was surveyed by RCAHMS. Here, there is a series of interconnected assarts running over a distance of some 8 km from Sufficient Hill in the south to Hartsgarth in the north, with further stretches visible over a distance of about 5.5 km to the north as far as Hermitage Castle, the estate centre, where there is a distinctive assemblage of monuments including a chapel and a park pale. A range of settlements that is associated with these assarts. Where post-medieval farming has not removed any earlier material, there are townships or fermtouns comprising six to eight buildings, usually on the same alignment, with turf-walls or turf-walls over stone-footings. Associated yards and gardens are rare, which may make it difficult to distinguish farmsteads from shielings. At Greenshiels and Foulshiels (Fig. 4) most of the buildings are long byre-houses with a hollow for the drainage of cow slurry, running out of the lower end of the byre and, presumably, a domestic upper end. These byre-houses range from about lOm to 20m in length. Scattered amongst the assarts at Kirk Hill, and indeed elsewhere in Liddesdale, there are smaller building-clusters, the buildings of which are usually smaller than the byrehouses. The status of these sites remains obscure. Equally, there is the occasional single building, e.g., that near Ettleton Church, which may be a medieval farmstead. Some of these may account for the unlocated farms that are recorded in the 1376 and 1541 rentals ofLiddesdale. Peel towers are few in number in Liddesda1e; only two (Foulshiels and Puddingbum) were located in the area of the survey with any degree of certainty, and that at Foulshiels had been robbed to its foundations. The sites of two others, Redheugh and Copshaw could not be located with any confidence. For the size of area there are fewer peels than might be expected, particularly in comparison with the numbers in Southdean. On Kirk Hill (Fig. 3) there are at least three core areas of settlement. These are represented by roughly D-haped enclosures, defined by deer-dykes, with several complex expansion phases that led to the interconnection of the assarts at their maximum extent. It is difficult here to define the full extent of the assarts, because of the complexity of the system and the attrition of modem farming activities on the archaeology on the lower slopes of Kirk Hill. However., the partly afforested assart on Sufficient Hill encloses about 25ha, comparable with the larger assarts in Southdean. As in Southdean, expansion was followed by a phase of contraction as the assarts on the east Settlement in the the Hunting Forests of Southern Scotland in the Medieval and Later Periods Fig. 4. - Assarts at Foulshiels and Greenshiels in the forest ofLiddesdale. 351 P.J. Dixon flanks of Kirk Hill and Park Hill were abandoned. Although not visible on the plan, a subsequent assart post-dates the abandoned assart on the north-east side of Park Hill. In 1718 the estate plan (SRO RHP) shows that the east slopes of Kirk Hill and Park Hill were occupied by the farms of Side, Milneholme and Copshawpark. Of these only Side farmstead could be located in the survey, since the sites of other two, as indicated on plan, lay on the haughland, which has subsequently been more heavily improved. The farms are depicted as strips of ground running from the River Liddel up on to the high ground, completely superseding the D-shaped enclosures found in the pattern ofassarts. The farmstead at Side is surrounded by rectilinear fields, enclosed by earthen-banks that overlie the arable rig of the assarts. Similar postmedieval field-systems around the abandoned township on the east slope of Kirk Hill that also overlie the old rig-systems. However, in this case, the rectilinear fields, it is suggested, were constructed as outfields for Milnholm Farm. Recent afforestation and the construction of a sheepfold have obscured the the character of the farmstead at Side. In the vicinity of Foulshiels (Fig. 4), there are at least three assarts. These comprise a core area based on Nether Foulshiels Tower, and two main phases of expansion, Nether Foulshiels and Greenshiels. A small additional expansion of about 4ha was made to the Greenshiels assart on either side of the Ryedale Bum, with a small settlement visible on the east side of the bum. Some of the land within the assart banks of the two Foulshiels has been smoothed by recent agriculture. However, the course of the assart bank enclosing Greenshiels may still be traced in more or less its entirety, as it runs from bum to bum, illustrating the way in which the topography was used to enclose an assart. At all three main settlements of Greenshiels and the two Foulshiels, there are buildings that either post-date the assart banks or are placed outside, which may imply that forest restrictions had been relaxed or were not rigorously applied when they were built. The greater part of the system collapsed between the late 16th and the end of the 17th centuries. The settlement at Greenshiels, and the two Foulshiels settlements of Over and Nether Foulshiels, are evident as late as about 1590, when Pont mapped them (Blaeu 1654), but only Foulshiels appears in the Hearth Tax of 1691. This appears to give a span of about 100 years in which Greenshiels and Over Foulshiels came to be abandoned. 352 The estate centre at Hermitage Castle, Liddesdale Hermitage Castle was occupied from the mid 13th century to the 17th century. It stands about 300m east of Hermitage Chapel, itself of 13th century date, on the north side of the Hermitage Water. It is all but surrounded by boggy ground, providing good defensive qualities. Starting as an earthwork castle, all the ramparts, apart from that on the west, were levelled in the mid-14th century to build a stone tower-house which was expanded and remodelled during the later 14th and 15th centuries. As an earthwork castle its role is difficult to understand in view of the continued occupation ofLiddel Castle, the original estate centre. However, by 1376, if not before, the greater part of the forest lands lay in the north and west parts of the valley, so that it may have been a conscious choice to build a hunting-lodge at this location, close to the accessible hunting. Whatever the original reason, the construction of the stone castle at Hermitage in the 14th century indicates a preference for this situation over that of Liddel Castle, which was abandoned, probably because of its closeness to the border. The castle stands at the centre of two systems of deer-dykes. There is an assart that runs off to the east, enclosing the haugh lands along the river, another to the west that runs up to the homestead enclosure on the west side of the chapel garth, with an extension to the west. From the north side of this assart, another deer-dyke runs in a great arc to the north-west, defining a funnel-shaped area with the ditches on the inside. This appears to be a deer park, but open to the hill ground on the west, with a narrower end by the castle, where there is a gully cut by a bum that provides an excellent killing-ground for the deer that are driven in to the park. This type ofhunting, with a drive conducted by hundreds of men from the estate, was a traditional method in Scotland which was still being used in the forest of Mar in the early 17th century (RCAHMS 1995). Few medieval parks have been recorded in any detail, so that there is too small a body of data to know if this was a common arrangement. Traditionally the park pale enclosed a reserve, but this type of park enclosure suggests that some medieval hunting parks may have been designed with an opening. Kincardine Park has an extension to the north that encloses a steep sided valley with an opening to the east, which could have served as a killing ground, or as a funnel to drive deer into the main part of the park (Guilbert 1979). Be that as it may, Hermitage Park was not an enclosed space that that could be used as a reserve for deer, or latterly for keeping stock, although it no doubt had grazing value! With the abandonment of the castle in the 17th century, a farmstead was established on the east side, Settlement in the the Hunting Forests of Southern Scotland in the Medieval and Later Periods Fig. 5. -Hermitage Castle with Park Pale and Chapel in the forest of Liddesdale. 353 P.J. Dixon comprising three buildings arranged around a courtyard and an enclosure. The White Dyke and the long long straight boundary dyke that forms its west side appear to be the boundaries of the post-medieval farm. At its east side the White Dyke post-dates the assart bank near the present Hermitage Farm. Another farmstead, which may also belong to this period, is the farmstead that stands on the south side of the perhistoric enclosure at Lady's Knowe. There are earthen-banked rectilinear fields in the vicinity of both of them, and there is another larger group of these post-medieval fields on Tofts Knowes to the south of the castle, possibly worked from the farmstead situated on their west side. In turn, all these settlements were abandoned in turn during the 18th century. Conclusions For the first time in Scotland, the archaeological remains of medieval assarting, the assart bank, have been recognised and recorded. To date they appear to be a feature of southern Scotland, there being little or no sign of them in any recent work in the hunting forests of the Highlands, e.g. the forests of Alyth, Clunie and Mar (e.g. RCAHMS 1990 and 1995). However, post-medieval phases of expansion in settlement may have obliterated any traces of them in these areas, or there may be other factors at work. The currency of the assart bank appears to be firmly medieval, but the exact date of its demise may vary from forest to forest, with the end of the forest administration in the post medieval period. The assart banks display a considerable complexity of development, that is to say, phases of expansion and contraction, as well as complete abandonment, which compliments the documentary evidence for a con- traction of settlement in the late medieval period, but also providing the only evidence of an expansion of settlement in the 12th to 14th centuries. Finally the landscape of the medieval estate centre at Hermitage indicates that the medieval park in Scotland may have had more variation in its design, according to the type ofhunting carried out, than hitherto appreciated. Bibliography BLAEU 1654: Atlas Novus, Amsterdam. ERS 1897: The Exchequer Rolls ofScotland, Vol17, 1897, Edinburgh. FLEMING A. & RALPH N. 1982: Medieval Settlement and Land Use on Holne Moor, Dartmoor: The Landscape evidence, Medieval Archaeology 26, 101-137. FRASER Sir W. 1894: Annandale family book of the Johnstones, Earls and Marquises of Annandale, Edinburgh, 2 vols. GILBERT 1979: Hunting and Hunting Reserves in Medieval Scotland, Edinburgh. RCAHMS 1957: An Inventory of the Ancient and Historical Monuments ofSelkirkshire, Edinburgh. RCAHMS 1994: Southdean, Borders: An Archaeological Survey, Edinburgh. RCAHMS 1995: Mar Lodge Estate, Grampian: An Archaeological Survey, Edinburgh. RCAHMS 1997 forthcoming: Eastern Dumfriesshire, Edinburgh. RHM 1853: Registrum Honoris de Morton, Bannatyne Club, Edinburgh. STAMPER P.A., 1983 The medieval forest ofPamber, Hampshire, Landscape History 5, 41-52. SRO RHP: The Scottish Record Office, Register House Plans. Dr. Piers Dixon Royal Commission on the Ancient and Historical Monuments of Scotland John Sinclair House, 16 Bernard Terrace Edinburgh EH8 9NX Scotland, UK 354 Inhoud Contents 06 Table de matiere In halt RURAL SETTLEMENTS - DE LANDELIJKE WERELD - LE MONDE RURAL LANDLICHE SIEDLUNGEN Preface 5-6 Regional settlement studies - Approches regionales - Regionale Siedlungsforschungen Mate) Ruttkay Early medieval settlement structure in Slovakia 7-15 Maria Beres Regionale Beziehungen der Dorfsiedlungen Si.idostungarns vom 10. bis 13. Jahrhundert 17-20 John L. Bintliff The archaeological investigation of deserted medieval and post-medieval villages in Greece 21-34 Giovanni Di Stefano Villagi tardo-bizantini degli Iblei: primo medioevo sicilliano 35-38 Andre Bazzana, John De Meulemeester & Andre Matthys Quelques aspects du peuplement medieval du Valle de Ricote (Murcie, Espagne) 39-54 Christopher Dyer Recent developments and future prospects in research into English medieval rural settlements 55-61 Mark Gardiner Trade, rural industry and the origin of villages: some evidence from South-East England 63-73 Carenza Lewis Medieval Settlement in Hampshire and the Isle of Wight 75-86 Terry Barry Recent research in Irish medieval rural nucleated settlement 87-92 Case studies -Etudes de cas - Einzeluntersuchungen Marco Milanese, Fabrizio Benente & Franco Campus Il progetto Geridu. Indagini archeologiche in un villagio medievale abbandonato della Sardegna 93-110 Maria Angels Ruf, Toni Vi/a, Xavier Sate, Xavier Llovera, Cristina Yafiez & Josep Maria Bosch Le site du Roe d'Enclar (Andorra): evolution des strategies d'occupation et exploitation d'un territoire du IVeme la fin du VIIIe siecle 111-119 Maruska Federici-Schenardi & Robert Fellner L'habitat rural du Haut Moyen Age de Develier-Courtetelle (Jura, Suisse) 121-130 Armelle Querrien L'exploitation du sol et du sous-sol autour d'un bourg castral du Berry, du Xle au XVe siecle 131-142 Isabelle Catteddu Le site medieval de Saleux "Les Coutures": habitat, necropole, et eglises du haut Moyen Age 143-148 a 355 Rene Proos Venray-'t Brukske, an early medieval settlement on the sandy soil ofLimburg 149-156 Dries Tys Landscape and settlement: the development of a medieval village along the Flemish coast 157-167 Marnix Pieters Raversijde: a late medieval fishermen's village along the Flemish coast (Belgium, Province ofWest-Flanders, Municipality ofOstend) 169-177 Christopher Loveluck Flixborough- the character and economy of a high status Middle Saxon settlement in northern England 179-194 M.A. Aston The Shapwick Project, Somerset, England 195-210 Neil Price The Gamla Uppsala Project: rescue and research in an early medieval ritual landscape 211-219 Sofia Andersson & Eva Svensson The local and regional arena of a medieval Swedish farm 221-231 Territorial organisation and central places- Organisation territoriale et sites centrauxGebietsverwaltung und Zentralorte Julian Richards Anglian and Viking settlement in the Yorkshire Wolds 233-242 Kaname Maekawa A deserted medieval village and the formation of a fortified town in Cambridgeshire, England 243-252 UdoRecker The medieval parish ofLohn. Preliminary report on a research project 253-262 Birgitta Berglund Changes in the power structure around AD 1100 on the northern Norwegian coast. The importance of waterways and of the organisation of trade in building and maintenance of power 263-270 Settlement and rural infrastructure- Habitat et infrastructure rurale- Siedlung und liindwirtschaftliche Rustung Julio Antonio Perez Celada Horticultura y molinos de agua en el curso medio del rio Carrion en la edad media 271-282 Antonio Fernandez Ugalde El almacenamiento subterraneo y la conquista feudal en la peninsula iberica: aportaciones de la arqueologia 283-289 Elena Serrano, Antonio Fernandez Ugalde & Leonor Pefia-Chocarro Los silos medievales en el Reino de Toledo 291-296 Settlements and landscapes- Habitats et paysages - Siedlungen und Landschaft Jurg Tauber Landliche Siedlungen in der Nordwestschweiz von der Merowingerzeit bis zum Mittelalter. Archaologische Quellen und historische Interpretation 356 297-308 Jean-Loup Abbe Le parcellaire rural des bastides du sud-ouest de la France: l'apport des sources ecrites et planimetriques 309-319 DirkMeier Der Wandel der Landschaft und Besiedlung von der romischen Kaiserzeit bis in das Mittelalter im Dithmarscher Ktistengebiet (Deutschland) 321-330 De/la Hooke The effect of English settlement in medieval North Wales 331-344 Piers J. Dixon Settlement in the hunting forests of southern Scotland in the medieval and later periods 345-354 357 MEDlEY AL EUROPE BRUGGE 1997 Internationaal congres over Middeleeuwse en Latere Archeologie 1 - 4 Oktober 1997 An International Conference of Medieval and Later Archaeology 1st - 4th October 1997 Patron The Government ofFlanders. werd georganiseerd door I was organised by fut organisee par I wurde veranstaltet von Stad Brugge Instituut voor het Archeologisch Patrimonium Vrije Universiteit Brussel Provinciebestuur van West- Vlaanderen S tad Brugge Instituut voor het Archeologisch Patrimonium Vrije Universiteit Brussel Stad Brugge Druk: ministerie van de Vlaamse Gemeenschap departement Leefmilieu en Infrastructuur afdeling Logistiek sectie Drukkerij ISBN 90-75230-12-5 Provincie West- Vlaanderen