ore Jankovi

The Slavs in the 6th century North Illyricum

ore Jankovi, Faculty of Philosophy, Belgrade* In the immediate surroundings of Vranje, several early Byzantine fortresses have been discovered, whose abandonment is linked to the Hun (Kutrigur) attacks in 539 and 544 A. D.; the same holds more or less true of other fortresses in southern Dardania and Macedonia Secunda. The fortresses in adjacent southern provinces were abandoned in 571-586. With the help of late Early Byzantine fibulae we can determine the provinces that lasted until the beginning of Emperor Heracliuss reign (610-641) in which, or along whose borders, also lived Slavs, i.e. Serbs. On the basis of written, archaeological, and topographic data, perhaps from as early as the mid 6th century, the Serbs, and possibly the Antes, i.e. Russians, lived in the upper Southern Morava Pinja Bregalnica Lepenac area. Key words: Early Middle Ages, Serbs, Slavs, Russians, Antes, Byzantium, Vranje. * This article is a result of work on the project Jablanica, Kukavica, Vranje and Krajite in 711 century funded by the Ministry of Science, Technology and Development of Republic of Serbia (file no. 1481).

Trial excavations of early Byzantine forts, conducted around Vranje in 2003, show that they were abandoned as early as in the course of the 6th century. These are the localities of Kale above Vranjska Banja and Gradite in Korbevac (, et al. 2004). Judging by surface findings, the case is the same with Gradite in Prvonek east of Vranjska Banja. The general impression is that all these forts are single-layered that they were built and demolished at the same time. The same roofing tiles on all three sites, the thinness of the cultural layer and the similarity of pottery show this. The absence of animal bones is notable. Kale was destroyed in a fire, whereas Gradite conveys the impression of having been abandoned. The found pottery fragments are similar to others from North Illyricum, dated to the second and third quarter of the 6th century; there is no such pottery in the strata of the time of the reconstruction of Emperor Mauritius ( 1981: 141-147; Bjelajac 1990: 163168). There are no data for the duration of other forts around Vranje, except for Kale at Klinovac, where a hoard was found with the latest coins from 544 (aj 1973: 2532). According to the inventory of the National Museum in Vranje, Markovo Kale was used in the 6th century, and later finds are not known. Kale at Veliki Trnovac to the South of Vranje ( 1986) seems to have been the centre of this area in the early Byzantine period, judging by the basilica, undoubtedly episcopal, with rich marble decoration, which is also in Vranje; there are no data suggesting its usage in the late 6th or early 7th century. These are not the only single-layered forts in this part of early Byzantine Illyria. Similar cases have been recorded adjacentllyria.in some adjacent regions. Skobaljic Grad near Leskovac also has only one early Byzantine layer (Jo 1987: 65). Short-lived forts were also found in Sirinika upa, where sondage was made into the gradite Zidinac above Gotovua and Rimsko Gradite at Brezovica. Although two layers were recorded at Rimsko

Gradiste, still the pottery finds show that it did not last into the late 6th century ( 1997b: 31-32). These two forts on the route Scupi Theranda could certainly have been destroyed in 548, when the Slavs forced their way into the region of Durazzo (BG III 29), for these forts were on the Dardanian borderland towards Prevalis and New Epirus. Scupi near Skopje was the capital of the province of Dardania, which included the region of Vranje. Procopius does not record Scupi, and there are no known written data showing that it existed as a Byzantine town in the second half of the 6th century. The last preserved written mention of Scupi dates from 518, when it was demolished in an earthquake which destroyed Dardania, but its inhabitants saved themselves by previously fleeing the city in anticipation of an enemys attack. This was recorded by Marcellinus Comes, as well as the fact that in 517 large parts of Illyricum, both Macedonias, Thessaly, and all the territory to Thermopylae and Old Epirus were plundered by the Gothic horsemen (Chron. 99-100). There, in basilica II, raised above the layer of the destroyed city, coins of Anastasius were found on the floor and of Justinian above ( 1979: 36, 44-45). Some of the towns of Macedonia Secunda could have been deserted in these events. In Stobi, the capital of Macedonia Secunda, two worn-out follises of Justin I were found in a building believed to have been destroyed by an earthquake ( 1979: 214-215). They could have found their way into the country some time during the reign of Emperor Justinian, when the whole city was destroyed. Later finds from Stobi are not known. On the basis of my knowledge concerning the pottery and other finds discovered at Gradite near Debrete, in the north of Pelagonia, there are no later finds than from the early reign of Justinian (, 1979: 141-148). Thus, with at least two or even six forts around Vranje which were not rebuilt after the conquest some time in the course of the 6th century, there are also two forts at Sirinika upa in the west; there are no data about the forts between the two. Northwards, at Skobalji Grad, life did not continue. In the south, in Scupi, the latest coins are those of Justinian, minted, by my guess, before the mid 6th century. This gives us an area 100 kilometers wide, in which, it seems, there were no Byzantine garrisoned forts since the second half of the 6th century, with a few exceptions to be discussed later. In addition to the facts already mentioned, the written sources recording the attacks of the Huns, that is, Kutrigurs, or Bulgars, also point to the period when these forts were conquered or finally deserted. The most reliable description of these attacks is that by Procopius ( II 4). They attacked twice. In the first attack in 539 they ravaged the area from the Adriatic (the Ionian Bay) to the suburbs of Tsargrad, conquered 32 fortresses in Illyricum, seized the city of Kassandria in Halkidiki in Macedonia, Chersonnes in Europe and even went across to Asia Minor. During the second attack they plundered Illyricum and Thrace, bypassed Thermophylae and plundered Hellas, excluding the Peloponnese. Repeated plundering affected the circulation of money and caused the locals to hide some hoards. Interestingly enough, contemporaneous hoards have not been recorded in the regions plundered during the attacks of the Gothic horsemen in 517-518. A reduction in the circulation is observed in Pernik in 538/9 ( 1981: 245-252; 1986: 33), which may have been destroyed then, judging by a large number of Danube-Illyrian fibulae with the rectangular cross-section of the curve ( 1981: 169-173; 1981: 169-171; ibid 1986: 136-138). At the same time the circulation was also reduced in the hoard from Selce near Sveti Nikola ( 1961). Hoards with the latest coins from 537/8 are from the vicinity of Prahovo on the Danube ( 1984: 84), Suva Reka in Metohija (aj 1973), Sekulica near Kratovo ( 1990), Oree near Veles (Petaki 1996) and Novo Selo near Strumica (Kondijanov 1996). Also there are hoards from Sadovik

near Pernik and Petelnica near Pazardik (Jurukova 1969: 259) with the latest coins from 538/9. Judging by the number of coins from the reigns of Emperors Anastasius, Justin I and Justinian respectively, in the hoards from Prahovo and Selce, the Suva Reka hoard could be the latest (containing the greatest proportion of coins of Justinian, instead of Justin). A few hoards contain the latest coins from 543/4: Dobra at Djerdap (M 1984), Klinovac near Vranje (aj 1984: 29-30), and also the hoard from Grnar near Gnjilane ( 1991);[1] the Suva Reka hoard should be contemporaneous. A few hoards from Bulgaria date from the same period (urukova 1969). It is certainly no accident that among the accidental coin finds from the forts in the very southwest of Dardania there are at least five coins of Emperor Anastasius, at least one of Justin I, eight of Justinian (527-538), two pieces from 538-544, one from 564/5, and only one or two pieces belong to the reign of Justin II or Emperor Mauritius ( 1996). Procopius notes that in 544 the Illyrian soldiers fled Italy for their homes because of the Hun attacks (BG III 11-12). Thus the second Hun attack is dated to 544. At the beginning of the fourth decade the Antes plunder Thrace, and then the Sklavenoi in 545, while at the same time the Antes were negotiating with the Empire over their settlement opposite Scythia Minor in order to prevent the Huns from crossing the Danube (BG III 14). That the agreement with the Antes was reached is shown, in addition to some archaeological finds, by the circumstance that the Kutrigurs invaded Illyricum via Pannonia. A massive Sclavonian campaign followed in 548, when they ravaged Illyricum all the way to the region of Durazzo, conquering forts that nobody defended because they were considered secure, looting, taking prisoners and freely searching through the terrain, while the Illyrian army followed, not daring to engage in open combat (BG III 29). This is the time when the Sirinika upa fortresses could have been destroyed. These Slavic attacks caused the hiding of the Dobra hoard, for the Danube banks and the road in Djerdap were not suitable for cavalry attacks in those days, but for those of infantry. Obviously, the surrounding area of Vranje was destroyed in the second Hun attack in 544, as shown by the hoards from the locality of Kale near Klinovac and also from Grnar. The disposition of the hoards corresponds to the described area of abandoned forts, which may indicate that all of them were permanently destroyed in 539-544. Hoards were also found further to the east, along the road used by the Huns (map 1). The hoards encompass areas along Europes key road: Philipopolis Sardica Naisus. They also encompass areas along the road Naisus Scupi Stobi Pautalia Sardica, in other words, the road connecting Illyricums capital Thessalonica and the areas close to the Danube. Thus we are shown the directions of the Hun campaigns, when the forts around Vranje were destroyed. Afterwards, between 550 and 559, this and the remaining areas of Illyria and Thrace suffered from devastating campaigns fought by both the Slavs and the Kutrigurs. However, there is very little numismatic evidence of these campaigns in the previously plundered regions there was no circulation of money, that is, no population. The exception in Illyricum is the hoard from Selce in the vicinity of Prilep, which can be dated to 551 at the earliest ( 1961). It seems that these events also had an affect on the system of government in the central regions of Illyricum. In Charter 11 of 535 signed by Emperor Justinian (527-565), which establishes Justiniana Prima, there is a mention of the province of Macedonia Secunda, whereas in Charter 131 of 545, confirming its rights, there is no such province. The most likely explanation is the distress between 535 and 545, certainly during the events mentioned above, when the area suffered a decline in population and economic power, and therefore in

importance, until eventually it was no longer recognized as a province. Simultaneously, Dardanias capital Scupi was not rebuilt either. Judging by the news announced in Procopiuss writings around 554, Justiniana Prima had become the capital of Dardania. Namely, in the province of the European Dardanians, in the vicinity of his birthplace, Emperor Justinian had built Justiniana Prima to become a city worthy to be the metropolis of the whole region. In addition, it has been chosen as the see of Illyricum (De aedif. IV 1). It is not improbable that earlier, in 535, Justiniana Prima was part of the Mediterranean Dacia, and later, together with its region, joined Dardania as its capital. At the time when Procopius announces the lists of forts, the Mediterranean Dacia is not represented by the capital city and the number of forts, but by the provinces of eight cities (Sardica, Kabetso, a town of an unknown name, Germania, Pautalia, Casseta, another unknown town, Remisiana), among which there is no mention of Justiniana Prima. The list of Dardanian forts mentions neither the capital, nor the cities which are previously said to have been built or restored by Emperor Justinian Justiniana Prima and Secunda, Justinopolis, and Scupi. Obviously, significant changes took place in the system of government, where more significance was attached to the cities ( 1981: 76-86). These changes are yet to be studied. The inhabitants of Scupi may have been settled in the newly founded Justiniana Prima. Be that as it may, it seems that in the mid 6th century the central part of Macedonia Secunda and the southeastern half of Dardania remained deserted and in some way united, not only as areas that were not rebuilt. The plunder of the fortresses in the border regions of the Vardar basin is also remarkable for the hiding of hoards, but not before the beginning of the reign of Emperor Mauritius (582602). Two hoards of coins have been found in Heraclea with the latest pieces dating from 583/4; the city was not rebuilt as a settlement until the 9th/10th centuries, as indicated by the pottery finds ( 2001). Another hoard with the latest coins from 584/5 has been found in the fort of Baba near Prilep ( 1977). A hoard of gold and bronze coins with the latest pieces dating from 584/5 was found at Bargala on the Bregalnica (A 1989: 65-67). A small hoard of bronze coins, which has not been made known to the public, was found at Markovo Krue on Vodan above Scoplje, with the latest coins from the time of Emperor Mauritius ( 1982: 51). There is no doubt that these forts were destroyed during the Slavic settlement further inland of Thessalonica, which was at first attacked by a small force, and later, in 586, a prolonged siege was laid to the city (Mir. I 12-14). That part of the Vardar basin with the river Crna and the Bregalnica was then permanently taken over by the Slavs (Popovi 1980). These Slav attacks were preceded by a Slav campaign into Macedonia, which ended in settlement in 571. It is not recorded in the written sources but was established on the basis of numismatic data (Popovi 1981: 111-125). As it is shown by the disposition of the hoards (Pirot, Cariin Grad, Thassos) and the reduction of circulation (Pernik, Amphiopolis, and the hoards mentioned above), the Slavs in Mediterranean Dacia split up in two groups, one of which penetrated from Ni to Praevalitana. Judging by a half-follis dating from 570/1,[2] the fort of Hisar at Leskovac, opposite Scobalji Grad, was destroyed in these events. The other group descended through the Struma valley to Amphiopolis, Thassos and the area of Thessalonica: afterwards they settled somewhere in the upper Strumica basin. This can be concluded on the basis of the inflow of money the inflow of money from Thessalonica into Pernik is afterwards limited (3 out of 10 coins in 570/1 to 576/7, in comparison with 26 coins minted in Thessalonica out of 32 in 565/6-569/0). On the other hand, in the Heraclea, Baba and Bargala hoards, coins from the mints outside Thessalonica occur only as an exception after 571 ( 1986: 39-41). In all probability, this Slavonic tribe participated in the subsequent attacks on Thessalonica. In Pernik there are no coins dating after 576/7, which

indicates that the city was most probably conquered during Slav campaigns against Greece, but that it was rebuilt and continued living until the beginning of the reign of Emperor Heracleus. In the area of hoards dating from 539-551, there are none from 571, nor from 584-586 (with the exception of those found on the border of this area Baba and Markovo Krue), which also shows that there could be no significant Byzantine settlements in these parts. Due to the insufficient number of explored archaeological sites of this era, invaluable data are obtained by mapping accidental fibulae finds, gathered in most cases by metal detectors. More and more is known about earlier and later Illyrian fibulae of the early Byzantine period ( 1986: 135-160). The use of fibulae was not customary in early Byzantine Macedonia, which applies to the whole Southern Illyria. Neither were they numerous in adjacent Dardania. However, a region has recently been delineated in the northwest of the Vardar basin, where several late fibulae were found. These are the sites of Markovo Krue and Gradite - Taor ( 1982: 51, . 26; 103, . 58) and Barovo ( 1996: . 1.2) near Skopje; in the north of Pelagonia Kale Brailovo ( 1993: 211, . 1.6); Isar Kale ipkovica, Gradite Podvis near Kievo ( 1996: 69, 72, 73, 76), Kale Izite at Pore ( 1996: 81-82), ajle near Gostivar ( 1996: . 1.5), Brikul Lukovica ( 1996: 59, 62; 1996: . 1.1). The manufacture of these fibulae in Aquis dates from after 586, and their use until the early 7th century; and of those with the broad curve and triangular foot from the beginning of the 7th century ( 1981: 172-174). The most conspicuous late fibula with a broad curve and a triangular foot has been found in southeastern Macedonia, in Budinarci near Berovo ( 1996: . 1.2). On the other hand, in the area of Macedonia we know of only one such Byzantine fibula from the lower basin of Danube. It is from Heraclea, which was finally destroyed around 585 ( 1988: 46-47, . 2). Consequently, these fibulae were used by the refugees settled there, most probably from the area of Aquis and other regions of Dacia Ripensis, subject to the Avars after 585/6 ( 1981: 63-65). In all probability, the movement of the refugees was prepared and carried out in 596-597, at the time when Byzantine commander Priscus made war against the Avars in those parts; at the same time settlement from the east and west can be perceived in the area of Aquis ( 1981: 158-161). However, we should not exclude the possibility of a settlement from the area of northeastern Dalmatia, i.e. from the Drina basin and the western regions of todays Serbia, as is indicated by some fibulae. In that case the migration also took place after 596, when the Avars were ravaging Dalmatia. In Pernik, where the circulation of money was terminated after 576/7, a large number of fibulae have been found, late ones, but also some earlier ones ( 1981: 166-174). This is why the question whether old or newly settled population lived in this area in the late 6th early 7th century remains unanswered. The circumstance that no coins of Emperor Mauritius or subsequent emperors were found, meaning that neither money-based economy, nor, consequently, central power were renewed, indicates that it was controlled by some Slavs. In the adjacent looted area of Macedonia finds of Slavonic fibulae have been made at an unknown location (Werner 1950: 155), in the surroundings of Bitolj and at Burlatica near Viniani next to Veles ( 1992: 24-25), which can be linked to the Slavs who settled from the lower Danube basin in the end of the 6th century (Fig. 1: 1-2). Recently a few finds of Slavonic cookware have appeared on the same territory, near the St. George church Gornji Kozjak (Bargala), Varnica Kruka Stari Karaorman and Kazandisko Malo tip ( 1996: 177180). All these finds can be roughly dated to the late 6th early 7th

century (Fig. 3: 1). In the southeast of Dardania, inside the walls of Bargala, gold coins of Emperor Phokas (602-610) were found, as well as Slavonic pottery ( 1989: 60, 65, . 72-75), which bears the most resemblance to pottery ascribed to the Serbs on the Danube ( 1997a: T. III-VI), from the late 6th and early 7th century (Fig. 2: 1-2). At the same time, in Scupi, situated at an easily accessible location, the Slavs lived as the prevailing or only inhabitants. Here, it seems, two Slavonic strata have been established. The older is represented by a dwelling situated in older walls, and determined by kneaded pots without decoration,[3] which on the Danube are dated to the late 6th and the beginning of the 7th century. The later stratum contains, it seems, pots decorated with a comb (Fig. 3: 4), similar to the cookware from the Danube, where it is ascribed to the Serbs of the late 6th 7th century. Two pots have been made known to the public, which could be dated to the late 6th - the beginning of the 7th century ( 1984: 54, no 24-25). One of them has its analogies on the Danube ( 1990: no. 101.7). Consequently, now we know that in the end of the 6th and in the early 7th century in the area from the Bregalnica to Skopje lived the newly settled Slavs, and that exposed, accessible old cities, situated in the valleys, were most probably held by the Serbs. To the north from the area of the abandoned forts on the line Vranje Sirinika upa, a line of rebuilt forts was formed in the central and northern regions of Dardania and in the whole Mediterranean Dacia. They lasted until the 7th century, until the beginning of the reign of Emperor Heraclius (610641). Judging by the late fibulae already mentioned (their number is growing, although they are not always well-dated), their southern border towards the region of the abandoned forts is on the line Babunica Ni Cariin Grad Obili ean Gradina above Pazarite ( 1999: 111). They also exist in Moesia Prima, as well as in the area of Katarakt, but they are particularly numerous in the area of Aquis ( 1986: 146160), and recently they have been discovered around Kruevac ( 2003). The circulation of money on that territory lasted even during the reign of the usurper Phokas, all until the beginning of the reign of Emperor Heraclius coins have been found in Ni, Cariin Grad, the surroundings of Jagodina, and, to all appearances, some hoards from the Danube with the latest coins of Emperor Mauritius belong to the same era ( 1986: 51-59). In Cariin Grad there are Slavonic artefacts. Two fibulae are known (- 1955: 168170; 1957: 313), designed, most probably, during the second half of the 6th century (Fig. 1: 34). They doubtless bear witness to the fact that the Slav female dress was worn in Cariin Grad. These two samples, in contrast to the two known contemporaneous Romaion ones, show that, presumably, a considerable number of inhabitants were Slavs before the city died out or was finally conquered. There may have been more Slavs than is indicated by the found objects, for if they used only Byzantine goods, they are hard to identify. Among the pottery from the two latest Early Byzantine layers (Bjelajac 1990: sl. 129-131), there are pots of what seems to be the Slav craftsmanship Fig. 3: 5-6); one of them is a pot with a handle Fig. 3: 6) pointing to the influence of the Byzantine craftsmanship although these are known to have been used by the Slavs on the territory of northern Illyric ( 1990: no 29.3). A small number of Slavonic potsherds have been recorded in the forts of the Aquis area ( 1979a), and probably there are some unnoticed and unpublished in the forts of Moesia Prima, Dardania and Mediterranean Dacia. A kneaded pot has accidentally been discovered near Trstenik (Fig. 3: 2) that could be dated to the 6th century ( 2002: 139-141). There are several early Slavonic finds near Jagodina, where coins of Emperor Phokas and Constantine IV from 643/4 and late fibulae have been found ( 1989: 84-85). Archaeological reconnaissance revealed potsherds that could be dated to the 6th century. In the village of Majura a fragment of a kneaded pot has been discovered, dark brown on the outside

and dark red on the surface of the fracture, made of clay with added sand, the sand containing mica and very little calcite (Fig. 2: 3). A potsherd was found on the Vrbica near the village of Dragocvet, which was also formed by kneading, but made of clay with a little calcite and fired in dark gray colours (Fig. 2: 4). They are kept in the Homeland Museum in Jagodina (inventory nos. 208 and 2144). The explored dwellings at Panjevaki Rit near Jagodina are not older than the mid 7th century ( 1993). In the second of the two explored early Slavonic settlements in Moesia Prima, at Reka near Vina, one dwelling belongs to the third quarter of the 6th century, i.e. to the period when this area was under Byzantine rule, and the other to the first quarter of the 7th century ( 1990: 17-20, 82-83). In the last stratum of some rebuilt forts throughout North Illyria (the Aquis area, Cariin Grad, the Raka basin, the coast), which lasted into the second decade of the 7th century, the usual finds are iron lids for baking bread in open hearths ( 1982: 144, 206; Bjelajac 1990: 182). Their appearance points to the Slavs, more precisely Serbs, since the use of such lids is their cultural characteristic ( 2001: 155-157). Consequently, the Slavs, i.e. Serbs, were settled on the land of Byzantium and in its forts in the last quarter of the 6th century as its allies and subjects, in Moesia Prima, the Aquis area, the urban areas of Mediterranean Dacia, in Dardania and elsewhere. The question is whether the regions of southern Dardania and Macedonia Secunda remained virtually deserted, as it seems, after the Hun attacks, whether Byzantium could have left them unpopulated, whether it could do something about it and whether it did it. According to written sources, this area grew in importance in 551, when the Slavs penetrated into the surroundings of Ni (it is not recorded whether it was conquered) with the intention to attack Thessalonica, but proceeded to Dalmatia (BG III 40). This area was also on the way of the Avars during their attack on Thessalonica. Undoubtedly, the defence of this road was important for Thessalonica, the capital of Illyria. The only way to keep, protect and make the deserted areas useful was at the time the settling of allies. It is known that the Gepids were settled on the Byzantine land after the defeat by the Avars in 567, and the Heruls before them. The area of Aquis on the Danube, from Donji Milanovac to Negotin and Gamzigrad in the inland area, may serve as an example of the settling of the Slavs. The Serbs, together with members of other Slavonic tribes, settled there around the middle of the second half of the 6th century ( 1997a). They were border guards - borderers, settled in the border villages and forts to defend the border and the road. Therefore, I assume that on the territory from the upper course of the Vardar with the Lepenac and southwards to the Bregalnica, across the Binaka Morava and its tributaries, eastwards probably to the watershed towards the Struma, some Slavs were settled as allies somewhere around the mid 6th century. Some toponyms mentioned in Procopiuss writings are interpreted as Slavonic, and precisely in the area destroyed in the Hun attacks: and in Dardania, as well as and in the region of Nis, in the region of the city of w (De aedif. IV 4; Niederle 1906: 179-180), which could be probably the same as Pernik. Of course, the theory of their Slavonic origin is refuted (Beevliev 1970:1-2), but in the light of archaeological discoveries it is becoming more and more convincing. If these are really Slavonic names, then they belong to some Slavs who had settled there before Procopiuss work was written, therefore in the end of the fourth decade of the 6th century, before 554, which is the probable date of its publication. The abovementioned migration of the Romans from the Danube to southwestern Dardania indirectly shows that the area of southeastern Dardania was settled by the Slavs. After 586, when a new population, including the Serbs and other Slavs, was settled in Coastal Dacia, in

fact in the Aquis area, the Romans may have been moved from there to the southeast of Dardania in 596-597, or they may have left the forts of northeastern Dalmatia (where afterwards the Serbs spread), again because of the Avar plunders. Their gathering only in the southwest of Dardania, judging by the known finds of late fibulae, indicates that southeastern regions were already settled, otherwise late fibulae would have been found there also. It is becoming more and more obvious that the Slavs started to take over Europes Byzantine territories much earlier and in a different way than it is usually thought. This conclusion is not compatible with the usual philological dating of the preserved geographical names of antiquity and foreign words in the Slavonic language.[4] Namely, opinions have taken root that the Slavs settled there in the beginning of the 7th century, taking over the valleys, while the Romanized indigenous population, later the Rumanians and Vlachs, and allegedly the Albanians, continued living in the mountainous regions. From them, allegedly, the conquering Slavs later took geographical names and a few words. However, the possible taking over of names from the Roman population may have taken place in the second half of the 6th century, and first of all by cohabitation in Byzantine forts. Interestingly, it is concluded that there are many more geographical names of pre-Roman origin in Serbian than preserved names of settlements of Roman times, except on the coast, where the Latins lived ( 1981: 131-132). To all appearances, the Romanized indigenous population moved away in the end of the 6th beginning of the 7th century. The written sources point to this conclusion, stating that the inhabitants of Thessalonica could recognize the Slavonic tongue as early as in 586, and explicitly claiming that refugees from Ni and Sardica lived there in the beginning of the reign of Emperor Heraclius, as well as from other provinces in 621-623 (Mir. I 12; Mir. II 2). Byzantium had an interest in grouping its population and refugees around Constantinople for the purpose of defense. That interest grew into a necessity when the Avars enslaved large masses of population and took them from the province of Europe, i.e. Constantinoples hinterland, in 623. There are records of Emperor Mauritiuss intentions to move 30 000 warriors and their families from Armenia to Thrace (Charanis 1959: 33-34). Of course, part of the old population remained in Illyricum and intermingled with the Slavs. Let us remember the peculiar cases of defection from the besiegers into Thessalonica (Slavs, Germans or?) and Prince Hatzcon, who was given shelter by some Slavs in 621 after the Slav siege of Thessalonica, led allegedly by him, was lifted (Mir. II 1). The obvious conclusion derived from the written and archaeological sources is that, despite the recorded Slav violence which occurred while they were seizing the Byzantine provinces in the Balkans, there was certain closeness between a part of Byzantine population and some Slavs. If the Sclavenoi were the Slavonic tribe attacking Byzantium, then it should be presumed that the Antes and Serbs were the ones collaborating with the Byzantines. These conclusions do not correspond to those of historiography, which differ more and more from exact, that is archaeological, facts.[5] Now let us consider the facts indicating that the Slavs who settled on the line ara - Kukavica - emernik were Serbs. First of all, the Serbs are mentioned as Byzantine allies in the reign of Emperor Heraclius, and later ( 1998: 131-136). If we presume that Priscuss Boiski, who, together with other tribes, defected to the Byzantine army in the middle of the first half of the 5th century (Frg. 1), in fact came from the land of Boiki where the Serbs came from according to Constantine VII Porphyrogenitus (DAI 32) - consequently that they were Serbs, then an alliance between the Serbs and Byzantium was recorded as early as then. The oldest mounds attributed to the Serbs, in the area of Grahovo and Drvar, date from the late 4th century ( 1998: 54-46). A settlement attributed to the 5th century Slavs has been

discovered in the vicinity of Viegrad (remonik 1970: 80). This alliance has also been archaeologically confirmed on the banks of the Danube, as already mentioned. The foundation of the settlements there discovered is dated to around the time of the Avar campaign in 584-586, which means that they were founded under Byzantine rule. This means that we are discussing the settling of the Slav allies for the purpose of defending the border. A remarkable Serbian ritual site dedicated to the dead and containing mounds has been discovered on the neighbouring Veliko Ostrvo ( 1998: 15-20). Representatives of other, different Slavonic tribes also lived in the settlements we have listed. Therefore, there is no reason not to assume that the Serbs, as well as some other Slavonic tribes, were Byzantiums allies in the mid 6th century also, and were as such settled in its interior. This is all the more true because of recorded and archaeologically confirmed examples of many Slavs being in Byzantine military service, not only as mercenaries, but also as military officers and commanders, therefore as Byzantine citizens, Romaions, certainly Christians, starting from Hilvud, the military commander of the prefecture of Thrace in the time of Justin I. The preserved toponyms and some recorded but as yet unexplored archaeological sites indirectly point to the fact that the Serbs settled in the area from ara via Kukavica to emernik. I have shown earlier that the Serbs within South-East Europe are characterized by some toponyms originating from the time of cremation of the dead ( 1998: 124-126). The most conspicuous for our purpose are those bearing witness to places called Igrite located on mountain tops between valleys, as gathering places of the inhabitants of mountainous districts (upe) on holidays. In the surroundings of Vranje as many as four Igrite are known: the first to the southeast from the village of urkovica, at the altitude of approx. 1000 m, near the peak of Motin (1307 m), the second to the east of Kriva Feja on the mountain chain that stretches out to the north from the Besna Kobila peak (1922 m), at the altitude of 1600-1700 m, then to the north of Vranje near Oblik (1310 m), at the altitude of approx. 1100-1200 m, and the fourth 11-12 km to the north, on the southern slopes of Mount Kukavica, under the peak of Lisac (1375 m), at the altitude of approx. 1080 m. The village of Igrite is situated to the north-east of Kukavica. There are a number of gumnite toponyms meaning a threshing-floor at similar locations. In the west of this area, above Muutite near Prizren, Igrite was recorded in the 14th century, which exists no more. The peak of Sveti Ilija (1270 m) towers to the west of Vranje. Above Uroevac, near the mouth of the Nerodimka, where it joins the Lepenac, rises Ljuboten 2499 m high, on which the local population gathered on the days of St. Eliah before the NATO occupation; this may have been a gathering place for Kosovo Polje, the Lepenac, Polog, Crna Gora, the Binaka Morava. There are more peaks like Sveti Ilija near Vranje, which may have served as gathering places for the upa (district) of Vranje and the upper Morava. Not far, in the east of Metohija, there is Zborite (gathering place) and there is also one to the south of Ni. No such grouping of toponyms of this kind has been preserved anywhere in the basins of the Morava and Vardar, which shows that they belong to the age before the Nemanjies. It is hard to imagine that at the time of King Stefan the First-Crowned (1217-1228), when these regions were finally freed, the population was replaced by newly settled Serbs, who, moreover, followed customs the Church could not have regarded with approval. Some details point to the Russians, i.e. Antes, as probable settlers in the areas around Vranje. In Krajite and Vlasina there are sagas of the extermination of Latin males in a battle, and of the settling of the Russians ( 1912: 165-167). Kijevac, the name of two villages near Babunica and near Surdulica, seems to confirm this legend, as well as the name of the village of Kijevo, although it is located on the opposite, western side of Vranje. There is also a village of Rusce to the south of Vranje between Lepinac and Klenik, and there are Rosuljak,

the Rusalijsko cemetery, etc. It was mentioned earlier that in 544 negotiations with the Antes took place about their settling on the border. Besides, Emperor Justinian himself had the name Antikos () in his title since 533 (Chron. Pasch. 632, 2). The Russian tradition records that Kiy, the founder of Kiev in the land of Poljani, conducted certain negotiations with the Byzantine Emperor, and that he erected Kijevec on the Danube ( 12-14). His sisters name is of interest because of , a fort in Dardania recorded by Procopius, with the river of Lab and the village of Lapunica to the east of Podujevo, situated close to unexplored ruins of a fortress. A village called Lebet exists on the Mala river, 13 km northeast of Vladiin Han. Turkish censuses of the 16th century record, along with the villages of Lebed, i.e. Lebet and Plavi Lebed, the toponym Bukovi Lebed (?), the villages of Rusce and Kijevo ( 1985: 149, 153, 160). The names of the village of Kijevci Poljani and the village of Hrvati, in the late medieval district of Brvenik on the Ibar, clearly refer to a migration from the region of the Dnieper and the Dniester. The question of the date of this migration remains unanswered. The archaeological witnesses of the late 6th century migrations from that direction are the above-mentioned Slavonic fibulae from Cariin Grad (Fig. 1: 3-4). One of them, decorated with little circles, belongs to a kind characteristic of the Dnieper basin and this is the only such specimen found south of the Sava and the Danube, although they exist in Pannonia (Werner 1950: 160-170). The study of these issues requires new generations of archaeologists, ethnologists and philologists, but the solutions are obviously within our grasp. The igrita (playing-grounds) as well as the toponyms Vrana Glava, Kobilja Glava and suchlike, which point to old beliefs, here also correspond with the occurrence of tumuli. The probable mounds (gromile), i.e. fairly small mounds that can be presumed to be medieval, have been recorded in several places south of the Southern Morava. There is one, covered with grass, on the mountain chain on the northern side of Pinja, about 2 km to the northeast from St Prohor Pinjski Monastery. On the right of the road descending to the Monastery there are a few heaps of stone that could be tumuli. Some 15 tumuli have been discovered on Mali emernik near the village of Mlaiste, on the mountain ridge between the estates of Golubovci and Kuitak, and there are even more on the neighbouring ridge. Since they have not been excavated, their dating is difficult, but they are unlikely to be younger than the 12th century. There is no doubt that the Slavs settled these regions very early. Whether they were Serbs or Antes, or both, and whether the settlement took place in the mid 6th century, can be established only by excavation. The solution lies, on the one hand, in the precise dating of the strata of early Byzantine forts. If they really remained abandoned on such a vast territory, they are a truly reliable indicator showing that a population of a different culture, one that did not inhabit their walls, settled there. On the other hand, only archaeological excavation of the Slav sites can incontestably determine when they were set up. The Serbian border settlements on the banks of the Danube exist side by side with the Byzantine forts or next to abandoned forts. Therefore, if, in addition, the Serbs as the subjects of Byzantium were the only barbaric tribe to inhabit its towns in such numbers as to have given them their name ( = Serbia, todays Servia, the Town of Serbs of an unknown location in Asia Minor), we may justifiably assume that they had previously settled in remote and depopulated regions, in the areas important for the defence of Byzantium, but closer to their land of origin. Hence the archaeological evidence of the Serbs settling in depopulated towns, as Scupi and Bargala were since the mid 6th century, is not unusual.

Illustrations

Map 1. Fortresses and hoards in North Illyricum abandoned towards the end of the first half of the 6th century: 1. Dobra; 2. Hajduka Vodenica; 3. Prahovo; 4. Grad Nikole Skobaljia; 5. Suva Reka; 6. Brezovica; 7. Gotovua; 8. Veliki Trnovac; 9. Markovo Kale; 10. Korbevac; 11. Kale; 12. Prvonek; 13. Klinovac; 14. Sadovik; 15. Pernik; 16. Scupi; 17. Sekulica; 18. Selce; 19. Oree; 20. Debrete; 21. Stobi; 22. Novo Selo; 23. turci. Legend:

Hoards from AD 550 Hoards from AD 544 Hoards from AD 539

Significant fortresses Archaeologically confirmed deserted fortresses from the time of Justinian Fortresses presumably abandoned during the time of Justinian Sites with fibulae designed before AD 585

Map 2. Slavic sites, hoards and finds of Byzantine fibulae in North Illyricum dating from the late 6th and early 7th centuries: 1. Vina; 2. Veliki Ostrov; 3. Dragocvet; 4. Majur; 5. Jagodina; 6. Trstenik; 7. Pazarite; 8. Ni; 9. Babunica; 10. Pirot; 11. Cariin Grad (Iustiniana Prima); 12. ean; 13. Obili; 14. Scupi; 15. Barovo; 16. Vodno; 17. Pernik; 18. ipkovica; 19. ajle; 20. Lukavica; 21. Iite; 22. Taor; 23. Podvis; 24. Brailovo; 25. Viniani; 26. Stari Karaorman; 27. tip; 28. Bargala (Gornji Kozjak); 29. Budinarci; 30. Bitola; 31. Heraclea.

Legend:

Decorated pottery characteristic of the Serb workmanship Slavic pottery without decoration and fibulae found outside fortresses Igrite toponyms Hoards from AD 571 Byzantine fibulae designed after AD 585 Forts with Slavic pottery without decoration or fibulae Forts with decorated pottery characteristic of the Serb workmanship Hoards from AD 584-585 Hoards from around AD 615

Fig. 1. Slavic fibulae from the last quarter of the 6th century: 1 Burlatica, 2 - Bitola ( 1992: T. 1); 3-4 Cariin Grad ( 1977: T. XII).

Fig. 2. Decorated pottery characteristic of the Serb workmanship, dating from the late 6th and early 7th centuries: 1-2 Bargala ( 1989: . 73, 75); Slavic pottery without decoration dating from the late 6th and early 7th centuries, 3 Majur, 4 - Dragocvet.

Fig. 3. Slavic pottery without decoration dating from the late 6th and early 7th centuries: 1 Stari Karaorman ( 1996: sl. 6); 2 ( 2002: . I 1); 5-6 Cariin Grad (Bjelajac 1990: . 129, 131); 7 - Scupi, a wheel-turned pot, 8 wheel-turned decorated pot of the Serbian type ( 1984: . 24-25).

SOURCES
(Quoted according to I 1955 and II 1959, ; Fontes latini historiae bulgaricae II, 1958)

BG: Bellum Gothicum in Procopii opera, ed. J. Haury, Lipsiae 1906. BP: Bellum Persicum in Procopii opera, ed. J. Haury, Lipsiae 1906. Chron.: Marcellini V. C. Comitis Chronicon, ed. Th. Mommsen, Berolini 1893. Chron. Pasch.: Chronicon paschale, rec. L. Dindorf, Bonnae, 1832 .

DAI: Constantine Porphyrogenitus, De admistrando imperio, Greek text ed. by Gy. Moravcsik, Engl. transl. by R. J. H. Jenkins, Budapest 1949. De aedif.: De aedificiis, in Procopii opera, ed. J. Haury, Lipsiae 1913. Frg.: Prisci Fragmenta, ed. L. Dindorf, Lipsiae 1870. Hist.: Theophylacti Simocattae historiae, ed. C. de Boor, Lipsiae 1887. Mir. I, II: Miraculorum s. Demetrii martyris liber I-II, ed. C. Byeus, 1891. : , 1950.

References

, . 1989 , . , ., , . , . 1979 - . 1977 ., Macedoniae acta archaeologica 5: 141-149. , . 1998 , in , ed. . , , 372-390, 681-728. , . 1996 , 14: 173-182. Beevliev, V. 1970 Zur Deutung der Kastellnamen in Prokops Werk ''De aedificiis'', Amsterdam. Bjlajac, Lj. 1990 La cramique et les lampes, in Cariin Grad II, ed. B. Bavant, V. Kondi et J.M. Spieser, Beograd Rome, 161-190. Charanis, P. 1959 Changes in the Byzantine Empire in the seventh century, Dumbarton Oaks Papers 13: 23-44. , . 1986 '''' , 10: 59-60. remonik. I, 1970 Istraivanja u Muiu i abljaku i prvi nalaz najstarijih slavenskih naselja kod nas, Glasnik Zemaljskog muzeja 25: 147-260. , . 1989 ( 1987), 5: 82-85. , . 1973 VI , 7: 25-37. , . , . 1979 II , M 5: 31-46. , . 1981 XII , in I, ed. . , 1981, 125-140. , . 2001 Heraclea Lyncestis, . , . 1981 VI VII , . o 1986 6-7. , , doctoral dissertation, Faculty of Philosophy, Belgrade. o 1997 VI VII , in , ed. . , , 133-149. o 1997 , 10: 31-35. o 2000 , in . , ed. . , , 45-68.

1998 , . 2001 , 1901-2001, ed. . , , 141-161. , . , . 1990 , . , . 1962 , II, . , . , . 1987 , 27: 61-67. , . 1981 , , ed. . , , 218-257. o 1969 Les invasions Slaves au sud du Danube daprs les trsors montaires en Bulgarie, Byzantinobulgarica 3, 256-262. , . 1977 , 3: 181-192. , . 1961 , , () 3: 53-64. , . 1990 . , 11: 199-213. o 1996 The Early Byzantine Hoard from Novo Selo, near Strumica, acedonian Numismatic Journal 2: 95-104. , . , . 1977 , . , . 1984 , . , . 1996 , , , , I, 2: 53-84. , . , . 1993 , M 13: 227-243. , . 1981 , , ed. . , , 107-200. Malcolm, N. 1998 Kosovo. Short History, London. , . 1988 , 12-13: 45-54. o 1992 , . -, . 1955 1953 1954 , 5-6: 155-180. o 1957 1955 1956 , 7-8: 11-328. , . 1979 , Studies in the Antiquites of Stobi 3: 205-215. o 1982 , . , . 1984 , 7: 16-17. , . 1912 , 18, 8, . Niederle L. 1906 Slovanske staroitnosti, Puvod a potky Slovany jinich, II 1, Prag. Petaki . 1996 The Early Byzantine Hoard from the Village Oree, MNJ 2: 87-92.
o o

Popovi, V. 1980 Aux origines de la slavisation des Balkans: la constitution des premires sklavinies macdoniennes vers la fin du VIe sicle, Acadmie des inscriptions & belles - letters 1980, 229-257. o 1981 Un invasion Slave sous Justin II inconnu des sources crites, 4: 111-126. o 1984 , 7: 83-90. Popovi, . 1999 Tvrava Ras, Beograd. , . 1991 , 14: 49-56. , ., , ., , . , . 2004 , 20 (forthcoming) , . 2003 , communication at the meeting of the SAD Section for Medieval Archaeology, Beej, December 2003. , . , . 2002 , 18: 137-155. , . 1985 XVI , . , . 1993 , , VII-X , 9: 207-210. Werner, J. 1950 Slawische Bgelfibeln des 7. Jahrhunderts, Reinecke-Festschrift, Mainz, 150-172. , T. 2000 , .

6.
2003. ( 2004), 6. , ( 1981: 141-147; Bjelajac 1990: 163-168). , , ( 1986), . , ( 1979: 36, 44-45) ( 1979: 214-215), ( 1997: 31-32), 548. (, BG III 29), (, 1979: 141-148), ( 1987: 65). 535. 545, , , (, De aedif. IV 1; 1981: 76-86), , . - 539. 544. (, II 4; BG III 11-12, 14). , ( 1961; 1973; , 1984: 29-30; Jurukova 1969; 1981: 245-252; 1981: 169-173; 1981: 169-171; 1986: 33, 136-138; 1984; 1984: 84; 1991; Petaki 1996; 1990; Kondijanov 1996),

 ( 1996). 544. ( 1973: 2532) . . , . , 586. ( 1982: 51), ( 1989: 6567), ( 1977), ( 2001), 571. ( 1981: 111-125; 1986: 39-41), . , 7. . 6. , . , a 586. , . , , , , . , , . . , 539-544, , . . 6. , : , , (, De aedif. IV 4; Niederle 1906: 179-180). , . - , , ( , ) , ( 1962: 39).[6] , , . 6. , ; ( 1981: 131-132). '' '' 6. - 7. (Mir. I 12; Mir. II 2). (Charanis 1959: 33-34). , ; (Mir. II 1). , , . , . , .[7] , , , 4. 5. ,

 6. . 6. - 7. ; , , . , . '''' '' '', . . ( = , ; = ), . ( 1912: 165-167). , () 533. (Chron. Pasch.: 146), 544. . , , , (: 12-14). , , . . 16. , , (?), ( 1985: 149, 153, 160). , , . (. 1: 3-4) (Werner 1950: 160-170) 6. .

[1] I did not find a village of that name near Gnjilane, but there is one east of Vranje. [2] According to information provided by researcher M. Stoji of the Archaeological Institute in Belgrade. [3] Based on my inspection of the finds excavated by B. Babi and M. Mandi, for which I thank them. [4] Philology, easily finding material for its studies, still cannot determine absolute dates of the perceived phenomena, but only relative, and thus depends on the outcome of archaeological research. For instance, the question of dating the toponyms attributed to the Romanized indigenous population remains unanswered. Todays Vlachs of eastern Serbia have neither preserved toponyms nor oral tradition speaking of historical places: instead of Aquis stands Slavic Prahovo, instead of Romuliana Slavic Gamzigrad, instead of Pontes Slavic Kladovo etc. In some places the names from the late Middle Ages are also lost: the fort of Viesav is today known as Rimski Bunar. This undoubtedly confirms the settlement of the Vlachs in 18th-19th centuries, recorded in the written sources. Let us also remember that all medieval charters of Vlachian and Moldavian rulers are written in Slavonic languages, using Cyrillic alphabet, as well as the inscriptions in churches, and only in the 16th century did the gradual Rumanization of the language begin, whilst the Cyrillic alphabet was abandoned in the 18th century. Hence only medieval toponyms and censuses may be used for such research, and used with caution. It has been a long time since K. Jirecek saw the survival of the Romans

through the names Ni Naisus, Ras Arsa, Ljipljan Ulpiana, Skoplje Scupi ( 1962: 39). But, to all appearances, Ni was the only city surviving until the 9th century as a Byzantine stronghold with Slavenized or Slav population (like Serdica); no native finds younger than the beginning of the 7th century were discovered in the mountainous forts explored in the Raka basin, despite the extensive explorations; the medieval Ljipljan is not the same as Ulpiana, for it is situated on top of some other Early Byzantine city (Justinopolis in Procopiuss writings?); the Slavs, not the Albanians, lived in Scupi since as early as the 6th century. How, on the basis of such facts, can we speak of the continuity of the life of the Romanized indigenous population, side by side with the Slavs, especially in the mountains? [5] These views are presented in the first three chapters of the work entitled The Settling of the Slavs and the Formation of the Serbian State (in ). In recent years H. Malcolm (Malcolm 1998) and T. ivkovi ( 2000) have further elaborated these attitudes, seeing the country of Serbs within the borders formed by the West Morava, the Ibar, the Drina, the Piva and the Tara. From there, somewhere in 11th-13th centuries, the Serbs are supposed to have spread to the east and pushed the Romans, ancestors of the Rumanians and Vlachs, from Kosovo (and Metohija) and the Morava basin. Allegedly, it is not clear how the Serbs, coming from the cramped space of the 7th century Servia, could have colonized the so much more spacious regions of Dalmatia. This ignores Srb in Lika, Gordoserba in Asia Minor, the Serbs on the Elbe and much more. The area mentioned is a hilly-mountainous region, suitable for cattle-breeding, whereas the regions where the descendants of the Romanized indigenous population allegedly resided are filled with spacious valleys suitable for farming, such as Metohija, Kosovo, Vranje and the rest of the Morava basin. This contradicts the view that the Slavs took the valleys, while the Romanized population took to higher regions and lived by cattle-breeding. And why should the space where the Serbs lived be limited by the Tara and Drina, when the facts given by Porphyrogenitus (in some cases accepted in their entirety if necessary, and in other cases rejected) testify that in the mid 10th century the Princedom of Serbia extended between Croatia (from Cetina) and Bulgaria? This also ignores the territory of Serbian (Serbo-Croatian), whish allegedly wedged itself among the South Slavonic dialects, and whose artificial division into Serbian and Croatian probably justified the fictitious Drina border. T. ivkovic also arrives at other conclusions that cannot hold water. Among other things, he claims there was no Slav settlement around Thessalonica or in the Balkan peninsula in general before the beginning of the 7th century. Today the known archaeological material must be considered as a historical source (let me draw your attention to the graves of Slav women buried on the Byzantine land, therefore Christians). In addition, there is unnoticed information, given by Theophylact Simocatta (Hist. VII 2), about the Slav settlements on the right bank of the Danube, when the commander Peter comes across the settlements of the Slavs. T. ivkovi treats archaeological material as if it were philological data, not knowing, as a historian, that archaeology uses methods which establish not only relative, but also absolute chronology: he suggests that the finds from Muii be dated to the 7th instead of to the 6th century, or later (and leaves out the fact that I. remonik has dated the older stratum of settlement to the 5th century), and for this he needs neither expert nor scientific proof. For the cemetery of Djonaj, which used to be dated to the 9th-11th, and now to 10-13th centuries, although it can hardly be older than the 11th century ( 1998: 720-724), he employs the same system of guesswork, suggesting it should rather be dated to the 9th and 10th than to the 11th century when the Romaion population, hitherto confined to the mountainous regions, starts returning to the valleys which are by that time occupied by Slavs. ( 2000: 140). And then he arrives at the absurd conclusion: the Romans lived in the area Tetovo

Timok as cattle-breeders and were forced to move from Kosovo (and Metohija?) by the farming Serbs (from mountainous regions!) to Rumania. And what kind of migration of Rumanian ancestors from the south of todays Serbia to todays Rumania or Thessaly could have occurred in the 11th century without being recorded in Byzantine sources? Physical anthropology, referred to by T. ivkovi, in which even poorer investments are made than in archaeology, still very much depends on archaeological interpretation, and first of all on dating. It should be remembered that the Slavs used to cremate their dead, so it is not clear what general conclusions about their anthropological type can be drawn on the basis of the available material. [6] , . . , , , . , 9. ; a 7. ; ; 6. , . [7] . (Malkom 1998) . (2000) , , , . , . , ; 9-10. , , , 11-13. . , ? . ; . , , . - ?.