WEEK FORD TIN MILLS, DARTMOOR By PHILIP NEWMAN The earthworks and structural remains of the tin mills at Week Ford were surveyed in April 1991. An upto-date record of the site is presented, including a review of the work of previous writers. Some new observations about the remains arising from the survey are offered and a simple chronological sequence is suggested. The site is particularly important in the historical study of Dartmoor archaeology as it was, in the 1880s, the first tin mill to be excavated or 'cleared'. INTRODUCTION In recent years controversy has focused on one of the most interesting and important tin mill sites in Britain, at Week Ford in Devon (Earl 1990, Greeves 1990) where a vigorous debate has developed around the publication of a radiocarbon date centering on the 8th century AD, given from a sample taken at the site (Earl 1989, 119). Debate is not new to Week Ford, however. In the late 19th century it was foremost among tin mill sites on Dartmoor as a place where first ideas about early tinworking processes were formulated, and from that time most studies of tin mills have included this site in their discourse. In 1991 the writer was given access to the mills by the present owner, Mr Anton Coaker, to undertake a detailed ground survey, the results of which are reproduced here. Although earlier surveys of Week Ford have been published by Burnard (1887-90) and Worth (1940), associated earthworks have not previously been considered in detail. The following paper is a description of the field remains, based on the new survey, and a discussion of the contributions of previous writers, with some reference to the recent dating debate. DOCUMENTARY EVIDENCE AND EARLY INVESTIGATIONS Week Ford, or Beara Mill as it is alternatively known, first enters the documentary record in 1608 when John Hannaford paid 3d for 'Wikeford Milles' and two acres (DuCo London, Dartmoor Proceedings 1203-1735, fol. 29). 'Weekeford mills' are mentioned again on leases and mortgages in 1730 and 1737 (DRO Accessions List, documents in private collection), although these may be references to the place and the parcel of land rather than to working tin mills. Little light is shed on the earlier history of the site from documentary sources and any pre-17th-century activity may only be assumed at present. The well-preserved nature and picturesque, accessible location of the mills at Week Ford led to their being among the first tin mills to come to the attention of antiquarian investigators in the 19th century. It was probably here that a tin ingot was discovered by Edmund Pearse in the 1830s, reported in a letter to the early Dartmoor writer Mrs E.A. Bray, which was subsequently published by her: ... In and about the old streamworks there are now to be seen several remains of the Phoenician smelting houses, called Jews Houses: from one of these near the confluence of East and West Darts [Dartmeet], about three years since, there was taken tin ore, which was redressed and smelted at Crowndale, by the present Tavistock Smelting Company, and not far from this place there was found a block of Jews tin, supposed to be the most ancient in existence, and now in possession of a gentleman of this town. The surface of this block betrays marks of great antiquity, being much corroded by the influence of those external agents to which it has been exposed (Bray 1836, 254). Although other tin mills are now known of in the locality of Dartmeet, identified from documentary evidence (French and Linehan 1963, 177), their field remains are insignificant and unlikely to have been recognised as blowing mills in 1830. The structures at Week Ford, which were undoubtedly in a very fine state of preservation at that time, are the most likely candidates for those which Pearse describes. In the late 1880s the lower mill at Week Ford received the distinction of being the first Dartmoor tin mill to be investigated archaeologically or `cleared' when Robert Burnard, the noted Dartmoor antiquary and founder member in 1892 of the Dartmoor Exploration Committee, searched through the tumbled stone of its interior. Burnard had previously published a survey and description of both mills in a paper of 1888 (Burnard 1887-90, 104-6), in which he described the structures and the furnace, and alluded to stones containing `curious circular or oval-shaped cavities'. In a second paper the results of the investigation were described, although not in any great detail (ibid., 226-8). Several more stones with circular cavities were uncovered among the tumble, in addition to two fragments of a mouldstone, which together form the mould surviving at the site today. A debate as to the function of the stones with circular hollows then evolved. William Crossing, who had first encountered the stones in 1885, having been told of their existence by a local farmer and having informed Charles Burnard (Robert Burnard's father), assumed they were moulds for casting ingots and tells of how he disagreed with a local labourer who believed them to have been formed by stamps (Crossing 1891, 9). Robert Burnard, while accepting that the stones were indeed for the purpose of crushing, believed mechanical stamps would have shattered the stones. He gave them the name of `mortars' and declared that they were for hand stamping, used primarily for crushing slags but also larger nodules of stream tin (Burnard 1887-8). In the Mining Journal for 1891 an anonymous writer gave support to the idea of the stones being used in conjunction with mechanical stamps, stating, `Some of the hollows ran in sets of three on stones which had evidently formed the beds of very simple and primitive "stamps", (Anon. 1891, 947). The next important investigator to visit Week Ford was R. Hansford Worth. Worth was the first person to carry out a programme of systematic fieldwork to research Dartmoor tin mills, recording and publishing over 40 such sites between 1914 and 1940. It was Worth's talent for thorough data-gathering and accurate recording at mills over the whole of Dartmoor which allowed him eventually to conclude that mortarstones had indeed been used in conjunction with mechanical stamps, although he did not cite evidence from Week Ford on this particular occasion. Worth did survey the two mills at Week Ford, however, choosing to draw them together at a smaller scale, to demonstrate their relative positions, and separately at large scale for the sake of detail. His description of the site was comprehensive, with measurements of all the structural features, and he mentioned an additional find of a polished axle-bearing stone.' He also observed that water from the wheelpit of the upper mill, having passed over the wheel, was diverted via the tailrace onto the wheel of the lower mill, thus, in his opinion, proving contemporaneity of use of the two mills (Worth 1940, 201). A final contributor to the mortarstone debate who chose to focus his discussions at Week Ford was Hamlyn Parsons (1956, 194-5). He considered the name `mortarstone' to be a misnomer, since when Burnard had first referred to them thus he had done so in the belief that the hollows had been deliberately formed to contain the ore during the crushing process, in a way similar to a pestle and mortar. Worth's research had by then demonstrated that the hollows were a consequence of wear rather than having been created to hold ore while being crushed, and Parsons suggested that the stones should in future be referred to as `anvil stones' but this term has not been adopted by successive writers. He described two examples at Week Ford which in his opinion had been discarded through excessive wear, and one which was only slightly worn. THE SITE (Figs 1-2) Week Ford is a stepping-stone crossing point on the West Dart river, approximately 1.3km upstream of Dartmeet, near the confluence of the 0 Brook and the Dart. The mills are sited at 0 Brook 0 0 O ~ E N a, L0 o~ 0 Cc E'° c o fO o m _' F _ O t Wa ~ J i 187 Week Ford Earl's find spot T; Mill A tailrace - %Q G iT =\\1111111 'SIC j.;. . ~~~~~ " "/ll lll'll 0''nl~\\\~I lr n~n \ '/'~i%, // l _ Mill B : hit ►111 ' 1 "IIIj1I III feat \\Illl/ embankment , j z, 1 11` reave • 0 Fig. 2. Week Ford: earthwork survey. 188 metres 20 the foot of a gently sloping northward-facing hillside, south of the Dart and about 80m west of the 0 Brook, at SX 66187232. The valley of the 0 Brook, which extends for approximately 6km to the south-west of Week Ford, has been extensively worked for tin, with much of the valley floor remaining today as dry channels and spoil heaps characteristic of streamworking, the area immediately adjacent to the Week Ford mills on the east bank having been particularly heavily worked. A large streamwork called Dry Lake, which joins the 0 Brook valley at SX 66107050, is mentioned in a document of 1240, making it one of the earliest-documented tinworks on Dartmoor (Rowe 1985, 165). Further south again there are large complexes of opencast lode workings at Henroost and Skur, mentioned in 1569 (Greeves 1986, 4), as well as the Henroost Mine which ceased working in the early 1920s (ibid., 6). The 0 Brook itself forms a section of boundary to the ancient Royal Forest of Dartmoor, placing the mills at Week Ford just within the forest. The site is within Ashburton stannary. The remains consist of two substantial ruined granite buildings, sited c. 10m apart (Fig. 2). Two teats served the mills; both enter the site from the east and would have carried water from the 0 Brook to the waterwheels. A shallow gully running north-to-south on the eastern side of the site may also have served as a leat supplying water to a dressing floor. There is some prehistoric evidence in the vicinity: a section of reave at the southern extremity of the site has been cut through by one of the ]eats and does not survive beyond that point. There is a hut circle c. 180m south of the mills. Its walls still stand to over lm in places; the hut has clearly been subject to post-prehistoric remodelling. MILL A (Pl. 1; Fig. 3) Mill A, the lower mill, is levelled into the base of the hillside at the point where the slope merges into the floodplain. It is among the largest tin mills on Dartmoor with overall internal dimensions of 10.2 x 4.7m. The north wall and a section of the eastern wall are built against an earthen mound which may represent a vestige of the hillslope into which the building was cut. The south wall is also revetted against the hillside, with a drop of c. 1.6m from the external ground level. Although the condition of the surviving walls is fair, much of the masonry has collapsed and the interior of the building is completely covered with tumble. (It appears that when `clearing' the structure, Burnard, who claimed only to have touched loose stone, either replaced the stones which he had removed, or did not remove them from the building but merely searched through them.) There are opposing entrances in the east and west walls. That in the west wall is lm wide and has two surviving jambs, both with vertical grooves to accommodate the woodwork of the door. The eastern entrance, which is approached from outside by a hollow path cut into the foot of the slope, has an opening of 1.5m and one upright jamb surviving. Both entrances have granite threshold stones. The furnace is sited centrally between the east and west walls, to the south of the entrances. It is constructed from granite slabs, with the back and east side surviving in place, and is built into a low revetment which traverses the building, creating a slightly raised floor in the southern portion of the mill. Although the western side upright of the furnace is missing, the opening would have been c. 0.8m wide. Built into the south-east corner of the furnace is a fragment of a broken mouldstone, with one corner of the mould surviving. The main mouldstone (Pl. 1) is located just within the western door; its upper surface is level with the mill interior. This position is very similar to those of other recorded moulds (e.g. the Upper and Lower Merrivale mills on the river Walkham), suggesting that it may still be in situ. However, Burnard reported finding two fragments of a mould which he re-united and placed in their present position. The fracture separating the two halves is still visible across the upper face. The wheelpit is attached to the western exterior of the building but is totally obscured by fallen masonry. Its position is revealed by the alignment of the ]eat and tailrace, as well as the presence of an aperture in the western wall, through which the drive shaft from the waterwheel to the machinery would have passed. This consists of a rectangular opening measuring 'idLC I. Cck l - olli: nluuiUstunc iii Mill A ( I m scale). 0.9 x 0.55m high, capped by a single stone lintel. The tailrace is also filled with tumble, though its course is clearer toward the northern end of the building where the channel is cut into the adjacent bank and runs parallel with the structure. In the south-west corner of the building a recess of 1.0 x 1.2m has been built into the wall. Its close proximity to the wheel-axle opening raises the question of whether it in some way accommodated part of the machinery. However it could alternatively have been a chute into which unprocessed material was tipped into the building from above, making use of the height to the south of the building. The mill has a second chamber attached to its south-eastern side measuring 1.3m wide. The northern wall of this chamber, if it had one, has now become obscured by the growth of a large tree which has disrupted the area. The function of this additional space is not clear but its construction belonged to the same structural phase as the southern wall of the mill. This is evident now that the main eastern wall has collapsed almost to ground level, without leaving a scar on the south wall, to which it had been butted and which still stands to 1.6m. The southeastern external corner of the mill is surrounded by an additional low wall and bank 1.5m from the building which appears to define an alleyway, perhaps giving access to the rear. Several mature deciduous trees have their roots within the structure and are something of a threat to the survival of the walls. Burnard noted that the largest tree, sited within the partition wall had (in 1888) a girth of 5%ft [1.7m]; he considered it to be of great age (Burnard 1887-90, 105). A sherd of unglazed pottery was discovered in this mill wall by Tom Greeves as a chance find in the 1970s (Greeves 1981, 92). The ware type was then identified as having a date range Week Ford Mill A 0 metres 5 Fig. 3. Week Ford; survey of Mill A. 191 Plate 2. Week Ford: general view of Mill B trom the north-east. from the end of the 15th century to c. 1550 (Allan, pers. comm.), which potentially offers an earlier working date for the Week Ford mills than the documentation would suggest, although without a reliable association such may only be surmised. MILL B (Pl. 2; Fig. 4) The upper mill stands 10m upslope from, and to the south of, Mill A and is an irregular foursided structure, levelled into the slope of the hillside, with the south rear wall acting as a revetment. The internal dimensions are 4m north-to-south by 6.3m and 5.3m east-to-west. The walls are constructed from large undressed moorstone and stand to a maximum height of 2m. They are on average lm thick though the western wall is battered outwards, measuring 2.3m at base. The floor of the interior is strewn with fallen masonry. There is a granite-lined recess, cut into the south wall, 1.3m wide by I m deep. This could have been a fireplace or, alternatively, an ore-chute into which cassiterite could have been tipped from barrows, making use of the higher ground level to the rear of the building. Within the recess lies a large fallen slab 0.6m wide, into which a lateral groove has been cut. Its purpose is unknown. A second large slab measuring 1.2 x 0.8m lies on the ground on the exterior of the recess, level with the top of the wall. The entrance is in the northern wall and is 0.8m wide. The western door jamb has a vertical groove which would have accommodated a wooden door lining. A flat-topped stone survives in place across the threshold. An exceptionally well-preserved wheelpit occupies the entire width of the exterior of the eastern wall. The wheelpit is 0.6m wide internally and survives to a depth of 1.6m. The south192 Week Ford MillB ht 2m \\ o 0o entrance ///JJ/// slab JI 0 metres 5 Fig. 4. Week Ford: survey of Mill B. east corner of the wheelpit has at some time been deliberately breached to allow water to flow through it from a leat which meets the wheelpit at an acute angle from the east, although it is unlikely to have been for the purpose of supplying this wheel (see below). The area immediately south and to the rear of the wheelpit is strewn with large fallen boulders which formed the final support for a wooden launder, carrying water onto the overshot or backshot wheel. The main teat embankment stands 5m south of the wheelpit. It is an earthen bank 13m long, now overgrown with gorse and heather, with traces of walling surviving on the exterior to the west. The bank is 1.5m wide, standing to 0.7m high at the lower end. A linear hollow to the east of the embankment probably marks the source of material from which it was built. 193 Ems ,.. a 'd 'w' "•4t' j7 ~~ ,r z a"~ &'~d '&' Sc Plate 3. Week Ford: three-holed mortarstone lying outside the entrance of Mill A. MORTARSTONES (PI. 3) Detailed measurements and descriptions of all the mortarstones at Week Ford would not be particularly useful in this survey. However the positions of all those mortars visible when the survey was conducted are marked on Figs 3 and 4. A total of 13 stones show evidence of having been used as mortars, including seven with three hollows on at least one face, indicating that three heads of stamps were used at the site at one time. Five of the mortarstones are within the structure of Mill A, with a further two just outside the eastern entrance. Another five lie scattered to the north of Mill B with a sixth positioned inside the wheelpit. It is notable that a large three-hollowed mortarstone recorded in the lower mill by Tom Greeves in 1975 (Greeves, pers. comm.) is no longer visible. LEATS AND OTHER ASSOCIATED EARTHWORKS The waterwheels of the mills were supplied via two leats (Fig. 1). The upper, southern, Teat is on the whole the better-preserved, being visible over its entire course. Its position demonstrates quite clearly that it was cut to direct water to the teat embankment of Mill B. The channel may be traced up to a point on the 0 Brook where the headweir would have been sited at SX 66277212, though this has now been destroyed. The lower Teat, although clear over the section close to the mills, has been effaced on the steeply-sloping section of hillside near the river. As the channel approaches the south-west corner of Mill B, it turns to the north where a breach in the corner of the wheelpit allowed 194 water to flow through the wheelpit and continue to the wheel of Mill A. via a short channel which runs between the Mill B wheelpit and the top of the scarp above Mill A. Thus it seems unlikely that this leat ever provided water to power the wheel of the upper mill. The tailrace of Mill B consists of a channel with steeply-cut sides, up to 0.8m deep, aligned with the wheelpit and running to the north, parallel with the structure of Mill A. A shallow gully running downhill on the eastern extremity of the site is likely to have conducted water to the dressing floors and any buddies which may have existed. Although no longer visible, these were probably sited on the level area to the north of Mill A, now overwhelmed by bog. DISCUSSION Experience has shown that attempts to establish a relative chronology at tin-milling sites on the basis of field evidence alone is difficult if not impossible. Excavations at Colliford (Austin et al. 1989) and currently at Upper Merrivale have revealed complex, multi-phase activity, far beyond anything imaginable from field evidence. At Week Ford the field evidence is ambiguous and two possible scenarios could be argued for the building sequence of the mills. The fact that the tailrace of Mill B makes such an obvious detour around Mill A could suggest both that Mill B was the later built and that Mill A was still in use in some form when Mill B was constructed. However, an equally convincing opposing argument could be put forward for Mill B as the earlier building, since the leat of Mill A breaches the wheelpit of Mill B and, along with Mill A which it supplied, thus appears to be of later date. If so, it would have been necessary to re-dig the tailrace from Mill B to avoid the new lower mill. However, despite the lack of clarity for the earliest phase, the existence of independent water supplies for each mill suggests both could have functioned together during later phases. Given that the furnace survives so well in Mill A. it seems very likely that this building was functioning as a blowing-mill in its final phase of activity. A fragment of broken mouldstone built into part of the furnace suggests also that the furnace was rebuilt at least once, using material from an earlier smelting phase. Moreover there are many heavily-worn mortarstones within Mill A which, although disturbed by Burnard, could indicate alterations to the fabric of the building. Recent excavations at Upper Merrivale. a site which has many similarities to Week Ford, have demonstrated that mortarstones within the interior tumble of tin mills are likely to have been used as building material where the fabric of the building has been altered or rebuilt after the mortarstones have been discarded and are so positioned because of the inward collapse of the walls within the surviving structure. They are not necessarily associated with a final stamping phase (Gerrard and Greeves 1992, 4). The interior of Mill B contains no visible mortarstone but several lie scattered around the exterior. One at least still lies in the position where Burnard observed it over one hundred years ago inside the wheelpit. There is no evidence of a furnace within this building, or of a mouldstone, and given the fine condition of the structure and that the floor debris appears comparatively shallow, one would expect such features still to be visible had they existed at the time of abandonment. It is still possible, however, that further artefacts lie obscured within the tumble. Despite the conflicting evidence for the building sequence, it should be noted that Mill A is in the better position. Being lower, it could make more effective use of the water supplied by the 0 Brook. It also appears to represent more phases of activity than Mill B with at least one structural rebuild and a refurbishment of the furnace, and perhaps a change of function. On this evidence it seems likely that Mill A was built first, though we cannot be certain of its function as a blowing-mill in its earliest phase. The upper mill (B), is not so well-positioned as Mill A, being higher, but is sited in the only possible location which would allow both mills to work simultaneously, using water from the upper and lower leat. The 0 Brook is a vigorous and fairly reliable water supply and should have been capable of providing sufficient water to do this. The mill contains no real evidence for smelting and was probably functioning as a 195 stamping-mill in its abandonment phase. It is worth noting that in the document of 1608 the mills at Week Ford are referred to in the plural, suggesting that both were in use at the same time (DuCo London, Dartmoor Proceedings 1203-1735, fol. 29). Our only hint of a date for either of the mills comes from this documentation although Greeves' pottery find could push it back to the early or mid 16th century. Abandonment is likely to have been during the 18th century or earlier, particularly for Mill A, as Burnard observed the oak tree growing from the eastern wall was already large and `ancient' by 1880 (Burnard 1887-90, 105). The radiocarbon date In 1989 Bryan Earl published a note on Week Ford in which he presented a calibrated radiocarbon date of AD 640 to 800 (60%) or AD 570 to 890 (95%) from a sample of charcoal, probably oak or ash, taken from the site (Earl 1989, 119). The sample had been collected from a peat bank to the north of Mill A, adjacent to the river, where a natural section has been created by water erosion and in which deposits of charcoal, tin slimes and slag have become compacted. The peat bank coincides with the position of the wheelpit tailrace from Mill A, the intervening section between the mill and the river bank now being covered by bog. The apparent association of the charcoal with pieces of slag also lodged in the bank led Earl to claim that smelting was taking place at Week Ford in the 8th century — a claim which Greeves considered to be ill-advised owing to the amount of disturbance which had taken place in the area, particularly during the creation of the tailrace, invalidating any association between the slag and the charcoal (Greeves 1990, 45-6). Moreover there are other problems in accepting this date, not least the fact that only one sample was dated.2 We also have to allow for the longevity of oak trees on Dartmoor. It has already been noted by Burnard that the tree growing in Mill A is of great age. If Earl's sample had come from the heartwood of a similar tree, possibly 300 years old, it could theoretically have been felled in the 13th century. Such doubts could be disregarded if several samples had been dated, giving a satisfactorily close spread of dates. Even in the event of this happening, however, the case for 8thcentury smelting activity and a thousand years of continuous tin-smelting at Week Ford, as claimed by Earl, is optimistic in the extreme as the charcoal was not reliably associated with any phase of activity, structure or artefact and was not retrieved using a recognised technique, which could have established any such association. NOTES 1. This stone was last reported as being housed in Totnes Museum where an anonymous caller had handed it in, fearing its safety from souvenir hunters (Masson Phillips 1986, 14). 2. The radiocarbon user's maxim `one date is no date' (Aitken 1990, 95) must surely apply here. ACKNOWLEDGEMENTS Many thanks to Mr Anton Coaker for kindly granting permission to undertake the survey in 1991. I am grateful to Dr Tom Greeves for drawing my attention to the Duchy of Cornwall references and for giving me access to his own field notes. Thanks also to Paul Glanville who gave me assistance with the survey. BIBLIOGRAPHY Unpublished sources DuCo Duchy of Cornwall: Dartmoor Proceedings 1203-1735 DRO Devon Record Office: Accessions Lists Printed sources Aitken, M.J. 1990 Science-based Dating in Archaeology. Anon. 1891 `Ancient Mining Remains on Dartmoor', Mining J. 22 Aug 1891, 947. Austin, D., Gerrard, G.A.M. & Greeves, T.A.P. 1989 'Tin and agriculture in the Middle Ages and beyond: landscape archaeology in St Neot Parish Cornwall', Cornish Archaeol. 26, 7-251. Bray, A.E. 1836 The Borders of Tamar and Tavy Vol. 3. Burnard, R. 1887-90 `On the Track of the "Old Men" of Dartmoor. Parts I & 2', Trans. Plymouth Institute 10, 95-112, 223-42. Crossing, W. 1891 'Crockern Tor and the Ancient Stannary Parliament', Western Antiq. 11 (Aug—Sept 1891), 9. Earl, B. 1989 `Tin Smelting at Week Ford, Dartmoor; a brief note', J. Hist. Metallurgy Soc. 23(2), 119. 1990 `Eighth-century tin smelting on Dartmoor; do we really have the evidence?' [reply to letter from Tom Greeves], J. Hist. Metallurgy Soc. 24(1), 47-8. French, H. & Linehan, C.D. 1963 `Abandoned Medieval sites in Widecombe-in-the Moor', Rep. Trans. Devonshire Ass. 95, 168-79. Gerrard, S. & Greeves, T.A.P. 1992 `Summary report on the excavation of Upper Merrivale tin blowing and stamping mill (SX 55197664)', Dartmoor Tinworking Res. Group Newsletter 2, 44. Greeves, T.A.P. 1981 `The archaeological potential of the Devon tin industry' in Crossley, D.W. (ed.) Medieval Industry, Counc. Brit. Archaeol. Res. Rep. 40, 85-95. 1986 Tin Mines and Miners of Dartmoor. -- 1990 `Eighth-century tin smelting on Dartmoor: do we really have the evidence?', J. Hist. Metallurgy Soc. 24(1), 45-6. Masson Phillips, E. 1986 [Note on mortarstones], Devon Archaeol. Soc. Newsletter 34, 14. Parsons, H. 1956 `The Dartmoor Blowing-House: some recent investigations', Rep. Trans. Devonshire Ass. 78, 189-96. Rowe, S. 1985 A Perambulation of Dartmoor. (Facsimile of 1896 edn.) Worth, R.H. 1940 `Notes on some Dartmoor Blowing-Houses', Rep. Trans. Devonshire Ass. 72, 201-8. N.B. Week Ford is sited on private land; there is no public access. This paper is published with financial assistance from the Dartmoor National Park Authority and the Council for British Archaeology. 197