SECTION 3.8:CLAYDON PIKE WORKED STONE by Fiona Roe
3.8.2: THE WORKED STONE FROM LONGDOLES FIELD 3.8.1 WORKED STONE FROM WARRENS FIELD (Fig. 3.8.1: Worked stone from Warrens Field )The worked stone from Warrens Field amounts to a dozen pieces, which have been summarized in Table 1, and described in the catalogue. Ten finds are from middle Iron Age contexts, while two unphased fragments probably also belong with the Phase 1 material. QuernsThe emphasis is very much on querns or quern materials, the only exception being a possible slingstone (1016). Both saddle and rotary querns were in use, though with more evidence for the use of saddle querns. Parts of two of these have survived (56 and 291), one of which is apparently a re-used piece from a larger quern. Its small size suggests that it was intended for a child, to use for grinding corn in order to practice an essential skill, and also no doubt to help out with the daily tasks (56, Fig. 3.8.1, no.1). There are also fragments from three rubbers for use with saddle querns 28, 57 and 154, one part of a hog-backed example, a typical Iron Age variety of rubber (28, Fig. 3.8.1, no.2). These, together with four fragments of saddle quern material 5643, 25, 276 and 294, account for two thirds of the Iron Age assemblage. Of note are two pieces from rotary querns. One of these (176, Fig. 3.8.1, no.3). was probably once part of a somewhat thick upper stone, of which the surviving depth is now 105 mm. It was a well made example, not particularly large, with a diameter of approximately 300 mm, and a rim that was carefully pecked to shape. The second piece of rotary quern (71, Fig. 3.8.1, no.4) is part of a lower stone that was also fairly thick, with a present depth of 85 mm. The diameter was about 330 mm. The stone used for the Warrens Field querns was nearly all imported to the site (Table 2). The local resources at Claydon Pike were limited to pebbles of hard quartzitic sandstone or quartzite from the gravels of the area, which were used for just one small rubber from S 7, and also for the possible slingstone. The Jurassic limestone lies in a band to the north of the site, with Oxford Clay to the south, so that it would have been necessary to bring in serviceable grinding stone from beyond these two areas. Nearly all the quernstone came from the same direction, either from the Silurian sandstone of May Hill 51.5 km (32 miles) to the north west, or from the Upper Old Red Sand stone of the Forest of Dean, some 64.4 km (40 miles) away. Two possible querns, represented now only by fragments (25 and 2643), were made from Lower Greensand with a source around Culham in Oxfordshire, and so these came from the opposite direction, from some 37 km (23 miles) down the river Thames. This greensand was very much a saddle quern material. The May Hill sandstone was also a traditional saddle quern material, and finds of this amount to one saddle quern (Fig. 3.8.1, no.1) two rubbers (Fig. 3.8.1, no.2 and 154) and a worked fragment (276). By contrast, the Upper Old Red Sandstone was more widely used for rotary querns, although saddle querns are not unknown and one occurred at Warrens Field (291). There are also two pieces of rotary quern made from Old Red Sandstone, representing part of an upper stone (Fig. 3.8.1, no.3) and part of a lower stone (Fig. 3.8.1, no.4). Another fragment of Old Red Sandstone (294) could not be further identified as to type. DiscussionThe contexts in which the middle Iron Age quern fragments were found are varied, and all suggest deposition of the most casual kind (Table 2). The most common occurrence is in ditches, with two finds from house gullies, and two others from clay-lined pits. Saddle quern fragments came from all three islands (Table 2), and so appear to have been in use throughout the life of the settlement. One rotary quern fragment (176) is from Island 3, while the other (71) is from Island 1. Island 3 is considered to be the earliest of the three (G. Edgeley Long, report on MIA pottery), so it would seem that the new technology for grinding corn with rotary querns was known quite early on at Warrens Field. However the changeover to rotary querns appears to have been gradual, with saddle querns continuing in use for some time. Thus Lower Greensand from Culham, which had long been used for saddle querns, was found in a house gully on island 1. The old ways persisted, whether from unwillingness to change the habits of thousands of years, or from lack of opportunity. The worked stone from Warrens Field can be compared with finds from other sites in the locality. The middle Iron Age assemblage from the nearby site at Horcott Pit, Fairford includes saddle querns made from both May Hill sandstone and Old Red Sandstone from the Forest of Dean (Oxford Archaeology, in prep (a)). The same two materials were also found in use for querns at the middle Iron Age site at Preston, Gloucestershire (Roe 1999 (a), 416). Finds of querns made from May Hill sandstone are common generally in the area, and are known from Thornhill Farm (Shaffrey 2004) and further sites in Oxfordshire such as Hatford Quarry (Oxford Archaeology, in prep (b) and Bampton (Ashmolean Museum). The Upper Old Red Sandstone has on the whole been less frequently recorded, but a fragment found at Gassons Field, Lechlade is from a late Bronze or early Iron Age context (Roe 1998). The small rotary quern found unstratified at Sherborne House, Lechlade may, if not Saxon in date, belong to the Iron Age occupation there (Roe 2003 (a), and so join the growing numbers of Old Red Sandstone querns found in later prehistoric contexts either along or south of the Thames. The fragments of Culham greensand are the only examples of this stone known to date from Gloucestershire; most finds are from Thames gravels sites, and have been recorded mainly from Oxfordshire (Roe, in prep). 3.8 2 THE WORKED STONE FROM LONGDOLES FIELDObjects (Figs 3.8.2: Worked Stone from Longdoles Field Phase 3, 3.8.3: Worked Stone from Longdoles Field Phase 3 and 3.8.4: Worked Stone from Longdoles Field Phase 4)There was heavy reliance on imported stone at Longdoles Field, as shown by the summary of 118 objects (Table 3), since 100 (85%) of the stone finds were brought to Claydon Pike from outside the area. By contrast, Cotswold stone would have been transported over only medium distances and there are 13 objects made from this limestone, most of which appear to be re-used pieces of building stone. Jurassic limestone was suitable in particular for making into different kinds of vessels. Only 5 objects are of materials that could have come from the immediate locality, but the site lies on a gravel island, which cannot have been a productive source of usable stone. A large proportion of the worked stone objects are artefacts for grinding, either querns and millstones or whetstones, which were often used additionally as point sharpeners. There are few spindlewhorls, only 2 of which were made from stone, with 3 of Kimmeridge shale, and another 7 made from fired clay. One third of the stone objects was unstratified, one of these being a stone axe which is likely to be of prehistoric date. Querns and millstonesThe querns and millstones found at Longdoles Field are nearly all now in a fragmentary state, and some 43% are unstratified. However even small pieces can usually be assigned to a specific source. It can be seen from Table 3 that Upper Old Red Sandstone from the Forest of Dean/Wye Valley area was particularly important as a quern and millstone material. There are also numerous whetstones made from Lower Old Red Sandstone, described below, so there could have been strong links with the Forest of Dean. The Upper Old Red Sandstone comprises two interbedded varieties of stone, sandstone and conglomerate (Welch & Trotter 1961, 49), and both were widely utilised. The quartz conglomerate contains pebbles which are mainly of white vein quartz, and these give the stone a distinctive appearance (eg 592, Fig. 3.8.4; 2981, Fig. 3.8.2). The rotary querns and millstones were manufactured by pecking into shape, and neat pecking round the rim is characteristic of the examples from Longdoles Field. The grinding surfaces of rotary querns were also prepared by pecking, although they often later became worn into concentric rings (2926, Fig. 3.8.2). Six upper stones could be identified, and another six lower stones, five of which were fully pierced (eg 2926 Fig. 3.8.2; 870 Fig. 3.8.4). At least six of the pieces that were found can be identified as millstones, on the basis of an estimated diameter of up to circa 750 mm (eg 3062, 2984, Fig. 3.8.2), or a slot for a rynd fitting (2575/2927). A nearby stream could have supported a watermill. Over one third of the Upper Old Red Sandstone was unstratified, but it is likely that it was in use throughout the Roman period, having already occurred in Iron Age contexts at Warrens Field. The Old Red Sandstone querns and millstones were transported to Claydon Pike some 64.4 km (40 miles) from the Forest of Dean, but they are in no way unusual, since they are found on all other Roman sites in the area (Saunders 1998). They have also been recorded, for instance, at the neighbouring sites of Thornhill Farm (Shaffrey 2004) and Roughground Farm (Allen 1993, 160 & Ashmolean Museum). Finds from nearby Roman towns include Old Red Sandstone both from older and more recent Cirencester excavations (Corinium Museum; Cotswold Archaeology, in prep) and from Wanborough (Buckley 2001 & Swindon Museum). Upper Old Red Sandstone, mainly in the form of rotary querns, has also been widely recovered from Roman sites in Oxfordshire (Roe in prep). Millstones made from Upper Old Red Sandstone are also known from other sites, and can be impressive, as is demonstrated by a pair of complete upper stones with grooved grinding surfaces from Woolaston, Gloucestershire (Scott Garrett 1938, Pl. IIB: Watts 2002 58 & Fig 21). It is surprising however how much Millstone Grit was also brought to Longdoles Field, coming from the Pennines around Sheffield, some 188 km (117 miles) north from the site. Only one rotary quern made from Millstone Grit has been identified (3009), and it appears that it was being utilised primarily as a millstone material. The finds from Longdoles Field have not survived in good condition, and although 6 pieces could be identified as coming from millstones, another 7 are of uncertain type. In four cases traces of pitting could be seen, this being the typical method for finishing off querns or millstones made from this variety of stone. Evidence from other local sites supports the theory that Millstone Grit was brought into Gloucestershire mainly as a millstone material. A large example of one of these millstones can be seen at the Chedworth Roman villa (Watts 2002, 62 & Pl. 9). Pieces of another large millstone were found at Frocester Court (Price 2000, 195 & Gloucester Museum), and part of a further one came from Cricklade Street in Cirencester (Barber & Walker 1997, 9 & Corinium Museum). In addition there is a probable millstone fragment from Somerford Keynes (Section 5.3) and fragments of Millstone Grit were found at Roughground Farm (Allen 1993, 160 & Ashmolean Museum). These millstones would have been valuable pieces of equipment, and one found at Wanborough, which was one metre in diameter, had been repaired with lead (Buckley 2001, 160 & Fig 62, 16). It is clear that although Millstone Grit is found in smaller quantities than Upper Old Red Sandstone on Gloucestershire sites, the trade from the Pennines must have been an important part of the Roman economy. Niedermendig lava does not survive well on gravel sites, and all the pieces found at Longdoles Field are all small fragments. There is some evidence from other Gloucester sites to show that it was used both for rotary querns and for millstones. A complete lower stone from a lava millstone was found at the Wortley Roman villa (Taylor & Bagnall 1989, 43 & fig 14, 121). Some lava rotary querns are known from local Roman sites, as for example from Stepstairs Lane, Cirencester (Cotswold Archaeology, in prep), and from Wanborough (Buckley 2001, 156), but it is more usual for unidentifiable, weathered fragments to be found, as at Longdoles Field. However, the number of Roman sites in Gloucestershire where lava has occurred (some 16 at the time of writing) suggests that it may have been quite widely used. There are only two querns made from Jurassic sandstone or limestone from the Cotswolds, and these may have served as a stand-by if supplies of imported quernstone failed to arrive when needed. Whetstones and point sharpenersThe large numbers of whetstones, especially from Phase 3, are unusual. They have been summarized in Table 4, which shows that there are 16 from phase 3 alone, and another 16 unstratified examples. These seem to hint at a rural economy in which they may have been needed for sharpening tools such as sickles or scythes, and two 2 sickles were in fact found in Phase 2 contexts. Thus the whetstones fit with the suggestion that haymaking was an activity particularly characteristic of this site. Two distinctive varieties of whetstone were found at Longdoles Field. Firstly, there are whetstones of rod type, and these are all made from light-coloured sandy limestones (eg 1030, 2141, Fig. 3.8.2). The second variety consists of whetstones of slab type, and these were all made from sandstones, and usually ones that are dark red in colour and slightly micaceous. The whetstones of the rod variety are mainly made from Kentish Rag, which is likely to have come from around Maidstone (1030, 2141, Fig. 3.8.2), but three are made from fine-grained, sandy limestone from the Cotswolds. Most of these whetstones are now well worn to a characteristic cigar shape (2141). Three of the Kentish Rag ones retain traces of grooves from the initial shaping of stone (1030), but this is not a particularly uncommon feature. Whetstones of Kentish Rag first appear in Phase 2, but the majority of finds are from Phase 3. After this they appear to tail off, although there are a few undated ones which could perhaps belong in Phase 4. However it does seem as though Kentish Rag whetstones were not freely available during the later part of the settlement, and whetstones made from Cotswold stone were apparently substituted. These are made from the same fine-grained Jurassic limestone that was used for some of the roofing tiles, and so probably came from the Great Oolite at no great distance from the site. The whetstones of the slab variety are mainly made from purple-brown Lower Old Red Sandstone tilestone from the Forest of Dean (Welch & Trotter 1961, 33). These tend to be unevenly shaped, relatively thin, re-used fragments, although at least two have been worn into a pyramidal shape (474, Fig. 3.8.4; 371). Two more whetstones appear on macroscopic examination to be Pennant sandstone (2979, 2747), another tilestone, and this may also have been obtained in the Forest of Dean, although the Bristol coalfield is another possible source. A few others were made from red sandstone of less certain origin which may however also have come from the Forest of Dean. These include as a Phase 2 example which is in the form of a more regular, rectangular block worn along the sides (1058). Four other whetstones/point sharpeners were made from light coloured quartz sandstone likely to be Jurassic sandstone with a more local source. The flat whetstones of Old Red Sandstone tilestone were used in a different way from the Kentish Rag ones, with wear on the main, flat surface, rather than longitudinally up and down the sides. The broken edges have usually been left unmodified. These whetstones could have been used for more heavy duty honing than the Kentish Rag ones. In addition, a number of the Old Red Sandstone ones were also used as point sharpeners (eg 474). The Lower Old Red Sandstone was in use during both phases 3 & 4, but half of the finds are unstratified, so that the full picture is not clear. A curious fact is that although the Old Red Sandstone whetstones appear to be made from re-used roofing tile, there are no definite examples of such roofing material from Longdoles Field, where all the stone roofing tiles were made from Jurassic limestone. There are only three possible pieces of Old Red Sandstone paving stone from phase 4 (3555, 3563, & 3574). These have been catalogued as building stone, but in fact the wear traces could be the result of some non intensive whetting. Further whetstones made from Old Red Sandstone tilestone were found at the nearby sites of Roughground Farm (Allen 1993, 161 & Ashmolean Museum) and Thornhill Farm (Shaffrey 2004), where again no roofing tiles made from Old Red Sandstone were recorded. However these roofing tiles from the Forest of Dean were being transported to other sites in Gloucestershire, since they were widely used in Roman Gloucester and other sites in that particular area such as Hucclecote (Roe 2003 (b), 51). It could be that broken pieces of tilestone from the source area were not wasted, but were instead traded out, along with querns and millstones, to sites such as Longdoles Field where they could be put to good use as whetstones. Other Roman sites in the region, as for example Kingscote, have some whetstones made from Coal Measures Sandstone, which may have been transported southwards from the Pennines in conjunction with the millstones of Millstone Grit (Gutierrez & Roe 1998 (b), 178). No doubt the Old Red Sandstone whetstones, coming from a nearer source, were less costly, and for this reason might have been preferred at Longdoles Field. Whetstones made from Pennant sandstone appear to be less common in the area, but were noted from Barnsley Park (Corinium Museum), and one was found at Asthall (Roe 1997, 101). These too could be fragments of roofing tile which were put to further use. As was seen with the querns and millstones of imported stone, the materials used for whetstones at Longdoles Field have not occurred in isolation at this site, but are part of a wider picture. The Kentish Rag whetstones, in particular, appear to have been very widely distributed and there is a comparable collection of 14 Kentish Rag whetstones from Barnsley Park (Peacock 1971, 153), while another ten examples were found at Kingscote, (Gutierrez & Roe 1998 (b), 178). Elsewhere in Gloucestershire and Oxfordshire they have frequently been recorded but in smaller numbers, as for instance at Asthall (Roe 1997, 100). Whetstones made from Lower Old Red Sandstone have been found in smaller quantities. Nevertheless, apart from their presence at the adjacent sites of Roughground Farm and Thornhill Farm, they have occurred at Barnsley Park (Corinium Museum) and Asthall (Roe 1997, 100), two further sites where the stone roofing tiles were made from Jurassic limestone (Williams 1971, 101; Booth 1997, 102). There are also examples from Somerford Keynes (section 5.3). There are further objects of imported stone from Longdoles Field which again demonstrate how certain lithic materials were being widely distributed during the Roman period (Table 5). Purbeck Marble from Dorset was one such variety of stone, being used especially for mortars and other vessels, but also on occasion for palettes. The fragment from Longdoles Field (2838, Fig. 3.8.2).) appears to be the first palette of this material to be recognized from Gloucestershire, where other known examples were made from imported Mediterranean marbles, as was the case for instance at Kingscote (Gutierrez & Roe 1998 (a), 167). However a quantity of Purbeck Marble was recovered from Roman Cirencester (Corinium Museum), and so the palette, or a piece of stone from which to make it, could have been acquired from the market there. Kimmeridge shale was also brought from Dorset, and there are fragments from 11 small objects. These amount to 3 spindlewhorls (423, 603, Fig. 3.8.2), 6 small bracelets, four of which are plain (76, Fig 3.8.3) and 2 decorated (660, 946, Fig 3.8.3), together with a ring and a small bead (2338, Fig 3.8.3). This collection is typical of what has been found on other Gloucestershire sites, as for instance at Barnsley Park (Corinium Museum), Frocester (Price 2000, 185 and Kingscote (Timby 1998, 220). Cotswold stone lends itself well to the manufacture of stone vessels, and three were found at Longdoles Field. A shelly variety of the Great Oolite was used for a well-made dish (1624, Fig 3.8.4), and this may have come from around Coln St. Aldwyns in Gloucestershire, along with limestone roofing tiles. Two other pieces may come from informally made mortars, one (4573, Fig. 3.8.3) apparently made from a re-used piece of building stone from the Corinium quarries, the other (2448, Fig 3.8.3) probably made from a limestone cobble which could have been found in the local gravels. Comparable limestone mortars were found at Somerford Keynes (section 5.3), and another came from Thornhill Farm (Shaffrey 2004).One spindlewhorl is made from a fine-grained limestone, which again is similar to stone used for roofing tiles, with a probable source in the Great Oolite. Pieces of chalk from 4 contexts, one of them shaped (3205) are of unknown use, but may have been brought to the site for craft work. A small cylindrical bead (331) of red, banded, siliceous stone is probably jasper, but the source is unknown. The few objects made of stone from the immediate locality demonstrate the limited local resources. Small flint pebbles which probably came from the river gravels were used for a possible marble and a possible counter (2336, 2864, Fig 3.8.3). A quartzite pebble, also probably from the gravels, was used for a small weight of Iron Age type (2572). Fine-grained calcareous mudstone from the Oxford Clay or Kellaways Beds was used for a spindlewhorl and a slab that may have been some kind of smoother (361, 750). A stone axe (199) was unstratified, but probably relates to earlier prehistoric activity in the area. Macroscopic examination has suggested that it is made of the Group VI tuff from Great Langdale in Cumbria, an axe material that was very widely used. Langdale axes account for almost one third of the stone axes found in Oxfordshire (Roe 1999 (b), 228) and for nearly half of all the grouped axes from Gloucestershire (Davis et al., 1988, 16). Building stone (Fig. 3.8.5: Structural stone from Longdoles Field)The building stone, summarized in Table 6, is a collection of roofing tile fragments, architectural pieces and samples of building stone, amounting to 62 items. Two main sources of Jurassic limestone were used. The roofing tiles are made from fissile varieties of the Great Oolite, which were probably obtained from around Coln St Aldwyns, Gloucestershire (Richardson 1933, 106). The limestone used for columns and other shaped pieces of masonry probably came from the Roman quarries at Corinium (McWhirr et al. 1982, 35). Nearly half of the building stone consists of pieces of limestone roofing tile, with just a couple of complete examples (eg 5840, Fig. 3.8.5). Ceramic roofing tile was also found at Longdoles Field, but some stone roofing was already being used during Phase 3. The majority of the finds, however, represent pieces of roofing tile discarded during Phase 4. Nearly all the fragments that were kept are diagnostic pieces with holes in them, and one of these, found unstratified, still has the iron nail in place (5847). All the stone roofing tiles were made from shelly varieties of the Great Oolite, probably the Forest Marble (Sumbler et al. 2000, 68). The shell fragments, lying parallel to one another, would have caused the limestone to divide easily into usable slabs, which however were often relatively thick, so that the roofing tiles were weighty. One not quite complete roofing tile (5831) weighs 3.0 kg, and another complete hexagonal tile (5842) weighs 2.375 kg. The limestone is variable in character, ranging from a fine-grained variety consisting of many small shell fragments to coarse-grained varieties which may contain large pieces of fossil shell (eg 5842). It should have been possible to obtain the full range of tilestone in one quarry. It is not possible to say exactly where this may have been, but such bulky items would not have been transported any further than was necessary. Comparable limestone roofing tiles were quarried in more recent times around Coln St Aldwyns, Gloucestershire (Richardson 1933, 106). Two Roman sites are known here, on either side of the river Coln where it is crossed by Akeman Street (RCHM 1976, 37, 97), and the river could have been used to transport the tilestone towards Claydon Pike. However the distance involved is not great, amounting to around 7.8 km (4 or 5 miles), so that a journey by country road should have been no great problem. The architectural stone from Longdoles Field amounts to parts of 4 columns (Fig. 3.8.5: 472, 2015, 2016), along with 2 column bases (Fig. 3.8.5: 2928, 2014). There are also 8 pieces of shaped masonry (Fig. 3.8.5: 3061, 2585) and four further unidentifiable fragments. A good quality limestone would have been needed for shaping into columns and masonry, and this was available in the Corinium quarries 17.5 km (11 miles) to the west of the site. This limestone has proved difficult to place within the local Jurassic sequence (Richardson 1933, 49), but appears to belong near the top of the White Limestone, or the lower part of the Forest Marble (Sumbler et al. 2000, 67). It is both oolitic and shelly, pale coloured when weathered, but a creamy shade when fresh. Some small fragments occurred in four Phase 3 contexts, but the main use of this quarried stone must have been for the Phase 4 villa. Although this villa was modest in size, it was still apparently adorned with columns, and not all plain ones, since some had carefully moulded bases (Fig. 3.8.5: 472, 2014). Two columns could be measured (471, 472) and were found to have diameters of 188 & 200 mm, which fall within the size range suggested by Blagg as suitable for domestic buildings (2002, 189). A column more slender than these two (Fig. 3.8.5: 2016) could have been part of a colonnade or veranda, while those slightly larger in size (eg Fig. 3.8.5: 472, 2014) could have belonged to a porch or entranceway. The corridor wall that was built during Phase 4 to join buildings 8 and 9 seems the most likely candidate for the positioning of a colonnade. The samples of building stone suggest that anything suitable may have been collected in the general local area and used for the stone structures at Longdoles Field, whether for walls of random rubble construction, or for mixed rubble infill. Pieces of Jurassic limestone appear to have been the main component. Some Jurassic sandstone was also used, and this probably came from the Kellaways Beds, which are known to contain sandstone doggers at South Cerney (Torrens 1982, 77). Three fragments of Lower Old Red Sandstone with slight wear traces may have belonged to paving stones. However there was no evidence of paved flooring from the villa, and these pieces from Phase 4 contexts (3555, 3564, 3574) may have been intended to be used for whetting. Roman building stone was in general selected on very much a local basis, so that wide comparisons with other sites cannot be expected. The vast Roman quarries at Cirencester must have been employed mainly to provide building stone for Corinium. A plain column very similar to one from Claydon Pike (471) was noted from The Avenue, Cirencester (Corinium Museum). Ready made pieces such as this may have been exported from Corinium, and no doubt both the stonework and the transport for them were costly. Although the villa at Longdoles Field was modest in size, money was evidently found for some architectural features that would proclaim the importance of those living there. The villa (or farmstead?) at Barnsley Park is only 7 km (4.4 miles) from Corinium, and yet here they obtained no such domestic adornments. However pieces of this quarried limestone were also found at the nearby sites of Roughground Farm (Allen 1993, 161 & Ashmolean Museum) and Kempsford (Oxford Archaeology, in prep (c), while at Somerford Keynes some of the same stone had been utilised for carved monumental stone (section 5.3). At Wanborough too there were 2 pieces of shaped masonry (Blagg 2001, 153 & Swindon Museum), which could have come from the same source. At none of these other sites were limestone columns recorded, but stone from the Corinium quarries seems to have been available for other purposes, for those willing and able to pay for it. Limestone roofing tiles comparable to those from Claydon Pike are also limited very much to a local distribution. Further examples were found at Roughground Farm (Allen 1993, 161 & Ashmolean Museum), with further fragments from Kempsford (Oxford Archaeology, in prep (c)). 3.8.3 GENERAL DISCUSSIONThe Iron Age worked stone from Warrens Field is best viewed not as an entity in itself, but as part of a much wider picture. The finds fit well with what is known of Iron Age Gloucestershire (Roe 1999 (c), 109) and Oxfordshire (Roe in prep), but include materials that were also distributed beyond the boundaries of these two counties. The May Hill sandstone has been found on Iron Age sites in both Worcestershire (Roe 2000, 102) and Warwickshire (Warwick Museum), while the distribution also extends eastwards across Oxfordshire as far as Abingdon. What is most striking about this distribution, though, is the fact that, as at Warrens Field, the querns are frequently associated with Malvernian pottery, and particularly that of the B1 variety (tempered with Palaeozoic limestone), and also with Droitwich briquetage. These three commodities, together with smaller quantities Upper Old Red Sandstone, appear to have been part of an extensive and well organized trade network. These goods, perhaps together with more perishable items, must have been moved across the Cotswolds, and the use of wheeled vehicles or packhorses over established routes seems likely. There is a contrast with the distribution of saddle querns made from Culham Greensand. The fragments from Warrens Field are the only ones recorded from Gloucestershire, since all the other known findspots are in Oxfordshire or Berkshire (Roe in prep). Nearly all these querns could have reached their destinations by river, from a starting point close to the Thames. The position of Warrens Field within reach of the Thames may account for the presence of these querns here. The Roman assemblage is quite different from that found at Warrens Field, where saddle quern materials such as May Hill sandstone and Culham greensand predominated, although the beginnings of a changeover to rotary querns made from Upper Old Red Sandstone could be detected. The main Roman quern material, Upper Old Red Sandstone, came from much the same area as the May Hill sandstone had before, and so the organization for its distribution may already have been in place. Other non perishable commodities that could have been part of the trade network include coal, iron and some Lower Old Red Sandstone and Pennant sandstone. The Upper Old Red Sandstone was now being carried much further from the source area. It was the main quern material at Silchester (Shaffrey 2003, 147), and the querns even reached some sites in Northamptonshire, Essex and Surrey. Whetstones made from Lower Old Red Sandstone were also found in some numbers at Silchester (Wooders 2000, 388). Millstone Grit and Niedermendig lava are typical of Roman sites in England generally, and so are in no way unusual. Both occur in some quantity on sites in eastern England, as for instance in Kent (Roe 1999 (d), 29), where transport by sea would have greatly facilitated the distribution. The Thames would have played a part in the distribution of Niedermendig lava, but use of this important routeway in the Roman period was nothing new. Smaller objects, such at the Kentish Rag whetstones, Kimmeridge shale and stone palettes could have been carried around much more easily by pedlars, and also sold in markets, so that it is no surprise to find them very widely disseminated on Roman sites. This was a period when nothing was wasted if a further use could be found for it, so it would have been normal for broken pieces of Old Red Sandstone tilestone, leftover from the manufacture of roofing tiles, to be sold on as whetstones. They were probably loaded onto carts along with querns and millstones, in order to make a bit more money out of the enterprise. What is remarkable is the amount of effort that must have been put into the transport of the much larger pieces of worked stone, and especially the millstones, which occur regularly on Roman sites. There is no discrimination between different types of site, since the need for good grinding materials would have been universal, whether at regional centre, a small town, a villa, a farmstead or a ‘native’ settlement. The use of particular materials may to some extent have been fairly arbitrary, because, when a broken quern or millstone needed to be replaced a journey to the nearest market, in the case of Claydon Pike probably one at Corinium, may have been necessary, and the choice of material could have been restricted to what was on the market stalls at that particular time. The efficient organization of supplies of grinding materials meant that a rural ‘cottage style’ villa such as the one at Claydon Pike could, over the years, have made use of all the best lithic materials that were in circulation at the time, including ones for millstones measuring up to a metre in diameter. AcknowledgementsThe work for this report has been dependent on seeing material in museums, much of which has not been fully published. I am very grateful to all the museum curators who made this possible: Tim Bridges at Worcester, Sue Byrne at Gloucester, Christine Edbury & Lauren Gilmour at Standlake, Paula Gentil at Corinium, Guy Kilminster who was then at Cheltenham, Nicholas Palmer at Warwick, Alison Roberts at the Ashmolean Museum, Oxford and Isobel Thompson at Swindon. Thanks are also due to Kathryn Ayres of the Surrey Archaeological Unit for making stone objects available for study. BibliographyAllen T 1993 ‘Stone objects’ in T Allen, T Darvill, S Green & M Jones, Excavations at Roughground Farm, Lechlade, Gloucestershire: a prehistoric and Roman landscape, Oxford Archaeological Unit, Thames Valley Landscapes: The Cotswold Water Park, Vol 1, Oxford Archaeological Unit & Oxford University Committee for Archaeology Anderson A S, Wacher J S & Fitzpatrick A P 2001 The Romano-British ‘Small Town’ at Wanborough, Wiltshire: Excavations 1966 – 1976, English Heritage & Britannia Monograph No 19, Society for the Promotion of Roman Studies Barber A & Walker G 1997 ‘Cricklade Street, Cirencester’, Cotswold Archaeological Trust Annual Review 7 (for 1996), 9 – 10 Blagg T F C 2001 ‘Carved and sculptured stones, clay and metal figurines’ in Anderson et al., 153-5 Blagg T F C 2002 Roman Architectural Ornament in Britain, British Archaeological Reports British Series 329, Oxford Booth P M 1997 Asthall, Oxfordshire: Excavations in a Roman ‘Small Town’, Oxford Archaeological Unit, Oxford Buckley D G 2001 ‘Querns and Millstones’ in Anderson et al., 156-60 Cotswold Archaeology, in prep Report on excavations at Stepstairs Lane, Cirencester (SLC.02) Davis R V, Howard H & Smith I F 1988 ‘The petrological identification of stone implements from south-west England: Sixth Report of the Sub-Committee of the South-Western Federation of Museums and Art Galleries’ in Clough T H McK & Cummins W A, Stone Axe Studies Vol 2: The petrology of prehistoric stone implements from the British isles, C.B.A. Research Report No 67, London, 14-20 Gutierrez A & Roe F 1998 (a) ‘Mixing palettes’ in Timby, 167 Gutierrez A & Roe F 1998 (b) ‘Stone objects: whetstones’ in Timby, 176-8 McWhirr A, Viner L & Wells C 1982 Romano-British Cemeteries at Cirencester, Cirencester Excavations II, Cirencester Excavation Committee Oxford Archaeology, in prep (a) Report on excavations at Horcott Pit, Fairford (FHP.02) Oxford Archaeology, in prep (b) Report on excavations at Hatford Quarry (HATQ) Oxford Archaeology, in prep (c) Report on Excavations at Kempsford (KEMUQ.01) Peacock D P S 1971 ‘Whetstones’ in B Cunliffe, Excavations at Fishbourne 1961 – 1969. vol II: The Finds, Reports of the Society of Antiquaries on London No XXVII, 153 Price E 2000 Frocester: A Romano-British Settlement, its Antecedents and Successors, Gloucester & District Archaeological Research Group, Stonehouse Richardson L 1933 The Country around Cirencester, Explanation of Sheet 235, Memoirs of the Geological Survey of England and Wales, H.M.S.O., London Roe F 1997 ‘Worked stone (except flint) in P M Booth, 100-101 Roe F 1998 ‘Worked Stone’ in R King, ‘Appendix 2, Excavations at Gassons Road, Lechlade, 1993’ in A Boyle, D Jennings, D Miles & S Palmer, The Anglo-Saxon Cemetery at Butler’s Field, Lechlade, Gloucestershire. Vol 1: Prehistoric and Roman Activity and Anglo-Saxon Grave Catalogue, Thames Valley Landscape Mono No 10, 278-9 Roe F 1999 (a) ‘The Worked Stone’ in A Mudd, R J Williams & A Lupton, Excavations alongside Roman Ermin Street, Gloucestershire and Wiltshire: The archaeology of the A419/A417 Swindon to Gloucester Road Scheme, Oxford Archaeological Unit, 414-21 Roe F 1999 (b) ‘Stone axes’ in A Barclay & C Halpin, Excavations at Barrow Hills, Radley, Oxfordshire. Vol I. The Neolithic and Bronze Age Monument Complex, Oxford Archaeological Unit, Thames Valley Landscapes Vol 11, 228-33 Roe F 1999 (c) ‘Quernstones’ in C Parry, ‘Iron-Age, Romano-British and Medieval Occupation at Bishop’s Cleeve, Gloucestershire: excavations at Gilder’s Paddock 1989 and 1990 – 1’ Trans Bristol & Glos Arch Soc 117, 109-10 Roe F 1999 (d) ‘The worked stone’ in A Boyle & R Early, Excavations at Springhead Roman Town, Southfleet, Kent, Oxford Archaeological Unit Occasional Paper No 1, 29-30 Roe F 2000 ‘Worked stone’ in R Edwards & D Hurst, ‘Iron Age Settlement and a Medieval and Later Farmstead: Excavation at 93 – 97 High Street, Evesham’ Trans Worcs Arch Soc 3s 17, 92,102 Roe F 2003 (a) ‘Worked Stone’, in C Bateman, D Enwright & N Oakey, ‘Prehistoric and Anglo-Saxon Settlements to the rear of Sherborne House, Lechlade: excavations in 1997’, Trans Bristol & Glos Arch Soc 121, 68-9 Roe F 2003 (b) ‘The Worked Stone’ in A Thomas, N Holbrook & C Bateman, Later Prehistoric and Romano-British Burial and Settlement at Hucclecote, Gloucestershire, Bristol & Gloucestershire Archaeological Report No 2, Cotswold Archaeology & Bristol & Gloucestershire Archaeological Society, 50-1 Roe F in prep ‘The Worked Stone’ in G Hey et al., Yarnton: Iron Age and Roman Settlement and Landscape, Oxford Archaeology RCHM 1976 Royal Commission for Historic Monuments (England), County of Gloucestershire, Vol 1: Iron Age and Romano-British Monuments in the Cotswolds, H.M.S.O., London Saunders R 1998 The Use of Old Red Sandstone in Roman Britain: a Petrographical and Archaeological Study, Ph D thesis, University of Reading Scott Garrett C 1938 ‘Chester Roman Villa, Woolaston, Gloucestershire’ Arch Cambrensis 93, 93-125 Shaffrey R 2003 ‘The Rotary Querns from the Society of Antiquaries’ Excavations at Silchester, 1890 – 1909’ Britannia 34, 143-174. Shaffrey R 2004 ‘Worked Stone; in D Jennings, J Muir, J Palmer & A Smith, Thornhill Farm, Fairford, Gloucestershire: An Iron Age and Roman pastoral site in the Upper Thames Valley, Thames Valley Landscapes Mono 23, Oxford Archaeology [page numbers to be added] Sumbler M G, Barron A J M & Morigi A N 2000, Geology of the Cirencester District, Memoir for Sheet 235, British Geological Survey, London, The Stationary Office Taylor B & Bagnall A 1989 ‘Finds’ in D Wilson, Excavation of a Romano-British Villa at Wortley, Gloucestershire, Fifth Interim Report, 35-46 Timby J 1998 Excavations at Kingscote and Wycomb, Gloucestershire: A Roman Estate Centre and Small Town in the Cotswolds, with notes on related sites, Cotswold Archaeological Trust, Cirencester Torrens H S 1982 ‘The Geology of Cirencester and District’ in J Wacher & A McWhirr, Early Roman Occupation at Cirencester, Cirencester Excavations I, Cirencester Excavation Committee, 72-8 Watts M 2002 The Archaeology of Mills and Milling, Tempus, Stroud Welch F B A & Trotter F M 1961 Geology of the Country around Chepstow and Monmouth, Explanation of Sheets 233 & 350, Memoirs of the Geological Survey of Great Britain, H.M.S.O., London Williams J H 1971 ‘Roman Building Materials in the South West’, Trans Bristol & Gloucester Arch Soc 90, 95-119 |