SECTION 6.3: WHELFORD BOWMOOR FINDS REPORTS

6.3.1 THE ROMAN POTTERY (Kayt Brown)
Introduction
Condition
Methodology
Fabrics
Fine and Specialist wares
The Samian (Brenda Dickinson)
Coarsewares

Forms
Discussion
Bibliography

6.3.2 THE COINS (Cathy King)

6.3.3 THE SMALL FINDS (Hilary Cool)
Personal Ornaments
Toilet Equipment
Transport
Structural Items from Buildings
Tools
Fasteners and Fittings
Agriculture
Miscellaneous
Overview

6.3.4 ROMAN GLASS (Jennifer Price and Hilary Cool)
Catalogue
Square bottles
Window glass
Melted glass

6.3.5 CERAMIC BUILDING MATERIAL (Leigh Allen)

Top of Page

6.3.1 THE ROMAN POTTERY by Kayt Brown

Introduction

The excavations produced a total of 3551 sherds, weighing 35092g, dated to the Roman period. The assemblage displayed a relatively tight chronological range restricted in the main to the 2nd Century AD, with a small quantity of early-mid 3rd century AD material. The restricted date range and unusually high proportion of imported samian and amphorae are in contrast with many other assemblages in the region.

Top of Page

Condition

The assemblage was in a relatively poor condition; the surface preservation of sherds was variable due to adverse soil conditions, and many sherds were abraded. Although average sherd sizes were generally good, the material from the phase 1 enclosure ditches has an average sherd size of just 4 g. Almost half (47% by sherd count) of the total assemblage was from topsoil and machining layers. There were few good groups of pottery from stratified contexts; of these only nine contexts produced 50 or more sherds. The majority of the pottery was recovered from the phase 2 building platform and associated layers and may represent dumping of domestic refuse. Poor surface preservation precluded the identification of any evidence of use on all but a few sherds where such evidence was restricted to sooting on the exterior.

Top of Page

Methodology

The assemblage was recorded following the standard Oxford Archaeology recording system (OA unpublished), summerised in the report on the pottery from Somerford Keynes (section 5.3.1). Sherds were examined by context and assigned to ware (or generalised ware group) and vessel type, where appropriate. (Unstratified material was only recorded to generalised ware group). Sherds from each context were then quantified by sherd count, weight, rim count and rim equivalents. Also recorded were decoration, condition, evidence of use/reuse and repair.

Top of Page

Fabrics

The proportions of the different wares are summarised in Table 1.

Fine and Specialist wares

These wares comprise, samian ware, amphora, finewares (imported and British), mortaria, white wares, and white-slipped wares. Combined these account for over 15% of the assemblage by sherd count and 25% by weight. Within this, amphora and samian are particularly well represented. Amphora comprised body sherds of southern Spanish amphorae, most probably form Dressel 20, while a sizeable samian assemblage was predominately central Gaulish and Antonine in date (see report by B. Dickinson below). British finewares were restricted to a single sherd of Oxfordshire colour-coat. Sources for mortaria are Oxfordshire white-ware and white-slipped ware, and a Cirencester white-slipped fabric. The number of white-slipped sherds maybe under-represented, as the poor soil conditions appear to have removed in a number of cases the surface of a sherd. Where slip does survive it is inevitably patchy. This may also in part explain the paucity of white wares. Only a single sherd of Verulamium white ware was recorded although again poor soil conditions and surface discolouration may be have led to some misidentification of this material.

Top of Page

The Samian by Brenda Dickinson

All of the samian in this collection is heavily eroded and much of it consists of tiny unidentifiable and undatable scraps. Most of the better preserved pieces are Antonine or later. There is also a sherd of late 1st-century La Graufesenque ware and another vessel is probably of the same provenance and date. Of the three pieces from Les Martres-de-Veyre two are Trajanic and the other is either Trajanic or Hadrianic. One Lezoux piece is Hadrianic-Antonine.

The Antonine ware covers the entire period. With the exception of one East Gaulish vessel, it is all Lezoux and includes two or three examples of the colour-coated ware commonly known as black samian. there was probably more samian in use on the site after c 160 AD than in the earlier Antonine period. There was also a little late 2nd- or early 3rd-century East Gaulish ware, apparently from Rheinzabern.

The evidence of this assemblage, such as it is, does not suggest a break in activity on the site from its foundation, perhaps in the Trajanic period, to the end of the 2nd, or middle of the 3rd, century AD.

Top of Page

Coarsewares

The ‘belgic’ type and other grog-tempered coarse wares, characteristic of the late Iron age/early Roman period at other sites within the region, are virtually absent within this assemblage. At Asthall such fabrics account for 22.5% (by sherd count) of the phase 2 (mid 1st - early 2nd century AD) assemblage (Booth 1997, 116 table 5.16), at Somerford Keynes the figure is 17% (by sherd count) (see section 5.3) and grog-tempered coarse wares were the dominant fabric types at Thornhill Farm in the late Iron age/early Roman period (periods C-F, AD1 - 120; Timby 2004). At Whelford Bowmoor only 10 early roman grog-tempered sherds were identified, as body sherds of handmade jars. There were a small number of later, romanised, grog-tempered coarse ware fabrics (O80, O84); such coarse sandy and grog-tempered fabrics were being produced at Purton, west of Swindon from the late 2nd century (Anderson 1979).

Regional, presumably local, coarse wares accounted for over 50% of the assemblage (by sherd count, 44% by weight). Within this group of wares un-sourced fabrics, which probably include sherds of Oxfordshire and north Wiltshire fabrics (the identification of which was again obscured due to the poor preservation conditions), are the predominant ware groups. Sources for the north Wiltshire industry have been identified to the west of Swindon at Purton, Whitehill Farm and Toothill Farm (Anderson 1979, 9), although it was not possible to assign fabrics within this assemblage to known production centres within the industry. Such material generally dates from the early 2nd-4th centuries AD. At Wanborough these fabrics accounted for 44% of the quantified assemblage, and occurred in deposits from the late 1st century AD, indicating that kilns may have been operating from an earlier date around Wanborough (Seager Smith 2001, 298). Also particularly well represented within the Whelford Bowmoor assemblage are sherds of Black-burnished ware imitation fabrics (9% by sherd count) (Fig. 6.3.2: Phase 2 pottery (6-16), no. 14), whereas Dorset black-burnished wares are comparatively poorly represented (only 9 sherds). At the nearby site of Kempsford black-burnished wares were the third best represented coarse ware (Biddulph forthcoming), they were the only significant non-local coarse ware at Wanborough (Seager-Smith 2001, 299) and at Claydon Pike (23% in Phase 3) (see section 3.2). A Savernake ware variant, R94, identified at Claydon Pike, was also identified, along with a finer fabric with distinct grog inclusions (R38) which may well overlap with R94. A small quantity of Severn Valley ware was also identified. One notable absence is the reduced sandy R37 fabric, which comprised 38% of the assemblage by sherd count at Asthall (Booth 1997, 114). This may in part be explained by the surface preservation of the sherds but could equally be a reflection of the distribution of this fabric as it is thought to have been produced in the Asthall/Wilcote area of Oxfordshire. As one of the distinguishing characteristics of this fabric is the dark surfaces and lighter core it may well be that any sherds present within the assemblage could only be assigned to the generalised R30 ware group. Within the Kempsford assemblage (Biddulph forthcoming) R37 was a very minor component of the coarsewares. Calcareous coarseware fabrics were the final significant coareware group and included early roman fabrics.

Top of Page

Forms

Jar forms, although the principal forms, comprise only 40% of the assemblage by estimated vessel equivalents. Jar forms are also quite restricted in range, where sufficient profile survived for rims to be assigned to a vessel type, everted rim jars (Fig. 6.3.2: Phase 2 pottery (6-16), no. 6) and medium mouthed jars (Fig. 6.3.2, no. 12) were the principal forms. No early jar forms such as bead rim, carinated or high shouldered ‘necked’ jars, common vessel types in 1st century groups within surrounding assemblages, were present within this material. At Asthall, for example, bead rim jars are also absent but shouldered jars occur in the ‘belgic’ type fabrics (Booth, 1997, 120). Both forms are present within the Somerford Keynes assemblage (Brown SKNB pot, section 5.3.1). Bowls are the next significant form group represented (16% by estimated vessel equivalents) a number of these forms are curving sided bowls; a reflection of the large proportion of samian wares within the assemblage as most of these vessels are Dragendorff forms 31, 31R and 37. This high proportion of bowls to jars is very significant. Most assemblages in the region that span the same period tend to show a much higher proportion of jars to bowls, which decrease over time as the proportion of bowls increase. Again at Asthall, in phase 2 (early Roman period) jars comprise 80% of the assemblage, while at Somerford Keynes jars account for 60%, and Old Shifford Farm 62% (Timby 1995, 129). Within the small assemblage at Kempsford, however, bowls formed the dominant vessel class (Biddulph in prep.). The third best-represented class are cups at over 8% of the assemblage as a percentage of estimated vessel equivalents. This figure again reflects the large number of samian vessels, comprising forms Dr. 33 and Dr. 27 (see Table 2). Other forms represented to a lesser extent include plates, dishes, mortaria and flagons, each class forming less than 5% of the assemblage by vessel equivalents. Identifiable mortaria forms were Young type M17 Oxfordshire white-ware mortaria, dated to AD 240-300, and some of the latest material within the assemblage. The only miscellaneous form present was the base of a triple vase occurring in phase 2 (Fig. 6.3.1: Phase 1 pottery (1-5), no. 5). Phase 1 forms were restricted to jars, bowls and lids in coarseware fabrics. A correlation between ware groupd and form type is shown in Table 3.

Catalogue of illustrated sherds

Figures 6.3.1 and 6.3.2 present a selection of illustrated vessels from Phase 1 and Phase 2.

Figure 6.3.1: Phase 1 pottery

1 Coarse ware jar, fabric R38, FT37/SCA/LR3

2 Complete profile of small bowl/dish, fabric R35, FT37/SCA/LR3

3 Rim and neck of flagon, fabric O32, 40/A

4 Jar with burnished zone on shoulder and burnished lattice decoration, fabric R10, 36/A/3

5 Part of triple vase, fabric O32, 13/A/2

Figure 6.3.2: Phase 2 pottery

6 Cooking jar, fabric B30, 4/1

7 Large jar with groove on upper shoulder, fabric R38, 4/1

8 Jar with cordon at base of neck, fabric O30, 4/1

9 Bowl with flat rim, fabric R30, 4/1

10 Mortaria with spout and groove along top of rim, fabric M32, 4/1

11 Lid, fabric R35, 54/B

12 Medium mouthed jar/bowl, fabric R30,106

13 Coarse ware jar rim, fabric O80, 106

14 Complete profile of bowl with burnished lattice decoration, fabric B30, 106

15 Flanged bowl (top of rim missing), fabric O30, 106

16 Mortaria, fabric M30, 106

Top of Page

Discussion

The ceramics at Whelford Bowmoor display a tight chronological range from the early 2nd century to early-mid third century, with a few sherds in a late Roman shelly fabric, probably of 4th century date. Two phases of activity were identified stratigraphically, the early - mid 2nd century and late 2nd-early-mid 3rd century. The material from each phase is summarised in Table 4.

Within phase 1 the pottery recovered from the ditches comprising the enclosures was in a poor condition, as may be expected from such feature types. Wares were restricted to local and regional coarsewares. By far the bulk of the assemblage, including all the imported material that can be phased, was recovered from the building platform and associated layer, dated to the mid 2nd century AD. This material was much more diverse in nature; the mixed nature of the fabrics, range of vessel forms represented, and large sherd size are all support the theory that this is domestic material from the site. The assemblage does contrast to other domestic assemblages in the region; in particular the proportion of bowls to jars and the unusually large quantity of imports are not characteristic of a low-status rural assemblage. As a proportion of the assemblage, the fine and specialist ware of total of 15% by sherd count is much higher than would perhaps be expected of a ‘rural’ site in the region, as defined by Booth (Henig and Booth 2000, 173), who has used the proportion of this material as a means to comment on site status. Rural sites such as Thornhill Farm (Timby, 2004), Old Shifford Farm (Timby 1995, 128), and Gravelly Guy (OA in prep) have all produced less than 1% of this material during the early Roman period, while for urban sites such as Ashtall the figure is 6.9% in phases 2-4 (Booth 1997, 134), the villa site at Roughground Farm, Lechlade produced 7.3% fine and specialist wares (Booth, 1993, 141) and at Claydon Pike the figure in the early Roman period is 7.7%. The combination, therefore of the high proportion of fine and specialist wares and the variety of forms, such as cups, plates and mortaria, would suggest either a high status site or a highly ‘romanized’ lifestyle of the inhabitants. Rural sites in the vicinity that show continuous occupation from the Iron Age into the roman period, would appear generally, to maintain a strong ‘native’ element within the ceramic assemblages until well into the 2nd century. The Whelford Bowmoor assemblage provides a good contrast to such a pattern.

Top of Page

Bibliography

Anderson, A S 1979 The Roman Pottery Industry in North Wiltshire Swindon Archaeological Society Report No 2

Biddulph, E, in preparation ‘Pottery’ in P Booth and D Stansbie, A Roman rural landscape at Kempsford Quarry, Gloucestershire

Booth, P M, 1993 Introduction to the pottery from the 1990 excavation, in Allen, T, Darvill, T, Green, S and Jones, M, 1993 Excavations at Roughground Farm, Lechlade, Gloucestershire: a prehistoric and Roman landscape, Thames Valley Landscapes: The Cotswold Water Park Vol 1, OUCA, Oxford, 135-142

Booth, P. M. 1997 Asthall, Oxfordshire: Excavations in a Roman ‘Samll Town’ Thames Valley Landscapes Monograph No.9, Oxford Archaeological Unit

Henig, M and Booth, P 2000 Roman Oxfordshire Stroud

Seager Smith, R 2001 The coarse pottery in: A S Anderson, J S Wacher and A P Fitzpatrick The Romano-British ‘Small Town’ at Wanborough, Wiltshire, Excavations 1966-1976 Britannia Monograph Series No. 19, 232-344

Timby, J, 1995 Pottery, in G Hey, Iron Age and roman settlement at Old Shifford Farm, Standlake , Oxoniensia 60, 1996, 124-136

Timby, J 2004 The pottery in: D Jennings, J Muir, S Palmer & A Smith Thornhill Farm, Fairford, Gloucestershire: an Iron Age and Roman pastoral site in the Upper Thames Valley Oxford Archaeology Thames Valley Landscapes Monograph 23

Young, C J, 1977 Oxfordshire Roman pottery, BAR Brit Ser 43, Oxford

Top of Page

 

6.3.2 THE COINS by Cathy King

The 24 coins recovered from the site at Whelford Bowmoor, situated within 1 km of the Claydon Pike and Thornhill Farm excavations, were scattered over an area of c 0.5 hectare. Two of the three silver coins are denarii of Severus Alexander and there is also a plated core of a denarius of Caracalla; they all belong in the years AD 193 to AD 260 and the single bronze coin of the earlier empire was produced in the late second century AD (Table 1). The only period of peak loss that is well represented is AD 260-AD 296 with six coins or 25% of the total for the site. The apparent under-representation of coins from the later third and fourth centuries is due in part to the illegible coins of these years (4 coins, 16.7%) and the small size of the group as a whole but it may be worth noting that there are no coins from the years AD 364 to AD 378 which figure so prominently at Claydon Pike and Leaze Farm. Although the coins from Whelford Bowmoor were recovered from a small area, their rather wide chronological distribution does not support their being a hoard.

Top of Page

 

6.3.3 THE SMALL FINDS by Hilary Cool

This report deals with all the small finds from Welford Bowmoor other than the coins and the objects made from stone. If the obviously modern items and the featureless fragments of metal from the topsoil are ignored, there are 149 items (see Table 1). Of the material considered, one item came from the 1988 season and two from that of 1983. All of the rest of the material was recovered in 1985.

The assemblage is biased in that no worked bone artefacts are present presumably because bone does not survive well at the site. As noted when discussing Somerford Keynes (section 5.3.3), such a lack is a serious loss.

A selection of objects is illustrated in Figs 6.3.3: Brooches, 6.3.4: Other personal items and 6.3.5: Transport .

Top of Page

Personal Ornaments

The personal ornament assemblage is dominated by brooches (Fig. 6.3.3: Brooches ). Where the forms that have been discussed previously in the report on the brooches from Somerford Keynes, the reader is referred there for detailed typological comment.

The earliest brooch was an example of a La Tene III brooch from an unstratified context (98). It is obscured by iron pan but appears to belong to a common variant with central mouldings (Oliver 1988, 35). This form was certainly in use in the 1st century BC and into the 1st century AD. Though they are occasionally found in post Conquest assemblages, this brooch would, on balance, probably indicate pre-Conquest activity. Another early brooch is 67 from a Phase 1 context. It belongs to the Nauheim Derivative family and seems related to the expanded bow form (Olivier 1988, 37 no 15) found in the south-west. Mackreth (1998, 129 no. 84) prefers a pre-Conquest date for the type. Though this brooch does not have the pronounced swelling on the upper bow, the slight widening on the upper bow and the concentration of the decoration there does give the impression of an expanded area. A date early in the floruit of Nauheim Derivatives, ie in the mid 1st century, might thus be preferred. These two brooches predate the suggested start date of the activity on the site, and it is difficult to imagine that they would still have been in use by the early 2nd century.

There are also two other unstratified brooches which suggest occupation prior to the 2nd century. The Light Polden Hill 33 (see Somerford Keynes 1077) is a mid 1st century form. The fragment 101 probably comes from a second Nauheim Derivative and, if so, a mid to late 1st century date would be appropriate. It is possible that the two penannular brooches found unstratified (69 and 130) should also be regarded as predating the 2nd century. They are both of Fowler Type D5, examples of which regularly occur in pre-Flavian assemblages (see Somerford Keynes section 5.3, 1082, 5036).

The other brooches have lifespans that would be consistent with occupation starting in the 2nd century. The two Colchester Derivative brooches (63 and 95) both belong to Hull Type 93 (see Somerford Keynes section 5.3, 49, 60, 146 etc) and cannot be more closely dated than to the mid 1st to 2nd century. Lower Severn T-shape brooches such as 224 were in use from the later 1st century and during the 2nd century. This example is from a Hull Type 111, a form not otherwise attested in the Cotswold Water Park sites (see Hattatt 1987, 102). There is also one example (88) of a Wroxeter brooch, a type in use during the 2nd century (see Somerford Keynes section 5.3, 969, 1143, 5129). This was the only brooch found stratified in a Phase 2 context.

Two penannular bracelets, both unstratified, are present (Fig. 6.3.4: Other personal items ). The remaining terminal of 225 has the typical mouldings used when depicting an Asclepian snake. Such bracelets were most fashionable in Roman Britain from the mid 2nd into the 3rd centuries (Cool 2000, 33). The other example (220) has simple grooved decoration and cannot be closely dated within the Roman period.

Both finger rings (97 and 106) belong to the simple expanded type typical of the 1st to 3rd centuries (Henig 1974, types II and III). The unstratified 97 (Fig. 6.3.4) may be more narrowly dated to the 3rd century as it retains a moulded glass intaglio (ibid 164). 106 though not closely typologically dated can also be placed in the later 2nd to 3rd century period as it came from a Phase 2 context. The fact that both finger rings may be dated to the period when other aspects of the material culture such as the fine pottery are suggesting that the inhabitants had greater access to more expensive items, is of some interest. Wearing a ring with an intaglio device, no matter how crude, suggests aspirations towards a Romanised lifestyle and this may be another strand of evidence to suggest that occupation in Phase 2 was of a different nature than that of Phase 1. It may also be noted that the only stratified hobnails (12) were also recovered from a Phase 2 context. The implications for changes in lifestyles that the adoption of Romanised footwear imply have been discussed in the Claydon Pike report. It is possible that here too the adoption of such footwear was a late choice. Unfortunately the assemblage from Welford Bowmoor is too small for this to be shown in a convincing was as was possible at Claydon Pike. The only other hobnail recognised (159) being found unstratified.

Top of Page

Toilet Equipment

This category is very poorly represented. The only item identified with certainty is a long-handled ligula (109) from a Phase 1 context. The fragment 16 from the Phase 2 building platform can be tentatively identified as part of the arm of a pair of tweezers.

Transport

Half of a two-link snaffle bit (99) was recovered from a Phase 2 context (Fig. 6.3.5: Transport). It belongs to the commonest type of horse bit in use during the Roman period (Manning 1985, 66). The fragmentary 29 has the type of loop on the back used on Roman harness fittings, but a Roman date for the piece is not certain. as it was found unstratified and harness fittings can be expected as casual losses up to the recent past.

Structural Items from Buildings

The material in this category is summarised in Table 2 ignoring such items of structural metalwork as were found unstratified. As ever it is dominated by nails which were in use on the site during both phases. The nailed bindings have been included here as some may have come from such items as hinges. The table hints at a slightly more diverse range of finds associated with the phase 2 activity. The presence of the unstratified fragment of cast window glass suggest the presence of a glazed building in the vicinity.

Top of Page

Tools

The poor quality of the preservation of ironwork on the site probably means that this category is under-represented and certainty over identification is not always possible. Two knives were recovered. 11 from a Phase 2 context is an example of a Manning type 13 knife, one of the commonest all purpose knife forms in use in Roman Britain (Manning 1985, 114). The Roman date of 57 is not certain as it was found on the spoil tip. The features it retains are consistent with it being an example of a Manning type 11 knife (Manning 1985, 114), but the type is long lasting. Given the scarcity of later material on the site, it is reasonable to assume it probably derives from the Roman occupation. There is also a possible smiths set (168) and a triangular blade (5) of a shape that would be consistent with it coming from either a small adze or a wax spatula (Crummy 2003, Type A). Given the general nature of the assemblage from the site, the former identification seems most likely. Both of these items were from the Phase 2 occupation.

Top of Page

Fasteners and Fittings (Fig. 6.3.6: Fasteners)

The items in this category are summarised in Table 3. As has been noted in the other Cotswold Waterpark sites, lead pottery repairs form a major part of the assemblage. Both the types of repairs used and the relative rate of recovery of the different forms are similar to those found at Somerford Keynes and Claydon Pike. As noted in the case of those two sites, where the repairs retain pottery sherds, it is clear that coarse pottery is being riveted. Calculating the riveting rate based on the total pottery assemblage of 3551 sherds produces a figure of 0.62%. This is considerably higher than the rate for Lowland sites calculated by Evans (in Booth et al 2001, 382) of between 0.05 and 0.2% and the rate found at Claydon Pike (0.1%). A high rate was also found at Somerford Keynes. There it could be surmised that the high rate was in part the result of the collection strategy, though even so an unusually high rate of curation seemed likely. this explanation does not appear the case here as a high proportion of the small finds have been found stratified (see Table 1). The high curation rate of coarse pottery here seems at odds with the proportions of finewares, amphorae etc which has led Brown to conclude that the site was either high status or highly Romanised.

Whether there were things on the site that their owners wanted to lock up is a matter of debate. There is one latch lifter (26) from a Phase 2 context, but as discussed in the Claydon Pike report these are more designed to close doors rather than secure them. Items such as the unstratified 73 are normally considered to be part of locking mechanisms (see Birley 1997, 30), though quite how they functioned is unclear. Frequently they have heads in the form of bell-shaped studs but examples with disc heads are known, see for example a mid 1st century one from a Boudiccan deposit at Colchester (Crummy 1983, 124 no. 4142).

It may be noted that there is a notable paucity of studs, rivets, miscellaneous bindings etc that normally make an appreciable part of a Roman small find assemblage.

Top of Page

Agriculture

A remarkable find is the billhook (4) found in the Phase 2 platform. It appears complete, though now broken in two, allowing for post excavation flaking. On the radiograph the outline of the organic handle inserted inside the iron socket can clearly be seen preserved by mineral replacement. Typologically it belongs to Manning’s Type 2 billhooks, though lacking the spike on the back (Manning 1985, 58), a form that was in use throughout the Roman period. At the break one, area of iron is bent back suggesting that the break may have occurred before deposition and is not the result of post excavation damage that has clearly affected much of the ironwork in the long interval between excavation and study. It is difficult to imagine how the angle of this strip, which bends back through approximately 1000, could be consistent with accidental damage. As with the saw from Claydon Pike it seems likely to have been deliberately damaged. The presence of such a large, complete and probably deliberately broken item is of considerable interest, and raises the possibility that this was not casual rubbish disposal but a form of deliberate structured deposition.

Top of Page

Miscellaneous

The miscellaneous items are summarised in Table 4. There are three items that could have functioned as weights for nets and the like. 112 is large enough to be considered as a steelyard weight. The Water Park sites have produced a variety of non-standard steelyard weights, but most clearly have the iron suspension loop embedded through the centre of the weight (see Somerford Keynes section 5.3.3, nos 179 and 614). This piece lacks this feature and so such an identity is unlikely. Lead whorls are also present, one of which was found stratified in a Phase 2 context, strengthening the conclusion derived from the Somerford Keynes assemblage that they were part of the material culture of the early Roman period in the area. Precisely what they were used for is unclear. Possibly they too were used as weights.

Top of Page

Overview

This is a small assemblage compared to those from Somerford Keynes and Claydon Pike, but despite this it does cast some interesting light on the site activity. The first thing to note is that the brooch assemblage does indicate some activity in the area prior to the suggested early 2nd century start date for Phase 1. The other items that can be assigned typological dates confirm the 2nd to 3rd century date suggested by the pottery. There is no evidence that occupation or even casual use of the area continued into the later 3rd or 4th centuries.

The range of finds recovered is curiously limited. If Table 1 is inspected it can be seen that only 7 different functional categories are present. This may be compared to the 14 as Somerford Keynes and 16 at Claydon Pike. Though this is a much smaller assemblage and suffers from bone not surviving well, this paucity of functional categories is probably more a result of the nature of the occupation on the site rather a collection problem. The site was metal detected and whilst this can lead to a bias in what is found (Cool forthcoming), it does not lead to systematic under-representation of particular categories with the possible exception of toilet equipment. A category that is conspicuous by its absence here is that of household equipment. The paucity of the normal stud etc element of the fastener and fitting range has already been commented on, and many of those items too would have come from objects found in a domestic environment. The small find evidence would thus appear to be at variance with that of the pottery where the level of samian and amphorae recovered hints at a site with aspirations above that of a basic level farmstead. Only the finger ring 97 hints at similar aspirations amongst the finds considered here.

Top of Page

 

6.3.4 ROMAN GLASS by Jennifer Price and Hilary Cool

Twelve pieces of Roman glass were found; four were fragments of vessel glass, one was window glass and seven were melted lumps. In addition one piece of post-Medieval flat glass, probably from a window pane, was recorded. The vessel glass fragments come from four square bottles of 1st- or 2nd-century date. For discussion of this form see Claydon Pike (Section 3). The window glass fragment came from a cast matt-glossy pane, also probably of 1st- or 2nd-century date. For discussion of window glass see Claydon Pike (Section 3). The melted lumps of glass are not closely identifiable; they probably come from a vessel or a window pane badly affected by heat.

Catalogue

Square bottles
Vessel Y, blue green; small fragment, side of prismatic bottle. 10287, 39 4

Vessel Y, blue green; fragment. side of large prismatic vessel. 1028¸ 2´ 54/A Vessel Y, blue green; fragment, side and base of prismatic vessel. One raised concentric circle on base. 10281 2· 54/A

Vessel Y, blue green; fragment, side of prismatic vessel 1029° 3 4/2

Window glass
Window glass B¬ blue green; fragment, cast matt­glossy window pane with thick rounded edge and very smooth matt surface 1029² 3· Layer ±

Melted Glass
Blue green; seven small melted lumps. 10293 15 f4/4

Top of Page

 

6.3.5 CERAMIC BUILDING MATERIAL by Leigh Allen

A total of 28 fragments of tile weighing a total of 1 kg were recovered from the Whelford ­ Bowmoor excavation. the fabrics present were in general the same as those found at Fairford, Claydon Pike. Fabric 1, whose origin is the kiln site at Minety, Gloucestershire, fabric originating from Warnborough, Wiltshire, fabric 2 provenance unknown, and fabric 3 (not previously identified at Fairford, Claydon Pike (an orange/red rough fabric with quartz, flint, iron ore and clay pellet inclusions, probably originating from the Swindon area in north Wiltshire.
Unfortunately the sample is so small and the fragments so abraded that it is not possible to distinguish the types of tile present nor is it conclusive evidence for the existence of a tiled building. these fragments are more likely to have been amongst rubble brought on to the site for use in the construction of a pavement or building foundation.

Top of page