SECTION 3.2:CLAYDON PIKE ROMAN POTTERY by Sarah Green and Paul BoothWith Contributions By Edward Biddulph, Brenda Dickinson, Kay Hartley, Peter Webster and David Williams3.2.1 INTRODUCTION 3.2.2 QUANTITIES AND GENERAL DISTRIBUTION 3.2.4 FORMS 3.2.5 GENERAL DISCUSSION 3.2.1 INTRODUCTIONThe excavations of Longdoles Field from 1979-1984 produced approximately 1,158,705 g of Iron Age and Roman pottery, ranging in date from the middle Iron Age to the late Roman period but concentrating principally on the late Iron Age and early and middle Roman periods. All the pottery was subject to a preliminary examination in which basic information was recorded for each context group. This information (comprising the total weight of pottery recovered, the date range and a qualitative record of the major fabric and form groups) was entered on an INGRES relational database at the Oxford University Computer Science Department and provided preliminary dating evidence for the major site phases. The weight of pottery from each context provided by this initial record was one of the factors considered in selecting feature groups for detailed quantification, and also provides some basis for a comparative examination of the concentrations of particular classes of pottery over the whole excavated area. Apart from fulfilling a minimum weight qualification (50 g was chosen arbitrarily), contexts were chosen for quantification on grounds of their stratigraphic significance (ie whether detailed examination could elucidate some of the problems of site phasing) and that they provided a representative sample of all feature types, from all areas of the site throughout its history. On this basis some 430 kg, approximately one third (by weight) of the total pottery recovered, was recorded in detail, context by context. All samian ware and amphora sherds were recorded regardless of context. This work was carried out by Sarah Green in the 1980s using a recording system with a unique series of codes for fabrics and vessel types. Large parts of a draft report were prepared by Sarah Green on the basis of these data. After work on the project was revived the quantification of the samian ware and amphorae was reworked (by Kayt Brown) to provide data comparable with those for other ware groups in the assemblage. All the form and fabric data were recoded to bring them into line with the pottery recording system applied to most OA sites in the region since 1990. A number of groups of key importance for the phasing of the site were reexamined and the fabric, form and phase data recalculated. The overall report was reworked, augmented and updated by Paul Booth. The original fabric series is subsumed within the OA regional fabric series held at Oxford and the pottery database (having evolved from INGRES to dBaseIII and then Access) and other data are lodged with the project Archive. AcknowledgementsThanks are owed to David Williams, Kay Hartley, Peter Webster and his dedicated evening class and Brenda Dickinson for their specialist advice. Wendy Page, Eleanor Beard and Danyon Rey prepared and mounted the pottery drawings that form the basis of the present illustrations. Simon Palmer, Sebastian Rahtz, Debbie Duncan and Lauren Gilmour assisted Sarah Green with the computing and quantification and Jackie Carvell deciphered the original manuscript. Edward Biddulph, Kayt Brown, Grace Jones and Alex Smith all provided significant assistance in the process of revising and updating the report. 3.2.2 QUANTITIES AND GENERAL DISTRIBUTIONThe breakdown of the pottery assemblage in terms of weight by area is shown in Table 1. The material concentrated in four main areas (Trenches 13, 17, 19 and 29), which together accounted for 99.5% of the total pottery. The sample examined in detail, drawn from these four areas and also including the great majority of the small group from Trench 27, the circular shrine, totaled some 35,225 sherds (Table 2). Table 1 shows that, as would have been expected, the highest densities of pottery were encountered in Trenches 13, 17 and 19, the focal areas of the settlement. These three trenches produced very similar quantities of pottery in relation to their excavated areas. Only the assemblage from Trench 29 (and the small group from Trench 27)shows a significant reduction in the quantity of pottery in relation to the area examined. This may reflect in part a lower level of examination of features in this Trench, but increased distance from the focal area of the settlement was probably also a significant factor. The two largest trench assemblages (13 and 17) had the lowest proportion of material extracted for detailed recording. Nevertheless the totals thus examined were broadly in proportion to the importance of these trenches in terms of area. The smaller trench assemblages were examined more intensively in order to provide fully quantified datasets that were sufficiently large to have some statistical validity. The three principal measures employed to assess the relative frequency of pottery in the main excavated areas, sherd count, weight and RE, show considerable consistency across the trenches. The average sherd weight for the site as a whole, 12.2 g, is not particularly high, particularly when it is considered that it includes amphorae, and suggests a generally well-fragmented assemblage. There is some variation in average sherd weight from one trench to another, with a low figure in Trench 19 (the assemblage from Trench 27 is too small for the figures to be particularly meaningful) and a relatively high one in the more marginal Trench 29. There is no obvious explanation for the latter figure, though it is possible that a lack of late Roman activity here resulted in a lesser level of redeposition and consequent reduction of sherd size than was the case in other parts of the site. A rough calculation based on the figures from the detailed sample recording suggests an estimated original sherd total for the whole excavation of c 100,000. As already indicated, it was hoped that the context groups selected for detailed recording would prove representative of the site assemblage as a whole. It is therefore worth noting the variation in representation of the two major fabric classes for which comparative data are available; samian ware and amphorae. The samian ware total for the whole site was some 1984 sherds, of which 1482 (74.7%) occurred in the fully recorded groups. The comparable data for weight (more valid, since this was the primary measure for the full site assemblage) indicate that 72.7% of the samian was from groups subsequently fully recorded, against a site average of only 37.1% for all pottery. The representation of amphorae in the fully recorded sample is not dissimilar, with 62.9% (by weight) of the material occurring in these context groups, though interestingly this sample only contained 42.9% of all the amphora sherds from the site. It does seem, therefore, that the selection of contexts for detailed recording involved a degree of ‘cherry picking’, with a bias towards groups containing samian ware and amphorae. This suggestion is supported by the high incidence of samian and amphora sherds in unphased contexts (samian accounted for 39.4% of all sherds from such contexts, for example), suggesting that these groups were selected for recording on the basis of ceramic interest rather than stratigraphic importance. This may have been subconscious, and is in any case understandable, but it may reduce the value of the assemblage as a whole in terms of comparison with other sites from the region (see further below). Any such bias should be less significant, however, in terms of assessment of intra-site assemblage variation. 3.2.3 FABRICSIntroductionSome 84 fabric types were identified, divided into 11 major groups for publication (for further discussion of this approach and more detailed fabric descriptions see general introduction and Table 3). Analysis was done on the basis of macroscopic identification but it is estimated that at least 40% of the sherds were examined using a x10 hand lens and 10% using a x20 binocular microscope (fresh breaks in all cases). Soil conditions at Claydon Pike were not ideal for the preservation of pottery, particularly of surface finishes. It is therefore possible that fabrics such as mica dusted, lead glazed, and some of the colour-coated wares have not always been recognised. The principal data relating to fabrics are tabulated. Overall quantities of each fabric by the three major measures are presented in Table 3. The relative representation of fabrics by trench (area) is shown in terms of sherd count in Table 4and the changing occurrence of fabrics by phase for all trenches (as a percentage of sherd count) is given in Table 5. Fabrics are discussed below in relation to the general ware groups to which they are assigned, in the sequence set out in Tables 3, 4 and 5 and drawing on the information in these tables. Specialist contributions on selected fabrics and ware groups (on samian ware, amphorae and mortaria), are placed within this sequence. This discussion covers only the main points and not all fabrics are mentioned individually. Samian Wareby P. V. Webster with assistance from the University College, Cardiff Department of Extra-Mural Studies Samian Group IntroductionThe samian pottery from Claydon Pike consists, in the main, of small fragments (and many minute scraps). Very few contexts yielded large numbers of sherds. Plain forms predominated with few large pieces of decorated ware. In these circumstance a complete published list of all samian is impractical. All sherds were listed by context on record sheets and this information transferred to the computer record of the site. In this account information derived from the archive will be found, where relevant, in the structural report above and only a complete list of potters stamps and details of all the larger fragments of decorated ware are given below. The remaining information is summarised in a series of tables with commentary below. AcknowledgementsI should like to thank members of the University College, Cardiff, Department of Extra-Mural Studies Samian group who worked on the Claydon Pike samian and in particular, P Davies, N Hancock, L John, and D Rees, A Smith, R Taylor and H Weeks upon whose work on the decorated ware part of this report is based. CommentaryPieces with their provenance distribution summarised in Table 6 comprise all fragments for which an identification of source and form could be ascertained. Most, but not all, identifications are certain (some are marked with a query in the archive). Where a single context contained several fragments all certainly from the same vessel these were originally counted as a single piece but with so many small fragments, many identifiable although not from rim sherds, no other means of ascertaining the minimum number of vessels represented has been practical. It should be noted that in subsequent re-recording separate fragments were counted as such to bring the records for samian ware into line with those for the rest of the pottery from the site. RE data were recorded for the same reason. It is recognised that these measures underestimate the original number of vessels present on the site, but achieving comparability with quantification techniques for other pottery types was considered to be of paramount importance. In Table 7 the occurrence of vessel types is presented by Trench and phase, with quantification by REs. The presence of vessels not represented by rims is also indicated, as is the total quantity of each type per Trench including the material not within the fully recorded sample. Comparison of these figures shows that there are no significant differences between the character of the samples and the complete assemblages, except in the case of Trench 19 in which, contrary to the general trend across the site (see above), rather less than half the total samian ware fell within the fully recorded sample. Even here, however, the relative representation of some types was affected, rather than the fundamental character of the group, and the total trench assemblage was insufficiently small to have any impact on understanding of the overall site samian ware assemblage. In Table 8 the relative percentages of vessel types by trench are juxtaposed to allow comparison of vessel incidence across the site, using the overall trench totals rather than the data from the fully recorded sample. The table also gives a correlation with broad vessel class categories for the purpose of integrating the samian types with the broader evidence for vessel use. These tables show that the overwhelming majority of the pottery was plain ware. Decorated forms formed only 5.7% of the RE total found and it will be seen that although there was a large variety of forms represented, many of these were present in small numbers only and that vessels in the form 18 to 31 range (over 42% of the RE total) and the cup forms 27 and 33 (36.5% of the RE total) together dominated the assemblage. Given the chronological range of the material (see below) the dearth of decorated forms is unusual and must, one presumes, reflect some unusual characteristic of the usage of the site. The possible religious nature of some of the structures found may be relevant here although religious groups who prefer the austere to the ornate in their belongings would, one would have thought, have been the exception rather than the rule. Can it simply be that we are seeing here a reflection of the relative poverty of this rural site? If, as seems likely, decorated samian cost more than plain ware, could the inhabitants of the locality only afford the latter? The overall quantity of samian represented is reasonably large although this is in the main a product of the very large area excavated. Individual contexts rarely yielded many fragments and a great many produced only small pieces or minute scraps. This renders the identification of both form and source difficult and greatly increases the possibility of small fragments being either residual or even intrusive. It must be said, therefore, that many of the pieces form an inadequate basis for the dating of their particular contexts. Overall, however, the assemblage should be more reliable in terms of the site dating as a whole and the assemblage from individual trenches seems also to be large enough to allow us to draw some reasonably reliable conclusions. To this end Tables 7 and 8 have been divided by trench. It may be noted that, even given the small quantities of decorated vessels recovered, numbers of form 29 are very low (there were no rim sherds of this form) compared with those of form 37. Numbers of form 15/17 are also low. Thus, although there are a few vessels which could be pre-Flavian (forms 24/5 and Ritterling 12) the overall pattern would suggest that occupation commenced in the Flavian period and perhaps not much before c AD 80. The plentiful Central and East Gaulish pottery shows that occupation continued through the 2nd century and the numbers of such forms as 31R and 45 (along with smaller numbers of 32, 72 and 79/80) all support occupation at least into the late 2nd century or beyond. If we compare the relative quantities of pottery from the different sources as they appear in each of the main areas excavated (Table 6) we see that South Gaulish pottery is present in markedly larger proportions in Trench 13 when compared with the other areas. We may also note that the early forms (24/5, Ritterling 12 and 15/17) also derive solely from Trench 13 and it would be reasonable to assume that this was the area first to be occupied. Trench 29 also has a relatively high proportion of South Gaulish pottery and so should have seen relatively intense later 1st century occupation. The proportion of South Gaulish ware from the Trench 19 and from Trench 17 suggest a 1st century occupation which was either shorter or less intense. The relatively lower quantities of Central and East Gaulish pottery in Trench 13 are probably less significant as all areas yielded pottery extending into the late 2nd century in date (compare for example the proportions of the late form 45 in each of the areas as seen in Tables 7 and 8). Illustrated Sherds (Fig.3.2.1: Samian )Plain Ware Only one fragment of plain ware was sufficiently unusual to be worthy of illustration: 1. Bowl in a pinkish orange fabric with an orange-red slip, probably of East Gaulish origin. The rim is missing but there is the beginning of a handle just below the offset. The vessel is most likely to be a variant of Dragendorff’s form 34 (cf. Oswald and Pryce 1920, Pl LII; also Stanfield 1929, Fig. 13, no. 61 and pp.140 and 148). Unlike the vessel illustrated by Stanfield this is in a normal samian fabric. Possibly Antonine. From Trench 13. Decorated Ware The following abbreviations are used: O. (figure type in) Oswald 1937 S & S Stanfield and Simpson 1958 From Trench 13. 2. Form 37. South Gaulish. A narrow ovolo has a straight tongue ending in a rosette. Below is a row of chevrons delineated by wavy lines above and below. A fragment of the main decorative scheme shows a half medallion containing a small animal figure (possibly a lion). The chevron may be that used by Meddillus (cf. Knorr 1952, Taf. 39, C and D and Knorr 1919, Taf. 55, B). Vessels of this general style are to be found in the Pompeii Hoard (see, for instance, Atkinson 1914, no. 55). c. AD 75-90. 3. Form 37. Central Gaulish (from les Martres-de-Veyre). The ovolo is replaced by a row of minute dolphins arranged head to tail, a motif used by Drusus I (Potter X-3) as in S & S, Pl. 12, no. 142 and Terrisse (1968), Pl. IX, no. 10066. The decoration contains a seated figure with a cup (a reduced version of 0.571; cf. Terrisse 1968, Pl. V, no. 10059) and a double and triple leaf ornament used in conjunction as ibid. Pl. XIII, no. 454. Drusus also used the six-petal rosette. c. AD 100-120. 4. Form 37. Central Gaulish. The ovolo has been replaced by a triple leaf ornament, possibly Rogers (1974) G.120. The central tongue of the ornament was beaded and is reminiscent of part of an ornament used by Vegetus (S & S, Fig. 14, 20). Substitution of decorative details for the ovolo appears mainly in the first half of the 2nd century. Early/mid 2nd century. From Trench 17. 5. Form 37. South Gaulish. Basal portion showing a basal wreath over which was panel decoration. A panel containing a hare (O.2079) and hound survives with a fragment of a panel containing a half medallion to the left. The vertical borders terminate in six-pointed rosettes. The basal wreath is formed from the triple leaf shown with a list of its users by Knorr (1952, Taf. 49, F with detail and list to right). c. AD 75-100. 6. Form 37. Central Gaulish. Fragments of a bowl stamped ALBUCI. The die is of Albucius of Lezoux. The ovolo is his ovolo 1 (S & S, Fig. 35) but without the border below. The freestyle decoration shows, from left to right, a panther (O.1511), a warrior and a further panther (O.1533). The warrior appears in S & S, Pl. 120, no. 3, a closely similar piece. c. AD 150-180. 7. Form 37. Central Gaulish. A triple leaf motif, possibly Rogers (1974), G217, has been placed on its side and top to bottom apparently to form a continuous frieze immediately below the ovolo and wavy line border. Below this frieze is a small circle and ?beaded circle and what appears to be the hind quarters of a very large animal (perhaps a large lion). The ovolo may have had a rosette at the end of the tongue. Probably early to mid 2nd century. From Trench 19 8. Form 37. South Gaulish. A triple tongued ovolo over panel decoration with one panel containing a running hound (probably O.1925). There is stylised grass below the hound and a tendril affixed to the wavy line border above it. For a similar motif see Knorr 1919, Taf.18, D (stamped by Calvus); also ibid. Taf.57, H (by Mercator). c. AD 75-90. 9. Form 37. South Gaulish. The vessel lacks an ovolo but originally had strap handles on the rim (cf. Stanfield 1929, no. 7 - there is no evidence for a spout on our example but it is sufficiently incomplete for such a feature to have been on the lost portion). The upper zone contains a running dog (possibly O.1925) and a fragmentary hare. The three-pronged design and the ornament below both the hound and the hare appear all to have been made using an ornament illustrated as Knorr 1919, Td.7, centre (OF COTOI; see also ibid. Taf. 27, 6 & Taf.44,10). The lower zone consists of large 13-petal flowers (cf. ibid Td.12 - Bassus, Coelus, Meddillus). c. AD 75-90. 10. Form 37. East Gaulish, Rheinzabern ware. The lower portion of a bowl showing panel decoration with to the left the edge of a medallion separated by a beaded border from a panel containing the archer O.272 (Ricken and Fischer 1963, M.174) and a running stag. The stag is incomplete but appears to be similar to one used by Comitialis V (Ricken 1948, Taf. 96-102 passim) who is among those potters who uses the archer. See also comments after no. 11 below. 11. Form 37. East Gaulish, Rheinzabern ware. Ovolo above a fragmentary medallion with an astragalus and ?half medallion to the right. The ovolo is similar to Ricken and Fischer 1963, E.17 used by a variety of potters including Comitialis V (cf. no. 10 above). 2385/C/2. Another East Gaulish bowl rim with a fragment of worn ovolo from context 2472 (*Trench 13) may be from the same bowl. It may be noted that although the motifs appearing on nos. 10 and 11 are none of them restricted to only one potter, all are used by Comitialis V and it is likely, given the small quantity of decorated samian reaching the site that all the fragments listed under 10 and 11 are from the same bowl. The date is likely to be late 2nd to early 3rd century. The Potters Stamps by Brenda Dickinson Each entry gives: potter (i, ii, etc where homonyms are involved), die number, form of vessel, reading, pottery of origin and context number. (a), (b) and (c) indicate: (a) A stamp attested at the pottery in question. (b) Not attested at the pottery in question, though the potter is known to have worked there. (c) Assigned to the pottery on the evidence of fabric, distribution, etc. 1. Attius ii 8b 18/31 ATTIVS[.F]. The fabrics associated with this stamp suggest that it comes from a die used at Les Martres-de-Veyre in the Trajanic or early-Hadrianic period. It may belong to the Attius who later worked at Lezoux. 2216 2. Borillus i 5b 31 [BORILL] IOF Lezoux (a). This stamp occurs widely in Antonine Scotland. It has also been noted in the late-Antonine Aquincum Hoard and was used on forms 18/31R, 27 and 79/80. c. AD 150-180. 687/11 3. Butturrus 2a 33 BVT[TVRRI] Les Martres-de-Veyre (a). A stamp of one of the later Les Martres potters. It occurs in Antonine Scotland and in the Verulamium second fire deposits (after c. AD 150). There is one example on form 79. c. AD 140-165. 1640 4. Celsianus 1a 31 [CELSI]NAI.OF Lezoux (a). A stamp used frequently on mid- to late-Antonine forms such as 31R, 79, 79R and 80. c. AD 160-190. 2430 5. Daminus 3b 80 DA[MINI.M] Lezoux (a). A stamp from a die used on forms 31, 33, 79 and 80. Mid- to late-Antonine. 2113/2 6. Dontio 6a 27 DONTIOIIICI La Graufesenque (b). Most of the site evidence for this stamp comes from Flavian foundations such as Brough-on-Humber, Caerleon, Carlisle and Chester. There is one example from the pre-Flavian cemeteries at Nijmegen. c. AD 65-90. 633/B/2 7. Germanus i 27c or c' 18 GERMA[NIF] or GERMA[NI] La Graufesenque (a). Stamps from both versions of this die occur at Flavian foundations, including Segontium (27c) and the fortresses at Nijmegen and York (27c'). c. AD 70-90. 522/3 8. Malledo 1a 33 MALLEDON retrograde Lezoux (a). Malledo’s stamps occur on Hadrian’s Wall, in the Verulamium second fire deposits and on forms 31R and 79. This particular stamp has been noted on form 80. c. AD 150-180. U/S Tr13 9-10. Marcellinus ii 2a 31; 79 or Ludowici Tg MARCELLINIT Lezoux (a). A stamp noted in late-Antonine groups from New Fresh Wharf (London) and Pudding Pan Rock, and at forts in northern Britain reoccupied c. AD 160. c. AD 160-200. 1200/2 11. MA[ on form 31R, almost certainly from the same die as the last. 1200/2 01/66 12. Muxtullus 1b 31 MVX[TVLLIM] Lezoux (b). This stamp occurs in early-Antonine groups at Alcester and Castleford and there are three examples from Camelon. c. AD 140-160. 2161/2 13. Pateratus 1a 44/81 (stamped diagonally across the collar of a small bowl).P TE[R TIOF] Lezoux (a). This stamp is commonest on forms 18/31 and 27. There are two examples from Castleford in a context of c. AD 140-150. c. AD 135-155. 2360/A 14. Primus iii 22g 27g OFPRI La Graufesenque (a). Most of Primus iii’s output is pre-Flavian, but a few of his stamps occur at Flavian foundations, including this one, which is known from Castleford and the Nijmegen fortress. c. AD 50-75. 2421/E/1 15. Priscus iii 9b 33 PRISCVS Lezoux (b). Priscus iii’s decorated ware is mid- to late-Antonine and his plain ware turns up in groups of late-Antonine Samian from Tac (Hungary) and Pudding Pan Rock. c. AD 160-190. 1253/F 16. Tertiolus i 2a 31 TERT[IOLVS] Lezoux (c). The use of this stamp on form 18/31 and of one of his others on form 27 suggests a Hadrianic or early-Antonine date. 504/19 17. Tituro 5b 31 TITVRONIS Lezoux (b). This stamp is common on the later Antonine forms, such as 79, 79R, 80 and Ludowici forms Tg and Tx. c. AD 160-190. 1213 18. Vagiro/Vagirus 4a 33 VAGIROF Lezoux (c). This stamp occurs mainly on form 33, though one example of form 27 is noted. One of his other stamps was used on form 80. c. AD 145-175. 559/J/1 19. Vitalis vi 2a 38 or 44 VIT L[ISF] Lezoux (b). This minor Central Gaulish potter made forms 31, 33 and 38. His site record includes Chesterholm (2), Ilkley and London (a group of late-Antonine samian). This particular stamp occurs at Malton. c. AD 150-180. 2818/E/1 20. A fish, on form 32 etc., replacing the more usual name-stamp. This is from a die used at Rheinzabern (Ludowici 1905, 288, M18). Late 2nd- or 3rd-century. 504 21. I III II on form 27, South Gaulish. Flavian-Trajanic. 2807/A/1 22. ......IM on form 27g, South Gaulish. A thumbprint obscures the rest of the stamp. Flavian-Trajanic. 2261 23. ]FEC on form 27, South Gaulish. Flavian or Flavian-Trajanic. 2092/A/2 24. ]V, followed by a swallow-tail end to the frame, on form 15/17 or 18, South Gaulish. Flavian or Flavian-Trajanic. 547/M 25. ]VRI on form 18/31 or 31, Central Gaulish. Hadrianic or early-Antonine. 2342/A/2 26. ]VMA? on form 33, Central Gaulish. Antonine. 1730/A/1 27. DOCC[ on form 31, Central Gaulish. Antonine. 2597/2 28. ]VS or S [ on form 31, Central Gaulish. Antonine. 1555/2 29. P .....SF on form 31, very eroded, Central Gaulish or Trier ware. Late 2nd or 3rd century. 541/1 30. ]S or S[ on form 31, probably Trier ware. Late 2nd or 3rd century. 1205 31. ]VSF on form 31, East Gaulish. Late 2nd or 3rd century. 687/1 19055/99630 32. ]VSVSFECI or ]VTVSFECI on a flat base, East Gaulish. Late 2nd or 3rd century. 2384/2 Fine wares (F)A fairly wide range of fine wares was present on the site, but many of the fabrics involved were numerically insignificant. The fine ware assemblage was dominated by Oxford colour-coated ware (fabric F51), which totaled 80% of all fine wares by sherd count and REs, and some 84% by weight. Minority components in the fine ware assemblage included two lead glazed fabrics - F21 (Central Gaulish) and F22 (local), both represented only by body sherds and neither recorded from the contexts chosen for full quantification. The single mica dusted fabric present on the site (fabric F31) was also probably a local product, used for beakers and a dish. It was confined entirely to Trench 13 and was present in Phase 2 as well as later. An early 2nd century date seems likely. Early Roman imported fine ware fabrics F41 (Lyons ware) and F42 (Central Gaulish) were also very scarce and confined to Trench 13. Fabric F43, Central Gaulish ‘Rhenish’ ware, was also surprisingly rare, but the slightly later Moselkeramik (F44) was more common, with a notable concentration of sherds in Trench 19, where they formed 3.6% of the Phase 3/4 assemblage. With the exception of Oxford and Nene Valley products British colour-coated wares are also notably scarce. In the case of New Forest and possible Colchester products (fabrics F53 and F55 respectively), at the limits of their distribution, this is not particularly surprising. More striking is the scarcity of other regional colour-coated wares, in particular those of the ‘south-western brown slip ware’ family (F61 and perhaps also F62), likely to have originated in Gloucestershire (cf Young 1980). A ‘local’ colour-coated ware, of broadly similar character (F63), was also present only in very small quantities. This can probably be assigned to the North Wiltshire industries (cf eg Anderson 1979, 11-12). All these fabrics were probably used solely for beakers, except F61, in which a jar form was also identified. Nene Valley ware (F52) probably reached the site from the Antonine period onwards, though it was best represented in Phase 3/4, and again was principally used for beakers. Flagons and ‘Castor box lids’ also occurred, but it is notable that late Roman colour-coated jars, flanged bowls and dishes were not present except, in the case of jars, possibly as body sherds. Jars, at least, did form a part of the Oxford ware (F51) repertoire. This dominant fine ware was intrusive in Phase 2 contexts and thereafter present throughout, with its highest representation in the 4th century Phase 4, when it totaled 9% of all sherds. A wide range of forms was present, including Young (1977) types C4, C5, C8, C13, C16, C18, C22, C27, C37, C44, C45, C47, C51, C54, C55, C68, C70, C71, C75, C78, C81, C83 and C88 Amphorae (A)by D F WilliamsAmphorae of forms Dressel 20, Dressel 2-4, Southern Spanish, Camulodunum 186sp and Camulodunum 185A were identified. A total of 1707 sherds weighing approximately 68.5 kilos was examined, of which Dressel 20 sherds form by far the largest percentage. The breakdown of this material by fabric is summarised in Table 9, while Table 10 shows the distribution of the amphorae from the fully recorded sample by phase within trenches. Dressel 20 (fabric A11)Twenty three rims were recovered from the site, some 16 of which can be paralleled roughly with examples from Augst illustrated by Martin-Kilcher (1983) in her scheme for the development of the Dressel 20 rim. The date range suggested by these rims extends from the mid 1st century (the earliest example is paralleled by forms dated c AD 45-65 by Martin-Kilcher) to perhaps as late as the mid 3rd century, but only three of these rims are likely to date after the middle of the 2nd century. This is the most common amphora type imported into Roman Britain (Williams and Peacock 1983). Dressel 20 amphorae were made in the southern Spanish province of Baetica, along the banks of the River Guadalquivir and its tributaries between Seville and Cordoba, and carried olive oil (Ponsich 1974; 1979). This type of amphora has a wide date-range, from the Augustan prototype Oberaden 83, with a fairly upright rim, a short spike and less of a squat bulbous body than the later form, to the developed well-known globular form which, with some typological variation, was in use at least up to c AD 260 (Zevi 1967, 234). Dressel 2-4 (fabrics A20 and A35)This amphora type was made over a wide area, eg in Italy, France, Spain and the Aegean, as well as Britain (Castle 1978), during the period from the late 1st century BC to the mid 2nd century AD. However, quantitative trends suggest that it was in decline by the later 1st century AD (Panella, 1973). Tituli picti indicate that the principal content carried was wine (Zevi 1966). Four of the Claydon Pike sherds are in a distinctive ‘black sand’ fabric (A35), caused by inclusions of dark-coloured augite (Peacock 1971, 164). An origin in the Latium region has been suggested for this fabric, on the basis of the presence of yellow garnet when viewed in thin section (Courtois and Velde 1978). However, yellow-brown garnet is also a feature of the sands further south, and a Campanian origin, in particular the area around Herculaneum, has been argued by Peacock (1977). Further analysis by Velde and Courtois (1983) using an electron microprobe has distinguished two separate compositional groups of yellow (melanitic) garnet, one source of which they propose is situated near to Rome and another in the region of Vesuvius. The latter proposal agrees with Peacock’s (1977) suggestion, but as yet there is no archaeological evidence for an origin near Rome for the ‘black sand’ fabric. A Campanian origin seems more likely, since examples of bricks and tiles in the Pompeii-Herculaneum region are in an identical fabric. An amphora body sherd possibly from a Dressel 2-4 amphora was discovered (in a Phase 4 Trench 13 pit fill 1577/A/3) with the characters LEG II A roughly inscribed on the inner surface. The fabric (subsumed under A20) was soft, fairly rough and sandy with frequent small inclusions of white limestone with a buff (Munsell 7.5YR 7/3) outer surface and light red (2.5YR 6/8) inner surface and laminated core. Thin section study under the petrological microscope shows frequent inclusions of quartz grains and fossiliferous limestone, together with flecks of mica and a little plagioclase felspar and quartz-mica-schist. The particular fabric of this Claydon Pike vessel is not one the writer has encountered before. Unfortunately the inclusions present are all fairly common and not suggestive of a specific area of origin. Serious doubts have been cast on the authenticity of the inscription. Camulodunum 186C (fabric A12)This amphora form was made along the coastal regions of southern Spain, between Cadiz and Malaga, where a number of kilns are known, and seems to have been mainly used to carry fish-based products from the Flavian period or shortly before to the early 2nd century AD (Beltrán 1970; Peacock 1974). It was widely distributed in the western Roman Empire. UndesignatedSeveral amphora sherds could not be assigned to known classes with certainty. Mortaria(M)with identification of imported and unusual types by K F HartleyAlthough fourteen different mortarium fabric types were present this ware group, like fine wares, was dominated by products of the Oxford industry (fabrics M22, M31 and M41), which amounted to 85% of all mortarium sherds (83.4% RE). Local white slipped and unslipped oxidised wares (M32, M33 and M52), probably deriving from the south Gloucestershire/north Wiltshire area, were also present: two vessels of Antonine date (Nos 311-2), a wall-sided vessel (No 320) and No 321, a vessel dated to the 3rd century are illustrated in M32, ‘Cirencester’ white slipped ware (Rigby 1982b microfiche 1, D03-D05), and a single vessel in fabric M33 (No 313) is assigned to Minety as it is in the same fabric as Claydon Pike tile fabric 1. These fabrics were scarce, however, and numerically no more important than the products of major industries from outside the region - Verulamium, the Nene Valley and Mancetter-Hartshill, all of which were present (fabrics M21, M24 and M23 respectively). It should be noted, however, that the initial unquantified recording of the pottery from all contexts gave a clear impression that the ‘local’ fabric M32 was present in larger amounts than the above mentioned industries, although in spite of the source presumably being closer to Claydon Pike than the Oxfordshire kilns, it was never a serious competitor. In the late Roman Phase 4 Oxford mortaria make up about 97% of the total with single sherds of fabrics M32, M16, M24 and M52 forming the remainder. This situation is quite different at Cirencester, at least in the area of the late Roman cemetery (Sites CS and CT - McWhirr et al 1982). Here in the mid 3rd century and later there was a far higher proportion of local and Mancetter/Hartshill mortaria. Rigby suggests that in combination with the Oxford industry, they shared most of the market evenly (Rigby 1982b, microfiche 1, D13), but in the 4th century assemblage at The Beeches, however, Oxford mortaria comprised 85% (of REs), with local fabrics constituting the majority of the remainder (Keely 1986, 188-189). Claydon Pike may have been too far east of Cirencester to benefit from the presumed use of the Fosse Way in the distribution of Mancetter/Hartshill products. It was also on the eastern periphery of the marketing area of the local white slipped mortaria and flagons - few if any of these have yet been found east of Faringdon. Equally scarce were imported mortaria, but these were notable for the relatively wide range of (Gaulish) fabrics involved: M11. Hartley Group I (Hartley 1977)Seven vessels belonging to Hartley Group I were identified, three of which are illustrated. 307 and 309 are dated to AD 50-80 and 308 is dated AD 65-100. One vessel (not illustrated) had the remains of a stamp which was too damaged to be illustrated. All but one of this group came from Trench 13, the majority being residual in surface spreads. M12. Hartley Group II (Hartley 1977)Four vessels were identified as belonging to this group, two of which are illustrated, 314 and 315, both of Gillam type 238, dated AD 70-100. Not illustrated are two further vessels of this form, one of which, dated AD 65-100, showed traces of a stamp, unfortunately illegible. The other similar vessel was dated AD 50-100. M14. Bushe-Fox 26/30 (Bushe-Fox 1913; 1914)Approximately six vessels were identified as Bushe-Fox 26/30. No 310 is dated c. AD 50-85 with a provenance of Central France. Two other illustrated vessels, Nos 316 and 317, are dated 80-150 and are either from Gaul or the Rhineland. This type is virtually never stamped. M15. Central France ?Only two sherds of this fabric recorded, one of which was drawn (No 304) and is the subject of the following thin section report: a fragment of a stamped mortarium with a lead rivet in a soft, rough micaceous fabric containing frequent discrete grains of quartz and white felspar throughout, pinkish-red (between Munsell 5YR 7/4 and 7/6) surfaces having a slightly lighter core. Thin sectioning and study under the petrological microscope shows a groundmass composed of frequent flecks of mica and some small subangular grains of quartz, set in an anisotropic matrix of baked clay. Also present are larger discrete grains of potash and plagioclase felspar, quartz and mica (including biotite). The general composition of this fabric suggests derivation from a biotite granite. A possible continental source has been suggested for this vessel on typological grounds, and the Massif Central, Italy, Belgium, Switzerland and Yugoslavia have been mentioned. Outcrops of granite are to be found in Italy (the Alpine region and in Calabria in the south), the Massif Central, in Switzerland and to a limited extent in Yugoslavia. The sherd is too small for the exact rim profile to be determined but it does not appear to be a normal mortarium of Hartley Group I type (Hartley 1977). No other examples of the fragmentary stamp ] OF are known. It would be possible for it to be from an unknown stamp of Fronto who worked in the period AD 50-85 and whose mortaria are known from London and Martinhoe, Devon. However, until further examples of the stamp are found this identification and date can only be regarded as conjecture, especially as the two known mortaria of Fronto are not in a similar fabric. M16. ?Central France (cf Gloucester TF 9AA, Hartley 1985)Only body sherds present Oxford mortaria seem to have reached Claydon Pike from shortly after the inception of their production as the range of white ware vessels includes M2 and M6, dated AD 100-170 (Young 1977, 68, 69). The overall list of Oxford white mortaria includes Young (ibid) types M2, M3, M6, M10, M11, M17, M18, M19, M22 and M23. Only three vessels have been illustrated, two examples of Young form M2 with illiterate stamps (Nos 305 and 306) and an example of Young form M6, mended with lead rivets (No 318). White and red colour-coated oxidised mortaria (fabrics M31 and M41) were also present. The latter were almost twice as common as the former in terms of sherd count, but M31 was better represented by RE measurements. All the common mortarium types in these two fabrics, ie WC4, WC5, WC7, C97 and C100, were present. These types, together with the higher representation of mid 3rd century and later white ware types (ie Young M17-M19, M22 and M23) indicate that the highest level of mortarium use at Claydon Pike was in the 4th century in Phase 4. A single example of Young type C97 is illustrated (No 319). White Firing Wares (W) (except mortaria)White wares formed a relatively small proportion of the Claydon Pike assemblage. They derived from two principal sources. Several fabrics were ascribed to the Oxford industry, including parchment ware (W11) and fine and coarse tempered white wares (fabrics W12 and W22 respectively). A burnt white ware fabric (W23) was less confidently equated with the Oxford product (Young 1977, 113). None of these fabrics is sufficiently diagnostic for it to be certain that all sherds assigned to them were Oxford products, but in the absence of evidence for comparable white ware production in the region this seems most likely. Specific Oxford (Young 1977) vessel types identified included parchment ware forms P5, P9.2, P16, P24, P25 and white wares W2, W33, W36, W46, W49, W54, W56 and W59. The major component of the white ware assemblage was a distinctive fabric, W24, of uncertain but probably relatively local origin. This comprised some 70% of white ware sherds, though it was slightly less well-represented in terms of REs (61.3%). This may be because W24 was not used for forms such as flagons which (occurring in fabrics W12 and W22) tend to be slightly over-represented by RE measurements. Fabric W24 was used very largely for jars (94% of REs in this fabric) with occasional bowls and lids (forms assigned were C, CC, CD, CE, CH, HB, HC and L). It was particularly prominent in Phase 2 contexts and remained the dominant white ware in the Phase 3 and 3/4 assemblages, but was in decline (and probably residual) in Phase 4. The Oxford white wares occurred throughout the life of the site, though fabrics W11 and W23, both dated after AD 240 should have been intrusive in Phase 2 and earlier Phase 3 contexts. The relatively high representation of W23 in Phase 2 might be an indicator that these sherds were not the standard Oxford burnt white product, however. White slipped wares (Q) (except mortaria)Three oxidised white slipped fabrics were present, together comprising 2% of the sherd total (and a little less by weight and REs). Fabric Q21, Oxford white-coated (WC) ware was a relatively minor component of the assemblage from Phase 3 onwards, used principally for flagons and bowls (the only specific Oxford types recorded were WC1 and WC2) though jars, beakers and dishes also occurred. The principal fabrics were both probably of relatively local origin. Q22, ‘Cirencester white slip ware’ (Rigby 1982b microfiche 1, D04-D06), had a mortarium fabric counterpart (M32) as did the Oxford fabric Q21 (with mortarium fabric M31). Flagons, jars and beakers, in that order of significance, were the vessel types represented, appearing fairly consistently from Phase 3 onwards. The known chronology of this fabric suggests that a low level occurrence in Phase 2 may have been intrusive. Fabric Q23, the most important of the three fabrics, was certainly in production in the 1st century. It was well-represented in Phase 2 contexts and the wide range of associated vessel types included the ‘honey pot’ (type CCv, eg No 79) thought to be a specifically pre-Flavian form (see below). Flagons (eg Nos 38 and 46), jars (eg Nos 126 and 151) and beakers (eg No 178) were all significant components of this production. Bowls and dishes were also present but were much less important. Early ‘Belgic type’ wares (E)These wares, ‘Belgic’ purely in the stylistic sense implied by Thompson (1982, 4), dominate late Iron Age to early Roman assemblages in the region. The principal component is a grog-tempered ware group (E80) which with one major exception (see below) was not subdivided in detail. In terms of fabric this incorporated considerable variation, though grog or clay pellets were always the principal tempering agent, but as none of these variations can be matched with known production sites their significance is uncertain and was thought not to merit systematic recording. These fabrics show a complete spectrum in the technology employed ranging from the completely handmade through wheel-finished to wheel-thrown. The division into handmade or wheel-thrown variants (as at Cirencester into fabrics 3 and 24) has not been made here. Fabrics in the E80 group are widely encountered in the Upper Thames Valley but the lack of known sources makes it impossible to determine whether this indicates a single production location or a widespread tradition using similar raw materials across the region. The latter seems much more likely, however. The chronology of this important ware group remains uncertain - its origins are pre-conquest but are not thought to predate the 1st century AD in the Oxford region (eg Booth 1997b, 81-2; but compare also eg Timby 1996, 125). One important individual fabric was identified within this group - fabric E83, assigned to the Savernake industry. This shared all the principal characteristics of the E80 group, but may possibly have outlasted it, perhaps surviving into the late 1st century whereas most E80 production is unlikely to have continued into the Flavian period (a point which is, of course, not demonstrable at Claydon Pike). The precise relationship between E83 and the reduced Savernake fabric R95 is unclear, however; the latter is not necessarily a linear descendant of the former. Minor components of the E ware group included a further grog-tempered fabric E89, a distinctive fabric (E39) tempered with large rounded sand grains, and a rare flint-tempered fabric (E61). Occurrences of E39 and E61 may have been entirely residual as neither fabric appeared in Phase 2 contexts. A relatively wide range of vessel types was present in these fabrics, but this disguises the fact that the repertoire was totally dominated by jars, which amounted to 93% of all E ware REs. The minor types were a flagon, ?tankards, bowls and dishes in E80 and bowls, dishes and a lid in E83. This last form also occurred in fabric E39. The fabric E80 bowls included very large open forms such as No 265. Oxidised ‘coarse’ wares (O)Oxidised wares formed a significant proportion of the assemblage, totalling 9.2% of sherds. These fabrics were mostly of relatively local origin, with the principal one (O31, 47.7% of all oxidised sherds) the equivalent of the main North Wiltshire reduced coarse ware fabric R35. Probable and possible Severn Valley wares were also quite significant, although there is some uncertainty about which fabrics could be assigned to that industry. Fabrics O32, O42 and O45 were initially so assigned, but the fine slightly sandy character of O32 might suggest an affinity with the North Wiltshire industry. Fabric O43, originally defined as ‘local’, is also reminiscent of the Severn Valley tradition. Fabrics O42, O43, O45 and O46 together comprise 20.3% of oxidised coarse ware sherds. If O32 is included in this group this rises to 25.9%, or 2.4% of the total sherds. These fabrics were quite broadly distributed across the site, except for a notable concentration of fabric O43 in Trench 19. Their chronological ranges are less clear cut - the principal chronological emphasis of fabric O42 was in Phase 2, of O32 in Phase 3 and O43 was most common in Phase 4, but all were present from Phase 2 onwards. With regard to vessel types the principal forms in fabrics O42 and O43 were jars, supplemented by tankards and beakers. The O43 repertoire also included bowls and a lid, while that of O42 included flagons. This last form was also important in fabric O45 and especially in fabric O32, in which it amounted to almost two thirds of REs, but as suggested above in relation to white wares these figures may exaggerate the importance of forms which may survive as complete rims, particularly when the fabric RE total is relatively small, as here. The presence of tankards in many of these fabrics is a characteristic both of the Severn Valley and North Wiltshire industries. The output of the latter, represented most clearly by fabric O31, was dominated by jars (59.4% of REs), though not to the extent of its reduced ware counterpart, R35. Tankards, flagons, beakers and bowls (in that order in terms of RE quantities) were also significant parts of the O31 repertoire, which spanned almost the whole range of vessel forms. This fabric classification probably represents the products of at least two kilns, however these kilns are all part of the same group in the Purton/White Hill Farm/Wanborough area (Anderson 1980) and any fabric differences are insignificant. Already well-represented in Phase 2 contexts, in which it constituted almost 5% of sherds, fabric O31 reached a peak in Phase 3 and was clearly in decline in Phase 4, at only 3.8% of all sherds, when it may have been largely if not entirely residual. Other probable local products such as fabrics O82 and O84 were used almost exclusively for jars, particularly large storage jars. The same was true of the only certain non-local oxidised fabric present at the site, fabric O81 from Buckinghamshire (Booth and Green 1989) although this fabric, a regular component of later Roman assemblages in the Upper Thames Valley, was not very common at Claydon Pike. Reduced coarse wares (R)Reduced coarse wares, comprising 41.4% of the total sherds at Claydon Pike (38% of weight but 44.4% of REs), were the dominant ware group in all phases of the late Iron Age and Roman site, although they would have been introduced only in the second half of Phase 2 (after the mid 1st century AD) and their real importance in this phase is difficult to judge in the light of the clear evidence for intrusive material. At their peak, in Phase 3, R wares totaled 46.4% of all sherds. This ware group was dominated by two contrasting North Wiltshire traditions. Savernake ware (R95) comprised 6.1% of the Phase 2 assemblage and was even better represented in Phase 3 in Trench 13, although it is not likely to have been in production after about the middle of the 2nd century (Timby 2001, 81). Elsewhere it was reasonably common in all the parts of the site occupied from Phase 3 onwards, with the exception of Trench 19 from which, strangely, it was all but totally absent. The most important component of the entire site assemblage, however, was formed by the sandy fabric group R35. As with the corresponding oxidised fabric O31 this must have included the products of several different kiln sites which cannot be distinguished readily. R35 amounted to 18.7% of all sherds in Phase 2, rising to 37.5% in Phase 3 and maintained at that level in Phase 4. Unsurprisingly it was more common (over 40%) in Trenches 17, 19 and 29, all occupied from Phase 3 onwards, than in Trench 13 with a longer occupation range in which the earliest assemblages were dominated by other fabrics. A very wide range of vessel types occurred in fabric R35, but it was dominated by jars, which together comprised 89% of REs in this fabric. Of the other forms only bowls (2.9%), flagons (1.8%), tankards (1.4%) and beakers (1.2%) amounted to more than 1% of the output. A similar fabric (R34) was presumably also a North Wiltshire product, though with (initially, at least) an early specialist focus, perhaps represented by vessels such as No 302 (cf Cooper 1998, 327). It is not certain that all the sherds assigned to this fabric are equivalent to the more closely defined Cirencester fabric 5. The range of forms, again dominated by jars, otherwise contained only bowls, dishes and lids. The only other significant reduced ware was the fine micaceous fabric R85. This was probably a fairly local product but in a tradition which is widespread in south-west England (cf Timby 2000, 143). R85 appeared throughout the life of the site, so there is no definition of its chronological range. The repertoire of vessel types was dominated, as for all reduced wares, by jars (here 72.5%) but of the other types present lids (eg Nos 332 and 338) were unusually common, at 16.1%, much the highest representation of this type in any fabric except for a few with very small (statistically invalid) RE totals.
|
Fabric |
Sherds |
% sherds |
Wt (g) |
% wt |
Eves |
% eves |
No. vessels |
% vessels |
Oxfordshire colour-coat ware (F51) |
28 |
5% |
90 |
1% |
0.32 |
4% |
3 |
5% |
Nene Valley colour-coated ware (F52) |
59 |
11% |
476 |
6% |
0.93 |
11% |
6 |
10% |
New Forest colour-coated ware (F53) |
3 |
1% |
44 |
1% |
- |
- |
- |
- |
Oxfordshire white ware mortaria (M22) |
3 |
1% |
184 |
2% |
0.01 |
<1% |
1 |
2% |
Oxfordshire red colour-coated mortaria (M41) |
4 |
1% |
46 |
1% |
- |
- |
0 |
- |
Oxfordshire fine oxidised ware (O11) |
20 |
4% |
110 |
1% |
0.1 |
1% |
1 |
2% |
General sandy oxidised ware (O20) |
13 |
3% |
68 |
1% |
0.1 |
1% |
1 |
2% |
General coarse oxidised ware (O80) |
7 |
1% |
153 |
2% |
- |
- |
- |
- |
General medium sandy grey ware (R30) |
200 |
36% |
3018 |
40% |
2.98 |
36% |
22 |
37% |
Black burnished ware 1 (B11) |
185 |
34% |
2470 |
33% |
3.08 |
37% |
18 |
30% |
Late shell-tempered ware (C11) |
30 |
5% |
640 |
8% |
0.85 |
10% |
8 |
13% |
TOTAL |
- |
- |
- |
- |
Pit 1989 yielded 7.5 kg of recorded Roman pottery from 14 contexts (a further 4 kg was recorded on site but cannot be located), though in actual terms, the pottery was recovered from deposits 1 to 6. In two instances, pottery had been assigned between two contexts (3-4 and 4-5). The dating sequence from the top fill (1) to the lowest fill with pottery (6) is as follows:
1 Mid to late 4th century
2 Mid 4th century
3 Early 4th to mid 4th century
3-4 Early 4th to mid 4th century
4 Late 3rd to early 4th century
4-5 Early 4th century
6 Early 4th century
A wide range of forms and fabrics are represented in this pit assemblage. Locally-produced grey ware (R30) inevitably formed the largest proportion in most deposits. This category included North Wiltshire material (particularly fabric R35) which was not distinguished individually. It is clear that non-local coarse pottery manufacture was providing serious competition to local producers since the volume of grey ware is almost matched by black-burnished ware from Dorset (B11). Plain or flanged rim dishes and cooking jars arrived from this source, and most examples are among the latest of Dorset products. From context 6 came a splayed rim jar current during the first half of the 4th century (Gillam type 13), while an undecorated plain-rimmed dish (Gillam type 84) came from context 2. Such vessels continued to be made well into the middle of the 4th century, though production of the fabric declined rapidly at this time. Shell-tempered ware (C11) is represented in quantity in the latest fills, particularly in context 1, almost certainly pushing the date of final deposition of this fill into the second half of the 4th century. Cooking pot type jars and dishes found here are typical of the fabric. Mortaria, all from Oxfordshire kilns, are not well represented; the single recognisable form (Young (1977) type M22) is an industry standard. Of the fine wares, Nene Valley colour-coated ware (F52), surprisingly, takes a significantly larger share of the assemblage than fabric F51 from Oxfordshire. The representation of F51 is broadly in line with the Phase 3 and 4 figures from the site as a whole, and at Cirencester, for example, fabric F51 only contributes significant amounts after c 350 (Cooper 1998, table 28). Context 2 in pit 1989 yielded the most F51, including Young bowl type C45. Even allowing for the possibility that some of the pottery assigned to F52 is actually F53 from the New Forest, the proportion of F52 remains extremely high and is completely anomalous in comparison with the overall site assemblage. At least some of the F52 sherds in feature 1989 were residual; among the pieces present is an indented beaker with so-called ‘butcher’s hook’ barbotine decoration from context 4, dated to the 3rd century (Perrin 1999, 94).
With a mean sherd weight (MSW) of 14 g, pottery from the pit is reasonably well-preserved. Throughout the pit, pottery was of uniform size and condition, except in context 1, where the MSW was 23 g, and context 6, which had a MSW of 10 g. In the remaining fills it was between 13 and 15 g. A small number of cross-context joins are evident between contexts 2 and 3-4, and between 4-5 and 6. This, along with the uniform condition, suggests only short intervals between some episodes of filling, though it is conceivable that some contamination (during excavation?) has taken place. However, the composition, condition, and date of context 1 is certainly different from the lower fills, suggesting a much longer pause separating deposition of that context and context 2. Given the uniform condition (generally) and the cross-context joins, the pottery, except for that in context 1, probably derived from the same source, presumably a midden, which continued to be added to even after relocation of some of its contents. This may help to explain the seemingly residual material, especially in the lower fills.
In summary, all pottery-yielding contexts were deposited during the first half of the 4th century, probably towards the middle of the century. The final filling took place in the second half of the century, though perhaps not too far beyond AD350.
1. O42, CC. 2092/2.
2. R95, CD. 2092/A/2.
3. R95, CD. 2092/2.
4. R35, CD. 2092/1.
5. O42, CD. 2092/1.
6. B11, CH. 2092/1.
7. R34, H. 2092/1
8. R34, HB. 2092/1.
9. B11, I, with faint traces of acute angle lattice decoration. 2092/1.
10. B11, JA, with faint traces of acute angle lattice decoration. 2092/A/2.
11. W24, JA. 2092/A/2.
12. R35, JA, with small hole drilled in base. 2092/1
.
1. E83, CD. 547/F/1.
2. R95, CD. 620/L.
3. R35?, CD, slight sooting. 547/D/1.
4. R95, CD. 547/B/1.
5. R35, ?CD. 620/K.
6.
B11, CK. 620/N.
In stratigraphic sequence
1. R30, CM (?Young type R38). 1989/A/6.
2. B11, JA. 1989/A/6.
3. R30, CM (?Young type R38). 1989/A/5.
4. B11, JA. 1989/A/5.
5. B11, CK. 1989/B/3-4.
6. B11, CK. 1989/B/3-4.
7. F52, E. 1989/B/3-4.
8. F52, E. 1989/B/3-4.
9. B11, JA. 19898/B/3-4.
10. M22, KD (Young type M22). 1989/A/3.
11. B11, CK. 1989/A/2.
12. F51, HC (Young type C45). 1989/A/2.
13. B11, JA. 1989/A/2.
14. R30, L. 1989/A/2.
15. C11, CK. 1989/A/1.
16. R30, CM (Young type R38). 1989/A/1.
17. C11, HB. 1989/A/1.
Anderson, A S, 1979, The Roman pottery industry in North Wiltshire, Swindon Archaeol Soc Rep 2
Anderson, A S, 1980, Romano-British pottery kilns at Purton, Wiltshire Archaeol Mag 72/73, 51-58
Atkinson, D, 1914, A hoard of Samian Ware from Pompeii, J Roman Stud 4, 27-64
Beltrán Lloris, M, 1970 Las ánforas romanas en España, Zaragosa.
Booth, P, 1991 Inter site comparisons between pottery assemblages in Roman Warwickshire: ceramic indicators of social status, J Roman Pottery Stud 4, 1-10
Booth, P, 1997a Asthall, Oxfordshire, excavations in a Roman ‘small town', 1992, Thames Valley Landscapes Monograph No 9, Oxford Archaeol Unit
Booth, P, 1997b Pottery and other ceramic finds, in C Mould, An archaeological excavation at Oxford Road, Bicester, Oxfordshire, Oxoniensia 61 (for 1996) 75-89
Booth, P, 2002 Late Roman cemeteries in Oxfordshire: a review, Oxoniensia 66 (for 2001), 13-42
Booth, P, Boyle, A, and Keevill, G D, 1994 A Romano-British kiln site at Lower Farm, Nuneham Courtenay, and other sites on the Didcot to Oxford and Wootton to Abingdon water mains, Oxfordshire, Oxoniensia 58 (for 1993), 87-217
Booth, P, Evans, J, and Hiller, J, 2001 Excavations in the extramural settlement of Roman Alchester, Oxfordshire, 1991, Oxford Archaeology Mono 1, Oxford
Booth, P, and Green, S, 1989, The nature and distribution of certain pink, grog tempered vessels, J Roman Pottery Stud 2, 77-84
Bushe-Fox, J P, 1913 Excavations on the site of the Roman town at Wroxeter Shropshire, in 1912, Rep Res Comm Soc Antiqs London No 1
Bushe Fox, J P, 1914 Second report on the excavations on the site of the Roman town at Wroxeter, Shropshire, 1913, Rep Res Comm Soc Antiqs London No 2
Castle, S A, 1978 Amphorae from Brockley Hill, 1975, Britannia 9, 383-392
Cooper, N J, 1998, The supply of pottery to Roman Cirencester, in N Holbrook (ed), Cirencester: the Roman town defences, public buildings and shops, Cirencester Excavations V, Cirencester, 324-350
Courtois, L, and Velde, B, 1978 Une amphore à grenat jaune du Latium à amathonte, Bull Corr Hellen 102, 977-981
Gillam, J P 1976 Coarse fumed ware in north Britain and beyond, Glasgow Archaeological Journal 4, 76-80
Green, L S, and Booth, P, 1993 The Roman pottery, in T G Allen, T C Darvill, L S Green and M U Jones, Roughground Farm, Lechlade, Glos.: a prehistoric and Roman landscape, Thames Valley Landscapes: The Cotswold Water Park 1, (OUCA) Oxford, 113-142
Hands, A R, 1993 The Romano-British roadside settlement at Wilcote, Oxfordshire I. Excavations 1990-92, Brit Archaeol Rep (Brit Series) 232, Oxford
Hartley, K F, 1977 Two major potteries producing mortaria in the first century AD, in J Dore and K Greene (eds), Roman pottery studies in Britain and beyond, Brit Archaeol Rep Supplementary Series 30, 5-17, Oxford
Hartley, K F, 1985 TFs 9AA-9AC, 9Z, imported mortaria, in H R Hurst, Kingsholm, Gloucester Archaeol Reps Vol 1, Gloucester, 72
Howe, M D, Perrin, J R and Mackreth, D F, 1980 Roman pottery from the Nene Valley: a guide, Peterborough City Museum occasional paper 2
Keely, J, 1986 The coarse pottery, in A McWhirr, Houses in Roman Cirencester, Cirencester Excavations III, Cirencester, 158-189
Knorr, R, 1919 Töpfer und Fabriken verzierter Terra-Sigillata des ersten Jahrhunderts, Stuttgart
Knorr, F, 1952 Terra-Sigillata-Gefässe des ersten Jahrhunderts mit Töpfernamen, Stuttgart
Martin-Kilcher, S, 1983 Les amphores Romaines a huile de Baetique (Dressel 20 et 23) d'Augst (Colonia Augusta Rauricorum) et Kaiseraugst (Castrum Rauracense). Un rapport preliminaire, in J M Blazquez and J Remesal (eds), Produccion y Comercio del Aceite en la Antiguedad. II Congresso, Madrid, 337-347
McWhirr, A, Viner, L, and Wells, C, 1982 Romano-British cemeteries at Cirencester, Cirencester Excavations II, Cirencester
Oswald, F, 1937 Index of figure types on terra sigillata, Liverpool
Oswald, F, and Pryce, T D, 1920 An introduction to the study of Terra Sigillata, London
Panella, C, 1973 Appunti su un gruppo di anfore della prima, media e tarda etá Imperiale, Ostia III: Le terme del Nuotatore: scavo dell’ambiente V et di un saggio nell’area Studi Miscellanei 21, Rome, 460-622
Peacock, D P S, 1968 A petrological study of certain Iron Age pottery from western England, Proc Prehist Soc 13, 414-427
Peacock, D P S, 1971 Roman amphorae in pre-Roman Britain, in M Jesson and D Hill (eds) The Iron Age and its hill-forts, Southampton, 161-188
Peacock, D P S, 1974 Amphorae and the Baetican fish industry, Antiqs J 54, 232-243
Peacock, D P S, 1977 Roman amphorae; typology, fabric and origin, Coll de L’École Française de Rome 32, 261-278
Peacock, D P S, and Williams, D F, 1986 Amphorae and the Roman economy, London
Perrin, J R 1999 Roman pottery from excavations at and near to the Roman small town of Durobrivae, Water Newton, Cambridgeshire, J. Roman Pottery Stud 8
Ponsich, M, 1974 Implantation Rurale Antique sur le Bas-Guadalquivir, Madrid
Ponsich, M, 1979 Implantation Rurale Antique sur le Bas-Guadalquivir, Paris
Ricken, H, 1948 Die Bilderschüsseln der römischen Töpfer von Rheinzabern, Tafelband, Bonn
Ricken, H, and Fischer, C, 1963 Die Bilderschüsseln der römischen Töpfer von Rheinzabern, Textband, Bonn
Rigby, V, 1982a The coarse pottery, in J S Wacher and A D McWhirr, Early Roman occupation at Cirencester, Cirencester Excavations I, Cirencester, 153-200
Rigby, V, 1982b The pottery, in A McWhirr, L Viner and C Wells Romano-British cemeteries at Cirencester, Cirencester Excavations II, Cirencester, 112-125 and microfiche
Rogers, G B, 1974 Poteries sigillées de la Gaule Centrale I: Les motifs non figurés, Paris
Stanfield, J A, 1929 Unusual forms of terra sigillata, Archaeol J 86, 113-151
Stanfield, J A, and Simpson, G, 1958 Central Gaulish Potters, London
Swan, V G, 1975 Oare reconsidered and the origins of Savernake Ware in Wiltshire, Britannia 6, 36-61
Symonds, R P, 1992 Rhenish wares, fine dark coloured pottery from Gaul and Germany, Oxford Univ Comm for Archaeol Monograph 23
Terrisse, J-R, 1968 Les Céramiques sigillées gallo-romaines des Martres-de-Veyre (Puy-de-Dome), Gallia Supplement 19, Paris
Thompson, I, 1982 Grog-tempered `Belgic' pottery of South-eastern England, Brit Archaeol Rep (Brit Series) 108, Oxford
Timby, J R, 1996 Pottery, in G Hey, Iron Age and Roman settlement at Old Shifford farm, Standlake, Oxoniensia 60 (for 1995), 124-136
Timby, J R, 2000 Pottery, in E Price, Frocester. A Romano-British settlement, its antecedents and successors, Volume 2 The finds, Stonehouse, 125-162
Timby, J, 2001 A reappraisal of Savernake ware, in P Ellis (ed), Roman Wiltshire and after: Papers in honour of Ken Annable, Wiltshire Archaeol and Nat Hist Soc, Devizes, 73-84
Tomber, R, and Dore, J, 1998 The national Roman fabric reference collection: a handbook, Museum of London Archaeol Services Mono No 2
Velde, B, and Courtois, L, 1983 Yellow garnets in Roman amphorae - a possible tracer of ancient commerce, J Archaeol Science 10, 531-539
Wallace, C, and Webster, P V, 1989 Jugs and lids in Black Burnished ware, J Roman Pottery Stud 2, 88-91
Williams, D F, and Peacock, D P S, 1983 The importation of olive oil into Roman Britain, in J M Blazquez and J Remesal (eds), Produccion y Comercio del Aceite en la Antiqedad. II Congresso, Madrid, 263-80
Young, C J, 1977 The Roman pottery industry of the Oxford region, Brit Archaeol Rep (Brit Ser) 43, Oxford
Young, C. (J), 1980 The late Roman finewares, in B Rawes, The Romano-British site at Wycombe, Andoversford; excavations 1969-70, Trans Bristol and Gloucester Archaeol Soc 98, 41-46
Zevi, F, 1966 Appunti sulle anfore roman, Archaeologia Classica 18, 207-247
Zevi, F, 1967 (Review of) M.H. Callender, Roman Amphorae, J Roman Stud 57, 234-238