PREVIOUS   NEXT   CONTENTS   ISSUE   HOME 

The Samian

The total samian assemblage comprises 4937 sherds, weighing 79237g, and forms just 1.2% by weight of the total pottery recovered. The samian was identified and dated by Brenda Dickinson and catalogues of the stamps and decorated vessels were produced (see below). The Arretine ware (Italian-type sigillata) was isolated and subsequently identified and dated by Joanna Bird (see below). Figure 309 shows cumulative vessel losses at 5-year intervals. As part of the English Heritage-funded Samian Project, Steve Willis accessed the samian data in advance of publication and produced an overview of the trends at Elms Farm compared with a range of assemblages recorded elsewhere. The following report encompasses both sets of work.

Abbreviations

Conspectus = form type in Ettlinger et al. 1990
D = figure type in Déchelette 1904
O = figure type in Oswald 1936-7
OCK = second edition of the Corpus Vasorum Arretinorum (Oxé and Comfort 2000)
ORL = Der obergermanisch-raetische Limes des Römerreiches
Rogers = motif in Rogers 1974

The Italian-type terra sigillata by Joanna Bird

Cite this as: Bird, J. 2015, The Italian-type terra sigillata, in M. Atkinson and S.J. Preston Heybridge: A Late Iron Age and Roman Settlement, Excavations at Elms Farm 1993-5, Internet Archaeology 40. http://dx.doi.org/10.11141/ia.40.1.bird

The Italian-type sigillata consists of twenty-five sherds representing a maximum of twenty vessels; eighteen platters, including seven examples of Conspectus form 12, and two cups. The majority of the pieces, including the single stamped base, are in a fine fabric with some colourless mica flakes visible in the section. Without analysis no certain source can be suggested for this fabric, which is a very pale brown or beige in colour, with a brownish tint to the slip. The material is mostly dated c. 15 BC-AD 10, with a few pieces that could date up to c. AD 20.

1. Three joining sherds from the floor of a large platter with rouletted circle and two-line radial stamp. The upper line of the stamp is almost completely lost, but may include the letter R; the lower line is probably complete, and reads TITI. Fill 20031, Pit 20030 (Group 248), Area L, Period 2B

Dr Philip Kenrick comments:

'The only such stamps for which I have radial examples are:

OCK 2224.2	LVCR/L.TIT	Lucrio L. Titi of Arezzo (?), c. 20 BC+
OCK 2251.1-3	SEX/TITI	Sex. Titius of Arezzo, c. 30-15 BC

My only initial offering for the fragment of line 1 was "..R" so Lucrio might be possible, but there is no facsimile for the type listed above for confirmation and I would be a little surprised if the stamp extended further to the left to accommodate the praenomen. The only other potential stamp with "..R" would be ANTER/TITI (OCK 2204, location undefined, c. 15 BC+), but I do not have either a radial stamp for him or identical facsimile/reading. Sex. Titius has more radial stamps, and might perhaps be possible.'

2. Platter, Conspectus form 12, and probably 12.3, but insufficient survives to be certain; diameter 300mm. Mid- to late Augustan. 4000 (Group 8021), unstratified, Area A1

3. Large (diameter approx. 460mm) platter, Conspectus form 12; cf. especially 12.1.3, though 12.1.2 is the only example illustrated of comparable size, may be the same vessel as 7148 below. Mid- to late Augustan. Fill 7134, Pit 7060 (Group 310), Area G, Period 2B

4. Floor sherd, large platter; may be same as 7134 above, and is certainly of the same date and origin. Fill 7148, Pit 7146 (Group 308), Area G, Period 2B

5. Platter base with bevelled foot and double groove on the floor, probably Conspectus form 12. Relatively dark reddish-orange fabric and slip, probably from an early Gaulish workshop. Mid- to late Augustan. Fill 7173, Pit 7174 (Group 853), Area G, Period 3

6. Platter, probably Conspectus form 12; diameter approx. 260mm. Mid- to late Augustan. Fill 7516, Stake-hole 7515 (Group 849), Area G, Period 3

7. Bevelled cup foot, too fragmentary to identify certainly, but cf. Conspectus forms 13 and, particularly, 14. The fabric may be one produced at Pisa, but the sherd is burnt. Mid- to late Augustan. Layer 5883 (Group 369), Area I, Period 3-4

8. Two sherds: i) platter floor fragment; ii) platter wall/floor sherd, probably as Conspectus 12.3.1. Both mid- to late Augustan. Fill 9585, Pit 9611 (Group 75), Area D, Period 2

9. Platter, Conspectus form 12, in a fine chalk-filled fabric. Mid- to later Augustan. Fill 10287, Pit 10288 (Group 303), Area F, Period 2B

10. Two joining sherds, small (diameter 180mm) platter, Conspectus form 12.2. Relatively coarse fabric, very micaceous, probably early Gaulish. Mid- to later Augustan. Fill 11277, Pit 11337 (Group 61), Area N, Period 2A

11. Large square foot from a platter (cf. Conspectus B1.6-10); the slip is almost completely lost. Augustan-Tiberian. Fill 8011, Pit 8012 (Group 296), Area E, Period 2

12. Four sherds: i) two sherds, probably from one platter floor; ii) floor sherd with at least one groove, from a large platter; slightly burnt; iii) platter floor sherd, burnt. All Augustan-Tiberian. Fill 9704, Pit 9792 (Group 288), Area D, Period 2B

13. Platter floor sherd. Mid-Augustan-Tiberian. Layer 9542 (Group 9016), Area D, unphased

14. Floor sherd, large platter; burnt. Mid-Augustan-Tiberian. Fill 14226, Pit 14225 (Group 36), Area L, Period 2A

15. Large (diameter 360mm) platter, Conspectus form 12.2; the fabric is rather pinker in tone than most of the material from the site. Mid- to late Augustan. Fill 20329, Pit 20481 (Group 42), Area L, Period 2A

16. Two small sherds, probably from one platter; the slip is mostly lost. Mid-Augustan-Tiberian. Fill 9011, Pit 8013 (Group 75), Area D, Period 2A

17. Conical cup fragment, form not identifiable, probably later Augustan-Tiberian. Fill 4497, Pit 4496 (Group 276), Area K, Period 2

18. Two sherds, probably the same vessel. Not terra sigillata but copying an Italian cup form with conical body and upright rim (cf. Conspectus form 23). The fabric is dark cream, with large inclusions of orange grog or haematite, and dense small mica flakes; the slip is thick, orange in colour but rather unevenly applied - perhaps deliberately, to give a marbled effect. The fabric suggests a Central Gaulish origin, the form a date in the early to mid-1st century AD. Layer 7445 (Group 871), Area G, Period 5; Fill 7453, Gully 7454 (Group 1349), Area G, unphased

The decorated ware and the potters' stamps by Brenda Dickinson

Cite this as: Dickinson, B. 2015, The decorated ware and the potters' stamps, in M. Atkinson and S.J. Preston Heybridge: A Late Iron Age and Roman Settlement, Excavations at Elms Farm 1993-5, Internet Archaeology 40. http://dx.doi.org/10.11141/ia.40.1.dickinson

Assemblage characteristics

A maximum of 4560 vessels, which can be assigned to a pottery or area of production, were recorded. Fifteen sherds that were either impossible to attribute to a source, or were not necessarily samian, have been omitted from the statistics. The 4560 vessels are divided by type, as follows:

Abbreviations: BR British; CG Central Gaulish; EG East Gaulish; SG South Gaulish;
AR Argonne; BAN Banassac; COL Colchester; HB Heiligenberg;
LG La Graufesenque; LM La Madeleine; LZ Lezoux; MN Montans;
MV Les Martres-de-Veyre; PUL Pulborough; RZ Rheinzabern; TR Trier

The decorated ware that could be relatively closely dated includes bowls with stamps of, or in the styles of, the following potters:

The most striking feature of the South Gaulish ware from La Graufesenque is that the f30 bowl form accounts for almost one quarter of the identified decorated ware and 37% of the commonest forms (these being bowls f29, f30 and f37). The cylindrical bowl f30 survived throughout the entire period of samian export, but normally makes only a modest showing against the more popular hemispherical f29 and f37. Most of the examples at Heybridge are Neronian and it is not impossible that they formed part of a single consignment. It is also curious that two complete bowls were buried together with a complete, but pierced, grog-tempered jar and a large lid. There are few other known instances of such an occurrence and no convincing explanation occurs (see also The structured deposition of pottery).

The globular jar, f67, is generally even rarer and the number found here (eleven (0.24%), with possibly three more), though modest, may seem worthy of mention.

The remaining eight South Gaulish pieces consist of a single vessel from Montans and seven from Banassac. First-century Montans ware is not particularly common in Britain, but small quantities have been noted in London and the west Midlands, particularly. The occurrence of Banassac ware is similarly sporadic, but rather more widespread, appearing on sites as far north as Carlisle and Old Penrith. The pieces found at Heybridge all belong to the first half of the 2nd century and it is not impossible that they arrived in a single consignment. The 1st-century assemblage also contained seven pieces of Lezoux ware, two of them decorated. This, again, does not occur in large quantities on any British site, apart from London, but its distribution is wide, and it has been noted on at least two sites in Scotland.

The samian supply to Heybridge diminished in the Trajanic period. This is not necessarily significant, as Trajanic ware from the Central Gaulish factory of Les Martres-de-Veyre was unevenly distributed in Britain, and a good many sites received noticeably less samian in the first two decades of the 2nd century than they did before or after. The same is apparent in the published samian from Colchester (Symonds and Wade 1999, 3 and 120). Increasing quantities of samian began to be discarded at Heybridge in the AD 120s with the arrival of Lezoux ware, but it was not until the early Antonine period that noticeably larger amounts were being discarded. This high level was maintained down to c. AD 180.

The bulk of the collection is 2nd century and so, as would be expected on a British site, consists mainly of Central Gaulish ware from Lezoux. This was supplemented by East Gaulish ware, particularly towards the end of the 2nd century and by British wares from Colchester and, probably, Pulborough. The 3rd-century supply is entirely East Gaulish. A notable aspect of the samian, which may again hint at the remains of a single consignment, is the popularity of the work of Cettus of Les Martres-de-Veyre, ten of whose bowls were found. Like some of the recovered plain ware from that factory, they are Hadrianic-Antonine in date.

The geographical position of Heybridge makes it a prime candidate to receive Colchester samian, and a maximum of 101 vessels was found. Only five of these were decorated. Unfortunately, the scarcity of comparable East Anglian assemblages makes it difficult to assess the importance of this find. The much smaller assemblage, of approximately 357 samian vessels recovered from Scole (Hartley and Dickinson 1977, 155-72) produced only 0.8% of the total vessels in Colchester ware, against 2.2% from Heybridge. This, of course, could be explained by its greater distance from the kilns. Colchester ware at Chelmsford was seemingly rare, amounting to just a trickle of sherds (Rodwell 1987b, 97). The other British pottery source whose wares almost certainly occur is Pulborough, represented by three vessels. The distribution of this pottery is confined mainly to Sussex, Hertfordshire and Essex, but occurrences are also known from Surrey, Norfolk (twice) and London, with one example as far afield as Sea Mills in Gloucestershire.

The earlier East Gaulish samian consists almost entirely of La Madeleine ware, which accounts for 7.2% of the East Gaulish assemblage. La Madeleine supplied samian to Britain in the Hadrianc and early Antonine periods and seems to have maintained a steady, if relatively modest, trade with the province. Also 2nd century are four vessels from the Argonne and single examples from Heiligenberg and, perhaps, Blickweiler. A larger proportion (14.9%), much of it 3rd century, was supplied by Trier, but the bulk of the East Gaulish ware (73.9%) comes from Rheinzabern and some of it is certainly 2nd century. A small proportion of the East Gaulish samian cannot be assigned to specific potteries.

Owing to the erosion of the sherds in much of this collection, it was only possible to detect cross-context joins between decorated pieces, though a few others were noted. No further work was done regarding cross-context joins since sherd links for the coarse pottery were only fortuitously recorded, because of the large volume of pottery involved.

Distribution

The highest concentration occurs in the central zone of the settlement, with a marked decrease eastwards and northwards (Table 12).

Distribution of the samian by area is shown in Table 13.

The fluctuations in the pattern of accumulation of discarded samian is normal for a British settlement occupied throughout the period of samian import into the province, with the bulk of the assemblage consisting of South Gaulish and Lezoux ware, with lesser contributions from Les Martres-de-Veyre, East Gaul and other factories whose wares are uncommon in Britain.

Summary

A small quantity of Arretine ware, with a date range of c. 20 BC to AD 25, was found. Unlike some of the other pre-Roman settlements in the south and south-east where it occurs, there is no contemporary samian. The absence (with a single exception) of typologically Tiberian forms is striking and, indeed, there seems to have been no significant build-up of discarded samian before c. AD 50. It had reached its peak in the 1st century by c. AD 60-65 and by c. AD 80 the level had dropped noticeably and did not recover until the early Hadrianic period. If the trend suggested by the samian is to be believed, the settlement did not suffer at the hands of Boudicca, and the burnt sherds of relevant date are no more than might be expected in a random collection of this size.

Most of the sherds have been dated according to the reigns of emperors or dynasties, but it was possible to date much of the decorated ware and the identified potters' stamps on plain ware more closely, and this is likely to give a truer picture. Figures 309 and 310 shows the discrepancies caused by the use of two different methods and the danger of assigning wide date ranges to large volumes of sherds. While the 1st- and 2nd-century peaks coincide, there is a conflict in the early 2nd century, with the broader dating suggesting an upsurge in discards c. AD 120, and a date five years later by the decorated and stamped vessels. The latter is more likely to be correct, reflecting a time when the supply of Lezoux samian to Britain had got into its stride. Similarly, the start of the final decline in the supply is more likely to have begun c. AD 180, as suggested by the decorated ware and potters' stamps, than AD 200, by which time the export of Central Gaulish samian to Britain had almost ceased and the quantities of East Gaulish ware reaching the province were comparatively modest.

Catalogue of decorated ware

D1 Bowl f29, South Gaulish. The distinctive feature of this bowl is the band of rouletting on the central cordon. This would normally indicate a Tiberian product, but the coarseness of the rouletting suggests a rather later date and the formal arrangement of the decoration is not typical of the pre-Claudian period. The lozenge-shaped leaf in the lower zone, used by more than one mould-maker, is on f29 bowls from Camulodunum (Hawkes and Hull 1947, pl. xxvi, no. 13) and Hofheim (Knorr 1952, Taf. 76E), cf. also D10, below. c. AD 40-55. Layer 13498 (Group 600), Area I, Period 3B; Fill 13545, Pit 13893 (Group 176), Area J, Period 2B; Fill 13546, Pit 13894 (Group 3024), Area J, Period 3; Fill 13547, Pit 13549 (Group 404), Area J, Period 2B; Layer 13576 (Group 600), Area I, Period 3

D2 Bowl f29, South Gaulish. The details in the upper zone are all on a bowl of the same form from Camulodunum, dated to the Claudian period by Hull (Hawkes and Hull 1947, pl. xxv, no. 21). The bud in this zone is on bowls stamped by Licinus, from Wiesbaden (Knorr 1919, Taf. 45A) and Colchester (Knorr 1952, Taf. 63E), the latter from a mould stamped by Volus. The large trifid motif in the lower zone is on a bowl from Camulodunum (Hawkes and Hull 1947, pl. xxxvii, no. 1) and the decoration of the whole zone may be identical to that of another from the same site (Hawkes and Hull 1947, pl. xxiii, no. 17). c. AD 45-60. Layer 24243 (Group 3044), Area M, Period 3

D3 Bowl f29, South Gaulish. The trifid motif, astragalus scroll-binding and rosette in the spiral are almost certainly the same as on a bowl from Colchester, which may have some connection with Murranus (Dannell 1999, fig. 2.34, no. 477). c. AD 45-60. Fill 17258, Pit 17412 (Group 330), Area Q, Period 2B

D4 Bowl f29, South Gaulish. The decoration of the lower zone consists of alternating panels containing saltires, involving tulip buds and trifid motifs, and single-bordered medallions. c. AD 50-70. Fill 15786, Pit 15757 (Group 900), Area M, Period 3; Fill 15892, Post-hole 15891 (Group 9013), Area M, unphased

D5 Bowl f29, South Gaulish. The delicate scroll in the upper zone, with its eight-petalled rosette, hollow bud and chevron-and-beads suggests a range c. AD 50-70. Fill 7267, Pit 7266 (Group 2114), Area G, Unphased

D6 Bowl f29, South Gaulish, stamped OFCRESTI retr. (see S31, below). The decoration of the lower zone is precisely matched on an f37 bowl of M. Crestio from the Cala Culip (Cap Creus) wreck (Mees 1995, Taf. 36, no. 7). The plant in the upper zone is also on one of his f37 bowls, from Castleshaw. The boar in this zone is probably Hermet 1934, pl. 27, no. 41. c. AD 65-85. Pit 20120 (Group 2042), Area L, Period 2B

D7 Bowl f29, Central Gaulish. The series of wavy lines, also on D16 below, occur on an f29 bowl at Lezoux with an internal stamp of Acapusos and on an f37 bowl from Colchester (GBS A526). Another vessel from Colchester, perhaps a lagena (1.81 G2850), has the tongueless ovolos. The animal in the upper zone is almost certainly a stag. The coarse, micaceous fabric and matt orange glaze belong to the 1st-century range at Lezoux. c. AD 70-85. Fill 14805, Pit 14806 (Group 707), Area L, Period 3

D8 Bowl f30, South Gaulish, in the style of Masclus. The single-bordered ovolo with rosette tongue (Dannell et al. 1998, fig. 2, IA) is probably the one on a signed bowl from London (Mees 1995, Taf. 107, no. 3). All the other motifs are known on his signed bowls. The scroll, frilled leaves, corded ring and ten-petalled rosette are on an (unprovenanced) bowl in the Kunsthistorisches Museum in Vienna (Mees 1995, Taf. 108, no. 1). The single medallion is on another bowl without provenance in Narbonne Museum (Fiches et al. 1978, fig. 14, no. 5), and the bud in the lower part of the medallion is on a bowl in a group of Claudio-Neronian samian found at Narbonne-La Nautique (Mees 1995, Taf. 113, no. 10). The dog and bird are not precisely identifiable. c. AD 50-65. Layers 14541, 14573, 14609 (Group 8005), unstratified, Area L

D9 Bowl f30, South Gaulish. The ovolo (Dannell et al. 1998, fig. 1, EE) occurs in association with motifs on f29 bowls stamped by a number of Claudio-Neronian and Neronian potters. It is on two bowls from La Graufesenque, one (Hermet 1934, pl. 73, no. 1) with the trifid motif and the chevron scroll (there used as an arcade), the other (Hermet 1934, pl. 73, no. 2) with the small rings. Though not strictly datable within the period c. AD 40-70, since the decoration of the f30 bowl is often archaic in style, this is most likely to be Neronian. Layer 5936 (Group 606), Area I, Period 3

D10 Bowl f30, South Gaulish, in the style of Martialis i. The ovolo (Dannell et al. 1998, fig. 1, Fdb) and lion with doe (Hermet 1934, pl. 25, no. 8) are on a signed bowl from Entraigues (Mees 1995, Taf. 103, no. 1). The trifid motif (Hermet 1934, pl. 14, no. 44), group of three (smaller) rosettes and, probably, the ten-petalled rosette are on a bowl from Usk with a mould-stamp of this potter (Mees 1995, Taf. 103, no. 4). The lion to left is a smaller version of Hermet (1934, pl. 25, no. 20), with a different tail, presumably added after the original was broken. The lozenge-shaped leaves are almost certainly the same as those on D1, above. c. AD 50-65. Layer 4148 (Group 732), Area K, Period 3

D11 Bowl f30, South Gaulish, in the style of Sabinus iii. The tongueless ovolo (Stanfield 1937, fig. 11, no. 24), polygonal leaf (Stanfield 1937, fig. 11, no. 90), outermost bud in the lower part of the scroll (Stanfield 1937, fig. 11, no. 58), innermost bud (Stanfield 1937, fig. 11, no. 86), larger striated spindle (Stanfield 1937, fig. 11, no. 61) and poppy-heads (Stanfield 1937, fig. 11, no. 73) are all known for Sabinus. The scroll, poppy-heads, spirals and the largest bud are all on a signed lagena from Rodez (Stanfield 1937, pl. xxi). c. AD 50-65. Layer 4148 (Group 732), Area K, Period 3; Layer 4518 (Group 8004), unstratified, Area K

D12 Bowl f30, South Gaulish, in the style of Masclus. The ovolo (Dannell et al. 1998, fig. 2, LL) and pointed leaf are on a signed bowl from Augst (Mees 1995, Taf. 110, no. 3), the arcade on one from La Graufesenque (Mees 1995, Taf. 109, no. 1) and the astragalus pillar on one from Richborough (Mees 1995, Taf. 111, no. 2). c. AD 50-65. Fill 8796, Pit 25221 (Group 663), Area P, Period 3

D13 Bowl f30, South Gaulish. The unusual leaf occurs on an f30 bowl from Aislingen (Knorr 1919, Taf. 95J). The decoration is also unusual in having saltires in adjacent panels. c. AD 50-65. Fill 20483, Pit 20199 (Group 336), Area L, Period 2B

D14 Bowl f30, South Gaulish. The lower zone features a large winding scroll with a chevron arcade in the lower concavity, containing a composite motif consisting of a beaded column supporting spirals and a trifid motif, and with tendrils with tulip buds springing from its base. Such arrangements are typical of the Neronian period. c. AD 60-80. Fill 4039, Pit 4008 (Group 756), Area K, Period 3

D15 Bowl f30, South Gaulish. The ovolo (Dannell et al. 1998, fig. 2, SE) appears on bowls with mould-stamps of Mommo and signatures of Memor, Primus iv and Tetlo. It occurs with the vertical panel of trifid motifs on a bowl from Colchester (Dannell 1999, fig. 2.29, no. 424), attributed to Memor. A slightly different inverted bifid motif was used by Primus iv, on a signed bowl from Camelon. The stag is on a signed f29 bowl of Mommo from La Graufesenque (Mees 1995, Taf. 147, no. 7). c. AD 70-90. Fill 10778, Ditch 25244 (Group 360), Area F, Period 3

D16 Bowl f37, Central Gaulish. The coarse wavy-line border is matched on a 1st-century Lezoux f29 bowl (D7, above). The leaf is perhaps the prototype for a range used at Lezoux in the Hadrianic and Antonine periods (Rogers G200-205). c. AD 70-110. Fill 20011, Pit 20010 (Group 707), Area L, Period 3

D17 Bowl f30, South Gaulish. The trident-tongued ovolo has not been recognised on stamped or signed bowls, but it, and the style of the decoration, suggest Flavian-Trajanic date. The couple in the first panel (O.374) are on a bowl from Aquileia (Knorr 1910a, Taf. VI, no. 3) and, unprovenanced, on one in Stuttgart Museum (Knorr 1910b, Taf. 1, no .8), which also shows the same column (Hermet 1934, pl. 16, no. 48). The erotic group (a larger version of Oswald, pl. xc, C) is on a bowl from Heidenheim (Knorr 1910a, Taf. VI, no. 1). The bird is probably Hermet (1934, pl. 25, no. 80). Oswald attributed the bowls from Heidenheim and Aquileia to Banassac because they have an ovolo which is known to have been used there, but it was also used at La Graufesenque, and the decoration points to manufacture there. c. AD 85-110. Fill 5146, Pit 5147 (Group 409), Area J, Period 3

D18 (Not illustrated) Bowl f37, Central Gaulish, in the style of Drusus (X-3) of Les Martres-de-Veyre. The decoration includes a blurred trifid motif impressed horizontally, followed by a tripod (Rogers Q15), over a wreath composed of anchor motifs (Rogers G395). For the wreath and tripod, see a bowl from London in the style of this potter (Stanfield and Simpson 1958, pl. 10, no. 121). c. AD 100-120. Fill 9344, Gully 9343 (Group 3063), Area D, unphased

D19 Bowl f37, from the South Gaulish factory of Banassac. The kneeling stag is a common figure-type, used at both La Graufesenque and Banassac and produced in many variants. There is no precise parallel for this one in the range illustrated by Hofmann (1988, 148, nos 217-221). Similarly, the lion to right is close to type 204, but is not exactly the same. The leaf, which is used both in the basal wreath and (with the tip only) as a filler, may be the same as one on a bowl from Banassac, where the wreath is formed by overlapping impressions of the base of the leaf, instead of its upper part. The ovolo and the upper wreath are too blurred to be attributable. The fabric and glaze of this small bowl make attribution to Banassac secure, despite the lack of parallels for the details. c. AD 120-150. Layer 4166 (Group 8004), Area K, unphased; Fill 4266, Post-hole 4265 (Group 3034), Area K, Period 3; 7000, unstratified, Area G

D20 Bowl f37, South Gaulish (Banassac). The ovolo is probably the one most often used by the Natalis Group, which appears on a bowl from Cannstatt with what are probably the same arrow-head motifs (Knorr 1905, Taf. X, no. 7). The animals are a bear to right (Hofmann 1988, 150, no. 261 and a boar to left (Hofmann 1988, 149, no. 238). The bear is on a bowl from Straubing by one of the Natalis group. The chevron wreath is not precisely paralleled, but many similar examples occur in the work of these potters. c. AD 100-150. Fill 10182, Ditch 25245 (Group 361), Area G, Period 3

D21 Bowl f37, Central Gaulish, in the style of Drusus of Lezoux. The rosette-tongued ovolo is not closely identifiable. The Vulcan (D.39 = O.66), slave (D.374 = O.647), dancer (O.363 variant) with lantern (Rogers Q65) and leafy column (a smaller version of Rogers Q5) are on bowls with mould-signatures from, respectively, Colchester (Stanfield and Simpson 1958, pl. 88, no. 1), Chester (Stanfield and Simpson 1958, pl. 88, no. 3), Southampton (Stanfield and Simpson 1958, pl. 88, no. 8) and Salzburg (von Koblitz 1926, Taf. III, no. 4). The eagle (not in D. or O.) is on an f30 bowl at Castleford, from a pottery shop destroyed by fire in the AD 140s (Dickinson and Hartley 2000, fig. 27, no. 518). On the Heybridge piece it apparently holds a lizard in its beak. The six-beaded rosette (Rogers C278) is Drusus's commonest one. c. AD 125-145. Fill 5146, Pit 5147 (Group 409), Area J, Period 3

D22 Bowl f37, Central Gaulish. The ovolo (Rogers B109), dog (O.1926A) and small medallion inside a larger one are all present in the work of Catussa. The medallions are on bowls and moulds with stamps of Rogers's Catussa II, which sometimes also carry mould-signatures of Gemenus (Rogers 1999, pl. 43). The dog is on a bowl from Lezoux with a mould-signature of Catussa and a mould-stamp of Cantomallus. The ovolo is on a signed mould of Rogers's Catussa I (1999, pl. 27, no. 2). It seems from the connections between the styles of Catussa I and II that they represent the work of two mould-makers working for the same man. The tree (Rogers N9) and the other motif (perhaps the obelisk Rogers P68) are not recorded for either style. c. AD 160-190. 4000, unstratified, Area A1

D23 Bowl f37, Central Gaulish. The mould-signature, ]an[ retr., upside down below the decoration, almost certainly belongs to Ianuaris i. Although the seated figure (D.527 = O.913) and the Pan-mask (D.675 = O.1214) are not known on signed bowls, all the motifs occur on signed bowls of Ianuaris, and the use of astragali (Rogers R7) placed diagonally across the borders is typical. The details are: single-bordered ovolo (Rogers B28), wavy-line borders (Rogers A24), eight-petalled rosette in a single medallion (Rogers C6), eight-beaded rosette (Rogers C281), beaded ring (Rogers C290) and leaf motif (Rogers L12). The borders, beaded rosettes, astragali and beaded rings in vertical series are on a bowl from Carlisle (Stanfield and Simpson 1990, pl. 170, no. 4) and the rosette in a medallion is on one from York. c. AD 125-150. Layers 24058 and 24138 (Group 8006), Area M, unphased

D24 Bowl f37, Central Gaulish, with mould-signature Tetturo... retr., upside down below the decoration. The details include: Pan (D.419 = O.717), putto (D.204 = O.394), leopard (D.969 ter = O.1564) five-petalled rosette (Rogers C120) and wavy-line borders (Rogers A26). The cornucopia is almost certainly Rogers U247, which occurs on a bowl in Tetturo's style in a pit at Alcester filled in the AD 150s (Hartley et al. 1994, fig. 50, no. 275). The festoon, a smaller version of Rogers F16, without the inner border, is on a bowl in his style from Corbridge. For further discussion of the potter, see S124, below. c. AD 130-160. Fill 20020, Pit 20019 (Group 711), Area L, Period 4

D25 Bowl f37, East Gaulish (La Madeleine). The ovolo (Ricken 1934, Taf. VII, C) appears on bowls from Camelon and Mumrills. The leaves consist of inverted trifid motifs (Ricken 1934, Taf. VII, no. 14) with added stems. Other details known to have been used at La Madeleine comprise a smaller trifid (Ricken 1934, Taf. VII, no. 24), acanthus (Ricken 1934, Taf. VII, no. 25) and festoon with acorn terminals (Ricken 1934, Taf. VII, no. 51). c. AD 130-160. Cleaning layer 5603, Area I

D26 Bowl f37, Central Gaulish, in the style of Cettus of Les Martres-de-Veyre. The ovolo and leaf (Rogers J144) are on a signed bowl from Les Martres (Terrisse 1968, pl.xx, no. 526) and the hare (O.2061) is on a bowl from Silchester with a mould-stamp in the decoration (Stanfield and Simpson 1958, pl. 141, no. 2). All the other details are on bowls, or moulds, in his style. They are: Apollo and chariot (not in D. or O.; Stanfield and Simpson 1958, pl. 141, no. 9, from Les Martres), small bear, D.820 = O1627 (Stanfield and Simpson 1958, pl. 141, no. 16, from London), large bear, D809 = O.1595 (Stanfield and Simpson 1958, pl. 144, no. 57, from Carlisle), lion and panther (D.766 = O.1450, D.809 = O.1570; both on a mould in Moulins Museum), 'tree' (Rogers Q5), on a bowl from Corbridge (Stanfield and Simpson 1958, pl. 143, no. 35) and medallion below the decoration, on a bowl from Carlisle (Stanfield and Simpson 1958, pl. 144, no. 57). The elongated, reversed S-motifs are extremely rare for Cettus, but occur on a bowl from Silchester. c. AD 135-160. Cleaning layer 5617 (Group 8002), Area I

D27 Bowl f37, Central Gaulish, in the style of Cettus of Les Martres-de-Veyre. The ovolo and bunch of grapes seem not to have been recorded for him before. The decoration includes a Pan (D.419 = O.717), as on a signed bowl from Colchester (Stanfield and Simpson 1958, pl. 141, no. 4) and centaur (D.436 = O.745), as on a bowl in his style from Leicester (Stanfield and Simpson 1958, pl. 142, no. 33). The large trifid motif (Rogers G13) is on a bowl from Corbridge (Stanfield and Simpson 1958, pl. 142, no. 23) and the small trifid (Rogers G340), which he rarely used, is on a bowl from Mumrills (Hartley 1961, fig. 80, no. 54). For the use of an astragalus across a panel border, see Stanfield and Simpson (1958, pl. 141, no. 14, from Corbridge). c. AD 135-160. Fill 9071, Ditch 9070 (Group 778), Area D, Period 3

D28 Bowl f37, Central Gaulish, in the style of Cettus of Les Martres-de-Veyre. The ovolo (Rogers B96) and a lion to left are on a bowl in his style from London. c. AD 135-160. Layer 5877 (Group 600), Area I, Period 3

D29 (Not illustrated) Bowl f37, Central Gaulish, in the style of Divixtus i. A panelled bowl, with 1) Double festoon or medallion, over a crouching panther (D.799 = O.1518), 2) Double festoon or medallion, over a small double medallion. The profusion of festoons and medallions and his typical ring-terminals (Rogers C132) make attribution to Divixtus certain, see Stanfield and Simpson (1958, pl. 116, no. 10, from Corbridge). c. AD 150-180. Fill 9444, Ditch 9434 (Group 777), Area D, Period 3

D30 (Not illustrated) Bowl f37, Central Gaulish, in the style of Divixtus i. A seated Abundance (D.472 = O.801) and a caryatid (D.656 = O.1199), in adjacent panels, are on stamped bowls from Corbridge (Stanfield and Simpson 1958, pl. 116, no. 17 and no. 10, respectively). Both bowls have his single-bordered ovolo with beaded tongue (Rogers B12). c. AD 150-180. Spread 7457 (Group 1323), Area G, unphased

D31 (Not illustrated) Bowl f37, Central Gaulish. The ovolo (Rogers B143), with straight line below, is on a bowl from Great Chesterford with a mould stamp of Secundus v (Simpson and Rogers 1969, fig. 2, no. 4). The Cupid with torches (D.265 = O.450) is on a stamped bowl from Toulon-sur-Allier and the goat (D.889 = O.1836) is on one from York. The warrior (D.117 = O.188) is on a bowl with the same ovolo, attributed by Rogers (1999, pl. 80, no. 21) to Pugnus, but more likely to be by Secundus. c. AD 150-180. Fill 20020, Pit 20019 (Group 711), Area L, Period 4

D32 Bowl f37, Central Gaulish. The ovolo (Rogers B223) is associated mainly with Cinnamus ii, but it appears, usually with a straight line below, on bowls in the style of Secundus v. The figure in the large medallion is a Cupid with torches, as on D31, above. The motif in the corner of this panel is probably the tail of a small dolphin (D.1057 = O.2401), which occurs on a stamped Secundus bowl from Great Chesterford (Simpson and Rogers 1969, fig. 2, no. 4). The figure in a festoon in the top of the panel containing a crouching lion (D.753) and a supine figure (D.553 = O.939) is almost certainly an Amazon (D.154 = O.243), which occurs on the same bowl. The supine figure occurs, with the ovolo and dolphin, on a bowl which is almost certainly by this potter. c. AD 150-180. Fill 10296, Ditch 10406 (Group 838), Area G, Period 5; Fill 16083, Well 6280 (Group 531), Area H, Period 3

D33 (Not illustrated) Bowl f37, Central Gaulish, with mould-stamp [CIN]NAMI retr.; Cinnamus ii of Lezoux, Die 5b. The panels include; 1) An ornament with leaves and dolphins (Rogers Q6), 2A) Double festoon, probably containing a bird, 2B) A dancer (O.819A, with broken left hand), 3) The potter's stamp and a triple leaf (Rogers L11), 4) A double-bordered medallion, containing a Pan (D.419 = O.717). c. AD 150-180. Layer 10310 (Group 1312), Area F, unphased

D34 Bowl f37, Central Gaulish. A freestyle bowl in the style of Cinnamus ii, with a stag (D.852 = O.1720), lion (perhaps not previously recorded) and dolphin (D.1050 = O.2382). The corn-stook which is exclusive to Cinnamus (Rogers N15) occurs on a stamped bowl from London, with the stag (Stanfield and Simpson 1958, pl. 163, no. 70). He is also known to have used the dolphin. c. AD 150-180. Fill 7154, Pit 7169 (Group 4048), Area G, unphased

D35 Bowl f37, Central Gaulish. The ovolo (Rogers B145) was used by Carantinus, Cinnamus ii and Illixo. As there is apparently a beaded border below it, the bowl is most likely to be by Cinnamus. c. AD 150-180. Cleaning layer 8239 (Group 8013), Area E

D36 (Not illustrated) Bowl f37, Central Gaulish. A panelled bowl, with 1) Minerva (a variant of D.77 = O. 126), owl (D.1020 = O.2331) and a trifid motif (Rogers H109), 2) An erotic group (a smaller version of Oswald, pl. xc, B). This is almost certainly by Cinnamus ii, who is known to have used all the details. c. AD 150-180. Cleaning layer 8239 (Group 8013), Area E

D37 (Not illustrated) Bowl f37, Central Gaulish. The ovolo (Rogers B182) is one of Cinnamus ii's less-common ones. The lower concavity of a winding scroll contains a kneeling stag (O.1704A) over an acanthus (Rogers K12), in a double medallion. The upper concavity has a polygonal leaf (Rogers J89) and another leaf (Rogers H13). The last is on a stamped bowl from London (Stanfield and Simpson 1958, pl. 161, no. 53) and all the other details, except for the polygonal leaf, are on another bowl from London (Stanfield and Simpson 1958, pl. 159, no. 25). c. AD 150-180. Fill 9444, Ditch 9434 (Group 777), Area D, Period 3

D38 Bowl f37, Central Gaulish. The ovolo (Rogers B85) appears on bowls in the styles of potters belonging to both the Cinnamus ii and Paternus v groups. This piece has links with the former, occurring on a stamped bowl of Cinnamus from Le Mans (Rogers 1999, pl. 32, no. 50a) and on one from Toulon-sur-Allier, by Secundus v. The figure-types are a philosopher (D.523 = O.905) and, probably, a dog (O.1974A), both known for Cinnamus. The candelabrum is made up of two elements which he used on stamped bowls, the dolphins on a basket (Rogers Q58; Stanfield and Simpson 1958, pl. 159, no. 26, from London) and the top part (Rogers Q43; Walke and Walke 1968, Taf. 36, no. 1, from Gauting). The six-beaded rosette (Rogers C278) is on a bowl in his style from Cambridge and the cornucopia (Rogers U245) is on one from Caerleon. On balance, therefore, this bowl is more likely to be by Cinnamus than Secundus, but the range will be c. AD 150-180, in either case. Cleaning layer 5617 (Group 8002), Area I

D39 Bowl f37, Central Gaulish, with mould-stamp in the decoration, ALBVC[I] (see S10, below). The details, a ring-tongued ovolo (Rogers B107), Cupid with torches (D.265 = O.450), bird (D.1010 = O.2316) and a baton (Rogers P3) have all been previously recorded for Albucius. c. AD 150-180. Layer 7073, Area G, unphased; Fill 7119, Pit 7118 (Group 852). Area G, Period 3

D40 (Not illustrated) Bowl f37, Central Gaulish, in the style of Albucius ii. Two sherds each show a panel, not necessarily adjacent, as follows: 1) Jupiter (D.4 = O.3), as on a stamped bowl from London (Stanfield and Simpson 1958, pl. 120, no. 5), 2) Venus (D.204 = O.338), apparently exclusive to Albucius, and appearing on a bowl from Bregenz (Stanfield and Simpson 1958, pl. 121, no. 16). c. AD 150-180. Fills 17188 and 17332, Post-hole 17230 (Group 942), Area Q, Period 3

D41 Bowl f37, Central Gaulish. A bowl in the style of Iullinus ii, with his smallest ovolo (Rogers B164). The figure-types include an Apollo (D.45 = O.77) and a dolphin (O.2394A), neither of which seems to have been recorded for Iullinus before. The pillar supporting the arcade (Rogers P21) is on a stamped bowl from Lezoux (Stanfield and Simpson 1958, pl. 125, no. 1). c. AD 160-190. 7000, unstratified, Area G; Fill 7071, Pit 7072 (Group 868), Area G, Period 4; Fill 7274, Gully 7273 (Group 1345), unphased

D42 Bowl f37, Central Gaulish, in the style of Censorinus ii. The layout of the decoration is typical of his work, particularly the use of an astragalus border (Rogers A10) below the ovolo (here Rogers B105) and of horizontal astragali to join borders, as on Stanfield and Simpson 1958, pl. 102, no. 11, from Corbridge. The motifs, a nine-petalled rosette (Rogers C194), column (Rogers P3) and astragalus (Rogers R7) are all known for Censorinus. The seated figure (D.527 = O.913) and the Mercury (D.289 = O. 529) seem to be unrecorded for him. c. AD 160-190. Fill 13818, Gully 13739 (Group 615), Area I, Period 3; Fill 13887, unstratified, Area J

D43 (Not illustrated) Bowl f37, Central Gaulish. The ring-tongued ovolo (Rogers B103) and zigzag border (Rogers A24) were used in conjunction by Martio i (Rogers's Martio II; cf. Rogers 1999, pl. 71, no. 5). c. AD 160-190. Fill 20013, Pit 20012 (Group 707), Area L, Period 3

D44 Bowl f37, Colchester ware. This appears to be in the style of Hull's Potter B, on whose moulds the rosette (1963, fig. 40, no. 70) and the roundels (Hull 1963, fig. 40, no. 79) appear. See figs 35, no. 8 and 36, no. 5, respectively. The border of squarish, separated beads seems to be a new detail for him. Potter B seems to have no connections with the group of East Gaulish potters who migrated to Colchester in the early Antonine period, and he is almost certainly later than them. c. AD 160-200? Fill 7123, Pit 7122 (Group 868), Area G, Period 4

D45 (Not illustrated) Bowl f37, East Gaulish (Argonne). The surface of the bowl is heavily eroded, but a row of lions can be seen, with other animals below. A basal wreath is composed of opposed bifid motifs (Oswald 1945, fig. 6, LVII), which occur on two bowls from Lavoye with mould-stamps of Tocca (Oswald 1945, fig. 9, no. 35-6). c. AD 140-180? Layer 9454 (Group 8012), Area D, unphased.

D46 Bowl f37, East Gaulish (Rheinzabern). A small bowl, probably in Style II of Belsus, though the details were all used by other contemporary potters. The ovolo (Ricken and Fischer 1963, E26), double medallion (Ricken and Fischer 1963, K20a) and stork with snake (Ricken and Fischer 1963, T221) are on a stamped bowl from Rheinzabern (Ricken 1948, Taf. 110, no. 8). c. AD 170-230. Fill 16083, Well 6280 (Group 531), Area H, Period 2

D47 (Not illustrated) Bowl f37, East Gaulish (Rheinzabern). The ovolo (Ricken and Fischer 1963, E7) and a leaf (Ricken and Fischer 1963, P79) are on a stamped bowl of Helenius (Ricken 1948, Taf. 175, no. 18). c. AD 170-240. Fill 8807, Pit 8748 (Group 44), Area P, Period 2A

D48 Bowl f37, East Gaulish (Rheinzabern). The details, a dolphin (Ricken and Fischer 1963, T194a), rosette (Ricken and Fischer 1963, O48) and corded border (Ricken and Fischer 1963, O248) were used by several Rheinzabern potters, but only Mammilianus seems to have used them all, and so the bowl is tentatively attributed to him. For the rosette and border, see a stamped bowl (Ricken 1948, Taf. 122, no. 1). The dolphin is on a bowl in his style (Ricken 1948, Taf. 123, no. 16). c. AD 180-240. Fill 7267, Pit 7266 (Group 2114), Area G, Period 2B

D49 Bowl f37, East Gaulish (Rheinzabern), with mould-stamp of Pervincus (S65). The ovolo (Ricken and Fischer 1963, E33) is on a bowl with the same stamp from Rheinzabern (Ricken 1948, Taf. 240, no. 10). The leaf (Ricken and Fischer 1963, P16) and the triple festoon (Ricken and Fischer 1963, O133) are on a stamped bowl from Heddernheim (Ricken 1948, Taf. 238, no. 2). The beaded festoon is similar to ones known for him. c. AD 200-240. Cleaning layer 5602 (Group 8002), Area I; Layer 13568 (Group 600), Area I, Period 3

D50 (Not illustrated) Bowl f37, East Gaulish (Rheinzabern). A bowl in Ricken's Julius I or Lupus style, with ovolo (Ricken and Fischer 1963, E46). The ovolo (Ricken and Fischer 1963, E42) and acanthus (Ricken and Fischer 1963, P145) are on a stamped bowl of Julius I (Ricken 1948, Taf. 154, no. 3) and a stamped mould of Lupus (Ricken 1948, Taf. 157, 7), both from Rheinzabern. The panel divider, on which the acanthus is set upside down, is Ricken and Fischer's O273, which occurs on an unstamped bowl in the same general style (Ricken 1948, Taf. 161, no. 10). c. AD 200-250. Fill 11139, Pit 10910 (Group 676), Area N, Period 5

D51 Bowl f37, East Gaulish, in Ricken's Victor II-Januco style at Rheinzabern. The ovolo (Ricken and Fischer 1963, E43) is on a stamped mould of Victor (Ricken 1948, Taf. 233, no. 1). The mask (Ricken and Fischer 1963, M5) and pillar (Ricken and Fischer 1963, O231) are on a mould in the style of these potters. The rosette above the pillar is apparently an unrecorded motif. c. AD 200-260. Fill 6012, Post-hole 6011 (Group 475), Area H, Period 7

D52 Bowl f37, East Gaulish. A substantially complete bowl in the style of Afer iii of Trier. The ovolo (Fölzer 1913, Taf. xxxii, no. 954), dog (Gard 1937, no. 76) and medallion are on a signed mould from Trier. The Diana with hound (Fölzer 1913, Taf. xxix, no. 478) is on a stamped bowl from de Meern (Holland). c. AD 200-260. Fill 10337, Ditch 10404 (Group 838), Area F, Period 5

D53 Bowl f37, East Gaulish, with a mould-stamp of Iulius viii (Ricken and Fischer's Julius II) of Rheinzabern (see S45, below). The decoration includes the ovolo (Ricken and Fischer 1963, E23), and a pedestal-like motif (Ricken and Fischer 1963, O161?) in an arcade (Ricken and Fischer 1963, KB73). See Ricken (1948, Taf. 205, no. 9) for a similar decorative scheme. c. AD 225-250. Layer 6020 (Group 573), Area H, Period 5

D54 Bowl f37, East Gaulish, in Ricken and Fischer's Julius II-Julianus style (= Iulius viii-Iulianus iii) at Rheinzabern. The pillar (Ricken and Fischer 1963, O221), Venus (Ricken and Fischer 1963, M51, but without the mask) and Hercules (Ricken and Fischer 1963, M86) are all on a stamped mould of Iulius (Ricken 1948, Taf. 208, no. 22). c. AD 225-260. Layer 12041 (Group 966), Area R, Unphased

D55 (Not illustrated) Bowl f37, East Gaulish. A badly eroded bowl in the style of Dubitatus of Trier. The ovolo is Fölzer 1913, Taf. xxxii, no. 954. The decoration seems to consist of two animals alternating, a stag (probably Gard 1937, no. 58, but with both front legs complete) and a dog (Gard 1937, no. 84). c. AD 225-260. Fill 6171, Pit 6169 (Group 546), Area H, Period 4

Potters' stamps

Each entry gives: Potter (i, ii, where homonyms are involved), die, form, reading, published example (if any), pottery of origin, date, context information. Stamps on the decorated ware, above, or from key pottery groups, are discussed in detail.

Superscript (a), (b) and (c) indicate:
(a) A stamp attested at the pottery in question.
(b) Not attested at the pottery in question, but other stamps of the potter known from there.
(c) Assigned to the pottery on the evidence of fabric, distribution, etc.
Ligatured letters are underlined

S1 Acurio 5a Cup f33 ACVRIO·I (Walke 1965, Taf. 40, no. 51-2) Lezoux^a. c. AD 150- 180. Fill 12009, Cremation Pit 12006 (Group 964), Area R, Period 3

S2 Advocisus 2a' Dish f79R or TgR [ADVO] ISI Lezoux^a. c. AD 170-190. Cleaning layer 14637 (Group 8005), Area L

S3 Aestivus 2a Cup f33 [A]IISTI[V]I:M Lezoux^b. c. AD 160-190. Fill 7126, Pit 7127 (Group 868), Area G, Period 4

S4 Aeternus 5a Dish f31 [AET]ERNII Lezoux^a. c. AD 155-185. Layer 10188 (Group 8014), Area F, unphased

S5 Albucianus 6a Dish f31 ALBVCIANI Lezoux^a. There are several examples of this stamp in a group of late Antonine samian recovered off Pudding Pan Rock, Kent. Otherwise, the only evidence of date is a single example on a dish f31R, from Catterick. However, the use of some of his other stamps on dishes f79 and f80 also points to a range c. AD 160-200. Fill 17037, Pit 17038 (Group 948), Area Q, Period 4

S6 Albucius ii 3b Cup f33 ΛLBVCIOF (Hartley and Dickinson 1981, 266, no. 2) Lezoux^a. There is no particularly useful dating evidence for this stamp of Albucius ii, but from his record in general Antonine activity is not in doubt. His wares occur at forts both on Hadrian's Wall and in Antonine Scotland and his stamps are found on a range of forms including cup f27, and dishes f31R, f79 and f80. c. AD 50-180. Fill 7390, Pit 7389 (Group 867), Area G, Period 4

S7 Albucius ii 6b Cup f33 AL[BVCI] Lezoux^a. This stamp is known in Antonine Scotland, from Balmuildy (Miller 1922, pl. xxxvii, no. 1), and an example in Chesters Museum almost certainly comes from one of the Hadrian's Wall forts. It was used on dish f42, which is unlikely to be later than c. AD 150, but also on dish Ludowici Tg, which suggests use of the die after AD 160. c. AD 150-180. Fill 4212, Pit 4211 (Group 756), Area K, Period 3

S8 Albucius ii 6b Dish f18/31R or f31R ALBVCI (Miller 1922, pl. xxxvii, no. 1) Lezoux^a. c. AD 150-180. Cleaning layer 5602 (Group 8002), Area I

S9 Albucius ii 6c Cup f33 ALBVC[I] Lezoux^a. c. AD 150-180. Layer 6053 (Group 506), Area H, Period 3

S10 Albucius ii 6h Bowl f37 ALBVCI (Stanfield and Simpson 1958, pl. 120, no. 1) Lezoux^a. Decorated bowls with this mould-stamp occur on Hadrian's Wall and in Antonine Scotland. There is also one in the Wroxeter Gutter deposit. c. AD 150-180. Layer 7073 (Group 1319), Area G, unphased

S11 Arilira 1a Dish f31R [ΛRI] I Λ (Dickinson 1986, 187, no. 3.12) Triera. c. AD 180-260. Fill 4798, Pit 4913 (Group 4016), Area K, unphased

S12 Attius ii 6a Dish f18/31-31 [ΛTTI]V ·EF Lezoux^b. c. AD 135-165. Fill 23123, Pit 23118 (Group 906), Area N, Period 3

S13 Banuus 3a Dish f31 BΛ VI·M Lezoux^b. c. AD 175-200. Layer 16081 (Group 8001), unstratified, Area H

S14 Banvillus 2a Cup f33 BANVI[LLIM] (Miller 1922, pl. xxxvii, no. 2) Les Martres-de-Veyrea. c. AD 130-155. Layer 12150 (Group 4031), unstratified, Area R

S15 Bio 2b Cup f24 BIOFECIT (Hull 1958, fig. 99, no. 3) La Graufesenquea^a. c. AD 50-70. Fill 7142, pit 7141 (Group 758), Area G, Period 2B

S16 Borillus i 10d Dish f18/31 BORI[L]LIM (Roosens 1976, Taf. 1) Lezoux^a. Graffiti cut inside the base, after firing, FIR and VET. c. AD 145-165. Layer 6316 (Group 547), Area H, Period 4; Fill 6603, Ditch 6825 (Group 3007), Period 3

S17 Calvus i 5b Platter f15/17 or f18 [OFCA]LVI, in a frame with swallow-tail ends (Bechert and Vanderhoeven 1988, Taf. 40, no. 95) La Graufesenquea^a. c. AD 70-90. Layer 13576 (Group 600), Area I, Period 3

S18 Calvus i 5m Platter f15/17 or f18 OFCALVI (Walke 1965, Taf. 40, no. 105a) La Graufesenquea^a. c. AD 70-85. Layer 4899 (Group 749), Area K, Period 3

S19 Campanus ii 2a Dish f79 [C]AMPANIO (Simpson 1987, 158, no. 34) Lezoux^a. c. AD 160-190. Fill 8009, Pit 9029 (Group 783), Area D, Period 3

S20 Celsus ii 1b Cup f27 [OFC]ELSI (Knorr 1921, Taf. IX, no. 46) La Graufesenquea^b. c. AD 80-110. Fill 6541, Pit 6543 (Group 654), Area H, Period 3

S21 Celsus iii 2a Bowl f38 CELSI M (Dannell 1971, 303, no. 24) Lezoux^a. c. AD 160- 190. Fill 6029, Pit 6030 (Group 553), Area H, Period 4

S22 Cerialis v 3a Bowl f37 CERI[ALIS] (Ludowici 1927, 240, c) Rheinzabern^a. c. AD 160-190. Fill 11302, Pit 11303 (Group 671), Area N, Period 5

S23-4 Cinnamus ii 5b Bowl f37 (2) CI[NNAMI] retr., [CIN]NAMI retr. (Walke 1965, Taf. 39, no. 11) Lezoux^a. Decorated bowls with this stamp occur frequently on Hadrian's Wall, but are even more common in Antonine Scotland. c. AD 150-180. Layer 4706, Area K, Period 3; Layer 10310 (Group 1312), Area F, unphased

S25 Cintussa 1a Dish f18/31R C·INT·VSSA Lezoux^c. c. AD 130-160. Fill 9016, Pit 9015 (Group 783), Area D, Period 3

S26 Cintusmus i 5a Dish f31R CINTVS[M] (Dickinson 1990, fig. 183, no. 11) Lezoux^a. c. AD 160-190. Fill 16083, Well 6280 (Group 531), Area H, Period 3

S27 Cobnertianus 1a Dish f18/31R-31R COBN[ERTIAN ] (Durand-Lefebvre 1963, 78, no. 238) Lezoux^c. c. AD 155-165. Cleaning layer 5602 (Group 8002), Area I

S28 Cotto i 3a Dish COTTO[F] (Bémont 1976, no. 133) La Graufesenquea^a. c. AD 45-65. Fill 13800, Pit 13809 (Group 397), Area J, Period 3

S29 Cracuna i 1a Cup f33 CRACVNA·F (Hartley 1972, fig. 81, no. 69) Lezoux^a. c. AD 125-155. Fill 20286, Pit 20185 (Group 711), Area L, Period 4

S30 Crestio 5b' Bowl f29 OFCRESTI[O] (Durand-Lefebvre 1963, 82, no. 250) La Graufesenquea^a. c. AD 50-65. Layer 6418 (Group 506), Area H, Period 3

S31 Crestus 1a Bowl f29 OFCRE TI retr. La Graufesenquea^a. Many instances of this stamp have been noted on sites founded in the early Flavian period, such as Rottweil, York and the Nijmegen fortress, and it is also known in Period IIA at Verulamium (Hartley 1972, fig. 81, no. 34). A single example of his work, a cup f24 with a different stamp, suggests some pre-Flavian activity. c. AD 65-85. Pit 21020 (Group 167), Area L, Period 2B

S32 Crestus 3a Platter f15/17 or f18 OΓ.[CRES] (Nash-Williams 1930, 173, no. 30) La Graufesenque^b. c. AD 70-85. Cleaning layer 5602 (Group 8002), Area I

S33-4 Divixtus i 9d Bowl f30 (2) [DIV]IX·F·, DIVI[ (Miller 1922, pl. xxxvii, no. 12). c. AD 150-180. Layers 6025 and 6118 (Group 573), Area H, Period 5

S35 Domitus i 1c Platter f15/17 or dish f18/31 DO[MITVSF]. Domitus i is known to have worked at both Les Martres-de-Veyre and, later, at Banassac. There is one example of this stamp from Banassac (Cavaroc 1964, no.30), but the fabric of the Heybridge dish and the heavy concentration of the stamp in Britain suggest that the die was also used at Les Martres. As several of his dies were used at both centres, this is not impossible. A dish in a London Second Fire deposit, from one of his other dies, was also made at Les Martres and, though unburnt, is almost the only evidence for the date of his activity there. c. AD 100-120. Fill 19150, Pit 19149 (Group 658), Area P, Period 3

S36 Donatus iii 1d Dish f31R [DONA]TVSF (Ludowici 1927, 214, c) Rheinzabern^a. c. AD 80-240. Fill 5864, Pit 5805 (Group 444), Area J, Period 6

S37 Felix i 2d Bowl f29 [OFFEI]CIS La Graufesenque^b. This stamp comes from a die that was used almost exclusively on f29 bowls. Dating relies largely on the decoration of these bowls, but the stamp occurs in the Boudiccan burning at Colchester (Hull 1958, fig. 99, no. 5). c. AD 55-65. Fill 9370, Pit 9218 (Group 768), Area D, Period 3

S38 Gabrus ii 2a Dish f31 GABRVS·F· (Hull 1963, fig. 48, no. 16) Colchester^a. The fabrics associated with this stamp, and its exclusively East Anglian distribution, suggest the die was used only at Colchester, though it is not impossible that this was the same Gabrus who worked at Trier and Lavoye. The forms stamped with Die 2a include dishes f18/31R, f31, f31R and f79/80, indicating activity mainly after c. AD 160. c. AD 160-190. Fill 7152, Pit 7118, Area G, Period 3; Fill 7202, Pit 7157 (Group 852), Area G, Period 3

S39 Gallio 1a Cup f33 Λ IO Lezoux^c. c. AD 140-200. Fill 8038, Pit 8037 (Group 789), Area E, Period 3

S40 Gippus 2a Cup f33 IPPI·M (Dickinson 1986, 189, no. 3.58) Lezoux^a. This stamp occurs in a group of samian from Tác (Hungary), almost certainly burnt in the Marcomannic Wars, and was also used on the mid-Antonine dish, f18/31R-31R. A range c. AD 155-185 is likely, therefore. Fill 7061, Pit 7060 (Group 310), Area G, Period 2B

S41 Gippus 2a Cup f33 IPPI·M (Dickinson 1986, 189, no. 3.58) Lezoux^a. c. AD 155-185. Surface 10104 (Group 813), Area F, Period 4

S42 Gnatos/Gnatius 7a Cup f33 NΛTOS Lezoux^c. c. AD 135-155. Cleaning layer 18737 (Group 8003), Area J

S43 Illixo 7a Bowl f38 ILLIXOF Lezoux^a. c. AD 160-180. Fill 7390, Ditch 7389 (Group 867), Area G, Period 4

S44 Iulius Numidus 2a Cup f33 IVL·NVMIDI Lezoux^b. c. AD 160-190. Ditch 5390 (Group 422), Area J, Period 4

S45 Iulius viii 3g Bowl f37 (mould stamp in the decoration) [I]VLIV[SE] retr. Rheinzabern^a. There is no internal dating evidence for this particular stamp, but the potter's work occurs in a group of wasters from Rheinzabern dated (provisionally) broadly c. AD 210/220-260 (Reutti 1983, 54-60). A range c. AD 225-250 seems appropriate for this bowl. Layer 6020 (Group 573), Area H, Period 5

S46 Iustus ii 3b Cup f33 IVSTI·M Lezoux^b. c. AD 160-190. Fill 10011, Pit 10012 (Group 811), Area E, Period 4

S47 Maceratus 2c Dish f31 MACERATI Lezoux^a. c. AD 150-180. Fill 8167, Well 8188 (Group 788), Area E, Period 3

S48 Macrinus iii 5b Dish f31 MACRINI (Walke 1965, Taf. 42, no. 209) Lezoux^a. c. AD 150-180. Fill 14634, Pit 14632 (Group 722), Area L, Period 6

S49 Magio i 1a Dish f31 ·[MΛGIONI·] (Dickinson 1986, 190, no. 3.85) Lezoux^b. Only the stop in the ansate beginning to the frame of this stamp survives, but it is distinctive enough to make attribution certain. The stamp is known on dish f31R and on f31 dishes dated to the later 2nd century. It has been noted from Chesterholm and Chesters. c. AD 160-190. Fill 4430, Pit 4487 (Group 739), Area K, Period 4

S50 Marcellus iii 11a Dish f18/31 or f31 MΛRCELLIVS (ORL B33, Taf. 19, no. 82) Lezoux^a. c. AD 130-155. Fill 9039, Pit 9038 (Group 806), Area D, Period 4

S51 Martialis ii 1b Dish f18/31-31 MA[R]TIALIS (Walke 1965, Taf. 42, no. 235) Lezoux^b. c. AD 125-145. Cleaning layer 5619 (Group 8002), Area I

S52 Martinus iii 7a Dish f31 M·ΛRTI (Durand-Lefebvre 1963, 143, no. 437) Lezoux^a. c. AD 160-190. Fill 8094, Well 8188 (Group 788), Area E, Period 3

S53 Masc(u)lus 19a Platter f15/17 MASCV[LVS] (Dannell 1971, 310, no. 64) La Graufesenque^a. The bulk of Masc(u)lus's output is Neronian, and this stamp has been noted in Period II at Verulamium (c. AD 60-75) and in the Oberwinterthur Keramiklager, destroyed in the early AD 60s. However, as it also occurs several times in Flavian contexts, it is likely to have continued in use in the AD 70s. c. AD 60-80. Fill 15787, Pit 15757 (Group 900), Area M, Period 3

S54 Masc(u)lus 19a Platter f15/17 or f18 [MASCV]LVS La Graufesenquea^a. c. AD 60-80. Fill 10376, Pit 10382 (Group 3068), Area F, Period 3

S55 Maternus iv 1a Dish f31R MΛTERNI Lezoux^a. c. AD 160-180. Fill 5676, Post-hole 13181 (Group 648), Area I, Period 6

S56 Miccio vii 1a Cup f33 MICCIO[·F] (Hull 1963, fig. 48, no. 26). The die from which this stamp came was used at both Sinzig and Colchester. This piece is in Colchester fabric. c. AD 150-180. Layer 6053 (Group 506), Area H, Period 3

S57 Minuso ii 1a Cup f33 MINVSOF (Hull 1963, fig. 48, no. 28a). The die from which this stamp came was used at Colchester, and also at Trier, where it occurs on dishes f18/31, f31 and, perhaps, f31R. This piece is in Colchester fabric. As the potter stamped the f27 cup at Trier, he will certainly have worked there first, but as he also stamped f32 dishes there, he will scarcely have left Trier before the middle of the 2nd century. This could be an instance either of a potter migrating to another area or sending a workman there with some of his existing dies. c. AD 155-170. Fill 7390, Ditch 7389 (Group 867), Area G, Period 4

S58 Minutus 3b Flat dish MINVTVSF Trierb. c. AD 180-240. 4000, Unstratified, Area A1

S59 Muxtullus 1a Dish f31 ·MVXTVLLI·M (Walke 1965, Taf. 43, no. 264) Lezoux^a. c. AD 160-180. Surface 6289 (Group 514), Area H, Period 3

S60 Muxtullus 1b Cup f33 [MV]XTVLLIM (Walke 1965, Taf. 43, no. 262) Lezoux^b. c. AD 130-150. Layer 12093 (Group 970), Area R, unphased

S61 Niger ii 3b'' or 3b''' Platter f15/17 or f18 FN[GR<I> La Graufesenquea^a. The stamp is more likely to have come from Die 3b'', since no examples of 3b''' have been noted on dishes. c. AD 55-70. Fill 18225, Post-hole 18242 (Group 9003), Area J, unphased

S62 Paterclos/Paterclus ii 6a Dish f18/31 [PΛTERC]LIM (Dickinson 1986, 193, no. 3.135) Les Martres-de-Veyreb, c. This potter is known to have worked at both Les Martres and Lezoux. The dish is in one of the fabrics in the Les Martres range. c. AD 100-125. Fill 13884, Pit 13883 (Group 595), Area I, Period 3

S63 Paterclos/Paterclus ii 10a Dish f18/31 [PΛTE]RCLOSFE (Allgaier 1992, no. 78) Les Martres-de-Veyre. c. AD 100-110. Layer 11206 (Group 8007), unstratified, Area N

S64 Pentius 1a Dish f79 or Tg PIINTII·M[Λ ] Lezoux^c. c. AD 160-190. Layer 6053 (Group 506), Area H, Period 3

S65 Pervincus 3f Bowl f37 PIIRVINCVS retr. (With N reversed) (Ludowici 1927, 243, d) Rheinzabern^a. c. AD 200-240. Cleaning layer 5602 (Group 8002), Area I

S66 Perpetus 5c Dish or bowl PERPETVS (Ludowici 1927, 226, e). The die for this stamp is known to have been used at Rheinzabern, like many of his others. However, the decoration of the bowl and the fabric both suggest origin at Trier. c. AD 180-240. Fill 15353, Pit 15354 (Group 701), Area M, Period 6; Fill 15355, Pit 15356 (Group 701), Area M, Period 6

S67 Pistillus 4a Cup f33 PISTILLII Lezoux^a. Apart from single examples of dishes f79/80 and f80, all the stamps recorded from this die are on f33 cups. There are nine cups in the Wroxeter Gutter deposit and one from Haltonchesters. This is the only example to show double I at the end of the stamp, the rest reading PISTILLI. The intrusive (fainter) stroke is presumably due to a scratch on the die. c. AD 160-190. Fill 4579, Pit 4526 (Group 729), Area K, Period 3

S68 Pont(i)us 8h Cup f27g O ONTI (Dickinson 1986, 193, no. 3.154) La Graufesenquea^a. Most of Pont(i)us's output is Flavian, with stamps from other dies occurring at sites such as Cappuck, Inchtuthil and the Saalburg. This stamp occurs mainly on f27 cups, but a few examples on cup f24 suggest that the die was in use in the late Neronian period. c. AD 65-90. Fill 5146, Pit 5147 (Group 409), Area J, Period 3

S69 Pridianus 7a Dish f18/31R or f31R PRIDFEC La Madeleineb. c. AD 130-160. Layer 4706 (Group 905), Area K, Period 3

S70 Priscinus 4b Dish f31 PRISC.../SF Lezoux^b. c. AD 150-170. 15000 (Group 8006), unstratified, Area M

S71-2 Reburrus ii 3a Cup f33, Dish f31 REBVRRI·OFF, REBV[R]RI·OFF (Dickinson 1996, fig. 143, no. 74) Lezoux^a. c. AD 145-170. 3830, unstratified, Trial Trench 5; 4000, unstratified, Area A1

S73-4 Reginus ii 1a Dish f18/31R, Cup f33 [REGI]NI·M, REGI[N]I·M (Hartley 1961, 107, no. 7) Les Martres-de-Veyreb. c. AD 115-145. 3999, unstratified, Area X; Cleaning layer 5610 (Group 8002), Area I

S75 Roppus ii 1a Dish f18/31 RO[PPVSFE] (Hartley 1970, 26, no. 56) Les Martres-de-Veyreb. c. AD 115-135. Fill 23019, Pit 23012 (Group 694), Area N, Period 4

S76 Rottalus 1a Dish f31R [RO]TTΛLIM (Dickinson 1986, 194, no. 3.176) Lezoux^a. Recorded from Benwell, Chesters and the Brougham cemetery, where most of the Lezoux samian is late 2nd century, and on dishes f79, f79R, f80 and Ludowici Tg, this stamp clearly falls within the period c. AD 160-200. Fill 8802, Pit 8801, Area P, Period 6

S77 Ruffus ii 2a Dish 18/31 RVFFI·M (Curle 1911, 240, no. 82) Lezoux^a. c. AD 125-145. Fill 12215, Cremation pit 12219 (Group 964), Area R, Period 3

S78 Rufinus iii 4c or 4c' Cup f27g [OF]RVFI[N] or [ F]RVFI[ ] (Durand-Lefebvre 1963, 204, no. 634) La Graufesenquea^a. c. AD 65-85. Fill 7623, Ditch 25045 (Group 5), Area G, Period 2A

S79 Rufus iii 3b Platter f18 [OFRV]FI (Ribeiro 1959, VI, no. 55) La Graufesenquea^a. c. AD 70-90. Layer 5693 (Group 609), Area I, Period 3

S80 Sacerus ii Uncertain 1 Dish f31 SΛCERIKI Lezoux^c. c. AD 170-200. Fill 4844, Pit 4913 (Group 4016), Area K, Period 4

S81 Severinus iii 3a Dish f32 SIIVIIRINVS (Ludowici 1927, 230, c) Rheinzabern^a. c. AD 180-240. Cleaning layer 5602 (Group 8002), Area I

S82 Sextus v 4c Dish f31 [SEXT]I·MA Lezoux^b. c. AD 160-200. Cleaning layer 5602 (Group 8002), Area I

S83 Sulpicius 3a Platter f18 OFSVIPIC La Graufesenquea^a. c. AD 80-110. Layer 5693 (Group 609), Area I, Period 3

S84 Sulpicius 8j Cup f27 SVLPICI La Graufesenquea^a. c. AD 80-110. Fill 24198, Pit 24197 (Group 696), Area M, Period 4

S85 Tarvillus 1b Cup f33 (complete) TΛRVILLIM Lezoux^c. c. AD 135-150. Fill 12200, Cremation pit 12203 (Group 964), Area R, Period 3

S86 Tasgillus ii 4b Cup f27 TASCI IV retr. (Terrisse 1968, pl. liv, second example). The die for this stamp was used at both Les Martres-de-Veyre and Lezoux. This pot was made at Les Martres. c. AD 110-125. 17000 (Group 8024), unstratified, Area A4

S87 Tauricus i 1a Cup f33 TAVRICIOF Lezoux^b. c. AD 150-180. 4001 (Group 8021), unstratified, Area A1

S88 Verus vi 2c Dish f31R ]/ERVSFEC, in guide-lines (Ludowici 1927, 232, b). The die for this stamp was used at both Rheinzabern and Trier. This piece comes from Rheinzabern. c. AD 180-240. Cleaning layer 5427 (Group 8002), Area I

S89 Virio-- 1a Platter f15/17 VIRIO La Graufesenquec. c. AD 55-75. Fill 4842, Pit 4843 (Group 1147), Area K, unphased

S90 Virthus 2a Dish R VIRTHVSFEC+ La Graufesenquea^a. c. AD 50-70. Cleaning layer 5611 (Group 8002), Area I

S91 Virthus 3a' Cup f27 <VI>RTHVS[FE<C>] La Graufesenque^b. c. AD 60-75. Layer 6420 (Group 509), Area H, Period 4

S92 Vitalis iii 2a Platter f15/17 or dish f18/31 [V]+A[LISM·S·F (Hartley 1972, 233, S8) Les Martres-de-Veyrea. c. AD 100-120. Layer 6316 (Group 547), Area H, Period 4

S93 ]CIS? on Platter f15/17 or f18, South Gaulish. Neronian. Fill 407, Pit 410, Area W, Period 3

S94 ]VC? on Cup f27g, South Gaulish. Neronian. Cleaning layer 14573 (Group 8005), Area L

S95 MS[ on Platter f15/17 or f18, South Gaulish. Neronian. Fill 10609, Post-hole 10745 (Group 822), Area F, Period 4

S96 ]I or I[ on Platter f15/17R or f18R, South Gaulish. Neronian or early-Flavian. Fill 13639, Pit 13640 (Group 594), Area I, Period 3

S97 ..OF..NI..on Cup f27, South Gaulish. Neronian or early-Flavian. Fill 8537, Pit 8524 (Group 656), Area P, Period 3

S98 III[ or ]III on Cup f27g, South Gaulish. Neronian or early-Flavian. Fill 20117, Pit 20116 (Group 2042), Area L, Period 2B

S99 OFPO[? on Bowl f29, South Gaulish. c. AD 70-85. Cleaning layer 5597 (Group 8002), Area I

S100 V·IIN or V·IIV on Platter f18, South Gaulish. Flavian. Layer 5149 (Group 3024), Area J, Period 3

S101 ]/I on Platter f18, South Gaulish. Flavian. Fill 20180, Pit 20174 (Group 707), Area L, Period 3

S102 VN[ on Cup, South Gaulish. Flavian or Flavian-Trajanic. Layer 5693 (Group 609), Area I, Period 3

S103 IIIV...VN retr. on Dish f18/31 (complete), Central Gaulish. Hadrianic. Fill 12199, Cremation pit 12203 (Group 964), Area R, Period 3

S104 CN[ retr. (?), mould-stamp in the decoration, Central Gaulish. Hadrianic or Antonine. Fill 8737, Pit 8736 (Group 902), Area P, Period 6

S105 SI[ or SE[ on Dish f31, Central Gaulish (Les Martres-de-Veyre). Hadrianic-Antonine. Fill 3587, Cremation Pit 3585 (Group 317), Area W, Period 2B

S106 ]I I\/ on Dish f31, Central Gaulish. Early to mid-Antonine. Cleaning layer 5610 (Group 8002), Area I

S107 GI[ on Cup f33, Central Gaulish. Early to mid-Antonine. Fill 10182, Ditch 25245 (Group 361), Area G, Period 3

S108 IXX[ or ]XXI on Dish f18/31R or f31R, Central Gaulish. Antonine. Fill 5537, Post-hole 5538 (Group 395), Area J, Period 3

S109 IVL[ on Dish f31, Central Gaulish. Antonine. Layer 6226 (Group 573), Area H, Period 5

S110 ]NI·M on Dish f31, Central Gaulish. Antonine. Layer 6227 (Group 573), Area H, Period 5

S111 ]IM or IMΛ on Dish f31, Central Gaulish. Antonine. Layer 10289 (Group 820), Area F, Period 4

S112 SE[ on Dish f31, Central Gaulish. Antonine. Fill 16148, Pit 16149 (Group 559), Area H, Period 4

S113 MAC[ on Cup f33, Central Gaulish. Antonine. Fill 10182, Ditch 25245 (Group 361), Area G, Period 3

S114 Perhaps TITVLI[ or TITVLL[ on Cup f33, Central Gaulish. Antonine. Layer 16187 (Group 573), Area H, Period 5

S115 SAT[ on Dish f31, Central Gaulish. Slightly burnt. Mid- to late Antonine. Fill 8094, Well 8188 (Group 788), Area E, Period 3

S116 Possibly ]ENIL.. \TI on Dish f31 or f31R, Central Gaulish. Mid- to late Antonine. Fill 5800, Well 5806 (Group 639), Area I, Period 4

S117 ΛT....M on Dish f31R, Central Gaulish. Mid- to late Antonine. Surface 7549 (Group 862), Area G, Period 4

S118 C[, O[ or ]O on Dish f31R, Central Gaulish. Mid- to late Antonine. Layer 15035 (Group 8006), Area M, unphased

S119 ]ΛPV [ (?) on Dish with concave base, East Gaulish (Rheinzabern). Late 2nd or first half of 3rd century. Fill 15073, Pit 15005 (Group 701), Area M, Period 6

S120 ]MITIV[? on Dish f32, East Gaulish (Rheinzabern). Perhaps a stamp of Primitius, but the reading is not certain. First half of 3rd century. Fill 4994, Pit 4989 (Group 1147), Area K, unphased

Graffito

S121 SE[ (S reversed) on Cup f27, Central Gaulish. Hadrianic or early Antonine. Inscribed in the centre of the base before firing, instead of a potter's stamp. Fill 5157, Pit 5158 (Group 409), Area P, Period 3

Mould signature

S122 ]an[ retr. on bowl f37, Central Gaulish. The signature was inscribed upside down below the decoration, before the mould was fired. This could in theory belong to either Quintilianus i or Ianuaris i, whose styles of decoration are sometimes indistinguishable, and who are both known to have used signed moulds. However, the lettering makes attribution to Ianuaris almost certain. c. AD 125-150. Layer 24058 (Group 8006), Area M, unphased

S124 Tetturo... retr. on bowl f37, from a mould inscribed upside down below, the decoration, before firing. Signed bowls of Tetturo are known from Toulon-sur-Allier and Rogers (1999, 254) believed that he only worked there. However, much of his known output is in Britain, and it is likely that he also worked at Lezoux at some stage in his career. The fabric of the Heybridge piece is rather redder than normal for Lezoux, but that is insufficient evidence to attribute it to Toulon. The provenance must remain doubtful, therefore. Bowls in his distinctive style from Alcester (in a pit filled by c. AD 160), Camelon, Corbridge and Inveresk suggest a range c. AD 30-160. Fill 20020, Pit 20019 (Group 711), Area L, Period 4

An analysis of some aspects of the samian pottery by Steven Willis

Cite this as: Willis, S. 2015, An analysis of some aspects of the samian pottery, in M. Atkinson and S.J. Preston Heybridge: A Late Iron Age and Roman Settlement, Excavations at Elms Farm 1993-5, Internet Archaeology 40. http://dx.doi.org/10.11141/ia.40.1.willis

Brenda Dickinson's report provides a substantive guide to the samian pottery recovered, identifying major trends and including a comprehensive catalogue. Her work enables some further aspects of the use and consumption of samian to be considered and compared with wider patterns discernible in Roman Britain.

Samian is, of course, a particularly useful artefact class for the archaeologist given: (i) its standardisation of form and fabric, (ii) its sequential typological development, change in decorative detail and stamps, that are well understood and facilitate relatively close dating, and (iii) its wide distribution and reporting, which enable comparative analysis. Various studies have demonstrated that samian was particularly valued or prized among contemporary communities. Distinctive and unusual in appearance compared to other contemporary pottery types, samian is perceived, by archaeologists, to have been a high-status commodity. Certainly across Britain, samian was in the vanguard of imports arriving at indigenous sites in the years following the conquest and circulated in a manner different from other pottery types (cf. Willis 1997a; 1998); some samian assemblages even appear to represent 'diplomatic' gifts (Haselgrove et al. in press).

Evans' study of graffiti on Roman pottery (Evans 1987) has shown that samian was much more frequently inscribed with names and marks than other pottery types, with marking evidently expressing a concern to denote ownership (see this publication for an alternative interpretation of X-graffiti ). Further, studies of the repair of broken pottery vessels via lead riveting or cleats shows that samian was repaired with disproportionate frequency compared to other types of vessels (Marsh 1981, 227; King and Millett 1993, table 16.5; Evans 1996a, 89; Evans 1996b, 62; Booth 1997, 123). In other words, a broken samian vessel was more likely to be repaired than any other type of pot, and this is a widespread pattern, identifiable at all types of site. The large majority of decorated samian vessels are bowls and it seems that these vessels may have been particularly valued, either because they were relatively expensive (reflecting the amount of labour taken to produce them, plus transport costs) or perhaps because the form and finish were attractive for use as (communal?) drinking vessels. Overall scrutiny of the incidence of samian seems to confirm that it was, indeed, a status symbol, relating to wealth, social and cultural identity (Willis 1998). Samian, therefore, can be a sensitive indicator of a range of processes to a degree that is not possible with other pottery types of the period. For these reasons it is potentially useful to explore further the character of the samian assemblage recovered, with the advantage of such a large sample of this pottery having been recovered.

Recent studies have shown consistent differences in the pattern of samian occurrence that seems to reflect site type strongly. Major towns and sites associated with the Roman military show a pattern of more frequent samian use and deposition, while at others (small towns, roadside settlements, religious foci and rural sites) samian is much less frequent relative to other pottery types. Variations in the frequency of samian at different settlement types appear to have been socially structured. Differences in access to quantities of samian may have had an economic basis (i.e. it may not have been easily affordable), and/or particular cultural attitudes may have been influential (e.g. at some types of site samian may have been considered an important regular mechanism of status display among certain social groups, and a symbol of 'Romantitas' and wealth, while at others, people may not have used sets of samian as a regular everyday status symbol, but less often, for specific purposes or events). Comparison of the samian assemblage from Elms Farm with patterns recently identified for other sites in Britain is potentially instructive.

Samian as a component of pottery groups

An analysis of the frequency of samian among pottery groups by phase

In order to establish the nature of samian use and consumption at Heybridge, several approaches have been adopted. A useful index of the general pattern of samian use and consumption can be established by quantifying its occurrence within a series of phased pottery groups. As noted elsewhere, the key groups are considered to be both representative of the pottery being consumed at the site through various phases and 'good groups' from a methodological point of view. The data from these groups have been supplemented here using those from the 'wider pool', that is, further good, representative and stratified groups, which are also used elsewhere in the analysis of the pottery assemblage. Combining the quantitative information from these groups by phase results in unusually large samples, which should provide a reliable guide to the overall frequency of samian consumption through time. Data on the occurrence of samian within these groups are presented in Tables 14 and 15. The data are derived only from groups in which samian sherds are present (which is a function of the data available to this author), and exclude, also, groups that appear to include structured special deposits, or, in one or two cases, some intrusive sherds. In effect, these criteria omit only a few of the key and wider pool groups (omissions from the key groups comprise the following: Ceramic Phase 4; KPG17 (20008), which appears to represent a structured element, and KPG15 (9214), which included no samian: Ceramic Phase 7; KPG26 (6280) which also seems to represent a structured element, and KPG27 (1589), which included no samian). That a small number of groups with no samian have been excluded from the analysis means, of course, that any figures indicating the percentage of samian pottery within phases slightly over-represents the actual frequency of this ware per phase vis-à-vis other pottery wares.

Table 14 shows the frequency of samian among pottery groups by phase when EVE is the measure; Table 15 shows the equivalent data where weight is the measure. Weight proportions are a good measure when the intention is to compare the composition of groups over time, between sites, etc., while EVE is also suitable for this purpose, it gives an impression of vessel turnover (cf. Orton 1989). Orton (1989) has recommended that both these measures be employed where possible. For reasons discussed below, one should not anticipate that these two measures will yield like results in terms of percentage figures etc. though they should, significantly, show similar trends. The data reproduced in Tables 14 and 15 show that despite the fact that, in absolute terms, a very large sample of samian was collected during fieldwork, it forms only 2.2% by weight of the pottery being deposited. It is noteworthy that although the quantities of samian within these groups are invariably modest, it is rarely absent from sizeable groups.

Considering, firstly, Table 14, a general pattern of low percentages is consistent through time and presumably reflects a comparatively low frequency of use of samian ware. Samian is present in Ceramic Phase 3 but forms only a tiny fraction of the pottery of that phase; this is not surprising since, away from Roman military sites and major aggregated centres, samian is generally only occasionally found among contexts dating to the mid-1st century AD (cf. Willis 1997a). The samian percentages for Ceramic Phases 4 and 5 are remarkably similar to each other and imply continuity in the consumption and deposition of samian through the early Roman period. These data are likely to be a reliable index, given the robustness of the sample (for these are large samples and combine data from six and nine good groups respectively).

Samian data from other sites in Britain (e.g. Marsh 1981) indicate that there was a peak in the supply and consumption of samian in Britain during the Flavian period (c. AD 70-100). Ceramic Phases 4 and 5 overlap with this peak, and it is possible that any effects on consumption at Heybridge during periods when the general supply of samian in Britain was lower, namely during the Neronian period (c. AD 55-70) and the Trajanic-early Hadrianic period (c. AD 100-125), are masked within the generality of the ceramic phasing. Brenda Dickinson notes that the period c. AD 100-120, during the time when the principal source of samian was Les Martres-de-Veyre, was one of markedly limited samian supply to Britain. Analysis of the samian from specific sites (e.g. Southwark), as well as general trends in deposition (cf. Willis 1998, 102-5), indicates that there was probably some careful curation of older samian vessels during this period of low supply. Again, this may be reflected in the consistency of the samian data from Ceramic Phases 4 and 5.

Moving into the 2nd century, the proportion of samian among the sample relating to Ceramic Phase 6 is 8.3%. This represents a doubling of the percentages for Phases 4 and 5, but probably does not reflect a change in cultural practice at Heybridge, but rather may relate to the general increase in samian appearing in Britain from the early Antonine period (from c. AD 140), supplemented in the Essex region by local production of Colchester samian from c. AD 155 (cf. Tyers 1996, 114-16). The frequency of samian dips dramatically, and unexpectedly, in the sample relating to Phase 7. It is likely that this anomaly is due to the sample being less representative than others, for it comprises, regrettably, data from just three groups, with one group numerically dominant (KPG25). It should be noted, however, that there is less pottery of all types in CP7.

The proportion of samian in the sample from Ceramic Phase 8 returns to a similar level as that of Phase 6. The majority of the samian from this first half of the 3rd century phase is Central Gaulish. One might suspect that a proportion of these items are residual from earlier phases, but it is also likely that the normal 'life cycle' of some later Central Gaulish imports meant that many of these vessels were still extant into this period, and, again, curation of fine ware vessels in a period of low samian supply is also highly probable.

When weight is the measure, a general similarity in trends is apparent (cf. Table 15). Percentages are very low by this measure. This is to be expected since samian is a fine ware and includes a range of comparatively thin-walled and small forms, including cups and dishes. Such vessels, when broken, produce comparatively light sherds when compared to other types such as some storage and cooking jars whose sherds may be individually comparatively heavy. What is particularly archaeologically significant is not that there may be correspondence between the results of these measures in terms of actual figures and percentages, but that they show trends in the relative frequency of pottery that assists comparative analysis (cf. Orton 1978).

Considering Table 15 overall, there is good agreement with Table 14 in terms of the comparative frequency of samian. The weight data confirm that samian comprises only a tiny fraction of the pottery of Phase 3. The samian percentages for Ceramic Phases 4 and 5 are again consistent, suggesting continuity in samian consumption and deposition through the early Roman era. Moving into the 2nd century, the proportion of samian among the sample relating to Ceramic Phase 6 is 1.8%. This proportion is very small, though as with the EVE measure this represents a doubling of the percentages for Phases 4 and 5. The frequency of samian dips in the sample relating to Phase 7, as it does when EVE is the measure, though by weight it is less marked than with EVE data. The percentage is higher than in the 1st-century groups. Again, the proportion of samian in the sample from Phase 8 returns to a similar level as that of Phase 6, a pattern seen also by EVE.

Comparison of the Elms Farm data by phase with data for other sites in Roman Britain

It is instructive to compare these data with others in Roman Britain. Table 16 documents the average percentages that samian pottery comprises, by weight, within groups from a variety of site types. This provides an approximate guide to the general frequency of samian c. AD 40 to 200 (unfortunately, there are insufficient data yet published nationally to enable a similar table to be generated where EVE is the measure, though see Table 14). Work for the English Heritage funded Samian Project has shown that there is a strong correlation between the status and identity of sites and the proportion of samian within the groups recovered from such sites (Willis 2005). By this measure, Heybridge compares well with the pattern seen at Roman 'small towns', roadside settlements and, indeed, rural sites.

Considering pottery groups from sites of 'small town' or roadside settlement status, several cases of higher frequencies of samian than at Heybridge can be noted. At Braintree, for instance, conventionally seen as a Roman small town, samian formed 5.8% of the pottery from well 102 and ditch 307 (group size: 4.4kg) at College Road, dated c. AD 150-250 (Martin 2000a). At Meole Brace, Shropshire, samian accounted for an unusually high 7.5% of the pottery from Phase 3 (group size: 15.4kg), c. AD 210/220-230 (Ellis et al. 1994). Similar percentages occur among groups from roadside settlements, for example in the case of a series of groups at Neatham, Hampshire (Millett and Graham 1986; Willis 1998, table 1), at Pomeroy Wood, east Devon, where samian forms 2.9% of the pottery from Phase 4i (47.4kg), dated c. AD 90/100-260 (Fitzpatrick et al. 1999), and, in Oxfordshire, at Wantage, Mill St, where samian accounted for 2.2% of the 8.4kg of pottery from Phase 1, dated to c. AD 70-160 (Holbrook and Thomas 1996). In Essex, a samian percentage of 1.1% is recorded from a pit at Great Dunmow (gravel pit 857) for the period c. AD 190-230/240 (Going and Ford 1988; Willis 1998, table 1) consistent with contemporary levels at Elms Farm.

Two samples from Coggeshall show marked variation, though this evidently relates to the context of these finds. A lower percentage than at contemporary deposits at Elms Farm is recorded from the St Peter's School site, where samian formed a mere 0.02% of the pottery from Phase 4.1 (16.1kg) dated to c. AD 65-150. This probably reflects the likelihood that the fieldwork was located away from the settlement nucleus (Clarke 1988). Conversely, at The Lawns, a sample dated to c. AD 140-200 (Phase 4.2) had a relatively high percentage for samian of 6.4% (Martin 1995). In this case, this may be a function of the small size of the sample (1.6kg) but the sample appears to be associated with a building of some importance (Isserlin 1995).

Some comparative data for rural sites in the region are available. At Buildings Farm, just west of Great Dunmow, a group of c. 12.2kg of pottery dating to c. AD 35-100 included samian, though this constituted just 0.6% of the group (Wallace 1997). Amongst a similarly sized group from Old House, Church Langley, dating to c. AD 120-165 samian accounted for 1.2% of the pottery (Martin 2000b). At both, the proportion formed by samian is lower than in the contemporary phases at Elms Farm, albeit marginally so. In sum, comparison with data from other sites demonstrates that the proportions of samian within phased groups at Elms Farm are in broad accordance with proportions at other smaller nucleated centres. Heybridge has somewhat higher relative frequencies of samian than occur at a range of rural sites, though the difference from proportions at rural sites is essentially marginal.

Data by EVE are available for various sites in Essex; this information, with regard to samian, is reproduced in Table 17. It can be seen that samian is generally rare in the early post-conquest period, but clearly becomes more frequent in the 2nd century. In other words its frequency over time at Heybridge reflects a broader pattern in the county. That the data relating to Ceramic Phase 7 at Elms Farm (later Antonine to early 3rd century) are anomalous is again emphasised when set against the samples from Rivenhall and Chelmsford. It is noteworthy that both the early to mid-3rd century groups listed have similar proportions of samian to mid- to late 2nd-century groups, emphasising that samian was apparently still being used to a significant degree into the 3rd century, a period when imports of new samian were more limited than previously.

Samian within individual groups at Elms Farm

As shown above, the data for the individual stratified groups were summed by phase to generate robust figures for comparative analysis. Some attention to the frequency of samian within the individual KPGs and wider pool groups is instructive. Again this is via EVE and weight. This analysis demonstrates the previously identified tendency for EVE data to produce indecisive results when sample size is small or moderate (e.g. Orton 1982). This occurs because the presence of a single rim sherd of a certain type within a group with a modest overall EVE size can have a significant impact on the percentage figures. Hence weight data are a more reliable indicator in this instance. The data by EVE are presented here as a matter of record and for methodological interest.

Table 18 lists the proportion of samian within each KPG and wider pool group, in which samian rims occur, by EVE (pool groups with intrusive material are excluded). Two groups including likely structured deposits are included here, however. There are six groups of Neronian-early Flavian date (Ceramic Phase 4). Excluding Pit 20008 which represents a likely structured deposit, the proportions of these groups formed by samian range from 1.7% to 15.9% and none is particularly near the mean for this phase of 3.4%. The groups with the lowest proportions of samian are the markedly larger groups, and the three comparatively high proportions are associated with quite small groups. Seven groups of Ceramic Phase 5 are represented, covering the period c. AD 80-125. As with Ceramic Phase 4, the data show a wide range of proportions of samian within these groups. Samian forms a conspicuously high 23% of the sample from Ditch 6646, though this is clearly an effect of a small EVE sample size (compare the percentage when weight is the measure: Table 19). Conversely samian forms only 0.5% of the large group from Pit 13771, but this reflects an absolute paucity of samian from this feature.

There are ten Hadrianic to mid-Antonine groups (Ceramic Phase 6) and these show marginally less variation than with the preceding samples. Eight of the ten groups have samian proportions of c. 6% or more. The highest proportion is 33% among the smallish sample from Pit 4211. There are only two groups available for the period c. AD 170-210, neither of which is likely to be representative of normal levels of samian consumption at this time. Indeed, the group from Well 6280, where samian forms c. 22% of the pottery by EVE, is evidently a structured deposit with several complete or near-complete vessels. The three groups of early to mid-3rd century date (Ceramic Phase 8) are all near to the mean for that phase.

Table 19 lists the equivalent data to Table 18 when weight is the measure. Of the mid-1st to early 2nd century groups (Ceramic Phases 4 and 5) most groups have proportions of samian similar to, or a little below, the mean for the phase, verifying the validity of this mean (cf. Table 19) as an indicator. (This contrasts with the EVE results of Table 18, which are affected by skewing owing to the small EVE totals of some groups). The highest proportions occur in Ditch 17086 and Pit 20174 but amount only, in both cases, to just over 3% of the group. The overall picture is consistent and emphatic; very small proportions characterise all fifteen groups. Turning to the ten Hadrianic to mid-Antonine groups (Ceramic Phase 6), more variation occurs. Pit 10044 has a very low proportion of samian (a single sherd). The highest proportion is 6.7% among the smallish sample from Pit 4211. Seven of the ten groups have samian proportions under 3%. The highest proportions are 3.9% within the sample from Pit 7118, which yielded a range of samian items, and 4.4% and 6.7% from comparatively small-sized pit groups in Area K.

Four groups are available for the period c. AD 170-210. These include the sample from Well 6280, wherein samian forms 15% of an evidently structured deposit. The remaining three groups, as with the four groups available for Ceramic Phase 8, c. AD 210-260, all have low proportions of samian, though with some degree of variance from their respective means.

Spatial distribution of the samian

In her report, Brenda Dickinson documents the proportion of all the samian recovered in each area of the site. This information is of intrinsic interest, but needs to be calibrated in order to overcome biases arising from such variables as differential levels of archaeological input. Only then can any actual patterns in the distribution of the remains be established. In order to achieve such calibration, one can examine the proportions of samian within groups from different areas of the site.

The thirty-five KPGs and wider pool groups listed in Table 19 come from various locations. In principle, plotting the spatial incidence of these groups by site area (and chronological phase) might be thought useful, carrying the prospect of identifying areas of above and below average frequency of samian, and other patterning, that might be significant. Other recent studies have plotted the proportions of samian occurring across sites in order to isolate trends with regard to different functional and status areas, as at London (Milne and Wardle 1993) and Lincoln (Darling 1998). In the event, little patterning is discernible from the Heybridge data. A larger number of samples, together with a more even spread of samples spatially and chronologically, would be desirable in order to establish the presence of any spatial trends. Only three areas have more than three samples available, namely Areas H, I and K, with the latter being the source of nine samples. The majority of the groups are located in the vicinity of the temple/core area at the western end of the examined area.

A few observations may be significant. Area K, with the largest number of sample groups includes both the group with the highest proportion of samian (excluding the structured deposit, well 6280 of Area H) and the group with the lowest proportion of samian. This may simply be a function of the fact that this area has the largest number of samples. Virtually all areas with two or more sample groups include groups with both comparatively high and comparatively low percentages of samian (i.e. Areas D, E, G, H, I, K and L). An exception is Area Q, where two samples available have relatively high proportions of samian, both at 3.2% by weight.

Finally, it is worth noting that sherds from a samian inkwell (Ritterling 13) were recovered (from fill 4212, Pit 4211). This form is functionally specific and is unlikely to have been distributed in the same manner as other samian forms. Samian inkwells are very rare site finds, and are almost invariably associated with Roman military sites and major civil centres where they occur at or near fora and other sites of business and record keeping (Willis 2005). In other words, their distribution is highly structured. Unlike writing tablets and styli, which were principally to do with the recording of information over the short term in a relatively cheap format, the presence of an inkwell implies the documentation of information for keeping over the long term (or long transit) and an investment of some wealth, owing to the relative expense of ink and the receiving medium, either vellum or papyrus. The vessel is South Gaulish and was recovered from a Ceramic Phase 6 context, so it was either residual or (not surprisingly) had had a long life. The association of this find with other forms of evidence is consistent with the possibility that better quality building/s existed here, by the western margin of the excavated area (Mark Atkinson, pers. comm.).

Analysis of the proportions of decorated samian ware

Table 20 records the proportions of decorated samian by source, based on the number of vessels represented. These data have a chronological dimension in so far as the sources of samian at Elms Farm are sequential. The table uses data for the whole recovered assemblage as identified by Brenda Dickinson and provides a potentially useful guide to the consumption of decorated ware (mainly bowls, but including some beakers and other closed forms), as opposed to plain samian forms (largely cups, platters and dishes). The present data (in this form) convey an instructive picture for the composition of the assemblage, which is helpful at a number of levels. Broad trends are apparent. Further study of the samian might look at the whole sample using EVE as the measure of frequency and at these data by phase; this would doubtless present significant information (but would take time to generate). The present data suffice as a general index.

Assessing levels of decorated forms among samian assemblages is a useful undertaking since it has been demonstrated that systematic differences in proportions occur at different types of site (Willis 1997a; 1998). Decorated samian vessels, such as bowl forms f29, f30 and f37, are often thought to have been comparatively valued items because of their unusual character, and also more expensive to produce and transport than plain forms. It may be that these vessels were prized not because they were decorated, but because they were bowls and potentially drinking vessels (Willis 1997b). If decorated vessels were valued more than other vessels, the incidence of such items may be an index of site status and identity. Alongside Table 20 relating to Elms Farm, it is possible to present, via Table 21, some comparative data for other sites in Britain. Table 21 is a 'short-hand' summary of more detailed information published elsewhere by individual site and stratified phase group (Willis 1998, table 3). Table 21, shows that there were clear differences in the character of samian consumption at different types of site. Among the fifteen samples from major civil centres in Britain, for instance, only one sample has a percentage for decorated ware below 20%. However, among sites lower down the settlement hierarchy, including small towns, roadside settlements, smaller nucleated centres and rural sites, proportions of decorated ware are very often below 20%. This is the case at Neatham, Hampshire, where decorated forms account for only 14.3% of the assemblage covering the period c. AD 150-235 (Millett and Graham 1986), and at Towcester, Alchester Road, (Phase 2, c. AD 170-270) where the equivalent figure is 13.2% (Brown and Woodfield 1983). In Building AJ at the 'small town' at Kenchester, Herefordshire, decorated ware formed 20.7% of the samian from Period 2c (c. AD 140/150-180/200), though the building was of some pretension (Wilmott and Rahtz 1985). These systematic variations between different types of settlement raise questions regarding samian form and function, as well as variations in the perception and use of samian vessels at different types of sites.

Comparison between Tables 21 and 20 suggests that Heybridge compares closely to the 'norm' for the small town, roadside settlement and smaller nucleated centres. Decorated vessels form just under 20% of the South Gaulish samian during the mid- and later 1st century AD. The proportion with regard to Les Martres samian is higher, probably since the industry seems to have produced disproportionately more decorated bowls than did the other industries, and because there may have been an unusual 'delivery' of Cettus bowls from this source (Dickinson, above). The proportion of decorated vessels falls somewhat within the mid- to later 2nd century material from Lezoux, with the proportion at c. 16%. Significantly, a similar pattern has been noted elsewhere, as at Catterick, North Yorkshire (Jeremy Evans, pers. comm.), and at Lincoln and Verulamium, insula XIV, where Darling (1998) has identified a general decline in the proportion of decorated sherds dating to the 2nd century. The percentage for East Gaulish samian is not dissimilar from that of the Lezoux ware at 17.2%.

Quite strikingly, only five out of a maximum number of 101 vessels in Colchester samian are decorated forms (Table 20). As Brenda Dickinson states, the paucity of other collections of Colchester samian of any size makes assessment of the finds from Elms Farm difficult. It is unclear whether this low proportion is a function of perhaps relatively fewer decorated forms being manufactured by the Colchester industry (cf. Bird 1999, 76), the status of Heybridge, or the preferences of its consumers and their perception/definition of Colchester samian. In other words, was there simply little decorated Colchester samian available, or were the consumers at Heybridge using Colchester plain samian ware but preferring their decorated bowls to be from the technically more accomplished Central Gaulish industry?

On the whole, the picture that emerges from Table 20 is one of broad consistency in the levels of decorated ware through time. Variance occurs only with the less important sources of supply (Colchester and Les Martres) and, in these instances, may relate more to the output of those industries rather than be specific to Heybridge. Significantly, when the numbers of vessels from all sources are aggregated, the percentage formed by decorated vessels is 16.9%, a proportion that is very close to the mean for six samples from other sites of middle rank scale and probable like functions (Table 21; these sites include Baldock (excluding structured deposits), Castor (Water Newton/Durobrivae), Kenchester, Neatham and Towcester). Discounting the Colchester samian results in a similar percentage of 17.2%.

Summary

Overall, a picture emerges of a familiarity with samian among the people living at Heybridge, or whose lives related to Heybridge, but samian was not commonplace. Samian was evidently supplementary to the ceramic repertoire at Heybridge, though seems to have been invested with particular significance by its users. Some vessels may have been in everyday use, others (perhaps the majority of vessels) may have been saved for special days and events. As elsewhere in Roman Britain, there was clearly no social restriction regarding access to samian. Heybridge, however, was almost certainly, vis-à-vis continental samian, at the end of a long chain of exchange, and evidently did not exert a particular 'pull' in acquiring samian. Indeed, it was probably at the tail of 'down the line exchange'. Nonetheless, it is significant to note that the frequency of samian at Chelmsford is hardly higher than at Heybridge, despite the fact that Chelmsford was almost certainly the main market centre for central Essex at this time.

Table 21 indicates high proportions of samian occur at military sites and sites at the apex of the settlement hierarchy; in short 'high' levels of samian consumption are associated with sites closely articulated with the inter-provincial/Imperial economic system, with users familiar with metropolitan Roman material culture, and centres of higher status. The pattern of samian evidence from Elms Farm shows a comparatively low level frequency of consumption, but one that is normal for a smaller nucleated settlement. In so far as samian was an indicator of social status, this seems to have been a particularly urban phenomenon; relatively low levels of samian per se, and of decorated vessels, at small towns, roadside settlements and smaller nucleated centres like Heybridge suggest there was 'less status' - or rather less individuals with wealth and status - at such sites, and/or that status within such milieux was not commonly displayed by means of the ownership of samian.

On the other hand, the selection of samian vessels for inclusion with burials and structured special deposits, where it often occurs with a frequency at variance to its representation in normal site deposits (e.g. Tables 18 and 19, Well 6280), indicates that here, as elsewhere in the Roman province, it was regarded somewhat differently from other ceramics.

In sum, the occupants of Roman Heybridge had an awareness of samian, which was available to users throughout the importing period (c. AD 20-260). Its prominence in the everyday use of ceramics was probably moderate, and, besides, it was not necessarily in common daily use. There is clear evidence that it was regarded differently from other contemporary pottery types. Overall, the patterns of samian consumption at Heybridge, defined through these analyses, accord with trends identified at other smaller nucleated sites of the period.

 PREVIOUS   NEXT   CONTENTS   ISSUE   HOME