Summary of the Samian Ware

Summary of the Samian Ware

There were 400 sherds in this collection, representing a maximum of 321 vessels, 6.88  estimated vessel equivalents (EVES), weight 4369 g. It will be helpful to make some comparison with the samian ware recovered from previous excavations at the neighbouring Faircloughs Homes site (925 sherds from a maximum of 633 vessels, 27.66 EVES; weight 17.23 kg). The Community Dig collection was in reasonably good condition, comprising sherds weighing a good average of 11 g, though this was considerably lower than at the Faircloughs Homes site (19 g). At the Community excavation, there were far fewer near-complete vessels: only four compared with 21 at the latter; two of the four were SG products and none of the four was produced later than c AD 150. There was also a large proportion of indeterminate fragments (29%).

The maximum of 83 moulded bowls constituted 27% of the collection, considerably less than at the Faircloughs Homes site (39%). 65 of the bowls retained at least some fragment of decoration and there was one beaker with decoration of the ‘cut-glass’ variety (Catalogue No 13). There were only four potters’ stamps, 1.2% compared with 4.9% at Faircloughs; however, this figure does not relate to the status of the site, but is likely to relate rather to its chronology, if it is the case that more potters were stamping more pots at certain periods (see Ward forthcoming).

The evidence of the samian ware suggests that occupation in this part of Middlewich began in the later first century and continued at a high level well into the Antonine period. The extreme shortage of late second-century samian on this site is highly significant. All the evidence of the samian ware points to reduction in occupation before c 160. The presence of six EG vessels, and one specifically dated within the earlier third century in particular, may be taken to reflect activity in the area in the third century.  

Table 1. Absolute values of samian wares by maximum vessel numbers, sherd weight and estimated vessel equivalents (EVES)

Table 2

Table 2. Relative quantities of samian by maximum vessel count, estimated vessel equivalent (EVES) and weight

Table 3

Table 3. Faircloughs site. Relative quantities of samian by maximum vessel count, estimated vessel equivalent (EVES) and weight

Origin and date

A maximum of 139 vessels, forming 43% of the collection, originated in South Gaul, 175 vessels (55%) originated in Central Gaul and 7 vessels (2%) were from East Gaul. These figures are summarized on Table 1. According to EVES, the proportions for SG, CG and EG wares were 28%, 71% and less than 1% (1.92, 4.92 and 0.04 EVES respectively). Sorted by weight, the proportions of 27%, 71% and 2% (1181 g, 3110 g and 78 g) are closer to those of EVES than of maximum vessel numbers. This discrepancy reflects at least partially the large proportion of indeterminate fragments in the collection; the combined evidence of EVES and weights may well indicate that numerous South Gaulish sherds belonged to the same vessel or vessels.

Unless stated, for reasons given elsewhere, quantification below is based on the maximum numbers of vessels represented on the site rather than EVES.

 Table 4

Table 4. Middlewich Community Dig: histogram of all samian vessels by half decade (maximum nos)

Table 5
 Table 5.  Faircloughs Homes site:  histogram of all samian vessels by half decade (maximum nos)

Evidence for 1st-century occupation
There were at least one or two vessels (eg Catalogue No 24) that were produced in the pre-Flavian era and were probably Neronian products of the period c AD 55-70; the one example of a pre-Flavian cup form (Ritterling 8) was most probably a Neronian product.  A further four vessels (two of them fragments of moulded bowl form 29) were dated only loosely within the range c AD 65-85. Thus the possible pre-Flavian material must have ranged between 1 and 4% of the SG ware. Many more SG vessels, both plain and decorated, were not datable closely within the Flavian period. In contrast, at the Fairclough Homes sites much of the SG ware had been produced in the early-Flavian period and specifically before c AD 85/90 (compare the histograms summarizing all samian ware from both sites (Table 4-5). The samian was taken at Faircloughs to suggest a slightly later date for foundation than was indicated at the Millennium site in the fort at Carlisle. At Carlisle, the greater proportion of pre-Flavian material (c 4% of the SG wares) and many Neronian-Flavian products were considered consistent with a foundation date in the early 70s (Ward, unpublished report). Amongst the plain samian from this Buckley’s Field site, there were none of the dishes of form 18 or 18R displaying the pronounced external offset at the junction of wall and base that is recorded commonly on these forms in Neronian deposits (eg the fortress at Usk; Tyers 1993, 133). This feature on such dishes is thought to be diagnostic of Neronian date or of the very early 70s at latest; it was noted at the Faircloughs site, though not here.

At the Faircloughs site, too, the ratio of the earlier bowl form 29 to that of the typically Flavian form 37 was 1:10. At the Community site that ratio was 1:15. Recent research has suggested that the SG form was being produced until at least c 90. Felicity Wild (2002, 268) has argued for a date in the 70s or early 80s for parity of the two forms; however, Willis (2005, has noted that, according to his database, parity between the two forms was reached specifically between AD 87 and AD 89 and, if correct, this would imply that the formation of the Middlewich material occurred later than AD 90. At Walton-le-dale, no precisely comparable data are available, but there was only one form 29 amongst 25.6 kg of samian ware (Wild 2002, 271) and at that site, foundation after c AD 90 was suggested. At any rate, the material under discussion appears in general to have been produced later than that on the Faircloughs site. Certainly, a higher proportion of form 29 would be expected if foundation of the Middlewich site occurred in the earlier-Flavian period and the main activity on this particular site appears to have begun in the later first century. .

The second century and later

The pronounced peak on Table 4 in the half-decade c120-125 owes much to the unusually large quantities of products from Les Martres-de-Veyre that are datable in the range c 100-125, combined with the bulk of Lezoux products that were dated in the Hadrianic and Hadrianic-Antonine periods. Bearing in mind that some of the Les Martres products dated c AD 100-125 will have been produced in the Hadrianic period, the evidence of the samian ware on this site nevertheless suggests that there was considerable activity in the vicinity during the Trajanic period.

Although, for whatever reason, there was a surprisingly small proportion of moulded bowls (17%) amongst the Les Martres products, the Les Martres ware as a whole constituted a very large proportion of the assemblage. There were 29 vessels, both plain and decorated, from the workshops that were operating there in the Trajanic-Hadrianic periods, including at least one of slightly later origin (probably contemporary with the potter Cettus who was active there in the period c AD 135-160). Thus, as much as 9% of the samian assemblage and 17% of the CG wares was produced at Les Martres, compared with only 4% of the samian ware as a whole (9% of the CG ware) at the Faircloughs site. Brenda Dickinson (2000, 204) commented that 8-10% is a fairly normal proportion of collections from on British sites. There appears therefore to have been considerable activity in Buckley’s Field in the Trajanic to early Hadrianic period.

Period Fabric Potter Vessels(max nos)
Trajanic Les Martres Igocatus (X-4) 1
Trajanic to early-Hadrianic   X-13 2
Hadrianic Lezoux Arcanus 2
Hadrianic-early Antonine   Butrio 1
    Large S Potter 1
    Attianus 1
  Lezoux Cinnamus early group 2
    Criciro 1
early- to mid-Antonine: c 150-170 Lezoux Cinnamus standard style 1
c 150-180   Cinnamus or Secundus 1
    Pugnus-Secundus 1
mid- to late-Antonine   Priscus-Clemens 1
    Attributable CG bowls: total 15    
Earlier 3rd c. EG Lupus iv (Rheinzabern) 1

    Table 6 CG and EG moulded bowls with decoration that was attributable to specific potters or groups (max nos)

There was a maximum of 83 moulded bowls, of which 66 had some part of the decoration surviving: 35 of the 41 CG bowls bore decoration, though most were mere fragments. Those CG bowls that could be attributed with some certainly to specific styles are listed on Table 6 above. Amongst the CG bowls from Lezoux, over a quarter of those attributable to specific potters’ styles were Hadrianic or Hadrianic to early-Antonine products. As at the Faircloughs site, there was one bowl in the style of Criciro and here, two were in the style of the early Cinnamus group. Following the dating of the Leeds index of potters’stamps, a date c AD 135-160 is given for this style (though Rogers 1999, 98 dated the Cerialis-Cinnamus partnership as early as c 135-145 and the ‘standard’ style as c 140-160+). One or two bowls represented the standard style of Cinnamus, dated here c 150-170 and 150-180, the latter representing the style of Cinnamus or Secundus; there was also one bowl in the Pugnus-Secundus style (again c 150-180).

One bowl, No 27, in the Priscus-Clemens style was dated firmly within the mid- to late-Antonine period (approximately 160-200). As at the Faircloughs site, the virtual absence of all other bowls of that period from this site would be compatible with a period of abandonment or inactivity in this area of Middlewich not long after the mid-second century; the same absence of later-Antonine bowls was noted also at Holt (Ward, 1998, 139). It is quite possible that this bowl, as well as the few later-Antonine plain vessels that were also present on this Middlewich site could have been in use still in the third century.   

Form SG CG EG Total
Ritterling 8 1     1
Curle 11 2 2   4
15/17 2     2
15/17 or 18 2     2
18 4     4
18 or 18/31 2     2
18 or 18R 6 1   7
18R 13     13
18R or 18/31R 4     4
18/31   12   12
18/31 or 31   4   4
18/31 or 18/31R   2   2
18/31R   11   11
18/31R or 31R   7   7
31   1   1
31R   3   3
31R group     2 2
27 14 15   29
33 4 8   12
29 4     4
29 or 37 1     1
30 2 2   4
30 or 37 1     1
37 34 39 1 74
35 4     4
36 3 2   5
38   10   10
72   1   1
79   2   2
Ludowici Tx   1   1
Indeterminate 36 52 4 92
TOTAL 139 175 7 321

Table 7. Total of vessel forms by fabric (maximum nos of vessels)

Amongst the plainwares, as at the Faircloughs site, the dish forms 18/31 and 18/31R that were Hadrianic-early Antonine products outnumbered greatly the Antonine forms 31 and 31R. Form 31R and its East Gaulish equivalents were the deep dish variety that developed after c 160 and are usually found in profusion on sites under steady occupation in the later second century. At the Community excavation, these developed forms constituted less than 2% of the total. Also, of the ten examples of the flanged bowl form 38, whose popularity reached a peak in the second half of the 2nd century, half of these may have been Hadrianic products and none of the remainder could be said with any certainty to have been produced after c 150.

Form 27 was the most popular of the cup forms made in South Gaul in the first century, but it lost popularity in favour of form 33 in the second century and went out of production by c AD 160. At both the King St excavations, there were more than twice as many form 27s as 33s. At the Faircloughs site, amongst the Central Gaulish wares, forms 27 and 33 were more or less equal in number, but in contrast Community Dig finds included 15 CG form 27s, but only 8 form 33s. It is conceivable that this reflects the small size of the sample, or it may signify some quirk in the supply to this site; however, it could suggest that activity on this site was reduced rather earlier than at the Faircloughs site. On the other hand, the fall-off at the latter site c 160 seems even more marked according to Table 5 than does the histogram based on the Community Dig finds (Table 4).

At any rate, at the Community Dig there were very few vessels of any form that were manufactured in the later-2nd century: there was one CG beaker with ‘cut-glass’ decoration, only two sherds from dishes of the late-second century form 79 (produced after c 160) and none of the mortarium forms that were produced after c 170 at the earliest.  The presence of samian mortaria and particularly the popular lion-headed form 45 is normally a good indicator of late-Antonine activity. The virtual absence of late second-century samian, stamped, decorated or plain, is highly significant and all the evidence of the samian ware appears to point to reduction in activity on the site before c 160.

Early 2nd-century evidence from other north-western sites, and specifically those ‘industrial’ sites with varying degrees of military involvement may be considered, too (see Ward, forthcoming). Samian collections from Holt, Prestatyn, Holditch, Wigan and a very small group from Ochre Brook, Tarbock were used to produce Table 8.

Table 8
Table 8. Other industrial sites with varying degrees of military involvement: all samian vessels, by half decade (maximum nos)

At the legionary works depôt at Holt, there was an abundance of Trajanic-Hadrianic samian ware, but the stamps indicated an ensuing reduction in activity there also (Ward 1998, 141, fig 8). At Holditch, none of the samian ware from recent excavation needs to have been produced after the middle of the second century (contra the editorial amendment in Ward 2007, 127). The chief periods of activity also ran out around the middle of the 2nd century at two more sites on the line of King St, not included on Table 8: the probably military depôt at Walton-le-dale and industrial site at Wilderspool (see Ward 1998a, 62). The evidence from Prestatyn and recent excavations at Wigan also appear to support this dating for sites where military presence is strongly suspected (Ward 1989 and Ward, unpublished samian report for the Grand Arcade site, Wigan). That there were changes afoot in this period at militarily linked establishments seems confirmed at the auxiliary tilery at Quernmore. The present sample from the Community Dig at Middlewich supports the suggestion that the area of the Faircloughs saw a reduction or change in activity in or very close to the early-Antonine period, with few vessels having been manufactured after c AD 160; much the same pattern is visible at other industrial sites with military involvement (cf Tables 4, 5 and 8). Certainly, the evidence of the samian ware can be taken to indicate little activity on the Community site at Buckley’s Field in the late second century

The general scarcity of EG wares on the western side of Britain is well known. The proportion of EG wares in the Community Dig collection (2%) is very low for a western site, even, when compared with, for example, the Faircloughs material (4%) or that from sites with a high level of third-century occupation such as the ‘mansio’ in Castle St, Chester (6%); see Bulmer 1980, 21; 1993, 16  (cf the much higher proportion from the north-eastern site at Piercebridge: 17%; Ward forthcoming, 2008). In this collection, there were no products of the Argonne factories of the Hadrianic- Antonine period as have been recorded at such sites as Lancaster, Ribchester, Chester and Carlisle. However, as at Holt (Ward 1998 and 1998a), the presence of not only the few late CG vessels (eg Catalogue No 27), but also EG products from Rheinzabern and/or Trier in this small sample could support renewed activity in the third century: in all, six CG and six EG vessels were produced after c AD 160. Of the later 2nd- or 3rd-century products from Rheinzabern and Trier at Middlewich, one of the Rheinzabern products was a moulded bowl in the style of Lupus iv, who was working in the earlier third century, probably within the range c AD 200-240; see No 4. This was the latest identifiable samian ware on this site (found unstratified in Trench 1). As noted above, this bowl along with a few other late CG and EG sherds could reflect third-century activity in the area. 
Condition: burning, wear, repair and re-use

Samian ware can provide much information about a site and its occupation. The significance of this information has been acknowledged for many years (Bulmer 1980a, 89; Marsh 1981, 227f; Ward 1989; Willis 1998, 121). Re-working or reuse of samian ware has been noted at many sites (see Ward 1993, 19ff). In this collection, 11% of the total showed evidence of wear, repair or secondary use; 1.6% of these had been re-worked or re-used and 0.9% had been repaired. Most of these repaired, re-used or re-worked vessels had been produced in the second-century and may well have seen considerable use. Samian ware must have been a relatively valuable commodity and, where possible, its life will have been extended in use, whether by cutting down and smoothing off, re-working or repairing of accidentally broken pieces

Marks of ownership also point to the value of samian vessels to their owners (see Tomlin 2002, 504). At least three vessels displayed graffiti or erasure of graffiti (implying second ownership; see Nos 12, 29). One in Trench 12 Phase 8 (No 12) was on a bowl  manufactured in the range c 150-180, the other two, produced in the earlier second century, were a (stamped) dish of Hadrianic origin (No 28) and a cup produced in the Hadrianic-early Antonine period (No 29), both found in Trench 12, Phase 9m (1217). The proportion formed by these graffiti (0.93%), although in a small sample, may be compared with 0.16% of the total at the Faircloughs site: this comprised a single graffito (on a dish stamped in the period c 150-175).

In Trench 12 Periods 8 and 12, there was a tiny counter, a large disc, a footring re-used upside down, a cut-down cup and at least one vessel that had seen, at the least, hard use (No 20). There were no examples of spindle-whorls, an item that is thought typical of the late-Roman period and particularly the fourth century (Cool 2000, 53).
There were three repaired vessels in Trenches 9 (Period 7) and 12 (again in Period 8). These constituted 0.9% of the total, compared with as much as 3% of the Faircloughs site assemblage. These figures may be compared against 0.3% at the Millennium site in Carlisle fort, 0.7% at Prestatyn, 1.4% at Piercebridge, 1.6% at Mitchells Brewery in the Lancaster vicus, and 2.8% at Worcester Magistrates Court (see Ward 1989, 154; 1993, 21 and 2007, forthcoming , and Ward unpublished reports). Willis (2005, 11.5 Table 73) found that the average for nine of his listed military sites was 2%. Certainly, repairwork on samian ware has been noted at other sites known to have had metal-working or industrial activities, with military or possibly military involvement (see Holt, Ward 1998a, 52 and, on Heronbridge, 73; Prestatyn, Ward 1989, 154).  At Piercebridge, there was the workshop of a metalworker who may have taken on the repair of broken samian vessels, in the northern vicus (Ward 1993, 19f and 2008). In this small sample from the Community Dig, two repairs were of the rivet variety and one used cleats. However, it is difficult to judge the success of repairwork, particularly when the vessel had finally broken through the repair and lacked its rivets or cleats.

In addition to the re-worked or repaired sherds recorded above, Trench 12 (1267) contained a small fragment from a Hadrianic-early Antonine CG product that displayed a sharply worn corner and scoured and scratched edges. This could possibly have been used as a scraper, but, as in the case of possible examples at the Faircloughs site, its secondary use was only a tentative suggestion (cf Ward 2008).

7% of the material was burnt. This was only a little lower than at the Faircloughs site (10%), where, unlike in the present collection, several pieces appeared to have suffered unusually from smoke (rather than burning or sooting).


The evidence of moulded bowls on the site may be considered. The products of Les Martres-de-Veyre always include a large proportion of these decorated bowls (31% at Carlisle Millennium site; Ward, unpublished report); see Darling 1998, 176. At the Faircloughs site, an exceptionally large proportion of the products of Les Martres comprised moulded bowls (48%) and on that basis it was proposed that moulded bowls were highly favoured in the vicinity of this Middlewich site in the early 2nd century (Ward 2008). However, at the Community Dig site, for whatever reason, the proportion was surprisingly small at only 17%.   

In contrast, we have seen above that the overall total of 83 moulded bowls constituted 26% of the whole collection compared with 37% at the Faircloughs site (see Fig 3). If we discount the many fragments of indeterminate form from the total, then the moulded bowls from the Community Dig could have comprised as much as 41% by EVES, though (for whatever reason) considerably their proportion was considerably smaller according to the maximum numbers of vessels: 35% (see Tables 1-2).

Thus, the figure of 41% comprising moulded bowls by EVES is very close to that at the Faircloughs site (43%). Willis (2005, Table 42), relying on the ‘number of vessels represented,’ noted that the average percentage of decorated samian bowls on thirty military sites was 30% and in extra-mural occupation at military sites 35%. In contrast, the proportion at exclusively civil sites was much lower, at 17-26%. It is unclear whether Willis included all indeterminate fragments in his totals, but presuming that he did, even so the Middlewich proportions would be high for an exclusively civil site. The assemblage may be taken to reflect the military presence in the area, whether or not there was specifically military involvement at Buckley’s Field. The moulded bowl proportion according to maximum numbers represented is clearly much higher than that of 25% noted in the Carlisle Millennium assemblage (Ward, unpublished samian report). So too, although it is difficult to compare collections quantified differently, moulded bowls appear more frequent at Middlewich than at Ribchester (30%) or indeed than the average as quoted there (10%; see Dickinson 2000, 204). Ruth Leary concludes concerning the pottery on this site that the level of decorated samian does indeed indicate that the settlement had a military-type assemblage, but that the characteristics of the group in terms of other luxury goods and table ware: kitchen ware ratios suggest the area was of fairly low status. It may be noted that Willis (2005, Table 45) has compiled data for types of samian vessel represented at sites of differing status and it is striking that the proportions of bowls, cups and dishes on the Middlewich site are closer to those recorded from extra-mural occupation sites at Roman military installations than they are to those from the military installations themselves (op cit, 8.2.3 Chart 14, cf 8.2.2 Chart 13).

Table 9
Table 9 The relative frequency of samian vessel types/functional categories, discounting indeterminate sherds (maximum vessel nos)

Table 10

Table 10. The relative frequency of samian vessel types/functional categories, discounting indeterminate sherds (by EVES)

Table 11

Table 11. Faircloughs site: relative proportion of vessel types (maximum vessel nos)

Although this was a relatively small sample, it may be noted that here there were none of the inkwells that were recorded at the Faircloughs site. Writing of graffiti on samian ware itself is taken to be evidence of a high status site (Evans 1987, 202). The proportion of 0.93% of the samian total formed by graffiti on this Community Dig site is relatively high. It may be compared with the figure of 0.16% at Faircloughs, 0.5% at Piercebridge and also at the large Mitchells Brewery site in the Lancaster vicus, or with 1.25% at the Millennium site in Carlisle fort (Ward, reports all unpublished or forthcoming). It is difficult to compare other, published data, since the proportions published are often based on all graffiti on all manner of ceramics. However, at Ribchester (Dickinson 2000), it appears that 26 graffiti comprised 0.86% of the samian ware. At the Faircloughs site, the present writer proposed that the proportion of 0.16% might reflect a low level of literacy amongst the average samian-owner in the vicinity and that this in turn could relate to the special, industrial function of the site. Clearly, there was greater evidence of graffiti in the present, smaller sample. There were still no obvious examples of the ubiquitous practices (not requiring literacy) of scoring bases with an X (Tomlin 2002, 504) or the widespread habit of nicking the standing surface of footrings.

As seen above, the samian ware on this site was clearly valued enough to warrant repair in order to extend life in use even though such vessels were probably abundantly available at the time of their repair (none was necessarily produced later than the early-Antonine period). Willis found that the proportion of repaired decorated samian vessels from rural sites is markedly lower than at military sites and smaller civil centres (Willis 2005, Table 75), though he admitted that rural sites had relatively low proportions of decorated ware anyway. Certainly Middlewich had high proportions of moulded bowls and this is no doubt reflected in the proportions of moulded bowls that were repaired. It would, however, not be surprising if the decoration made the vessel more highly prized and valued enough to make its repair more worthwhile. It may be that a craftsman’s services became necessary in order to extend the life of pots at a time when supplies perhaps ran low –  during mass-movement of troops to Scotland in the early-Antonine period, perhaps.



The writer wishes to thank Ruth Leary for her help and for sharing her expertise throughout this project.