Southwell Archaeology’s Ceramic Research Project

A report on a project supported by the Geoffrey Bond and The Thoroton Society Research Award, our sincere thanks to both for the funding without which these findings would not have been possible. Our thanks also to Imogen Wood (https://www.potsimogenwood.com/ for the detailed ceramic petrology work.

Introduction

The Research Award was used to find answers to two research questions: whether there were Romano-British ceramic building materials (CBM) fired locally, and further, was there a distinct medieval pottery workshop in Southwell, Nottinghamshire, dating from around the late 11th to the mid-14th centuries?

Southwell is the site of what some archaeologists believe to be the largest Romano-British ‘villa’ in the East Midlands two portions of which are now a scheduled monument. A large amount of ceramic building material (CBM) has been recovered from this site suggesting a substantial and opulent building. From our excavation work we have also recovered a considerable amount of medieval pottery material, particularly spanning the early 12th to early 14th centuries. East Midlands pottery expert Jane Young has argued that there was a local Southwell type series produced during this period. This proposition is based on a combination of analysing fabric characteristics under a x20 microscope along with forms, and glazing, and also upon the probable similarity to the Romano-British CBM, which given its density, was likely to have been made locally. Working within the East Midlands Historic Environment Research Framework some key questions for Southwell Community Archaeology group have been whether the Romano-British CBM and the Medieval pottery were manufactured from a local clay source.

On Jane Young’s recommendation thin section petrographic analysis of materials was undertaken by Imogen Wood a specialist who understood not just the geology, but the processes clays go through when fired.

The findings below are summarised from Imogen’s detailed technical report. Published pottery ware types are given shortened codes for database purposes and we have followed an established type series for the Nottinghamshire area (see samples below).

Thin-section methodology and samples

Thin-section petrology involves expertly creating thin sections of the pottery sample in order to examine them under a polarising microscope. In this case, using a polarising petrographic microscope, with a range of X2-X40 magnification under plane- and crossed-polarised light the samples were grouped based upon the nature of their inclusions, clay groundmass and voids. Compositional, textural and shape criteria were used to detect the presence of specific practices, including the addition of temper and clay mixing.

Twelve samples were selected:

The aims of the analysis were to:

Results

Six of the ceramic samples are derivative of a source local to Southwell. Five of the ceramic samples are not manufactured from local clays.

The Southwell ceramic fabrics fall into two main groups: the Southwell Micrite (SWM) characterised by the presence of dolomite crystals or rock fragments and the well-rounded buff micrite (carbonate mudstone); and the Southwell Sandy Micaceous fabric (SWSM) characterised by micaceous siltstone and sandstone.

The Gunter mudstone member outcrops to the west of Southwell and is the likely source of the clays in the local fabrics. A siltstone formation lies in the upper reaches of the local River Greet and alluvial sands and silts overlie the geology of the Greet Valley, which helps to explain the variations in the size of quartz sands within the Roman building materials identified by Young. Given its constituents, the clay sample itself had most likely lain in a waterlogged marsh or pond environment. The two tegula samples are derived from the local clay with the addition of micrite (calcareous mudstone) macroscopically identifiable as light buff inclusions (fabric SWM). One of these samples was high fired to over 800 degrees centigrade. The medieval tile also demonstrates a local source with micrite and micaceous sandstone inclusions (fabric SWSM). The box flue tile was visibly of a shell-tempered fabric and unsurprisingly indicated a source outside the area. Young suggested Northamptonshire or Bedfordshire as possible production sites. However, the fabric in this sample ‘is more comparable to fossiliferous shell fabrics from Triton knoll and others in East Lindsey’.

The pottery samples provide some interesting results. The Southwell Area Medieval Green Glaze type (SOAMG) does indeed derive from a local source (fabric SWSM). This sherd was wheel made and fired at around 800 degrees centigrade. The dating generally given is circa early 13th to late 14th century.

The two sherds initially thought to derive from the Beverley area of Yorkshire (BEVO1T, BEVO 2T) are in fact of local Southwell origin. They lack the clay matrix and chert inclusions typical of the Beverley production area but contain the distinctive well-rounded micrite pellets and dolomite of the Southwell locality. The fabric in these two sherds was fine, a result of sifting out inclusions above 0.5mm (fabric SWF). Both vessels are wheel made and fired over 800 degrees centigrade; our initial thinking was that BEVO IT was an earlier fabric (early 12th to early 13th century) than type 2. Given the slight differences in the coarseness of inclusions seen at x20 this is still likely to be the case.

The results from three sherds thought to be of Southwell origin (SDOXMG, SACSPL, SNSPT) demonstrate derivation from a sandy igneous geology characterised by the presence of metamorphic minerals and rounded quartz garnet rock fragments. Some detailed analysis suggests two fabric types — SIGI and SIG2. Recent work suggests multiple clay sources incorporating glacial till deposits probably lying east of Lincoln in the East Lindsey district. However, the original fabric codes were assigned on the basis that the ware types are different in form, glaze and other details to those of east Lincolnshire.

Sherds SNSPT (fabric SIGI) and SACSPL (fabric SIG2) are coil made and low fired around 800 degrees centigrade. The manufacturing results are probably in line with the earlier medieval dating for this pottery (circa early 12th to early 13th century) as indicated by immature pocked glazing resulting from the use of lead powder to produce a glazed decoration (a splashed glaze). Sherd SDOXMG (fabric SIGI) indicates coil making with wheel finishing and a higher firing temperature, the dating usually given as circa early 13th to early 14th century and indicates a suspension glazing technique.

The NSP sherd (circa late 11th—mid 13th century) is also a splashed glaze type and not unexpectedly falls outside of the Southwell area for manufacture, with a fabric of sandy igneous dolomite extraction. Alan Vince’s analysis of NSP suggests a clay source different to that used in earlier and later Nottingham production but its origin is uncertain.

Conclusions and further questions

Although some caution needs to be exercised when attributing provenance to ceramics using geological analysis the thin section petrography on these samples is strong evidence that there was a local ceramic production of both Romano-British CBM and Medieval pottery. Some samples believed to be of Southwell origin may well be imports from outside the locality. However, these wares retain some distinct features akin to local production in form and glazing that makes them different from east Lincolnshire types. Following discussion with Jane Young we propose to retain the original codes but add a suffix, eg SDOXMG-IW-SIG, indicating analysis done by Imogen Wood.

Two sherds identified as Beverley ware fabrics are products of refined clays from the Southwell area, but the same convention will be applied to these so that we can add BEVO1- or BEVO2-IW-SWF. This will enable us to add to the knowledge base of ceramic specialists while not foreclosing further discoveries and changes as more pottery groups are discovered.

Although this project does not overly affect our dating of the ceramic material it does demonstrate that we know so little about where these kiln sites were located. In addition, it may well be that suitable pot making clay sources were quickly worked out and are no longer there for discovery. Across the East Midlands there are a huge number of fabric variations, and while Nottingham has a fairly well-defined type series for Medieval pottery, ceramic discoveries in Southwell add another layer of local complexity.

A further line of research is to ascertain whether known Medieval kiln sites in the Newark area are possible origins for local Southwell pottery. There are similarities in geology between the Gunthorpe Member series around Southwell and the Edwalton Member on the opposite side of the Trent underlying Newark. More samples from along the Trent valley would be needed to affirm or rule out possible production in the Newark-on-Trent area rather than, or as well as, Southwell.

Another research question is the origin of the box flue tiles, of which there are other similar shell-tempered fragments from the ‘villa’ site. Is there a possible specialist production centre for these tiles and if so, can it be located within a literature search? It raises some fascinating questions about how this Romano-British site was located within trading networks. 

Dr Stephen Rogers

< Previous