Mapping Roman pottery in Britain

P A Tyers

The maps in Roman Pottery in Britain (1996)

Part II of Roman Pottery in Britain (RPB: Tyers 1996b) contains an 'Atlas' of almost 100 types of pottery. The Atlas pages, and their maps, cover the main classes of Roman pottery made or found in Britain, but were chosen principally to illustrate various points in the chronological and economic discussions in Part I of the same volume. The maps were subsequently recreated using essentially the same data set, most recently in 2004, and made available from this website.

The principal sources available in 1996 for the distribution maps were:

• Detailed studies of major Romano-British pottery industries such as those by Young (1977) and Fulford (1975) on the Oxfordshire and New Forest industries. The data in these works were collected in the 1970s.
• Regional surveys such as that by Pollard (1988) on Kent, or major studies of urban assemblages that incorporate significant data on their regional background (e.g. Holbrook/Bidwell 1991, on Exeter and the South-West).
• The lists of pottery types in the ‘Roman Pottery Bibliography’, published in the Journal of Roman Pottery Studies vols. 1-5 (1986-92).
• The distributions of mortarium stamps reported by K. F. Hartley in numerous published studies (e.g. Hartley 1984, on Verulamium).
• Lists of pottery types compiled from other published sources.

Mapping pottery

The ‘English Landscape and Identities’ project (‘Studying English archaeology from 1500BC and AD1086’) sought to map aceramic, or minimally ceramic, zones across the country. For the Roman period, the data were extracted from the individual maps in RPB described above and combined together in a series of plots illustrating the number of fabrics represented (or ‘richness’) of the assemblage, divided by broad class (sigillata, amphoras, mortaria, fine wares, coarses wares etc) or by date (Green 2015).

A number of (potentially) ‘aceramic’ (or ‘less ceramic’) zones were thus identified, though some of these reflect the choice of wares for the original maps – chosen in part to illustrate points in the accompanying narrative – rather than any real absence of Roman pottery on the ground. A certain bias towards earlier wares, imports (particularly from Gaul), and the pottery of southern Britain is evident in the original selection of pottery types mapped in 1996, to the detriment, somewhat, of later Roman pottery, and the products of northern England. These biases are naturally reflected in any summary maps derived from these (Green 2015; Green/Creswell 2021, 10, 82).

An updated set of maps (2021)

Roman Pottery in Britain was published at a moment when web-based publications and archives in archaeology were still in their infancy. The Atlas pages on amphoras were made available on the newly created Internet Archaeology digital journal (Tyers 1996a) when none of the data used in their creation had been available digitally. It had to be transcribed manually from conventional printed publications.

Considerable quantities of distribution data on Roman pottery types are now available digitally. Significant additions include the following resources:

• A Survey of Samian pottery in Britain: Willis (2005)
• The Study Group for Roman Pottery Bibliography: Study Group for Roman Pottery (2014), incorporating data from the earlier Journal of Roman Pottery Studies bibliography, with additions up to 2012
• The Gallo-Belgic pottery database: Timby/Rigby (2007)
• ‘Names on terra sigillata’ (samian stamps): Hartley/Dickinson (2008–2012). The original data, corrections and additions can be downloaded from RGZM (2021).
• Corpus of amphora stamps: online at CEIPAC (2021).
• The Rural Settlement in Roman Britain project: Allen et al. (2018). This study summarises the data from some 2500 individual settlements, often using information derived from ‘grey literature’ reports. These data, available from the Archaeology Data Service, record, at a minimum, the presence of sigillata, mortaria or amphoras. For some records there are further details, and while some of these are rather limited (e.g. ‘Roman grey ware’) others list the principal wares present¹, or note more unusual vessels, particularly imports, and with some indication of quantities. More data could no doubt be extracted from this data set by returning to the original ‘grey literature’ reports.²
• The Pottery Kilns of Roman Britain: originally available as a monograph (Swan 1984), with the detailed gazetteer on microfiche. These data and additional records of more recently excavated kilns are now (2021) available from a dedicated website through an interactive map³. The kilns provide information on the location of the production sites for many of the Romano-British wares plotted in the maps.

Summary of the data

The database currently (2021) includes some 84000 records (for Great Britain), though this includes duplicates where the same vessels are recorded in different data sets. So a particular vessel may appear in the original excavations report, then summarised in the JRPS biblioghraphy entry, and finally listed in a synthetic regional survey or a study of a particular fabric.

Geographically this is spread across 1238 hectads (10 x 10 km squares), of which 1064 are in England – for comparison, England contains c. 1430 hectads⁴ – with the remainder in Wales (104) and Scotland (70).

Some 2000 different ‘fabrics’ are reported in these data, though this figure needs to be taken with a pinch of salt. The same ware may appear under different labels reflecting regional naming schemes, or may change from one publication to another to reflect the developing understanding of the material or genuine disagreements about the classification of the material by different practitoners. Furthermore much ‘legacy’ data on some vessel categories such as amphoras may only record a vessel form, with limited indication of fabric, although numerous studies demonstrate that some forms were manufactured and imported from several different regions. Cleaning, consolidating and refining the fabric grouping in the original data continues.

Increasingly the fabric codes developed for the National Roman Fabric Reference Collection (Tomber/Dore 1998), now available on-line from the NRFRC web site, are being used in pottery reports for the principal classes of imports and major regional and traded wares. These provide a certain level of consistency between reports. The presentation of important regional fabric type-series – often developed by now-defunct regional units during the 1980s – remains a work-in-progress. The on-line fabric series for Gloucester or Worcestershire demonstrate the continuing importance and potential of these collections.

The accompanying maps show the overall distribution of all records of Roman pottery in the data set as a point distribution map and a hectad map of the ‘richness’ of the assemblage – broadly the total number of ‘wares’ represented in each square. The ‘hot-spots’ (coloured red) on the latter highlight both the major urban and military sites in the province, and those with a high intensity and long-history of publication and study of Roman pottery – which is often the same thing.

Colophon: notes on software

The maps in RPB in 1996 were developed from data stored in a set of comma-separated text files, manipulated using standard UNIX text processing utilities such as grep, awk and perl. The images themselves were generated by the X11 drawing program tgif. The process is described in outline on p.206 of RPB. The maps in the online Atlas pages were rebuilt in 2004 from these data using the Generic Mapping Tools.

The current version holds the data in a postgresql database and the maps are generated by the GIS program qgis using the Biological Records Tool (Field Studies Council 2021). While the latter has been developed for biological and natural history recording it can be readily adapted for archaeological records such as those considered here. Sets of maps at different 'resolutions', such as simple points, or grids of different sizes (e.g. 2, 5 or 10km) can be generated and compared⁵.

May 2021; Updated October 2024