Introduction

The study of trade networks in Roman Britain has long relied on mapping the distribution of distinctive pottery types. Black-burnished ware from Dorset, Severn Valley ware from the Welsh borderlands, and Nene Valley colour-coated ware from the East Midlands each followed distribution corridors that reveal the economic geography of the province.

This study applies network analysis to a newly compiled database of 4,200 pottery assemblages from excavation reports published between 1970 and 2023. Rather than mapping individual ware types, we model the relationships between production sites and consumption sites, identifying hub settlements and gateway communities.

Methodology

Data Collection

We compiled pottery data from 4,200 stratified assemblages across 312 Romano-British sites. Each assemblage was classified by fabric type, form, and date range. Minimum vessel equivalents (MVE) were used where available; otherwise sherd counts were weighted by average vessel mass.

Network Construction

Sites were modeled as nodes in a directed network, with edges representing the flow of pottery from production zones to consumption sites. Edge weights were calculated using a gravity model incorporating assemblage proportions and the inverse of inter-site distance.

Analysis

Community detection (Louvain algorithm) was applied to identify trade clusters. Betweenness centrality identified gateway settlements mediating between clusters. Results were cross-referenced with the known Roman road network.

Results

Trade Communities

The analysis identified seven distinct trade communities, broadly corresponding to known economic regions but with several surprises. The most striking was a previously unrecognized cluster linking sites along the Jurassic Way between Cirencester and Lincoln, suggesting that this geological ridge served as a major trade corridor independent of the road network.

Gateway Settlements

High betweenness centrality identified several gateway communities not traditionally associated with major trade. Alcester (Warwickshire) emerged as the third-most important gateway in the network, mediating between the Severn Valley production zone and the East Midlands market.

Discussion

These findings challenge the model of a primarily road-driven distribution system. While the major roads undoubtedly facilitated long-distance trade, the network structure suggests that natural corridors and river systems played an equally important role in shaping regional exchange patterns.

References

Tyers, P. (1996). Roman Pottery in Britain. London: Routledge.
Fulford, M. and Hodder, I. (1974). "A regression analysis of some late Romano-British pottery." Proc. Prehist. Soc. 40: 159-167.
Allen, J. and Fulford, M. (1996). "The distribution of South-East Dorset Black Burnished Category 1 pottery." Britannia 27: 223-281.

17:[["$","script",null,{"type":"application/ld+json","dangerouslySetInnerHTML":{"__html":"{\"@context\":\"https://schema.org\",\"@type\":\"ScholarlyArticle\",\"headline\":\"Inland Trade Networks in Roman Britain: Evidence from Pottery Distribution\",\"description\":\"Spatial analysis of ceramic finds from 200+ excavation sites reveals a previously unrecognized network of inland trade routes connecting major Romano-British settlements. Contrary to the dominant model of coastal-dominated commerce, our clustering analysis shows that river systems and road networks facilitated substantial inland exchange, particularly for coarseware pottery produced in the Midlands.\",\"author\":{\"@type\":\"Person\",\"name\":\"Elena Rossi\",\"affiliation\":{\"@type\":\"Organization\",\"name\":\"University of Bologna\"}},\