Villa and Town in Roman Britain

Eleven years ago Mr A. L. F. Rivet observed with reference to die distribution of Romano-British villas that 'the main development often takes place not near the original cantonal capital but near a somewhat smaller town'. The important social implications of this observation seem to have gone unnoticed. In so far as villas represent the dwellings of an upper class this distribution ought to give us a hint of the habits of the gentry in Roman Britain. The best clue is perhaps analogy. The seat of the duke of Omnium is Gatherum Castle. The duke's family may appear at social functions (sparingly) or more commonly on the hustings in the local town but can hardly be imagined to have had a house there. The town houses of people of this sort are found in London, the spas and other such centres of fashion and politics. At a less exalted level we may imagine the gentry of the Romano-British civitates having their villas in the countryside within convenient ränge of a market town and their town houses in the civitas capitals. Those who moved on a higher social and political plane may well also have had houses in the provincial capital and it is probably at places such as London, Cirencester and York that we should look for die urban counterparts of Lullingstone and Bignor.