1. Introduction

It seems clear that post-Roman British occupation lasted for as much as 40 years in Kent and considerably longer in parts of Sussex, Hertfordshire, Hampshire and Wiltshire. Where are the British artefacts for this period in the south of Britain? Some items have been recognised but very few for what is an appreciable period of time. Is there something wrong with our archaeological dating? In this article I will seek to show that there is.

The old assumption that the use of coinage ceased in Britain with its withdrawal from the Roman Empire may be one of the root causes of the problem. As long ago as 1954, work on a mainly unstratified 4000 coin sample of the large number of Theodosian nummi from Richborough detected two minted in 421-3 and one each minted in 423-5 and 425-35 (Kent 1954): other early 5th-century nummi have also been identified from St Albans, Wroxeter, Dunstable and elsewhere (Abdy 2006). Such base coinage has no value outside of a full monetary economy unless its usage had become socially acceptable.

This suggests that some parts of Britain, at least, continued to use base coinage well into the second quarter of the 5th century (Moorhead and Walton 2014, 112-14) and perhaps even later. The bulk of this circulating coinage continued to be the 395-403 dated Theodosian issues with occasional later coins: there is also some evidence from the latest hoards for the re-monetarisation of older 4th-century nummi and late 3rd-century radiates. These hoards include those from Weymouth, Wroxall, three from Richborough and one from Caerwent (Robertson 2000, nos 1525, 1540, 1543, 1547, 1611 and 1632).

If a monetary economy of sorts did survive into the mid- or even late 5th century in certain areas, then it may be that the latest Roman pottery in southern Britain should be dated to c. 370-435/50 rather than c. 370-410. This would go a long way towards explaining the apparent lack of artefacts datable to the period c. 410-450 in the south of Britain, as associated pottery and coinage is used to date other types of objects, such as the peculiarly insular Tortworth strap ends and horse-headed buckles, to the very end of the Roman period when they may in fact be entirely sub-Roman in date.