Site-Finds in Roman Britain

Abstract

Britannia iii I was able to give a very summary and outline picture of Roman coinfinds in Britain. 1 A year later this picture was extended to France and Italy, 2 and various similarities and differences were noted. At the same time John Casey was working particularly on military coin-finds, and he came independently to very similar conclusions. 3 A framework had been established: there was a British pattern which was different from the continental pattern and all British sites, to a greater or lesser extent, lay within this pattern. The most obvious features, drawn out in diagram form in Britannia iii, 4 included low coin-loss on every site from the Claudian conquest to A.D. 260, high coin-loss from 260 to 294, a dip from 294 to 330, a major peak from 330 to 348, and a general high level of coin-loss in the fourth century. In the last twenty years various methods have been tried in order to separate out the very different types of sites within Roman Britain-forts, temples, towns, and villas for example-into different patterns of coin-loss. The assumption on which this was based, though never made explicit, was that since there was reason to believe that different activities went on at forts and temples this ought to result in different patterns of coin-loss, and these in turn ought to be visible in the archaeological record. There have been some signs of success but, in general, any divisions found were seldom clear-cut and the methods used were judged to be complicated. Several review articles and chapters have summarised this work in detail, with full bibliographies, so that only a short summary will be given here; more details can be found in the works cited. 5 A major move forward came when it was recognised that, within the overall British pattern, towns had a pattern of coin-loss which was different from all other sorts of site. Since all sites other than towns are by definition in the countryside this gave a separation out into urban and rural patterns of coin-loss. The characteristics of the two groups depended on the number of coins from the late third century (260 to 294) compared with the number of coins from the later fourth century (330 to 402). Towns, in general, have equal numbers of coins from the two periods; other sites have more coins, sometimes three-or four-times more, in the later period. It must be noted that this differentiation 1 'A short survey of the Roman coins found on fourteen sites in Britain',

Quantifying artefacts over time: Interval estimation of a Poisson distribution using the Jeffreys prior

“… Quantification of coins by mintage date from Roman Peasant Project sites, 2009–14, according to the periodization of Reece ( , ). …”

“…As such, the topic has obvious applications in archaeology, and has been of increasing interest the more that computational methods are applied in the field (Crema et al 2010;Crema 2011Crema , 2012Crema , 2015Bevan et al 2013). Essentially, the procedure entails dividing a quantity of material over its possible time periods, and then summing its fractions per unit of time: that is, if one has two artefacts that could have been in use over a span of Reece (1995Reece ( , 1996. 100 years, their quantity would be 2/100 in any year.…”

Exploring Romano–British finds assemblages

“…It is of some interest to be able to compare these formally as it has long been clear that different types of sites have different patterns of coin loss (Reece 1993). A variety of attempts were made to do this (summarised in Lockyear 2000, 397) culminating in a method which presented cumulative percentage curves adjusted for the British mean (Reece 1995a). The results were sorted by eye and similar groups plotted together.…”

Military Identities in Late Roman Britain

“…As before, the coin assemblages are the easiest to quantify, if not to interpret. Coin data from a number of sites has been compiled to plot against Caerleon (Gardner 1997, 38;Gardner forthcoming), using an extension of the methods described earlier, again following Reece (1995). Figure 5 allows direct comparison of site assemblages with each other, while retaining their relationships to the British Mean.…”