4. The Distribution of Late Roman Coinage in Britain

The distribution pattern of coinage issued in the early 5th century and found in Britain is restricted. Gold coins of the late 4th century AD are found as far afield as Corbridge in Northumberland and Holyhead in Anglesey (Bland and Loriot 2010, 42-3, figs 33-34) whereas 5th-century pierces are largely confined to the south-east, with concentrations from Kent to East Anglia, and in the south from West Sussex and the Isle of Wight to Dorset (Figures 1 and 2) The majority of late Roman gold coins come from military or urban contexts (Bland and Loriot 2010, 54, table 12).

Figure 1
Figure 1: Roman coin hoards from Britain with a terminus post quem of AD 402 or later. Note that the Traprain Hoard (East Lothian, Scotland) and Coleraine hoard (Northern Ireland) are off the map
Figure 2
Figure 2: Single finds of Roman coins in Britain, struck after AD 402

The majority of Theodosian siliquae and clipped siliquae are also found in southern Britain (Figure 3) in the 'lowland' zone to the south and east of the Fosse Way with an extension into East and parts of North Yorkshire. Finds in Devon and Cornwall, Wales, the west Midlands and the north-west and north-east are rare. Interestingly, hoards and stray finds of siliquae are common in rural areas, with very few specimens being found as site-finds on military or urban excavations; Richborough, and possibly Cirencester, are the notable exceptions (Reece 1972, table 1a).

Figure 3
Figure 3: Silver siliquae of the period AD 388-402 (Reece Period 21) found in Britain (It is not recorded if the Welsh siliquae were clipped or not.)

Indeed, the distribution of nummi is in marked contrast to the distribution of siliquae. Although on excavations, stray finds of bronze nummi far outweigh the number of silver siliquae (316:1 at Richborough; Table 2), the rural dataset provided by the PAS shows a much larger proportion of siliquae to nummi. This suggests that siliquae were more common in the countryside than on urban sites. This may be because detectorists do not report these small and often virtually illegible coins, but the authors' experience of dealing with large, complete, rural assemblages does seem to suggest that the finding is valid. It is possible that silver coinage circulated in rural areas because of its intrinsic value, whereas bronze nummi only played a role in the economic spheres of urban and/or military life.

One can also question whether the finds of nummi (whether as hoards or site finds) at certain urban centres, for example Dorchester-on-Thames and Canterbury, are intimately linked with a military presence (Moorhead, Anderson and Walton in press; Wacher 1995, 203-5; Burnham and Wacher 1990, 121-2). In this regard, it is interesting to note that a large number of major assemblages and hoards of Theodosian bronze coins are found in coastal or estuarine regions. Does this suggest strong links still existed between the Continent and major centres in Britain that were easily accessible from the sea? Interestingly, in the Netherlands and Belgium, after a lacuna of coin-loss in the mid-4th century, there is a sudden, albeit short-lived, spate of hoarding of Theodosian bronze coins in the late 4th and early 5th centuries. Perhaps the Roman authorities were trying to maintain maritime links between centres in southern Britain and the Rhineland (Moorhead, Anderson and Walton in press; Stroobants 2013).

Table 2: Relative proportions of nummi and siliquae from site assemblages and recorded with the Portable Antiquities Scheme (see Bland et al. 2013, 131, tables 6a-d)
Assemblage Reference Nummi Siliquae Ratio
Richborough Reece 1991 22,750 72 316:1
Non-Richborough sites Reece 1973 1,949 103 189:1
Caerwent Guest and Wells 2007 3463 1 3463:1
Wales, non-Caerwent Guest and Wells 2007 63 6 10.5:1
PAS England queried 12.12.2012 1,004 271 3.7:1
Figure 4
Figure 4: Single-finds of Theodosian nummi, AD 388-402 (Reece Period 21) and bronze hoards with a terminus post quem of AD 402
Figure 5a Figure 5b Figure 5c
Figures 5a-c: Maps showing assemblages with above-average coin-loss for the late 3rd and early 4th centuries (Fig, 5a), mid-4th century (Figure 5b) and late-4th century (Figure 5c). The assemblages were divided into three sub-groups based on their chronology using DMax Cluster Analysis. For more detailed discussion of average values, see Walton 2012, 103 passim.