Extract from 'Iron Age Communities in Britain' Prof. Barry Cunliffe

Much has been written on the subject of currency bars (e g more recently Tylecote 1962 206-1 1; D.F. Allen 1968b), leading to the conclusion that there were many regional variations and that, since many of them were carefully hoarded, they were likely to have been of value. The simplest explanation is therefore that the bars represented a medium of exchange or barter, the iron content providing an actual as well as a perceived value. Further implications might be that they were manufactured by specialist iron-smelters and were traded to the neighbouring communities within restricted regions, the bars being then either hoarded as a form of wealth or manufactured into implements.
Analysis of the slag inclusions in bars from three hoards, Beckford, Danebury and Gretton, showed that it was possible to distinguish three different sources but the task of tracking down the actual location of the ore would be very much more difficult (Hedges and Salter 1979
The very existence of the bars, and the fact that each hoard was probably produced from a single source (at least in the three cases examined) is sufficient to show that iron extraction had become a specialist skill in south-east Britain by the second century BC. Although exact dating of furnaces is difficult and the number known is not large the majority seem to belong to the period before the second century. It is tempting therefore to suggest that there may have been a move away from small-scale local production to more centralized production in the south-east as the Iron Age progressed. Such a change would have required the development of more organized redistribution mechanisms. In this context it is interesting to note the concentration of currency bar hoards in hillforts and the evidence from Danebury that bars were being cut up there for manufacture into tools and weapons. The appearance of the bars, implying specialist production. may have taken with it the development of specialist smiths working from selected locations.
Eventually there may be enough analytical data to test this hypothesis.