If the standard assessment is correct, then in English armies the warbow was not challenged on the battlefield by the handgun until the sixteenth century. However, given the significant changes that had happened in the technology of handguns by the 1480s compared to the 1460s, and the inadequacy of the documentary record, this is a question that may, in the last resort, only be finally answered by battlefield archaeology. Bosworth provides our first opportunity to address the issue in a systematic fashion.

Unfortunately it is not straightforward to establish which bullets within an assemblage, if any, relate to a battle and which do not.⁵⁶ The problems are as varied as the fact that lead ball was used with a catapult for birding, and the difficulty of identifying whether heavily distorted lead objects were actually bullets at all. The overriding difficulty is that right across England, from the sixteenth to the nineteenth centuries, lead ball was deposited from small calibre guns during sporting activity. This provides ‘background noise’ from which battlefield bullets must be isolated. When dealing with the later fifteenth century one must add to this our very poor understanding as to the calibre of handguns of the period. Thus it is impossible to search for relevant peaks in the calibre graph in the way one can, at least to some degree, for mid-sixteenth century bullets based on the data from the Mary Rose, or for mid-seventeenth century assemblages based on data from the Littlecote Collection.⁵⁷ For the later fifteenth century we currently lack a wreck assemblage or data from a large assemblage of closely dated handguns surviving in arsenals.

While the unusually large calibre bullets fired from hand-cannon should be easily distinguishable from the background noise, those for the much smaller bore arquebus may only be identifiable where they were used in large number. To tackle the issue we first need to compare the Bosworth calibre graph with data from surveys where the bullets are unrelated to military action (Figure 7.4).⁵⁸ There is some correlation of the Bosworth graph with that from Kings’ detecting in Northamptonshire, which we can use as an initial guide to background noise. Interestingly, when one compares the calibre graph for lead projectiles from Towton it also appears similar to King’s data, apart from the 22 mm and 30 mm projectiles which are discussed below.⁵⁹

That the bullets scattered right across the Bosworth survey are largely from sporting activity is supported by data from a small area close to Sutton Cheney. Here, far from the battlefield, at least eight ‘mould brothers’ have been identified – bullets with identical deformities which show they come from the same mould.⁶⁰ This may be a good indicator of sporting activity over a long period, where one individual has cast his own bullets, for no such grouping of mould brothers was found in the large bullet assemblage from Edgehill battlefield.⁶¹ Interestingly, when the Bosworth data is resampled to exclude the rest of the survey area (Figure 7.5) then different peaks appear in the calibre graph, although it is possible that the reduction in sample size means random factors are distorting the pattern.