The first task is to locate the battlefield, so we must begin by studying the landscape within which the battle was fought. Unfortunately, as we have seen, there is very limited evidence in the primary accounts and battle names from which we can work. Indeed, it is this paucity of detail, together with the degree of landscape change that has occurred since the medieval period in this area, which is the principal reason why the Bosworth battlefield has proved so hard to locate.
The difficulties are compounded by the concern of historians that authors of late medieval chronicles may often, in the absence of authentic information, have provided idealised representations of combat drawn from military manuals or from accounts of earlier battles. If this is true then even the few topographical details that do exist for Bosworth might prove spurious. However, there is an alternative explanation for such similarities – that they simply reflect the frequency with which comparable situations were encountered in battle and, for good tactical reasons, elicited similar responses. Indeed, it is almost certain that the actions of commanders, in responding to such threats or opportunities, were influenced by their reading of those very manuals and by their knowledge of how earlier commanders had acted in similar situations. This is, after all, a principal way in which senior officers have always learned their trade – by studying the actions of their predecessors, and reading military manuals.
When investigating any medieval battle it is therefore essential that all the topographical and other detail from the primary sources is included in the analysis, unless wholly discredited. Once a battlefield has been located, the terrain reconstructed and the battle archaeology explored, then often obscure details that previously have been ignored can prove of the greatest significance.
What then of the topographical clues from the original accounts that were reviewed in the previous chapter? Most important are the references to a marsh, which appear independently in two reliable sources as well as in the ballads. Several of the battle and related names also support the idea of a low lying, waterlogged area. Thus the ‘mor’ of Redemore indicates poorly drained ground, while Sandeford indicates a road crossing a watercourse. In contrast the heath-related battle names indicate dry ground and thus, at first sight, they seem to define a quite different landscape. However, they can be considered from a different perspective, for all three (marsh, moor and heath) represent marginal land use. In the landscape of highly nucleated settlement seen in the East Midlands in the medieval period, such marginal land typically lay at the periphery of townships, whereas the villages tended to lie towards the centre of their township or at least at the heart of the best agricultural land. If the battle was fought on marginal land at the periphery of several different townships, then this might also explain why the immediately adjacent settlement names did not figure prominently in the naming of the battle.
Thus we are probably seeking a low lying marginal tract of land, on the periphery of several townships, which encompassed fen, moor and heath, and where a road crossed a watercourse. Significantly, wetland was not common or extensive in medieval Leicestershire, while the proximity of heathland further narrows down the options.
Then there are hills reported in the accounts. The most significant is that upon which Henry was crowned after the battle. It is only mentioned by Polydore Vergil, but he gives it a very prominent place in both time and space, making clear it was near to the battlefield. Then there is the high hill upon which stood a windmill where, in the rout, the duke of Norfolk was supposedly killed and his son captured. This appears in The Song of Ladye Bessiye, which is a highly suspect source where mountains and high hills abound. Yet, while these features may, to some degree, have been added for dramatic effect, there seems little obvious reason to invent a windmill. So, perhaps that is a genuine association recorded in folk memory. While a windmill may be expected in most townships during the medieval period, if those in use in the fifteenth century could be located then they would provide just a handful of very precise locations.
Finally we have the association with Dadlington, which arises solely within the documents relating to the chantry for the dead, created in the first years of Henry VIII’s reign.
Whereas Redemore and Sandeford, and indeed the fen, moor and heath, may be expected to identify a relatively restricted area for the battlefield, the reference to Dadlington provides a fairly broad canvas, taking us to a township covering more than 4 km2. In addition, given the importance of Ambion Hill in earlier interpretations, it was essential that this be encompassed by our study, if only to enable it to be ruled out as the site of the battle on the basis of secure evidence. The area finally chosen for reconstruction of the historic terrain was based around the zone across which Saxton’s battlefield area appeared to extend (see Figure 1.10). However, for practical reasons relating to the way in which the medieval landscape was organised and documented, it was also necessary to encompass complete townships. This gave a study area comprising the whole of Sutton Cheney, Dadlington, Shenton, Stoke Golding and Upton. Later we had to add small parts of adjacent townships, most notably Higham on the Hill. With hindsight, it can now be seen we should also have encompassed other land on this western periphery, so as to incorporate the whole of the upper catchment of the Sence brook.
There were two objectives in undertaking the reconstruction. The major concern was the terrain of the battlefield itself, so that the site could be located and the events placed within the landscape as it was on the day. However, consideration was also given to the wider context, which would assist in understanding the routes taken by the armies as they approached the battlefield. This can be particularly important for the study of medieval battles, where the topographical details in the primary accounts are so sparse.
The principal influence on the movement of medieval armies will have been the network of major roads. They will normally have provided the most direct, widest and easiest of routes, particularly where they ran through land that was wooded or already enclosed. Also significant were the bridges, for only on the major roads will they normally have been sufficiently wide and strong to allow the passage of the many heavy wagons and artillery of the army’s train. In contrast, in the manoeuvring which immediately preceded the action, speed and ease of movement may not have been a commander’s only or even his primary consideration. Thus, for the immediate context of the battlefield one would ideally wish to understand other aspects of the wider landscape – whether the ground was open or enclosed, and where land use such as woodland or heath might have impeded or facilitated the movement of large numbers of troops.
Similar considerations will have influenced the choice of ground by the army which deployed first. The commander will have exploited features which favoured the strengths of his forces, provided protection for his troops or placed the enemy at a disadvantage. Thus, a reconstruction must consider not just the features mentioned in the battle accounts but also the broader aspects of the historic terrain, which may have influenced tactical decisions.
The methodology for mapping the historic landscape is well developed, but its application to the study of battlefields poses unusual problems, for one is seeking to reconstruct a single day in the life of a landscape. Moreover, while in most landscape analysis one can choose case studies from a wide range of places on the basis of survival of relevant evidence, for battlefields the landscape to be explored is pre-determined. This may severely limit the potential for reconstruction, for one is dependent on the vagaries of survival of documentary and physical evidence.
The complexity of the task is also affected by the way in which the landscape has evolved. Midland England has seen a variety of changes: sometimes large scale planned transformations; sometimes numerous small changes having major cumulative effect over decades or centuries; whereas at other times there may have been long periods of relative stability. The ‘marsh’ reported in the primary sources for the battle of Bosworth provides a clear example. It will not have been stable in character or extent ever since the last ice age, when the landscape of England was completely redrawn. There will have been a process of natural evolution, especially progressive silting. This will have been hastened by the impact of woodland clearance in prehistoric and Roman periods, and by the massive expansion of arable cultivation in the late Saxon and early medieval period, which mobilised soils and greatly increased the volume of silts being laid down on valley floors.1 Then there is the impact of intentional human action, particularly through drainage schemes. To compound the difficulties of our task, much of the physical evidence upon which one can draw has a chronological precision of decades or more often of centuries, whereas we are concerned with the detailed character of the landscape on one day in 1485.
The problems of reconstruction will vary dramatically depending on the period being studied, the region, and the particular type of landscape within which the battle was fought. This is because the regions of England often have a very different landscape history. For much of the medieval period the majority of Leicestershire, like the adjacent counties of Warwickshire and Northamptonshire, was under arable cultivation with a separate open field system in each township. This is ideal, because an effective archaeological methodology for the mapping of such landscapes has been demonstrated by Hall.2 Problems do of course still arise. For example, while one can be confident when mapping the landscape at the height of the agricultural expansion, in the late thirteenth and early fourteenth century, there was reversion of some land at various times thereafter from arable to heath, furze or most often to open or enclosed pasture. This was particularly true where marginal land was involved. Usually it is only where there is good documentary survival that such issues can be adequately addressed. The potential for terrain reconstruction can also vary dramatically between adjacent townships. This depends not only upon when and how enclosure took place, but also upon the vagaries of documentary creation and survival. Thus, for example, if land was owned by a monastic house or a major secular lord, or if a legal dispute had taken place then records are more likely to have been compiled and to have survived.
Fortunately for our study, in this region at least, relatively little enclosure had taken place prior to 1485.3 Of course, in the limited areas where land was already enclosed in hedged fields, even if on a small scale, this could have major tactical implications. However, there is nothing in the primary accounts of the battle to suggest that it was fought in anything other than wholly open ground.
The reconstruction of the historic landscape of the five townships was undertaken by a team of specialists in historical, archaeological, palaeo-environmental, soils, and place name survey.4 The present chapter represents a synthesis of the evidence in their reports, but it cannot be considered definitive. A search for primary written sources is rarely exhaustive, while the archaeological detail for the open fields and the palaeo-environmental evidence for the changing extent and character of the wetland over time could all be enhanced. Some of the limitations arise because the terrain study was completed before the true location of the battlefield was established and so it was not possible to effectively target the work. This is most obviously the case with the documentary sources for Upton township, which were not originally seen as a high priority as the township seemed to be peripheral – now it is clear that it encompasses the heart of the battlefield.
The starting point for any reconstruction is the underlying physical geography. This comprises the relief and drainage pattern, together with and influenced by the underlying geology and the geomorphology. In the present study relief and drainage for the wider context of the battlefield is provided from the (out of copyright) 1 inch scale New Popular Edition Ordnance Survey mapping (Figure 4.7). For the finer detail required for the battlefield itself we have used the NEXTMap Britain digital terrain model (dtm) at 5 m resolution, from which a contour map has been generated (Figure 4.1). As data was not filtered to remove the distortions caused by woodland there are one or two aberrations, most notably in Ambion Wood. The data has also been somewhat over interpreted, to produce contours at 25 cm height interval, in order to give a clear visual impression of the minor changes in relief in the relatively level ground where the battlefield lies.
Relief has been largely stable over the last 500 years. While there has been no major quarrying, the construction of the canal, under an Act of 1794, and the railway in the nineteenth century have both caused localise alterations to the landform and have impacted on drainage. In the late twentieth century a small group of ponds were also dug in the low lying ground to the north-west of Crown Hill. Other earthmoving was undertaken in the 1940s and in the late twentieth century on the former airfield at the south western corner of the map. In the heart of the battlefield, as now understood, there has been almost no alteration in ground levels outside the farm complexes.
The study area focuses on a wide, flat bottomed basin with impeded drainage, from which the water issues northwards along a single channel running through the village of Shenton. This is surrounded by an arc of higher ground which is particularly distinct in the area of Stoke Golding and Dadlington on the south-east and Ambion on the north-east.
There have been substantial man-made changes to the drainage pattern since 1485, first in the form of drainage channels and then the installation of an extensive pattern of land drains in the nineteenth and twentieth centuries.5 In addition there have been natural changes over a much longer timescale, which mean that the soils and palaeo-environmental evidence partly reflect earlier conditions rather than the situation in 1485. The watercourses are depicted here from the first edition 6 inch scale Ordnance Survey mapping of the 1880s (Figure 4.1). While the main watercourse, the Sence brook, has a recorded name, the three tributary brooks on its south-west side, now essentially man-made drainage dykes, seem to lack specific names. Further research may identify their correct names but, for practical purposes, they are described here as the Fowlismere, Fen and Fen Meadow brooks, from significant features through which they flowed.
For the surface geology we have drawn upon the 1:10,000 scale mapping by the British Geological Survey. For much of the study area this comprises drift deposits, ranging from fluvio-glacial sands and gravels or clays, through extensive areas of boulder clay (till), to deposits of alluvium in the lowest lying areas beside the streams. Across limited areas the solid geology of mudstone and sandstone reaches the surface to influence the character of soils and drainage. This is particularly true in the in the south-western part of the basin, where the battlefield is now known to be centred. This varied geological base will have influenced the development of soils – including soil type, pH and drainage – which in turn will have influenced land use and the survival and condition of artefacts and of human remains. However, the resolution of the geological mapping is too coarse for the detailed analysis required here and so it was necessary to turn to soils mapping.