Williams also shows the site of Sandeford, the place where the royal proclamation states that Richard was killed, which he bases on Hollings’ identification of 1858.³⁹ But Hollings had no pre-nineteenth century source and none which actually named the location as Sandeford. The site is not mentioned by Hutton and Nichols, who claim to have collected local traditions, and it lies far from the spot where they said Richard was killed.⁴⁰ The construction of the Visitor Centre in 1976 on Ambion Hill was followed by the placing of the flags and monuments at key locations in the landscape, based on the interpretation provided by Williams. A few Ricardian enthusiasts are even said to have had their ashes scattered beside the new monument – the ‘death stone’ – where Richard supposedly fell. Thus by the time of the 500th anniversary of the battle the process of anchoring the events within the modern landscape had been completed.

On that anniversary, in 1985, Colin Richmond dealt a dramatic blow to this comfortable picture. Following-up on Gillingham’s observations in 1981, Richmond showed that the character of Ambion Hill did not match that of the battlefield terrain as described in the contemporary accounts. He also noted that the chantry to the dead had been established in the chapel at Dadlington and the bodies were buried in its graveyard, so the battle should be expected to have been fought somewhere in Dadlington chapelry.⁴¹

The challenge was now taken up by Foss. He explored the issue in detail, using his skills as a local historian and his knowledge of the history of the Hinckley area to examine the wider landscape, as well as returning to the primary sources for the battle. His initial articles were superseded in 1990 by a more comprehensive study, which was reprinted with minor amendments in 1998 (Figure 1.15). Additional material about Dadlington, relevant to the battle, was subsequently published by Parry.⁴² Foss sought to reconstruct the historic terrain and then to fit the military history into this framework, identifying the most reliable topographical evidence from the primary sources by a process of detailed source criticism. Finally, he brought together the evidence of chance archaeological finds reported by antiquaries and others. Foss’s slim volume is the most fully referenced of all the studies on Bosworth, enabling his arguments to be re-evaluated.

Foss demonstrated, beyond reasonable doubt, that the battle was not fought on Ambion Hill and he presented a persuasive case that the battlefield lay somewhere within the low lying basin between Ambion, Dadlington, Stoke Golding and Shenton. Here his field name study showed a tight distribution of wetland and moorland in the historic period. He also consulted Hartley’s sketch mapping of ridge and furrow across Leicestershire, to identify extensive areas that could not have been marsh. Foss also refined Richmond’s discussion of the Dadlington chantry, where the battle is even called Dadlington Field. Finally, he noted a document of 1283 which proved that Redemore, the first name for the battle, lay in or immediately adjacent to Dadlington township.

The study was let down only by the scope, quality and accuracy of the plans, in which he made only a minimal attempt to place the events accurately within the landscape. The armies were positioned only very vaguely, with unrealistic distances between the forces and not even a guestimate of the likely frontages of the battle arrays. Thus, for example, when his plan is transferred to a modern map base he is seen to have placed Northumberland 1.5 km from Richard’s other forces, based on the dubious evidence in Drayton’s poem Poly-olbion. Similarly, Henry’s rearguard is positioned almost a mile (1.6 km) to the rear of the vanguard under Oxford. Such distances run contrary to good military practice of the time.⁴³ It is also interesting to note that even Foss had not completely abandoned Ambion, for it is on the lowest slopes of the hill that he has Richard and his vanguard deployed and from which the attack is launched. So, although Williams’ criticisms of Foss’s analysis were unreasonable and can be largely dismissed, many uncertainties did still remain. This was certainly not the ‘definitive, grass-roots reconstruction of the site and the action’ suggested by Bennett, but Foss had completely changed the context for all future discussion of the battle.⁴⁴

In achieving this Foss was following in a rarely acknowledged tradition of local historians making significant contributions to the study of battlefields – as with Twemlow at Blore Heath in 1912, and Page at Sedgemoor in 1935. Applying their skills to unravel the history of the landscape within which the events took place, they were able to provide insight which military historians were rarely able to bring.⁴⁵ But, unlike them, Foss had been able draw upon recent advances in the methodology of battlefield study, demonstrated by Newman in his work on Marston Moor in the 1970s. Here Newman had not only explored the historic terrain in a more coherent way than ever before but, in his own admirable but idiosyncratic way, he had also investigated the battle archaeology. This was at a time before even Scott had begun his pioneering work with metal detectors on the battlefield of the Little Bighorn, which is rightly considered to be the birth of modern battlefield archaeology.⁴⁶

When English Heritage compiled the Register of Historic Battlefields, in 1994 (Figure 1.16), the report on Bosworth reviewed the main primary sources for the battle and the evidence for the two alternative sites – that proposed by Foss and that on Ambion Hill. It also mentioned a third alternative, first suggested by Starkey in 1985, which lay near Atherstone, but it concluded that there was too little supporting evidence for the latter. The Register entry, finally published in 1995, avoided the difficulties by simply encompassing both Ambion and the Foss site in a single battlefield boundary, which also extended up to and included Crown Hill.⁴⁷

Following publication of the Register and in consultation with English Heritage, in 1995–1996 LCC established a steering committee, under Chris Brook (head of Historic Building Conservation) and Peter Liddle (County Archaeologist), to oversee a programme of archaeological work to prove that the battle did indeed take place on Ambion Hill.⁴⁸ In the absence of a project design the intended methodology is unclear, although minutes of meetings show the work was to include a metal detecting survey, fieldwalking, remote sensing, historic map analysis, production of a digital terrain model and core sampling. In the event, work apparently ceased in 1996. Although no records survive, those involved claim that nothing was achieved, with just a small area intensively metal detected, according to an unspecified random sampling method, and a slightly wider area fieldwalked, all focussed on the top of Ambion Hill. In 1998, in collaboration with Natural Environment Research Council, LCC acquired thermal imagery in three transects across the battlefield, from Fen Lane to Ambion Hill. This was subsequently analysed as part of PhD research programme, seeking evidence on the location and extent of the marsh and of mass graves. The work failed to reveal any useful information not already visible on the RAF vertical air photographs of the 1940s.⁴⁹ The one significant outcome of the new initiative was the revival, in 1999–2002, of the metal detecting and fieldwalking survey. This was re-established on a wider scale, after limited discussion with others working on battlefield sites in England, and was carried out by a volunteer team supervised by Richard Mackinder, then a ranger at the Visitor Centre.