A Guide to the Archaeology of Conflict

© T.L.Sutherland 2005

Towton: An example of the successful archaeological British medieval battlefield


In August 1996, building work at Towton Hall, Towton, North Yorkshire, disturbed part of a medieval mass grave from the battle of Towton (fought between the rival factions of the Houses of Lancaster and York, AD 1461). The builders are reported to have removed 24 skulls and the bones from an indeterminate number of skeletons. In September of the same year, a team of osteologists and archaeologists from the University of Bradford and WYAS Archaeological Services excavated the remaining part of the grave. This resulted in the recording of parts of at least 38 skeletons (Sutherland 2000a).

In early 1997 Tim Sutherland, a member of the original excavation team, began research into the battle of Towton, specifically aimed at placing the evidence from the mass grave within the archaeological battlefield context. In late 1997, Sutherland began intensive archaeological investigations as part of a PhD thesis in an attempt to locate and record physical evidence of the medieval battle (Sutherland forthcoming 2), under the title of the ‘Towton Battlefield Archaeological Survey Project’.


The project initially began by assessing the landscape in which the battle was fought. In most battles, the faction that chooses the location of the battle should have a distinct advantage, as it can use landscape features, such as steep terrain, wet ground, enclosed land and settlements to defend its flanks and rear. At Towton, the steep slopes leading down to the River Cock protected the Lancastrian right flank (see Figure 2), who were on the field first, whilst the low-lying wetter ground, which was also highly visible from the raised central ground, protected their left.

If the unknown location of a battlefield is sought, then it is important to analyse the landscape topography and the contemporary road network: the larger the army, the greater their dependence on a good road for the movement of large quantities of equipment and provisions. Existing major road networks allowed access and the movement of troops and, more importantly, baggage trains, to the site of engagement - there are few large historic battles that did not utilise well-constructed roads.

The archaeological survey work at Towton began by looking for ferrous artefacts using archaeological geophysical magnetic surveys. However, when excavating a sample of the ferrous anomalies discovered during several of these surveys, it was concluded that the artefacts were either unrecognisable or were generally a result of modern farming or dumping practises. They were thus perceived of as a form of 'contamination' within the battlefield assemblage. This form of battlefield prospection survey was therefore considered inefficient (Sutherland 2000a).

An archaeological field walking survey was also initiated but, once again, virtually all of the artefacts recorded were found to be either unidentifiable, related to manuring practices, or were from a period other than that of the battle. Additionally, in order to examine each unidentifiable ferrous artefact more effectively, expensive and time-consuming radiographs would have to be taken. This procedure was considered to be prohibitively costly. It was therefore initially considered that this method of locating evidence of battle might be too time-consuming as a prospection method alone. However, it later proved effective in the detailed analysis of smaller areas, so its use was adapted (see below).

A decision was then made to alter the method of investigation from the search for ferrous artefacts to those of non-ferrous metals, as non-ferrous artefacts usually have a better rate of preservation compared to ferrous items and can therefore be more easily identified. It was also considered that the site was likely to contain fewer nonferrous metal artefacts, be they fifteenth century or once again 'contamination', so that they would be more straightforward to isolate.

As magnetometers will not locate non-ferrous metals, electromagnetic (EM) prospection surveys were initiated using a specific EM survey instrument (White's TM 808, a large metal detector). The advantage of this instrument was that it could be 'tuned' to an infinite number of settings and, with a certain amount of expertise; a logger could be attached to it in order to record the results. Although this instrument was then in its initial stages of development this 'loggable' metal detector proved productive in its ability to locate archaeological features and record non-ferrous metal artefacts.

At an early stage of the research (1997) a metal detectorist (Simon Richardson) was approached, who had already been searching the battlefield for several years and making sketch maps of artefact scatters he found that related to the battle. He was asked if he would join the survey project and transfer all of his data on to 1:2500 scale OS maps. From then onwards he marked the location of where he found any fifteenth century artefacts on these maps. Although time spent searching the battlefield in this manner was limited, due to it being a part time, and unfunded occupation, it proved successful in that details of artefact scatters, not noticed before, became evident. This eventually developed into a more accurate survey method by employing a small hand held satellite navigation instrument to record a relatively accurate location (plus or minus a few metres) for each artefact. However, a balance had to be achieved between randomly searching the battlefield when fields became available to be detected and a systematic survey of the area. The latter is difficult to achieve, when only a single individual is searching the land and the former produces a bias within the data, as the searcher tends to remain longer in areas that produce a greater quantity of artefacts.

Concurrent with the metal detecting, an additional multidisciplinary array of archaeological prospection techniques was instigated, including aerial photographic analysis, excavation of test pits and the full archaeological excavation of individual features, in an attempt to locate evidence of the battle. Earthworks were also analysed, former fields systems that would have been contemporary with the battle were recorded, and the geology and topography of the landscape was mapped out to assess which areas of land were most likely to assist or hinder medieval armies as they prepared to engage in battle. The geology of the landscape was also analysed to gauge what the rate of artefact decay would be for artefacts buried in the soil covering the battlefield - there would be little point in searching for artefacts that would have corroded long before they could be located and recorded. Fortunately for the survey results, the buried soil at Towton (soil lying on the surface is open to additional metrological effects) is excellent at inhibiting the corrosion rate of buried metal artefacts and degradation of human bone.

An important part of the investigation was the search for the historically documented mass graves from the battle, the location of which had been lost for over 150 years. The historical archive was searched for primary documentary evidence of the battle and the results analysed. Subsequently, all of the historically documented sites of the suspected mass graves were identified and investigated using geophysical survey and trial excavation. All of these previously recorded sites were, however, found to be erroneously associated with the battle. Many proved to be either prehistoric or Romano-British features that had no association with the conflict. The historical archive, where appropriate, was therefore reinterpreted. Eventually it was determined that the search for the graves would have to continue on alternative, and so far un-investigated areas of the battlefield.

Meanwhile, newly identified (mainly non-ferrous) artefact concentrations were proving interesting. Rather than the ferrous military artefacts scatters, which were initially envisaged as marking the general site of the battle, it was confirmed that the non-ferrous, non-military artefacts in fact highlighted the battlefield site (Sutherland 2003). Rather than arrowheads and weapon fragments pointing to the areas of conflict, clothing fasteners and buttons, belt fittings and badges suggested where further investigations should take place (ibid).

These non-ferrous artefact concentrations sometimes formed specific patterns within the landscape. These patterns were subsequently investigated using the previously mentioned array of archaeological prospection techniques, with successful results. For example, one concentration of artefacts formed a relatively narrow line across the landscape in, what appeared to be, an ideal location for medieval armies to be arrayed in the initial positions of engagement for battle. This area was subsequently re-investigated using metal detector surveys, recording both non-ferrous and ferrous artefacts in an attempt to find small ferrous fifteenth century artefacts. Although the ferrous 'contamination' was still great, a number of medieval ferrous arrowheads were discovered. Further surveys were carried out in other areas close-by, resulting in the identification of a distinctive square anomaly, fifty metres across, which was defined by arrowheads alone. Further geophysical surveys were undertaken over this 'square'. Additionally, as the surveys were carried out, other important finds found on the surface, were also located and recorded, thereby carrying out an alternative, simultaneous field walking survey, which produced important results. During one of the metal detector surveys, the lower part of a human arm bone (distal ulna) was located within this 'square' feature. Additionally, during a subsequent earth resistance survey, two human teeth were discovered. These teeth were seen to coincide with the location of a long narrow rectilinear anomaly found using a magnetic survey, which was interpreted as a possible linear pit. A trial trench was excavated across this anomaly and one side of a ditch was encountered. Within the fill of this feature were hundreds of disarticulated human bones and teeth. It is probable that this feature, a former grave, contains the disarticulated human remains that were left following a clearance operation described in an unpublished grant by King Richard III. This refers to the removal of the bones from Towton Field in 1484, 23 years after the battle. The full excavation of this grave is due to take place in 2006.

The discovery of this former mass grave and the concentration of artefacts, and particularly the arrowheads, provide compelling evidence for the location of the battle of Towton. This proves that the methodology used has been successful in locating extensive archaeological evidence of a medieval battle. Furthermore, there is no reason why the application of a similar methodology to examine and identify other historic battlefields should not be as successful, if the evidence is still preserved and can thus be located.

Once significant archaeological evidence of a historic battle has been located, it should be possible to reevaluate the historical evidence to assess whether or not it is verified by the former. The research at Towton, for example, has proven invaluable, in that features formerly associated with the conflict have been shown not to be so. Additionally, new information has subsequently been used to interpret historical features within the landscape that had not previously been fully understood. For instance, it has since been established that the location of the graves on the battlefield lies very close to a feature marked on early Ordnance Survey maps as 'Lord Dacre's Bur Tree', the site of a legend associated with a leading Lancastrian noble who fought and was killed in the battle of Towton. It is therefore probable that the approximate location of this former tree stood on the site of a grave marker of the mass graves. It is therefore also possible that similar features or structures on other medieval battlefields mark the location of other unknown mass graves from those conflicts.

The Towton Battlefield Archaeological Survey research is the first multidisciplinary archaeological survey of its type, which has successfully located physical evidence of a medieval battle. It has proved so successful that other similar surveys have been, or are about to be initiated on other medieval battlefields (Azincourt, France; Bosworth, England) using the methods formulated and information gathered during the Towton survey.

Concise Methodology
  1. Locate the approximate area of a former battle by analysing the geology, topography and road networks and comparing this information with historical descriptions of the site
  2. Locate and record diagnostic artefacts from the battle using field walking or metal detector surveys
  3. Collect more of these artefacts until patterns become apparent within the assemblage
  4. Once a pattern or an area of interest become apparent attempt to understand the reason for this patterning by carrying out other types of survey. Analyse these patterns with a multidisciplinary array of archaeological techniques specifically designed to maximise the available data for the type of evidence already gathered
  5. Prove that the evidence is from the battle in question. For instance, people also shot arrows and fired lead balls from firearms during target practice or during hunting, resulting in small non-battlefield related assemblages. Human remains of those who engaged in combat provide good evidence. However, these are more often found individually in cemeteries rather than on battlefields and so they will be difficult to locate. Fragments of disarticulated human remains may indicate locations of former conflicts but may also indicate the presence of an unknown and ploughed out former cemetery. The early dating of any human remains discovered in a potential search area is therefore very important. Human teeth can remain in a good state of preservation and can be easily identified even when they have become disarticulated from the rest of the skeleton. The search for human teeth rather than human bone might aid the location of former mass graves.
  6. Once a battlefield has been identified re-check the data with historical sources to tie the two types of information together. It is possible that two battles were fought on the same site and the evidence obtained is not from the battlefield under investigation (e.g. at Towton there is potential evidence of three different conflicts).

© 2005 Tim L. Sutherland & Simon H. Sutherland