INTRODUCTION

Geophysical survey was undertaken at Vespasian's Camp, Amesbury, Wiltshire, during 7th-10th August 1995, in response to a request from the Monuments Protection Programme (MPP) of English Heritage. This Iron Age univallate hillfort occupies a defensive position on a prominent spur immediately above a meander of the River Avon west of Amesbury. The site is now divided into two unequal areas by the modern road leading from the village to Stonehenge (Stonehenge Road). Some 11.5 hectares of the fort lie to the north of the road and are densely forested whilst the remaining 3.5 hectares to the south have been developed within the last century and contain a number of residential houses, their gardens, and a small field. Although the majority of the hillfort is currently protected as a scheduled ancient monument (SM 10360), approximately half of the southern partition remains unprotected and it is this area that is currently under consideration. The geophysical survey was undertaken, therefore, in an attempt to assess the archaeological potential of this southern area.

The area surveyed, centred on SU 14 41, lies over Upper Chalk which is partially overlain by a drift deposit of Valley Gravel (British Geological Survey, 1950).

METHODS

In an attempt to maximise the information recovered from the site both magnetometer and resistivity surveys were carried out.

Magnetometer survey is a well established archaeological prospecting method and provides a rapid, non-invasive means of investigating many types of rural occupation sites (Clark 1990). The magnetometer is capable of detecting a wide range of archaeological features such as buried pits, ditches, gullies, kilns, ovens, and hearths.

Resistivity survey is a similarly well established technique also capable of identifying pits and ditches but is particularly well suited to the detection of sub-surface building foundations and other masonry features (see Note 1).

To complement the magnetometer survey, soil samples were retrieved along two transects of auger holes (see Fig 1) in order to measure their magnetic susceptibility (MS). MS is a natural attribute of soils which becomes artificially enhanced when those soils have been in contact with fire, as would be the case on an ancient occupation site (Tite 1972, Cole et al. 1995). It is the enhanced MS of soils infilling archaeological features such as pits and ditches, that allows them to be detectable with a magnetometer. Relatively high MS values can thus, in their own right, sometimes be an indicator of former occupation (and ancient industrial activity) and MS is also a valuable aid to the interpretation of detailed magnetometer survey.

Due to the awkward shapes of the areas available for survey two separate survey grids based on 30m squares were established, each on a roughly north-south alignment (see Fig 1). Each 30m square was then surveyed with a Geoscan RM15 resistivity meter using the Twin Electrode configuration. Measurements were taken at 1.0m intervals along traverses 1.0m apart and the data was periodically down-loaded to a microcomputer in the field. The resulting data is illustrated in this report using greyscale plots (see Fig 2). An abrupt change in background resistivity was encountered within the data (see below) and so the data has been statistically treated using a Wallis contrast enhancing filter (see Fig 2; Scollar et al 1990).

Each grid square was subsequently resurveyed using Geoscan FM36 fluxgate gradiometers. Measurements were recorded at 0.25m intervals along traverses 1.0m apart. Presentation of the data has been enhanced by the application of a local median filter to reduce the intense response to ferrous material (Scollar et al 1990). Greyscale and graphical trace plots of this data appear in Figure 3.

RESULTS

Resistivity Survey (Figure 2)

As is readily apparent in the plot of the raw resistivity data (Fig 2, plot 1) there is a marked change from high readings (pale greys on the plot) directly over the chalk bedrock (grid squares 1-9) to significantly lower readings (darker greys) over most of the remainder of the survey area, overlying gravel. This basic variation in response to the underlying geology is also apparent in the bi-modal distribution of the data (Fig 2 plot 1, histogram). The northern boundary of the drift deposit is most apparent in the data running eastward along the southern edge of grid square 7.

More clearly of archaeological significance is an arcuate high resistance linear anomaly enclosing a semi-circular area approximately 30m in diameter in grid squares 14, 15, 19 and 20. Whilst this might usually be interpreted as a wall or foundation, it is more probable in this case that it represents a cut or negative feature such as a ditch. That this is so is indicated by the magnetic response from the same feature (see below). The high resistance response presumably arises from a ditch with a coarsely textured or stony fill which is less moisture-retentive than the surrounding substrate. Whilst this feature may represent the outer perimeter of a former dwelling, it is larger than might be expected.

Evidence for other archaeological features is limited to the very perimeter of the survey area where this coincides, in part, with the hillfort defences. The rampart is clearly visible as a high resistance anomaly in grid square 16, flanked to either side by narrow bands of lower resistance. Parallel to these, within the hillfort, there is a narrow band of higher resistance values which may represent a revetment or a counterscarp feature. High readings along the extreme southern edge of grid squares 17-20 probably represent the inner foot of the rampart. The generally lower readings adjacent to this, forming a diffuse band within the hillfort perimeter, may represent a modest build-up of soil against the rampart.

Recent activity on the site must account for much of the variation in resistivity readings. Most distinct is the response to a water pipe running northwest-southeast through grid squares 3, 8 and 9 for which there is clear correlation with an intense magnetic anomaly (see below). To the east, a broad band of high resistance can be seen running west-east (through grid squares 22 and 26) which is a response to terracing thought to be associated with a former field boundary. Other more random variation in resistivity is due to recent landscaping, building and gardening activities. The location of an irrigated putting green can be seen as a discrete zone of low resistance in grid squares 23 and 24.

Magnetometer Survey (Figure 3)

The response of the magnetometer survey has been greatly affected by the presence of modern ferrous metal, particularly within the gardens to the east and north of the survey area. This disturbance is due to iron fencing (see the northern edge of squares 12-15), a modern ferrous pipe (characterised by its fluctuating positive and negative magnetic signature running through squares 3, 8 and 9), and miscellaneous buried litter. Landscaping and gardening must have also contributed, making the response in these areas particularly confused. The character and extent of modern interference is perhaps best demonstrated in Figure 3 (plot 2) where the sharp vertical deflections in the traces represent modern ferrous material.

Beyond these areas of disturbance the magnetic response is both weak and very uniform. Indeed, analysis of the frequency distribution of the data (see histogram on Fig 3) shows that the majority of the readings recorded lie within 1 nanotesla (nT) which is close to the maximum sensitivity of the instrument. Nevertheless, there are some weak magnetic variations likely to be of archaeological significance. In particular, the magnetometer has detected the semi-circular feature mapped by the resistivity survey (see above) as a very weak positive anomaly (1-2 nT).

The survey has also responded to the hillfort defences by detecting a discontinuous positive magnetic anomaly in square 16 which turns east to follow the southern edge of the survey area. Along this same edge there are also some weak and discontinuous linear anomalies which do not form any obvious pattern.

The results of the magnetic susceptibility measurements are presented in Table 1. In nearly every instance the auger was stopped at a depth of 20-30 cms, either by chalk, or by flint, the latter prevailing over the area mapped as Valley Gravel (see above). Values of topsoil MS were generally found to be low and consistent with the magnetometer response, the majority falling within the range 20 - 30 SI x10-8 m3 kg-1.

The topsoil samples from auger holes 1-9 were heated in a furnace (following Cole et al 1995) in an attempt to achieve their maximum magnetic enhancement. Their MS was measured before and after heating in order to calculate their fractional conversion (see Table 1). Samples 1-5 show a low and uniform level (averaging 3.6%) while samples 6-9 show a marked increase of approximately 6% again to fairly uniform values (averaging 8.9%). These results suggest that the soil samples had only achieved a small fraction of their potential enhancement in situ, the jump in the readings being due to the change in geological substrate. This lack of enhancement, allied to there being no major contrast between topsoil and subsoil MS, helps to explain the lack of variation in magnetic signal over the areas not already disturbed by modern magnetic interference.

CONCLUSIONS

The following tentative conclusions may be drawn. Only one archaeological feature has been detected with any certainty. This is the semi-circular arc visible in both sets of data in grid squares 14, 15, 19 and 20. Whilst it might possibly represent the site of an Iron Age dwelling, its size is more consistent with a Bronze Age barrow, other examples of which are believed to be present elsewhere within Vespasian's Camp. No record exists, to our knowledge, of any more recent structure in this area.

Elsewhere within the areas surveyed there is little or no geophysical evidence for buried archaeological features. The area least disturbed by recent interference, the field belonging to Sky House (grid squares 12-20), is characterised mostly by a subdued magnetic response interrupted only by a scatter of reactions to ferrous litter. The resistivity response, whilst slightly more lively, does not reveal any recognisable archaeological patterning. There could very well be concealed features here, but they are presently undetectable.

Beyond this area, nearer the houses and within the smaller gardens, the interference created by twentieth century activity has largely masked the geophysical response from any underlying features. The reported wartime military presence in parts of the south-western corner of the fort (Mr J Purvis pers comm) has left no obviously abiding traces, however.

Magnetometry and resistivity surveying are undoubtedly the prospection methods best suited to investigation of Iron Age hillforts. However, it should be stressed that the effectiveness of these techniques is much reduced when they are applied across a mosaic of small and irregular land parcels where there is already a high level of superficial recent disturbance. The interruption of the survey coverage, and of the instrument signals, may thus very easily mask or obscure the response to more subtle underlying archaeological patterning. It should be further stressed, as a fundamental tenet of any such survey, that a lack of an obvious response to buried archaeological features cannot be taken as evidence that such features do not exist. For instance, certain highly significant categories of archaeological features are usually too small or too poorly contrasted with their physical surroundings to be detectable: post-holes, stake-holes, cremations and inhumations being examples of such features. Larger features, too, may be indistinguishable should local soil or geological conditions be unsuitable.

Acknowledgements: we are very grateful to the various property owners who agreed to tolerate our survey activities in their gardens, and to Michael Pearce who kindly much facilitated the survey of land owned by Mr and Miss Purvis.

Notes

1. For a more detailed description of these techniques please refer to Clark (1990) or Scollar (1990).

References

Clark, A J 1990 Seeing Beneath the Soil: Prospecting Methods in Archaeology, Batsford, London.

Cole, M A, Linford N T, Payne, A W & Linford, P K 1995 Soil magnetic susceptibility measurements and their application to archaeological site investigation, in Beavis, J (ed), Science and Site: Evaluation and Conservation, Archaeological Sciences Conference 1993, London (in press).

Institute of Geological Sciences 1950 1" map Geological survey of Great Britain, Sheet 289, Salisbury - Drift.

Scollar, I et al 1990 Topics in Remote Sensing 2: Archaeological Prospecting and Remote Sensing, Cambridge.

Tite, M S 1972 The Influence of Geology on the Magnetic Susceptibility of Soils on Archaeological Sites, Archaeometry 14 , 229-236.

Surveyed by: M Cole, P Cottrell, A David, N Linford, A Payne
Date of survey: 7-10th August 1995
Reported by: M Cole
Date of report: 18th August 1995
Ancient Monuments Laboratory report number: 44/95

Return To:
On-Line Reports Index Database Home AML Home