LOWER FARM, NUNEHAM COURTENAY, OXON.

Interim report on the continuing geophysical survey.

INTRODUCTION

Magnetometer survey was first undertaken at Lower Farm during the summer of 1991 (Bartlett 1991). This survey was commissioned following the discovery of large quantities of Roman pottery by the Oxford Archaeological Unit (OAU) during the laying of a Thames Water pipeline. Subsequent investigations by the OAU within the pipeline easement revealed evidence of a substantial industrial site although no actual kilns were found (Booth et al 1994). Only a very limited area was accessible for the magnetometer survey but, although no anomalies likely to represent kilns were located, the results did suggest that a survey of the surrounding area would be profitable.

Due to the potential importance of the site, more extensive magnetometer survey was undertaken by the Ancient Monuments Laboratory (AML) during the summer of 1992 (area A on Fig 1; Cole 1992). The results of this work amply demonstrated the suitability of the underlying Jurassic Cornbrash to magnetometer survey. Not only were a number of Roman kilns, pits and ditches detected, but evidence of both earlier (prehistoric) and later (medieval) features was also located. The clarity of the response was such that integration of the results of this survey with those from the excavations has offered perhaps the most detailed plan of an Oxfordshire Roman kiln site to date (Keevill & Cole 1995).

Both of the initial magnetometer surveys were conducted within a block of surviving medieval pasture. Given that the site was under consideration for scheduling, a clear priority was that a measure of its full extent be obtained. The results suggested that Roman activity might continue into the arable fields to the east. The survey was extended again, therefore, in November 1994 to investigate the surrounding fields and amplify the results of a programme of field walking (undertaken by the OAU) during 1995. Additionally, the survey offered an opportunity to assess any contrast in preservation which might be visible in the data between the pasture and surrounding arable fields.

Magnetometer data was collected from three areas (B, C and D on Fig 1; centred on SP 53 00, SP 53 00 and SP 53 00 respectively) each of which overly the same substrate of Jurassic Cornbrash.

METHOD

Three separate grids of 30 x 30m squares were laid out, each oriented so as to best fit the field boundaries of that particular field (see Fig 1). Each square was then surveyed using Geoscan FM36 fluxgate gradiometers. Measurements were recorded at 0.25m intervals along traverses spaced 1.0m apart. The resultant data is presented in this report in the form of greyscale and graphical trace plots (see Figs 2, 3, 4, & 5). Presentation of this data has been enhanced by the application of a local median filter to reduce the intense response to ferrous material (Scollar et al 1990).

RESULTS

A summary plan of all of the AML magnetometer data (which now encompasses an area of just under 10ha, including the 1992 survey) is presented in Figure 2. From the latter it is evident that archaeological activity continues well beyond that mapped by the 1992 survey (area A on Fig 1) and indeed clearly must extend beyond the limits of the 1994 survey area.

The majority of the activity in the 1994 extension of the survey is concentrated within area C, clearly reflecting the pottery distribution mapped by the OAU (G Keevill pers comm). A pattern of enclosures can be seen lying either side of a central west-east road or trackway which may be a continuation of the trackway located in 1992 running along the eastern edge of area A. There is also the suggestion of a fork in this trackway at the eastern edge of area C (grid square 13). Other interesting features include a large sub-rectangular (?IA) enclosure (in grid squares 20, 23 & 24) which has an obvious entrance to the east (see note 1).

A number of additional kiln-like anomalies have been detected (all within area C), apparently focused into as many as 6 discrete groups. These anomalies all share the same characteristic form and are a similar strength as those identified as kilns in area A which were subsequently confirmed as such by augering (Cole 1992). The response of the magnetometer to ditches is exaggerated in the immediate vicinity of the suspected kilns presumably due to their local in-filling with soil which has been strongly enhanced magnetically as a result of its association with the industrial activity. A component of this increased magnetic response may also be due to an accumulation of ceramic debris within the ditches.

Significantly, the response to buried archaeological features can be seen to be of a similar magnitude within the area C (which has been under prolonged arable cultivation) to that recorded in the unploughed pasture field to the west (area A). In comparison to area A, the response to ridge and furrow in area C is much subdued thereby improving the clarity of the response to the underlying features. However, faint traces of ridge and furrow can still be made out to the centre of area C, despite it having been completely levelled by cultivation.

Ring Ditches

A number of annular and penannular anomalies have also been detected in addition to those already located in area A. The magnetic response to these features varies widely. Two (in area C; grid squares 23 & 27) have produced clearly defined anomalies up to 20 nanotesla (nT) in strength while 3 others in the southern third of area C are only just discernable at 1-2nT. Whilst all of the ring ditches visible in areas A and C are between 10 and 20m in diameter, there is a very subtle circular anomaly in area D (squares 2, 3, 6, & 7) which, if complete, would be approximately 30m in diameter.

The circuit of the more northerly of the two ring ditches located in grid squares 23 and 27 is interrupted to the east suggesting a possible "entrance", an interpretation to which further weight is provided by the more pronounced response to the ditch at the terminals to either side. None of the other ring ditches appear to have similar gaps although a number are too indistinct for this to be certain. The visibility of some is further confused by other anomalies relating to a different, presumably later, period. There is evidence of internal features within some of the ring ditches (particularly those in squares 23 and 27). Some caution must be voiced with regard to this interpretation, however, given the multi-phased/period activity at the site.

In area B three incomplete sub-circular features have been detected clearly as anomalies up to 8nT in strength. This clarity suggests that, rather than being a partial representation of more complete ring ditches, these anomalies do indeed reflect their true extent. As the open sides of all of these features face eastward, a possible interpretation might be that they represent wind-breaks against the prevailing westerly winds (see note 2).

Adjacent to the latter, and slightly to the west, well defined, but irregularly shaped areas of magnetic enhancement have been detected. These may well represent the response to shallow quarrying activity, possibly for clay for use in pottery production.

Surprisingly, despite the high degree of detail that has been recorded by the survey, there is no obvious evidence for any buildings or foundations within the entire survey area. Whilst the magnetometer will not necessarily detect the presence of buried walls or foundations, the excellent response afforded by the underlying Jurassic geology does offer the optimum conditions for their detection (see for example the survey of Roman buildings over Jurassic limestone at Kingscote, Gloucs. - Cottrell & Payne 1993). This suggests, therefore, that if there are Roman buildings at the site they may be fairly insubstantial (perhaps timber built; see note 3) or simply have not been encountered thus far.

The survey has been affected by modern ferrous material in places. This is most obvious where a service pipe cuts southeast-northwest through area C, and the particularly strong disturbance created by the electricity pylon situated just to the east of the pipeline corridor and between areas A, C and D (see Fig 2).

CONCLUSION

The naturally high iron oxide content of soils derived from Jurassic limestone has provided ideal conditions for highly informative magnetometer results. A detailed plan of a major Roman kiln site has been revealed in great clarity. The distinct pattern of enclosures and trackways mapped by the initial survey in 1992 can now be seen to be only a part of a far more extensive system. Detail of many more kiln sites has been located, revealing an apparent focusing into as many as 6 discrete groups within the field to the east. A number of further ring ditches, of various morphologies, have also been mapped underlining the multi-period nature of the activity at the site.

Analysis of these results, combined with those from field walking, indicate that the Roman activity continues well beyond the eastern limit of the survey. Further survey work, allied to more comprehensive interpretation, and publication is anticipated.

Notes

1. Iron Age pottery forms a small component of the overall pottery assemblage for the sites so an attribution to this period is not unfeasible.

On-Line Reports Index Database Home AML Home