In July 1990 aerial photographs revealed the parchmarks of two buildings (County Monument HWCM 10000) in a permanent pasture field at NGR SO 519458. One appeared to be a rectangular post-built structure of considerable size, as indicated by the marks of post-pads (east building), and the other an equally substantial building with a clearly defined ground plan consisting of several rooms (west building). In addition to these features there is a complex series of earthworks and cropmarks in the surrounding area representing the remains of an abandoned water management system and the 15th-century moated manor of Freens Court (HWCM 314), demolished in 1945, which occupied land now under cultivation to the north of the newly discovered parchmarks (Figure 1).

The parchmarks were equated by their finder (J Pickering) with the presumed site of a Mercian (8th century AD) royal palace associated with the murder of King (later St.) Ethelbert by King Offa in 794 which, according to local tradition, vague documentary references¹ and placename evidence² is thought to have been located somewhere in the parish. Typologically the buildings cannot easily be dated and their attribution as a Mercian palace is disputed. The size and form of the structures and quantity and quality of metal detector finds from the vicinity suggest a high status residence but the material recovered spans a wide date range in the Medieval period. Alternative interpretations of the site are that it represents an earlier precursor of the Medieval manor of Freens Court or subsidiary agricultural buildings such as barns contemporary with the former. However these buildings are not marked on early 18th C. maps³ depicting Freens Court and the finds are not compatible with a purely agricultural function. Despite this uncertainty concerning the early Medieval origins of the site its archaeological significance is not in question.

In April 1991, when this potentially important site became threatened by the activities of treasure hunters and uncontrolled digging, it became apparent that action was needed to secure its future protection. As a consequence the case was referred by the Archaeology Section of Hereford and Worcester County Council to English Heritage for consideration for scheduling. More accurate information was required from the ground in order to initiate monument protection measures, and the Ancient Monuments Laboratory was therefore commissioned to undertake a geophysical survey of the site, carried out in June 1991.

The site is located on the flood-plain of the River Lugg on the boundary of alluvial deposits and silty soils of the Bromyard Series developed on Lower Devonian reddish silty shale, siltstone and sandstone (Soil Survey of England and Wales 1983).

A detailed resistivity survey was carried out over the area encompassing the parchmarks discovered in 1990, and extending north to cover some of the topographical features interpreted as water management works that cover a substantial area of the site. There is clearly scope for further coverage of the larger surrounding archaeological landscape including the extensive earthwork complexes and the moated area containing the site of Freens Court. Unfortunately the latter was not accessible due to crop coverage, and although resistivity survey was chosen as the most suitable geophysical technique for investigating the area, (given the perceived nature of the archaeological features and the geology (see below)), it was not sufficiently quick to allow the full archaeological potential of the site to be explored. A pilot magnetometer survey was tried, over the area of the western building, but in common with other sites in the Marches the local geological conditions did not favour the use of this faster technique. It is probable that the mapping and interpretation of the archaeology surviving as earthworks would be better achieved by topographical survey.

Instrument readings were based on a grid of 30m squares, measured in to the field boundaries (see Figure 2).

An initial survey was carried out with a reading interval of 1m along successive 30m long traverses placed 1m apart. The results obtained were encouraging and the decision was made to conduct a further higher resolution survey over the most striking anomalies, with a reading interval of 0.5m along the traverses and 0.5m separation between traverses. In both cases a Geoscan RM15 Resistivity Meter was employed, which was connected in the Twin Electrode configuration with a mobile probe separation of 0.5m. The data was logged in the internal memory of the RM15 and periodically transferred in the field to diskette on a portable micro-computer for storage and data verification. The resulting reconstructed data has been interpolated and is presented here, in both raw and enhanced form, as greytone plots (Figures 3, 4, 7 and 8). For comparison, the 1m data has also been plotted in colour (Figure 5) and as graphical (X-Y) traces (Figure 6). A more detailed list of the various plots included and the methods of image enhancement used is provided in a later section of the report.

A single 30x30m square was surveyed (see Figure 2) with a Geoscan FM36 fluxgate gradiometer, which recorded the vertical magnetic field gradient at intervals of 25cm along successive parallel 30m long traverses placed 1m apart. Readings were recorded at 0.1 nanotesla (nT) sensitivity and the traverses were orientated approximately north-south. The data was captured in the internal memory of the FM36 and processed in the field in a similar manner as the resistivity data. The resulting raw data, after reassembly in the laboratory, is presented here in the form of greytone and X-Y traces (see Figure 9 and later section on plots enclosed).

The topography of the site has obviously been much modified in the past by human activity and this influence is clearly marked in the raw data by areas of contrasting resistance. The higher ground where the parchmarks occur is characterised by correspondingly higher background resistance, whilst the surrounding areas of damp and low lying flood-plain have very low resistance values. The results suggest that the boundary between these areas is artificial. A system of earthwork embankments and former water channels has been clearly resolved north of this division in the coarse (1m) survey. The survey evidence is not sufficient, however, to determine the relationship of topography to the parchmarks but there is some repetition of alignments. All the features could have been constructed simultaneously or alternatively the water system could have been planned at a later date around a pre-existing settlement site. A former pond and possibly part of two sides of the moat of Freens Court have also been located.

The features responsible for the parchmarks have been resolved with exceptional clarity by both surveys. This is not altogether surprising as the parchmarks suggested that the sub-surface features lie at a shallow depth and that they are substantial and well preserved. Nevertheless the image provided by the survey is a striking demonstration of the effectiveness of this survey method when used in ideal conditions. The patterns of high resistance are indicative of stone wall footings in the case of the western building and stone supports for a timber-framed east building. The results show the exact location of the two buildings and reveal detail of their ground-plans, size and layout, thus significantly enhancing the information available from the AP records.

The survey results confirm the general description of the west building as derived from the AP evidence:

"A stone walled building of linear type with a frontage about 100m long. In the centre is one main hall, with a room across the full width of it at the north end. To the north of this are two similarly paired rooms, but only about two thirds of the width. To the south of the hall is another on the same frontage, narrower and on a slightly different alignment. There may in addition be slighter structures on the east side of this range"

(Hilary White 1991)

The geophysical evidence does however show that it is slightly smaller than previously estimated (75m from north to south rather than 100m) and additional features have been detected to the north and north-west of the main building range. The latter consist of an open-ended irregular structure, not obvious on the APs, on the west side of the main building range, as well as further less well defined linear features - again perhaps walls - in the north-west of the complex. These additional features significantly increase the size and complexity of an already substantial building.

The second post-built structure in the south-east of the survey area, has a long axis of 36m and a width of around 10m between two parallel rows of post-pads. The post-footings appear to be placed at intervals of 3-4m along the sides. The interior of the building gave higher readings in comparison to the outside, suggesting the presence of a resistive surface such as a floor or layer of collapsed building materials. The ends of the building are not as clear as the sides, but there is perhaps a hint of a semi-circular shape at the north-west end and also a suggestion that the southern corner of the building may have been disturbed by a later pond.

A further L-shaped group of high resistance anomalies hint at the presence of ?walls linking the two buildings and forming two sides of a sub-rectangular enclosure, bounded on the west by the east frontage and long axis of the west building.

Although it was centred over the position of the west structure (as shown by the resistivity), the magnetometry provided no indication of the presence of a building. The reason for this is a lack of contrast between the magnetic susceptibility of the soil and the structural remains it surrounds - not unusual for the area. More surprising however is the lack of evidence for industrial and domestic thermally magnetised features such as hearths, ovens and furnaces associated with the building and thus there are no clues to the types of activities being carried out on the site. A few very localised strong magnetic anomalies are compatible with the presence of small iron objects - potentially artefacts associated with the occupation of the building but also equally likely to be of more recent origin.

The resistivity survey has provided exceptionally clear definition of the extent and layout of the structural remains previously seen on APs. The survey has also identified further associated structures not readily apparent from the APs. In summary the results show a substantial and potentially well preserved building complex probably constructed on stone foundations linked (perhaps by a walled yard) to another large but simpler building indicated by a rectangular pattern of stone-settings to support timber posts. The additional 0.5m survey provided slightly sharper definition of some wall alignments, but no real increase in the information contained by the 1.0m reading interval survey. In contrast to the resistivity, negative results were obtained from a trial magnetometer survey over the position of the main building. On the basis of the resistivity results, detailed description of the site is now possible, and this has enabled the initiation of accurate measures to secure the long term preservation of this potentially very important site.

Leland, John. 1910. Itinerary in England & Wales, ed. Lucy Toulmin Smith, vol III (London)

Linington, R. E. 1970. A first use of linear filtering techniques on archaeological prospecting Prospezioni Archeologiche 5 43-58

Linington, R. E. 1971. Further tests on non-symmetrical filtering systems. Prospezion Archeologiche 6, 9-20.

Scollar et al, 1986 Display of archaeological magnetic data, Geophysics, Vol. 51, No. 3, 623-633

Scollar et al, 1990 Archaeological prospecting and remote sensing, Topics in remote sensing 2, Cambridge.

Sheppard, J. n.d. The origins and evolution of field and settlement patterns in the Herefordshire manor of Marden. Queen Mary College, University of London, Occasional Paper 15.

Soil Survey of England and Wales, 1983 Soils of England and Wales, 1:250,000 Soil Map Series, Sheet 2 Wales.

White, H A, 1991 HWCM 10000 : A suggested Mercian palace at Sutton, Herefordshire. Unpublished assessment report, Hereford & Worcester County Council Archaeology Section.