South Cadbury Environs Project

Woolston Manor Farm

Woolston Manor Farm is centred on a basin-shaped combe 1 km to the east of North Cadbury, Somerset. The modern farm buildings and former farmhouse nestle on a rise some 200 to 300m north of a tributary flowing westwards into the Cam. The farmland comprises slopes of every aspect formed from sandy beds, parenting light soils well suited to cultivation.

Map showing Woolston Manor Farm in both 1725 and 2006

Woolston Manor Farm map in 1725 and in 2006

View Larger Image

Plan showing Gradiometry results at Woolston Manor Farm 2006-7

Gradiometry at Woolston Manor Farm, 2006-07

View Larger Image

The very distinct outline of the modern farm-holding is mirrored by a map from 1725 and suggests a much earlier estate boundary. Within the outline many of the field boundaries and names are little changed in nearly three centuries. A further indication of conservative regimes in the past are the fine surviving earthworks in two fields. Trinities, a three hectare field in the south west of the farm, has been protected by scheduling since the identification of the earthworks from 1947 air photographs. Earthworks in Great Cowleaze appear to have gone unnoticed by archaeologists.

Until 2005 the project aimed to cover a total of approximately 20% of each sampling locality with the three main techniques. However, there was a deliberate change of policy at Woolston where our aim was to cover all accessible land on the farm with gradiometry and regular test pits (Tabor 2008b ). This revision was made because the farm presented a rare opportunity to explore a succession of agricultural systems which have served a central habitative area for two millennia or more.

Geophysical survey covered in excess of 59 hectares, fieldwalking covered 17 hectares and 51 regular and a total of 59.5 square metres of targetted test pits have been excavated at Woolston Manor Farm. In addition, three small trenches were opened for a University of Bristol training excavation on the Plain of Slait.

The results of the gradiometry show a prevailing trend of positive linear anomalies on an approximately west south west to east north east orientation apparently in a co-axial relationship with west north west to east south east linears. On closer examination of the plots it is clear that there are several subtle variations within the prevailing orientation, some of which cannot be contemporary judging by the character of their intersections and overlaps. Comparatively few linears are wholly at odds with the prevailing scheme. The following analysis attempts first to identify coherent boundary systems before offering hypothetical dates for them in the light of the evidence from test pitting and excavation. To avoid duplications of maps the anomalies are presented in the hypothesised sequence.

Distribution Map showing Early Bronze Age features at Woolston Manor Farm

Early Bronze Age features

View Larger Image

Two targetted test pits on the plateau recovered six small grog-tempered sherds, two of beaker type, from the upper fills of truncated V-profiled ditches making up the east and south of an approximately 15m x 15m enclosure. The only other finds were three flint flakes. The fill matrix of very red sandy silts suggests a lack of organic material and is consistent with an early date. The enclosure has a small annexe on its east side, a form most closely parallelled by an example at Milsoms Corner, South Cadbury.

Distribution Map showing Late Bronze Age features at Woolston Manor Farm

Middle Bronze Age features

View Larger Image

The Middle Bronze Age phase is defined by a system recognised exclusively on the plateau in Plain of Slait and Lady Field 1. It clearly derives from the previous system and the long narrow parallel boundary pattern is similar to that dated as Early Bronze Age at Sigwells, Charlton Horethorne. A 2m x 1m test pit targeting an area of enhanced magnetism bounded within a space of 16m x 16m produced 20 sherds, most from a single biconical vessel of probable middle Bronze Age date, from an organically rich fill formed on the abandonment of a depressed floor area.

Distribution Map showing Late Bronze Age features at Woolston Manor Farm

Later Bronze Age features

View Larger Image

The very thin evidence for the dating of a complete re-orientation of boundaries in the Late Bronze Age is by analogy with a pattern observed around Cadbury Castle, most particularly at Milsoms Corner. There, the fills in the two surviving ditches of a south to north oriented rectangular enclosure have been firmly dated to the Middle and Late Bronze Age (Tabor 2008, figure 40).

The most surprising but best attested discoveries were two 6 m wide zigzag parallel anomalies which form part of a system including an open curvilnear, partly obscured by a pipeline. The latter is respected by a much smaller ditch forming the southern boundary of the Middle Iron Age system. The sole targetted test pit investigated the southern edge of the south zigzag ditch. The character of the hillwash sealing the ditch fills suggest that the upper part of the feature has been ploughed away. The fills suggest that the ditch had filled slowly and included flint, grog and shell-tempered, cordoned Bronze Age pottery and a hammerstone. Given the lack of later finds it is likely that the pottery date reflects the period of the feature's use.

The gap between the zigzag and the curvilinear is probably the entrance to an enclosure, presumably following the contour to take in the ridge which extends for another 400m west of Oat Croft. The substantial scale of the ditches suggests that Cadbury Castle was not the first local hilltop to be enclosed by earthworks.

A regular test pit in Little Eldridge recovered two calcite and shell tempered sherds which are sufficiently diagnostic to indicate activity of the period. They were associated with a subsoil formed directly over the natural and were not in the fill of a cut feature but the test pit was immediately adjacent to one of several continuous and intermittent linear anomalies running along the bottom of the valley, reflecting its use as a boundary marker over a long period. To the south west, close to the bottom of the valley's north west side, two of three sherds from a buried soil sealing another feature were flint-tempered.

A rich Iron Age assemblage from a targetted test pit in Great Cowleaze included Early Iron Age jar and bowl types (JB2 and BA2).

Distribution Map showing Middle to Late Iron Age Features features at Woolston Manor Farm

Middle to Late Iron Age features

View Larger Image

The targetted test pit in Great Cowleaze included a total of 160 sherds with shell mixture inclusions. The range of forms included jars (JC2 and JC3), saucepan pots (PB1) and bowls (BC3.3 and BD6) of the Middle to Late Iron Age, amongst them a fine example of the South West decorated style. They derived from a ditch fill, occupation horizons and a dark hillwash immediately below the topsoil, forming over a platform cut into an early hillwash. Much smaller comparable assemblages came from other test pits in the field.

The sheer density of high magnetic anomalies obscures the coherence of the systems they represent. The earthworks were recorded in a plane table survey but it did little to clarify the phasing as there are platforms ranging in date from the Iron Age to the 14th century AD.

Another field notable for Middle Iron activity was Little Eldridge. A regular test pit produced 12 sherds tempered with shell mixtures from a rapidly-formed rubbly hillwash which almost certainly derived from a quarry at the top of the valley side. The sherds have only a coincidental relationship with a much earlier ditch.

By analogy with the magnetic signature from the north west of Sigwells, a looser but extensive pit group, lying in a slight saddle extending across Card's Piece and Plain of Slait, probably started to form during this period. A third dense group of pits was identified at Hicknoll Slait. All three groups share a topographical preference for cutting off a steep-sided spur from a large plateau. All three spurs overlook Cadbury Castle but the views from the pit groups at Woolston and Hicknoll Slait are obscured by the slight saddles.

Plan showing Middle to Late Iron Age pit groups overloking Cadbury Castle

Middle to Late Iron Age pit groups overlooking Cadbury Castle

View Larger Image

A rectangular ditched enclosure encroaches slightly on the south east side of the pit group and may first have been dug in the Iron Age. The inception of a sequence of square enclosures linked to the west of the Sigwells pits was in the Middle Iron Age.

Distribution Map showing Late Iron Age/Romano-British features at Woolston Manor Farm

Late Iron Age to Early Romano-British features

View Larger Image

The pit group at Sigwells grew most quickly during the later first century BC and the first half of the first century AD, when there appears to have been an abrupt cessation of activity. The period of activity in the Woolston valley is broadly similar.

A track in the southwest, in Oat Croft, shares its orientation with the boundaries along the south facing slopes of the Woolston valley. Although only a slight trend in the gradiometer data, a regular test pit showed at least two sustained phases of use for what in effect became a ridgeway in and out of the settlement. It takes no account of any hilltop enclosure so the Late Bronze Age earthworks must already have been in an advanced state of decay, at least in places.

The western most targetted test pit in Great Cowleaze revealed a Late Iron Age to Romano British ditch sequence, the earlier having been filled before being replaced by the latter. The ditches appear have bounded one side of a track which serviced the higher ground. A targetted test pit 30 m to its east offered a longer sequence cut by a slightly later track. An Iron Age ditch was deliberately backfilled and subsequently re-cut and backfilled (covering a deliberately deposited Durotrigan bowl) in the 1st century AD. This was re-cut by a Romano-British ditch with a slightly different orientation, providing the basis for dating similarly oriented boundaries elsewhere in the locality.

Distribution Map showing Middle Romano-British features at Woolston Manor Farm

Middle Romano-British features

View Larger Image

A large assemblage of second to fourth century AD pottery from a targetted test pit in Rye Close provided a firm foundation for dating this extensive remodelling of the previous system. The pottery was recovered from two ditches, one of which had been re-cut, and a sealing black abandonment horizon into which semi-articulated groups of roofing slates had fallen, presumably through sliding off a decaying roof.

The density of finds from Rye Close is a clear sign of habitative settlement which can be traced at least 140m further south east through pottery in a ditch found in a regular test pit in Quar Close.

Distribution Map showing Late/Post Roman features at Woolston Manor Farm

Post Roman features

View Larger Image

Distribution Map showing Early Medieval features at Woolston Manor Farm

Early Medieval features

View Larger Image

There are no finds dated to the first five centuries following the Roman occupation. However, the gradiometer and earthworks surveys, coupled with test pit data, reveal some key horizontal stratigraphic relationships. It is clear that the initial rectangular scheme of the plateau enclosure was altered by an arcing double-ditched track but that the greater part of the enclosure remained in use and was almost certainly re-cut after the track was made. This is demonstrated by the much greater magnetic enhancement of the enclosure ditch readings on the north west side. Given the enclosure’s Late Iron Age to Roman span we can infer that the origins of the track are either later Roman or soon enough after that period for the enclosure to be an observable, even functional, feature in the landscape.

There can be no reasonable doubt that to its east two double ditched tracks converge in a droveway leading north from the settlement. To the south, in Great Cowleaze, it can be seen that there is a direct continuation of one of the tracks. The use of the latter during the Roman period, and the re-cutting of at least its east ditch in the 13th or 14th century, are a clear sign of its persisting significance and the continuity of habitative settlement in the core area.

In Card's Piece the system takes account of the natural contour, having a slight curve away from an orientation of origin based on linears extending south from Rye Close.

A strong magnetic response occurred in the lee of a short length of low bank with a slightly different orientation from the Iron Age and Romano-British systems. A targetted test pit produced sherds of Late Saxon and Saxo-Norman pottery high in hillwashes sealing a deposit of carbonised grain, lying over a waterlogged Romano-British abandonment horizon. The deposit may be Early Mediaeval.

Photograph showing Carbonised grain deposit

Carbonised grain deposit sandwiched by hillwashes including Romano-British (lower) and Late Saxon and Saxo-Norman (higher) pottery

View Larger Image

Distribution Map showing Medieval features at Woolston Manor Farm

Medieval features

View Larger Image

The modern road serving Woolston enters it from the west through a deep holloway, yet it is quite clear that it is not an ancient route as it slices acutely across all the pre-mediaeval systems. Even now the layout of fields north of it owe more to their Bronze Age ancestry than to the 'new' route. The Little Eldridge valley continued to form a natural boundary but the ditches in it were re-oriented to become perpendicular to the new road.

Three targetted test pits in Great Cowleaze provided pottery ranging from Late Saxon to the 14th century AD. Most notably one revealed that the east boundary of the main track leading to the higher ground was re-cut by a large Medieval ditch. The lower part of the ditch appeared to have stabilised after initial silting which included 13th century pottery. The middle fills included most of an articulated bovine skeleton. The extremely poor condition of the bone suggested that the body had remained exposed in the ditch for a considerable period, probably at a time when the settlement had been abandoned. Subsequently, a very rapid fill included over 800 14th century pot sherds, many glazed and decorated including near complete Ham Green vessels, in turn sealed by slower-forming silts. Two small stoney features may relate to subsequent late Medieval activity. It seems likely that this was an area of settlement abandoned rapidly following a period of intense stress (possibly the Black Death) which had been systematically cleared at some time after the original crisis. The presence of a fine fluted, handled, 13th century mortar fashioned from Portland Stone implies the proximity of a high status or religious building.

A test pit targetted the top of the southern edge of a platform where several linear geophysical anomalies converged. A richly organic Romano-British abandonment horizon was cut by a cess pit of probable Medieval date, in turn cut by the trench for very substantial footings. Despite their substance it seems likely that these foundations are Medieval although their extent is not known as the structure did not show as a geophysical anomaly.

Trinities, as a Scheduled Ancient Monument, was known to contain the surviving part of a Medieval landscape. Permission was granted for geophysical and earthworks surveys, and for the excavation of three 1 m square regular test pits and three 1 x 2 m test pits to be targetted on the basis of the geophysical results.

The survey shows that the system was set out as a single conception, perpendicular to the modern road, with only slight subsequent modifications to access. It is likely that there were several buildings on the platforms but only one, to the west, has a substantial area of enhanced readings within it. A targetted test pit exposed two small ditches on the north east edge immediately to its east. They appear to have silted up slowly until deliberately levelled upwards with organically rich brown soil which included a moderate amount of structural ceramic lumps and several large lumps of slag. It is very probable that the associated structure was used for metalworking. A targetted test pit approximately 15 m further north east established that a ditch defining the east boundary of a track was probably of 12th to 13th century AD date.

Trinities' earthworks include part of a rectangular feature in the north east of the field which butts onto the modern stream and was surely an earlier mill pond. The situation has long been the best suited to that purpose and the mill mentioned in the Domesday Book was surely close by.


The south facing slopes and light soils of the north part of Woolston Manor Farm may have attracted Early Bronze Age farmers who were introducing a larger arable component to their subsistence strategies, although traces of long, narrow, boundary systems are more often interpreted as ranch-like, hence for animal husbandry. The thin spread of pottery over varied topography implies varied settlement choice and a mixed subsistence pattern, although it should be noted that pottery on the east side of Lady Field 2 might well derive from ploughed out barrows.

The small Middle Bronze Age enclosure structure in Lady Field 1 surely contained a fairly substantial dwelling in an exposed part of the landscape. As a magnetic anomaly it looks very similar to the earlier example in Card’s Piece so there may well have been habitation on the plateau, perhaps used by people tending to flocks.

The south to north re-orientation of land division at Milsoms Corner appears to have been part of a process of creating more, purpose specific, enclosures, possibly as the acquisition of favourable land became more difficult. It is worthy of note that in all three areas where the pattern was adopted it succumbed soon enough for succeeding schemes to be based on those of the Earlier Bronze Age, although the process was a great deal slower, and of a different character, at Sigwells. By the opening centuries of the first millennium BC the distribution of pottery within the sample localities has shrunk to on or close to three key hilltops: Cadbury Castle, Poyntington and the Oat Croft ridge.

There may have been a struggle for ascendancy between three communities, although any conflicts were surely resolved by the later Middle Iron Age when Woolston, in particular, appears to have become a thriving community able to farm substantial tracts of land, much probably under rotational ploughing. This might reasonably be regarded as the developed landscape which complemented the contemporary developed hillfort. It was a landscape designed to grow crops and enclose animals, to collect and re-distribute their manure, and to reduce the impact of soil movement.

As noted at Sigwells and South Cadbury there seems to be a distinct rupture, even an hiatus, towards the Middle of the first century AD before existing boundaries were re-engraved on a subtly different orientation. The landscape was structured much as before, and the processes of production were surely similar. At Woolston there are some traces of a southward expansion. A far greater change is observed in the highly speculative Post Roman landscape. Many enclosures appear not to have been maintained, whilst the expansion of tracks and introduction of a droveway in the north, creating a route to and from the settlement, may indicate at once a change in affiliation to other settlements and a proportional increase in the importance of animal husbandry.

However, we should note the physical evidence of cereal production during the early part of the period in Great Cowleaze. Subsequent deep charcoal flecked occupation deposits imply the continuation of a busy and successful settlement during a time which is widely regarded as very unstable when viewed from the prevailing perspective of urban decay.

By the 12th century AD craftsmen and their families appear to have had an allotted habitation and work space to the south east of the core settlement, on terraced slopes over looking the stream. So far we can identify the Smiths and the Millers! Within two centuries the industrial area had been abandoned, probably at the same time as parts of the core in Great Cowleaze, where an abundance of glazed and decorated pottery illustrates the material and geographical division between classes. The settlement was revived soon afterwards but the manor house was probably re-sited to the south of the modern road, where the eighteenth century manor house now stands.

The core habitation area on the lower south-facing sides of the basin, within a strip of approximately 400m from west to east and 100m from south to north, had remained such for over 2000 years by the time the present farmer sold his house to a commuting solicitor in the 1990s. Indeed, it only expanded westwards beyond an original 250m long strip after the abandonment of the Great Cowleaze platforms in the 14th century AD

Go to top of page