Marshfield – An archaeological survey of a southern cotswold parish

By V. Russett


This survey is part of the Marshfield Project set up by Avon County Planning Department in 1982 to examine the development of a southern Cotswolds parish from its earliest use by man up to the present. The other main element of the Project was the excavation of an Iron-Age and Roman settlement at Ironmongers Piece. That excavation will be published in the British Archaeological Reports British Series.

Some 13 people worked on the survey, though not all at the same time, over a 2-year period. They were supervised by Vince Russett who undertook the documentary research prior to starting the survey. The survey team was established as part of the ACCES scheme sponsored by Avon County Council and financed by the government through the Manpower Services Commission. ACCES (Avon County Community Environment Scheme) is a scheme concerned with providing practical work for unemployed adults on a variety of conservation and environmental improvement projects to the benefit of local communities and the County as a whole.

The general purpose of the survey was to see how an area of Cotswold landscape had been used and modified by man in the past, and not just to look for new ‘sites’. This is the first time that such a systematic fieldwork project has been undertaken in this region. Marshfield parish was chosen because it has a large amount of arable (20% of the parish), where finds collecting could be undertaken, and pasture land where it was already known there is a good survival of earthwork remains. Another aspect of this detailed survey was to assess both the various methods of carrying out fieldwork, and what data should be obtainable, particularly for the later prehistoric and early medieval periods.

An important part of the Marshfield Project was to involve local people from Marshfield and elsewhere in the County of Avon. This survey could not have been undertaken without the goodwill and assistance of local people in various ways. As far as possible both the results and methods of the Project have been made available in presenting exhibitions, open days and information booklets. Indeed this publication is intended not only as an academic record, but also one which will be of interest to all members of the local community who might wish to gain a deeper knowledge and enjoyment of the history of their local area. For that reason the detailed information about archaeological finds is published not in the main account, but in microfiche in the envelope in the back pocket.

R. lies


Marshfield is a south Cotswold parish of about 5,500 acres (2,230 ha.). It lies in the centre of the eastern edge of the county of Avon, though formerly its south-eastern corner was the meeting point of the counties of Somerset, Gloucestershire and Wiltshire. (FIG 1)

Its position twelve miles from the centre of Bristol and five miles from Bath makes Marshfield a very desirable commuter village, but its population has remained around the 1100- 1300 mark for the past 200 years, as the numbers living in the countryside have dwindled, but those living in the village have increased in number.

The core of the medieval town, that is High Street, Market Place and Hay Street is a Conservation area. The main street consists of 17th and 18th century house fronts (many of which conceal older cores) making it one the most interest (and in certain wind directions, colder) of villages in the district.

What the casual visitor does not see, however, is the spread of farmsteads throughout the parish, some of which lie on sites that have been occupied since Roman times or before. (FIG 2) Most of these are mentioned in the text.

The countryside around this village might be a model for the typical English parish of the last quarter of the twentieth century, with its varied and intensive farming, and many different landscapes, and it was this very variety that led to the choice of Marshfield for this survey. The questions we set out to answer were:

  • Can we discover all the archaeological sites in the parish?
  • Can we explain them?
  • Can we understand how the landscape of Marshfield has developed, and what the factors were that gave rise to the varied and beautiful landscapes of this Cotswold parish?


The map entitled ‘Physiography’ shows the contours of the parish (in metres) and the streams in the parish. (FIG 3) Most of the Cotswold plateau is poorly watered, and this, along with the thin, easily drained soils, make it very suited to modern arable farming.

The parish divides neatly into six zones or physiographic units, which all have sets of common factors in their make-up, the topography, geology and soils being strongly inter-related. (FIG 4)

1. The Oldfield plateau. This is a broad, largely flat or slightly rolling area in the north-west of the parish, mostly enclosed in large fields with stone walls, poorly wooded and watered, and the small valleys that do exist have minute streams in them, which swell greatly in wet weather.

2. The valley of the Broadmead Brook. This 40m deep valley lies between the two large flatter areas of the north of the parish. It is almost entirely pasture land, and is wetter than the uplands with generally clayey soils. Strip lynchets all along the valley show that it was nevertheless ploughed; there is little woodland in the western part, although Harcombe Wood lies just outside the parish to the north, and Shirehill Wood lies in the eastern part. The fields are mostly smaller and hedged. The water of the Broadmead Brook itself has been important in the past, both for watering animals, and as a source of power for a mill.

3. The broad flat Marshfield plateau, a typical dry Cotswold arable area, with large walled fields, very few trees and a general ‘open’ appearance. Much of Marshfields arable capacity lies in this unit.

4. The valley of the Doncombe Brook. This is mostly heavily wooded, the steep slopes being unusable for agriculture. The woodlands are a mixture of the remains of ancient semi -natural hardwood forest, and planted conifers.

The Ashwicke – Fuddlebrook plateau, flattish dry country, but with small parkland, pasture and woodland elements. The fields are
mostly large in size and hedged.

The valley of the St Catherines Brook. This is a steep sided feature, now entirely pastoral, with a little
surviving semi-natural woodland. The slopes are of clayey to silty soils, very prone to slipping, and all fields are hedged. The Brook itself is a
source of water, and once drove three mills in 5 km, all of which are now in ruins.


Marshfields solid geology is very strongly linked to its soils and the dissected nature of”its southern part, and the hummocky valley sides, are entirely due to the underlying Jurassic rocks.

The plan shows the layout of the geology of the parish. (FIG 5)

The Great Oolite Limestone, the most extensive formation in the area, is a
hard white shelly limestone, occasionally used for building stone in areas where its quality is high enough (it is probable that the quarry at The Rocks Home Farm was dug in an attempt to exploit the potential building stones of the formation). Interspersed with this are grey, lumpy, areas of limestone composed almost entirely of shell debris, which are very obvious when field walking.

The Jurassic clays need little comment; they are mostly overlain by the Evesham soil series, and are wet, unstable areas.

The Fullers Earth rock is a similar hard rock to that of the Great Oolite limestone, but occasionally posesses within it the bands of Fullers Earth that gives it its name. No outcrops or workings of this were seen in Marshfield.

The Inferior Oolite seen here is a rubbly limestone, unsuitable for building purposes, but otherwise, similar to the Great Oolite in appearance. It also tends to support more freely draining geologies, hence the modern use of this area for arable cultivation.

The Midford Sands outcrop in one small area near Beeks Farm in the south east of the parish. They are yellow, micaceous, fine sands, and can be seen clearly in modern exposures there.

The valley of the St Catherines Brook has a chaotic geology. The steep slopes expose the clay layers, which have flowed over geological time, in turn undermining the higher rock strata, which have then collapsed. Disturbance by agriculture also results in very quick erosion (p.14).

The small Head deposit near Ashwicke is composed of rubble and clay from the Oolite. The lack of exploitable mineral resources (other than building stone) has contributed to the lack of industrial development in the parish, especially as its height and position at the extreme southern end of the Cotswolds, with only steep slopes between it and Bath, made transport into and out of the parish difficult and expensive before the era of the heavy lorry.

Over most of the parish, the soil is derived from the underlying Jurassic limestones and clays. (FIG 6) It is largely calcareous, and was probably so when the soils were first cultivated, so liming of the soils in this parish for agricultural purposes has been very limited. The soils are generally light and well-drained on the higher ground (in physiographic units 1, 3, and 5) and heavier and more clayey, with more water in the valleys, and on the slopes.

The soils of the Sherborne series (including most of the Sherborne -Haselor and some of the Yatton-Sherborne) are typical; fine loamy to clayey soils, usually stony, and derived direct from the parent rocks below (or in the case of the Yatton-Sherborne, from a scree or Head deposite derived from that limestone). This brown soil has been almost entirely ploughed in this parish, giving to a large extent an artifically stone-free soil. It was apparent in field walking, however, that modern cultivation is ploughing into the bedrock on this soil series.

The Evesham soils lie at a lower level, and are generally over Jurassic clays on hillslopes. They are clayey themselves, and generally wetter, making them much less suitable for arable cultivation, and the land-use in 1840 makes it clear that pasture and woodland occupied practically all of this soil. It is subject to landslips over much of its range, and certainly, within the woodlands of the Doncombe valley, there is chaotic landslipping, making the description of the geology there well-nigh impossible. (FIG 6)

The Atrim soil, present as a small unit along the St Catherines valley, is a loamy soil, derived from underlying fine sandy and silty Jurassic rocks. This is particularly clear in the valleys between Nailey and Beeks Farms, where the presence until 100 years ago of Aglescombe ‘Common has meant little or no cultivation has occurred. It is very easily eroded, leading to the rounded contours of the valley sides seen in that area, and thus is seldom ploughed, although massive strip lynchets in the St Catherines valley north-west of Beeks Farm show this was not always so.

Finally, the Martock series. These are fine silty soils over Jurassic silt-stones and clay, wet and very easily eroded. They lie in the bottom of the St Catherines valley, and are almost exclusively used for pasture. Where they have been disturbed in the past, massive landslips have occurred, in some cases undermining the higher Yatton – Sherborne series, as has happened on the steep hillslopes east of Ayford Mill.

These soils have clearly determined the course of Marshfield’s agricultural history, and it is quite likely that the tendency to slippage of the soils in the St Catherines valley has removed much of the pre-Medieval evidence there. (REF 1)


This map (FIG 7) shows the land use of Marshfield parish in 1982-3. The map indicates the high proportion of pasture at present in the parish, somewhat surprisingly perhaps, to a casual visitor who would see an almost completely arable landscape from either of the main roads through the parish. As would be expected, the majority of the arable lies in the drier flatter areas, and pasture occupies the valley sides and wetter Evesham soils. In one or two cases, where pasture occupies higher, flatter areas, this is as a result of fairly recent farm management decisions, and historically, these areas too have mainly arable farming.

The woodlands of Marshfield are fairly unremarkable, much of that in the well-wooded south-eastern corner of the parish having been planted as conifer woods, or converted from semi-natural broadleaves to conifer, in the last 40 years.

Certain areas were not walked by the survey teams, for a variety of reasons. These varied from the area being built up (such as Marshfield town, where the only field work done was the watching of building sites and other earth moving), to lack of manpower at the critical times of the year when the field was ploughed. Fortunately, all major landowners, with one notable exception, readily gave permission for our fieldwork. The land-use has naturally had a great effect on the nature of the archaeological evidence found, both in generating that evidence in the first place, and assuring its survival or destruction. Clearly, most of the evidence for prehistoric and Roman settlement, for example, has come from artefact scatters, and these are simply not available in pasture areas, unless earthmoving is undertaken. Unfortunately, the converse does not apply; one ploughing of a site can completely remove any earth works on
some of the more delicate sites.

Changing land-use, although possible, does not pose too great a threat to archaeological sites in Marshfield (most of them have been ploughed already!); in general, the agricultural community seem well aware of the importance of what little remains in the form of earthworks.

This vast time span (about 9,000 -10,000 years) normally yields little information in arable areas, unless in the form of extensive artefact scatters. Where an area has been intensively farmed as Marshfield has for the last two thousand years, survival of all but the contents of deep ditches is unlikely. Flint scatters of various prehistoric dates were found, however, and these have been studied by Anne Everton, whose report on the assemblages appears at the end of this section. (FIG 8)

Apart from these flint assemblages, Marshfield once possessed one of the finest barrow cemeteries on the Cotswolds. It lay on downland and was flattened for an extension of cultivation in 1947. At least seven, and possibly eight barrows of ‘Bronze Age’ (probably 1500 – 1100 BC) date existed (FIG 9) Accounts are confused, however, and I am grateful to Kim Collis for making some sense out of the conflicting maps and evidence. (REF 2, 3, 4)

Barrow 1 (ST 79417450) was examined by fieldwalking in 1947, and again in 1983; it proved to be of limestone rubble and red earth, but the only finds made were 5 flints found in 1947, including a scraper and a core.

Barrow 2 (ST 79467453) was also fieldwalked in 1947 and 1983. The bulldozer driver reported that this barrow had possessed a ■stone retaining wall’, and in 1947, five tiny pieces of ‘pale brown ware’ were recovered, along with calcined bone, and this was interpreted as a ploughed out cremation burial in a Bronze Age vessel. Needless to say, no trace of either category of find was made in 1983. A few unworked flint flakes were found on both occasions.

Barrow 3 (ST 79527456) was excavated totally in 1947 – 1949. It was confluent with the barrow 4, on its north-east, and the earlier of the two. There was a central primary cremation (possibly cremated at a site under the barrow mound) which may have been of a female. Finds made included 2 amber beads, a small bronze awl, two shale buttons and a shale bead, and a total of 42 flints, including three scrapers, two cores, and a poor petit-tranchet derivative arrowhead. The barrow was surrounded by a well-built retaining kerb of local oolite blocks.

Barrow 4 (ST 79537457) was also excavated during 1947-8. It too possessed a dry-stone wall retaining kerb, although this was more plough damaged than that of barrow 3. There was a primary cremation in the centre of the barrow, probably of ‘a youth in his teens’, mingled with much burnt stone and sherds of a ‘Middle Bronze Age collared urn’. One fragment of this vessel was illustrated in the final report, and an estimate of its size made. There were at least three or four ‘token’ cremations of a few fragments of burnt bone, in pits in the south-west sector of the barrow. 124 flints were also found in the excavation, including a fine geometric microlith of Mesolithic date (and thus residual) three scrapers, a portion of a fine flint knife, two cores, and a barbed-and-tanged-arrowhead.

Barrow 5 (‘St Oswalds Tump’) was also fieldwalked during 1947 and 1983. In 1983, only two flint flakes were found, compared with 93 in the 1947 walk. These included a very fine flint knife, seven scrapers, a barbed-and-tanged arrowhead and three cores. The bull-dozer driver took the remains of two bronze daggers found during the removal of the mound, into Bristol Museum in 1958. They were of a type then dated around 1500 – 1350 BC, but would now be dated a little later, perhaps 1250 – 1100. (REF 5) He reported that a stone cist, 1.8m x 0.5m x 0.5m had been found, with the daggers and a quantity of calcined bone. Although the mound has disappeared, the ditch which surrounded it still survives, filled in, and still shows, even from ground level, as a prominent crop-mark. References to St Oswalds Ring (as in Aubreys Monumenta Britannica) may be describing this ditch, rather than any suggested earthwork.

Barrows 6-8 (ST 79607465 approx.) are described as ‘three low mounds’. These may not have been barrows; no trace of them was found during fieldwalking in 1983.

Barrow 2a (ST 79487454) may have existed, or may be simply a reference due to confusion in the reports.

Barrow 9 (ST 79437451) is similar; LVG reported a mound at this location (REF 4), but there was little or no sign of it in fieldwalking in 1983.

The rest of the field yielded little in the way of flint scatters, either in 1947 or 1983; presumably implying that the flints in the barrow mounds are not simply chance incorporations from the soil used to build the mound.

The relationship of this barrow cemetery to other sites in the area is unknown; no other barrows or barrow≠like mounds are known in the parish, although we were not able to consult air photographs of sufficient quality to be sure that the crop-marks of ploughed out barrows and other earthworks cannot be seen. This work remains to be done.

Apart from flint scatters, no other arefacts of the Bronze Age period were found in fieldwalking, with the possible exception of one sherd of pottery of ‘Bronze Age’ character seen in the topsoil from the excavations at Ironmongers Piece (H.A.W.). This is not entirely surprising; the pottery of the period is very unlikely to withstand the centuries of subsequent arable cultivation.

The subsequent centuries, up until the Roman arrival in the area in c47 AD, are little better known. The excavations at Ironmongers Piece have shown that the Iron Age period may be well-represented in the archaeological record, if only we knew where to look; certainly no extensive scatters of Iron Age pottery have been found during fieldwalking in the parish, which is perhaps slightly surprising for an area that had such dense settlement in the succeeding Roman period.

As mentioned in the Roman section. Iron Age pottery, mainly of fabrics that need not be any earlier than 200 BC, and are more likely to be considerably later, was found at several of the Roman sites examined. (FIG 8) It is very difficult to be sure whether the tiny amounts found are significant; they may represent the work of local potters early in the Roman period, before Roman ceramic technology took hold; they could be residual; but in view of their rather poor resistance to weathering, compared to the Roman sherds found, it is possible that they do represent a genuine late Iron Age presence on the sites in question; only excavation can solve this problem. At Ironmongers Piece, the finding of a very small number of Iron Age sherds in the initial fieldwalking (less than 20), and the finding of only 14 more in subsequent gridwalking in 1982/3, was not a good indicator of the fact that deep field ditches of 50 BC-50 AD underlay the later Roman site, and that even some domestic structures may have been on the site before the Roman period. (REF 6)

Fieldwalking at the Harcombe Roman site also yielded small numbers of sherds in Iron Age fabrics, 11 in all, and thus representing less than 1% of the pre-Medieval pottery found.

The pottery was also found at the Hams site (2 sherds) and at the Blackies site (3 sherds). Unfortunately, the excavations carried out in the Iron Age sites in the immediate area (such as Bury Wood Camp, Colerne) are not helpful in interpreting the pottery, nor in assessing the nature of the Iron Age landscape.

If the excavations at Ironmongers Piece are typical, it may be that Iron Age field systems underlie much of the north of the parish, but this is not provable, of course.

No Iron Age casual finds appear to have been made in this parish.

This is a rather unsatisfying picture, given the quantities of information for later and earlier periods, and possibly further work on the Roman sites of the parish will yield some extra information about late prehistory.

One final puzzle remains. In 1779, Rudder, the Gloucestershire County historian, in describing the pasture of Marshfield Down said that ‘…visible traces of intrenchments remain there…’. (REF 7) Unless he was talking about the ditch around St Oswalds Tump, which seems unlikely, it is not clear what site he was describing. The one large soil and crop mark on Marshfield Down, which is not a road marked on early maps, is a broad mark about 20m across, running from the Down Road at ST 79307426 to ST 79707445, on the parish boundary. This appears to be crossed by a crop-mark of one of the eighteenth century roads on air photographs, and assuming it is not a geological boundary (it does not tally with any such shown on geological maps), this could correspond to the ploughed-out remains of Rudders ‘intrenchments’. The area certainly deserves a geophysical survey.

The Flint assemblages by Anne Everton

During the course of the survey, over 6,0 00 flints were recovered, some occurring in heavy scatters but many fields producing few or none. One of the problems of interpretation is that the soils of the Sherborne series (P.13 ) generally produce larger quantities of flints, due to the thinness of the soil. It was very clear, however, that even within this area, certain fields stood out as
having much denser scatters.

In the thicker soils around Marshfield Town, flint finds are much less common, presumably still being deep down in the soil, which has been built up above the prehistoric layers by centuries of manuring. Other factors affecting the presence of flint scatters include colluviation. This is a process by which soils gradually ‘creep’ downhill to fill shallow valleys or depressions, taking their content of artefacts with them. This would be especially evident in the St Catherines Brook and Broadmead Brook valleys, where the clay soils and rock make soil movement and slumping inevitable. This process is often aided, or even caused, by cultivation.

The quality of the flint recovered is generally rather poor, with some re-use of material from earlier occupations, so for example, flakes of polished axes are not an infrequent find, and often, cores from earlier periods have been reflaked. Tools are also sharpened for re-use; it is even possible that scatters of discarded flint of an early prehistoric period attracted late settlers to the area, in this region without natural flint deposits.

Generally, the later prehistoric flint tools are smaller than those of earlier periods, implying that increasing difficulty was found in meeting demand, as sources originally used were no longer available, possibly coming under different political control as a result of increased organisation and control of the landscape, rather than exhaustion of the original sources.

These sources were probably chalk flint from the Wiltshire Hills, Combe Rock deposits also from Wiltshire, with small elements of Greensand chert and Portland chert. There is very little evidence of use of rolled flint from river gravels.

The earliest flint found in the survey is represented by a few (less than 5) microliths of the Mesolithic period (8,000 – 4,000 BC), probably from the latter half of the span, suggesting little exploitation of the area, unless processes such as colluviation have removed the evidence. It is also worth bearing in mind that sites of this date are not very common, and that only a small percentage of the parish fieldwalked was arable.

There is, so far, little evidence of
earlier Neolithic occupation, except that the finds of re-used polished flint axes imply possible clearance in
the general area, and the scatter of leaf arrowheads may begin in this period, although this form of tool was used over a long time span, as late as
1600 BC in the early Bronze Age, in some cases.

Most of the flint scatters suggest a late Neolithic/Early Bronze Age settlement in the area, as shown by the petit-tranchet and barbed-and-tanged arrowheads, and some of the
scraper forms, and lies on present evidence in the date range 2500 – 1800 BC. (FIG 8)

Flints of the later prehistoric period are not evident in this material, but these may anyway be difficult to recognise in an assemblage of poor quality and heavily re-used material of this type.

One enigmatic class of finds is a group of unpatinated, blocky flints, with crushed edges, best described as ‘bashed lumps’. They are not corticated, having been roughly prepared to produce a number of facets before use. The distribution of these is interesting, as they lie almost exclusively in areas adjacent to medieval occupation, so for example, this was the only class of flint find in field 634 at West End Town, where at least eight occurred. It is quite possible that these may be medieval or later strike-a-lights.

There are also two post-medieval gunflints.

Ironmongers Piece produced evidence during the excavations, and during field walking – before and during excavation – of occupation from the earlier Neolithic into the earlier Bronze Age (perhaps 3,500 – 1600 BC), which stretches into the field to the north.

The flints from fields 11 (near the
Shirehill Roman site) and field 91, near the Harcombe Roman site, suggest that their main period of occupation was during the later Neolithic to
early Bronze Age (2200 – 1600 BC), whereas field 73, in the north of the parish, near Harcombe Wood, was occupied solely in the Neolithic period.

The chisel and oblique arrowhead forms in fields 192 (near the Hams Roman site) and 180 (about 200m north≠east of the last), ‘suggest their main
occupation period was the later Neolithic.

Other flint scatters occur in the parish for example, the fairly dense scatter first recorded in 1912 to the north of Beeks Farm (REF 8), and are also generally on the lighter soils of the Sherborne Series.

There is not enough evidence from these sites to justify suggesting any firm evidence for occupation sites at present, although, of course, large areas of the parish may have fallen within areas covered by shifting settlement sites of the late Neolithic or early Bronze Age periods.

The range of tool types in these scatters seems limited, the majority of them being the.ubiquitous scrapers, generally smaller in size than in other areas, a few knife forms, few borers, and the occasional notched scraper.

One laurel-leaf of the early Neolithic period was found, in the site at Ironmongers Piece.

The impression generally is of an impoverished flint industry, but this may be an unjustified assumption, as in areas where flint is so scarce, the larger prestige items that would denote a wealthier population were probably more likely to be looked after, and retained longer.

The knives found in the soils of the
barrows on Marshfield Down, however, were almost certainly prestige tools deliberately incorporated in the material of the mounds, presumably implying that they did at least exist in the area.

Unfortunately, the restricted range of tool types, with consequent difficulty in identifying different forms of agriculture and land use, makes it almost impossible to assess the nature and appearance of the landscape of the parish in the various prehistoric periods.

Almost the only environmental evidence from the area is the report on the charcoals from the barrow cemetery, which simply suggested that oak, ash, beech, were present in the middle Bronze Age.

Any further work on the prehistoric period in Marshfield and the southern Cotswold area should include some attempt to find environmental evidence for the period involved, as the flint typology approach appears to be of limited use in this case.

Within the context of only 20% of the parish of Marshfield being fieldwalked arable land, the discovery of at least nine Roman settlement sites thereon (along with at least four other possible settlements in the pastoral part of the parish) must be highly significant. Due to the very nature of our methods, in fieldwalking present arable land, the finding of pot scatters of Roman date has been largely confined to the flatter, higher areas of the parish, (presumably as well suited for arable farming in the Roman period as today), and thus possibly in the area where settlement was thickest in that period, it is surely significant, however, that finds of Roman pottery have occurred at Beeks Farm and Oakford (FIG TO), in. both cases, on the valley slopes of the St. Catherines Brook, where the emphasis is largely pastoral today (although it has not always been so; massive strip lynchets exist in several places in the valley to indicate past cultivation, presumably in the early medieval period). This hints, perhaps, at the former existence of a populous Roman landscape in the valleys of the parish, too. Evidence of this will necessarily be difficult to find – not only is modern agricultural practice against this, but at least in the St. Catherines Brook Valley, soil slumping on a large scale has probably removed and/or buried the Roman horizons.

There is also the probability that the Roman landscape contained more woodland than at present, and if the landscape at Marshfield was as intensively farmed and occupied as that of other counties (e.g. Northamptonshire (REF 9) or somerset (REF 10, 11) then the most likely site for this woodland would be on the less agriculturally attractive valley slopes. Early post – Medieval documents are unambiguous in their references to clearance of woodland on a large scale (e.g. REF 12) and while some of this may have been woodland regenerated over Roman farmland, its location largely on valley sides (along with the surviving woodlands in those positions today), strongly indicates at least some wooded element in the Roman landscape.

As well as the settlement sites definite and possible, mentioned above, there are scatters of Roman pottery outside of these areas, presumably arriving in these fields as a result of Roman manuring practices, spreading them with the household refuse. When these sherds are also plotted on the maps, some form neat ‘halos’ around the settlement sites -in particular around the two West End Town sites, and the two sites near Shirehill – which could be from the manuring of fields close to the farm.

Other fragments are less easily explicable. A few sherds of pottery at Whiteshill (ST 771757) may perhaps be manuring relics from a site, as yet undiscovered, to the north in West Littleton, while scatters such as the ten sherds found in fields around ST 779748 could indicate a site as yet undiscovered, since they seem rather far from the nearest known site at Castle Farm. Equally it seems odd that sites as prolific of pottery as the Harcombe and Blackies sites (FIG 10) should have so little pottery in the fields around them; Harcombe none at all, and Blackies only 2 sherds and indeed, very little scatter within field 656 either. It is particularly unfortunate that certain arable areas could not be walked, either because permission to fleldwalk was not forthcoming or lack of manpower at the appropriate period, and a more extensive walking of fields in the parish is needed before any real conclusions can be arrived at, concerning the number of settlements that were present in the Roman period. The closeness of sites at West End Town and Shirehill shows how high the potential population density could be; the total absence of Roman pottery from many fields shows caution is needed in estimating it.

Bearing all the above in mind, an estimated total of 30 Roman sites in the parish may be on the conservative side.

There would have been no nucleated village in the parish in Roman times (indeed, there was probably none until the 13th century). This was typical of pre-9th century England; the settlement pattern of scattered farmsteads and hamlets still to be seen in Wales or the West Country was common throughout England then.

Neither does there appear to have been any ‘villa’ in the normal sense of the word, at Marshfield. While several of the sites may have been well-to-do (as judged by the existence of fine tablewares, painted plaster etc.), none appears to be of sufficient scale to merit the title of villa, only the Ironmongers Piece or Harcombe sites perhaps approximating to it.

The economy of Marshfield in the Roman period was presumably dominated by its agriculture, the light soils being as suitable then, as now, for arable farming, and the river valleys for pastoral farming and woodlands. Excavated data from the farmstead at Ironmongers Piece certainly show the presence of cattle, sheep/goats and pigs, their relative importance varying at different dates, and the existence of the ‘corn dryer’ -possibly a malting oven – implies barley production, or at least some form of grain crop.

Undoubtedly, Bath would form an important market for Marshfield, both for the sale of agricultural produce, and the purchase of pottery, salt, tools, etc., which were unobtainable locally. The Fosse Way was built very early in the Roman period, and forms the eastern boundary of part of the parish today, and this would have made for easy communication with Bath and Cirencester.

The remainder of the Roman road system of Marshfield is not clear, but much of the current system could have Roman, or even prehistoric, origins. The lane known today as Rushmead Lane/ Castle Lane, is a good case in point. As late as the 16th century (REF 13) its course probably ran from the A420 along the southern side of the valley of the Broadmead Brook to west End Town, then along the currently metalled road to its junction with the Tormarton road at ST 78547525, whence it ran across the fields beyond, where it was known as the Greenway, eventually along what is now the southern side of the field containing the Shirehill Roman site, and down to the stream. This last section has now been largely obliterated, although it survives as an earthwork in the field next to the Tormarton road, and can be seen as a crop-mark in some of the fields beyond. (The name Greenway also seems to apply to other sections of this road, as field names ‘Greenaway Tyning’ next to it imply). This line takes the road within 100m of three Roman settlements, and within 30 0m of two more, making it a likely candidate for a road in use in the Roman period.

Some roads must have served the Beeks Lane site and the two other possible sites nearby: the present Beeks Lane seems a likely choice. Whether the hollow-ways leading to Ayford (still in use) or Oakford (now largely disused) date back this far, is unclear. Unfortunately, the rate of formation of hollow-ways is notoriously difficult to estimate, but they are at least plausible candidates for roads with Roman or pre-Roman origins.

Other Roman sites, now some distance from roads (such as the site at Oldfield) must have at one time had tracks linking them to the larger roads nearby, although discovering them after as much as a thousand years of subsequent ploughing seems inpossible. Sites on the eastern side of the parish may have communicated direct with the Fosse Way – as in the Hams site, where the Roman access route may have been the present footpath to Colerne, a road only 200 years ago, and still the parish
boundary for part of its length.

Other elements of the Roman landscape are even less clear from the field evidence.

The farm buildings were of local stone foundations, and with superstructures of stone, timber, or possibly wattle and daub. Roofs of Pennant sandstone and/or ceramic tile are recorded from several sites, and no doubt thatch was also used. Almost certainly, as at any date in history, many short lived timber buildings and lean-tos would have been found around the more permanent stone buildings, although they leave little or no trace in the archaeological record. Apart from the excavated site at Ironmongers Piece, this is about all we can tell of the buildings.

The appearance of the fields themselves is unclear. From evidence in other parts of the country, upland Roman field systems are often composed of mainly small, rectangular fields such as survive at Springs Farm today (FIG 16); it is likely that much of the parish would have been laid out in these fields, since obliterated by many centuries of ploughing. There is some anecdotal evidence that similar fields once existed around the Blackies site (REF 14), and it is just possible that more may be detectable in ideal crop-mark conditions.

Certainly, from the evidence above of ‘halos’ of pottery around the Roman sites, there was probably cultivation of areas near the farmsteads; despite everything, we know so little of Roman agricultural practice that it is impossible to go further than this in describing the comtemporary landscape.

1. Shirehill Roman Site (Field 11) (ST 793763)

This site was first fieldwalked (it was actually the first field walked during the Parish survey) after ploughing and harrowing in October 1982 revealed two rectangular stone scatters in the north centre of the field, visible from the south. This site was only fieldwalked, not gridwalked, due to problems with availability of manpower at the appropriate time in 1983 and 1984.

This was sufficient, however, to indicate a Roman presence in the area, as well as the earlier prehistoric site discussed above.

In all, 214 sherds of Roman pottery were found in the field, ranging from 1st to 4th/early 5th century. No pre-Roman pottery was found, nor any of the 5th – 10th century ‘Dark Age’ period. (Finds of these dates, however, tended to be made only in the more intensively studied gridwalked fields). The Roman period produced no other finds, except perhaps the one tiny sherd of turquoise glass found, but this could be post-medieval.

The pottery is biased towards fine wares; the ratio of fine to coarse is approximately 1:10. Of the fifteen identifiable sherds of Samian ware, 2 rims of Drag. 18/31 bowls were identified, along with one fragment of a rim of another bowl type, probably of first century date. There were also four fragments of New Forest colour-coated indented beakers of the second half of the fourth century, and two other unprovenanced colour-coated vessels.

The coarse ware followed the usual pattern of sites in the Marshfield area; roughly 25% is BB1, and the rest assorted wares, presumably of local origin.

There appears to be no sign of roofing materials (except one possible fragment of Roman ceramic roofing tile), which raises some awkward questions. The possibilities for this site would seem to be:

1) That the site has as yet been largely undamaged by ploughing, and the pot scatter is from an area
peripheral to the main site.

2) That the occupation site is slightly outside of field 11 -presumably in the woodland to the north, as the fields to the east, west and south have all been examined, and produced very little Roman pottery.

3) That ploughing and robbing subsequent to the Roman period have almost entirely destroyed the site, leaving only a pottery scatter.

Option 3 seems the most likely, the small numbers of very abraded 12th/ 14th century sherds in this area point to manuring, and therefore, probably cultivation in this area in the Medieval period (see below).

2. Ironmongers (Field 14) (ST 798760)

Following the discovery of a flue of a corn-drying kiln by the farmer, Mr Charnaud, this field was grid walked by R. lies in 1981 (REF 6). The area with the greatest concentration of Roman finds was excavated for ACCES by Blockley and Day (forthcoming).

The evidence from the excavation has been useful in the interpretation of some of the other sites in the parish, along with its possible implications for the pre-Roman landscape.

During the course of the excavation, the rest of the field was gridwalked twice, once in very poor visibility, when unsurprisingly, the results were fairly meagre, and once in very good conditions. Results from the two, however, were in excellent agreement. In both cases, coarse wares were overwhelmingly preponderant; Samian ware formed 4.36% and 4.78% of the total pottery found in the two surveys, while other fine wares comprised 1.09% and 0.97%. Overall, this represents an ratio of 1:18 fine to coarse wares, much lower than (say) the site at Shirehill. This may reflect the fact that the main part of the settlement was excavated and so was not available for fieldwalking, if we can accept that fine wares are more likely to be found closer to the buildings themselves.

The numbers of sherds found vary from greater than 100 per 20m square to 0, the greatest concentration being to the immediate south of the excavated area, presumably coinciding with the southern half of the cemetery which has remained unexcavated (FIG 11). No trace of settlement was found anywhere else in the field. Along with the evidence from field 13 to the north (14 Roman coarse ware sherds only from fie’ld-walking) and, the negative results of a cursory examination of the field to the east in Wiltshire, this would seem to indicate that the excavated area represents the majority of the Roman site; if there are any structures outside this area, they are such as to leave no trace in the form of pot scatters.

There were few finds of metal, glass or other Roman artefacts in the gridwalking; three bronze coins occurred, all of which were illegible but recognisably Roman in date. One
other bronze small find occurred.

The total of 27 sherds of Iron Age pottery found is high for a site in Marshfield. 14 of these were found in gridwalking the main field, and these nearly all occurred in the area to the south and south-west of the excavated area, again confirming that the excavation covered the majority of the site (although a chance observation in the heavily waterlogged conditions of December 1982 of a part ring of patches of soil retaining water in the extreme south west of the excavated area may indicate the presence of Iron Age structures there – such as a ring of postholes or the like). None of this pottery from fieldwalking need be dated outside the range 50 BC – 50 AD.

Finds from the end of the Roman period only consisted of two sherds of heavily shell-tempered fabric, and in view of the proven post-demolition activity at Ironmongers Piece, the findings of these few sherds may be significant in the implications of the few sherds found elsewhere in the parish.

The Roman sherds were about 5.5% fine wares (1:18), consisting of 46 sherds of Samian (including 6 recognisable forms, i.e. 2 Drag. 27 bowls (2nd century), 2 Drag. 18/31 bowls (undecorated and thus having a wide dating bracket up to 3rd century), 1 Drag. 33 cup (2nd century) and 1 Drag. 38 flanged bowl of the late 2nd century). There were also 9 sherds of New Forest colour-coated vessels of late Roman date (probably 4th century), and 16 other fine colour coated sherds, mostly of Oxford provenance. Of the coarse wares something of the order of 20% were BB1.

The evidence of the gridwalking is quite clear concerning the occupation area in the field (FIG 12). An area 100m square at ST 779756 contains about 65% of the Roman pottery found in the field, and this can be quite clearly seen in the diagram to fall into two groups. These may represent two buildings set at right angles, one parallel to the south east side of the field, and one parallel to the north side. The area is covered with fragments of Pennant sandstone roofing tiles, many having nail holes, and some clearly burnt. Although not recorded in detail, the majority of the Pennant recovered from the field occurs in this area. This was the main roofing material at Ironmongers Piece, and was at least an important subsidiary material at the Ham

Apart from indicating the sites of these buildings, it is probable that at least some clues to function can be obtained from the gridding data; in the eight grid squares covering the possible eastern building, 17 sherds of mortarium, representing at least ten vessels occur, including two squares with eight and five sherds apiece, while no other square in the entire field contains more than two sherds, and there were only 38 sherds of mortarium from the whole field (152 grid squares all told). This concentration of these sherds in the eastern area of the potsherd scatter may indicate some domestic function for the hypothesised eastern building.

Well outside the main concentration of Roman finds, a bronze finger ring of Roman date was found.

As in so many of the other Roman sites in this parish, there has been subsequent long and heavy ploughing. This has probably destroyed the majority of structures in the field, and bedrock is being ploughed in several places. It would be interesting to do a small exploratory excavation be undertaken to test the survival of the Roman structures on the site.

4. The Ham (Field 192) (ST 788734)

This site was already known from fieldwork carried out by R. Knight in the 1970′s, before the survey began. It was nevertheless walked in strips in an attempt to delimit the area of Roman occupation. This only revealed 37 sherds of Roman pottery, but also showed the extraordinary quantity of ceramic rooftile of Roman date that was lying on the surface. A sample of 30 fragments was taken, and proved to comprise both tegulae and imbrices, the two components of Roman ceramic roofs.

The site was subsequently gridwalked in January 1983, and 33.5 Kg (73 lb) of Roman ceramic tile was recovered.

Subsequent plotting of weights of tile recovered per square neatly pinpointed the site of the Roman buildings (FIG 13) and this was broadly confirmed by the plotting of Roman potsherd counts (FIG 14), although these showed a distribution slightly skewed to the NE of the area indicated by the tile evidence; presumably other buildings roofed in Pennant sandstone and other roofing materials also existed on the site, and the ceramic roofing tile only indicates the site of one.

Of the total of 226 sherds of Roman pottery recovered, no less than 22 (9.7%; 1:10) were of fine wares. The 16 sherds of Samian included 2 Drag. 18/31 bowls (mid 1st – 3rd century), 1 Drag. 27 cup (1st/2nd century), 1 Drag. 36 bowl (2nd/3rd century) and a decorated sherd from a Drag. 37 bowl (2nd century). Other fine wares included 2 New Forest colour-coated beakers, one indented and late 4th century, a sherd of a late 4th century rosette-stamped Oxford bowl, and three other finewares, one probably an imitation of a Samian Drag. 33 cup in a black colour-coated fabric.

Other Roman finds included a limestone drain block, several Pennant sandstone roofing tiles, including an almost complete specimen, and two Roman bronze coins, both badly worn, although one is identifiable as probably 3rd century.

Two sherds of a possible Iron Age fabric were found to the north of the Roman site, and about 20m away; neither need fall outside the range of 50 BC – 50 AD.

5. Oldfield Site (Field 676) (ST 750734)

This site was discovered during routine field walking in 1982, when a coin of 330 – 337 AD was found, along with 89 sherds of Roman coarse wares. They mostly occurred in the top 30m of this field, which slopes steeply from north to south. With this in mind, a band 60m wide at the top of the field was gridwalked in February 1983.

A further 407 sherds of Roman date were recovered, containing only a minute proportion of fine wares (1.22%; 1:81). Among these were a Samian Drag. 33 cup fragment, and a large part of a black-on-orange colour-coated mortarium, probably an imitation of the Samian Drag. 45 vessel. There were also sherds of New Forest and Oxford colour-coated vessels. One sherd of amphora was found, along with two of ■cheesepresses’ – ceramic colanders -in a local, soft orange fabric.

The distribution of pottery in the field is distinctly odd. (FIG 15)

195 sherds (and the coin) came from two 20m squares in the extreme north­west of the field, and only three other squares examined held over 20 sherds. This presumably means that the Roman site is not actually in field 676, but somewhere to the north. It is certainly not to the north-west beyond the ruined 19th century cottages, as this field (673) has also been walked, and only 3 Roman sherds were found. The most likely explanations are:

1) That the Roman site detected lies under the track to the north of the field (where a further 20 or so Roman sherds have been found) or even in the pasture field beyond.

2) The source of the Roman material is soil dumped from elsewhere.

While the second posibility cannot be dismissed out of hand, it seems unlikely; the level of the soil in the field is actually about 1m below that in the track, due to ploughing in medieval and modern times.

A further walking in 1984 revealed 130 more Roman sherds, but these included only 1 colour-coated fragment, confirming the very low ratio of fine coarse wares.

6. Blackies Site (Field 656) (ST 754742)

This site was discovered by R. Knight. Circumstances prevented our gridwalking this field, but parts of it were strip-walked, and the large collection of finds made over a number of years by R. Knight was examined at Castle Farm, Marshfield.

Altogether, about 600 sherds of pottery have been recovered from an area 100m x 30m at ST 754742. There is an extraordinarily small percentage of fine wares (1%; 1:100), very similar to that at Oldfield (see above). The fine wares include two joining fragments of a Samian stamped Drag 33 cup, and two sherds of New Forest colour-coated vessels. The coarse wares are about 50% BB1 or local imitations, and only one sherd of mortarium was found. Overall, the impression from the pottery is of a poor site; however, the findings of a silver coin (Roman but not available for study at the time of the survey), and a bronze bow brooch of AD 40 – 70, may indicate otherwise; metalwork was conspicuous by its absence on the ether sites walked in Marshfield. The only other Roman find in the field was a small turquoise glass beat.

Three sherds of Iron Age pottery were also found, all near or on the Roman site.

7. Beeks Lane Site (Fields 454/485 north-east) (ST 767723)

This site was brought to our attention by Mr Hayes of Marshfield, who had recovered 64 sherds of Roman pottery there in 1976. These were kindly lent to the survey for recording. These finds also included 8 fragments of Medieval ceramic roofing tiles, and a number of fragments of Pennant Sandstone tile.

As the finds came from very close to the road (Beeks Lane) in field 485 North-east, the site was walked twice, but not gridwalked. Field 454 was under pasture at the time, but this was also walked (not gridded) in 1983, when ploughed.

Altogether, a further 123 sherds of Roman pottery werS recovered, despite exceptionally adverse soil conditions when field 454 was walked. These sherds were unexceptional, many were unabraded, although none large, and as the soil is only about 15cm thick above bedrock, it seems unlikely that any structures have survived the documented heavy medieval and post-medieval ploughing of the area. The ratio of fine wares to coarse recovered was about 5% (1:20.5), and the Samian included sherds of a Drag. 33 cup (1st – 3rd century), and a Drag. 18/31 bowl (mid 1st – 3rd century). There was one large sherd of an Oxford mortarium, along with two in other fabrics.

A ‘halo’ of Roman sherds occurs in the fields around the site; Field 453 contained 7 sherds all in the western corner, and thus nearest to the site, while fields 479 and 480 also has shreds of Roman date. A coin of Allectus (296 – 298) was recovered
from field 454, in excellent condition.

Although there is not a large quantity of pottery from this site, this may only reflect the lesser attention given to the non-gridwalked fields, and there appears to be sufficient evidence to suggest the existence of an almost totally ploughed out site here.

8. Springs Farm Site (Field 588/614) (ST 766747)

This site is unusual in Marshfield in having related earthworks. Peter Fowler recorded a ‘Roman field system, with stone walls’ on Middle Down, Marshfield (REF 15), although his grid reference is a little to the north≠
west of the actual site- on the ground. The site is to the east of Springs Farm at ST 766747 and consists of a series of lynchets and low banks (FIG 16), which may indeed be tumbled stone walls, none of which are more than 1m high, and much mutilated by later ploughing.

The wall running across the centre of the site between fields 588 and 614 is butted up against the wall at its northern end, and is not shown on the Tithe Map of 1840, thus for all intents and purposes the earthworks have only survived in one field.

There are three parallel lynchets on the south-facing slope at the south of the site, petering out at their western end, and a fourth some way to the north on the level ground, now surmounted by a stone wall. All these lynchets turn at some point in the eastern end of the site to run more or less north-south, disappearing at the edges of the field, where they are replaced on the east at least by the earthworks of 19th century narrow-rig ploughing, and on the north by a featureless pasture ground, ploughed less than 10 years ago. In the south≠eastern corner of the field, the bottom two lynchets are abruptly terminated by a small area of narrow rig ploughing (18th/19th century) which has obliterated them, although traces of the lynchets can be seen in very low-angle lighting conditions. With the exception of the centre north of the site, no cross-banks exist on the lynchets, and yet there appear to be vague indications that the earthworks once consisted of smaller, rectangular fields, especially at their eastern end. The farmer reported that parch marks are often visible there, and in the summers of 1983 and 1984, the parch-marks of two cross-banks were seen, and recorded. Possibly these are the remains of Roman fields, later ploughed along the slope (probably in the medieval period.

The disturbed soil area between the topmost lynchet, and that below it/ is the area where Roman pottery was found by the survey team. 138 sherds were recovered altogether, including 6 Samian fragments and 1 colour-coated sherd, and pieces of two Pennant sandstone roofing tiles. These all occurred in soil disturbed by cattle, and the site remains as permanent pasture. In the next field immediately to the north of the site, R. Knight recorded many large unabraded sherds of Roman pottery when the last ploughing took place, although this pottery was not available for study at the time of the survey. This would seem to indicate the existence of a Roman building somewhere in the immediate vicinity.

9. West End Town Site (Field 623/634) (ST 768743)

The Roman site at West End Town was discovered as a consequence of gridwalking the previously discovered medieval site there in June 1983. Initially, our interest was simply in the large quantity of 11th – 15th century pottery found in field 623, and as we hoped it would be possible to identify medieval house sites by gridwalking, a 10m grid was used.

For the Roman period, 126 sherds were recovered, adding to the 62 already recovered in stripwalking in December 1982. The distribution was odd; the stripwalking certainly suggested that the majority of Roman sherds were in the southermost 3 0m of the field, rather than the next (50 sherds to 12), and this was borne out by the gridwalking (76:36). This also showed a very clear pattern; the Roman pottery all fell to the east of a line running north-south across the centre of the field, beyond this there was practically none.

A stripwalking of field 634 across the road, when ploughed in August/ September 1984, gave remarkable confirmation to this picture. The north-westernmost 20m of this field yielded 77 sherds of Roman pottery, while the rest of it (about 95% of the area) yielded only 21 sherds, all much smaller, and all abraded.

The implications for the Roman site at West End Town are clear. It almost certainly must lie somewhere along the strip to the north and south of West End Town Street, as only five other sherds have come from soil disturbance in fields along this road as opposed to the large amounts of medieval pottery found there, and it would appear that the Roman site probably lies around the eastern side of West End Town Farm.

No pottery of definite “Dark Age” date was found on this site, although sherds of early 11th century date were identified among the stripwalking material by M. Ponsford, and would seem to indicate some early medieval interest in the area.

The Pancras Close field name at the other end of the farm has excited some speculation about a possible early church dedication. The evidence does not exist – our earliest reference to the chapel is in 1584 (REF 16), and the Pancras name is not recorded until the 18th century. (REF 7) It would be quite in order as a medieval dedication.

However, if West End Town was once the major settlement in the parish, as seems likely, then the eccelesiastical centre would have been here, and no other dedication is available for the chapel.

10. Castle Farm Site (Field 578) (ST 771743)

This site has been identified by finds made by Mr R. Knight over a period of some years while carrying out pipeline and building works around the farm.

In total, about 170 sherds of Roman pottery have been recovered, but having a low ratio of fine wares
(about 3% 1:33). Included in this count are ten mortarium sherds, some unabraded, and presumably indicating the existence of a Roman building nearby, or under, the present day farm.

Two sherds of the shell-tempered ‘Dark Age’ pottery of post-Roman date were found (one the rim of a baggy pot). In conjunction with the site at West End Town (see above) this argues a possibility of continuity from the Roman into the medieval landscape. Other finds from the site have included a stone spindle-whorl, almost certainly of Roman date, and part of a bronze Roman spoonhandle.

All finds are in the possession of Mr Knight at Castle Farm.

Other Possible Roman Occupation

11. Oakford

Amongst the medieval finds from Oakford (see below), sherds of Roman pottery were found. Nine came from molehills close to Oakford Barn, along with medieval pottery, and 36 from the earth moving below the barn (see below under Oakford). Many of these sherds were large and unabraded, while this is clearly not the Roman site itself, there is a strong possibility of one close by (? under Oakford Barn). These fragments included three sherds of Samian, 2 of colour-coated vessels, and a mortarium fragment. As with most pastoral sites, the small amount of pottery found belies its probable signifance..

12. Beeks Farm

An enigatic scatter of small numbers of Roman sherds around this farm implies Roman interest in the area, if not settlement. In the ‘lower garden’ of the farm, 7 sherds of Roman pottery occured, including a large fragment of the footring of a Samian Drag. 18/31 bowl; in the field across the lane to the north, seven sherds were found in the area within 15m of the bottom (south) of the field; a further four sherds occurred in the lowest part of field 479 to the west. It is quite noticable that no Roman pottery occurs in the rest of the fields between Beeks Farm and the Beeks Lane site; a small occupation site around Beeks Farm is possible, or else these sherds could reflect an interest of the Beeks Lane site in this area, and consequent manuring spread of Roman material.

13. Holly Barn

43 sherds of Roman pottery (including 2 Samian and 2 New Forest colour-coated ware) were found in one tiny area of this field (c.40 of the sherds came from an area about 20m x 20m). While this clearly is not sufficient for an occupation site, such clustering is distinctly odd (cf 25-60 sherds/400m^ recovered on the Roman site at Harcombe Barn), and must presumably reflect some Roman interest in the site. Like the nearby Beeks Lane site (to which it is presumably related in some way) this site has been ploughed for centuries, building up a lynchet c.2m high at the bottom of the field.

14. Building Site (Field 680)

10 large sherds of Roman pottery occurred among the 239 sherds from this site.

Roman Pottery Distribution in General

Apart from a quirk in field 595 in the extreme north-west of the Parish (9 Roman sherds, possibly reflecting Roman manuring, or a site across the parish boundary to the north), and the ‘halo’ of fields with 1-10 sherds in around Down Thorns Farm (? manuring) there appears to be little Roman pottery found in any other fields more than 100m away from a known Roman site. From his fieldwork in the Midlands, Chris Taylor has stated that ‘every field has a few fragments of Roman pottery’, but out of 181 arable walked, only 56 have turned out pottery of this date, and removing from their number the 19 fieldwalks on fields that have Roman sites in, or very close to them, evidence for Roman manuring over the rest of the area (37 fields, 20% of those walked) is not very good. Where there is evidence for this (e.g. see above) the fields tend to form areas where manuring may have occurred, and large areas (despite the apparent density of Roman settlement) have no Roman pottery at all.

With the passing of Roman administration in the early years of the 5th century, the sources for our understanding of landscape development became very thin. Assuming there had been no catastrophic events in late 4th century Marshfield (and there is no evidence for such) the area probably entered the 5th century with its full Roman landscape layout, with small farms (now possibly sliding down the social scale, if current evidence elsewhere is applied) dotted among the many small fields.

When documents first appear, in the early Middle Ages, Marshfield has developed a full-blown Medieval open field system, with two enormous open fields (imaginatively called North Field and South Field!) and a small number of hamlets, and is acquiring a nucleated settlement in the centre of the parish.

Can the developments in between be followed? It seems impossible to imagine what could cause so radical a change in emphasis and landscape, but the timescale is very long (c650 years), and changes in other parts of the country (and, indeed, in Europe) can help us to understand something of the processes involved.

It has become clear over the last 20 years that the end of Roman administration in Britain did not necessarily mean an end of the Roman way of life, and in Bath (REF 17) and Cirencester (REF 18), there seem to be definite signs of communal life continuing into the mid-5th century. This would of course, mean the Fosse Way keeping a vital role as communication between the two, and hence forming an active part of the post-Roman landscape at Marshfield. Indeed, some of the individual Roman sites in the area seem to have traces of 5th century activity in and around them.

At Ironmongers Piece, excavation showed quite clearly a post-demolition phase, after the main 4th century building had collapsed or been destroyed. While consisting of only a hearth, a few post-holes and some very late Roman pottery, this may be significant. Two sherds of a very heavily shell tempered fabric, which may be 5th century in origin, were found in fieldwalking and a sherd of grass-tempered pottery from the other side of the field wall north of the site, may indicate later interest still. At Frocester Court (REF 19) and Barnsley Park (REF 20) this pottery is dated 5th century -presumably the central two quarters -and apparently only occurs there in an agricultural context as manuring remains in the fields after abandonment of the Roman sites and not in domestic contexts. This is likely at Ironmongers Piece, as no domestic occupation seems to occur after the presumably brief post-demolition phase. This site illustrates the problems of Dark Age archaeology all too clearly.

At Harcombe and Castle Farm sites, sherds of 5th century shell-tempered pottery were found, including the rim of a hand-made baggy pot at Castle Farm, and the continued importance of the West End Town, along with its early medieval pottery (? early 11th century) implies a degree of continuity.

In the small group of Roman sherds at Whiteshill (see above), there was one grass-tempered sherd, again of a possible post-Roman character.

The fate of the other Roman site is unclear. At Oakford, substantial Medieval occupation is documented in 1224, and some of the pottery from the site probably originates in the 12th century, so continuity is clearly possible, but unproven.

None of the other Roman sites so far discovered have large quantities of Medieval pottery on them (although observation of builders trenches at ST 777736 just south of Marshfield Town did reveal a scatter of Roman and Medieval pottery) and so presumably had ceased to exist fairly shortly after 400 AD.

Again by analogy with other areas, a post-Roman landscape in Marshfield might be expected to maintain its Roman character for some centuries, and indeed through the surviving features of that landscape (roads, settlements etc) to at least mould some of the .subsequent landscape development.

What effect the nearby Battle of Dyrham (577 AD) would have had is unclear. The anglo-Saxon Chronicle statesï- (REF 21).

’577. In this year Cuthwine and Ceawlin (who was the nominal King of Wessex) fought against the Britons and slew three Kings, Coinmail, Condidan and Farinmail, at the place which is called Dyrham, and they captured three cities, Gloucester, Cirencester and Bath’.

Whether the kings and cities can be equated is unclear but what this does mean is that Marshfield undoubtedly came under Saxon political control around 577; again, to judge from other areas this may not have greatly affected the landscape, although it must have been uncomfortable news for surviving Celtic peoples in the area!

At this point, the real Dark Age closes on the area. No archaeological evidence has yet been recognised, for the next 450 years; no documentary evidence relates directly to Marshfield until 360 years later.

All that can be deduced is based on evidence from elsewhere. The agricultural systems surviving from the Roman period would be more than adequate to feed Marshfield’s own population, and while Bath survived as little more than a monastery and its few attendant buildings (after 676), little outlet for Marshfield’s surplus goods (if any) would be available. The gradual change to a more market-oriented economy, the growth of individual prosperity, and various other changes, detailed elsewhere (eg REF 22), led eventually to the setting-up of the great Medieval common field systems. This is generally attributed to the 9th-11th centuries; as later evidence suggests that West End Town was the settlement at the head of the South Field, and not Marshfield itself (dating c1265), the division may have been carried out from West End Town.

Traces of an earlier system may exist in the place-name Oldfield, to the south-west of West End Town. In other areas of the country (REF 23) this often refers to a pre-common field arrangement, and while the name is only traceable back to 1584 in Marshfield (REF 16), the fact that ploughing is recorded there in 931 (see below), and that Oldfield never became part of Northfield or South Field, is surely significant.

One of the impulses that led to the common-field system in Marshfield may have been the growth of a market in Bath, where by the 10th century at the latest, an urban character was re-asserting itself. The other often-quoted reasons for movement to highly organised communal farming systems (pressure on pasture availability, and organisation of chaotic land-holding systems) could both have played a part; lack of pasture, though, seems unlikely at this date.

The sole documentary evidence surviving from the period is an important Saxon charter of 931 AD, detailing the bounds of a grant of 5 hides in Cold Ashton to Bath Priory. This gives details of the west of Marshfield in the early 10th century, although it nowhere mentions the parish by name. The boundary runs roughly north to south and the relevant parts run (REF 24) -

…..Andlang herpothes oth Aethelmodes wudu. Thanne suth be wudu oth Eddes dene. Of thaere Dene middlewearde licgath twegen aeceras on north healfe. Th’eft on tha ilcan dene. Andlanges dene eft on Maeiweg. Andlang Maerweges on Mapodor Leage. Be eastern thaere leage twegen aeceras. Of thaere twegen aeceras on hlemmes dene. Andlang dene. Andlang dene on hlamnys wylle. Andlang broces to thaes Cinges gemythan’.

In Grundy’s translation, this runs:-

“….Along the highway as far as Aethelmods Wood. Then south by the wood as far as Eddes Dene. From the dean about halfway down lie two strips of ploughland on the northern side. Then again on the same dean. Along the dean to the boundary way. Along the boundary way to the lea of the maple trees. To the east of the lea, two strips of ploughland. From the ploughland to ‘hlemmes’ dene. Along the dene to ‘hlamnys’ wylle. Along the brook to the Kings Watermeet. (after this point the Cold Ashton boundary runs away to the west from Marshfield).

Most of these points are identified by Grundy, and they seem clear enough. Some points, however, cast a valuable light on the 10th century landscape, and are thus worth examining in more detail.

The line of the boundary is shown (FIG 17) and detailed in FIG 18; it is clear that the boundary has not substantially altered here in the past thousand years, raising some interesting questions about the age of Marshfield itself as a land-unit.

Firstly, if the 931 boundary follows the 1984 boundary exactly, the section of ‘herpoth’ from ST 743741 to ST 744735 has disappeared, and is now only marked by a hedge on a low bank. A hollow-way exists on the boundary with Doynton just beyond this, which may be a remnant of the ‘herpoth’ mentioned.

Aethelmods wood, which presumably lay in the region of Pennsylvania (ST 744735) has disappeared without trace. It is not actually clear from the charter which parish it was in, but woodland is documented nearby, as late as the 17th century (REF 25).

Eddes dene (clearly the valley of the Broadmead Brook) has left a slight trace of its name in the landscape, or at least, the Edde element, in the form of the Tedswell field names, clearly derived from OE (Aet Eddes wella) ‘at Eddes spring or stream’, presumably the Broadmead brook itself. These field names are clustered around ST 762737.

The ‘two ploughstrips halfway down the dean’ are also intriguing. If aeceras can indeed be translated as ploughstrips (but see REF 26) this confirms that ground was being ploughed right out on the edge of the parish, on a hillside, in the area called Oldfield as early as 931. It is surely not coincidence that immediately to the north of the boundary in Marshfield at the point referred to, there are two low flat broad lynchets, probably those very ‘aeceras’.

The Maerweg (= boundary road), now the course of the A420, is still the Marshfield boundary, and this section of it forms part of the Greenway road, still traceable from Shirehill south west as far as Hill Farm in Cold Ashton parish; it is possible that the maer- element in the names of the road and the parish are connected.

Mapodor leage does not survive as a place-name, or a land-unit. It is however, an interesting clue to the nature of the area; presumably open with maple trees in or around it.

The ‘twegen aeceras” east of this ‘leage’ are not visible today, but the next point, Hiemmes dene, is a good confirmation that this boundary is correct. In 1679 (REF 27) John Goslett let a piece of land called “lemsdens ditch alias Gostletts ditch’, with a further piece of land ‘on the west of Bully hill, adjoining Gostletts ditch’. The ‘Lemsden’ name is very clearly a derivation of Hiemmes dene and the area is located by the second piece of information, as Bully/Bulls hills are to the north of this tiny valley (around ST 766731); satisfactorily, ‘Gostletts ditch’ is shown as a field on the 1768 map of Marshfield at the point where the tiny stream meets the St Catherines Brook at ST 766730.

‘Hlamnys Wylle’ is presumably corrupted from a ‘hlemme’ form; its location is approximate on the diagram.

The ‘broce’ referred to in the charter is clearly St Catherines Brook, the ‘Cinges gemythe’ is undoubtedly at the present site of Beeks Mill. The name Kings Watersmeet is unusual, as this watersmeet is a pretty modest affair.

Unfortunately, no evidence is known for any of the other boundaries of Marshfield until the perambulation of 1810. (REF 28)

The boundary with Batheaston is along the St Catherines Brook, and it would seem an ideal early boundary.

The Fosseway Way is certainly also a neat early boundary between parishes and counties, as is the Broadmead Brook in the extreme north of the parish.

The 1810 perambulation gives the name of a field on the south boundary as Merkham: if this is to be trusted as an early name containing the element maer-(0E:boundary-) it again implies the existence of an early boundary. It seems at least feasible that Marshfield of 1984 shows a common boundary with Marshfield of 931. (Clearly one or two details are not that old; the small (5m) deviations in the boundary wall at Ironmongers Piece, for example, are almost certainly due to the boundary running between ploughstrips in an unrecorded phase of cultivation there). Whether or not Marshfield is an earlier land-unit still remains to be seen. It has no recognisable ‘central place’ unlike many of the other land-units supposed to be earlier.

In 1086, Marshfield was owned by Edward the Confessors’ wife, Edith, and is recorded as a prosperous manor. (REF 29)


Without any real knowledge of the nature and extent of post-Roman and pre-medieval settlement, it is very difficult to discuss the origins of the farms and hamlets of Marshfield (or even that of the town itself), and its first date of occupation. (FIG 19)

While it is possible to say, for example, that Fuddlebrook had its first houses built between 1768 and 1830, from the Marshfield Estate map, and the OS First Edition 1″ map, and that the Rocks was built between 1584 and 1650 (see The Rocks), it is not so clear if Oldfield (first mentioned as a settlement c1650), Ringswell (first mentioned about 1575) or even Down Thorns Farm (shown on the map of about 1630) can be assigned origins earlier than the post-medieval period.

However, West End Town is definitely referred to in 1263, and by name in 1303, and has yielded 11th century pottery (see p.53); Marshfield was founded in 1265, on a site of unknown age, which had a church rebuilt in 1242 on the site of a possible Norman predecessor; Ashwicke is mentioned in 1287 (REF 30), and some 12th or 13th century pottery came to light here; Ayford has produced 13th century pottery; Oakford is mentioned in 1224 (REF 31) and Beeks appears to have had a functioning farm and mill by 1584 at the latest, and probably long before. (REF 32)

Shirehill is problematic; the mill there is first referred to in 1629, but the scatter of medieval pottery around the site makes it a possible medieval settlement.

Some deeds and documents of the early post-medieval period are exasperatingly clear in references to certain furlong names (such as Cherry Furlong, Hare Furlong, Charlocke Furlong, the Furlong by the Leaping Stock etc.) but not enough occur in the documents so far examined to reconstruct the furlong patterns in the medieval North and South Fields of the parish.


The origins of the settlement now the site of Marshfield town are not known. Roman pottery was found during the survey on a building site in Sheep Fair Lane, but only in small quantities. A single sherd of Roman pottery was found close to the church about 20 years ago, but in view of the existence of the Hams Roman site nearby, this is hardly surprising. The status of the settlement then, and its subsequent status up into the 13th century, are still unclear.

The Domesday Book entry for Marshfield (1086) mentions an estate of 14 hides, with 34 ploughs, a very large number for this part of the country, implying a fairly prosperous farming community. Although it has been suggested that the Domesday entry is describing Marshfield (town), and that a one-hide unit mentioned therein as being held by a priest is that for West Marshfield (=West End Town), and that this entry was the result of an error, and that West End Town should have been listed separately (REF 29), there is no evidence for this view, and the prime settlement at Domesday may well have been West End Town. Further work is needed to clarify this problem.

Few scattered references to Marshfield survive between 1086 and 1265. The manor passed to the Bishops of Wells shortly after Domesday, and John de Villula, bishop, gave it for the maintenance of Bath Priory in 1106. (REF 7) It subsequently passed to the Earl of Gloucester, and hence (c.1168) to the Abbot and convent of Keynsham, who retained it until the Dissolution in” the 1530s. Details of these changes of ownership are brief, and no help in understanding the development of the settlements in Marshfield.

During the year 1220, a case between one Roger de Sancto Laudo (St Loe) and the Abbot was heard, concerning ten librates of land, later described as half a knights fee (REF 33). The details tell us nothing about Marshfield, but do mention that the Abbot initially avoided responding to the summons because it had mistakenly been made out to the Abbot of Eynsham!

This case, and others resembling it – a squabble over lands illegally seized in the parish in 1233 (REF 34), and the concerns of Alfred of Marshfield with his lands in Marshfield in 1224 – 1232 (REFS 31, 34) are typical of a normal farming community such as existed at Marshfield in the early 13th century.

There must have been some increase in prosperity during this period, however, and the Abbot applied for a charter for a fair of three days, and weekly Tuesday market in 1234 (REF 34). This was granted under the usual conditions i.e. that it did not interefere with the profits of neighbouring fairs, and unfortunately for the Abbot, strenuous objections by Bristol succeeded in suppressing this fair.

Thirty years later, the Abbot again applied, and this time the charter was granted, and survived. It granted a fair of three days ‘…on the vigil, fe’ast and morrow of St Oswald’ and a weekly Tuesday market. This charter was later confirmed, and by 1462, a second fair on the feast of St Philip and St James had been added, and the market day changed to Friday. A copy of this 1462 confirmation is among the Codrington archives in the Gloucester Record Office.

The granting of the fair was the occasion for a spurt of growth for Marshfield Town, and by 1334 it was the fourth most prosperous town in the county, after Bristol, Gloucester and Cirencester. (REF 36)

The Abbot meanwhile, had been granted a writ of Free Warren by the King, around 1280, and in this document, along with the right to punish wrong-doers, it is significant that the Abbot was granted the taxes of bread and beer in the town. (REF 37) This may be hinting at a pre-eminent role for barley in the economy of the town as early as 1280; certainly sheep do not appear to be mentioned at all in Marshfield’s scanty surviving medieval documents.

By 1352, some idea of the agriculture of the small rectorial mannor held by Tewkesbury Abbey can be gained from a rare surviving compotus. (REF 38) This lists -’…Received: 15 pounds 6 shillings and 8 pence from all the profits of the manor of Marshfield this year…’ ‘…Expenses: For cheese bought and sent to Marshfield at harvest time: 4 shillings. For expenses of four men at Marshfield for the harvest: 12 pence. For threshing at Marshfield: 6 pence…’

(This manor was given to Tewkesbury by Robert Fitzhamon, and could conceivably have been the one hide held by the priest at Domesday).

By the mid-15th century, the documents become a little more common, and it is probably significant that the first occupation quoted in the Calendar of patent Rolls for an inhabitant of Marshfield is ‘milleward’, and that shortly after, numbers of ‘maltmen’ are mentioned (REF 39, 40, also see p86), both references being in the middle of the century.

At this point in the history of Marshfield town, it should be possible to begin to use information from the actual buildings, as undoubtedly, some interiors of 15th century date survive in the High Street, but little or no published study has been carried out into the buildings of Marshfield, and this was outside the scope of the present study.

While details about the debts of individuals in Marshfield, and their misdemeanours which led them to be recorded in the national archives are no doubt fascinating, they tell us very little of the town itself, and it is not until the dissolution of Keynsham and Tewkesbury Abbeys in 1539 that the established order begins to change (and be recorded). As detailed under the sections of Ayford and Oakford, the Gostlett family eventually acquired Marshfield, after its passing through a number of hands in the twenty years following the dissolution.

A few of the documents in the Codrington archives cast a little light on the town in the immediate post-medieval period. The vicarage of Marshfield by this date, in the possession of New College, Oxford, was described in 1584:

‘…The vicarage House lyeth between the house of Thomas Crispe, and John Hedges the younger and hath these many rooms within it as followith: One ffair parlour Wainscotted round about, one little buttery, one hall, one great buttery, one little chamber, one kitchen with lofts over all the said house…one gatehouse with a stable, and a barn, a garden, and a little orchard containing…half an acre… one other house commonly called the Old Vicarage, letten out by the said vicar to Thomas Crisp…’. (REF 13)

A picture of the same buildings is given in 1629 (REF 13): ‘…the parsonage House with the Ccurt enclosed by the stable and other necessary housing to the same belonging, one little piece of garden ground on the East side of the said house, with three barns and a pidgeon house in the middle of the backside…” (and the positions of the barns and the surrounding properties are described). The ‘pidgeon house’, of course, is the dovecot that still remains.

These sorts of descriptions are not uncommon in the 17th and 18th centuries in Marshfield, and certain of the premises should be identifiable by diligent tracing of the strings of owners recorded in the deeds, both in the property concerned, and frequently, on either side as well.

It is interesting to find, for example, a deed of 1749 which mentions a piece of waste ground, 22 feet long, 9*5 feet wide at the north end, 64 feet wide at the south end, with a ‘little brewhouse erected on it by Henry Shadwell’, and leased to George Bennett, maltster, whose tenement was immediately to the east. This property should be easily identifiable (and it may be the record of the blocking off of one of the alleys between the burgage plots, perhaps?). (REF 41)

Similarly, a deed of 1674 mentions a ‘…tenement called The Shambles, with its stalls,, stallholdings and bulks, and the rights to set up standings on any part of the waste ground adjoining to the north…” (REF 42)

The map of Marshfield made in 1744 shows no real details in the town (except a grotesque parody of the church, and generalised plan of Home Farm), and the 1768 map, although clearer, is no real improvement, the Tithe Map of 1840 being the first map detailed enough to identify individual buildings.

Even at this late date, however, the plan of Marshfield has hardly changed from that of its medieval predecessor, only the addition of the buildings on the Inner Hayfield in the 17th century, the alteration around Weir Lane (see above) and the buildings of the Almshouses by Elias Crispe in 1612 – 1619 being major differences, and subsequent alterations have been relatively small, although with the building of the Marshfield by-pass in the 1960s/1970s, the possibility of new development along Back Lane has virtually turned it into a major route, with estate development in the fields between it and the by-pass.

The lack of early documentation also makes Marshfield’s agricultural history hard to follow. Whereas the evidence for the smaller settlements is just about enough to see some details of the layout of the medieval fields (for example), as mentioned in the introduction to this section, post-medieval evidence only points in the vaguest way to the layout of Marshfield’s two-field system, and tells a little of its roads, but little else.

Marshfield c1600

As in the attempts made elsewhere to show the landscape of Ayford and Oakford cira 1600, a note of caution must be concerning this plan.

The only settlements shown are Beeks, West End Town and Marshfield Town. It is known that others existed, but they are not important in the administration of the parish.

Oldfield is noted specifically as The details are mainly inferred from documents of later date, which may not record drastic changes; on the whole, however, the major changes in Marshfield’s landscape were happening around 1600, and this map attempts to show the landscape just as there changes where beginning. (FIG 20).

Separate from North Field: the mention in 1696 for example, is clear enough (REF 43), and the shape of the area is based on the Oldfield Farm of the 18th century. Many documents refer to the North Field of Marshfield, and Steward Sleight is generally referred to as being in North Field. The Hayfield is probably a late division of the North Field, but was in existence in 1584 (REF 13). Marshfield Down (alternatively called Oswalds Down) was the common of Marshfield up until 1851, as was Down Thorns Common, and the earliest map in 1630 shows them as portrayed, with Down Thorns Farm.

The South Field, which does not include the hamlets of Ayford and Oakford, but probably did include Ashwicke, is also often referred to. The document of 1584 (REF 13) makes it quite clear that the area between West End Town and the present line of the A420 is in South Field, although in later times this area was called the South-West Field, or Townsend Field (for example, in 1698, REF 44). The area around Beeks Farm is described in that section; the main uncertainty there is the exact whereabouts of Egglescombe Wood.

The woodland shown does not include Foxholes Wood in Oldfield, as there is insufficient evidence to locate it; the amount of woodland left at this date was surprisingly small, especially compared to that left in the smaller settlements.

The two mills portrayed were probably both working by this date; Beeks certainly was, since it was recorded in 1584 (see p.73), and Shirehill Mill was recorded not long afterwards, in 1629. (REF 13)

This map is not very detailed, as the Codrington archives do not present the appropriate information; it is possible that there are further records in the Public Record Office in London which could cast some light on the medieval landscape of Marshfield.

The Planned Medieval Town

The town of Marshfield shows clear signs of having been deliberately laid out, presumably around the time of the granting of the fair and market to the Abbot in 1265. This process is well-known all over the country, but is seldom recorded in any detail. Similar layouts exist in nearby historic Gloucestershire and Somerset (such as Wickwar, Chipping Sodbury and Axbridge), and the plan of Marshfield in 1840 (FIG 21) shows clearly the remaining elements of a medieval town’s layout.

The church and manorial farm lie together at one end of the town, with their great barn and dovecot of later dates. This area may have existed before the planned town, and it has been suggested that the Saxon nucleus from which the town developed could lie around the church, or slightly to its south-east, in the region called Little End. Watching of building works in the area during the survey has not been able to confirm this (indeed, the notable feature was the lack of any medieval pottery in the soil at all). The most striking feature of this end of the town is the regularity of the parallelogram of streets surrounding the church and farm; this may be part of the planned layout, with the streets at north and south running parallel to the two back lanes of the town (see below). It could, conceivably, even be a relict feature from an earlier landscape, but this can only be resolved by further fieldwork.

In front of this area is the market place, reached by a dog-leg corner that re-routes the old London-Bristol route through. This triangular area was the nucleus of the fairs, and markets along with the Lower Market Place, now represented by the wide street leading towards the south and Ringswell. The full extent and layout of the original market place is difficult to see; the buildings around the Nelson Inn now in the centre of the area are almost certainly not original, and they may perhaps be the descendants of the buildings which the Abbot of Keynsham put up in 1407. “…Thomas Clyve, Abbot of Keynsham, made a purpresture in the town of Marshfield, by setting up three shops with solars and cellars above and below, in the kings highway, without license,…to the nuisance of the king and his people. They are worth 6 shillings and four pence yearly, and the Abbot has taken all the issues since they were set up…’. (REF 45)

Stretching away from the market place towards Bristol is the High Street, the main thoroughfare of the medieval town. For about 350m this street is very straight, and then bends slightly to the south. The straight length is probably the original planned element. It is broader towards the market place, and with the market place, was the main site for the booths and temporary stands on fair and market days. In the broad end of the street, or possibly just inside the market place itself, stood the market cross; a little lower, and probably opposite the entrance to the cul-de-sac that leads to the church, was the lower market cross, referred to in 1575 -1715 (REF 32 and REF 46).

On either side of the High Street were laid the burgage plots, long narrow strips of land all originally of the same size, and made this way in order to maximise the number of premises that could front onto the High Street. These ran back to a pair of back lanes, now called Back and Weir Lanes, behind the plots, which meant that heavy traffic delivering and collecting to and from the workshop areas behind the houses did not have to pass through the High Street, and goods could be delivered direct, without having to pass through the front of the property. This in turn meant that the house and shop could occupy the full width of the plot, without necessity for access ways to the rear.

A number of small alleys also connected the back lanes and High Street, most of which are still traceable, but few remain open. The south-east corner of this plot layout, between Weir Lane and the road to Ringswell, has been destroyed by later development. This must have taken place by the early 17th century at the latest, as a number of houses in the area (not least Weir Farm) are early 17th century in date, and clearly laid out after the new pattern was established. The original line of Weir Lane is preserved by the property boundary running from the corner in the road to Ringswell towards Weir Farm, shown very clearly on FIG 21.

Beyond the planned medieval town along the High Street there is more development which occurred during the period 1260 – 1600. There are larger, more irregular plots, significantly with no back lanes behind. These plots followed the curving pattern of strips in the common field that were presumably developed piecemeal. In view of Marshfield’s rapid rise in prosperity, it seems likely that this development was quite rapid, and could conceivably have all occurred during the 13th century.

The buildings on the north side of Hay Street, opposite the manorial farm, were probably only erected in the late 17th century, as a deed of 1703 records a lease by Andrew Waterford, maltman, of ‘…a message, tenement, garden and backside, and buildings upon the Inner Hayfield made by William Hosey…’ (REF 44), implying a fairly recent construction date.


For the purposes of the present study, this place is taken to include the settlement lying around West End Town Farm and Castle Farm, and thus stretching from the Broadmead Brook at one end (ST 763, 741) to the West Littleton Road at the other (ST 773, 746).

The two Roman sites, mentioned above as coinciding with these farms, imply a possible early importance for this area: unfortunately, without proper excavation of either site, no clue as to the size and wealth of these settlements can be found.

The existence of two sherds of possible 5th century pottery amongst the Roman material at Castle Farm, and the 11th century pottery at West End Town Farm, along with much other evidence (see p 53) implies that there may have been some continuity of settlement at West End Town from the 4th century on, but obviously, this is not proven. Without any clue as to the status of what is now Marshfield town pre-1265, it is difficult to say which settlement was the Meresfelde of Domesday Book: the grant of the charter for a fair, and the development of the planned medieval town of Marshfield, however, soon made Marshfield itself pre-eminent.

By 1303 at the latest. West End Town was known as Wesmaresfeld (REF 47): Richard de Heydon held one quarter of a knights fee therein, in the Hundred of Thornbury. Unfortunately, the ‘Knights fee’ is not a measure of area, but that area ‘…as was esteemed sufficient to maintain a knight with a suitable retinue… supposed to have contained 8 carucates, or 680 acres …’(REF 48). If this is a true measure, (and not all authorities quote this figure), Richard may have held c170 acres all told in West- End Town.

The documents are confused concerning this land holding. The earliest reference to it is in 1263 (REF 49) when one-third of a knight’s fee is held here by Richard de Clare, Earl of Gloucester and Hereford. In c1280, this is described as held by ‘the heir of Anselm of Rochesford’ (REF 50), and in 1314 (REF 51) by Richard de Heydone. This is almost certainly the same mentioned above; perhaps the difference between one-quarter and one-third of a knight’s fee is a legal definition, rather than a true reflection of area, for in 1346 (REF 52), Richard de Heydon’s holding is again described as a third of a knight’s fee. Later that year, Jqhnannes de Sodbury, cleric, holds ‘from the said Earl of Gloucester, a quarter of a knight’s fee in Marsfeld, which Richard of Heydon once held’.

Despite the different name, this seems likely to be a description of the same piece of land. (REF 47) In 1387 (REF 53), Jacobus de Berkeley is said to hold one-third of a knight’s fee in Littelmarsfeld (as West End Town has now become), and in 1393, (REF 54), an order is given to Robert Whytington to give Anne, widow of Thomas, Earl of Stafford, as her dower ‘… the third part of one knight’s fee in Lytelmarsfelde held by Jacobus de Berkeley at £10…(etc)…’. In 1398, the same Jacobus de Berkeley relinquished this land to William, brother of Thomas, the Earl of Stafford (REF 50) and in 1404 (REF 50) it passed to Edward, Earl of Stafford. This is also (presumably) the same block of land that was claimed by Nicholas Foyntz in 1481 (REF 55), and was held by the Poyntz family in 1517 (REF 56).

The further fate of this block of land is unclear; the Poyntz deed of 1481 mentions ‘…200 acres of (plough) land, 60 acres of meadow, 12 acres of pasture, 3 acres of wood and 100 shillings of rent there…’, and this implies a large block of land.

At the time of the reformation, the lands presumably passed into the hands of the Gostlett family, but the Manor of Little Marshfield (or ‘Marshfield foreigne’) remained in their hands until 1650, unlike other parts such as Oakford and Ayford, which were sold in the 1570′s. (see pp 57 and 65 ).

Another big difference between these settlements, is that at least by c1600, West End Town does not have its own field system, but its lands were intermingled with those of Marshfield itself, demonstrable even as late as the 1768 map.

Few other 16th century records of West End Town exist: a reference in 1549 makes it clear that others owned small portions of the area (REF 57): this deed is a grant to Thomas Horton of Iford, Wilts and Richard Byllet of Cossam (? Corsham) of ‘…the cottage and two closes of land and pastures in the tenure of William Robyns, in Weston, alias Westmershefeld, which… were among the possessions of Hortons chantry in Bradford-on-Avon parish church.

The other important references are the two in terriers of New College lands in Marshfield for 1584 and 1585 (REF 13). Two items in the first run “item…one acre in the South field, shooting east and west lying on the north side of a piece of land of Richard Reeds of Littleton and on the south side of a Meer that is between the Chapel and the Windmill…item one other acre lying between that and the Chapel shooting southward…at the west end of the furlong shooting from the Hinnock towards the Windmill’. In the 1585 document, the second item is more or less identical, but the first runs ‘item, one acre in the South Field on the south side of a Green Meere that is between West Marshfield and the windmill in Cold Ashton parish…’.

This could hardly be clearer. The windmill was at ST 756731, where a windmill mound can still be seen,- the (H)innock at the south-east of West End Town, as the field names clearly show, West End Town street was presumably already in existence, hence the ‘green meer’ mentioned (a ‘meer’ is a grassy balk in otherwise arable land) was almost certainly along an extension of the present track from West End Town (ST 76607415) to ST 76547395 on the A420. The chapel at West End Town was therefore at the west end of the hamlet, and it seems likely that it was in the field called Chapel Close (ST 766741), where a large building platform can be seen.

There are only a few other references to this chapel. In 1682 (REF 58). ‘Chappie Tineing at West End Town, on the south side of a house of Richard Chambers, lately converted into French grass…’ was mentioned. (French grass was a common name for sainfoin, a leguminous crop very popular in the 17th and 18th centuries). Fifty years later, Joseph Chambers was leasing from Henry Harrington ‘…all that east part of a message called the Chappell in Weston Town, containing 3 bays of buildings, and also all that west part of the same message with 1 bay, late… of William Harding of Upton, together with the little close of meadow adjoining…” (1728; REF 59).

In 1768, the field at ST 765741 is also called Chapel Close, but the house in it belongs to the Taylor family, so it seems most likely that the chapel stood in the present Chapel Close.

The field name ‘Saint Pancras’ is not known before Rudders 1779 repeat of the legends about Saint Pancras Well, so the assignment of this dedication to West End Town chapel is hazardous to say the least. The well is not mentioned before 1779, and is not known as such today.

The rest of the settlement is more difficult to understand . The findings of sherds of at least 20 medieval glazed ridge tiles in field’ 623 implies a building of some quality nearby (see p 53); the stone window frame implies a late medieval building. In view of the field name (‘Court & Close’) it seems possible that a manor house stood in this area or nearby, during the 14/15th centuries. Documentary evidence for this is completely absent. (see p49)

The farms and cottages of the settlement are also problematic. There was obviously a building at Castle Farm in the 16th century; it still survives. West End Town farm is obstensibly of 17th century date, while the house ‘Moonrakers’ next door to the farm is so altered as to be undatable.

Early maps of Marshfield are not much clearer: the 1768 map only shows the buildings to the south of ‘West End Town Street’, as the area to the north belonged to a different landowner (FIG 22): the Tithe Map of 1840 is the earliest large-scale plan of West End Town (FIG 23) and shows the area very much as it is today.

This is a far cry from the relatively large settlement of the 13th century, and as well as the chapel, other buildings are also referred to as going out of use; for example, in 1692 John Gostlett leased to Richard woodward ‘…all that one late messuage…now converted into a barn, together with the pasture on which it stands …” From the rest of the deed, it is clear that this building stood on the site of the (later) ruined barn at ST 76947415, opposite Westend Farm. This building was almost certainly occupied during the medieval period, as a pottery scatter was found in an area of dark earth around this site in fieldwalking in 1984. (REF 44)

Another building which disappeared before 1768 was at ST 76457415, described in 1703 as a ‘…copyhold messuage or tenement called Smiths Hold, a ground of Joseph Davies on the west’. This building thus stood in the long narrow field called ‘Smiths Hold’ in 1840, and there is a great deal of stone on the surface there today, presumably remains of the building. (REF 60)

Other building sites are indicated at ST 76577415 and ST 76777419 by the 1768 map: the age of these is not known, but the second is shown on the map with an enclosure around it, and this corresponds well with an area of black soil that yielded an extensive scatter of medieval and post-medieval pottery and clay pipes when walked in 1984.

The earthwork surveys in Saint Pancras Close and at the east end of West End Town Farm (FIGS 24 and 25) show more house sites; an excavation in 1981 north of West End Town Green at ST 76947425, to test the nature of earthworks there, was inconclusive.

At the other end of the settlement, the history of Castle Farm is obscure. The existence there of a long-house of at least 16th century date implies earlier settlement: certainly the fields to the north-east and south of the farm have produced scatters of medieval pottery, and a farm of some importance and wealth must have preceded the present building, as excavations several years ago in the garden of the present farm revealed a stone-lined pit, containing quantities of 18th century glass (mainly wineglasses and ale flutes) and 18th and early 19th century pottery. (REF 61). An immense quantity of fish bones were also found, along with many of duck, chicken and pigeon, and the dating of at least most of the material implies a wealthy farm here in the 1750′s. Very little other information exists concerning the actual buildings of West End Town.

The fields of this settlement appear to be largely interspersed with those of Marshfield itself so, for example, a three acre plot of land at Harcombe Wood, over 1*5 miles away from West End Town, is described in 1669 as ‘sometime parcel of a tenement called Smiths Hold’. (REF 62) Neither does West End Town have a separate two field system like Oakford and Ayford. Equally, the intensively farmed nature of this landscape means that all its woodland had been cleared by a very early date. Neither is there much evidence of small closes around the settlement, except the plots in which houses stand, and the plan of the enclosure around the cottage at ST 76777419 on the 1768 map makes it very clear that this is a late enclosure from medieval strip cultivation.

The reasons for the decline of West End Town from its 18th century hamlet status are unclear. Certainly, much of its area came under the ownership of the codrington family as did much of the rest of parish; it may be that the malting industry of the 18th century, with its emphasis on arable farming, and the effects of the rise of non-agricultural employment caused a drift away from this hamlet, so that by 1840, only the two farms remained, and the cottages had entirely disappeared.

Note: It has been suggested (REF 29) that in the Domesday entry for Marshfield, the ‘one hide owned by a priest’ corresponded with the Manor of West Marshfield, and that West End Town had been omitted from the Domesday account by accident. While this is an attractive hypothesis, there is no evidence for it. The Manor of ‘Marshfield’ was given by the Earl of Gloucester to Keynsham Abbey in c1168: as detailed above, in 1263, West End Town was owned by the Earl of Gloucester, presumably having been retained by his forebears for some reason in 1168.

The clear earthworks in this field, at ST 765742 are part of the former medieval settlement at West End Town. On the south of the site against the bridleway between West End Town and Brookhouse Green, is a large rectangular enclosure, marked by a bank and lynchet, about 30m x 60m, or possibly 30m x 80m if the area near to the south east corner of the site (and just inside the gate of the field) is included. Running across this area at right angles is the base of a stone wall, A.

This area presumably represents the site of a dwelling and its small enclosure, similar to, but not as large as, the surviving ‘Moonrakers’ house and enclosure on the east of this site, and to similar features shown on the 1768 map on the other side of the road. The rest of the area of the survey has a few low unclear earthworks, which may have building sites among them. At a later date, quarrying (at B) has dug away part of the site.

The western side of the site is marked by a low lynchet, separating the settlement remains above from the ploughed slopes beneath. This lynchet connects the end of the planned enclosure and that of the ‘Moonrakers’ plot. Over the whole area, cattle disturbances show the soil to be black, and quantities of medieval pottery have been found (see ‘finds’ below).

At the other end of West End Town Farm, slight earthworks at ST 76827423 in the corner of a field, indicate the site of a possible building and associated enclosure.

Other Earthworks

At ST 76227413, a large house platform has been much ploughed down. The last time this field was ploughed, about 15 years ago, the farmer noticed much stone, along with claypipes and pottery. The site is now pasture. It is shown as the site of a house on the 1768 map, and called Chapel Close on the Tithe Map of 1840: this building may very well have been the chapel referred to in 1584 (REF 13).

The houses shown on the 1768 map otherwise appear to have left no earthwork remains.

Finds at West End Town

As would be expected on the site of a deserted medieval settlement, finds were numerous when two fields there were ploughed. Some mention has already been made of this in the Roman section (see p. 33)ï When the field at ST 768743 was ploughed in 1983, the area was gridwalked in an attempt to locate the areas of medieval building.

The field had probably not been ploughed often before: a 1930′s oblique aerial photograph (REF 62) shows a linear bank inside the eastern wall of the field, and apparent earthworks inside this, which were just visible’in the ploughed field in 1983. They coincide roughly with one of the areas of medieval roof tile scatter (see below), but the other area lies in shadow on the photograph. A plot of these earthworks was made some years ago (REF 63), but it was clear when the field was ploughed that most of the features are due to dumping of stone.

The field was divided into a 10m grid, and the numbers of sherds of both medieval ceramic roof tile, and medieval pottery, plotted, with entirely unexpected and extraordinary results.

The numbers were not large, but the roof-tile plot showed two definite peaks, at approximately ST 76807430 and ST 76787425. A plot of total weight cf medieval ridge tile gave a similar result (not unexpectedly, as the size of the original rooftiles was probably broadly similar). However, when the plot of medieval potsherds was drawn, this revealed a completely different picture, with the vast majority (70%) of the pottery occuring in the 20m strip nearest the road, and with no peak in the distribution at all in the areas where the rooftile was found.

It is of course, possible that the roof tile fragments represent two buildings that for some reason, left no pot scatters to mark their sites; but this would be very odd indeed.

More likely is the possibility that the tile fragments (and associated stone scatters) are the result of dumping of building rubble at some date, to fill hollows in the field, and that the pottery is the true indicator of the position of the medieval building(s).

This would imply that the buildings at west End Town lay along the line of the present road (as indeed, the 1768 map shows those on the south side to do).

The status of the building in this area may have been high. Part of the tracery of a medieval window (14th/15th century) was found in the wall at ST 76907418 nearby; this would only have been found in an important domestic building (manor house?) or chapel. The rooftile fragments may also derive from the demolition of this building, and as further evidence, the field gridded was called ‘Court and CLose” in 1840 (often meaning “land by, or containing a large house’ – although it can mean ‘land by the cottages’).

Between the roadside edge of the field gridwalked, and the road itself, is a large area about 0.8m below the level of both, rapidly being dump-filled. Apart from the present farm building now standing, a second building was shown here on the Tithe Map of 1840. It is not possible to tell whether the manorial dwelling of West End Town was here, or on the site of the ostensibly seventeenth century West End Town Farm, but it was presumably in the immediate area.

In the other ploughed field, to the south of the road through West End Town, extensive scatters of medieval pottery were again found, mostly in two groups – the first around the barn at ST 76957415 in a rectangular area of black earth, and presumably indicating the site of a dwelling or enclosed garden, and the second at the northernmost tip of the field close to the road, again in a patch of black soil, presumably indicating occupation.

In the other fields around west End Town Farm, medieval pottery also occurs. A pipe-line trench to the north of the farm in 1982/3, in a field significantly named Stone Close, revealed c90 sherds of 12th – 15th century pottery, and cattle disturbance of the earthworks in pancras Close (see above) revealed c135, these figures clearly revealing the extent and archaeological potential of the area around West End Town Farm.

The settlement at Brookhouse Green is included here, although strictly speaking it is not clear whether it was regarded as part of West End Town in the past. As no early records exist (in any recognisable form,) no early name is known before the first mention of Gory Mead (see below).

Earthworks at Brookhouse Green (FIG 26)

A complex of undocumented earthworks occur in the centre of this field. between Brook House Green, and the spring that rises at ST 76087401.

Along the north side of the site, a lynchet marks the former course of the road running from West End Town to Oldfield farm, still visible over much of its course today, and shown as a road as late as 1768. This road ran around the northern and western sides of the field to a point where earthworks of a small quarry appear, presumably for walling or roadmending stone. At this point, the road branched, one continuation going to Oldfield Farm, and a second continuing along what are now hedgelines to the west end of Marshfield town. This is represented by a hollow-way 3m deep in places, but not included in this plan, beginning about 20m beyond the southernmost point shown.

Along the east side of the site, a low lynchet/bank runs, resulting from ploughing away of the soil from the lower side. This was presumably a field boundary at some time; it was not such in 1768 or afterwards, above and below it being the same pasture field at the time of the Tithe Map in 1840.

The rest of the earthworks are very confused and slight (although we were fortunate enough to survey in absolutely ideal conditions, with the grass reduced to no more than 1-2 cm high by sheep grazing). A appears to be the site of a building, as its rectangular shape, and traces of a dividing wall, imply. A slight holloway to its north is presumably the approach road from the north-west, and this seems to continue beyond the building to the spring.

The earthworks between A and the spring are presumably quarrying; they are deep, but irregular, and have possible dumped material within them. The three low lynchets to the north and west of A are presumably the result of ploughing, either as unfenced strips (and thus ?medieval in origin) or as part of a large garden to the building. The former seems more likely. It is just possible that A is only a quarry (although its regularity makes this unlikely); a small trial excavation is recommended to identify the nature of the site.

Earthworks at Brookhouse Green 2 (FIG 27)

This area of earthworks lies around a spring and a walled common called Brook House Green (ST 762741). The site has been investigated before, and a rather confused plan published (REF 63).

Within the common itself, a low earthwork (A) may possibly be a house site, but it is tiny and very difficult to plot. On both sides of the common, and across its centre, run deep hollow-ways, presumably being the feature followed by the later walls rather than vice-versa; they converge on the spring and broad shallow crossing-point of the Broadmead Brook on the east of the site. In the south of the site, in the pasture ground, there are exceptionally clear earthworks of a large building or house platform at B, shown by lynchets 1m high at back and front. This overlies two earlier earthworks, the lowest lying east-west, and the second almost north-south. This may represent three phases of re-building of the site. The north edge of this area is marked by a low stony bank, presumably the remains of a stone wall. Up the slope, and alongside the hollow-way on the other side of the present stone wall, are four flattish areas with steep banks below; these may be small house sites, but it is difficult to be sure as the area is heavily trampled and the soil has slumped badly. The lowest two lynchets may represent an old entrance trackway into the field behind the building.

Documentary evidence for the Brookhouse Green sites

There does not seem to be an old name for any settlement here, and presumably the buildings were included within either Oldfield or West End Town. A document of 17 03 (REF 60) refers to ‘…that roofless tenement and one close called Brookhouse … containing…seven acres … with Joseph Davis’s ground on its east…’. The deed is labelled on its reverse side ‘Allies tenement’. The term ‘roofless tenement’ is a legal phase which broadly speaking describes land where a house once stood: it is thus quite clear that one of the buildings above was abandoned before 1703.

It quite clear from another deed of 1703 (REF 60) that this refers to the second site, as it is referred to as west of Joseph Davis’ ground, which is itself west of a messuage or tenement called smiths Hold, which is quite clearly defined on the 1768 and Tithe Maps.

In 1733 a pasture called ‘Brookhouse Close or Gory mead…’ (REF 64) was sold to Sir William Codrington by Henry Harrington, and the name Gory Mead is traceable back to 1650, when John Gostlett was selling it to John Harrington, ancestor of Henry. The land is described there as ‘…Bramble and Gory Mead, near Oldfield…’ and therefore, refers to one of of the two Brookhouse Green sites. It is not possible to go further than this on the basis of presently known documents; nor has any pottery or other artefact been found on the site which would date the buildings.


Oakford has been briefly mentioned in the section on Roman Marshfield as a possible Roman occupation site. Like the other two sites of occupation in the south of the parish, it lies where a smaller stream joins the St. Catherines Brook, and had, at least in medieval times, a mill, although contemporary sources call this St. Catherines Mill.

The Roman occupation site lay about 50m below Oakford Barn; details of the finds are given on p34. Any continuity between this possible site and its undoubted medieval successor is unproven. There is no separate mention in Domesday of any manor of Oakford, despite its later status (see below), and the earliest mention so far traced is in 1224 ‘…..

Matillis, daughter of Robert de Okeford, petitions against Alfred de Marisfeld concerning half a hide of land, with its appurtenances, in Okeford…’ (REF 31). This half-a-hide is not traceable in the present landscape, and as the Oakford land-unit was formerly very much larger than would now appear, it seems insoluble.

Some hint of oakford’s medieval prosperity may be gained from the next reference, a jury list of 1247/8 (REF 65) in which the names William, Nicholas, Richard and ‘Ets’ de Hocford are mentioned under the manor of Marshfield. in the same document, only 3 jurors are named from ‘Marshfield’ (= West End Town, plus whatever lay on the present site of Marshfield, pre – 1264), and 3 from ‘Agmatha’ (very probably Ayford).

In 1331, John de Bagewode held a messuage and land in Oakford, which had previously been held together with land in West Hanham (REF 66), the majority of which had been granted by himself, and one William de la Greene, to the Abbot, and convent of Keynsham. Again, this messuage is lost, and untraceable today.

By 1554, a long grant of lands to New College, Oxford, includes (REF 67) – ‘…. a waggon-house of two bays called okeforde, and two closes of land and pasture adjacent….’.

Twenty years later, Thomas, Earl of Sussex, who was in possession of the manor of Marshfield, which he had received in 1572 for war service in northern England (REF 68), received a license (REF 69) to alienate ‘…the manor of Marshfield…Okeforde…(etc. etc.)… to John and William Gostlatt, and the heirs of William…’. It may be significant that in this deed, Oakford (along with Ayford, Ashwicke, west End Town, and a mysterious ‘Effoord’) are detailed separately to ‘Marshefelde’ itself, implying a degree of autonomy by this date at least. . Lists of these lands exist, but are not printed in the calendar of Patent Rolls, and have not been examined yet.

The following year, the Gostletts in turn alienated lands in oakford (together with others areas) to both Nicholas Webbe, subsequently of Ashwicke (REF 69), and to John Chambers, snr. and John Chambers Jnr., subsequently of Oakford (REF 69). A copy of this grant survives in the Gloucester Record Office (REF 70), and this details the lands granted. They include ‘…all that grange and farm lying in Okeford in Marshfield known as Okeford Grange,… and … one tenement, toft and half-yardland … held by Goslat of cullerne … two tenements, and two half yardlands … held by Thomas Wynyard … and two other tenements, and two half yardlands in Okeforde … held by Thomas Woodward … one messuage and tenement held by William Hopkens … (etc. etc.)…’ . This document tells much about Oakford. Although the tenements are clearly in oakford itself, only the ‘toft’ mentioned implies the existence of a house other than the Grange, although it may even be itself used in the sense of ‘land where a house once stood’. The implications seem to be of a desertion of the site of Oakford – save for the Grange – by 1575, perhaps supported by the ‘wagon-house’ being the only building mentioned in the deed of 1554 (above). This is also borne out by the pottery finds from Oakford (see p.-62), which decline dramatically from the 16th century onwards.

Thomas Woodward, mentioned here, is almost certainly a forbearer of the George Woodward ‘…of Okeford’ mentioned in the documents of 1630 (see below). This presumably implies that the family eventually took over the Grange, as no other messuage seems to have survived locally.

William Hopkens, who held the only messuage mentioned, (and who, with Thomas Woodward, signed the 1585 terrier of New college lands in Marshfield, Thomas Woodward in his capacity as churchwarden) gave his name to ‘Hotkins’, and ‘Hotkins Hill’ (ST 788709) on the Tithe Map of 1840, and a ‘cottage and garden in Hotkins’ are also referred to there (Tithe Apportionment No. 1041). This cottage, as clearly shown, still remains as a very overgrown ruin near the hydraulic ram in Oakford Lane, at ST 78927089 (FIG 56) . Earlier in this century, it was known as Hop House. While it is just possible that this name is a ludicrous coincidence, and the house was used for drying hops, a crop recorded in the Tithe Apportionment, it seems likely to again be derived from William Hopkins name.

In 1630, (KEF 12), Marshfield is noted as containing ‘…oakford, a hamlet, Okefords two little fields, are called fourth ffield, the other … (the name is illegible in the text)…’. A second reference, the New College terrier of 1629 (REF 13), gives in long, and painstaking detail, a confirmation of this fact, and the name of the other field as Rodlie ffield, now Rodney.

These two references make it abundantly clear that oakford, like Ayford, had its own two – field strip system, although by 1629, piecemeal enclosure was going on in Rodlie ffield at least.

The disposition of names on the Tithe Map, and earlier documents make it plain that oakfords land-unit stretched well into the lands of the 19th century Ashwicke Park (see FIG 33). Quite what the relationship between oakford and Ashwicke was, is unclear) few documents concerning Ashwicke appear to have survived (see p 80)- Foss Field, however, seems to have been largely to the north of the Marshfield to pixtongreen road, now an earthwork running from ST 78607200 to ST 79387133 across the former Ashwicke Park, but used as a main road to Bath up until 1840 at least. Rodlie field was in the extreme south of Marshfield parish, perhaps between oakford Farm, the Rocks, and the parish boundary, and now to a great extent, featureless pasture, although there are some large lynchets under the hedges of these fields, implying cultivation in the past.

By 1629, however, the actions of the Meredith family, who later built, and occupied the Rocks (q.v.), had begun. The southwood and southdowne of 1575 (REF 70) were being removed, and enclosed, and undoubtedly, this had a dramatic effect on the landscape of Oakford. Subsequent documents treat Oakford and the Rocks as two separate units, both owned by the same family, in 1686 (REF 71) for example, the lists of land owned with the Rocks (given on p 78) and those with Oakford Farm are separated. By this Oakford Farm had 40 acres in the still largely open Foss field, and twelve acres in Rodlie field, but the other holdings in ‘Rodley’ were now closes; Upper Rodley, and Bushy Rodley, for example, totalling twelve acres, and another close called ‘The Tyning above Rodlie field’. Although enclosure was still not complete, it had obviously been proceeding slowly on a piecemeal basis all through the 17th century, the usual state of affairs in Marshfield.

Unfortunately, the 1744 and 1768 maps of Marshfield do not include Oakford, and its first large scale map is the Tithe of 1840 (FIG 28). This shows enclosure completed, and ‘gentrification ‘ begun. Ashwicke Park has already begun to affect the Foss Field, and the Rocks Park is also prominent.

The ‘cottages in Motcombe Bottom …’ are best mentioned here. Although it is not clear whether the area around these cottages was within Oakford or not, the old hollow road from Ashwicke to Oakford ran down the hillside nearby, and the remains of two cottages survive as earthworks in the field here (FIG 32). They are apparently not mentioned in the Oakford lists but had disappeared by 1840. No dating is possible.

Earthworks at Oakford Barn (FIG 29) (ST 784701)

Oakford barn (shown in stipple) is a 19th century stone-built barn (the date 1844 appears on one roof-truss; and its layout is different from that on the Tithe Map of c1840), incorporating earlier features. It stands on a foundation of stone blocks up to 1 metre cube, possibly from an earlier building; a two-light, pointed-arch stone window of ? medieval date is incorporated into one end; a stone is built into the north≠west porch with the date 1731 and ‘Iohn C’ carved on it, and now upside down; two stones with sets of drilled holes are built in about 3 metres up the north-west wall in a re-used position; other stones bearing graffiti of one sort or another are re-ir.corporated in the present iuildir.g. llearly a stor.e structure :: s:re as =;-ewr.ere r.ear, and three other smaller agricultural buildings (all of 19th century build) are still standing close to the barn, although two lack roofs.

As can be seen in the plan Oakford 1840 (FIG ’28) the layout was once slightly different, and another building once stood between the present barn, and one of the subsidiary sheds. The buildings all stand in a yard surrounded by a stone wall about 1 metre thick, i.e. nearly twice as thick as the walls of the present buildings: perhaps this is a surviving feature of earlier building complexes on the site?

A further small square building stood at A on the plan. This is shown quite clearly on the Tithe map, and may have been a dovecot (the adjoining field is called Pigeon House Ground in the Tithe Apportionment). The site of this building is not traceable today.

Around the barn are a number of earthworks, of probable medieval -19th century date range. There is the tail end of a lynchet in the field next to the barn, which runs to an upright stone (marked ‘stone’ on the 1: 2500 map), but which is quite c1early an old gate post.

Below the barn are two shallow hollows partly from surface quarrying, although the eastern is large enough to be a house platform. Between these is a hollow-way, now filled in (1983/4) in its lower course, but still visible as a shallow ditch about 9-10 metres wide below the barn. This hollow-way formerly ran down the field, over a tiny stream at ST 78306996, then around the present southern end of Pigeon House Ground to join Oakford Lane. All the known sites of medieval occupation at Oakford lie along this hollow-way (see below), which is blocked by a wall where it enters the yard at Oakford Earn. Feature B is a convincing house platforms of ? medieval date. The two marked drops in level within the yard at Oakford Barn are odd: the southern may be connected with a track across the yard.

Medieval and Roman pottery (in small quantities) have been found to the east and south of the barn in mole-hills and tractor disturbance. Unfortunately, no earlier maps than 1840 exist, so the earthworks have no documentation, unless they are remnants of houses within the tenements referred to in 1575 (REF 70) .

Earthworks at Oakford (FIG 30) (ST 78326992)

This site is bounded on the south-east by Oakford Lane, and on the south and west by the track leading to Oakford Barn, which drops away very sharply on the west of this site. To the north of the site is a steep natural slope of about 5-8 metres, which yields a small quantity of 12th – 16th century pottery when trampled by cattle. The discovery of 12th -14th century pottery among the earthworks in small quantities here led to a survey.

To the east, the site of a building, with three or four courses of stone still standing on the north side, was revealed during earth moving in connection with the construction of the wessex Water Authority reservoir some years ago (REF 72), the east side of it being destroyed by those works, and the south presumably joining the fragmentary wall traceable along Oakford Lane under the hedge.

The earthworks at the western end of this site are less clear: they may be garden terraces, or perhaps the site of one or two more small buildings.

This plan is a sketch survey based on aerial photography and pacing. Although the earthworks are fragmentary, the interpretation is fairly clear, except where they are overgrown with scrub, depicted as light stipple on the plan.

Low down the slope towards Oakford Lane, at the north-west of the site, is a slight bank forming two sides of a rectangular enclosure, with a discernable entrance. This is of unknown date, but may be a garden or small field.

The massive lynchet at the centre west of the site (up to 4 metres high in some places) is larger than those associated with the medieval fields of the area; it may be a relict lynchet from aji earlier prehistoric field system.

A small square enclosure formed entirely by lynchets is in the south≠west end of the field, and is clearly visible from the bend in oakford Lane. This may be a medieval or post-medieval small cultivation plot -the footpath to the Rocks passes straight by it.

The southern end of the site is a large valley, fairly smooth, with just a few earthworks at its eastern, upper end. The main one of these is the entrance to a hollow-way running up the hillside through what is now a conifer plantation, to emerge in the fields near the Rocks.

The upper part of the field – the entire centre and north, and east areas – is taken up with low fragmentary lynchets on either side of a shallow hollow-way. There are almost certainly remnants of medieval strip cultivation, which was still continuing after a fashion in 1629 (REF 13), although amalgamation of strips and enclosure were well under way by then.

Earthworks at Motcombe Farm (FIG 32) (ST 787711)

Having been informed (REF 73) that there was a story of cottages in the valley below Motcombe Farm, the earthworks in the lowest field there (at ST 78687106) were carefully inspected, and seem to represent the remains of at least one, and possibly two buildings. They are well overgrown, but the remains of walls protrude at two places, and a large amount of cut building stone is piled on the far side of the wall from the earthworks, in open woodland. No cottage is recorded here on Tithe Map (1840), and no earlier records exist.

These earthworks are very close to a bend in the hollow-way which formerly carried the road from oakford to Ashwicke, which survives as a track and footpath to this day in places, but in others is completely overgrown (see below).

Finds at Oakford

There are very few ploughed fields at Oakford; the hilly nature of the area lends itself well to pasture. Opportunities for dating sites by fleldwalking finds are thus strictly limited. Apart from the finding of about 30 Roman sherds and perhaps 20 medieval sherds (all small and abraded) in mole-hills around Oakford Barn, no finds of significance were made until the hollow-way between Oakford Barn, and its crossing of the stream was in-filled, and its sides dug away for provision of a modern road in the winter of 1983/4.

At least one dwelling site was located as a dense scatter of pottery. This consisted of c600 sherds of pottery of the range 12th – 15th century in date, some very large, including about 80 rims and bases, mostly of ‘cooking pot” vessels, but also a few jugs and bowls. No ceramic rooftile was found. These virtually all came from an area about 20 metres in diameter (at 78327005) about 30 metres below the junction of the hollow-way with the field in which Oakford Barn stands, and at a point where three footpaths converge. Occupation may have been continued after the medieval period; however, only 22 sherds of 16th – 19th century date occured, including some combed slipware and later sgraffito wares .which may very well be intrusive. Other finds included two hones/ whetstones, one of pennant and one of ? Jurassic Limestone.

Of earlier dates, there were 18 worked flints, including one scraper of indeterminate date, and 36 sherds of Roman pottery, including 3 samian, 2 colour-coated fine wares, and a mortarium sherd, all appearing relatively unworn.

Without a doubt, the area of Oakford Barn, its environs, or the environs of the former hollow-way would repay excavation.

Oakford in 1600 (FIG 33)

There is just sufficient information for Oakford to attempt a reconstruction of the landscape in about 1600. This is not (and cannot) be guaranteed to be 100% correct, but the major details should be.

The road system was rather different to that of today. The Marshfield road past Ashwicke and Ashwicke itself (both removed during the emparkment at Ashwicke C1840) are well attested as late as the Tithe Map; the road connecting Foss Field and oakford, via Southwood, is represented on one edge of the 1768 map of Marshfield; the hollow-way connecting ‘Wm. Hopkens’ and Ashwicke, although presumably medieval, had probably not yet gone out of use; the track from Southdown to Oakford via Rodlie Field is not proven, but indicated by earthworks for at least part of its course; the road from Oakford Barn to the present road was present as a hollow-way until 1983; the short track up into Rodlie Field is a hollow-way in current use, and the remainder of the road system is current metalled highway. Conversly, the road from ‘wm. Hopkens’ to ‘Pickstones’ direct (present Oakford Lane) is not in existence as late as 1840; the road from Ashwicke Farm to ‘Pickstones’ is a creation of the 1830′s (REF 73) while the Rocks park and its avenue were planted around 1760 (p78).

The two fields, Foss Field (Voss, force, fourth, etc. etc.) and Rodlie Field, are located by their names appearing on the Tithe Map of 1840. This is not a completely accurate method, and the main problem is the area marked ‘?’ between Banderwells Wood and Foss Field. The earliest reference to this area is simply a name ‘Gaisons’ in 1840, presumably ‘garstons’ (= enclosed cattle pasture). It is not known what the status of this area was 240 years before; it may have been part of Foss Field.

The woodlands are fairly straightforward; both documentary (REF 12) and botanical evidence for the former positions of Southwood, Banderwell Wood, and a wood later called Dicknick Wood – its name in 1600 is not known – are clear. The only qualm I have about this plan is that there may have been more assarts into the south of Dicknick Wood, above Oakford, by this date. Certainly, there were by 1686 (REF 71), but these give the impression of recent clearance, a process that was still going on in 1840, but is now reversed by the coniferisation there.

The evidence for the settlements is given above; the medieval settlement of Oakford was probably almost deserted by 1600, but whether the. buildings in Motcombe Bottom were yet occupied is unknown.


Unlike Oakford (q.v.) no finds of Roman date have been made at Ayford, and, surprisingly, there seems to be very little medieval documentation. When, in its later history, it formed part of the manor of Marshfield, its story becomes a little clearer.

In 1247-1248, a jury list of the manor of Marshfield (REF 65) mentioned, along with four jurors from ‘Hocford’ and at least three from Marshfield, three more from a settlement called variously Agmath, Agmatha, and Agmtha. The full name is probably the second. The elements of this name are OE (ge) mySe (= mouth of river, watersmeet), and probably OE eg (= island, water-meadow). While it is just possible that this name derives from the name (Cinges) gemyd’e , applied to Beeks Mill area in the 10th century (REF 24), it seems more likely that the known settlement of Ayford is intended. The ‘eg’ element in local place-names is common, after all: apart from Ayford and ‘Agmatha’ themselves, there is also Nailey Farm (from an earlier field name ‘Nailyffeld’ (1676; REF 42) – almost certainly from OE atten eg leah (= at the water-meadow pasture or clearing) – and probably Aglescombe Common (form in 1584 is Eylescombe (REF 32)). It is thus not unreasonable to suppose a change of name to Ayford. if this change did occur, it had happened before 1574 (REF 69), as at this date, Thomas, Earl of Sussex, was disposing of the manor of Marshfield, along with ‘Okeford … Ayfoord … (etc.)’.

This 1574 deed throws up a problem, in the form of a place name ‘Effoord’, given as a place of the same status as Oakford etc. It is quite possible that this is a simple mistake, in effect ‘Ayfoord’ written twice; but later, it was associated with the mill at Shirehill, right at the other end of the parish. It is quite conceivable that this gave rise to the confusion between Ayford and Shirehill (or Players’) Mill; they are certainly not identical, despite a deed of 1723 to the contrary (REF 74).

At Ayford, however, the existence of a medieval predecessor to the hamlet is undeniable: pottery of 13th – 19th century date is found on many sites in the area.

In 1630, (REF 12) Ayford is referred to as ‘…Aiford, a hamlet, and Aifords two little fields…’, and a great deal is hidden in this laconic description. It certainly means that Ayford, like Oakford, had by this tine acquired its own common fields, maintained separately from the rest of Marshfield. Both at Oakford and Ayford, it seems extraordinarily unlikely that these fields would have originated in the 16th century, durring the sale of parts of the manor of Marshfield held by the Gostlett family, and more likely, they were formed in the early Middle Ages (11th century – 13th century?) to serve the small communities at both these hamlets.

In the case of Ayford, no names for the ‘little fields’ are quoted in 1630, but it is clear from constant references in the 17th and 18th centuries that one was called Eastfield, where many strip lynchets survive in good condition, illustrating how these remarkably steep hillsides were cultivated. The use of this area, of course, reflects the very small areas of available flat land for cultivation, which in turn, presumably reflects on the status of Ayford earlier in its history as a separate land-unit from Marshfield: if it had been part of the same land -unit, there would surely have been no need to cultivate this area for such a long period of time, and it could have reverted to pasture in the Middle Ages, as did ploughed hillslopes in Marshfield proper. These lynchets lie mainly around ST 779712, and are connected to Ayford by a trackway surviving as a double lynchet along the hillside to Ayford Farm.

The second field of the 1630 reference is unclear: it may be Nailyffelde, as mentioned in 1676 (REF 42). This name was later given to Nailey Farm (built between 1768 and the late 1820s, but using some older materials, such as its windows, which possibly came from demolished buildings on the site of Ayford hamlet), but there is some confusing evidence concerning the name Southfield. In 1676, John Goslett was selling a large block of lands, and ‘…a messuage or tenement called Bircombes…’ to Thomas Berratt. In this is reference to ‘…ten acres of land in South Field (? of Marshfield)’ but also ‘…3 acres of meadow in the Southfield by the Tyneinge at Sweetmeade’. Sweetmeade is a meadow next to the stream by Ayford bridge; presumably Ayford either had a Southfield, or else there was some confusion by this date as to where the South field of Marshfield extended.

The rest of the lands included a barn, (the earthworks of this, and the house are still identifiable), and orchard, two gardens, and closes called Home Close (4 acres), Rideinge (5 acres), the Tyneinge in Eastfield (2 acres), a meadow called Sweetmead (3 acres), 10 acres in the Southfield, 3 acres of meadow in the south field by the Tyneinge at Sweetmeade, an acre of arable in ‘Nailyffeilde, near grove Leaze, 5 acres and 1 farundell in Eastfield (that is, 4 acres and 1 farundell near Champion Hill, 1/2 acre by the Little Grove, H acre near Punters Gate, H acre in Eastfield Bottom, a wood (1 acre and 1 farundell) called Rideinge Grove, and a small coppice in Hillgrove).

By 1728, this tenement was being sold to Richard Bramble of west End Town, and although the house still existed in 1744, it seems likely that the sale was the first step towards abandonment. It was gone by 1840. The next year (1729), a roofless tenement (plot of land which formerly held a house) known as Hammons, with a sheep-house on it, was sold to John England, along with another messuage nearby, and a large block of land (REF 75).

Both sales imply that Ayford was already declining: by 1744, only eight buildings (probably houses), and a mill survived (FIG 34), and by 1840, only two farms, a house, a barn, and a mill; and today only two farms remain, farming the entire area of Ayford (FIG 36).

Although Ayford is shown clearly as a hamlet on the early county maps of Gloucestershire, and adjoining counties (such as the 1575 map of Somerset), its decline can be judged by the fact that in the late 18th century, even its name was changed, Tunnicliffe’s map of 1789 calling it Longley (the name of one of its fields), and Carys map of 1798 following suite.

The landscape of the area thus underwent several profound changes in the period 1400-1800. At the beginning of this period, it was probably a small hamlet with two open fields, as recorded much later, and bounded by the St. Catherines Brook, Egglescombe Common, Halldoor Lane, the woodland at the head of the Combe in which Eastfield stood, and the surviving boundary bank lying between the lands of Ayford, and oakford, which shows today as a straight hedgeline running from the St. Catherines Brook right up to the top of the valley side.

Almost certainly this layout was little changed by 1600, but around this time, enclosure had begun, as in Oakford (q.v.), and it is noticeable that although there were still arable strips in use in 1676 (REF 42), there were also many recorded closes, most of which seem to have been around the hamlet, or by the stream. By 1744 at latest, however, the whole area was enclosed, albeit mainly in large closes, and it is possible that strip cultivation may have lasted a while longer in the large closes – although there is no evidence for this.

The 1744 map probably marks the high water mark of enclosure, however, as many hedges had gone by 1840, presumably as a result of amalgamation of small arable fields into larger pasture units as the status of Ayford declined.

Earthworks at Nailey Farm (ST 772711)

This survey was carried out at 1:1250 by the Avon Historic Sites Team from Avon County Planning Department. I thank them for permission to reproduce it here in modified form (FIG 37).

The earthworks are dominated by two tracks: from their fork on the extreme east of the plan (where they leave the earthworks of Ayford itself), the lower runs first as a- hollow-way then as a double lynchet, across the south part of the slope to cross the small stream on the west of the area, close to a structure of large stones in the stream (? possibly a former sheepwash) at ST 77057104. The upper runs as a hollow-way (with enormous elder trees on either side), and then as a double lynchet trackway, around into the fields now lying below and to the west of Nailey Farm.

In between there lie three lynchets: the topmost marks the line of a hedge on the 1744 map, but the two below are not shown there: there are presumably either lynchets due to strip cultivation within the fields, or to hedges removed before 1744 (which really amount to the same thing!).

At the north-western end of the survey, a double lynchet marks the site of a double hedgeline in 1744 perhaps a track then: beyond this, and just before the tiny stream course, a low bank running down the field marks a field boundary on the 1744 map.

Practically all the field features in this survey, then, are accounted for by structures still in existence in the mid-18th century: even then, they were probably relict features of a medieval landscape.

This confused mass of earthworks (FIG 38) bears no resemblance to any features shown on maps of 1744 onwards, and thus are probably earlier. Having said this, they are mostly puzzling.

At the north of the site, immediately below the two farm buildings, a double row of large blocks of stone, together with three large blocks to their east, and a small section of hollow-way beyond this, indicate the course of a track shown on the 1744 map as leading into Ayford East Field, which survives for another 200m or so east as a faint double lynchet next to the hedge, and about 8 – 10m wide.

To the south of this, the earthworks have all been complicated by a shallow hollow-way which runs up through them from the direction of Ayford Mill.

What remains are three lynchets, varying in height from 0.5 – 1m, running across towards the break of slope on the east. There may be remains of strip cultivation in Ayford East Field (the area is shown as part of the field to the east in 1744 – see FIG 34), or even conceivably remains of enclosures connected with house sites, now destroyed by the buildings in the 18th century of Ayford Farm.

This small area of earthworks is also part of Ayford hamlet (FIG 39). The very large break of slope (up to 3m high in places) at the east, is a boundary to the garden or homestall mentioned (and shown) on the 1744 and 1768 maps (see FIG 34), while the western group are clearly the house platforms remaining after the disappearance of the building shown on the site as late as c1880 (REF 76). The site is very wet indeed today: presumably it was better drained while in use as a habitation.

Earthworks at Ayford (ST 773700) (FIG 40)

The earthworks at the main site of Ayford hamlet contain house platforms of various periods, and some garden earthworks, along with a few surviving walls. At the western end of the site, at A, a large wall of blocks up to 0.8m across forms the edge, with the earthwork running north from itswestern side, of an enclosure shown as a garden on the 1744 map. A fragment of the house shown in the enclosure on that map survives as a stone foundation to the north of the bend in Ayford Lane, and a break of slope to the west of this may also represent part of the house site. The picture here is confused, however, by a building of 1859 (B), Ayford Lodge, which was built over the earlier site, but now only survives as earthworks (about 1m high) itself. The date stone of this building is now at Nailey Farm.

Up the slope and behind this, a double lynchet marks the track which formerly split into two, and ran to Beeks (on the south) and to the fields west of Nailey Farm (on the north) . These are shown clearly on the 1744 map. House platforms survive at sites C – H, although only C, G and possibly E were still occupied by buildings in 1744, so presumably the others had been already abandoned by this date.

The other tracks on the earthwork plan are -not very clear. The track between B and D, a hollow-way today, and J, also a hollow-way, are both parts of unfenced tracks on the 1744 map, but the other track going north out of the site towards the future site of Nailey Farm on that map cannot be traced on the ground, presumably, the operations in building Ayford Lodge destroyed it.

The three hollows behind F are probably small quarries, and the two large ‘lynchets’ to the north of this must also have been dug although probably cultivated at a later date. The 1744 map also shows a small field, probably an abandoned house and garden, to the north of where the earthworks of Ayford Lodge stand today. This is not detectable in the earthworks. The linking of map and earthwork survey is very satisfying; the earthworks are as well preserved as any deserted medieval village in Avon.

Other points of interest in the comparison of the 1744 map, and the earthworks:

1) Ayford Bridge is not shown on the 1744 map; presumably the eponymous ford was still in use.

2) A thick wall against the road corner at C, under the hedge, may be a surviving wall of house C.

3) The finds of medieval pottery only occur in the centre of the Ayford site, at the stream crossing where extensive trampling by cattle has brought material of past medieval date to light, and on the rotovated soil below house site c. One further sherd was found in another part of Ayford site.

Ayford Mill (ST 77547086)

The origins of Ayford Mill are not known: the lack of medieval documents for Ayford means that no direct reference to the mill can be found until a deed renting land to Phillip West, miller at Ayford, in 1725 (REF 77). An indirect reference ‘…Mylards Way, leading from Marshfield to Eyford…’ in 1629 (REF 13) means the mill was almost certainly in existence then, and both the existence of a hollow-way track leading to the mill (implying long usage), and the existence of a local population in the medieval period, as well as the name given above, implies the existence of a mill here in the late 16th century at least, without excavation, the matter cannot be taken further.

All that can be said about dating of the features seen on the survey,is that the earthworks reveal a set of structures in existence before 1744, as they are clearly shown on the map of that year (FIG 41).

The building remains may be later. The mill itself could easily be 18th century or earlier: no structural features remain that make this certain, although some at least of the walls were reparied with brick at some date. The millhouse is clearly post -1800, and indeed, its plan in 1744 and 1840 show clear changes. These were probably due to a disastrous fire C1790 (REF 8) which a local newspaper claimed did £500 worth of damage, as far as is known, incidentally, this mill has always been a grist-mill: in the 1720s, Philip West is a ‘miller’; in 1790, the fire destroyed a quantity of corn, and the last use of the site (c187 5?) was for grist; one of the mill-stones from this period is in the garden at Nailey Farm.

The earthworks are as follows:- A wall of massive stones (each up to 2m long and 0.8m thick) was built in the St. Catherines Brook at the top of the site, raising the level of the stream and obviating the necessity for an enormous mill-leat. The stones still lie where they collapsed, in the stream bed. The water left the stream and ran along a leat to the millpond. This leat was probably not as wide as the current earthworks indicate: the bank on its southern side may only be remains of repeated scourings of the leat. The millpond (as indicated) was a structure standing above ground-level, with walls of large blocks up to 1.5m thick. From here, the water could take two routes. When not required, it ran out of the sluice at the north-west end of the mill, and into the stream again. The water that powered the mill machinery, ran across the building at its south-east end, and over to turn a wheel in a now filled wheelpit. This in turn drove machines inside the mill buildings, and the water returned to St. Catherines Brook.

The large platform a, is the site of a building, shown there in 1744 and 1840, and perhaps one of the buildings referred to in the will of Philip west in 1729 (REF 77). The millhouse is mostly a heap of rubble today: it was clearly a large building, as its 1840 plan shows (FIG 35).

In the extreme south of the site, where a track runs on out into the fields, a stone wall with a large break of slope below may be a loading and unloading platform; a flight of steps leads up to it from the mill house. The only later features on the site are the large early to mid 20th century rubbish dump (the lemon-shaped hump between leat and millpond) and the holes made in the millpond walls during heavy machinery movement in the past 20 years.

Ayford in 1600

As at Oakford, a reconstruction of the landscape of Ayford around 1600 is possible, although some of the fine
detail is questionable. (FIG 42)

Egglescombe wood (earliest known reference 1584) is largely connected with Beeks, and in the late 16th century, a~ least sixteen acres was owned by Thomas Crispe of Beeks. The south side of this wood at Ayford is probably what was destroyed by the England and Tipper families C1600-30, and it is possible that some of the area marked ‘ ?’ on the west of Ayford was woodland previously.

Beyond this was Egglescombe (or Aylescombe) Common, separated from Ayford by the deep gully of the stream running down the centre of the valley. To the north of Ayford, a high bank under the south hedge of Halldoor Lane runs all the way along the north boundary, probably created as a boundary feature.

Nailey Field probably had much the outline shown on the map, which is based on the areas of strip lynchets visible in the vicinity of Ayford hamlet, and including the flatter areas on top of the spur on which Nailey Farm now stands. Between Nailey Field and Halldoor Lane, an area is called ‘Old slait’ and ‘Ayford Hill’ in the 18th century. These may have been cultivated in 1600 but there is no definite evidence for this, and rough grazing is a good possibility (along with one or two small coppices on the steeper ground).

East Field is clear enough; references to it are unambiguous, as are the areas of strip lynchets contained within it. There are references to strips in the present closes between East Field and Halldoor Lane (now called Lippiatt), and these were probably part of the Field in 1600.

In the south of the area, the woodland (Rideinge Grove and part of Dicknick Wood) is well-attested into the 19th century. The line of hedge between Ayford and Oakford has a bank under it, with a standing stone c1.8m high on top, half-way up the valley side. It is clearly medieval or earlier.

The pastures alongside the St. Catherines Brook are not well-attested in the documents, although the existence of a direct communication between Ayford and oakford implies some early clearance, and Rideinge Wood probably derives its name from OE rydding (= a cleared area through woodland suitable for riding along), a name usually applied to early medieval clearances. The closes around the mill and hamlet are well attested in the documents; these were often used in areas with extensive open field systems for gardens and horticultural purposes.

The road system is slightly more debatable: the Marshfield – Ayford road, the roads into Eastfield, to the mill, and towards Egglescombe common are all deep hollow-ways, clearly medieval or earlier in date. The road to Oakford is as well, as evidenced by the place-names (see above). The roads towards Beeks and Egglescombe Common beyond the hollow-ways exist as double lynchet tracks, and thus are probably ancient too. The track across Nailey Farm is more problematic; it existed in 1744, so may be old, but this is unclear.

Ayford Hill was almost certainly in existence by this time – the road to it was called Mylards way in 1629 (REF 13), – and almost certainly owes its position to the presence of a tiny tributary of the St. Catherines Brook, which runs out of Eastfield bottom, and has had stone walling built into the sides of its course in places to direct it to the millpond.


This area, in the south-west of the parish, of Marshfield has an intriguing history, well worth recording as a separate unit in Marshfield, post – 1500. After the recorded prehistoric and Roman finds at Beeks or nearby (p 31 and p 34 ), there is no evidence of later occupation until the 16th century. It is however, close to the meeting point of the parishes of Marshfield, Cold Ashton and St. catherines/Batheaston; the charter of 931 leaves the common boundary of Marshfield and Cold Ashton at this point, and calls it Cinges Gemy e (Kings watersmeet), a curious choice of name for a meeting of two small streams. This is, of course, next to Beeks Mill today.

The next point in the recorded history is a licence in 1575 (REF 69) to John and William Gostlett who had possession of the entire manor of Marshfield at the time, to alienate lands and rents in ‘Marshefelde” to Thomas Crispe. In 1584, there is a record of the Gostlett family selling to ‘Thomas Cryspe, of Marshfield, yeoman … a messuage called Beekcs, another called Tylyes, a gryest mill next the last messuage, now in the occupation of William Tylie, also several closes of wood, pasture and arable … containing ten acres … in the tenure of William Gibbes … also 24 acres, 3 roods, 18 perches of wood and wood ground, of which 16 acres and 2 roods lie in Eylescombe …’. (REF 32). The messuage called Beekcs is clearly a forerunner of either Beeks Mill Cottage, or of the present garage and boat store across the road from the mill – probably the latter. Beeks mill, or rather its predecessor, was clearly being worked by a member of the Tyley family, who are still very much alive in Marshfield today.

None of the current buildings at Beeks appear to have any traces of 16th century architecture in them, neither was any medieval or 16th century pottery found in field-walking nearby, although the fields to the north of Beeks Farm did yield a respectable amount of 17th and 18th century pottery, but only within about 100m of the farm. A complicating factor here, however, is that in the New College terrier of 1584, reference is made to ‘….John Blanchard the younger… (and) … his farm of Bykes…’ What farm? If a messuage was part of the farm, could it be that he held the present Beeks Farm, and Crispe, Beeks mansion? Whatever this means, Crispes were certainly in possession of Beeks Farm by 1687, as they sold it in that year. The first Thomas Crispe died in 1601 (REF .7) ‘…seized of Egglescombe Wood’, although in 1630 (REF 12) it was claimed that at least eighteen acres of Egglescombe wood, belonging to Thomas Tipper, had been destroyed by Edward England sometime in the past 30 years. This confusion is insoluble at present – suffice it to say that the two tiny copses on the western side of Halldoor Lane, at ST 772720, and ST 773718, were still called Greater and Little Tippers Wood in 1840, and that the England family left their name in fields in the area between Nailey Farm and Beeks Farm – Long England, Little England, and Great England. Along the western, uphill side of Little England runs a low bank, about 80cm high and 3-4m across, which looks very much like a medieval wood boundary bank, so it is probably in this area that the destruction occurred.

A later Thomas Crispe, in 1687 (REF 78) sold Beeks Farm, and the holdings of George Russell, Thomas Berratt (plus 20 others) to John Vyner. This is of great interest, because slightly earlier, in 1662, a piece of ground (“five acres in Tacely’) is described as being ‘…on the south side of the way leading to Russells Mill’ (REF 27). The name is repeated in 1729 (REF 75) ‘…one acre in Tacely … against the highway leading to Russells Mill…’. While there was, of course, a road leading to the mill just across the parish boundary in Cold Ashton, which is shown on the 1744 map, the ‘Tacely/Tasly’ fieldnames do not reach it, whereas they straddle Beeks Lane. Russells Mill is presumably, therefore, an earlier name for Beeks Mill, perhaps after the George Russell as mentioned.

The farm passed through various hands in the 18th century, and it was probably during the course of this century that Beeks Mansion disappeared. This large house (see ‘Folklore’) was on or near the site presently occupied by Beeks Cottages, and an enormous hollow-way (10m across and 3m deep) leads from the yard of Beeks Farm to this site. No description or reliable picture of the house seems to have survived, although it is depicted conventionally on the ■ap of 1744 (FIG 43). If this exceedingly dubious sketch does not simple resemble contemporary ideas of how a modern house should look, it should indicate that it was of no great age in 1744, as the house depicted is probably late 17th or early 18th century in date. There remain a few physical relics of the house at Beeks. A small garden behind the pre ent cottages is retained by 4-6m high walls, rising directly from the pasture behind the cottages. There is one doorway leading into this garden with post-medieval mouldings on the joints; the head of the doorway is missing.

The cottages stand on a terrace some 2-3m above this garden (see FIG 44), and set into the wall supporting the upper terrace are two openings, one now built up with brick and mortar, and a second leading to a small vaulted stone cellar, approximately 3m high and 2m wide, (but with rubble infill on the floor), with a drain leading from it to a well in the garden. This is probably a small icehouse or wine-cellar.

There are a number of well-dressed stones built into the (? early) 19th century cottages now on the site, and it is of course, possible that much of Beeks Farm incorporates worked stones from the House. Nevertheless, it is surprising that so little has survived of this house, if it was really as important as folk memory would have it (see p. 95 )

One other scrap of information about the house is given by the existence of a large direction indicator of 18th century date in Beeks Lane at ST 76997270. This structure has carved letters on two sides, to be read from Beeks Lane, or from Halldoor Lane (north-west and north-east). Both sets of writing are about 3m above the level of the roads, presumably to be read about 3m above the level of the roads, presumably by coach drivers. On the north-west side, the letters are ‘To Beeks Farm’ while on the north-east, they are TO BEEKS HOUSE/Turn on the right track/Over the down/Drivers of carriages/are desired to/keep the road made/Over the down. The track ‘over the down’ (called Beeks Down on the 1744 map) is clearly shown on this, and the 1768 map, and is still used for part of its length as a field access track today. Possibly the ‘Druidical stones’ mentioned on the 1817 OS map (see ‘Folklore’) were markstones on this track, to indicate the way in snow.

One of the buildings at Beeks Farm has a ’1646′ date stone, but the house itself looks earlier, perhaps early 17th century, or even late 16th. Without detailed work inside, it is impossible to be sure.

Beeks Warren (ST 769716)

The site of the rabit warren was first identified in November 1982 (in early morning sun) from the field to the north of Nailey Farm. The low sun clearly picked out at least ten pillow-mounds: subsequent survey (FIG 45) identified 13 in the field.

The plan shows Trulls wood to the north-west, on a very steep east-facing slope. Below the wood, a long low bank runs right across the site dividing the warren from the wood, and presumably this once held some fence or similar structure to pevent the rabbits entering the wood. Immediately below this, the field itself lies on a fairly gentle east-facing slope, on the west side of the former Aglescombe Common.

The sinuous outline of the field reveals it as an early enclosure; it is certainly pre-1744, as the manorial map of that year shows it clearly, along with other features discussed below (FIG 46).

The mounds are low structures, none more than 1m high, and generally short, and rectangular, maximum dimensions 20m x 5m, and many clearly show the side ditches from which the earth to make them was obtained. One is smaller, and elliptical in shape; this may not be, but the others very clearly are, pillow mounds (artifical mounds of earth raised for the purpose of breeding rabbits) a fairly common landscape feature, but nowhere for miles around occurring in such numbers.

The fencing which ran around the site is not clear; there is no mound under the hedgeline to north, east and south – perhaps the rabbits were allowed onto the common.

At the western end of the site, a track leads to Beeks Farm, and earthworks here clearly show a small garden enclosure (also shown on the 1744, 1768, and 1840 maps) along with the footings of a small, two roomed building. A further building partially survives, buit into the bank in the western end of Trulls Wood.

The documentary evidence for this warren is conspicious by its total absence. The 1744 map (a simplified version is shown in FIG 46) shows a large enclosure from the commons belonging to Wm. Swymmer, who held Beeks at the time. This enclosure took in not only the warren, but also Trulls Wood, the field to the north of Trulls Wood (‘Dry Leaze’ in 1840), and an enclosure which is not lost in the two fields north-west, and west of Trulls Wood (called ‘Gunnings Leaze1 in 1840). The layout had become as at present by 1903: the area may have been enclosed during the 1850s. The field containing the warren is unhelpfully called Yeeles on the 1840 Tithe Apportionment.

The buildings in the western corner of the site are also recorded on the maps of 1744, 1868, and 1840 (‘a cottage, and garden’) and the cottage was lived in until the present century (REF 79). The second building is never shown, and may have been a recent outhouse or toilet. Finds of pottery from the tractor tracks near the cottage begin with one sherd of a medieval fabric (? 15th century), but those of 18th/19th century date are most common. Presumably, this cottage originated as a warreners cottage, although it is never called such.

The pillow-mounds themselves look relatively recent – earlier ones are usually longer. It is just possible, therefore, that the warren was a short-lived experiment in the late 19th century (the field was ‘pasture’ in 1840), but is more likely to have been earlier. In the absence of any further documentation, there is no point in speculating further. Although carefully searched for, no evidence was found of vermin traps, another possible indication that this warren is late, i.e. after the introduction of control predators (‘vermin’) by gun. The individual pillow-mounds were surveyed separately in order to make comparison with other sites easier.

Beeks Mill (ST 76237125)

The history of the mill has been dealt with above: it was certainly in existence in 1584, but none of the buildings are of this date. The mill buildings are only poorly preserved, and have been altered greatly since the mill ceased production in the 1880s.

The earthworks (FIG 47) show that a dam must have stood in the stream at A, to raise sufficient head of water to feed the short leat leading to the mill-pond. This was largely above ground, and the thickness of its supporting walls, and amount of rubble from its demolition show it was a large structure. Excess water escaped through the sluice next to the mill; the water to drive the machinery passed on, and into the building. Whether this was the original form of the mill is unclear, but certainly no earlier leat is visible upstream from the present faint remains.

In the building itself the water turned a wheel of c.5m diameter in the wheelpit at A, which drove machinery close by – possibly on a floor above, as a stone shute survives in the wall next to the water-feed to the wheel. The other parts of the building are “more modern, being pigsties built from re-used materials in the early part of the century.

Two mill stones survive, one as a doorstep into the building across the lane at ST 76267123, and one as a lower doorstep into the early 18th century house above, now known as Beeks Mill Cottage.

The building across the road from Beeks Mill has several 18th century windows.



The two houses now at Ringswell Common are late 18th or early 19th century in date, but they stand on an older site. There are no ploughed fields in the immediate vicinity, and so no pottery for dating.

The earliest reference to Ringswell is in 1575, when in the sale of Beeks Farm to Thomas Crisp, is included ‘…a messuage, and a half-yard (about 12-15 acres) of land at Ryngeswell in the tenure of Thomas Clement…’. It is clear that at this time, the lane leading to Ringswell is already built up around Pitt Farm, as both it, and a tenement on either side are mentioned in the above deed (REF 32). There is no further record of buildings at Ringswell until the 1744 map.

In the field next to Ringswell Common, however, are the clear earthworks of at least two buildings, on the steep south-facing slope west of the Common. These are superimposed on a pattern of medieval strip-lynchets and, so are presumably 13th century or later (FIG 48).

There do not appear to be any references to these buildings, and they had disappeared by 1744. The map of this date (FIG 49) shows two houses at Ringswell Common with a third building close behind one of them. The two houses stood further back from the road than the present buildings on the site, and remains should be detectable in the gardens there. These buildings (at ST 78437327 and ST 78417331) were replaced before 1840 with the present. The third building, in the field at ST 78327337 has also disappeared, leaving very vague earthworks behind. In 1840, another small building (probably agricultural) was at ST 78357334, but this has also gone (FIG 50).

Below the close, at ST 78377335, a spring is marked on the 1:2500 OS map, and this proved to have a walled spring-head around it, of a hollow ‘beehive’ shape. Although no early reference exists, this is almost certainly the Ringswell of the place-name.

Ringswell Common was first mentioned (as Ringwells Green) in 1713 (REF 46); it was a small area on either side of the former Marshfield – Ashwicke road, an important through-route before the 18th century. The northern part, next to the Ringswell houses, has several large ‘lynchets’ running across it.

These may be natural,or perhaps man-made, and pre-common; the flora of the common is very rich, and if it was ever ploughed, it certainly has not been so for many centuries. It is possible, of course, that these terraces have been accentuated by road traffic.

In the southern, and smaller part of the Common, on the south side of the Doncombe Brook, a hollow-way under the eastern wall remains of the old roadway, together with various other banks and lynchets almost certainly caused by road traffic. In the field on the east of this a series of parallel hollows run down the hillside to a stony bank about 1m high above the stream. At present, these odd earthworks defy explanation.

Permission to survey this area was not forthcoming.

There is no justification for any suggestion that Ringswell or its area were the original settled areas of Marshfield.

The Rocks

The earliest history of the house (ST 790705) and adjoining area which became the Rocks is obscure, and despite folklore suggesting a very early medieval military use for this building (p.96 ), there is no documentary or other evidence for any building on the site before c1600. Before this period, the area was almost certainly part of the Oakford land-unit, as both later documents, and the lay-out of fields and road system seem to imply.

A document of 1575 (REF 70) mentions Southwood, and describes it as on ‘the west side of Southdowne’. These areas are clearly identifiable as those now on the west and east side of the Rocks, but the text does not mention or imply a house or castle here; if a substantial dwelling did exist, it would have surely been mentioned in this deed. (Southdowne is presumably so called to distinguish it from Marshfield Down in the north of the parish).

Despite the demonstrable mid-17th century date of the buildings formerly at the Rocks (and shown in photographs before the demolition of 1957) there seems to be no documentary record of them until 1686 (REF 71) although there is a reference in 1630 to a ‘wood called Southwood of 40 acres … destroyed by … Michael Meredith …’ This was the family who held the Rocks, and it is just conceivable that this destruction may have occurred during, and because of, the initial construction at the Rocks (REF 13).

In 1686, the Rocks was firmly linked with Oakford, as it was in subsequent documents well into the 19th century.

The building of the Rocks presumably occurred, therefore, as a new Great House for the owner of the estate in the 17th century. Although this part of the house was believed to have been completely destroyed in 1957 leaving only the extensive 19th century embellishments, recent renovations by the present owner have revealed a complete 17th century dovecot, two storeys high, which from its external dripmouldings, must once have been free standing building, but which had been completely absorbed into the later alterations.

In 1686, the area between the Lane and the Fosse Way was called the Downes, (South Downes of 1575), and was described as ‘a close of 60 acres’. Together with the ‘Home Close’ of 30 acres, this made up the area known in the 19th century as ‘The Park’, and corresponding to the pasture area still recognisable as the land of Home Farm, and visible on the map as a discrete unit between the Rocks, and the Fosse Way.

By 1712, the Downes had become two closes of arable called Great and Little Downe (42 acres and 18 acres respectively), and these were subsequently divided up into the seven fields shown on the Tithe Map (FIG 51). The avenue of lime trees leading to the Rocks had two trees felled in 1982 when both gave ring counts of 220 plus 2 years, suggesting a date of c1760 for their planting. This may be confirmed by a stone in the wall against the Fosse Way bearing the inscription ‘E x R 1761′ (see p 96 ), possibly commemorating the building of this wall or its predecessor.

This is almost certainly the period at which the division of the Downes occurred.

Further landscaping of the Park occurred post – 1840 when new tree -clumps were planted (along the east, south, and north boundaries, and at ST 79457018 for example) and field boundaries again altered slightly. A large barn indicated on the Tithe Map at ST 79077040 disappeared, and the road leading from Oakford to the Rocks was connected to Pixtongreen. All these alterations were completed before 1903 (OS Map), but the emparkment has now given way to pasture farming once more.

The map of 1840 (FIG 51) is the earliest available of the buildings at the Rocks; the complex plot of gardens shown has mostly been overgrown or destroyed, the barn at A has been removed, and only a deep well and some amorphous earthworks remain to mark its site. The barn at B survives as part of Home Farm.

The road system around the Rocks is clearly influenced by the presence of the Great House; its connection with Oakford, however, is circuitous as opposed to the footpath between the two which is more or less direct.

As the Oakford – Rocks road is shown on the Tithe Map it runs up the present Oakford Lane as far as ST 78917082 where it branches off, and runs up the course of what is at present a hollow-way, under a 19th century ‘antique’ bridge (carrying a footpath from the Rocks garden past Home Farm), and up to Home Farm itself. The 2m depth of this hollow-way seems unlikely to have been entirely developed through access to the Rocks – presumably it formerly gave access to the area of South downe, and then to the Fosse Way. Apart from the avenue, the other redundant road shown as present on the Tithe Map is the extension of the way from the Ashwicke Road across the avenue at ST 79337039. The course to the south of this is still detectable as a stony bank running across the field in to the wooded area beyond, and may eventually connect up with the beginning of a hollow-way, near Oakford Farm (FIG 31) although its traces in the woodland between are by no means clear.

The new road from Oakford to Pixtongreen (see above) has a predecessor that ran from Oakford to Ashwicke (see Oakford). It also runs in a hollow-way up the side of the hill, which the new road conspicuously does not.

These hollow-roads presumably represent the medieval (and possibly earlier) communications around Oakford.


Because of the present land-use of this area (largely rough pasture, parkland, or otherwise land that gives no finds on fieldwalking), no sites of date earlier than 13th century are known here. The Fosse Way forms the eastern border of this area, however, and it seems unlikely that there would be no Roman occupation here, given the high density elsewhere in the parish.

Few medieval references survive for Ashwicke, but the site of the pre-19th century Great House (see below) has yielded a few large sherds of pottery of 12th – 15th century date, during alterations made by the present landowners.

Ashwicke, in common with the rest of Marshfield, was held by the Gostlett family in the late 16th century, being alienated in 1575 to Nicholas Webb, and John Chambers along with Oakford (REF 69). One of the documents (REF 70) refers to ‘…a messuage, tenement, and one yardland in hagh Ashwike held by Martin Osborne…’, and presumably High Ashwicke was the name applied to the settlement on the site of the pre-19th century great house.

As in the rest of the parish, the late 16th and early 17th century was , a time of great landscape change at Ashwicke. Between 1570 , and 1630, 20 acres of Razes wood were destroyed by William Webb, the then owner, for pasture, and 40 acres of ‘Pickstones Wood’ (near the site of the present Pixtongreen) were destroyed by the owner Charles Gostlett (REF 12).

At the same time, William Webb was enclosing areas of the former open field there. He enclosed 27 acres in 1590 (at ‘New Tyning’ and ‘Foards Breech’), 8 acres called Bowhens Breach in 1619, 23 acres at or near ïRoadyate’ – the position of this last in unknown – and other smaller areas.

The lack of other documents prevents any discovery of the process of enclosure at Ashwicke, and particularly, makes it difficult to discover when the assumed medieval hamlet of Ashwicke was replaced by the pre-19th century great house, although by analogy with the rest of the parish, a date around 1600 seems likely.

Whatever happened in the early post-medieval period, by 1840, when the first map of the area is available the area is enclosed, with woodland, fields, a Great House, arid its outbuildings,and a Home Farm of 17th century date.

This landscape was greatly disrupted in the period around 1840-5 when emparkment took place, resulting in (more or less) the present landscape. This involved re-routing the Bath-Marshfield road (which now exists as a fence alignment, line of trees, and hollow-way in farmland), demolishing the old great house and building a new (in 1859), laying out parkland around the house, with a new approach drive, erection of two lodges, and ornamental garden features etc., most of which still exist.


After the prehistoric, Roman, and Saxon periods, the history of the Oldfield area is obscure. As in so many of the river-valley fields of Marshfield, the sides of the Broadmead Brook at Oldfield have the remains of strip-lynchets running along them. In some areas, such as the field containing the Oldfield Roman site (Field 676; FIG 15), there appear to be two phases of lynchets, one parallel to the hill-slope, overlying an earlier which ran up the slope at an angle. These are not strictly dateable; apart from the two lynchets dated by the 931 charter, the other two sets must fall within the 10th -18th century bracket, although the fields are not shown as having cultivated lynchets by 1768, and are almost certainly medieval in date (10th – 14th century).

There were no finds of medieval material of any significance in the field-walking, despite the apparent importance of the name. The road system which formerly led to Oldfield from Marshfield Town (and was abandoned between 1768 and 1817) now only survives as a massive hollow-way, presumably dating from at least medieval times, although the earliest mention is not until 1651.

The earliest documentary record of Oldfield is in 1584, when it is simply mentioned in a list of lands held by the Rectory (REF 13); the earliest mention of Oldfield Farm is in 1650 (REF 79), and although many documents of the post-medieval period specify rights of common in ‘Oldfield and Backwells Corner’, the exact area concerned is unclear.

The earliest references concerning the landscape of Oldfield show great changes in the early post-medieval period; between 1570 and 1630, 80 acres of woodland had been cleared (Foxholes and Park Wood are mentioned – these may have been descendents of the ‘Aethelmodes wudu’ of 931; Park Wood was clearly in the immediate vicinity of Oldfield Farm – see below – but the site of Foxholes Wood is unknown, unless the names of the fields such as Folly/Fally around Pennsylvania (ST 745736) identify it; This would be in the correct location to identify with the Saxon woodland). The same reference (REF 12) adds that in 1625, Humphrey Hooke had enclosed 300 acres of Oldfield, this being quite an early date for an enclosure on this scale in Marshfield.

It is clear from a deed of 1650 (REF 79) that the Gostlett family who owned most of Marshfield from the 16th century onwards, also owned much of Oldfield. This document is a lease of much land at West End Town, plus “… 323 acres together with the messuage, barn, stables, and … (closes) … Oldfields Crofte ^8h acres),’ Backwell Corner (17% acres), land called Parkwood lying at the end of ground once wood called Park Wood (8a), (plus) the wood, and underwood there … Broadeclose (16a) … 4^ acres of former wood at the south side of the barn … also lands south of Oldfield … called Parke Orchard, Parkewood, Foxalls Wood, Parkmeade …’ (plus a number of other fields) between the Gostletts, and John Harrington of Kelston, Somerset.

Two years later, the “… lands south of Oldfield’ were excepted from a lease of the manor of Great Marshfield, when John Gostlett was leasing it to one Mr Erisey (REF 79).

In 1690, a further one year lease (by Mary Gostlett to John and William Harrington) mentioned all these grounds, and incidentally, mentions ‘… a ground called Reedscroft lying near Buggrow wall, on the north-east bounded by the field of Great Marshfield …” giving at least one of the recognised boundaries of the Oldfield area.

In 1729, the land passed to the Codrington family who held it until 1843 (REF 80, 81) and Codrington extinguished the common rights by careful leasing. In 1789, the Tunnicliff map of Gloucestershire shows Oldfield as a separate settlement, still with road access via Brook House Green, rather than as today.

The implications for the landscape _ of Oldfield are quite clear; during the medieval period, ploughing occurred, presumably a continuation of use from the pre-11th century landscape: in the early post-medieval period, the landscape changes occurring elsewhere in the parish (such as at Oakford, q.v.) were also happening here, and the early enclosure in 1625 – almost certainly the 323 acres mentioned in 1650 (see above) – was presumably by means of stone walls that still exist. All the landscape except the section around the A46 – then still open common – was enclosed by the time of the 1768 map.

The ‘Park’ names at Oldfield are intriguing. They occupy a small area between Oldfield Farm, and the A420, and the ‘road from Marshfield to Oldfield Farm’ marked the eastern limit of the area. The names include Park Wood, which was at least 20 acres in size, Parkmeade, Park Orchard and ‘Park Piece in Windmill Leaze’ which was in Cold Ashton parish, next to the area concerned. Two other ‘Park’ field names occur just across the parish boundary in Cold Ashton. These names may indicate the site of an enclosed medieval park, although there is no trace on the ground of the expected earthwork bank delimiting its area.


At the end of the Roman period c400 AD, the site at Ironmongers Piece (see above) after a short post-Roman occupation had no later occupation at all, and it is possible that this northern corner of Marshfield parish was entirely deserted for a long period.

It is of interest to note,however, that a group of ploughed fields, all between 100 and 300m from Shirehill Mill (ST 788764), each yielded a few sherds of medieval pottery on field-walking (12th-15th century), while the fields nearer to Marshfield town did not. This may imply some medieval settlement here; the pottery was all abraded, and almost certainly from manuring.

A reference of 1629 (REF 12) calls the site William Plaiers Mill, and the lane leading across the fields to it, Players Mill Way. The lane remained until it was taken into the enclosures around 1750.

A map of c1630 in the Gloucester Record Office shows a large bulding in an enclosure on this site, called ‘Steward House’. There is no building as early as this now on the site, and the name Stoward was later associated with a traditional cattle drinking place further downstream.

The only other pre-1840 reference is in 1723 (REF 74) when John Harrington of Kelston sold to Stephen Ford of North Wraaxall “… one copyhold messuage or tenement, and one water-grist mill known as Efford or Players Mill …* together with several pieces of land, for £430. Included in the lands are ‘… one close of pasture near the river at Players Mill …’ and “… one close of arable or pasture at the upper side of the mill adjoining the mill pond…’

This document involves the name Efford, also mentioned in a deed of 1585 (REF 69), and nowhere else. It seems odd that these two mentions should be made 150 years apart, if both are mistakes, and it is possible that Efford is yet another name for the Shirehill/Stoward area, lost because of the confusion with Ayford. Certainly, some other deeds call Ayford Mill (q.v.) ‘Ayford or Players Mill’!

A small area of land, roughly coinciding with the spread of medieval pottery, is called Shirehill, or Sherril, and is associated with the mill in the documentary sources, the earliest reference occurring in 1651 (REF 46), where it is called ïSherwell’.

No land near the mill is in arable cultivation today, and so no direct dating by pottery scatter can be made.

Today, the site consists of a farmhouse (20th century), and close to the stream, a large mid-17th century stone house, still occupied, in the centre of a complex of farm buildings (FIG 52). At the southern end of this complex, a Cotswold stone barn, skilling, and associated sheds are mostly early 19th century, but at the northern end, the millhouse and mill buildings of early 17th century date survive relatively intact- The house was occupied until fairly recently, but is now used for general farm purposes, and the mill is converted for use as a barn. Internal floors, the wheelpit, and the hole in the wall through which the water formerly entered to turn the wheel are all missing or infilled, and no machinery survives. The building is set into the hillside, so that the water enters the mill at ground level outside, but is c1.5m above the floor inside. The plans (FIG 53, 54) shows the ground, and first floor of the two buldings, and the added room at the western end. Some internal details of the house are visible, and much of the original roof woodwork remains, all typical early 17th century work.

FIG 55 shows the earthworks associated with the mill. Although there is a large pond below the mill, this is modern, and the original millpond was tiny, and survives only as an earthwork. This shows clearly in FIG 55, lying between the mill building, and the present course of the stream. Today, the stream runs out through the original overspill sluice, now overgrown and damaged: parts of the stone setting for the sluice doors is now in the pool below.

The mill leat runs for 230m along the hillside; from a very low bank at the north-east end, it gains in size until at the south-west end, just before the water enters the millpond, it carries water a good 2m above the original stream course, which still survives as a wet area in the field below.

The wet area, and leat stream rejoin close to the mill, and just beyond this point, a series of massive stepping-stones (up to 1m square) carry a footpath across the stream. The water which powered the mill originally left by a culvert which is still visible in the garden of the house, although now largely dry. One millstone survives in the garden, and a second behind one of the sheds in the farmyard.

Apart from the mill, only the woodland at Shirehill is really of landscape history interest: this woodland is botanically very rich, and may be a surviving relict woodland, although early documentary references are lacking. There do not appear to be any giant coppice stools in the wood, but this is general in Marshfield.


In a parish such as Marshfield, the economy has always been based on agriculture, or related industries. It is often said that Marshfield owed its medieval prosperity to wool, and while there may be some truth in this, it is also certainly true that the growing of barley and malting have been underestimated in their importance in the Middle Ages.

The effects of this industry on Marshfield’s landscape is really indivisible from the usual arable farming landscapes; large fields, late enclosure (in this case) having ensured that many retain the curve; shape of the medieval fore-runners in the open fields.

The industrial sites themselves I not survive well in this intensively farmed area (FIG 56), and the malthouses which formerly stood behind the houses in Marshfield have likewise disappeared.


Malting is the process of convert barley grain into fermentable material. The grain is soaked, allowed to begin to germinate, and then killed and dried in special premises called ‘malthouses’ which have some means for spreading out the wet germinating grain, and pasing hot air through it.

Malting in Marshfield is documented over 500 years ago; in the Calendar of Patent Rolls, 19th November 1458, a pardon of outlawry for not appearing at court on a charge of trespass was given to Richard Porter of Marshfield (town) ‘maltman’ – he was impleaded with Roger Freman and John Herford, both also described as ‘maltmen’. (Oddly, someone who is almost certainly the same person as the last (John Colerne, alias Herford, late of Merschefeld, ‘yoman’) later appears in 1477 on a charge of debt of £20 in Missenden (REF 81). Richard Porter also was in trouble again later – he received further pardon of outlawry for breaking a contract of employment in May 1460, when described as a ïlabourer’ (REF 82). These three extracts point up that malting was probably seasonal, with maltmen having other occupations in other parts of the year. (This was not always the case; Richard Herford of ‘Marchefelde’ was described as a maltman when arraigned for burglary in May 1549) (REF 57).

In 1608, (REF 83), 21 maltmen were recorded out of 147 males quoted, over 14% the able-bodied male population, and more numerous than any other category except ‘labourer’ (21%). Various references to ‘maltmen’ can be found in deeds from the mid-17th century onwards, and the industry seems to have been booming in the 18th century, when certain individuals are referred to many times. Joseph Woodward, for example occurs from 1726 to 1759 at least, during which time he also farms extensive arable lands, and in 1728, one Joseph Barker rents “… the east part of a messuage … the Hall and Chamber with the lofts over them, and the malthouse adjoining and also part of the outlet and backside … sixty-three feet long, sixty and a half feet wide, and the well in the brewhouse belonging to the … cottage on its east, and the passage to the well …” (REF 59).

This indicates that at least some maltsters may have been small brewers as well, although this is not well documented.

By 1779, when Rudder wrote his County History, malting had ‘formerly been a great business, though now declining’, and he added that the Tuesday market was mainly frequented in the malting season. Unfortunately, the map of 1768 does not mention malthouses individually, the first clear map being after some further decline, in 1840, when the Tithe Map mentions 18 malthouses (see FIG 57).

Interestingly, the Woodward family connection continues, and although the earliest mention so far found was in 1692 (Richard Woodward, REF 44), in 1840, Isaac Woodward (in partnership with John Golding) held land, rented the house and malthouse at 59, High Street – now the Post Office – at the rear of 92, High Street, and at the rear of 3, Market Place, as well as owning the White Hart and Crown Inns, and they also ran a brewery. The success of Woodward & Golding was such as to elevate them in the 1858 Slaters Directory of Gloucestershire to be numbered among the gentry!

Malting appears to be in rapid decline at this date; by 1870 only 8 maltings and 1 brewery remained (that of Woodward & Golding) divided among five maltsters; and by 1897, only 3 remained.

The process and industry which provided much of Marshfield’s wealth for hundreds of years has left little physical trace; most malthouses have completely vanished, and the many fragments of floor tiles of malting kilns which are found in fields around the town testify where they have gone.


Two tollhouses formerly stood in Marshfield, of which only the west survives. It is a hexagonal – fronted stone building, added to a previously existing structure, and stands at ST 77237370. A second formerly stood at ST 78227382, the site now occupied by the Police Station. Its original board, giving prices, is preserved in the Tolzey Hall, at Marshfield.

Of the milestones along Marshfield’s roads, seven survive.


On this, the London road, milestones 103 and 104 from London survive. 103 at ST 78237381 is now a cast-iron plate reading ‘Chippenham 9 miles, Bristol Bridge 12 miles and 1 furlong’ set on a concrete stand, while 104, at ST 76517355 is a stone block c1.5m high, with an inset for a cast iron plate. This is missing.


One milestone survives on this road, re-set by the road verge at ST 74677457. It is a late 18th century
ashlar block with a segmental head, and round-headed cast-iron plate with the legend ‘XXVI to Cirencester XVI to Tetbury VII to Bath’.

Tormarton Road

Two ‘mileposts’ remain on this road, At ST 78427473, a low stone ‘milepost’ stands, about 0.8m high and 0.5m square. It has no inscripton. The field next to it is called ‘Half Mile Stone1 on the 1840 Tithe Map .

At ST 78507532, the second stone stands on the east side of the road outside of the farm entrance with ‘Rushmead Farm’ cut in it. It is a triangular slab – dm high. Two fields about 20 0m north are called Milestone and Milestone Piece in 1840, so possibly the milestone has been moved.

Beeks Lane

At ST 77647331, a small defaced milestone stands on the west side of the road junction. This is not a mile from Marshfield, but may have been a direction post.

At ST 76247124, a small milestone is made from a block of ashlar. This has a deep, square groove down the back, and carved roughly in the front ‘Marshfield 2′. This looks like a locally carved stone.

Direction Posts

1. The 18th century direction post in Beeks Lane, leading to Beeks House and Farm has been mentioned in the section on Beeks.

2. At ST 797708 a field is called ‘Direction Post Ground’ in 1840. This presumably refers to a structure on one of the two junctions on the Fosse Way nearby, although no trace of it remains now.

Boundary Stones

1. At ST 74767484, a small permanent slab, round-headed c0.8m high, with WB 1793 D & H carved on the west face.

2. At ST 76187519, a small sandstone, round-headed slab, c0.8m high, with the front face scaled away.

3. At ST 79737456, a ‘stone’ recorded on the modern 1 :2500 map (1973). This was not examined.

4. At ST 79637394, a boundary stone is recorded on the 1:25000 map, but not on the 1973 1:2500 map. It is no longer there.

5. At ST 76037339, a small ashlar slab about 0.8m high, just inside the fence, still survives.

Other Features

Two windpumps were recorded in the parish. One, at ST 78917165, still stands more or less complete, and a second at ST 77227473 has now been removed.

Ashwicke House formerly had its own private gasworks, although the site of this is unknown, and nothing appears to remain of it.

In Oakford Lane, around ST 78917087, are a hydraulic ram, which since its installation (c1910) has pumped water up to a reservoir at the Rocks, and the remains of a waterwheel seating which formerly performed the same function. Part of the wheel still remains in situ.


These (Shirehill Mill, Beeks Mill and Ayford Mill) are described in
their appropriate sections.

Quarrying and Limekilns

In an area where virtually all buidings are built of local Jurassic limestone, many with roofs of limestone slabs, and many field boundaries are local stone walls, much quarrying of local stone must have occurred.

Small quarries for local building and/or walling purposes are legion; for example, there are large hollows in the fields behind West End Town Farm, which are almost certainly quarries of unknown date.

A few quarries are larger and more significant:-

1. Hoopers Quarry (ST 78457465) lies by the east side of the Tormarton road, and is first recorded as Hoopers Quarr in 1725 (REF 77). Although there are further references in the 1720s, the quarry is not shown on the 1744 map. The field is knownas Hoopers Quar Piece in 1840, and consists nowadays of a filled-in-hollow about 10 0m across, showing as a stone scatter when the field is ploughed. 16th and 17th century pottery was found here during fieldwalking.

2. Oldfield Gate Quarry (ST 74757470). This is shown by the side of the turnpike (A46) road on a map of 1817 in the Gloucester Record Office (REF 84), and at this time was a roadstone quarry for the Commissioners of the Bath and Cirencester Turnpike Trust. It is shown on the 1903 6″ OS map, but is now returned to agricultural use.

3. Oldfield Quarry (ST 74837420). A long rocky ridge in this field is the remains of a quarry shown on the 1903 6″ OS map. This was presumably a roadstone quarry, but is now returned to agricultural use.

4. Eeeks Lane (ST 76677208). A small depression, about 30m across, and now being infilled in a field called Quarry Tyning in 1840, is almost certainly a quarry for walling stone.

5. Home Farm (ST 784738). There were undoubtedly small quarries here: the enclosure was called ‘Butthayes and the Quarrs’ in 1732 (REF 85). Now infilled and built over.

6. Downthorns Farm (ST 79687503). A large hollow, 40m across, by the Down Road, being infilled in 1982/3.

7. Pixtongreen Quarry (ST 791712). A large quarry shown here in 1903 was probably for freestone, but is now infilled.

8. Rocks East Woodlands (ST 79527090). A quarry shown here in 1903 has now been infilled (in 1950s).

9. Rocks Home Farm Quarries (ST 79257054). The south quarry is shown on the 1903 6″ OS map, but is rapidly being infilled. The north has an underground quarry leading off to the north, which connects to an air shaft nearby (REF 86).

10. A large quarry was made in a field adjacent to Castle Farm (ST 772744) at some stage in the 19th century. It was unused by 1903, and may have simply been for walling stone, or for rubble stone for the building of Castle Farm.


Two limekilns are known to have existed in Marshfield, but neither survives today.

1. One shown at ST 79247299, near Woodleaze Barn, is shown on the 1881 25″ OS map, but only the quarry hollows, over an area of 60m x 2 0m, remain. The construction of this kiln may have been linked with the 19th century reconstruction of Woodleaze Barn.

2. The field at ST 764720 was known as Limekiln in 1840, but not before (‘Taceley’ in 1768). No remains are visible there.

Other Industries (FIG 57)

Small rural industries must have been legion in Marshfield, but again, most have disappeared: the undoubted medieval woodworking industries which would have existed to use the woodlands, not least those in the valley of the Doncombe Brook, and which left little or no trace behind them (except a possible sawpit at ST 79857195 in Razes Wood) had dwindled by 1840 to William Jefferies’ timber yard in East End next to the churchyard.

William Downs occupied a cooperage at 67/68 High Street in 1840, and there was a blacksmiths shop, run by John Hulbert, at the rear of 12, Hay Street.

Several other industries and trades are listed in (for example) Pigots Directory of 1842, which gives, say, George Andrews, hat maker, at the rear of 5, Sheep Fair Lane, and although many of these smaller trades have left no trace (and few of the actual premises remain), they give at least some picture of the trade and industry of the small rural town nearby 150 years ago.


Military structures are common in the countryside, whether they date from Neolithic period, or from any date between them and the second World War. Only recently has the study of structures of the 1939-45 war become archaeologically respectable; it is sobering to think that these structures, which have little or no documentary record, and along with the old aerodromes and military camps of the period, reflect an episode of great historical importance and its effects on the contemporary landscape, are being destroyed indiscriminately without any record being made.

Marshfield was quite important during the years, being as it was, right next to the aerodrome at Colerne, and fairly close to Charmy Down. This meant that searchlight and gun emplacements would undoubtedly have been made in the area; a collection of local information and reminiscences of the period would be a great contribution to local history, and it is with the intention of providing an impetus to this end that this chapter is included.

The survey recorded several structures of the period during fieldwork, and made some notes concerning the existence of several more, but undoubtedly, many of the sites were of a very temporary nature, and without local knowledge of the period would be mistaken for agricultural structures. (FIG 38)

Only two sites were surveyed in any detail during the 1982-4 period, and one of these was destroyed literally days later.

1. Ashwicke Park site (FIG 59). In the strip of woodland next to the Fosse Way by the East Lodge of Ashwicke Park (ST 798717) is a collection of structures. They include at least eleven rectangular concrete hut – bases, an air-raid shelter, a round brick built rubbish burner, and two small slit trenches of unknown purposes. This site is not overgrown.

2. Ashwicke Park Drive site. Close to the above, and running beside the drive into Ashwicke Park, the survey recorded seven underground structures, presumably air-raid shelters, in 1983. These have since been demolished. They were rectangular or square in plan, and brick built, with reinforced concrete roofs (FIG 60).

3. At Oldfield, two rows of stone-heaps at right angles in the large field on the top of the hill at the west end, appear to have been blocks
to prevent the landing of aircraft.

These are common on Mendip. The rows at Oldfield were present in 1946 (air photo) but appear to have been removed shortly afterwards.

4. At ST 75277448, next to the Middledown Road, a small square enclosure shown on the 1950 6″ OS map, but now removed, was the site of a searchlight battery. Two Nissen huts formerly on the site were removed to Middle Down Farm by the farmer, and are still there (REF 87).

5. Near the Rushmead Farm crossroads at ST 785 752, a decoy aircraft lit by ‘flashing lights’ was deployed. No physical trace survives.

6. An earthwork platform, about 3m square, at ST 79287123 was pointed out to the survey as the site of a searchlight battery. Near as it is to Colerne aerodrome, this does not seem unreasonable; the plaform is certainly later than the medieval strip lynchets on which it is constructed.

7. A small building at ST 79467275, on top of Henley Hill, may have been constructed as a watching post, connected to the aerodrome.

8. In the woodland at ST 724730, an extension of the runways of Colerne aerodrome onto the Marshfield side of the Fosse Way survives as a series of short metalled tracks in what is now conifer plantation. In the aerial photographs of 1946, these are shown as leading to hangars (the demolition rubble from which is still on the site), and the runways are crowde with ‘mothballed’ aircraft.

9. Next to the bridge crossing the Doncombe Brook (and strictly, just outside of the parish) is the only pill-box recorded during the survey.

It is hexagonal, with reinforced concrete roof, and was presumably put there to guard the narrow valley road at this point.


During the course of our fieldwork in Marshfield, we collected many Stories about archaeological sites in the parish. This, of course, is nothing new, even for Marshfield: Aubrey recorded tales in the 17th century concerning the large (and now destroyed) bell-barrow called St. Oswald’s Tump on Marshfield down, and Rudder, the Gloucestershire County historian repeated the stories concerning St. Pancras Well, at West End Town, in 1779.

Some of these stories proved fruitful, especially those concerning the area about West End Town; those that were not, conformed on the whole to well-known motifs.

While no mention at all was made of West End Town in the CRAAGS publication ‘Historic Towns in Avon’ (REF 36) (the Marshfield section selecting Little End, at the eastern end of Marshfield Town, as the probable original site of Marshfield), several sources immediately informed us that ‘West End Town is the original Marshfield’ or ‘Marshfield began at West End Town’ or similar. The existence of the town place-name element could be enough to give rise ïto such a story, but see below. Other West End Town stories were:

1) That there had been a Chapel which was confirmed by the field name St. Pancras Close and a Holy Well called St. Pancras Well at the west end of West End Town Farm (confirmed by fieldwork – in fact, it is a spring which rises in the lane). One informant said that the reason nobody could identify the site of the chapel was that there had been a church on the other side of the road (this was not confirmed by our work!). The former existence of chapels, priorities, abbeys etc. is a common folk-lore motif, often used to explain the presence of prominent earthworks that are clearly those of buildings, as at West End Town, and so the definite references we found to the existence of the chapel were interesting. Its late survival as “… a building known as the Chappel …” to 1729 at least may help to explain the persistence of the story, especially as the field name ‘Chapel Close’ also remains, close to West End Town Farm.

2) That a ‘golden calf was buried somewhere at West End Town. The golden calf/coffin story is more usually associated with prehistoric monuments (see below) but examples are known of its association with Christian sites (as at Congresbury, Avon, where St. Congars bones are supposed to be buried in a golden coffin beneath the ancient yew-tree in the churchyard there.

3) That there had been a Manor House at West End Town (for evidence of this see ‘West End Town’).

4) That at Brookhouse Green, nearby, there ‘were once cottages’ -confirmed in our survey ( p.56 )

5) That the track from Oldfield Gate (ST 747748) to West End Town was once used by pack mules bringing corn from Dyrham to be ground at Beeks Mill. This is very likely to be true: the direct route from West End Town through Green Lane (ST 773735) to Beeks Lane is quite a plausible medieval mill-way.

The area of Beeks Mill and Beeks Farm has its own crop of legends, mainly concerned with Beeks Mansion. This was a great house, formerly standing roughly on the site of the present Beeks Cottage (ST 76617143), where the present garden in front of the cottages is pointed out as the exact site of the house. The building is historically well-attested, although its date of destruction is unknown; it must have been between 1768 and 1840 however, as it is shown on the map of the former date, but not on that of the latter. Among the stories concerning this house were:

1) That it burnt down. This is a persistent story, and may well be true, although the few remaining areas of stonework are undamaged by fire. These were not part of the house itself however.

2) That the mansion had twelve doors, fifty-two rooms, and three-hundred and sixty-five windows. Again, this could be true: Laboured symbolic designs like this do occur in the post-medieval period – but fifty-two rooms?

3) That a tunnel connected the mansion with Marshfield church. Such stories are the most common folk-lore of all those connected with ancient buildings; despite the occasional surprising example of long tunnels existing (as at Kelston, Avon) these are usually exaggerations based on the existence of features such as large drains, or here possibly the ? ice-house at the back of the present cottages.

Other stories exist about the area of Beeks Farm. There is supposed to be a golden calf buried nearby: this is one of the three reputed to exist in the parish. There were reputedly ‘Druidical stones’ near Trulls Wood, just to the north of Beeks Cottages. These were indicated in Gothic type on the 1817 OS 1″ map, but no trace exists of them today, unless the stone by the side of the track leading from Halldoor Lane to Beeks Farm (at ST 77037242) was intended. There are other large stones in the field to the east of Trulls Wood (the rabbit warren; FIG 45), but these are almost certainly natural blocks from the outcrops in the wood. Possibly these stones influenced later suggestions that Egglescombe Common (1575) which formerly surrounded this area, was a name derived from the Welsh ‘eglwys’ (a church). No trace, documentary or archaeological, can be found of this church, and in view of the fact that at this date, the reference gives the name as Eylescombe Common, it may be that the name is derived from an original Old English – eg form (= water -meadow or island). The really early forms that would settle this do not appear to have survived.

At Bullshills (about ST 77107308), there is reputed to be a double mound (one large, one small) that is the grave of a man and his horse ‘from the battle of Lansdown’. Although the site of this mound was clearly indicated to us on a 1 :2500 map, and had been apparently clearly visible in the 1920′s, we were unable to find any such mound. The site today is a steep hillside, heavily overgrown with blackthorn bushes, so mounds of very low profile might be missed. The battle of Lansdown is still remembered at Marshfield: several relics, including a large cannonball, are reputed to be from the battlefield.

The Rocks (ST 790705) has its share of folklore (although curiously, the third Marshfield great house, Ashwicke Hall, does not appear to).

Despite the Rocks almost certainly being post-medieval in origin (post-1584 to be precise), there is a tradition that it was a ‘Welsh marcher castle’ called Southernwood Castle, and that an iron beacon hanging over the entrance to the house (shown there in photographs of the early 20th century, and now preserved at Castle Farm Folk Museum, Marshfield) was lit to provide warning of marauding Welsh. The first house on the site was called Southwood House in 1686 (REF 71), the name presumably deriving from Southwood, first recorded in 1575 (REF 70), but no military use is recorded at any time in its history.

Stories of a secret tunnel running from the now overgrown garden at the Rocks, under the Rocks itself, and on to reach Home Farm, are at least partly true; a tunnel about 2-3m high runs about 35m into the cliffs from the garden, (in an easterly direction), and at its innermost end, appears to be blocked by a fall as it rises towards the house.

A stone built into the wall of the former boundary of ‘the Park’ at the Rocks (against the Fosse Way at ST 79647015) bears the deeply-carved legend ‘ExR 1761′. This is supposed to mark the spot where a highwayman shot and killed a traveller. This is, of course, possible, but a more mundane explanation would be that this stone marks the building of the present wall, or a predecessor; this date is exactly that of the trees in the avenue which leads from the Fosse Way to the Rocks ( p.79 ), and may indicate the date of one phase of the emparkment of this area.

At the south-eastern corner of the park at the Rocks, also the south-east point of the parish, and of the old county of Gloucestershire, are the Three Shire Stones. Even academics have been guilty of romanticising these stones; it has been claimed that they are the remains of a burial chamber recorded on Bannerdown by Aubrey in the 17th century. Mr Hayes of Marshfield (pers. com. ) has collected the stories (and remarkably inaccurate descriptions) about the stones, and also quotes the bill of 1859 expressly stating that they have been quarried and set up in that year.

One of the sites surveyed in 1982/3 was that of the cottage(s) in the valley below Motcombe Farm (FIG 30). These were found after reference had been made to ï… a story about old cottages in Motcombe bottom …’ Whatever these buildings were, they certainly disappeared by 1840.

The former barrow cemetery on Marshfield Down (p .18 ), which was destroyed in 1947, contained a large bell-barrow called locally ‘St. Oswalds Tump’. John Aubrey recorded in his topographical collections that earth had been taken from the mound to work cures on man and beast: he drew a sketch-plan of the barrow, recorded the legend that St. Oswald was crucified there, and recorded on his plan the stone on which Oswald kneeled, and that to which he was fastened (REF 91). Bede records a similar story about the site of St. Oswald’s martyrdom, and the battle at which Oswald died (Maserfeld) has been identified with Marshfield in the past. This identification is not now generally accepted, but it is curious that in 1234, the abbot and convent of Keynsham applied for a charter for a fair of three days on the vigil, feast and morrow of St. Oswald. As fairs often coincided with the festivals of the saints to whom local churches and chapels were dedicated, this may indicate that some identification of the two sites was already being made (REF 34). The hollow from which earth had been taken was, incidentally a prominent feature until the bulldozing of the mound.

Other stories about this mound -that a party of men trying to level it were scattered by a thunderstorm, and that a golden calf is buried there -are fairly typical folklore motifs concerned with barrows.

The other mounds in the cemetery (at least seven, perhaps more) were locally known as ‘The Soldiers Graves’ (REF 92).

A few scattered fragments of folklore also exist: REF 92 records a story associating Oakford with the oak under which St. Augustine met the British bishops in 603, but this sounds more like antiquarian speculation than genuine folklore.

A few field-names in the parish indicate possible folklore, although we did not record any tales connected with them: Hangmans Ash is a field on the Fosse Way; Cuckoo Park is supposed to be the field in which the Cuckoo is first seen, (as perhaps is Cuckoo Tyning), but these names are often corruptions of ‘cuckold’ – implying some clandestine sexual activities in the field!

There are probably many more items of folklore to be collected concerning Marshfield’s historic sites; their value, if treated very cautiously, should be self-evident from the above.


Unfortunately, like many other parishes formerly in South Gloucestershire, Marshfield is poorly documented before 1550. This problem is compounded by the fact that Keynsham Abbey was the major landowner in Marshfield in the Middle Ages, and Tewkesbury Abbey the next largest, and that virtually every medieval document relating to these abbeys has disappeared. This means that all the early forms of field names, settlement names and so on, which are vital to the correct interpretations of their origins, are not available. Some work has been done on the county records in the Public Record Office by the English Place – Name Society, but this has not proved very fruitful, and any earlier versions of names published by the EPNS (REF 30) occur in the Codrington archives in the County Record Office.

The main name in the parish, Marshfield, is recorded in Domesday Book (1086) as Meresfeld (REF 29), and the EPNS volume suggests a derivation from O.E. ‘maeres-feld’, a ‘tract of open land on the boundary’. This presumably refers to a county boundary or similar; the existence of a road called ‘Maerweg’ in 931 (= the boundary road) perhaps implies that the road is also named from this boundary. Derivations from OE ‘mersc’ (= marsh) seem unlikely, as very little of Marshfield actually is marshy today, except in excessively wet weather. A manor of ‘Meers’ is occasionally quoted from Rutter (1779) as being related to the parish name, but there is no other record of this, and Rutter may have been confused by a similar name in Falfield parish.

West End Town, the ‘second settlement’ of the parish has a complex place name history. It may have been the original Meresfelde; certainly by 13 03, it was known as Wesmaresfeld (REF 47); in 1549, it was ‘Weston, alias Westmershefeld’ (REF 57), and it had become Weston Towne by 1644 at the latest. These names all refer to the position of the settlement, west of Marshfield town. At the same time, it was known as ‘Litle Marshfield’ – in 1387, for example, it was Littelmarsfeld (REF 89), and in 1574, it was called Little Marshefeld, alias Westnershefeld (REF 69).

Ayford and Oakford both contain the Old English element ‘forda’ (= a shallow rive cross place); the name Ayford is discussed under the section describing its settlement history (p.65 ); Oakford means exactly what it says, ‘the ford by an oak tree’, the earliest known reference being in 1224 (REF 31).

Ashwicke is clearly explained by its earliest form (REF 30) – Ayswyk – the farm (‘wic’) near the ash-tree (‘aesc’), recorded in 1287. It is occasionally recorded as High or Higher Ashwicke, implying the existence of another, lower settlement which is not documented, and was not found during the survey (unless it is represented by the cottage remains below Motcombe Farm).

Some of the farms in the parish also have old names. The ‘farme of Bykes’ for example, is mentioned in 1584 (now Beeks Farm) (REF 13), and presumably this is the name of a former owner.

Oldfield is also first recorded in 1584 (REF 13), and Oldfield Farm is referred to as “… Oldfields Croft, with the messuage, barne and stables …’ in 1650 (REF 79), and the name means precisely what it says (see p.81 )

But the vast majority of the place-name evidence for Marshfield is in the form of field-names. Many of these are recorded on the Tithe Map of 1840, and some earlier versions of others can be found.

An analysis of the 819 names recorded in the survey archive, showed some interesting factors involved in the naming of fields.

179 of the names (22%) are simple ownership names. Many major and minor place-names begin as ownership names, and the late enclosure of much of Marshfield, with subsequent disruption of traditional ways of naming or describing land, have meant that many of the fields have the names of relatively recent owners, and sometimes those of families still living in the parish.

Long England and Little England, for example, are names only recorded relatively late, on the 1840 Tithe Map. They are fields near Ayford, and their name is due to one Edward England having owned them in the 17th century – in fact they were probably part of the 18 acres of woodland he had destroyed between 160 0 and 163 0 in Aglescombe Wood (REF 12). Many similar names exist: Paynes Corner, for example, is recorded in 1584 (REF 13), and was the name of a furlong in the North Field in 1725 (REF 77), and Hosey Mead, Doctor Smiths Ground, Lites Leaze, Tippers Wood, and Fords Breach are similar names.

The topographical names are next most common (13.9%). They include examples such as The Coombs (ye Combes 1629 REF 13), and Henley Hill (Henly hill 1729 REF 75) which are self explanatory, the ‘hen -’ element in the second simply meaning ‘bird’. ‘Head’ or ‘top’ are recorded, as in Egglescombe head (1629; REF 13) or Harcombe top (1840), meaning ‘the place at or near the top of the area named’ thus Egglescombe head was a tiny field by Halldoor Lane, at the top of the valley in which Egglescombe Common lay. Similarily, Northslade bottom (1692,- REF 44), meaning ‘piece of land in Northslade in a valley bottom’ is recorded.

Names from land-marks are common, too (9.6%). Such examples as ‘Monument Piece’ – the field next to the direction post in Beeks Lane, showing by its name that the post was no longer recognised as such, and was thought to be a ‘monument’; Milestone piece, and Mill acre are all recorded in 1840, all referring to a prominent nearby landmark. This common form of name can sometimes give a clue to the existence of a landmark no longer visible, such as Chapel Close at West End Town, or Ruxleigh cottages, formerly Rux poole, or Ruggs poole (1728; REF 59) referring to a pool drained before 1744.

Land use is a common element in field names, and sometimes informative. Names such as Oxleaze, Cowleaze and Horse Ground are all found in Marshfield, and simply imply pasturing of the named animals in those fields at some undefined date in the past. Washpool, a name at West End Town, refers to the existence of an artifical pond in the Broadmead Brook for the dipping of sheep. The name is at least 140 years old, being recorded on the Tithe Map, and some of the structure still survives. French ground (although occasionally an ownership name) usually means ‘field sown with French grass’ (sainfoin), a leguminous crop popular in the 17th and 18th centuries, and a case is recorded under West End Town.

Burnt hill, or Burnt bake, are almost certainly names recording an agricultural practice involving burning: the second refers to the practice of paring the soil off, burning it, and ploughing in the ash, a common way of breaking up pasture in the pre-fertiliser era.

The name Vincents draught (1692; REF 44) or Vincis throat (!) (1728; REF 59) is recorded at Rushmead Lane before enclosure of the area; it is almost certainly a corruption of ‘drift’ (= a way along which cattle are driven), and refers to the Greenway, along which cattle were driven to water.

Field size name elements are also common in late enclosed landscapes, and Twenty-seven acres, Four acres. Three acres, and so on are common here, forming 7.3% of the total names.

Other categories of field names involve field shape (e.g. Square leaze. Shoulder of Mutton ground, or even Long ground); names recording former administrative and legal status of land (e.g. Innock 1840, the Innocke 1585 (REF 13) = a piece of land temporarily enclosed for some purpose, from ME ‘inhoke’, or Hitchin 1840, Hitchings 1687 (REF 78) ‘ a piece of land on which a different crop was grown from that in the rest of the (open) field” from OE ‘heccing’), soil descriptions (Honeylands. Poor Hill, Clay Hill, or Filhorn are all fairly self-explanatory).

They may describe plants or crops growing in the area; for example Taceley 1840 is Tasly 1629 (REF 13), and almost certainly is from Old English ‘taesl-leah, meaning ‘the ground where teasels grow’ or Withy Mead, from Old English and current dialect ‘Widig’, meaning a willow tree. One of the earliest names of this type in Marshfield, mapoldor leage (931) is referred to on p Crops include oats (e.g. Oate lands in Foss Field 1629 (REF 13), sainfoin (see above) and possibly peas, although the tiny, narrow nature of both fields called Piss over in 1840 implies that their size, and not their crop of peas, is being described(!)

Other names of possible historical interest include Cripple Hole (= a field with a hole in one wall through which sheep can pass), a lost field name ‘Chestlehulk’ meaning ‘the shed in the field with very stony soil’. ■Chestle’ as a name element sometimes refers to ploughed-up building stone of a buried archaeological site, or sometimes simply to a gravelly or stony soil. There is no way of checking this name (which is first recorded in 1725) as its position is not known (REF 90).

Some names defy explanation, and any earlier forms coming to light may eventually help. One such is ‘Dicknick’, first recorded in 1686 (REF 71), as a close of wood and coppice, although the name seems to have been attached to the woodland above, originally.

‘Stoward’ is also extraordinary. It earliest known appearance is in ‘Stoward House’, appearing on the early 17th century Marshfield map in the Gloucester Record Office, and it is consistently so spelt. It is clear from later documents that the name is actually attached to a traditional cattle-watering place on the Broadmead Brook.

The enigmatic ‘Fuddlebrook’ suffers from too many possible interpretations, rather than too few. It appears to be derived from ‘fuddle’ (= to intoxicate) as is suggested by the EPNS, but the earliest form ‘footell brook’ (1629) (REF 13) may imply some other meaning. It certainly seems extraordinary to describe a stream as intoxicating, but unless an early form of ‘footling’ (= tiny, worthless) is intended (describing the tiny headwaters of the Doncombe Brook, which rises here), the name seems inexplicable.

A full record of all place-names found during the survey is preserved in the archive, along with provisional interpretations.



The normal method of fieldwalking arable land was as follows. The team lined up along one edge of the field, approximately 5 m. apart (although this distance varied slightly) and walked across the field picking up all the artefacts, including recognisable modern objects, within one or two paces either side of their track. By far the easiest way found was to walk along the plough furrows, thus making it easy to concentrate on the collecting of artefacts.

These were then put into ‘Minigrip’ bags carried by the individual walker, and the bags of finds pooled at the end of each strip walked. If there were few finds in a field, they were usually all pooled at the end of the walking of that field.

If any dense scatters of artefacts (e.g. potsherds) or obvious building materials, etc. were seen, the area they originated in was again inspected, and if the amount seemed worthwhile, the field was later gridwalked (see below).

Each bag of finds was labelled with a plastic label (white plastic sheet ïin 5 cm. squares) marked with waterproof felt pen (black ‘Pentel’). This proved to be quite adequate for identification purposes, only three bags of finds being eventually unprovenanced from the two years work

- and those not greatly important ones

- and only one bag lost, which was put right by re-walking the field in 1983/4.

During 1982/3, additional details like weather, time taken, state of soil and light, etc., were recorded, and an analysis of these will appear elsewhere.


On four Roman sites, and one Medieval site, the amount of artefacts on the surface was deemed sufficient to attempt to identify sites of This was carried out as follows:

A base-line was laid out, usually along one edge of a field, and a grid of 20 m squares (for the Roman sites, or 10 m for the Medieval site) was laid out. This grid was measured with two tapes, the rapidly evolving pattern of marker-canes allowing visual checks on accuracy. In general, accuracy to within 20 cm. was easily achieved. A sketchplan of the field was then drawn, and each square given a letter-and-number reference (B4, hi etc.). One person then controlled events. Each square was walked for 10 minutes the one person in each square again picking up all artefacts to be found in that 10 minutes. As they walked, the controller issued water-proof labels to each walker in turn as well as the bag for the next square. At the end of the ten minutes, the bag was put down and the persons moved to their next square.

At the end of each session, all the bags were collected up, and taken away. Only one of the sites (field 91) required two days work, and no interference with our grid-markers was experienced.

This method proved to be very successful indeed.


The normal method of fieldwalking pasture was as follows. The team, having been trained in recognition of features of archaeological interest (banks, ditches, old walls, large stones, etc.,) were instructed to first walk over the field, noting any features of interest.

When such were seen, a sketch plan of the field was drawn, from the 1:10,000 or 1:2500 map, and the features drawn in by eye. (“sketch surveying’) This was usually supplemented by pacing, and sometimes compass bearings. Usually, most features proved to be identifiable with map or aerial photograph features (and thus could be plotted very accurately onto a 1:2500 map), or if the site proved mo-re than usually interesting, further surveys could be carried out.

In the first year, details of weather etc. were recorded (as for arable-walking).


Any sites of earthworks or other structures thought worthy of more detailed study, were surveyed with tapes, using a tape baseline and 90 offsets. Initially an optical square was used, but this was found (in the hands of non-experts!) to have unfortunate capacities for error and 3:4:5 triangles were used to make tape right-angle offsets, and eventually, it was discovered that for lengths of no greater than 3 0m. using the corner of a drawing board for establishing the right-angle was perfectly satisfactory, so long as back-checks were made every so often. This proved easy, quick and accurate. Overall slopes in sites (generally less than 10 ) were ignored as of little consequence.

Initially, one or two surveys were carried out at 1:1250 using scale rules, but this proved grossly inadequate for most sites, and eventually 1:400 was our preferred scale, using graph film under the normal plastic field-recording film, for ease in recording and drawing the measurements made.

The field copies were drawn in pencil, and any notes (such as heights of lynchets, etc.) added, and hachured there and then. At the end of each survey, the whole site was re-walked with the plan to make certain nothing had been missed.


There were generally restricted to measuring of plans with tapes, usually at a scale of 1:100. Frequently, they were recorded in ‘science books’ as use of drawing boards was unnecessary. Photographs were used to record details of buildings.


These were walked in winter (when the undergrowth was low) in much the same way as pasture ground, as far as possible marking down such items as woodbanks, pre-wood earthworks, etc. In general, due to the very overgrown nature of most of Marshfield’s woods, fieldwalking was unrewarding, and areas coppiced or felled should be examined for earthworks etc. wherever possible.


Please note that all references to the map of 184 0 are to the Marshfield Tithe Map, in the Bristol Record Office. All references to the maps of 1744, 1768 and 1630 are to the manorial maps in Gloucester Record Office, (D1610/P34; D1610/P35; D1610/P18; D1610/New College microfilm). GRO, WRO and BRO are used below as abbreviations for the Gloucestershire, Wiltshire and Bristol Record Offices.


1. Findlay 1976 ‘SOILS OF THE SOUTHERN COTSWOLDS’ (Soil Survey of Great Britain)


3. Grinsell, L.V. 1958 TBGAS 77, 151

4. Grinsell L.V. and O’Neil H. 1960 ‘GLOUCESTERSHIRE BARROWS’ IN TBGAS 79



7. Rudder, S. 1779 ‘A NEW HISTORY OF GLOUCESTERSHIRE’ (1779)

8. Avon County Sites and Monuments Record in the County planning Department.


10. Aston, M. and Burrow, I. 1982 ‘THE ARCHAEOLOGY OF SOMERSET’ (Somerset County Council)

11. Leach, P. ‘ROMAN SETTLEMENT. IN SOUTH SOMERSET’ (Unpublished thesis)

12. GRO Dl610/New College microfilm

13. GRO D1610/R5

14. Knight, M. Westland Farm

15. Fowler, P. In ‘EARLY LAND ALLOTMENT’ (British Archaeological Reports)

16. GRO D1610/R5

17. Cunliffe, ed. 1979 ‘EXCAVATIONS IN BATH 1950-75* (C.R.A.A.G.S.)


19. Gracie, H. and Price E.G. 1979 ‘FROCESTER COURT ROMAN VILLA: SECOND REPORT) IN T.B.G.A.S. 97, 9


21. Anglo-Saxon Chronicle.


23. Aston, M. Pers. Comm.


25. GRO D1610/T2 0


27. GRO D1610/E621

28. GRO D1610/M3

29. Moore, J. 1982 ‘DOMESDAY BOOK: GLOUCESTERSHIRE’ (Phillimore)

30. Smith 1964 ‘PLACE-NAMES OF GLOUCESTERSHIRE’ (English Place-Name society)

31. Curia Regis Rolls _1J_(Public Record Office (P.R.O.))

32. GRO D1610 T25

33. Curia Regis Rolls 8_ (P.R.O.)

34. Calendar of Close Rolls (C.C.R.) Hen. 3 (1231-4) (P.R.O.)

35. GRO D1610/T18

36. Leech, R. 1975 ‘SMALL MEDIEVAL TOWNS IN AVON’ (C.R.A.A.G.S.)

37. Placita de quo Warranto (P.R.O.)

38. Potto-Hicks 1933. T.B.G.A.S. _55 249

39. Calendar of Patent Rolls (C.P.R.) Hen. 5 (1416-22) (P.R.O.)

40. C.P.R. Hen. 6 (1452-61) (P.R.O.)

41. GRO D1610/E74

42. GRO D1610/E62

43. GRO D1610/T21

44. GRO D1610/E63

45. Calendar of Inquisitions Miscellaneous (C.I.M.) (P.R.O.)

46. GRO D1610/T26

47. Feudal Aids 2^ (P.R.O.)

48. Adams 1976 ‘AGRARIAN LANDSCAPE TERMS’ (institute of British Geographers)

49. C.C.R. Hen. 3 (1261-4) (P.R.O.)

50. Calendar of Inquisitiones Post Mortem (c.I.P.M.) J3_ (P.R.O.)

51. C.I.P.M. 5 (P.R 0. )

52. C.I.P.M. 9 (P.R.O.)

53. C.I.P.M. (First Series) Volume 3. (P.R.O.)

54. C.I.M. 6 (P.R.O.)

55. C.P.R.Hen. 7 (1485-94) (P.R.O.)

56. G.R.O.D1086/M1

57. C.P.R. Ed. 6 _2

58. G.R.O.D1610/T32

59. G.R.O. D1610/E70

60. G.R.O. D1610/E64

61. Russett, White and Morgan (unpublished) ‘THE PIT AT CASTLE FARM’

62. Air photo in Avon county Planning Department.

63. Williams 1981 ‘WESTEND TOWN, MARSHFIELD’ in ‘BARG Review 2′

64. G.R.O./D1610 T38

65. T.B.G.A.S.

66. Calendar of Inquisitiones ad quod damnum

67. C.P.R. Phil. & Mary 1554-5 (P.R.O.)

68. C.P.R. Eliz. 1569-72 (P.R.O.)

69. C.P.R. Eliz 1572-5 (P.R.O.)

70. G.R.O. D1610/T24

71. W.R.O. 529:8/59

72. R. Earl, Oakford Farm pers. comm.

73. G.R.O. Q/SH 1832 C

74. G.R.O. D1610/T29

75. G.R.O. D1610/E71

76. ‘The Rocks Estate’ – map lent for copying by Mr p. Headley-Smith, The Rocks.

77. G.R.O. D1610/E69

78. G.R.O. D1610/T30

79. G.R.O. D1610/T20

80. G.R.O. D1610/T23

81. C.P.R Ed 4 (1476-85)

82. C.P.R. Hen. 6 (1452-61)

83. Smith, J. (1608) MEN AND ARMOUR FOR GLOUCESTERSHIRE Alan Sutton 1980

84. G.R.O./D1610 P38

85. G.R.O./D1610 P72

86. Pollard, D. (pers. comm)

87. Frankcom, A. Middle Down Farm, (pers. comm)

88. G.R.O. D1610/T27

89. C.I.p.M 16

90. G.R.O./D1610 T54


92. Typescript history of Marshfield, loaned for copying by Mr J. Andrews, Marshfield.


This survey would not have been possible without the co-operation and interest of the farmers and landowners of Marshfield: the families of Adams, Andrews (R and J), Andrews R, Anstey, Ball, Bond, Brown, Charnaud, Church, Creed, Crooke, Drew, Driver, Durnell, Earle, England, Fishlock, Frankcom, Fuller, Fry, Gardener, Godwin, Greenland, Guild, Harris, Hawking, Hitchcock, Hunt, Hurford, Jarrett, Knight (of west Littleton), Knight M, Knight R, Lane, May, Mays, Miflin, Park, Philips, Pullin E, Pullen L, Pullin N, Rawlings, Reynolds, Robinson, Saville, Smith, Taylor, Tiley, Watkins, Wayn, Winter- Alsop and Worlock. The following institutional landowners also readily gave permission to the survey: Avon County Council, Marshfield Parish Council, the former Ashwicke Hall Estates, Fountain Forestry and the International school of Chouiefat.

Only one landowner refused permission. The staff members of Avon County Community Environment Scheme (ACCES) who worked on the survey were Tina Barnett, Pete Buckley, Albert Chivers, Ally Minch, Graham Eelbeck, Steve Kingston, Nick Morgan, Andy McElwee, Andy Sutherland, Sharron Tennant, Alison Toyn and Hilary White who was deputy supervisor.

Many other people gave help and encouragement including Harold Hayes, John Humphries, Anne Everton, David Dawson, Mike Ponsford and Terry Pearson. A special word of thanks is due to Dick Knight who himself has done much fieldwork in Marshfield and readily gave his findings to the survey team. He also helped in establishing a list of landowners in the parish.

The survey was carried out on behalf of Avon County Planning Department by MSC workers of ACCES, funded by the Manpower Services Commission and sponsored by Avon County Council.

Please note that the majority of the Sites mentioned in this book are privately owned and have no general public access.

Prepared by Avon County community and Environment Scheme for the Avon County Planning Department, designed by the Graphics Section, Avon County Planning Department. Published by the County of Avon Public Relations and Publicity Department, P.O. Box 41, Avon House, The Haymarket, Bristol. BS99 7NF.

ISBN 0-86063-249-0
PRICE £3.60