<<

02 Mudd:Layout 1 03/07/2013 08:49 Page 3

A and Complex and its Early Medieval Reuse: Excavations at Netherfield Farm, South Petherton, Somerset, 2006

andrew mudd and mark brett With contributions by Hugo Anderson-Whymark, Stuart Black, Sarah Cobain, Jonny Geber, Frances Healy, Nathalie Marini, E. R. McSloy, Elaine L. Morris, Kathryn M. Price, Sylvia Warman, Geoff Warren, Nick Watson, Keith Wilkinson and Tim P. Young

An early Neolithic causewayed , a middle Neolithic long enclosure and an earlier Bronze Age open enclosure were among a group of prehistoric features discovered and examined by excavation at Netherfield Farm, South Petherton during archaeological mitigation work ahead of the construction of a natural gas pipeline between Ilchester and Barrington, Somerset, in 2006. Of particular interest were burnt deposits within the long enclosure ditches and a possible Bronze Age field system. Assemblages of and flintwork contribute to the understanding of these features and a programme of radiocarbon dating has amplified the chronology of activity on the site. Evidence from a group of burnt and unburnt pits and a partial enclosure reveal the reuse of the site between the fifth and eighth centuries AD.

INTRODUCTION The discovery and partial excavation of a group of prehistoric at Nether - field Farm, South Petherton, in Somerset provides new information on the distribu - tion and use of monumental architecture in Neolithic and Bronze Age Britain. The group comprised an extensive spread of remains, including a causewayed enclo sure, a long enclosure and a linear arrangement of ring-ditches, and is the first of this type of complex to be found in Somerset. The small size of the and the low frequency of material found retrieved from its ditches explain its incon spicuous presence in the modern landscape. Conversely, the long enclosure contained relatively abundant charred plant remains and other finds. The nature of both these monuments makes the complex unusual in national terms since causewayed enclosures are usually the more archaeologically productive of the two. The pits and ditches on the site of

Archaeol. J., 169 (2012), 3‒86 02 Mudd:Layout 1 03/07/2013 08:49 Page 4

4 excavations at netherfield farm, south petherton

the causewayed enclosure returned radiocarbon dates between the fifth and eighth centuries ad and indicate its reuse during the early medieval period. This evidence conforms to a wider pattern of early medieval activity found in and around prehistoric monuments identified across the county and beyond. As elsewhere, the extent, nature and motives for this reuse are not easily determinable from the scant remains recovered.

illus. 1 Site location. Scale 1:125,000

BACKGROUND Cotswold carried out a programme of archaeological work along the route of a new gas pipeline between Ilchester (NGR ST 50952295) and Barrington (ST 37851860) in Somerset (Illus. 1) between April 2005 and September 2006. Following several stages of earlier desk-based and field survey, field evaluation took place as mitigation work along the 17 km pipeline route between April and June 2006. Significant archaeological remains were identified in nine fields. Two of these fields, Areas 47 and 48, contained archaeological features that are considered to be of national 02 Mudd:Layout 1 03/07/2013 08:49 Page 5

excavations at netherfield farm, south petherton 5

significance and are the subject of this report. The other archaeological findings mainly relate to Roman settlement and land use and are summarized elsewhere (Brett and Mudd forthcoming). The initial blanket magnetometer survey along the entire 20 m-wide pipeline ease - ment revealed a circular enclosure (later defined as a causewayed enclosure) and a long enclosure on land north of Netherfield Farm (Sites 48A and 48B: Illus. 2). Two long but relatively narrow evaluation trenches partially confirmed their form and date. However, the potential importance of the discovery led to the decision to fully define the extent of the features by magnetometer and excavate a larger area than would be affected by the pipeline. To compensate for this additional work it was agreed that the complex of features to the north (Area 47) would be examined in a watching brief, rather than by excavation. The extended magnetometer survey commenced in the southern part of the site and fully defined a small causewayed enclosure comprising a single, incomplete, circuit (Illus 2 and 3). Two partial enclosures lay immediately to the east and north- east, apparently respecting both the causewayed enclosure and each other. The block of mottled geophysical terrain further north-east contained little that was clearly identi fiable except two north-west aligned linear features which correspond tolerably with former field boundaries of post-medieval date. To the north-east again, the excava tions defined the full length of the long enclosure (c. 80 m) and a linear arrange- ment of ring-ditches. There were linear ditches here as well, some of which respected the ring-ditches. Other linear ditches were only faintly defined but appeared to include a small double-ditched enclosure in the northern corner of this area. The magnetometer survey suggested a further extension of this pattern of enclosures, but this was not conclusively defined within the pipeline easement. The overall results indicated a group of earlier prehistoric monuments in these fields, potentially dating from the earlier Neolithic to the Middle Bronze Age, and representing a ceremonial complex so far unique for Somerset. The trial trenching established that the archaeological features were poorly pre - served and no stratigraphy survived above the surface of the natural substrate. Accordingly, the overlying topsoil and subsoil (altogether a little over 0.4 m thick) was stripped to the top of the substrate under archaeological direction by means of a mechanical excavator using toothless grading buckets. The whole area of the cause - wayed enclosure and adjacent features was exposed in this manner (Illus. 7) together with the western end of the long enclosure, including the ditch terminals (Illus. 4). The archaeological features were planned by hand on drafting film. Fifty per cent of all discrete features were excavated following a general excavation sampling strategy. One hundred per cent excavation was adopted for pits containing burnt material. A more limited intervention, comprising the hand-excavation of six sections through the ditches, was required for the long enclosure because only its western end was affected by the course of the pipeline. This report contains edited specialist reports, full versions of which are held in archive. All radiocarbon dates are calibrated and given at 95% probability unless other- wise stated. Posterior density estimates derived from Bayesian modelling of the radio- carbon dates are expressed in italics. 02 Mudd:Layout 1 03/07/2013 08:49 Page 6

6 excavations at netherfield farm, south petherton

illus. 2 Areas 47 and 48, showing geophysical survey results (Archaeological Surveys Ltd). Scale 1:4000 02 Mudd:Layout 1 03/07/2013 08:49 Page 7

excavations at netherfield farm, south petherton 7

illus. 3 Areas 47 and 48, showing abstraction and interpretation of geophysical survey results. Scale 1:4000 02 Mudd:Layout103/07/201308:49Page8

illus. 4 Plan of Site 48B, long enclosure, showing archaeological features. Scale 1:1000 02 Mudd:Layout 1 03/07/2013 08:49 Page 9

excavations at netherfield farm, south petherton 9

SITE DESCRIPTION The site lies on a gently inclined north-west-facing slope north of Netherfield Farm, approximately 500 m south-east of the hamlet of East Lambrook in the parish of South Petherton. The excavated area is situated just above the 20 m contour within 200 m of Lambrook Brook which drains into the River Parrett 1 km or so to the north-east. The geology here is a Late colluvial deposit which comprises a pale clay- silt with varied flint pebble inclusions deriving from the underlying Dyrham Forma- tion mudstones and sandstones (Wilkinson 2011). The poor preservation of bone here results from the sediment’s moderate acidity. The complex of Neolithic and Bronze Age features at Netherfield Farm was shown to spread over a distance of c. 600 m south-west/north-east along the east bank of the Lambrook Brook. The group comprised a causewayed enclosure, an adjacent and later U-shaped enclosure, a long enclosure about 150 m further north-east with a linear arrangement of ring-ditches close by, and a pattern of Bronze Age ditches and enclo - sures still further down the valley to the north-east. Linear magnetometer anomalies in Area 46 further north-east, examined by trial trenching, were not archaeological. neolithic Pit 48644 Lying between the two arms of the long enclosure (Illus. 4), pit 48644 was the earliest dated on the site. The pit was oval in shape and was 1.5 m long, 1.1 m wide and 0.6 m deep (Illus 5 and 6). Each of the three identified fills contained quantities of charcoal, dominated by oak but also containing elm, hazel/alder and cherry (Cobain this article; Tables 8 and 9). The upper fill contained a fairly substantial lump of oak charcoal, but a consideration of the form of the pit, without any recognisable post- setting or in situ burning, suggests that this does not represent the remains of an in situ post, but rather a secondary deposit of charred wood. A few crumbs of quartz- tempered pottery from the upper fill are only datable as Early to Middle Neolithic (McSloy this article), but two radiocarbon determinations on Prunus charcoal and hazelnut shell from the same context (48645) returned virtually identical dates in the range 3780–3660 cal. bc (NZA-35810, NZA-35816), making this feature substantially earlier than the long enclosure with which it appeared to be associated (Table 12; Healy, this report). The assemblage of lithics comprised twenty-nine pieces domi nated by utilized flakes and including fifteen flakes and blades of Portland chert.

Causewayed Enclosure The causewayed enclosure was approximately 55 m in diameter and lay on a gentle north-west-facing slope. The monument was formed by a single circuit of eleven separate ditch lengths arranged almost symmetrically, with a wide gap on its north- western side (Illus. 7). The sections of ditch varied in length; some appearing pit-like, while the longest, ditch 48j, was c. 23 m long. Significant truncation of archaeological features across the site resulted in the ditches only surviving to an average depth of between only 0.1 m and 0.3 m, and rarely any deeper. Therefore, while it is possible 02 Mudd:Layout 1 03/07/2013 08:49 Page 10

10 excavations at netherfield farm, south petherton

illus. 5 Site 48B, section AA through pit 48644. Scale 1:20

illus. 6 Pit 48644, half-sectioned, looking north-east. The soil block contained a lump of charcoal. Scale 1m 02 Mudd:Layout 1 03/07/2013 08:49 Page 11

excavations at netherfield farm, south petherton 11

illus. 7 Plan of site 48A, causewayed enclosure, showing archaeological features. Scale 1:1000 02 Mudd:Layout 1 03/07/2013 08:49 Page 12

12 excavations at netherfield farm, south petherton

that the north-western side of the monument had always been open, shallow ditches were probably present originally but subsequently lost to ploughing. Only one pit inside the enclosure may have been contemporary with the monument; the others were demonstrably later. The causewayed enclosure ditches had rounded terminals and varied in width from 1.2 to 1.9 m. The majority had a shallow concave profile with gradual or moderately steep sides (Illus. 8), although a few had broader, flat bases. The ditches had been left to silt up with little accumulation of cultural material, indicated by their largely sterile fills. From the earliest phases of silting within ditches 48g, 48n and 48p, small quantities of sherds of Early Neolithic Plain Bowl and less diagnostic Early to Middle Neolithic pottery were recorded. Meanwhile, ditch 48h contained a far larger num- ber: seventy-seven sherds from five different vessels. Ditch 48h also contained seven flint flakes. Two of these flakes were re-fitting pieces, indicating that the group derived from activity within or immediately adjacent to the ditch. A small assemblage of worked flint and chert, mainly flakes, came from the primary fills of the other ditches. However, a group of five artefacts, three of them end-scrapers, came from the terminal of ditch 48j (fill 48129). The scrapers comprise half the total from this monument, suggesting they were an intentional deposit. A radiocarbon determination on alder charcoal from the same primary fill returned a date of 3660–3635 cal. bc (NZA-35798). A far older radiocarbon determination on oak heartwood from fill 4832 of ditch 48g is likely to have an old wood offset and is not considered reliable as a date for the construction or use of the monument. After an unknown duration of silting, eight of the ditches (48f, 48g, 48h, 48j, 48m, 48n, 48p and 48q) showed evidence of re-cutting in certain sections as far as, or close to, the base of the original ditch (Illus. 8, sections BB, DD, EE). The extent of the re-cutting could not be traced throughout the full length of the ditches because the homogeneity of the ditch fills made the presence of re-cuts difficult to identify. More - over, later truncation of the features may have removed shallower re-cuts completely. The evidence suggests the re-establishment of the enclosure at least once after its initial construction. The later fills were again generally uniform, but the re-cuts of ditches 48j, 48p and 48q contained darker soils containing the most cultural material, comprising small quantities of Plain Bowl ware. A notable assemblage of fifty-one flints came from ditch 48j (Illus. 8, section DD; fill 48136). Among them were a chisel and two end-scrapers, as well as a number of flakes struck from the same Levallois core that appear to represent activity in or adjacent to the ditch (Anderson-Whymark this article). A radiocarbon determination on ash charcoal from ditch 48p (fill 48246, equivalent to 48207 or 48208, Illus. 8, section BB) returned a date of 3640–3520 cal. bc (NZA-35803).

Long Enclosure The magnetometer plot and the excavation show that the long enclosure was c. 85 m long with an internal width of 20 m. Opposing entrances were situated at the western and eastern ends (Illus. 4). The excavation of the western end of the enclosure 02 Mudd:Layout 1 03/07/2013 08:49 Page 13

excavations at netherfield farm, south petherton 13

illus. 8 Site 48A, sections BB to EE through causewayed enclosure ditches. Scale 1:20 02 Mudd:Layout 1 03/07/2013 08:49 Page 14

14 excavations at netherfield farm, south petherton

revealed that the southern ditch, 48s, survived up to 4.5 m wide and 1.15 m deep. Its northern counterpart ditch 48t, was more modestly proportioned, surviving to a maximum width of 3.65 m and a maximum depth of 0.6 m. This variation is original as there is no significant change in the topography to explain the difference in size of the two components. A similar sequence of deposits were revealed in the two sections cut through ditch 48s (Illus. 9, sections FF, GG). In each were clean basal deposits (48612, 48611 in the terminal section FF and 4877, 4876 in section GG). The grain-size and geochemical composition of the basal deposits in the terminal (Section FF) revealed that they were naturally deposited from two different formation processes. The primary clayey fills consisted of soil-derived particles deposited soon after the ditch was cut while the later sandier deposits had eroded from the ditch sides (Wilkinson 2011). The thicker deposits in the terminal, when compared with the middle ditch section, are presum - ably a result of a greater surface area exposed to erosion. The only limited evidence of activity consists of some burnt flint recovered in a soil sample from 48612 and three flint flakes from 48611.

illus. 9 Site 48B, sections FF and GG through long enclosure ditch 48s. Scale 1:50 02 Mudd:Layout 1 03/07/2013 08:49 Page 15

excavations at netherfield farm, south petherton 15

The basal layers were succeeded by shallow lenses of scorched earth and charcoal, 48610 and 4875. While this burning layer was not conclusively identified as con- tinuous throughout the ditch length, it occupied a similar horizon in the fill sequence of different ditch sections. Each was covered by thin layers of charcoal largely com - prising oak (Cobain this article). Slight scorching was also evident on the base of the overlying layer (Illus. 10) perhaps indicating an attempt had been made to extinguish the fire or cover the smouldering embers. Subsequent silting deposits in this section (Illus. 9, section GG) were without cultural material, but this was followed, towards the top of the sequence, by darker grey-brown silts containing pottery and flintwork. The upper fill of the terminal (48604) contained eight sherds of undiagnostic quartz-tempered pottery and seven pieces of struck flint, including an end-. Upper fill 4872, however, contained a large assemblage: among thirty-six sherds of pottery were a sherd from a Fengate- type cup and several sherds of other Fengate-type wares. Meanwhile, this upper fill contained 180 lithic artefacts including an extensive variety of forms such as blades, cores, chisel scrapers and a . The sequence of deposits in ditch 48t was similar to that in ditch 48s (Illus. 11). Gravelly deposits within the lower fills towards the western end of the ditch (Illus. 12) may indicate that an internal bank had once been present, although this evidence may simply reflect minor changes in geology. The lower natural infilling also comprised yellow-brown or orange-brown silts with little cultural input. Small quantities of hawthorn and blackthorn charcoal from the terminal (fill 48614) were selected for two

illus. 10 Section through ditch 48s, looking south-west. Scales 1m and 2m 02 Mudd:Layout 1 03/07/2013 08:49 Page 16

16 excavations at netherfield farm, south petherton

illus. 11 Site 48B, sections HH and II through long enclosure ditch 48t. Scale 1:50

illus. 12 Section through ditch 48t, looking north-west. Scale 2m 02 Mudd:Layout 1 03/07/2013 08:49 Page 17

excavations at netherfield farm, south petherton 17

radiocarbon determinations yielding comparable and consistent results in the range 3630–3190 cal. bc (Wk-22459) and 3494–3346 cal. bc (NZA-35805) which, at face value may give a range for the date of the construction of the ditch, although it appears to be anomalously early when modelled with the dates from the later fills (Healy, this article). This early silting was followed by two thin layers of charcoal separated by clean silts. The charcoal layers were not consistently present in all the sections and there are uncertainties in seeing them as uniform, ditch-wide events. They were clearest in Section II (Illus. 11) and both present, but less extensive, in Section 289 (Illus. 12). Conversely, in the terminal (Section 265) only one charcoal layer was present. In con- trast with the burning episodes in ditch 48s, these charcoal layers did not appear to be within cuts. Instead, they spread more widely, albeit intermittently, across the width of the ditch and not accompanied by notable scorching. It appears they may not relate to in situ fires but may be deposits of burnt or burning wood from elsewhere close by. The charcoal deposits contained no artefacts, although small quantities of pottery and flint were recovered from four contexts associated with the intervening periods of silting. These comprised eighteen sherds of Neolithic pottery, including impressed wares, and thirty-six worked flints, twenty-six of which came from fill 4866 (Illus. 11). The upper fills of ditch 48t appeared to relate to the cutting and filling of a series of shallow pits along the central line of the ditch. The pits appeared to have been used as , or other locations of fire, since activities resulted in the scorching of earlier deposits into which they had been cut, and they were filled with charcoal-rich soils (Illus 13–14). There was more than a single phase of these burnt pits (not shown on Illus. 9) and their locations were tightly controlled. Associated charcoal (Table 9, samples 20, 29) showed that it derived from a wide range of trees and shrubs, includ - ing notably potential food species such as those in the Pomoideae family, Prunus and hazel, and there was relatively little oak. Among the seeds, hazelnut was particularly prevalent, while crab apple, hawthorn, elder and possibly cherry were also present (Table 8, sample nos 10, 20, 29). This suggests the fires were related to food prepara - tion and would not seem out of place in a domestic context. The charcoal fills also contained significant quantities of pottery and worked flint. Ninety-eight sherds of pottery were recovered from deposits associated with this phase of activity, including a small number of sherds from bowls of Mortlake and Ebbsfleet type. A relatively large sherd from the primary fill of pit 48642 (Illus. 13) exhibited traces of burnt residue. A total of 135 worked stone artefacts were recovered from these features, an assem- blage that included groups of knapping . In contrast to ditch 48s, a large proportion of the lithic assemblage (37.4%) comprised Greensand chert. No specific activities can be surmised from the group of material, but it is apparent that they included the manu facture of , particularly chisel arrowheads (Anderson- Whymark this article).

Pits to the South of the Long Enclosure Pits 4857 and 4859 were circular and each survived to a depth of 0.15 m (Illus. 4). Both contained heterogeneous soils with charcoal and significant quantities of artefacts. 02 Mudd:Layout 1 03/07/2013 08:49 Page 18

18 excavations at netherfield farm, south petherton

illus. 13 Plan of burnt pits within ditch 48t. Scale 1:50

illus. 14 Section JJ through pits in ditch 48t. Scale 1:20 02 Mudd:Layout 1 03/07/2013 08:49 Page 19

excavations at netherfield farm, south petherton 19

Pit 4857, slightly larger at 0.9 m across, contained pottery (including probable Fengate ware) and two worked flints, and would seem to have been contemporary with the long enclosure. Pit 4859 contained daub, burnt fragments of animal bone and eleven flakes and blades, and would appear to be of broadly Neolithic date. There was also a cluster of three small pits adjacent to the western entrance to the long enclosure and a shallow ditch, 48u. These were without finds and contained naturally accumulated silts. Their dates are not known.

Pits around the Causewayed Enclosure There were a number of pits scattered around the causewayed enclosure (Illus. 7). Several were without any dating evidence but four contained pottery and/or flint- work which might date them to the Neolithic or Early Bronze Age. There is, how - ever, no consistency to their form or contents. Pit 48106 lying quite centrally in the south-eastern part of the U-shaped enclosure, contained a few grams of possibly earlier Neolithic plain ware and a collection of forty-five flints, mostly comprising utilized flakes, many from the same core, but also including a disc scraper and two serrated flakes. The assemblage would not be out of place in a Middle to Late Neolithic con - text (Anderson-Whymark, this article). It is possible that the pit was in some way related to the U-shaped enclosure rather than the causewayed enclosure (and con - tained some residual earlier pottery). The flints might indicate intentional deposition, although the feature was unremarkable and its purpose is unclear. The pit was 0.95 m in diameter and 0.25 m deep and is unlikely to have held a post. On the south-east side of the causewayed enclosure, pit 48183 was a broad shallow feature containing nine flints, including two unfinished arrowheads of chisel type. Pit 48268 was located on the causewayed enclosure’s western side, just off the alignment of the causewayed enclosure ditches; it was a more irregular oval and produced a small collection of flakes and blades.

bronze age U-shaped enclosure The U-shaped enclosure on the north-west side of the causewayed enclosure com - prised five individual lengths of ditch of different form (Illus. 7). Ditches 48a and 48b took the form of deep, almost vertically sided, palisade trenches up to 0.85 m wide and 0.7 m deep (Illus. 15, sections KK, LL). Ditch 48c was also steep-sided but rather shallower (0.45 m). Ditches 48d and 48e were all less substantial with gradually sloping sides; generally between 0.1 m and 0.35 m deep, and their form means that they are not readily interpreted as timber placements (Illus. 15, sections MM, NN). The depth of ditch 48a in particular suggests that the ‘enclosure’ was substantially complete and the open form was not the result of truncation on the north-western side. Although appar ently forming three sides of a single feature, the differences between the com- posite ditch lengths indicate a degree of complexity to its development. However, in the absence of stratigraphic relationships or refined dating, the monument’s develop - ment cannot be resolved from the excavated evidence. Forty sherds of Bronze Age 02 Mudd:Layout 1 03/07/2013 08:49 Page 20

20 excavations at netherfield farm, south petherton

pottery recovered from ditches 48a, 48c and 48e provide the dating evidence for these features. The pottery lacks diagnostic forms and is dated in broad terms to the Early or Middle Bronze Age based on fabric and firing characteristics. A large sherd came from primary fill 4823 of ditch 48a (together with a residual sherd of Peterborough Ware) and smaller sherds from 48203, the primary fill of ditch 48c. Most of the rest came from the later infilling from ditches 48a and 48e. The collection of thirty-six worked stone tools is not diagnostic of either date or function and may include residual Neolithic pieces. There is therefore no good indication of the sequence of construction within the Early to Middle Bronze Age but it would seem that con - struction began well after the abandonment of the adjacent causewayed enclosure.

illus. 15 Sections KK to NN through U-shaped enclosure ditches. Scale 1:20 02 Mudd:Layout 1 03/07/2013 08:49 Page 21

excavations at netherfield farm, south petherton 21 1 : 1000 47 ; abstraction and interpretation of geophysical survey results in relation to excavated features. Scale Area illus. 16 02 Mudd:Layout 1 03/07/2013 08:49 Page 22

22 excavations at netherfield farm, south petherton

illus. 17 Area 47, sections OO and PP through Bronze Age ditches. Scale 1:50

Pit 48104 Pit 48104, on the eastern side of the site was irregular and very shallow and contained a bladelet with a few small sherds of grog-tempered Bronze Age pottery.

Ditches north of the long enclosure The magnetometer survey conducted to the north of the long enclosure identified a number of ditches in the pipeline easement and the area to the east (Illus 2–3). Evalu- ation trenches 4703, 4704, 4705 and 4706 targeted these ditches, and were followed by an area excavation centred on the curvilinear ditches 47c and 47e (Site 47; Illus. 16). Bronze Age pottery was recovered from ditches 47403, 47405, 47503 (Illus. 18), 47d and 47e (Illus. 16). Curving ditch 47e at the northern end of the group may have formed part of a ring- ditch about 13 m in diameter, although its northern side would have lain under the modern hedgerow and it was not identified. The ditch measured about 1.6 m wide and 0.35 m deep with inconsistent sides and a relatively flat base (Illus. 17, section 00). Relatively securely dated by a group of twenty-one sherds of pottery in the Wessex biconical urn tradition, which date to the later part of the Early Bronze Age, 47e was cut by ditch 48d which was slightly deeper and contained a group of four Bronze Age sherds of similar type. Curving ditch 47c appeared to mimic the line of 47e and was of similar form and dimensions; it contained several fragments of animal bone and a piece of amorphous fired clay. 02 Mudd:Layout 1 03/07/2013 08:49 Page 23

excavations at netherfield farm, south petherton 23

illus. 18 Area 47; Trenches 4703, 4704 and 4705, showing archaeological features and geophysical survey results. Scale 1:1000

A double-ditched enclosure of sub-rectangular form was identified by magnet- ometer survey and examined in Trench 4704 (Illus. 18). The trench found three ditches in approximately their expected positions. Ditches 47405 and 47407 were of similar dimensions, 1.0 m–1.35 m wide and 0.25–0.35 m deep with moderately sloping 02 Mudd:Layout 1 03/07/2013 08:49 Page 24

24 excavations at netherfield farm, south petherton

sides, while ditch 47403 was slightly smaller (Illus.17, section PP). Three sherds of unspecific Bronze Age pottery came from ditch 47403 and another sherd from 47405. Two pieces of worked flint came from ditch 47407. To the north, parallel to ditch 47403, ditch 47503 was of similar dimensions and contained two sherds of Bronze Age pottery. To the south of this group, ditch 47304 contained just two worked flints. When viewed against the background of the wider magnetometer survey, the excavation results suggest the presence of a Bronze Age rectilinear field system (Illus. 2). This does not preclude some of the ditches being later, but this does not seem to be the case generally. Ditch 47407 in particular appears to be physically part of the field pattern evident in the wider magnetometer plot (Illus. 18), while other ditches are connected to the ring-ditches in the central part of the site in a manner that implies that the ring-ditches were still visible (Illus. 2). The link between these field ditches and the curving Bronze Age ditches at the northern end of the site (Area 47) is not evident from the area under investigation, but it is possible that the apportioned landscape developed directly from funerary or domestic enclosures here. early medieval features Pits A group of pits containing burnt material and evidence of scorching were found within and close to the causewayed enclosure (Illus. 7). Two of these features, 48265 and 48272, are best described as fire-pits, and three others are pits of distinctive lozenge shape (48144, 48170, 48192). Another pit, 48221, appeared to comprise a burnt fire-pit with a shallower rake-out pit on its southern side (48235). None of these features contained artefacts but their dating to the fifth to seventh centuries ad is confirmed by a series of seven radiocarbon dates from charred grain from five of these features. In addition, L-shaped ditch 48k has a likely seventh- to eighth-century ad date based on two radiocarbon dates derived from bones (Table 12). Fire-pit 48265 was 0.85 m in diameter and 0.1 m deep while fire-pit 48272 was slightly smaller and deeper. Both of these pits showed heavy scorching of the substrate (Illus 19–20). Fire-pit 48272 contained a carbon-rich primary fill overlain by two distinct deposits containing burnt flint and fired clay. These deposits are likely to have been associated with the use of the feature and were sealed by a layer of clean silt which accumulated after it went out of use. The charred plant remains included char- coal from a range of species, particularly of hawthorn type, and seeds dominated by grains, particularly barley, bread wheat and oats (Table 10). The shallower fire-pit 48265 had similar charcoal, burnt flint and clay fills and yielded a very similar range of crops and charcoal species. The three lozenge-shaped pits were of similar dimensions, 2.45–2.75 m long, 0.8 m–1.1 m wide and 0.35–0.4 m deep (Illus 21–22). They contained a similar range of contents which reflected the crops, weeds and fuel found in the fire-pits, but their purpose remains unclear. Despite the presence of burnt material in their fills, labora- tory analysis of geoarchaeological samples from pit 48144 indicates that there is no evidence of in situ burning. Moreover, the pits were not clay-lined. It appears there - fore that the fills comprised secondary burnt material. Four radiocarbon dates on 02 Mudd:Layout 1 03/07/2013 08:49 Page 25

excavations at netherfield farm, south petherton 25

illus. 19 Section QQ through fire-pit 48272. Scale 1:20

illus. 20 Fire-pit 48272, half-sectioned, looking south-west. Scale 30cm 02 Mudd:Layout 1 03/07/2013 08:49 Page 26

26 excavations at netherfield farm, south petherton

illus. 21 Section RR through pit 48144. Scale 1:20

illus. 22 Pit 48192, half-sectioned, looking north-east. Scale 1m

charred wheat grains and dogwood charcoal from the primary fills provided dates consistent internally and with the series from the fire-pits (Table 12; Illus. 31). Pit 48221 was slightly different in form, being circular with a diameter of 0.9 m and 0.3 m deep. A shallow hollow was exhibited on its southern side. Its western side was almost vertical, but on the east it had been cut away by ditch 48k. Like the other pits it was without finds except charcoal. The underlying substrate showed evidence of intense scorching. Geoarchaeological sample throughout the sequence of fills showed that the lower fill (48232) had heat-fused clay and sand grains, indicating a temperature 02 Mudd:Layout 1 03/07/2013 08:49 Page 27

excavations at netherfield farm, south petherton 27

in excess of 650º Centigrade. Chemical analysis found unusually high concen trations of zinc, copper and silver in the sediment. Despite the absence of slags, the pit might have been used for metalworking.

Ditch 48k Ditch 48k was an L-shaped partial enclosure to the east of the causewayed enclosure cutting pit 48221 (Illus. 7). The ditch was a substantial feature, 43.5 m long, up to 2 m wide and 0.75 m deep with rounded terminals, and generally moderately sloping, uneven, sides. The base was generally concave. The primary silting was up to 0.55 m thick and contained small quantities of early prehistoric and Roman pottery, worked flint and a quantity of cattle bone fragments. Two upper deposits contained a range of finds which included ferrous slag and a tiny sherd of probably medieval, chert- tempered, pottery. The animal bone assemblage included both cattle and sheep. Two radiocarbon dates on cattle bones from the primary fill yielded mutually consistent dates in the seventh to eighth centuries ad (Table 12) suggesting that the chert- tempered sherd was intrusive. Probably contemporary with ditch 48k, a large pit or ditch terminal 48123 was located only partly within the excavation area to the south-east. It was at least 5.3 m long, 1.2 m wide and 0.4 m deep and contained a similar range of material to ditch 48k. This included cattle bones, a piece of ferrous slag and iron objects including a binding strip. Nearby, three pits (48212, 48213 and 48214) were partially enclosed by ditch 48k and may well have been contemporary with it although none contained dating evidence. The largest, 48214, was 1.15 m long, 0.9 m wide and 0.45 m deep, and the others slightly smaller. There was no clue as to their functions.

NEOLITHIC AND BRONZE AGE POTTERY by E. R. McSloy incorporating thin-section analysis by Elaine L. Morris The assemblage of Neolithic pottery from the causewayed enclosure, long enclosure and associated features consists of 369 sherds, weighing in total 1455g and preservation is typically poor. While moderately small in national terms, the Neolithic group is one of very few known from this part of Somerset and the location is significantly west of the best-known sites of Wessex which have produced comparable material. A total of seventy-five sherds of Earlier Bronze Age pottery weighing 802 g came from the U-shaped enclosure and the ditches to the north (Area 47). For the most part fragmentation is high and preservation poor; and there are few sherds exhibiting diagnostic features. The group from Area 47, however, includes some better- preserved sherds, some of which can be identified as belonging to the Wessex biconical urn tradition. The pottery has been recorded respecting guidelines produced by the Prehistoric Ceramics Research Group (PCRG 1997). 02 Mudd:Layout 1 03/07/2013 08:49 Page 28

28 excavations at netherfield farm, south petherton

fabric descriptions The assemblages were sorted into thirteen fabrics (below). Ten sherds from the Neolithic group were selected for thin-section analysis with the aim of refining macroscopic observations and investigating possible sources for the major types. The results of this analysis are held in archive with summaries of the identifications abstracted here, and a discussion included below.

Neolithic F1 Dark grey with buff-coloured outer margin. Soft with irregular break and harsh feel where inclusions protrude. Common, poorly-sorted (3–10 mm) angular, fire-crazed flint. Dating: Middle Neolithic. 10 sherds; 43 g G/A1 Dark grey throughout. Soft with finely irregular break and soapy feel. Common, well- sorted (1-2 mm) sub-rounded, grey-coloured grog or clay pellet. Dating: Middle Neolithic. 1 sherd; 3 g Q1 Dark grey throughout. Soft with irregular or laminating break and soapy feel. Abundant fine/silt-sized quartz sand; sparse/very sparse, well-sorted (0.5–1 mm) sub-angular quartzite; sparse red-brown iron oxide and sparse, sub-rounded clay pellet/argillaceous inclusions. Dating: Middle Neolithic. 15 sherds; 75 g Q2 Dark grey throughout. Soft with finely irregular break and slightly sandy feel. Common well-sorted (0.3–0.5 mm) sub-rounded quartz sand; sparse red-brown iron oxide and sparse, sub-rounded clay pellet/argillaceous inclusions. Confirmed by thin-section sample 10. Dating: Middle Neolithic. 2 sherds; 55 g QT1 Dark grey throughout. Soft with irregular break. Surfaces are commonly smooth with coarse inclusions not often protruding. Irregular fractures. Common moderately-sorted (2–4mm) angular quartzite. Two sub-types identified by petrographic analysis, QT1A quart zite only (thin-section samples 4 and 5, and QT1B (thin-section samples 6–9) quartzite and chert Dating: Early and Middle Neolithic. 125 sherds; 508 g QT2 Dark grey-brown throughout. Soft with irregular break. Surfaces sandy or harsh feeling. Common and well-sorted (1–2 mm) angular quartzite. Dating: Early and Middle Neolithic. 3 sherds; 20g QT3 Dark grey throughout. Soft with finely irregular break and smooth feel. Sparse well- sorted (in range 1–2 mm) angular quartzite. Dating: Early and Middle Neolithic. 7 sherds; 22 g SI Grey with red-brown surfaces and margins. Soft with fine break and smooth/powdery feel. Largely free of coarse inclusions other than abundant silt-sized quartz. Banded appearance at break and fine, burnt-out organic material suggest use of untempered natural clay. Confirmed by thin-section sample 1. Dating: Early Neolithic.18 sherds; 51 g VES1 Dark grey throughout. Soft with irregular break and soapy feel. Common rounded or sub-rounded voids (typically 1–2 mm), some edged with white or pale yellow; sparse red- brown iron oxide. Thin-section samples 2–3 identified disintegrating micaceous mudstones and mudstone voids resulting in the vesicular character of the fabric. Dating: Early and Middle Neolithic. 62 sherds; 161 g 02 Mudd:Layout 1 03/07/2013 08:49 Page 29

excavations at netherfield farm, south petherton 29

VES2 Grey throughout. Soft with irregular break and soapy feel. ‘Corky’ with common large rounded or sub-rounded voids (in range 2–42 mm), some edged with white or pale yellow; sparse red-brown iron oxide. Dating: Early Neolithic. 42 sherds; 76 g V/F1 Grey with buff exterior and margin. Soft with irregular/laminating break and smooth feel; harsh where inclusions protrude. ‘Corky’ with common large rounded or plate-like voids; sparse angular flint (1–2 mm) and sub-rounded quartz. Dating: Middle Neolithic. 5 sherds; 16 g

Bronze Age GROG Dark grey throughout or with light brown exterior and margin. Soft with finely irregular break and soapy feel. Common moderate or well-sorted, self-coloured grog, 1–2 mm. May contain sparse quartzite, 1.5–2 mm. Dating: Early Bronze Age. 112 sherds; 821 g GROGm Dark grey throughout. Soft with finely irregular break and soapy feel. Common poorly sorted (1–3 mm), self-coloured grog and sparse, poorly-sorted (2–6 mm), sub-rounded grey or buff-coloured mudstone. Dating: Early Bronze Age. (Site 47 only): 2 sherds; 191 g

discussion Early Neolithic The Early Neolithic pottery relates almost entirely to the site of the causewayed enclosure and mainly from the upper fills of the causewayed enclosure ditches. The condition of the pottery is varied; the dominant ‘vesicular’ fabrics are weakened through the loss of mineral ‘soft rock’ inclusions and typically they are well- fragmented. Surface preservation however tends to be good and it is a feature of the Early Neolithic group that vessel surfaces are smoothed and well-finished. Radiocarbon dates from section 48j deposit 48129 and section 48p deposit 48246 (3693–3541 cal. bc and 3637–3523 cal. bc respectively) are fully consistent with ‘South-Western’ style Early Neolithic ceramics of the kind represented (Whittle 1977, 77–82; Ladle and Woodward 2009). These dates correspond to the main period of construction for causewayed enclosures in the 38th to 34th centuries bc (Healy, this article). Pit 48644, between the terminals of the long enclosure, from which the earliest radiocarbon determinations were obtained (both dates c. 3780–3660 cal. bc; NZA-35810 and NZA-35816), produced only featureless scraps of pottery in Neolithic quartzite fabric QT1 (below). All the Earlier Neolithic group (128 sherds; 437 g), barring the nine scraps from pit 48644 and two sherds from pit 48106 within the Bronze Age U-shaped enclosure, relate to the filling of the causewayed enclosure or its re-cuts. Quantities from the separate lengths of ditch making up this feature are shown in Table 1. The largest con- text group of seventy-seven sherds, from the primary filling of ditch 48h, represents a maximum of five well-fragmented vessels. Overall the level of fragmentation in the Early Neolithic group is high and reflected in a mean sherd weight of only 3.4 g. 02 Mudd:Layout 1 03/07/2013 08:49 Page 30

30 excavations at netherfield farm, south petherton

table 1 Early prehistoric pottery quantification by sherd count (ct) and weight (wt) in g CW=causewayed enclosure, LE=long enclosure, U=U-shaped enclosure

fabric si qi q2 ves1 ves2 v/f1 qt1 qt2 qt3 f1 grogm grog

Pit 48644 ct 9 wt 3 CW ct 18 2 56 42 7 2 3 wt 51 5 276 76 47 17 6 LE ct 232 513 2 ditch s wt 15 53 16 44 8 LE ct 13 22 75 2 4 ditch t wt 70 60 370 9 9 Pit 4857 ct 21 1 4 1 wt 44 3 26 3 Uct 141 wt 3 252 Pit 48106 ct 2 wt 3 Pit 48104 ct 6 wt 3 Site 47 ct 235 wt 191 365 Total ct 18 15 2 112 42 5 125 3 6 10 3 82 wt 51 75 15 392 76 16 508 20 18 43 194 620

Notwithstanding the limited sampling of the causewayed enclosure, there are no indications that the pottery recovered from this feature represents any form of structured deposit. Diagnostic Early Neolithic sherds are small in number and all are drawn (Illus. 23, 1–5). The group’s fragmentary state, in particular the absence of lower body profiles, precludes full classification of form according to Cleal’s scheme (2004, figs 4–5) and the group adds little to the understanding of Neolithic bowl typology. The vessels appear to be a mix of ‘open’ (nos 1–3) and ‘neutral’ bowl forms with either simple or rolled-over rims. Vessel diameter is in the range c. 200–280 mm. One vessel, the neutral-profile vessel number 3, features at least one vertically perforated, ‘pinched- out’, lug. This vessel is narrower, the diameter around 160 mm. All vessels are well made, typically with well-smoothed surfaces and thickness most often in the range 5–7 mm. Two of the vessels (nos 1 and 2) exhibit clear evidence for coil-manufacture. 02 Mudd:Layout 1 03/07/2013 08:49 Page 31

excavations at netherfield farm, south petherton 31

Vesicular (Mudstone-tempered) types (VES1 and VES2) are most abundant, num- bering ninety-seven sherds (347 g), although the number of vessels is considerably fewer. Igneous types are absent and ‘hard rock’ types are present as a small number of sherds in quartzite-tempered QT1 fabric. The seeming dominance of the vesicular fabric, albeit in a small group, is at odds with Early Neolithic groups from the more easterly counties of Wessex (Cleal 1995, 187; Darvill 2004, 194–5) where flint- tempered types typically make up well over half.

Middle Neolithic The Middle Neolithic group (172 sherds, 684 g) mostly come from the ditches forming the long enclosure and associated pits. Small quantities (five sherds) come from Site 48A; from one re-cut upper filling (48207) of ditch 48p of the causewayed enclosure (Illus. 23, no. 6), and (residually) from Bronze Age U-shaped enclosure (ditch 48a). The prevalence of impressed and other decoration indicates that all material belongs within the widespread Peterborough ware style (Middle Neolithic Impressed ware). The radiocarbon dates fall within the Middle Neolithic (Healy this article) and are well within the range to be expected for the Peterborough Ware series which extends c. 3400–2500 cal. bc (Gibson and Kinnes 1997). The condition of the pottery is generally poor and with a few exceptions only small body-sherds are represented. Average sherd weight is a little higher than that for the causewayed enclosure, though still low at 4 g. This, together with a general absence of joining or same vessel sherds across or within context, is again more suggest ive of a gradual accumulation rather than anything more structured such as deliberate place - ment. Surface preservation tends to be poor, more likely the result of deterioration in the ground rather than abrasion through the mechanical actions of re-deposition. Sherd thicknesses are notably in excess of those for the Early Neolithic group, most com monly in the range 10–12 mm and to a maximum of 20mm (Illus. 24, no. 16). The majority of sherds with features are illustrated (Illus. 23, 24; nos 6–17), omitting only smaller body-sherds with indistinct decoration. A collared rim-sherd from ditch 48s (Illus. 23, no. 7) can with confidence be attributed to the Fengate style. Further vessels from deposits which feature twisted cord and/or incised decoration, nos 8 and 9, might also belong to the Fengate sub-style; as might a vessel number 6 from a late (re-cut) filling of ditch 48p of the causewayed enclosure. The remaining vessels probably derive from round-based bowls but reliable attribution to sub-style is not always possible. The simple rim forms, distinct carination and ‘understated’ decoration indicate that bowl forms including numbers 11–13 belong to the Ebbsfleet sub-style. Sherd thickness and/or profusion of decoration suggest that the Mortlake type is also represented (Illus. 24; nos 16, 17). Surface deterioration results in the decoration being in some instances indistinct and the means of execution not always clear. Use of whipped cord (nos 6, 12, 16), twisted cord (nos 8, 10), fingernail (nos 7, 11) and impressions of a small circular-ended imple- ment (no. 17) are evident. On the large sherds from pit 48642 the circular impressions are made in repeated rows, the effect being not unlike basketry (no. 17). 02 Mudd:Layout 1 03/07/2013 08:49 Page 32

32 excavations at netherfield farm, south petherton

Represented fabrics and possible origin are discussed below. There is significant overlap with Early Neolithic types, with both the vesicular (mudstone) and quartzite/ chert-tempered types clearly continuing. Such continuity reflects a pattern previously noted from studies in the Wessex area (Cleal 1995, 187–9).

Bronze Age Site 47 (thirty-seven sherds; 556 g) produced the best-preserved site group and included the only vessels attributable with certainty to particular Bronze Age type or style. Vessels nos 18–20, from ditches 47d and 47e in the north part of the trench, are each consistent with the Wessex biconical urn tradition of the ‘later early Bronze Age’. The nature of deposition from linear and curvilinear ditches is suggestive of a domestic, rather than a funerary group. The assemblage is too small for useful com- parison, though there are good parallels for domestic use of biconical vessels from Brean Down, Somerset (ApSimon 1961, 67–136; Woodward 1990, 123–6) and Shaugh Moor, (Wainwright and Smith 1980). The Brean Down assemblage was associated with radiocarbon determinations in the range 1861–1422 cal. bc (Bell 1990). The Bronze Age pottery from Site 48A (forty-two sherds; 255 g) derives from features in the north portion of the area and principally from ditches forming the ‘U-shaped’ enclosure. Featured sherds are limited to base sherd no. 22 (Illus. 25) and the broad Early to Middle Bronze Age dating is based on the fabric and firing characteristics. stylistic affinities and comparanda Early Neolithic The South-Western regional style of Early Neolithic pottery is largely absent from Somerset, but is associated with the well-dated (Cleal 2004, fig. 5). The largest groups are from the Wessex region, including Maiden Castle causewayed enclosure (Wheeler 1943), Flagstones enclosure, Dorchester (Cleal 1997, 86–102), the ‘Coneybury Anomaly’ group from the area of (Cleal 1991, 45–57) and elements from Windmill Hill causewayed enclosure, near , Wilts. (Smith 1965). There is overlap also with the wares known mainly from Devon, although the Netherfield Farm group lacks examples of trumpet lugs common to Hembury vessels and similarly contains no evidence for the use of Cornish gabbroic clays. Analysis of the Early Neolithic fabrics suggests the exploitation of locally available resources with some limited evidence for long-distance movement of ceramics (Morris, below). The geological origins of the fabrics differ from those of the flint or calcareous-tempered types of Early Neolithic date described from / (Cleal 1995, 187). A parallel tradition of potting utilizing minerals offering similar hard- or soft-rock properties would seem likely. The absence of flint is perhaps surprising given its contemporary prevalence elsewhere in Wessex and moreover that knapping waste, which Cleal (ibid.) suggests may have been widely utilized, must have been freely available. 02 Mudd:Layout 1 03/07/2013 08:49 Page 33

excavations at netherfield farm, south petherton 33

Middle Neolithic The Middle Neolithic Peterborough Ware group is among the more westerly known examples from , although the type occurs sporadically in Devon (Laidlaw and Mepham 1999, 44–45; Quinnell 2007). Regarding decorative range and application, the assemblage resembles contemporary material from across southern England and Wales. The assemblages from West Kennet (Piggott 1962) and Windmill Hill/ Avebury (Smith 1965), large groups with all of the sub-styles represented, provide parallels in terms of vessel form and use of decoration. Both assemblages include cup- sized Fengate-type vessels of which no. 7 (Illus. 23) is an example, although a better parallel in terms of decoration is a vessel from near Ratcliffe on the Wreake, Leics. (McSloy 2008, fig. 6, number 5). Analysis of the pottery fabrics demonstrates the exploitation of local clays and other resources and suggests that the common types represent a continuation of earlier traditions. Some use of potentially more distant resources is hinted at (Morris, below), perhaps suggesting contact and exchange across communities. However, there is no evidence for movement of ceramics over longer distances. Notwithstanding the small size of the sample there are notable differences between the pottery from each of the two arms of the long enclosure (ditches 48t and 48s), both in terms of relative quantity and composition (Illus. 35). This apparent separation is marked by a greater abundance of material in ditch 48t, a group that consists of round- bottomed Ebbsfleet/Mortlake-style bowls, compared to a smaller group from ditch 48s comprising the typically vase-like vessel forms of the Fengate style. This distribution may reflect a wider pattern of discrete deposition which can be apparent from individual feature groups, if only rarely from site assemblages. Evidence for chronological progression in the Peterborough ware series has not been upheld by dating programmes (Gibson and Kinnes 1997). The three Peter- borough sub-styles are all potentially contemporary and began as early as the 35th century bc. Given the potentially extended period of use for the monument, the compositional differences across 48s and 48t could still relate to chronology, as a consequence of deposition tens or even hundreds of years apart. Differences in the deposit formation and use, evident between 48t and 48s, may also be of relevance here, in particular the extensive evidence for burning in the upper fills of 48t, which suggest that varying disposal practices were responsible.

illustrated vessels (Illus 23‒25) Early Neolithic 1 Open bowl, simple rim. Coil break. Causewayed enclosure ditch 48j; primary filling 4832. Fabric QT1. 2 Open bowl, rolled-over rim. Coil break. Causewayed enclosure ditch 48h; primary filling 4830. Fabric VES1. 3 ‘Neutral’ bowl with horizontal (vertically-perforated) lug; everted/out-curved rim. Causewayed enclosure ditch 48h; primary filling 4830. Fabric VES1. 4 Open bowl, rolled over rim. Causewayed enclosure feature 48q; secondary filling 48175. Fabric VES1. 02 Mudd:Layout 1 03/07/2013 08:49 Page 34

34 excavations at netherfield farm, south petherton

5 Open bowl, simple/flattened rim. Causewayed enclosure ditch 48p; primary filling (48198) of ditch re-cut 48197. Fabric SI.

Middle Neolithic 6 Flake, Middle Neolithic impressed ware/Fengate style(?). Whipped cord impressions to rim upper. Scored and indistinct impressed decoration to rim outer. Causewayed enclosure feature 48p; secondary filling (48207) of ditch re-cut 48206. Fabric Q1. 7 Flake, Middle Neolithic impressed ware/Fengate style. Small cup (diam. c. 130 mm). Fingernail notches to rim upper. Scored arc decoration to ‘collar’. Long enclosure ditch 48s; secondary filling (4872) of ditch cut 4871. Fabric I2. 8 Flake, Middle Neolithic impressed ware Fengate style(?). Twisted cord decoration, interior/exterior. Incised decoration to exterior. Long enclosure ditch 48s; secondary filling (4872) of ditch cut 4871. Fabric QT1. 9 Sherd with incised decoration, Middle Neolithic impressed ware Fengate style(?). Long enclosure ditch 48s; secondary filling (4872) of ditch cut 4871. Fabric Q2. 10 Middle Neolithic impressed ware. Twisted cord decoration, exterior. Pit 4857; primary filling (4858). Fabric QT1. 11 Middle Neolithic impressed ware/?Ebbsfleet style bowl (diam. c. 140 mm). Indistinct impressions and fingernail impressions interior/exterior. Long enclosure ditch 48t; burnt deposit 48639. Fabric QT1. 12 Middle Neolithic impressed ware/?Ebbsfleet style bowl (diam. c. 160 mm). Whipped cord impressions to carination; indistinct decoration to rim outer. Long enclosure ditch 48t; burnt deposit 48639. Fabric Q1. 13 Middle Neolithic impressed ware/?Ebbsfleet (?) style bowl. Whipped cord(?) impression below rim. Long enclosure ditch 48t; burnt deposit 48608. Fabric QT1. 14 Middle Neolithic impressed ware/?Ebbsfleet style bowl. Indistinct impressions to rim interior. Long enclosure ditch 48t; burnt deposit 48608. Fabric VES1. 15 Middle Neolithic impressed ware bowl. Indistinct impressions to exterior. Long enclosure ditch 48t; burnt deposit 48633. Fabric QT1. 16 Middle Neolithic impressed ware. Thick-walled ?Mortlake-style bowl sherd. Whipped cord decoration, exterior. Long enclosure ditch 48t; ditch fill 48638. Fabric QT1. 17 Middle Neolithic impressed ware/?Mortlake style bowl. Decoration as small circular impressions arranged in rows. Long enclosure pit 48642; primary filling 48643. Fabric QT1.

Bronze Age 18 Rim sherd from biconical-urn with internal concave bevel. Decoration consists of fingernail slashes to rim bevel and applied cordon at neck carination with deep fingernail impressions. Site 47; ditch 47d, fill 4727. Fabric GROGm. 19 Body-sherd from ?biconical vessel with fingernail impressions. Site 47; ditch 47d, fill 4727. Fabric GROG. 20 Body-sherd from ?biconical vessel, probably a bowl. Site 47; ditch 47e, fill 4725. Fabric GROG. 21 Body-sherd with ?horseshoe-shaped handle scar. Site 47; ditch 47e, fill 4725. Fabric GROG. 22 Base sherd. Biconical or Collared Urn? U-shaped enclosure, ditch 48a; secondary fill 48158. Fabric GROG. 02 Mudd:Layout 1 03/07/2013 08:49 Page 35

excavations at netherfield farm, south petherton 35

illus. 23 Early Neolithic pottery (nos 1–5). Middle Neolithic pottery (nos 6–12). Scale 1:2 02 Mudd:Layout 1 03/07/2013 08:49 Page 36

36 excavations at netherfield farm, south petherton

illus. 24 Middle Neolithic pottery (nos 13–17). Bronze Age Pottery (nos 18–20). Scale 1:2 02 Mudd:Layout 1 03/07/2013 08:49 Page 37

excavations at netherfield farm, south petherton 37

illus. 25 Bronze Age pottery (nos 21–24). Scale 1:2

discussion of the neolithic fabric groups by Elaine Morris The Neolithic causewayed and long enclosures lie in an area of Trias, Lias and alluvium geological formations (Welch and Crookall 1948, fig. 1). The extremely fine quartz grains in the layered clays of silty fabric (SI) undoubtedly derive from alluvium. Within the Triassic Keuper formation specifically, Tea Green and Grey Marls contain clays and mudstones (ibid., 42), either of which may have been the source for the vesicular mudstone fabric (VES1). Therefore, these two Early Neolithic pot-making fabrics existed naturally in the local landscape and were selected for their suitability to make vessels not requiring deliberately tempered fabrics. The successful firing of pots made from such a fine, dense, layered clay, as found in the silty fabric, indicates that the potter understood how to control the speed and rise of heat in an open fire. The porosity of mudstones in the vesicular fabric would have been easier to work with under similar firing conditions. In contrast, the quartzite-tempered and quartzite and chert-tempered fabric (QT1) may have originated from Carboniferous deposits located in the Mendips where a very hard, fine-grained grit band known as a ‘quartzite’ has been used as a top dressing for modern roads (Welch and Crookall 1948, 24). An extremely similar chert displaying chalcedonic, radial-fibrous quartz replacement is also known from the Upper Jurassic deposits in Dorset to the south (Adams et al. 1984, figs 187–88), while chert beds are also found in the Upper Greensand deposits in Somerset (Wilson et al. 1958, 146–48, 152–55, 157, and 163). Further research is required to investigate these and other 02 Mudd:Layout 1 03/07/2013 08:49 Page 38

38 excavations at netherfield farm, south petherton

poten tial sources in order to locate a fabric source with the combination of both quartzite and chert. The infrequent presence of naturally occurring but angular quartzite, rounded ferruginous sandstone and angular probable chert or flint in the otherwise medium-grained, sandy fabric (Q2) indicates that the clay selected to make it was different from that used to make the silty fabric (S1). The resource may have been located in a river valley sediment near that of, or similar to, the quartzite and chert source. Therefore, at least two fabrics (SI and VES1) were most likely to have been made from local, untempered clays, while two others (QT1 and Q2) appear to have been produced from sources outside the local area. Local procurement of resources for pottery production amongst sedentary, agri - cultural communities has been defined as normally up to 7 km for the clay and 10 km for tempering additives (Arnold 1985; Morris 1994; Morris and Woodward 2003, 289). This model may be unsuitable, however, for mobile or semi-sedentary groups during the earlier Neolithic in Britain. Very little investigation of Neolithic pottery fabrics from south-west Britain has been conducted using petrological analysis other than to identify gabbroic fabric pottery (Morris and Woodward 2003, table 1, fig. 1), but quartzite and chert have been recognized in Neolithic fabrics from west Cornwall (ibid., 285). In contrast, Early Neolithic bowls from the Sweet Track in Somerset had been made from fine sandy fabrics and grog-tempered fabrics (Williams 1976). No examples of large, angular quartzite and chert-tempered, or mudstone-rich fabrics were identified in that assemblage, which consisted of significant parts or fragments of seven or eight pots. However, at the Sweet Track the deliberate selection of particular vessels from amongst a much wider array of vessel forms and fabrics used in daily activities may have skewed the deposition of vessels (Smith 1976, 63–64). Continuity of pottery production from Early Neolithic to Middle Neolithic is demonstrated by the use of a quartzite-tempered (QT1) fabric, which involved adding large pieces of crushed silaceous rock to very fine clay with naturally occurring mudstone. The presence of the chert in subgroup QT1B suggests that a slightly different source deposit was selected during the Middle Neolithic period although the technological concept remained the same. It may be significant that quartz ite and chert are made from silica, a hydrated form of flint (Rye 1981, 34). In the Wessex region, flint-tempered fabrics represent c. 46–60% of identifiable Early and Middle Neolithic vessels (Cleal 1995, fig. 16.2). At Maiden Castle in particular, where quartzite-tempered pots are absent, flint-tempered fabrics make up 35–45% of the large Early Neolithic assemblage and 32–37% of the small Middle Neolithic (Peter- borough Ware) assemblage (Cleal 1991, tables 54 and 62). The silica-based fabrics made with quartzite/chert and with flint would have had similar technological properties, such as those required for thermal and mechanical shock resistance (Morris 1991). The choice of silica rocks as tempering agents would not have been taken lightly by the potters. The rocks required heating and then considerable energy to crush them to the required size-range. The Maiden Castle Early Neolithic assemblage also included a small number of sherds from vessels that had been made from fabrics with specifically non-shell voids (1.2%) and it would be useful to re-examine these thin sections to determine if any of the voids had once been mudstones. In addition, it is recognized that mudstones, 02 Mudd:Layout 1 03/07/2013 08:49 Page 39

excavations at netherfield farm, south petherton 39

particularly if they are angular in shape, can be misinterpreted as grog temper. Reviewing the grog-tempered fabrics identified amongst the Sweet Track assemblage may therefore prove informative.

STRUCK LITHICS by Hugo Anderson-Whymark (né Lamdin-Whymark) and Kathryn M. Price A total of 670 worked flint and chert fragments were found from the site of the Neolithic causewayed enclosure (Area 48A) and long enclosure (Area 48B). A detailed characterization and a catalogue of all pieces is retained in archive. Classification follows standard morphological descriptions (Bamford 1985, 72–77; Healy 1988, 48–49; Bradley 1999, 211–27; Butler 2005). All percentages cited exclude chips, as these can distort figures due to differential collection. condition The majority of artefacts are in reasonably fresh condition and free from cortication, although many exhibit edge-damage resulting from use. This indicates the majority of artefacts were deposited shortly after use and were not subjected to extensive trampling or disturbance. Only thirty-eight lithics exhibit evidence of burning (5.6% of the collection), but 301 lithics were broken (44.4% of the collection). The propor - tion of broken pieces is relatively high due to the presence of many flakes intentionally broken by percussion and flexion (Anderson-Whymark 2011). raw materials Raw materials suitable for knapping are scarce in the landscape here. Four lithic raw materials were identified comprising derived flint, Chalk flint, Greensand chert, and Portland chert. Flint from derived sources, such as fluvial gravels, was the most common, accounting 80.7% of raw materials. The second most common material was Greensand chert (16.5% of the group). This may have been obtained from numerous outcrops to the south, east and west, 10 km or more distant. Chalk flint and Portland chert were sparsely represented with nine (1.3%) and ten (1.5%) artefacts respectively from just two contexts in each case. The closest Chalk outcrops are c. 10 km to the south, but this raw material is likely to have been obtained from considerably further afield. The Portland chert is likely to have been obtained from, or close to, the Isle of Portland c. 50 km south-east. In addition, two flakes derive from the re-working of polished implements of light grey flint probably from a Chalk source. lithic assemblages Early Neolithic Causewayed enclosure The primary and secondary fills of the causewayed enclosure yielded only fourteen and nineteen lithics respectively, and the majority of these comprise undiagnostic flake debitage (Table 3). Two flakes in ditch 48h, primary fill 4830 refitted, but no other refits were found. The only retouched artefacts in the primary fills were three end 02 Mudd:Layout 1 03/07/2013 08:49 Page 40

40 excavations at netherfield farm, south petherton

table 2 The lithic assemblage from prehistoric pits on Site 48A

Pit Pit Pit Pit Material Category type 48104 48106 48268 48183 Total Derived Flint Flake 39 3 6 48 22 Bladelet 11 2 Irregular waste 11 Chip 11 Unfin. chisel arrowhead 22 Disc scraper 11 Retouched flake 11 Chalk Flint Serrated flake 22 Greensand chert Flake 11

Total 1 45 6 9 61

table 3 The lithic assemblage from the causewayed enclosure by phase of filling

Causewayed enclosure Raw not Primary Secondary 1st recut material Category type attributed fills fills fills Total

Derived Flint Flake 4 9 14 39 66 Blade 44 Bladelet 246 Blade-like 1225 Chip 145 Levallois core 11 Chisel arrowhead 11 End scraper 325 End and side scraper 11 Notch 11 Greensand Flake 11 Chert Blade 11

Total 5 14 19 59 97 02 Mudd:Layout 1 03/07/2013 08:49 Page 41

excavations at netherfield farm, south petherton 41

scrapers from ditch 48j, fill 48129, whose occurrence together may indicate that they were intentionally deposited. The ditch re-cuts contained a larger assemblage of fifty-nine artefacts, fifty-one of these from a single discrete group in ditch 48j, fill 48136. This group included a fine Levallois core (Illus. 26.1), a crude chisel arrowhead, two end scrapers (one intention - ally broken: Illus. 26.2) and a notched flake. The Levallois core and chisel arrowhead indicate that this lithic assemblage dates to the Middle or Late Neolithic, probably dating this re-cut to after 3300 cal. bc.

Early-Middle Bronze Age U-shaped enclosure The U-shaped ditch yielded twenty-one flints from its primary fills and fifteen flints from its secondary fills (Table 4). The only retouched was an end scraper and there was no material diagnostic of date.

Pits 48104, 48106, 48183 and 48268 Pits 48104, 48106, 48183 and 48268 yielded one, forty-five, nine and six flints respect- ively (Table 2). Pit 48106 (containing Early Neolithic pottery) was dominated by flakes, including several large ones (c. 60 mm) that appear to have been struck from the same core. However, none of these flakes refit and all exhibit use-damage indicating that they are not primary knapping debris, but pieces that were used and discarded. Similarly, two flakes of Chalk flint may have been struck from the same core, but these did not refit and both are well-used, silica-glossed, serrated flakes (Illus. 26.3). In addition to the serrated flakes, a disc scraper with a spur (Illus. 26.4) and an edge- retouched flake was found. The assemblage from this pit therefore represents a group of related flakes resulting from activities that perhaps included plant-working. Pit 48183 (fill 48184) contained two chisel arrowheads, including one with a refit- ting bulbar end demonstrating that it was abandoned in the process of manufacture

table 4 The lithic assemblage from the U-shaped enclosure by phase of filling

U Enclosure Raw material Category type Primary fills Secondary fills Total Derived Flint Flake 11 10 21 Blade 426 Bladelet 22 Blade-like 11 Chip 33 End scraper 11 Greensand Chert Flake 112

Total 21 15 36 02 Mudd:Layout 1 03/07/2013 08:49 Page 42

42 excavations at netherfield farm, south petherton

(Illus. 26.5). The second example is quite thick and exhibits use damage on the blade- edge (Illus. 26.6) and it may rather have been a wedge-shaped tool. The presence of a chisel arrowhead indicates a date in the Middle to Late Neolithic (Green 1980).

Long enclosure and nearby features In total, 445 lithics were recovered from the area of the long enclosure (Site 48B), the majority from the fills of the ditches 48s and 48t. Three pits also yielded stone artefacts (Table 5). ditch 48s The early fills of this ditch yielded only a few undiagnostic flakes that cannot assist in dating or interpreting the primary function of this ditch. However, the upper fills contained a greater number of artefacts, and fill 4872 alone yielded 180 artefacts (Table 6). The assemblage from fill 4872 included seven flakes of Chalk flint (4% of assem- blage), but artefacts of Greensand chert were comparatively scarce (nine, 5.2% of assemblage). In addition, eight pieces of Portland chert were recovered (4.6% of assemblage), including a fine Levallois flake with a facetted butt (Illus. 26.12). The artefacts of Chalk flint and chert were well-used indicating that they probably reached the site as tools. The derived flint includes several flakes resulting from the preparation of a Levallois core, although the core itself was absent. While the flakes may have been knapping debris, many show use-damage and were presumably employed for tasks nearby.

table 5 The lithic assemblage from pits on Site 48B by raw material and artefact type

Pit Pit E Neo Pit Raw material Category type 4857 4859 48644 Total Derived Flint Flake 27514 Blade 415 Chip 66 Spurred piece 11 Retouched flake 11 Greensand Chert Flake 66 Blade 44 Bladelet 22 Blade-like 11 Chip 11 Retouched flake 11

Total 2 11 29 42 02 Mudd:Layout 1 03/07/2013 08:49 Page 43

excavations at netherfield farm, south petherton 43

table 6 The lithics from long enclosure, ditch 48s by phase of filling

Ditch 48s

Secondary Tertiary Fourth Fifth upper Total Raw fill fill (burnt) fill fills material Category type 48611 4875 48605 4872 48604

Chalk flint Flake 77 Derived flint Flake 2 1 106 1 110 Blade 42 6 Bladelet 44 Blade-like 55 Chip 62 8 Rejuv. flake core face/edge 11 Levallois flake 11 Single platform flake core 33 Multiplatform flake core 11 Chisel arrowhead 33 Unfin. chisel arrowhead/blank 11 End scraper 11 2 Side scraper 33 End and side scraper 22 Spurred piece 22 Serrated flake 33 Polished flake knife 11 Retouched flake 77 Wedge-shaped tool 11 End scraper on Levallois flake 11 Greensand Flake 18110 chert Retouched flake 11 2 Portland Flake 55 chert Levallois flake 11 Unclass./fragmentary core 11 Retouched flake 11

Total 3 1 1 180 7 192 02 Mudd:Layout103/07/201308:49Page44

table 7 The lithics from long enclosure ditch 48t, by phase of filling

Ditch 48t Raw Category type Tertiary Middle fills Tenth fills Upper Pit fills Top Total material fill burnt fill fills* 48609 48638 48625 4866 48616 48643 48635 48617

Derived flint Flake 2 1 12 11 36 7 12 81

Blade 14117

Bladelet 132 6

Blade-like 11 2

Chip 1 1 1 33 36

Flake from ground implement 22

Tested nodule/bashed lump 11

Single platform flake core 11

Multiplatform flake core 11 2

Side scraper 11

Spurred piece 11

Other knife 11

Retouched flake 11 2

Hammerstone 11

Burin 11 02 Mudd:Layout103/07/201308:49Page45

Greensand Flake 2 3 13 12 8 38 chert Blade 43 29

Bladelet 14 16

Blade-like 12 3

Chip 11

Rejuv. flake core face/edge 1 1

Single platform flake core 11

Multiplatform flake core 11

Keeled non-discoidal flake core 11

Chisel arrowhead 11

Side scraper 11

Retouched flake 22

Core on a flake 11

Total 1 4 6 26 39 99 11 1 24 211

*Contexts 48607, 48618, 48608, 38633, 48634, 48639, 48641 and 48636. N.B. only one of these flints was burnt. 02 Mudd:Layout 1 03/07/2013 08:49 Page 46

46 excavations at netherfield farm, south petherton

In addition to a large collection of waste and utilized flakes, ditch fill 4872 yielded a broad range of retouched tools, accounting for 12.1% of the assemblage. There were seven scrapers of various forms, including one with a spur (Illus. 27.20) and another that was also used as a serrated flake, which was manufactured on a plunging (i.e. failed) Levallois-style flake (Illus. 27.19). Also found were three petit tranchet derivative (PTD) chisel arrowheads (Illus. 27.13–15) and one part-finished example with retouch along only one side (Illus. 27.16). Other tools include an unusual backed wedge- shaped tool (Illus. 27.17), the proximal end of a polished backed knife (Illus. 27.18) and numerous edge-retouched flakes. Nineteen artefacts (10.9% of the assemblage) were burnt, and seventy-four were broken (42.5% of the assemblage) with several pieces apparently snapped into wedge-shaped segments. These snapped segments may represent the debitage from manufacturing chisel arrowheads. The use of the Levallois knapping technique along with the presence of the chisel arrowheads and backed knife confirm a Middle Neolithic date for the assemblage. The presence of related knapping debris and a broad range of tools and utilized flakes, indicate that this material resulted from activities not far from the largely silted ditch.

ditch 48t In total, 211 lithic artefacts were recovered from ditch 48t (Table 7). It is again apparent that the primary fills yielded few artefacts while the upper fills yielded larger groups. In contrast to ditch 48s, Chalk flint and Portland chert were entirely absent, while Greensand chert forms a far greater proportion of the assemblage (sixty-five pieces, 37.4%). Several of the upper fills contained burnt material, but only three flints (1.7% of the assemblage) were burnt indicating that the lithics were not deliberately burnt. The large number of broken lithics (seventy-one, 40.8% of the assemblage) may reflect the manufacture of artefacts (e.g. chisel arrowheads), but no diagnostic manu- facturing debitage or unfinished arrowheads were recovered. The range of retouched tools comprised two side scrapers (Illus. 26.10), three edge-retouched flakes, a chisel arrowhead (Illus. 26.8), a spurred piece, a broken knife (Illus. 26.7) and a (Illus. 26.9), but they form only 5.8% of the assemblage. In addition, two flakes from grey polished flint were recovered from fill 4866. pits 4857, 4859 and 48644 Pits 4857 and 4859 each produced a small number of flakes, most of which bore signs of use. Pit 48644, the earliest dated feature on the site, contained a larger assemblage of twenty-nine artefacts dominated by utilized flakes, including a good proportion of Greensand chert. Three retouched artefacts were recovered. These comprise a spurred piece and two edge-retouched flakes none of which are chronologically diagnostic.

discussion The lithic assemblages from the primary ditch fills of these earlier prehistoric features are exceptionally limited and insufficient to characterize either the technology or activities associated with the early use of these monuments. In contrast, the lithic 02 Mudd:Layout 1 03/07/2013 08:49 Page 47

excavations at netherfield farm, south petherton 47

assemblages from the upper fills of the monument ditches and isolated Neolithic pits are more substantial and provide some insight into activities undertaken. Middle Neolithic radiocarbon dates from the upper fills of the long enclosure are consistent with the use of the Levallois knapping technique, and the presence of chisel - heads and polished . The frequent occurrence of Levallois reduction techniques raises a question over the products manufactured from Levallois flakes. Notably, none of the chisel arrow - heads in the assemblage from ditch 48t were manufactured from Levallois flakes, nor do any of the snapped flakes, which may have been debitage from the manufacture of chisel arrowheads, appear to be fragments of Levallois flakes. This indicates that chisel arrowheads were typically manufactured from regular flakes with an appropriate chisel edge, but this does not rule out the possibility that fine chisel arrowheads were being manufactured from Levallois flakes here, but not deposited. Studies have shown differences between arrowheads deposited in contexts and those present else- where (Devaney 2005). In this light, it is possible that the monument complex was considered unsuitable for the deposition of fine arrowheads manufactured from a Levallois flake. It can also be noted that the only example of a Levallois flake in the group is of Portland chert, and this was probably manufactured elsewhere. It may have been used as an unretouched tool such as a knife. It is possible that the Levallois flakes made from flint here were intended for use in this fashion rather than for modifica - tion, and may have been removed from site rather than deposited. However, the nature of the artefacts manufactured from the Levallois flakes knapped is unknown and the question here must remain open. The artefacts in the upper ditch fills were discrete groups of lithics, including knap- ping debris, much of which had also been utilized, and a broad variety of tools with an emphasis on scrapers and arrowheads. It is notable that lithic deposits in the ditches are of comparable composition to the assemblages recovered from some of the pits, which may indicate that the assemblages originated from similar activities and that the acts of deposition were guided by similar principles (Anderson-Whymark and Thomas 2012). In essence, the upper fills of the earlier monument ditches were receiving deposits of artefacts that might otherwise be buried in pits. Deposit 48136 in ditch 48j, appears to have been the contents of a discrete pit cut into the top of the much earlier ditch of the causewayed enclosure. The placement of discrete deposits of artefacts during the Middle Neolithic, both in the upper fills of the ditches and in the pits around the monuments, is a striking feature of the site. These deposits may be viewed in the terms of revisiting earlier monuments, with the acts of deposition serving to appropriate and reinvent these places. The revisiting of earlier Neolithic monuments in the Middle and Late Neolithic is common across southern England and many monu ments have yielded comparable artefact assemblages. For example, the upper fills of the chambers in West Kennet , Wilts., yielded deposits of Peter- borough Ware and associated lithics (Piggott 1962); the upper fills of a long barrow ditch at Badshot, Surrey, yielded a discrete group of Peterborough Ware associated lithics (Keiller and Piggott 1939; Lamdin-Whymark 2008); and, the upper fills of Staines causewayed enclosure, Surrey, contained chisel arrowheads and a small quantity of Peterborough Ware (Healey and Robertson-Mackay 1983; 02 Mudd:Layout 1 03/07/2013 08:49 Page 48

48 excavations at netherfield farm, south petherton

Robertson-Mackay 1987). Although they might superficially appear to result from mundane activities, the lithics from the Netherfield Farm causewayed enclosure and long enclosure were likely to have been selected and deposited at these specific locations as acts.

illustrated lithics (Illus 26‒27) 1 Flint Levallois core weighing 52 g. The scars of two Levallois flake removals can by observed on the top surface. The first removal ended in a hinge termination, but despite this the distal end of the core was not reworked before the second removal was made. This ensured the second flake would also end in a hinge termination, indicating this was the desired flake form rather than a knapping error. Causewayed enclosure ditch 48J recut, fill 48136. Later Neolithic. 2 Flint end scraper with intentional proximal break resulting from a blow against an anvil. Causewayed enclosure ditch 48J recut, fill 48136. Later Neolithic. 3 Serrated flake manufactured on a chalk flint flake. A slight silica gloss can be observed on the rear of the teeth. Site 48A, pit 48106. Neolithic. 4 Flint disc scraper with spur. Slightly irregular form, but bulbar end removed and the majority of the perimeter exhibited abrupt retouch. Site 48A, pit 48106. Neolithic. 5 Unfinished chisel arrowhead with refitting debitage manufactured from a flake of Greensand chert. Site 48A, pit 48183, fill 48184. Later Neolithic. 6 Unfinished flint chisel arrowhead or wedge-shaped tool. Bulb removed by retouch and distal end deliberately snapped. Site 48A, pit 48183, fill 48184. Later Neolithic. 7 Broken flint knife with invasive retouch. The break is probably intentional. Long enclosure ditch 48t, fill 4866. Later Neolithic. 8 Chisel arrowhead manufactured from a flake of dark brownish grey Greensand chert. Long enclosure ditch 48t, fill 48616. Later Neolithic. 9 Burin on the distal end of a broad flint flake. Long enclosure ditch 48t, fill 48643. Later Neolithic. 10 Double side scraper manufactured on a light orange grey Greensand chert flake. Long enclosure Ditch 48t, fill 48616. Later Neolithic. 11 Flint flake core weighing 53 g. The platform has been formed by the removal of flakes from several directions and bears affinities with Levallois style working. Long enclosure ditch 48t, fill 48616. Later Neolithic. 12 Levallois flake of fine black Portland Chert. Note the finely facetted platform. Burnt and broken. Long enclosure ditch 48s, fill 4872. Later Neolithic. 13 Flint petit tranchet derivative chisel arrowhead with slight modern damage to blade edge. Long enclosure ditch 48s, fill 4872. Later Neolithic. 14 Flint chisel arrowhead with slight barb on blade edge. Long enclosure ditch 48s, fill 4872. Later Neolithic. 15 Flint petit tranchet derivative chisel arrowhead. Burnt with slight damage to right hand edge. Long enclosure ditch 48s, fill 4872. Later Neolithic. 16 Unfinished flint petit tranchet derivative chisel arrowhead. Bulb removed by retouch onto the dorsal surface and the distal end has been snapped. Long enclosure ditch 48s, fill 4872. Later Neolithic. 17 Wedge-shaped flint tool. Medial segment of a deliberately snapped flake with semi abrupt retouch on the right hand edge; the left hand edge has use damage. Long enclosure ditch 48s, fill 4872. Later Neolithic. 02 Mudd:Layout 1 03/07/2013 08:49 Page 49

excavations at netherfield farm, south petherton 49

illus. 26 Flint (nos 1–12). Scale 2:3 02 Mudd:Layout 1 03/07/2013 08:49 Page 50

50 excavations at netherfield farm, south petherton

illus. 27 Flint (nos 13–20). Scale 2:3

18 Deliberately broken, polished backed flint knife. Small areas of striated polish to dorsal ridges and bulbar surface, but the blade edge is not polished. Backing retouch along right hand side and use-damage on left hand edge. Long enclosure ditch 48s, fill 4872. Later Neolithic. 19 Flint end scraper and serrated flake combination tool manufactured from a Levallois-style flake removed from a discoidal core. The removal, however, plunged and truncated the platform edge. Long enclosure ditch 48s, fill 4872. Later Neolithic. 20 Flint end scraper with a pronounced spur manufactured on a thick hammer flake. Long enclosure ditch 48s, fill 4872. Later Neolithic.

PLANT MACROFOSSILS AND CHARCOAL by Sarah Cobain A total of thirty bulk soil samples were processed and assessed for plant macrofossil and charcoal remains. These samples were taken from the Netherfield Farm Neolithic causewayed enclosure ditch; the Neolithic long enclosure ditches and pits; and the early medieval pits and fire-pits on the site of the Neolithic causewayed enclosure. Following the post-excavation assessment, eighteen of these samples were selected for plant macrofossil analysis and eight for charcoal analysis to provide additional 02 Mudd:Layout 1 03/07/2013 08:49 Page 51

excavations at netherfield farm, south petherton 51

information regarding the function of features sampled, activities undertaken, and to infer the composition of the local woodlands and flora (Carruthers 2009, 116, Gale 2009, 125). The processing and examination of the samples followed standard pro - cedures (Berggren 1981, Anderberg 1994 and Cappers et al. 2006). Nomenclature follows Stace (1997). The samples analyzed were well preserved with thirty-eight plant species and eleven tree/shrub species identified. As plant macrofossil and charcoal assemblages from the South West are relatively scarce, these samples provided a rare opportunity to investigate activities undertaken during the Neolithic and early medieval periods in the region. Identifications are presented in Tables 8–11. early neolithic pit 48644 Pit 48644 contained four carbonized hazelnut shells and a self heal (Prunella vulgaris) seed (Table 8). Charcoal consisted of frequent sessile/pedunculate oak (Quercus petraea/robur) with smaller quantities of alder/hazel and elm (Ulmus glabra) fragments (Table 9). The assemblage is typical for the period and suggests the consumption of wild foodstuffs. The self-seal, a grassland/disturbed ground species, may suggest areas of woodland clearance close to the site. early neolithic causewayed enclosure Ditch 48j contained a single indeterminate cereal grain and modern fat hen seeds and four additional samples from ditch segments contained no plant macrofossils (Carruthers 2009, 117–19, table 18). It is not possible to deduce information about this feature from this paucity of remains. middle neolithic long enclosure A total of thirteen samples were taken from long enclosure ditches 48t and 48s. Six samples contained no plant macrofossil material and one contained a single indeterminate tuber. A range of tree/shrub species, including carbonized hazelnut shells, cherry and sloe pips, crab apple, elder, hawthorn and ash seeds came from ditch 48t (fills 48607, 48608, 48628, 48636 and 48639, and fill 48643 of pit 48642 cut into the top of ditch 48t) (Table 8). Species from disturbed ground included carbonized hop trefoil and cleavers. A single carbonized pale persicaria seed indicates a marshland environment. Cultivated species comprised a single bread wheat (Triticum aestivum) cereal grain in deposit 48636. The excellent preservation and the relatively wide variety of species identified are rare in Neolithic assemblages. Charcoal recovered from fill 48639 in ditch 48t, pit 48642 and fill 4875 in ditch 48s are listed in Table 9. They consisted of alder/hazel, hazel, alder, sessile/pedunculate oak, ash, hawthorn/ rowan/crab apple (Maloideae species), cherry, blackthorn and willow/poplar. There were relatively large numbers of carbonized hazelnut shells from ditch 48t, with smaller amounts of elderberry and hawthorn seeds, and cherry and apple pips. Similar wild foodstuffs were found at the Neolithic long enclosure at Castle Hill, East Devon (Clapham 1999, 51) with hazelnut shells, sloe and elder present. These car - bonized remains are likely to represent discarded waste from the exploitation and 02 Mudd:Layout103/07/201308:49Page52

table 8 Plant macrofossil species identifications for pit 48644 and long enclosure

Feature E. Neo Pit Long enclosure Context number 48645 48607 48608 48643 48636 48628 48639 Feature number 48644 48606 48606 48642 48606 48628 ‒ Sample number 16 9 10 20 24 28 29 Habitat Sample volume (L) 34 21 22 34 34 8 16 Code Flot volume (ml) 57 11 16 37 25 27 21 Family Species Common Name

HSW Adoxaceae Sambucus nigra Elder 211 HSW Sambucus nigra Elder (modern) 1 D Amaranthaceae Chenopodium album Fat hen (modern) 11 HSW Betulaceae Corylus avellana Hazelnut 3 4 215 27 2 87 21 D Stellaria media Common chickweed (mod.) 11 G/D Fabaceae Trifolium campestre Hop Trefoil 2 G Lamiaceae Prunella vulgaris Self heal 1 D Malvaceae Malva spp Mallow (modern) 15 30 5 2 HSW Oleaceae Fraxinus excelsior Ash 1 cf E Poaceae Triticum aestivum Bread-type wheat 1 D Polygonaceae Persicaria spp Persicaria spp 1 HSW Rosaceae Crateagus monogyna Hawthorn 21cf 1 HSW Malus sylvestris Crab apple 1 HSW cf Prunus spp cf Cherry spp 11 HSW Prunus spinosa Blackthorn/sloe 1 cf D Rubiaceae Galium aparine Cleavers 1 Total macrofossils identified: 21 4 254 34 9 87 26

Key: HSW = Hedge/scrub/woodland species; C = Arable/crop species; D = Disturbed ground species; E = Economic species; G = Grassland species; M = Marshland species 02 Mudd:Layout103/07/201308:49Page53

table 9 Charcoal species identifications for pit 48644 and long enclosure

Feature Early Neo. pit Long enclosure Context number 48645 48645 48643 48639 4875 Cut number 48644 48644 48642 ‒ 4871 Sample number 16 33 20 29 30 Sample volume (L) 348341610 Flot volume (ml) 57 70 19 21 55 Family Species Common Name Betulaceae Alnus glutinosa Alder 2 Alnus glutinosa/Corylus avellana Alder/hazel 4 2 14 15 Betula spp Birch Corylus avellana Hazel 25 Fagaceae Fagus sylvatica Beech Quercus robur/petraea Sessile/pedunculate oak 27 76 6 40 Quercus robur/petraea Sessile/pedunculate oak h/w 92212 39 Oleaceae Fraxinus excelsior Ash 1 Pomoideae Maloideae spp * Hawthown/rowan/crab apple 19 14 Rosaceae Prunus avium/padus Wild/bird cherry Prunus spp Cherry spp 12015 Prunus spinosa Blackthorn/sloe 43 Salicaceae Salix spp/Populus spp Willow/poplar 2 Ulmaceae Ulmus glabra Elm 5 Indeterminate 6810

Total fragments identified: 46 100 74 60 79

* Crateagus monogyna/Sorbus spp/Malus sylvestris 02 Mudd:Layout103/07/201308:49Page54

table 10 Plant macrofossil species identifications for Site 48A early post-Roman pits

Feature Fire-pit Fire-pit Pit Pit Pit Context number 48266 48275 48274 48277 48278 48279 48145 48194 48172 Feature number 48265 48265 48272 48272 48272 48272 48144 48192 48170 Sample number 11 17 12 13 14 15 5 21 22 Habitat Sample volume (L) 40 25 20 36 32 7 32 41 34 Code Flot volume (ml) 44 65 69 109 112 38 88 98 50 Family Species Common name HSW Adoxaceae Sambucus nigra Elder 1 D Amaranthaceae Atriplex patula Common orache 1 D Chenopodium album Fat hen 33370679305 D Chenopodium album Fat hen (modern) 13 D Chenopodium polyspermum Many-seeded goosefoot 15 3 1 8 D Chenopodium rubrum Red goosefoot 7 D Chenopodium rubrum Red goosefoot (modern) 9 M Apiaceae Apium spp Fool’s-water-cress 41 C Bupleurum rotundifolium Thorow-wax 241 1 G Asteraceae Cirsium spp/Carduus spp Thistle spp 1 C Anthemis cotula Stinking chamomile 11330738 D Hieracium spp Hawkweeds 41 HSW Betulaceae Corylus avellana Hazelnut 212 D Brassicaceae Brassica/Sinapsis spp Mustard/Cabbage/Charlock 11403 D Caryophyllaceae Stellaria media Common chickweed 221 D Stellaria media Common chickweed (modern) 1 M Cyperaceae Carex spp Sedge 11 G Fabaceae Vicia spp/Lathyrus spp Vetches/vetchlings 233311212 4 G Lamiaceae Prunella vulgaris Self heal 2 E Linaceae Linum usitatissium Flax seed 1 cf E Poaceae Avena spp Oat 51 137 50 6 325 33 13 C Bromus sect bromus Chess 622 E Hordeum vulgare Barley(mostly cf 6-row hulled) 133 236 86 172 91 200 237 32 77 02 Mudd:Layout103/07/201308:49Page55

E Hordeum vulgare Naked barley 1 3 1131 E Hordeum vulgare Hulled barley 264144591 E Hordeum vulgare Barley 6-row hulled 23 220 124 166 194 184 57 12 10 E Hordeum vulgare Barley 2-row hulled 461451 E Triticum spp Wheat 43 3 E Triticum aestivum Bread wheat 19 134 91 26 168 254 65 E Triticum dicoccum Emmer 3 E Triticum spelta Spelt 7 549 E Triticum spelta Spelt glume bases E Triticum dicoccum/spelta Emmer/spelt 212 E Triticum dicoccum/spelta Emmer/spelt glume bases 15 E Triticum Wheat 63 E Triticum spp Wheat glume base 1 E Triticum spp Wheat rachis 1 E Poaceae Indeterminate cereal grain 155 45 121 249 126 181 94 36 E Poaceae Culm node 67 E Poaceae Straw 2 G/D Poaceae Grass spp 7 13 105 152 6 16 13 D Polygonaceae Fallopia convolvulus Black-bindweed 8 C Persicaria amphibia Amphibious Bistort 52732184513 M/D Persicaria lapathifolia Pale persicaria 51 4 2 D Polygonum aviculare Knotgrass 113422 D Rumex spp Dock 7169214533 D Rumex acetosa Common sorrel 76 D Rumex acetosella Sheep’s sorrel 1 M Rumex congolmeratus Clustered dock 363 6 D Rumex crispus Curled dock 2 HSW cf Prunus spp cf Cherry spp 2 D Rubiaceae Galium aparine Cleavers 2241 Total plant macrofossils identified: 334 600 688 1320 707 448 1204 477 264

Key: HSW = Hedge/scrub/woodland species; C = Arable/crop species; D = Disturbed ground species; E = Economic species; G = Grassland species; M = Marshland species 02 Mudd:Layout 1 03/07/2013 08:49 Page 56

56 excavations at netherfield farm, south petherton

consumption of wild food during visits to the site. In common with other Neolithic sites in southern Britain, Netherfield Farm shows very limited evidence for cereal cultivation at this time. The single bread wheat-type grain from the upper fill of ditch 48t may have been intrusive as bread wheat is not thought to have been widespread in Britain until the Roman period (Hagan 2006, 32). It has, however, been recorded at Hazelton North long barrow in Glos. (Straker and Wilkinson 2008, 68). The charcoal assemblage is represented by a dominance of oak in deposit 4875 from ditch 48s, in contrast to deposits 48639 and 48643 from ditch 48t which contained a much more varied selection of fuel woods. It is possible that the dominance of oak in ditch 48s was due to a burnt oak timber where the sample was taken. The varied selection of fuel woods, when considered in conjunction with the plant macrofossils, may suggest that fires were being used to prepare hand-collected wild foods. The similar assemblage from the Neolithic long enclosure at Castle Hill (Gale 1999, 194), with oak, hazel, blackthorn, Maloideae species and poplar/willow present, indicate that similar species were used as fuel here. Oak often dominates Neolithic assemblages, as at Castle Hill (ibid.). Indeed, oak-hazel woodland dominated the landscape in Somer - set in the fourth millennium bc (Rackham 2001, 29) and both these species are likely to have been easily accessible. The woodland surrounding the site at Netherfield Farm can be seen to have consisted of oak, ash and elm, with the under-storey and wood- land margins including hazel, hawthorn/rowan/crab apple, alder, cherry, backthorn, elder and poplar/willow.

early medieval pits on site of the causewayed enclosure Pits 48144, 48192, 48170, and fire-pits 48272 and 48265 all contained similar assem - blages of plant macrofossils (Table 10). These included arable weeds such as thorow- wax, stinking chamomile, vetches and chess. Species from disturbed ground were also identified, including Chenopodium species, hawkweeds, mustard/cabbage/char lock, black-bindweed, dock and cleavers. Grassland species included thistles, self heal and grasses. Fool’s-water-cress, sedge and clustered dock are marshland species, while hedges/scrub and woodland species included elder, hazelnut and cherry. Finally, cultivated species were recognized, including flax (Linum usitatissium), oats, barley, bread wheat (Triticum aestivum), emmer wheat, spelt wheat and cereal chaff (culm nodes, straw, wheat rachis and glume bases). Fire-pits 48272 and 48265 contained charcoal from alder/hazel, hazel, ash, Maloideae species, birch, wild/bird cherry (Prunus avium), cherry and blackthorn (Table 11). Early medieval communities usually cultivated a variety of crops, possibly as a strategy in case a crop failed in any given year (Pearson 1997, 3). The plant macrofossil assemblage from this site follows this trend. The most common cereal identified was six-row hulled barley, followed by bread wheat and oats, with smaller quantities of two-row hulled barley, spelt wheat and emmer wheat present. There were also small amounts of cereal chaff and arable weeds, which suggest that the fire-pits may have been used to dry the grain before threshing/winnowing to aid separation of the grain from the ears, or to harden the grain prior to milling and pounding. Since there is no in-situ burning in the lozenge-shaped pits, it is most likely they contained waste from 02 Mudd:Layout 1 03/07/2013 08:49 Page 57

excavations at netherfield farm, south petherton 57

table 11 Charcoal species identifications for Site 48A early medieval pits

Feature Fire-pit Fire-pit Context number 48278 48279 48275 Cut number 48272 48272 48265 Sample number 14 15 17 Sample volume (L) 32 7 25 Flot volume (ml) 112 38 38 Family Species Common name Betulaceae Alnus glutinosa Alder Alnus glutinosa/Corylus avellana Alder/hazel 14 15 25 Betula spp Birch 1 Corylus avellana Hazel 4333 Fagaceae Fagus sylvatica Beech Quercus robur/petraea Sessile/pedunculate oak Quercus robur/petraea Sessile/pedunculate oak h/w Oleaceae Fraxinus excelsior Ash 11 2 Pomoideae Maloideae spp* Hawthown/rowan/crab apple 17 60 4 Rosaceae Prunus avium/padus Wild/bird cherry 38 Prunus spp Cherry spp 57 1 Prunus spinosa Blackthorn/sloe 46 Salicaceae Salix spp/Populus spp Willow/poplar Ulmaceae Ulmus glabra Elm Indeterminate 10 5 Total fragments identified: 48 100 66

*Crateagus monogyna/sorbus spp/Malus sylvestris

the fire-pits. There is no discernible pattern to the differences in the quantities or species of cereals recovered from the burnt and unburnt pits and it is likely that the deposits in the unburnt pits represent debris from previous harvests. The dominance of barley is not usual for the early medieval period when bread wheat tended to be more common, for instance at West Cotton, Northants. (Campbell 1994, 67) and a fifth-century at Sladwick, Somerset (Campbell et al. 2007, 878). A dominance of barley was however found at early to mid-Saxon settle - ment of West Heslerton, North Yorks. (Carruthers and Hunter 2001, 9). It is likely that local arable regimes varied regionally. The damper climate in the South West, together with heavy clay soils, may have meant that a variety of more hardy cereals, such as barley, spelt and oats, were cultivated in case bread wheat crops failed. 02 Mudd:Layout 1 03/07/2013 08:49 Page 58

58 excavations at netherfield farm, south petherton

A single carbonized flax seed was recovered but it is difficult to determine whether this indicates that flax processing was taking place. Flax processing is a destructive process, and it is possible that it was being undertaken routinely, leaving only minimal remains. It is also possible that it represents a local weed. Flax was used to produce linen and from the plant fibres and linseed oil from the seeds (Gale and Cutler 2000, 152–53). The plant macrofossils suggest a predominantly arable environment with species such as stinking chamomile indicating a heavy soil. Other species present are indica - tive of a mixed disturbed and grassland environment and include mustard/cabbage/ charlock, pale persicaria, common chickweed, grasses and thistles (Rose 2006, 150, 168, 200, 463). These would most likely be found in uncultivated areas around the edges of fields. There are also a small number of species from a marshland environ- ment such as fool’s-water-cress, sedge and clustered dock (Rose 2006, 84, 163, 344), most likely brought to site from the nearby river margins. The charcoal assemblage was mixed. In fire-pit 48272 Maloideae was the most common represented, followed by alder/hazel. An absence of oak and an only minor presence of ash may be explained by the lower fire temperature required for drying cereals for threshing or milling. Species such as Maloideae or alder/hazel are anatomi - cally less dense and burn at lower temperatures than oak and ash (Gale and Cutler 2000, 34, 88–89, 120, 205; Grogan et al. 2007, 29–31, 48). The absence of oak and ash from this site may alternatively be attributed to increased pressure on woodland resources, meaning that communities would have had to rely on wood from hedgerows and on scrub that developed on abandoned Roman fields (Rackham 2001, 42). The range of species appears to be typical for this period, although it is likely that oak and ash were also present where woodlands existed. A similar assemblage came from the Church Field’s site, Shapwick Heath, Somerset in a slightly later, seventh to eighth-century, context (Gale 2007, 892). Here, alder/hazel, blackthorn, Maloideae, oak and willow/poplar were present.

ANIMAL BONES by Jonny Geber and Sylvia Warman Two burnt cow-sized and sheep-sized long bones and some fragmented sheep teeth were the only identifiable bones from the Neolithic features. All came from the long enclosure ditches. Soil samples from these ditches also yielded a number of unidenti- fiable fragments of bone, many of which were burnt. There were no bones from the causewayed enclosure or pits. Bronze Age ditches 47b, 47c, 47d and 47e yielded seventeen fragments (281 g) of bone. Of these, a cattle metatarsal was identified to species, the rest being cow-sized. A cow-sized humerus found in ditch 47e was an unusual size suggesting that it came from a large bull. Animal bones from seventh-eighth century ad ditch 48k and pit 48123 comprised an assemblage of 307 specimens (457 g), although only 18 of these were identifiable to species. Teeth, mandible fragments and lower limbs bones of cattle and caprovine were identified in the fill of ditch 48k. In terms of body parts, most specimens were teeth or mandible fragments, although some lower limb bones and metapodials were 02 Mudd:Layout 1 03/07/2013 08:49 Page 59

excavations at netherfield farm, south petherton 59

also present. Animal bone was also recovered from pit 48123, which contained cattle teeth, large mammal long bone and flat bone fragments, and further unidentifiable bones. The presence of both meat-rich and meat-poor elements indicates that the remains constitute both butchery and food waste and animals were likely to have been slaughtered on site. No slaughter or butchery cut-marks were identified on any of the bones, although this may be due to taphonomic factors because erosion of the periosteal surface and root etching was present on most of the fragments.

RADIOCARBON DATING OF THE CAUSEWAYED ENCLOSURE, LONG ENCLOSURE AND EARLY MEDIEVAL PITS by Frances Healy Of the twenty-five radiocarbon measurements from Netherfield Farm, sixteen were from the Neolithic monuments and associated pits, seven from the early medieval pits and two from the seventh-eighth century ditch. All measurements were obtained by Accelerator Mass Spectrometry (AMS) on single fragments of charred plant remains and animal bones. The samples were prepared and measured by the Waikato Radio- carbon Dating Laboratory, New Zealand and the Rafter Radiocarbon Laboratory, National Isotope Centre, Institute of Geological and Nuclear Sciences, New Zealand. The aims of the dating programme were (i) to refine the dating of the causewayed enclosure, (ii) to establish the date of the artefact-free pits and fire-pits and the undated ditch on the same site, and (iii) to exploit the opportunity offered by three successive horizons of burnt material in the ditches of the long enclosure for achieving increased precision by Bayesian modelling of radiocarbon dates. results and calibration Details of all the radiocarbon measurements from the site are listed in Table 12. The results reported there are conventional radiocarbon ages (Stuiver and Polach 1977), quoted according to the standards established by the Trondheim convention (Stuiver and Kra 1986). The calibrated date ranges (95% confidence intervals) were calculated by the maximum intercept method (Stuiver and Reimer 1986). The probability distributions of the calibrated dates (Illus 28, 31) were calculated by the probability method (Stuiver and Reimer 1993). All calibrations, as well as the analysis described below, were undertaken using the program OxCal v4.1 (Bronk Ramsey 1995; 1998; 2001; 2009) and the INTCAL09 dataset (Reimer et al. 2009). Quoted dates are at 95% probability unless otherwise stated. the bayesian approach The Bayesian approach to the interpretation of archaeological chronologies is described in varying degrees of detail by, among others, Buck et al. (1996), Bayliss and Bronk Ramsey (2004) and Bayliss et al. (2011). In essence it combines absolute dating evidence, such as radiocarbon measurements, with relative dating evidence, such as 02 Mudd:Layout103/07/201308:49Page60

table 12 Summary of radiocarbon dating results * all samples are single charred fragments

Calibrated Posterior density Posterior density Laboratory date estimate (95%) estimate (95%) number Date BP Context Identification*(95%) 13C(‰) from model 1 from model 2

Pit 48644 NZA-35810 4949±20 48645 Prunus sp. charcoal 3780–3660 bc -28.0 3775–3660 cal bc – NZA-35816 4951±20 48645 Corylus avellana shell 3780–3660 bc -26.1 3780–3660 cal bc –

Causewayed enclosure Wk-22462 5237±45 Ditch 48g, 4832 Quercus sp., ?heartwood 4240–3960 bc -25.6±0.2 4230–4190 (9%) – 4180–3960 (86%) cal bc NZA-35798 4854±20 Ditch 48j, 48129 Alnus charcoal 3660–3635 bc -26.5 3695–3680(3%) – 3670–3630 (85%) 3560–3535 (7%) cal bc NZA-35803 4782±20 Ditch 48p, first recut, Fraxinus charcoal 3640–3520 bc -27.0 3645–3620 (31%) – 48246 3605–3525 (64%) cal bc

Long enclosure NZA-35799 4475±20 Ditch 48s, 4875 indeterminate tuber 3340–3030 bc -28.2 3340–3210 (75%) 3340–3210 (87%) 3195–3150 (16%) 3190–3155(8%) cal bc 3140–3110 (4%) cal bc NZA-35801 4703±20 Ditch 48s, 4875 Quercus roundwood 3630–3375 bc -24.1 3630–3595 (13%) 3630–3595 (13%) 3530–3495 (22%) 3530–3495 (22%) 3465–3375 (60%) cal bc 3465–3375 (60%) cal bc NZA-35814 4518±20 Ditch 48s, 48605 Maloideae charcoal 3355–3100 bc -25.0 3280–3265 (2%) 3280–3265 (2%) 3245–3095 (93%) cal bc 3245–3095 (93%) cal bc Wk-22461 4699±44 Ditch 48t, 48614 Maloideae charcoal 3520–3340 bc -24.9±0.2 3620–3610 (1%) 3500–3350 cal bc 3525–3350 (94%) cal bc NZA-35805 4600±20 Ditch 48t, 48614 Prunus charcoal 3495–3350 bc -25.2 3500–3460 (47%) 3495–3465 (17%) 3380–3350 (48%) cal bc 3380–3350 (78%) cal bc 02 Mudd:Layout103/07/201308:49Page61

NZA-35804 4592±20 Ditch 48t, 48607 Corylus avellana charcoal 3490–3345 bc -25.9 3375–3335 (77%) 3375–3340 cal bc 3210–3190(9%) 3155–3130 (9%) cal bc Wk-22459 4454±44 Ditch 48t, 48608 Prunus sp. charcoal 3350–2925 bc -24.8±0.2 3350–3210 (64%) 3360–3150 (93%) 3200–3090 (31%) cal bc 3140–3110 2%) cal bc Wk-22460 4491±44 Ditch 48t, 48639 Pomoideae charcoal 3365–3020 bc -24.1± 0.2 3300–3040 cal bc 3310–3050 cal bc NZA-35806 4398±20 Ditch 48t, 48639 Prunus spinosa charcoal 3095–2920 bc -25.2 3265–3250 (3%) 3265–3250 (2%) 3100–3045 (92%) cal bc 3100–3045(93%) cal bc NZA-35809 4438±20 Ditch 48t, pit 48642, Corylus avellana shell 3315–3015 bc -23.0 3260–3250 (1%) 3095–3015 fill 48643 3100–3010 (94%) cal bc NZA-35815 4441±20 Ditch 48t, pit 48642, Corylus avellana shell 3315–3015 bc - 27.4 3100–3015 cal bc 3090–3020 cal bc fill 48643

Early Medieval pits Wk-22463 1505±35 Fire-pit 48272, Crataegus charcoal ad 430– 640 -25.4±0.2 cal ad 520–645 – fill 48277 NZA-35808 1510±15 Fire-pit 48272, barley grain ad 535–605 -23.8 cal ad 540–595 – fill 48279 Wk-22464 1465±35 Pit 48144, fill 48145 Cornus sp. charcoal ad 540–655 -23.9±0.2 cal ad 550–645 – NZA-35797 1430±15 Pit 48144, fill 48145 wheat grain ad 600– 655 -21.2 cal ad 595–650 – NZA-35800 1404±15 Pit 48170, fill 48172 wheat grain ad 615– 660 -20.6 cal ad 610–660 – NZA-35802 1428±15 Pit 48192, fill 48194 wheat grain ad 600–655 -22.5 cal ad 605–655 – NZA-35807 1557±15 Fire-pit 48265, wheat grain ad 430–555 -22.4 cal ad 465–570 – fill 48275

Early Medieval ditch 48k NZA-36713 1368±20 Fill 48162 Cattle calcaneous ad 641–675 -21.6 –– NZA-36712 1287±20 Fill 48162 Cattle mandible ad 670–772 -22.7 –– 02 Mudd:Layout 1 03/07/2013 08:49 Page 62

62 excavations at netherfield farm, south petherton

stratigraphic relationships between radiocarbon samples. An OxCal model specifies the known or assumed relative ages of the radiocarbon samples. The program calculates the probability distributions of the individual calibrated radiocarbon results (Stuiver and Reimer 1993) and then attempts to reconcile these distributions with the relative ages of the samples. This is achieved by repeatedly sampling each distribution to build up the set of solutions consistent with the structure of the model. This process produces a posterior density estimate of each sample’s calendar age, which often occupies only part of the calibrated probability. The program is also able to go beyond indi - vidual dates by calculating the ages of events that have not been dated directly, such as the beginning and end of a continuous phase of activity (when it is represented by several radiocarbon results), and for the durations of phases of activity or intervals between such phases. The resulting posterior density estimates, whether for individual dates or for estimated parameters such as the start of an episode of activity, are not abso lute; they are interpretative. The estimates will change as additional data become avail able or as the existing data are modelled from different perspectives. By conven- tion, they are always expressed in italics. samples and their contexts All the samples dated were single fragments of wood charcoal from short-lived species, or of charred grain or tuber, with three exceptions; one sample of oak charcoal initially dated from the causewayed enclosure (Table 12: Wk-22462) and two fragments of animal bone from the L-shaped ditch 48k (NZA-36713; NZA-36712). Those from the causewayed enclosure were effectively all that were available from a shallow, truncated ditch without distinct burnt deposits. Those from the pits in and beside the causewayed enclosure were all from substantial deposits of burnt material which, whether burnt in situ or dumped, seemed to represent single events. The individual bones from the primary fill of ditch 48k were the only samples available and, while potentially insecure as dating samples, they are mutually supporting and considered reliable indicators of the date of the feature in broad terms. Those from the long enclosure and pit 48644 in its entrance came, with two exceptions (Table 12: Wk-22461, NZA-35805), from coherent deposits of burnt material. The single fragments eliminated the risk of combining material of different ages in the same sample, the coherent deposits, where they were available to be sampled, reduced the probability of the fragments being re-deposited or intrusive, and the dating of more than one sample from the same context made it possible to check against this (Ashmore 1999). analysis and interpretation A model for the causewayed enclosure, and for pit 48644 and the long enclosure, is shown in Illus. 28. All three are modelled independently of each other. The bimodality (in some cases polymodality) of the probability distributions, not only for the individual dates but also for many of the estimated parameters, stems from wiggles in the calibration curve for these centuries. This reduces precision, which would be greater in earlier or later periods. 02 Mudd:Layout 1 03/07/2013 08:49 Page 63

excavations at netherfield farm, south petherton 63

The Causewayed Enclosure One date, measured on oak charcoal, which was possibly of heartwood, is inherently likely to be older than its context and is modelled as a terminus post quem (Illus. 28: Wk-22462). It is statistically inconsistent with the two dates on short life samples from other segments (Illus. 28: NZA-35798, -35803; T’=88.7; T’(5%)=6.0; m=2). Neither measurement was from a coherent deposit of charred material, so that both could have been re-deposited. On this slender foundation, the construction date of the enclosure might conservatively be estimated at 3645–3620 cal. BC (30% probability) or 3605–3525 cal. BC (65% probability), probably 3640–3625 cal. BC (25% probability) or 3585–3535 cal. BC (43% probability; Illus. 28: build causewayed enclosure).

The Long Enclosure and Pit 48644 Some 165 m to the north-east, the two ditches making up the long enclosure (48s to the south and 48t to the north) were separated by entrances at the north-west and south-east angles. The sequences in the two ditches may thus not have been identical, and they are therefore modelled separately. In ditch 48s there were two samples from context 4875 in the first horizon of burnt material (Illus. 28: NZA-35799, -35801; Illus. 9, section GG) and one from context 48605 in the second (Illus. 28: NZA-35814; Illus. 9, section FF). NZA-35799 and -35801 are statistically inconsistent (T’=65.0; T’ (5%)=3.8; m=1). Date NZA-35801, the older of the two, is therefore modelled as a terminus post quem for the context. The remaining two dates are in good agreement with the stratigraphy. In ditch 48t the sequence is longer. There are two statistically consistent (T’=0.4; T’ (5%)=3.8; m=1) results from context 48614, near the base of the ditch below the first horizon of burnt material (Illus. 28: Wk-22461, NZA-35805). The taphonomy of the samples is, however, problematic. Fill 48614 was a stony deposit, silted from interior of ditch terminal, among otherwise almost stone-free fills. Charcoal fragments in it were sparse and small, in contrast to those in the burnt horizons. Their statistical consistency argues for a single origin, but that single origin could have been a pre-existing context, whether the old land surface near the edge of the ditch or part of that land surface incorporated into a bank that silted into the ditch. Two models have therefore been constructed, identical except in that model 1 (Illus. 28) treats them as re-deposited in their context while model 2 (from which key parameters are shown in Illus 29 and 30 and listed in Table 13) treats them as contemporary with it. Higher up in the fills, there is a sequence of samples through contexts 48607, 48608 and 48639 (Illus. 11, section HH; Illus. 13), all in the second horizon of burnt material (Illus. 28: NZA-35804, Wk-22459, Wk-22460, NZA-35806). The two results from context 48639 (Illus. 28: Wk-22460, NZA-35806) are statistically consistent (T’=0.0; T’ (5%)=3.8; m=1). Finally, pit 48642, part of the third horizon of burnt material, yielded two statistically consistent (T’=0.0; T’ (5%)=3.8; m=1) results (Illus. 28; NZA-35809, -35815; Illus. 13). All the measure- ments from 48t are in good agreement with the stratigraphy in both models. The construction date, and hence the duration of use, of the long enclosure depend on the interpretation of the context of the two samples from context 48614. Alternative sets of results are listed in Table 13 and shown in Illus 29 and 30. 02 Mudd:Layout 1 03/07/2013 08:49 Page 64

64 excavations at netherfield farm, south petherton

illus. 28 Model 1 radiocarbon probability distributions of dates from the causewayed enclosure, pit 48644 and the long enclosure Each distribution represents the relative probability that an event occurred at a particular time. For each date, the distribution in outline is the result produced by the radiocarbon measurement alone and the distribution in solid black is based on the chronological model used. The other distributions correspond to aspects of the model. For example, the distribution ‘build long enclosure’ is the estimated date for the construction of the long enclosure on site 48b. The structure of the model is defined by the brackets down the left hand side of the diagram. ‘After’ denotes that a date has been modelled as a terminus post quem. Model 2 is identical, except that Wk-22461 and NZA-35805 from context 48614 are modelled as contemporary with their context rather than as termini post quos for the overlying layers. 02 Mudd:Layout103/07/201308:49Page65

table 13 Summary of radiocarbon results for long enclosure from models 1 and 2

Model 1 (samples from Model 2 (samples from 48614 treated as 48614 treated as redeposited)Model 1 contemporary with Model 2 Estimated parameters 95% probability 68% probability context) 95% probability 68% probability build long enclosure 3555–3335 cal bc (76%) 3460–3345 cal bc (66%) 3615–3360 cal bc 3515–3475 cal bc (17%) 3300–3130 cal bc (19%) 3215–3200 cal bc (1%) 3450–3365 cal bc (51%) use long enclosure 80 to 230 years (21%) 290 to 340 years 300 to 470 years 310 to 360 years (48%) 260 to 360 years (73%) 420 to 460 years (20%) 380 to 400 years (1%) abandon long enclosure 3250–3235 cal bc (1%) 3055–2970 cal bc 3080–2875 cal bc 3055–2980 cal bc 3085–2825 cal bc (94%) pit 48644 to construction 160 to 590 years 230 to 390 years 100 to 390 years 200 to 350 years of long enclosure

Pit 48644 to construction 30 to 210 years 70 to 90 years (10%) 30 to 210 years 70 to 90 years (10%) of causewayed enclosure 100 to 180 years (58%) 100 to 180 years (58%) 02 Mudd:Layout 1 03/07/2013 08:49 Page 66

66 excavations at netherfield farm, south petherton

illus. 29 Summary of start and end dates for the long enclosure, derived from models 1 and 2

illus. 30 Durations of use for the long enclosure, derived from models 1 and 2 02 Mudd:Layout 1 03/07/2013 08:49 Page 67

excavations at netherfield farm, south petherton 67

illus. 31 Probability distributions of dates from early medieval pits in and near the causewayed enclosure, modelled as forming a single phase of activity. The conventions are the same as in Illus. 28

Pit 48644, in the north-western entrance of the long enclosure, appeared on spatial grounds to be contemporary with it (Illus 4, 5, 6). Two statistically consistent (T’=0.0; T’ (5%)=3.8; m=1) results for short-life samples from its charcoal-rich upper fill, however, date to the later 38th or earlier 37th century cal. bc (Illus. 28: NZA-35810, -35816), providing a date for the context of 3760–3660 cal. BC, probably 3740–3690 cal. BC (52% probability) or 3685–3660 cal BC (16% probability; Illus. 28: 48645). This predates the construction of the long enclosure by 160 to 590 years according to model 1 and by 100 to 390 years according to model 2 (Table 13: pit 48644 to construction of long enclosure). It is also earlier than the very approximately estimated construction date of the causewayed enclosure. Early Medieval Pits and Fire-Pits There are seven dates from five features, the two pairs from single features, fire-pit 48272 and pit 48144, both being statistically consistent (Table 12). If all the dates are modelled as forming a single phase of activity, the overall agreement is good (Illus. 31). Following this model, the activity represented by these features lasted from cal. AD 415–565 (Illus. 31, start early medieval features) to cal. AD 620–725 (Illus. 31, end early medieval features), lasting 65–185 years. At 68% probability, the activity ran from cal. AD 495–550 to cal. AD 635–675, over a period of 80–125 years. 02 Mudd:Layout 1 03/07/2013 08:49 Page 68

68 excavations at netherfield farm, south petherton

Seventh- to Eighth-Century Ditch The two dates from the primary fill of ditch 48k are mutually and stratigraphically consistent and indicate that the datable artefacts from this feature are not contextually sound, being residual in the case of the prehistoric and Roman potsherds, and intrusive in the case of the medieval potsherd. discussion The Causewayed Enclosure Tentative as the date estimate for the construction of the enclosure is, it falls within an overall estimate for the construction and primary use of causewayed enclosures in southern Britain of 3765–3695 to 3330–3255 cal. BC (Whittle et al. 2011, 683, fig. 14.1). It also falls within an overall esitimate for the currency of South-Western style bowl pottery, which is present in this enclosure, of 3810–3690 cal. BC to 3340–3275 cal. BC (Whittle et al. 2011, 768, fig. 14.101).

Pit 48644 and the Long Enclosure Whatever the source of the samples from the upper fill of pit 48644, they are the earliest date for human activity in the area, given that the oak charcoal dated by Wk- 22462 in ditch 48g of the causewayed enclosure would already have been old when burnt. The interval between these dates and the construction date of the long enclosure is too great for there to have been any relation between them, unless the pit was marked in some enduring way. When it comes to the long enclosure itself, model 1, with its later start and a shorter duration, is the more persuasive of the two offered, since there is no evidence for the cleaning out of the ditch, and its clayey silts would surely have accumulated quite rapidly. The possibility that the samples from context 48614 were re-deposited, already inherent in the scant and scattered presence of charcoal in the context, is strengthened by NZA-35801 from context 4875, the sample for which was clearly re-deposited, showing that there was activity in the immediate area at about this time (Illus. 28). Whichever model is preferred, the enclosure is one of the most effectively dated of its kind, since, like monuments, long enclosures tend to be poor in both artefacts and organic material. The presence of the burnt horizons, and their effective sampling, has placed the monument in third quarter of the fourth millennium cal. bc and its abandonment at the turn of the fourth and third millennia. This accords with to the apparent later fourth millennium cal. bc age of other, more imprecisely dated examples, including one of two smaller long enclosures on Castle Hill, Devon, where a bulk sample of charcoal from short-lived species from a single surviving ditch fill was dated to 3630–3190 cal. bc (4630±50 BP; AA-30670) and an unidentified charcoal sample from a secondary fill elsewhere in the ditch to 2920–2620 cal. bc (4220±60 BP; Beta-78183; Fitzpatrick et al. 1999, 18–68). Farther away, a long enclosure of comparable dimensions to Netherfield Farm, that at Raunds in Northants., has an estimated construction date of 3350–2890 cal. BC, the breadth of which reflects the paucity of available samples (Harding and Healy 2007, 98). The later fourth 02 Mudd:Layout 1 03/07/2013 08:49 Page 69

excavations at netherfield farm, south petherton 69

millennium is also the period of cursus monuments, of comparable plan to long enclosures but often much larger. They are equally difficult to date, so that only a very approximate estimate can be offered, for construction between 3640–3380 and 3260–2920 cal. BC at 95% probability (Barclay and Bayliss 1999, 25). The dating of the Netherfield Farm long enclosure extends to its associated worked stone, including chisel arrowheads (found in the context dated by NZA-35799 to the early fourth millennium bc), an edge-polished knife, and the use of , and to the presence of Peterborough Ware. The Peterborough Ware association is replicated in a second, undated long enclosure on Castle Hill (Fitzpatrick et al. 1999, 24, 63), as well as in cursus monuments (Barclay and Bayliss 1999, 21).

Early Medieval Features The dating of six features without diagnostic artefacts to between the fifth and eighth centuries cal. ad emphasizes that activity of this period may sometimes be identified only by absolute dating.

SUMMARY OF ANALYTICAL REPORT ON GEOARCHAEOLOGICAL SAMPLES FROM THE NEOLITHIC LONG ENCLOSURE AND EARLY MEDIEVAL PITS by Keith Wilkinson, Nick Watson, Stuart Black, Geoff Warren and Nathalie Marini Geoarchaeological samples were taken from the Neolithic long enclosure (ditch 48s at the terminal (Illus. 9 section FF)), and from early medieval pits 48144 (Illus. 21, section RR) and 48221 (section not illustrated) on the site of the causewayed enclosure (for details, see Wilkinson et al. 2011). The aim of the analysis was to investigate sediment formation processes and as far as possible, the function of the features. The techniques used comprised grain size analysis, magnetic susceptibility, and major element identification using inductively coupled plasma optical emission spectroscopy (ICP- OES). long enclosure ditch, terminal of 48s An overall assessment of the data indicates that the lower fills, below burnt layer 48610, show natural sediment formation processes with little anthropogenic input except in the upper part of fill 48611 which showed higher levels of phosphorous, consistent with human or animal activity. early medieval pit 48144 The fills are largely derived from the geological substrate, although the basal fill may have incorporated a lining of sand and gravel (rather than of clay). The magnetic susceptibility and geochemical readings show no in situ burning. The red and black discolouration of the lower fill (48145) is likely to have been the result of natural weathering. 02 Mudd:Layout 1 03/07/2013 08:49 Page 70

70 excavations at netherfield farm, south petherton

early medieval pit 48221 The magnetic susceptibility measurements clearly demonstrate in situ burning in the base of the pit. The heating caused the surface of the geological substrate to develop a reddish colour and was of sufficient intensity to fuse clay-silt and sand grains in the lower fill. It is likely that the Curie point of the key iron minerals was exceeded, suggesting a temperature in excess of 650o Centigrade. High concentrations of zinc, copper and silver in particular were observed, with these elements more than twice as high as in the geological background. In the absence of indications of organic debris from human or animal waste, this might be a product of metalworking.

IRONWORK AND IRONWORKING RESIDUES FROM EARLY MEDIEVAL FEATURES iron binding strip by E. R. McSloy Three fragments from an iron binding strip, 45 mm wide and 160 mm long with three rivets in situ, came from the primary fill 48124 of pit 48123. The main interest relates to its likely seventh- or eighth-century dating. An early medieval date is plausible; the use of riveted binding strips as structural fittings or furnishings is common across the Roman and medieval periods. Late Roman and early medieval parallels include examples from Cannington, Somerset (Rahtz et al. 2000, 345, fig. 234 no. 61).

ironworking residues by Tim P. Young The hand-picked macroscopic slag material amounted to three fragments, two from ditch 48k (fill 48138) and one from pit 48123 (fill 48124). The piece from context 48124 is certainly incomplete, and the smaller piece from context 48138 is probably so. All three pieces contain almost identical slag and all three may be interpreted as smithing hearth cakes, in each case comprising a dimpled basal surface, overlain by dense but vesicular slag, with a thin veneer of paler, more glassy, slag on the upper surface. Where visible, the fuel employed was charcoal. The larger specimen from ditch 48k is apparently a whole, if fragmented, example. The weight of 200 g may have been reduced a little from the original weight, because of leaching, but is a typical weight for a smithing hearth cake derived from black- smithing (the end use of iron for forming or repairing artefacts). The cake appears to be transverse and formed along the wall below the blowhole, without signi ficant attachment to the wall. Alternatively, the elongate form may have been accen tuated by folding, if the cake was extracted from the hearth when still hot and plastic. The slags are indicative of ironworking in the general area. They are not indicative of any particular date, and have a potential range, based on their intrinsic nature, from the early pre-Roman to modern, although the increased use of coal and coke rather than charcoal as fuel since the eighteenth century and the progressive 02 Mudd:Layout 1 03/07/2013 08:49 Page 71

excavations at netherfield farm, south petherton 71

introduction of iron tuyères since the medieval period implies that a post-medieval date is unlikely. The seventh/eighth-century radiocarbon dates from ditch 48k are therefore compatible with these finds. The sieved residues from ten sediment samples from the early medieval fire-pits and pits (excluding pit 48221) were examined using a low-powered binocular microscope. The residues were dominated by natural rock grains. Their magnetic properties would indicate they have been heated but there is no evidence to link that with metallurgical activity.

DISCUSSION This is the first such early prehistoric complex found in Somerset. Nationally, around seventy causewayed enclosures were recorded by English Heritage’s overview (Oswald et al. 2001) most in southern England, the closest certain example being Whitesheet Hill, Wilts., although Hembury, Devon, is almost equidistant (ibid., fig 1.1). Only the enclosure underlying the at Hembury, and the enclosure at Raddon Hill in Devon, are certain identifications lying further west. The Netherfield Farm (South Petherton) enclosure, at 0.25 ha, is one of the smallest discovered, similar to that at Radley, Oxon., which at 60 m across is among the smallest listed by English Heritage (Oswald et al. 2001, 73), and smaller than that at Down Ampney, Glos. (90 m across). The Netherfield Farm enclosure is comparable to both of these in its simplicity of design (Illus. 32). The ditch circuit is incomplete, whether due to later truncation is not known, but the ‘unfinished’ appearance of causewayed enclosures is not uncommon among this class of monument and there can be no doubt as to its identification or its early Neolithic date. The ditches, although shallow and probably much truncated, have a dished or relatively flat-based profile typical of this type of monument (Oswald et al. 2001, fig. 3.8). The monument is not sited in an elevated location and difficult to envisage as a prominent landmark. However, Netherfield Farm is typical of a num ber of cause wayed enclosures in occupying a slight tilt in the land, in this case towards the west, with a directed rather than panoramic view, and in English Heritage’s terminology is ‘lowland oriented’. The dominant entrance faces north-west and there is an approximate symmetry along a south-east/north-west axis, although no particu lar significance can be attached to these orientations. The date of construction of the Netherfield Farm causewayed enclosure has been modelled at 3645–3525 cal. BC from two radiocarbon dates, one from a primary fill and one from a re-cut (Illus. 28). For the long enclosure, the availability of a sequence of charred wood in the ditches has enabled a more detailed model of construction, use and abandonment to be proposed for this monument (Table 13). This has been cal- culated at 3555–3130 BC (95%) for construction and 3085–2825 BC (94%) for abandon- ment following use for around 260–360 years (73%). Despite a degree of precision to the dating of these two monuments, the sequence of activity at the complex as a whole is poorly understood and there is a lack of clear indication of the use for any of the monuments. The causewayed enclosure is not the earliest feature on the site. Two mutually consistent dates of c. 3780–3660 bc (NZA-35810; NZA-35816) from pit 48644 in the 02 Mudd:Layout 1 03/07/2013 08:49 Page 72

72 excavations at netherfield farm, south petherton

illus. 32 Comparative plans A. South Petherton, B. Radley (Oxon.) and C. Down Ampney (Glos.). B and C after Oswald et al. 2001. Scale 1:5000

north-west entrance of the long enclosure indicate activity here pre-dating the causewayed enclosure by 30 to 210 years (Table 13). Despite the pit’s location, the dates show that it does not seem to have been related to the construction of the long enclosure as might have been suspected. Although it did not show the characteristics of a post-hole it can be assumed that activity here in some way marked the site for much later monument construction. There is no indication as to what this activity was, but, typically for the period, the charred plant remains comprised predominantly hazelnut shells and oak charcoal. There is an inconsistently early date on oak roundwood from ditch 48s of the long enclosure (NZA-35801) that may have been re-deposited from earlier activity, and some of the undated pits may belong with this ill-defined pre-enclosure phase. The initial excavation of the causewayed enclosure ditches may have been a single phase of activity, although this is essentially unknown. There has been doubt as to whether causewayed enclosures should be viewed as designed monuments, or whether they were always ‘projects in progress’, with no overall design in mind and ditch segments dug on an occasional basis until the monument lost its significance (Oswald et al. 2001, 75). This latter view has not found support in the radiocarbon modelling of British enclosures (Whittle et al. 2011; Frances Healy pers. comm.), 02 Mudd:Layout 1 03/07/2013 08:49 Page 73

excavations at netherfield farm, south petherton 73

illus. 33 Comparative plans A. Southwick (Northants.), B. Roughton (Norfolk) and C. Buckland (Oxon.). After Oswald et al. 2001. Scale 1:5000

although it is possible that construction timescales are too short to be detected this way. Elsewhere there is evidence from Staines, Surrey, where two fragments of the same human arm came from different sections of ditch, that the monument was probably of one phase (Bradley 2004, 118). At Windmill Hill, Wilts., joining potsherds from different ditch circuits led the excavator to suggest activity in different parts of the monument at broadly the same time (Smith 1965, 14). However, a note of caution is offered by Pryor regarding similar situations at Etton, Cambs., where, despite pottery from different contexts refitting, residuality could not be discounted (Pryor 1998, 376). At Netherfield Farm, the initial enclosure may have been of one phase, and essentially ‘complete’, since the subsequent phases of activity involved re-digging at least parts of the existing ditch segments, rather than completing the circuit (if the open north-western side were an original feature), or adding more circuits. The 02 Mudd:Layout 1 03/07/2013 08:49 Page 74

74 excavations at netherfield farm, south petherton

re-digging may have been widespread and may have resulted in the joining up of small segments to create longer ditches, although the uniform nature of the infill deposits made them difficult to distinguish one from another and they were not everywhere defined with confidence. The radiocarbon dating, such as it is, suggests that re-cutting might not have been undertaken much after the original ditches were dug (Table 12; NZA-35798, -35803), but there is contrary evidence from the flint assemblage from a recut in ditch 48j (Illus. 8, section DD) which contains a Levallois core and a chisel arrowhead unlikely to date before cal. 3300 bc (Anderson-Whymark this article). Of probably similar date is a sherd of Middle Neolithic pottery from the recut of ditch 48p (Illus. 8, section BB). Thus, re-digging of the largely silted in ditches appears to have continued until the long enclosure was built, and may have been undertaken concurrently with activity at that new monument. The material remains from the causewayed enclosure were meagre and give no clear indications of the activities carried out. The prevalent view of causewayed enclosures is that they were for occasional, perhaps seasonal, gatherings of individuals and groups who participated in activities that included the digging and re-digging of ditch segments for the placement of cultural material. There is, however, a diversity of evidence that can support other views. The quantity of material from Staines, for instance, led the excavator to envisage the site as a permanent settlement (Robertson- Mackay 1987, 124–25), while a defensive function might be surmised from the association of arrowheads and burnt timbers at Crickley Hill, Glos. (Dixon 1988, 82). The possible funerary significance of the Main Causewayed Enclosure at , Dorset, is evident from the human remains in the ditches (Mercer and Healy 2008). The absence of surviving bone in the Netherfield Farm enclosure greatly restricts any interpretation, so feasting or funerary roles cannot be dismissed even though there is no firm evidence for them. A degree of patterning to deposits in the ditches is present, with most pottery and flintwork coming from restricted locations, particularly the terminals of ditches 48h, 48j and 48q (Illus. 34). It appears that this would have derived from activities carried out close to these locations, perhaps simply because internal banks (which may have been present, although there was no direct evidence for them) largely prevented material finding its way into the middle ditch sections. There is little evidence for ‘placed’ deposits generally, the fragmentary nature of the pottery, with few joining pieces, being characteristic of more random collections of material from nearby, rather than items deliberately deposited. In a few cases there are small groups of flint, such as three scrapers from ditch 48j, which may suggest intentional placement. Spatial patterning has been observed at other causewayed enclosures such as Etton, where it has been proposed that there are left and right divisions of activity corresponding with funerary rites on the one hand and episodic settlement on the other (Pryor 1998, 367–69). A reconsideration of the Staines enclosure suggests that there was a distinction between depositions of pottery and flintwork predominantly in ditch terminals in the north and north-western part of the enclosure, and human remains in the southern and south-eastern area (Bradley 2004, 116–19). The deeper ditches of the Netherfield Farm long enclosure contained more material, particularly charred remains, but the monument presents its own problems 02 Mudd:Layout 1 03/07/2013 08:49 Page 75

excavations at netherfield farm, south petherton 75

illus. 34 Finds distributions, causewayed enclosure. Pottery quantification is by weight (g); flint quantification is by number. Scale 1:1000 02 Mudd:Layout 1 03/07/2013 08:49 Page 76

76 excavations at netherfield farm, south petherton

of interpretation. Even nomenclature is not straightforward since this is a type of monument that might be classified as a large mortuary enclosure. In common with most ‘mortuary enclosures’ there is no evidence for a mortuary function and the more neutral term ‘long enclosure’ is preferred, putting the monument into the ‘cursus continuum’ of overlapping types that do not fall naturally into classes (Healy pers. comm.). The Netherfield Farm long enclosure could not have enclosed a mound made from the upcast of the ditches, which were too far apart, and while it seems likely to have had a bank on one side or other, there was no strong evidence of this. Superficially, it bears a resemblance to the Inchtuthil long enclosure, Perths., both in its size (at 50 m long and 10 m wide a little smaller than Netherfield Farm), and the dark nature of the upper ditch fills which in both cases were rich in charcoal (Barclay and Maxwell 1991). There is some similarity of detail as well, particularly the layers of charcoal and scorched earth which were held to be evidence of a burning fence-line that had been deposited in the ditch (ibid., 38). At Inchtuthil there was good evidence that the ditches had held a series of upright posts which had been burnt, but there was no such evidence from Netherfield Farm where the wider, flat-based, ditches were not likely to have been for post placements, and the sequence of fills for the most part indicated a natural accumulation of silts. The lower fills of both ditches 48s and 48t were largely sterile, but intermittent thin layers of charcoal, comprising mostly oak in 48s, suggest there may have been deposits of burnt or burning wood deriving from activity within the ditches or outside them. In terms of material remains it appears that activity increased immensely, and possibly changed in nature towards the end of the use of the ditches. When ditch 48t had largely infilled, a series of ‘fire-pits’ of more than one phase were dug along the centre line associated with charcoal and much larger quantities of flint and pottery. These pits were too broad in relation to their depths to have been post-holes, and the scorched interface with the lower deposits indicates their use as hearths. It is also notable that the charred remains included a range of tree and shrub species as well as seeds of fruits such as hazelnuts, cherry, sloe, crab apple and elder. At face value this may indicate gatherings for the preparation and consumption of wild foods. The large collection of worked flint and chert from the upper ditch fills does not easily translate into specific activities. However, the common presence of deliberately broken flakes, which may have been blanks for chisel arrow - heads, and the less common although notable collection of complete and unfinished chisel arrowheads, may indicate a particular emphasis on making . Activities were perhaps similar to those envisaged at Windmill Hill causewayed enclosure where a similar range of charred remains, together with large groups of cattle bones, have led to an interpretation of that site as a location for gathering and feasting at significant social events (Whittle et al. 1999, 384–85). Also of interest is a possible patterning of pottery distribution showing mutually exclusive distributions of Fengate ‘jar’ forms (in ditch 48s) and Ebbsfleet/Mortlake bowls (in ditch 48t) which may relate to distinctive uses to which these vessels were put. This pattern has an echo in the different ditch dimensions, fill sequences and distribution of Greensand chert (Illus. 35) and may sug- gest a slightly different range of practices undertaken at each ditch or side of the enclosure. Inclusions in the pottery fabrics indicate that both local and moderately distant sources of raw material had been used, with quartzite and chert suggesting 02 Mudd:Layout 1 03/07/2013 08:49 Page 77

excavations at netherfield farm, south petherton 77

illus. 35 Finds distributions, long enclosure. Pottery quantification is by weight (g); flint quantification is by number. Scale 1:1000 02 Mudd:Layout 1 03/07/2013 08:49 Page 78

78 excavations at netherfield farm, south petherton

manufacture in the Mendip area or further afield. The use of nodular flint, Greensand chert and Portland chert as raw material for artefacts also suggests contacts with, or visitors from, some way away. While this type of monument is not uncommon in the South West, here, as elsewhere in Britain, there is considerable difficulty understanding how they were used. Two long enclosures at Castle Hill, Devon, have been interpreted as a ‘long mortuary enclosure’ and the terminal of a cursus whose length remains unknown (Fitzpatrick et al. 1999, 213–16). The ‘long mortuary enclosure’ is of a similar date but of slightly smaller dimensions than the Netherfield Farm long enclosure with a shallow, discontinuous ditch open at its east end. Its interpretation as a monument associated with secondary mortuary rites is based upon its similarity of form to long barrows as well as the lack of anything definable as domestic occupation. Pairs of long enclosures have also been identified through aerial photography at North Tawton and Nether Exe in Devon (Griffith 1985). While these regional comparisons may be valid in a formal sense, it is perhaps worth emphasising the distinctiveness of the Netherfield Farm monuments, where the relative abundance of evidence for occupation at the long enclosure, and the sparseness of remains at the causewayed enclosure, is the reverse of normally encountered patterns (J. Pollard pers. comm.) Following the abandonment of the long enclosure perhaps around 3000 cal. bc and at latest some time before c. 2800 cal. bc, the focus of attention may have shifted to the U-shaped enclosure adjacent to the causewayed enclosure. In many ways the U-shaped enclosure is the most difficult to understand. The date for the construction of this feature is not clear, the pottery from the primary fills only datable to the Early to Middle Bronze Age. The absence of Late Neolithic or Beaker pottery from the site may indicate that there was no continuity of activity, but the position of the U-shaped enclosure suggests an active respect for the Neolithic monument, and like it, the earlier monument had an open western side facing the Lambrook Brook. The posi- tion ing of a pit centrally on the long axis, containing Middle to Late Neolithic flint- work, suggests that this feature had a role in determining the location of the U-shaped enclosure. There is a curious similarity with the way in which pit 48644 appears to have influenced the layout of the much later long enclosure. It is perhaps a significant aspect of this site that the biographies of the individual ‘monuments’ appear to have been linked by far less conspicuous pits, showing that activity may have involved a complex process of demarcation and re-iteration throughout the complex as a whole. The ditches of the U-shaped enclosure seem to have been dug in different phases as they are of different characters with only the three ditches on the southern and eastern sides apparently dug for palisades, although it is possible that all three sides were enclosed by timbers. Palisade enclosures of various forms are known from the Neolithic and may have formed screened ‘sacred precincts’ for purposes that remain obscure (Whittle 1997). The palisade at Haddenham causewayed enclosure (Cambs.) formed an integral part of that monument as an inner concentric line behind the causewayed ditch (Evans and Hodder 2006). This and the examples from West Kennet, Wilts. (Whittle 1997) and Mount Pleasant, Dorset (Wainwright 1979) are, however, on a far more monumental scale and may not have any link with the later and smaller type here. In the present context it is possible that it was contemporary 02 Mudd:Layout 1 03/07/2013 08:49 Page 79

excavations at netherfield farm, south petherton 79

with the ring-ditches that were aligned on the long enclosure and may therefore have had a role in mortuary preparatory to burial. In view of the apparently inconspicuous size and location of the causewayed enclosure, it is surprising that it appears to have become the focus of monument construction throughout the Neolithic and into the Bronze Age. The site is certainly not a ‘landmark’ inviting persistent reference like that of many upland causewayed enclosures, and neither does it lie, like monuments elsewhere, within a major river valley or other conspicuous routeway that are thought to have guided the movements of early farmers (Oswald et al. 2001, 107). The middle Neolithic long enclosure lies 150 m distant and why this location was preferred to somewhere more adjacent is not clear. It is possible that significant Neolithic features, without any clear magnetic signal and therefore not defined, occupied the intervening ground and perhaps also the higher unsurveyed ground to the east. There would appear to be a likelihood of features in the field immediately north of Netherfield Farm where aerial photographs suggest the presence of one large and one small circular feature (SHER DAP YT25–28). The cropmarks cannot, however, be interpreted with any confidence and it is perhaps significant that none of the features identified in the geophysical plot are evident in the photographs. That the causewayed enclosure remained visible long after its abandonment is shown by the likely re-cutting of some of the causewayed enclosure ditches in the Middle Neolithic period and the location of the Bronze Age U-shaped enclosure immediately adjacent to it. The ring-ditches, aligned with respect to the long enclosure are likely to have been for Bronze Age and imply a direct continuity of ritual observance, if not mortuary practice, at the site. The excavated Bronze Age ditches at the northern end of the site appear to be linked physically and chronologically to the ring-ditches, although a domestic rather than funerary purpose seems likely from the form of these features and the associated pottery. There are several examples of monuments grouped in this way. Sites illustrated by English Heritage include Buckland in Oxon., Roughton in Norfolk, and Southwick in Northants. (Illus 32 and 33). In none of these cases is the identification of associated features secure, and no precise comparisons can be made. The monument complex at Barrow Hills, Radley in Oxfordshire has been analyzed in detail and it has been suggested that the development of the linear Bronze Age barrow cemetery was conditioned by ‘readings’ of the existing cultural landscape, including the Abingdon causewayed enclosure and other nearby Neolithic barrows and other features, although these did not necessarily follow straightforward alignments (Garwood 1999, 298). There are also a few examples of an evident link between causewayed enclosures and later , such as at Etton, Cambs., where the causewayed enclosure may have provided the focus for the later ‘ritual landscape’ (Pryor 1998, 110–12, 373, fig. 117). There are a number of other instances of the association of round barrows with causewayed enclosures, but for topographically prominent sites such as Windmill Hill, and Whitesheet Hill, Wilts., this may be coincidence (Oswald et al. 2001, 137). The configuration of prehistoric features at Netherfield Farm cannot be accounted for in this way. There appears to have been a direct relationship and spatial respect between the causewayed enclosure and succeeding monuments that implies abiding significance for a location that is without any obvious topographic 02 Mudd:Layout 1 03/07/2013 08:49 Page 80

80 excavations at netherfield farm, south petherton

distinctiveness. Why this location was chosen and how it was marked over such a length of time are matters of conjecture. The causewayed enclosure would presum - ably have retained visible banks and ditches, although these were on a relatively small scale. The long enclosure, by contrast, was defined by quite substantial ditches for this category of monument although their importance in general terms has been linked to their extent and effort taken to construct them, rather than their imposing visibility (Loveday 2006, 143). Places that became foci for repeated actions over the long term are seen as having been anchored in the sacred geography of prehistoric peoples, and may for instance have been the physical setting for ancestral myths or social memories (Bradley 2002, 32–34). The re-occupation of the site in the fifth or sixth century ad is interesting additional evidence for activity in this enigmatic period and would not have been identified without material suitable for radiocarbon dating. There is widespread if not prolific evidence in the region for the reuse of prehistoric in the immediate early medieval period, although these are normally considered to be elite residences because of the material evidence that sometimes includes a range of imported luxury items and metalwork (Burrow 1981, Dark 2000, 136–49; Webster 2008, 171–75). Leav ing aside the uncertain status of the occupation at Netherfield Farm, it does not seem likely that the causewayed enclosure would have been of sufficient prominence to be chosen with the need for defence or enclosure in mind, and the siting is more likely to have been an opportunistic reuse of a visible landscape monument that provided a framework for the activities undertaken. That the causewayed enclosure was still visible as an earthwork is implied by the location of the seventh/eighth- century ditch 48k. The features include fire-pits that may have been used for drying crops, and there is a good range of evidence for the types of cereals cultivated and the wood used as fuel, which appears to derive mainly from hedgerow or scrub species. The form of the fire-pits does not seem to have been specific to a crop-drying func- tion, and there are difficulties in ascribing to them a primary purpose. The seventh/ eighth-century ‘grain driers’ at Chantry Fields, Gillingham, Dorset, were large, rec - tangular, stone-lined pits with the capacity to separate the fire from the grain, assumed to occupy an above-ground structure, to prevent burning of the crop (Heaton 1992). It is not at all clear that the Netherfield Farm pits would have operated in the same way and they may have been more general-purpose domestic ovens (again with an assumed upper structure) or pits containing the fire and the cooking vessel. There do not appear to be close comparisons. The pits were of similar form to, but slightly smaller than the equally enigmatic ‘flint-burning pits’ on the sixth/seventh- century settlement at Kilverstone in Norfolk, where their use as cooking pits has been doubted for practical reasons (Lucy 2006, 186). Like the Netherfield Farm fire-pits they were not directly associated with domestic residences and they may have had a more occasional usage for an undefined industrial process (ibid.). There was one pit at Netherfield Farm that had been used for a high temperature process or processes, which the chemical analysis of the sediments suggests might have included non- ferrous metal working. The other non-burnt pits are of unclear purpose, but may have been backfilled with burnt material from the fire-pits. Iron slag from smithing came from the slightly later features but it is not known where this activity was carried out. 02 Mudd:Layout 1 03/07/2013 08:49 Page 81

excavations at netherfield farm, south petherton 81

The impli cations of these findings are not clear since little is understood about the produc tion and distribution of metals in this period, and the extent to which metal - working of any sort was specific to a particular type or status of site is not presently known. Iron smithing and non-ferrous metalworking has been identified at the ‘elite’ settle ment at Cadbury Congresbury (Rahtz et al. 1992, 116–31), but smithing evi- dence from Bywood Farm, Dunkeswell, and North Hill Glider Field, nearby on the Black down Hills in Devon (Webster 2008, 175) suggest that metalworking may have been carried out, perhaps commonly, on a range of sites. The greater use of chemical analyses may be one way by which the problem of identifying types and processes of metal working can be explored, but the data are not self-explanatory and may be open to other interpretations. The overall form, extent and duration of the occupation were not defined in the excavated area, but the evidence from the crops alone perhaps indicates that this was part of a settlement rather than a location of more temporary activity. As far as the evidence allows, the site may be seen to be typical of the unstructured settlements of the fifth and sixth centuries, lacking defined limits to either individual properties or to the settlement space as a whole (Reynolds 2003, 103). It is interesting that the only ditch is likely to be slightly later than the pits, and seems to conform to the national pattern in which boundary features first appear in the later sixth century onwards. The more highly regulated use of space has been linked to the development of legal culture from this time in the context of the consolidation of the early English kingdoms (ibid., 131). The site’s location away from Roman and medieval settlement is significant and contrasts with occupation of this date found elsewhere where there is some evidence for continuity from, or re-occupation of, Roman sites, both rural and urban (Webster 2008, 171–74) and the coalescence of early medieval occupation to form villages from the seventh and eighth centuries, as at Shapwick (Gerrard and Aston 2007, 964, 971–73). Such slight and apparently formless and artefact-poor occupation is known from elsewhere and like the evidence from Netherfield Farm is normally discovered as a by-product of investigations directed at other remains, such as the Roman settle- ment at Frocester, Glos. (Price 2000). Recent radiocarbon dating of human bones from the Roman sites at Winthill (Banwell) and Portbury, both in North Somerset, have demonstrated burials of the sixth to eighth centuries ad (Aston 2011) and shows that radiocarbon dating is one means with the potential to make considerable con - tribu tions to identifying the people and settlements of this period.

ACKNOWLEDGEMENTS

The work was undertaken at the request of Laing O’Rourke, the pipeline constructors, on behalf of National Grid, to satisfy the mitigation proposals for archaeology and cultural heritage contained in the Environmental Statement which accompanied the application to construct the scheme. We are most grateful to National Grid for providing the opportunity and funding to undertake the archaeological excavations and the post-excavation work leading to this publication, and to the pipeline constructor, Laing O’Rourke, for and lending their support to the archaeological works. Linda Bonnor, archaeo- logical adviser to National Grid, advised on, and facilitated, the works throughout, and Steven Membery monitored them on behalf of Somerset County Council. We are grateful for their support 02 Mudd:Layout 1 03/07/2013 08:49 Page 82

82 excavations at netherfield farm, south petherton

and interest. The initial desk-based studies were written by Richard Morton of Cotswold Archaeology and the geophysical surveys undertaken by Archaeological Surveys. The fieldwork was managed for Cotswold Archaeology by Mark Collard. Mark Brett led the fieldwork, while the excavation results were compiled and analyzed by Mark Brett under the overall management of Annette Hancocks and latterly Andrew Mudd. Contributors at the post-excavation assessment stage are acknowledged in that report (Cotswold Archaeology Report 9022) available in the site archive and upon request. Specialist contributions have been edited and summarized for this publication, and fuller reports can be found in the site archive. The illustrations were prepared by Lorna Gray. We are grateful to English Heritage for permission to reproduce plans of the causewayed enclosures shown in Illustrations 32 and 33. We are also grateful to Somerset Historic Environment Record for providing copies of the aerial photo - graphs of the site. The project archive (both finds and records) is to be deposited with Somerset County Museum Service under accession code 77/2007, subject to the agreement of the legal land- owners.

PRIMARY REFERENCES SHER (Somerset Historic Environment Record) aerial photographs DAP YT 25–28 (25 July 1995). Possible circular cropmarks at Netherfield Farm, South Petherton, Somerset

BIBLIOGRAPHY Adams, A. E., MacKenzie, W. S., and Guilford, C. 1984. Atlas of Sedimentary Rocks under the Microscope, Harlow: Longman Anderberg A-L., 1994. Atlas of Seeds; Part 4, Uddevalla: Swedish Museum of Natural History Anderson-Whymark, H. 2011. Intentional breakage in the British Neolithic, Lithics, 32, 16–22 Anderson-Whymark, H. and Thomas, J. (eds) 2012. Beyond the Mundane: Regional Perspectives on Neolithic Pit Deposition, Oxford: Oxbow ApSimon, A. M. 1961. The stratigraphy and archaeology of the late glacial and post-glacial deposits at Brean Down, J. Bristol Speleol. Soc., 9, 67–136 Arnold, D. E. 1985. Ceramic Theory and Cultural Process, Cambridge: Cambridge University Press Ashmore, P. 1999. Radiocarbon dating: avoiding errors by avoiding mixed samples, Antiquity, 73, 124–30 Aston, M. 2011. New Radiocarbon Dates for Early Medieval Somerset, Proc. Somerset Archaeol. Nat. Hist. Soc., 154, 185–89 Bamford, H. 1985. Briar Hill: Excavation 1974–1978, Northampton: Northampton Development Corporation Barclay, A. and Bayliss, A. 1999. Cursus monuments and the radiocarbon problem, in A. Barclay and J. Harding (eds), Pathways and Ceremonies: the Cursus Monuments of Britain and , 11–29, Oxford: Oxbow Barclay, G. J. and Maxwell, G. S. 1991. Excavation of a Neolithic long mortuary enclosure within the Roman legionary fortress at Inchtuthil, Perthshire, Proc. Soc. Antiq. Scot., 121, 27–44 Bayliss, A. and Bronk Ramsey, C. 2004. Pragmatic Bayesians: a decade integrating radiocarbon dates into chronological models, in C. E. Buck and A. R. Millard (eds), Tools for Constructing Chronologies: Tools for Crossing Disciplinary Boundaries, 25–41, London: Springer Bayliss, A., van der Plicht, J., Bronk Ramsey, C., McCormac, G., Healy, F. and Whittle, A. 2011. Towards generational time-scales: the quantitative interpretation of archaeological chronologies, in A. Whittle, F. Healy and A. Bayliss. (eds) Gathering Time: Dating the Early Neolithic Enclosures of Southern Britain and Ireland, 16–60, Oxford: Oxbow Bell, M. 1990. Brean Down: Excavations 1983–1987, London: English Heritage Berggren, G. 1981. Atlas of Seeds; Part 3, Arlöv: Swedish Museum of Natural History Bradley, P. 1999. Worked flint, in A. Barclay and C. Halpin, Excavations at Barrow Hills, Radley, Oxfordshire. Volume 1: The Neolithic and Bronze Age monument complex, 211–27, Oxford: Oxford Archaeology 02 Mudd:Layout 1 03/07/2013 08:49 Page 83

excavations at netherfield farm, south petherton 83

Bradley, P. 2004. Causewayed enclosures: monumentality, architecture, and spatial distribution of artefacts — the evidence from Staines, Surrey, in J. Cotton and D. Field (eds), Towards a New : Aspects of the Neolithic in South-East England, 115–123, York: Counc. Brit. Archaeol. Res. Rep., 137 Bradley, R. 2002. The Past in Prehistoric Societies, London: Routledge Brett, M. and Mudd, A. forthcoming. Prehistoric, Roman and later discoveries in South Somerset: the archaeology of the Ilchester to Barrington Gas Pipeline 2005–6, Proc. Somerset Archaeol. Nat. Hist. Soc. Bronk Ramsey, C. 1995. Radiocarbon calibration and analysis of stratigraphy, Radiocarbon, 36, 425–30 Bronk Ramsey, C. 1998. Probability and dating, Radiocarbon, 40, 461–74 Bronk Ramsey, C. 2001. Development of the radiocarbon calibration program Oxcal., Radiocarbon, 43, 355–63 Bronk Ramsey, C. 2009. Bayesian analysis of radiocarbon dates, Radiocarbon, 51(1), 337–60 Buck, C. E., Cavanagh, W. G., and Litton, C. D. 1996. Bayesian Approach to Interpreting Archaeological Data, Chichester: Wiley Burrow, I. 1981. Hillforts and Hilltop Settlement in Somerset in the First to Eighth Centuries AD, Oxford: Brit. Archaeol. Rep. Brit. Ser. 91 Butler, C. 2005. Prehistoric Flintwork, Stroud: Tempus Cappers, R. T. J., Bekker R. M., and Gronigen, J. E. A., 2006. Digital Seed Atlas of the Netherlands, Archaeological Studies 4, The Netherlands Online version — www.seedatlas.nl. Campbell, G. 1994. The preliminary archaeobotanical results from Anglo-Saxon West Cotton and Raunds, in Rackham 1994, 65–82 Campbell, G., Smith, W. and Straker, V. 2007. The charred plant macrofossils, in Gerrard and Aston 2007, 269–888 Carruthers, W. 2009. Appendix 10: Charred plant remains, in M. Brett, M. Collard, and A. Mudd (eds) Ilchester to Barrington Gas Pipeline Somerset, Post-excavation Assessment and Updated Project Design, 113–119, Cotswold Archaeol. Rep. 09022 Carruthers, W. and Hunter, J. 2001. West Heslerton. The plant remains, Unpublished Palaeobotanical Rep. Clapham, A. J. 1999. Environmental analyses, in C. A. Butterworth, A. P. Fitzpatrick, and J. Grove (eds) Prehistoric and Roman sites in East Devon: the Honiton to Exeter Improvement DBFO Scheme, 1996–9, Volume 1, 51–59, Salisbury: Wessex Archaeology Cleal, R. M. J. 1991. The earlier prehistoric pottery, in N. M. Sharples, Maiden Castle: Excavations and Field Survey 1985–6, 171–85, London: English Heritage Archaeol. Rep. 19 Cleal, R. M. J. 1995. Pottery fabrics in Wessex in the fourth to second millennia bc, in I. A. Kinnes and G. Varndell (eds), ‘Unbaked Urns of Rudely Shape’: Essays on British and Irish Pottery for Ian Longworth, 185–94, Oxford: Oxbow Cleal, R. M. J. 1997. Earlier Prehistoric Pottery in J. C. Smith, F. Healy, M. J. Allen, E. L. Morris, I. Barnes and P. J. Woodward, Excavations along the Route of the Dorchester By-pass, Dorset, 1986–8, 86–102, Salisbury: Wessex Archaeology Cleal, R. M. J. 2004. The dating and diversity of the earliest ceramics of Wessex and south-west England, in Cleal and Pollard 2004, 164–92 Cleal, R. M. J. and Pollard, J. (eds) 2004. Monuments and Material Culture, Salisbury: Hobnob Coles, J. M. and Orme, B. J. 1976. The Sweet Track, Railway Site, Somerset Levels Papers, 2, 34–65 Gale, R. and Cutler, D. F. 2000. Plants in Archaeology — Identification Manual of Artefacts of Plant Origin from and the Mediterranean, Otley: Westbury and the Royal Botanic Gardens Kew Dark, K. 2000. Britain and the End of the Roman Empire, Stroud: Tempus Darvill, T. 2004. Soft-rock and organic tempering in British Neolithic pottery, in Cleal and Pollard 2004, 193–206 Darvill, T. 2010. , London: Routledge Dixon, P. 1988. The Neolithic Settlements on Crickley Hill, in C. Burgess, P. Topping, C. Mordant and M. Maddison (eds), Enclosures and Defences in the Neolithic of Western Europe, 75–87, Oxford: Brit. Archaeol. Rep. Int. Ser., 403 Evans, C. and Hodder, I. 2006. A Woodland Archaeology: Neolithic Sites at Haddenham, Cambridge: MacDonald Institute for Archaeological Research Fairburn, A. S. 2000. On the spread of plant crops across Neolithic Britain, with special reference to southern England, in A. S. Fairburn (ed.) Plants in Neolithic Britain and Beyond, Neolithic Studies Seminar Paper, 5, 107–21, Oxford: Oxbow 02 Mudd:Layout 1 03/07/2013 08:49 Page 84

84 excavations at netherfield farm, south petherton

Fitzpatrick, A. P., Butterworth, C. A. and Grove, J. 1999. Prehistoric and Roman Sites in East Devon: the A30 Honiton to Exeter Improvement DBFO, 1996–9, Salisbury: Wessex Archaeol. Rep., 16 Gale, R. 1999. Charcoal, in Fitzpatrick, Butterworth and Grove 1999, 59–62 Gale, R. 2007. Charcoal, in Gerrard and Aston 2007, 889–94 Gale, R. 2009. Appendix 11: Charcoal, in M. Brett, M. Collard and A. Mudd, Ilchester to Barrington Gas Pipeline Somerset, Post-excavation Assessment and Updated Project Design, 113–19, CA report no. 09022 Garwood, P. 1999. Radiocarbon dating and the chronology of the monument complex, in A. Barclay and C. Halpin, Excavations at Barrow Hills, Radley, Oxfordshire, Vol I: The Neolithic and Bronze Age Monument Complex, 293–309, Oxford: Thames Valley Landscapes, 11 Gerrard, C. and Aston, M., 2007. The Shapwick Project, Somerset: A Rural Landscape Explored, London: Soc. for Medieval Archaeol. Monogr., 25 Gibson, A. M. and Kinnes, I. A. 1997. On the urns of a dilemma: radiocarbon and the Peterborough problem, Oxford J. Archaeol., 16(1), 65–72 Green, H. S. 1980. The Flint Arrowheads of the British Isles: A Detailed Study of Material from England and Wales with Comparanda from and Ireland, Oxford: Brit. Archaeol. Rep. Brit. Ser., 75 Griffith, F. M. 1985. Some recently discovered ritual monuments in mid Devon, Proc. Prehist. Soc., 51, 310–15 Grogan, E., Johnston, P., and O’Donnell, L. 2007. The Bronze Age Landscapes of the Pipeline to the West: An Integrated Archaeological and Environmental Assessment, Bray: Wordwell Hagan, A. 2006. Anglo-Saxon Food and Drink, Ely: Anglo-Saxon Books Harding, J. and Healy, F. 2007. The Raunds Area Project: a Neolithic and Bronze Age Landscape in Northamptonshire, Swindon: English Heritage Healey, E. and Robertson-Mackay, R. 1983. The Lithic Industries from Staines causewayed enclosure and their relationship to other earlier Neolithic industries in southern Britain, Lithics 4, 1–27 Healy, F. 1988. The Anglo-Saxon Cemetery at Spong Hill, North Elmham. Part VI: Occupation in the seventh to second millennia BC. Gressenhall: East Anglian Archaeol., 39 Heaton, M. 1992. Two mid-Saxon grain driers and late medieval features at Chantry Fields, Gillingham, Dorset, Proc. Dorset Nat. Hist.Archaeol. Soc., 114, 96–126 Heller, I., Kienast, F., Schoch, W. and Schweingruber, F. H. 2004. Wood Anatomy of Central European Species, www.woodanatomy.ch. Keiller, A. and Piggott, S. 1939. Badshot long barrow: a report on the excavation, in K. P. Oakley, W. F. Rankine, A. W. G. Lowther, A. Keiller and S. Piggott, A Survey of the of the Farnham District (Surrey), 133–52, Frome: Surrey Archaeological Society Ladle, L. and Woodward, A. 2009. Excavations at Bestwall Quarry, Wareham 1992–2005. Volume 1: the Prehistoric Landscape, Dorchester: Dorset Nat. Hist. Archaeol. Soc. Monogr., 19 Laidlaw, M. and Mepham, L. 1999. Pottery, in Butterworth, Fitzpatrick and Grove 1999, 43–51 Lamdin-Whymark, H. 2008. The Residue of Ritualised Action: Neolithic Deposition Practices in the Middle Thames Valley, Oxford: British Archaeol. Rep. Brit. Ser., 466 Loveday, R. 2006. Inscribed Across the Landscape: The Cursus Enigma, Stroud: Tempus Lucy, S. 2006. The Early Anglo-Saxon settlement and cemetery, in D. Garrow, S. Lucy and D. Gibson, Excavations at Kilverstone, Norfolk: an Episodic Landscape History, 170–201, Cambridge: East Anglian Archaeol., 113 McSloy, E. R. 2008. Neolithic Pottery, in Moore, R. ‘Ashby Folville to Thurcaston: The Archaeology of a Leicestershire Pipeline. Part 1: The Prehistoric Sites’, Trans. Leics. Archaeol. Hist. Soc., 83, 5–10 Mercer, R. and Healy, F. 2008. Hambledon Hill, Dorset, England. Excavation and Survey of a Neolithic Monument Complex and its Surrounding Landscape, 1, London: English Heritage Morris, E. L. 1991. Ceramic analysis and the pottery from Potterne: a summary, in A. Middleton and I. Freestone (eds), Recent Developments in Ceramic Petrology, 277–87, London: British Museum Occasional Paper, 81 Morris, E. L. 1994. Production and distribution of pottery and salt in Iron Age Britain: a review, Proc. Prehist. Soc., 60, 371–93 Morris, E. L. and Woodward, A. 2003. Ceramic Petrology and Prehistoric Pottery in the UK, Proc. Prehist. Soc., 69, 279–303 Oswald A., Dyer C. and Barber M. 2001. The Creation of Monuments. Neolithic Causewayed Enclosures in the British Isles, Swindon: English Heritage 02 Mudd:Layout 1 03/07/2013 08:49 Page 85

excavations at netherfield farm, south petherton 85

PCRG 2010. The Study of Prehistoric Pottery: General Policies and Guidelines For Analysis and Publication, Prehistoric Ceramics Research Group Occasional Papers 1 and 2, Third edition revised, http://www.pcrg.org.uk/Publications1–2.htm Pearson, K. L. 1997. Nutrition and the early-medieval diet, Speculum, 72(1), 1–32 Piggott, S. 1962. The Excavations, 1955–56, London: H.M.S.O. Price, E. 2000. Frocester: a Romano-British settlement, its Antecedents and Successors, two vols, Kings Stanley: Gloucester and District Archaeol. Res. Group Pryor, F. 1998. Etton: Excavations at a Neolithic Causewayed Enclosure near Maxey, Cambridgeshire 1982–7, London: English Heritage Archaeological Report, 18 Quinnell, H. 2007. A Peterborough sherd from the Beach at Westward Ho!, Proc. Devon Archaeol. Soc., 65, 231–33 Rackham, J. (ed.)1994. Environment and Economy in Anglo-Saxon England, York: Counc. Brit. Archaeol. Res. Rep., 89 Rackham, O. 1980. Ancient Woodland — Its History, Vegetation and Uses in England, London: Edward Arnold Rackham, O. 1994. Trees and woodland in Anglo-Saxon England: the documentary evidence, in Rackham 1994, 65–82 Rackham, O. 2001. Trees and Woodland in the British Landscape: The Complete History of Britain’s Trees, Woods and Hedgerows, London: Weidenfeld and Nicolson Rahtz, P., Woodward, A., Burrow, I., Everton, A., Watts, L., Leach, P., Hirst, S., Fowler, P. and Gardner, K. 1992. Cadbury Congresbury 1968–73. A Late/Post-Roman Hilltop Settlement in Somerset, Oxford: Brit. Archaeol. Rep. Brit. Ser., 223 Rahtz, P., Hirst, S. and Wright, S. M. 2000. Cannington Cemetery, London: Britannia Monogr. Ser., 17 Reimer, P. J., Baillie, M. G. L., Bard, E., Bayliss, A., Beck, J. W., Blackwell, P. G., Bronk Ramsey, C., Buck, C. E., Burr, G. S., Edwards, R. L., Friedrich, M., Grootes, P. M, Guilderson, T. P., Hajdas, I., Heaton, T. J., Hogg, A. G., Hughen, K. A., Kaiser, K. F., Kromer, B., McCormac, F. G., Manning, S. W., Reimer, R. W., Richards, D. A., Southon, J. R., Talamo, S., Turney, C. S. M., van der Plicht, J. and Weyhenmeyer, C. E. 2009. Intcal09 and marine09 radiocarbon age calibration curves, 0–50,000 years cal BP, Radiocarbon, 51, part 4, 1111–150 Reynolds, A. 2003. Boundaries and settlements in later sixth to eleventh-century England, in D. Griffiths, A. Reynolds and S. Semple (eds) Boundaries in Early Medieval Britain, Anglo-Saxon Stud. in Archaeol. and Hist. 12, Oxford: Oxford University School of Archaeology Robertson-Mackay, R. 1987. The Neolithic causewayed enclosure at Staines, Surrey: excavations 1961–3, Proc. Prehist. Soc., 53, 23–128 Rose, F. 2006. The Wild Flower Key, London: Frederick Warne Rye, O. S. 1981. Pottery Technology; Principles and Reconstruction, Manuals on Archaeology 4, Washington DC: Taraxacum Shepherd, W. 1972. Flint: its Origin, Properties and Uses, London: Faber Simms, M. J., Chidlaw, N., Morton, N. and Page, K. N. 2004. British Lower Jurassic Stratigraphy, London: Joint Nature Conservancy Council Smith, I. F. 1965. Windmill Hill and Avebury: Excavations by Alexander Keiller 1925–1939, Oxford: Clarendon Press Smith, I. 1976. The Pottery, in Coles and Orme 1976, 63–64 Stace, C. 1997. A New British Flora, Cambridge: Cambridge University Press Stone, D. J. 2006. The consumption of field crops in late medieval England, in D. Serjeantson, T. Waldron and C. M. Woolgar, Food in Medieval England, 11–26, Oxford: Oxford University Press Straker, V. and Wilkinson, K. 2008. Neolithic and early Bronze Age environmental background, in C. J. Webster (ed.), 63–74 Stuiver, M. and Kra, R. S. 1986. Editorial comment, Radiocarbon, 28, part 2B, ii Stuiver, M., and Polach, H. A. 1977. Reporting of 14C data, Radiocarbon, 19, 355–63 Stuiver, M. and Reimer, P. J. 1986. A computer program for radiocarbon age calculation, Radiocarbon, 28, 1022–30 Stuiver, M. and Reimer, P. J. 1993. Extended 14C data base and revised CALIB 3.0 14C age calibration program, Radiocarbon, 35, 215–30 Wainwright, G. J. 1979. Mount Pleasant, Dorset: Excavations 1970–1971, London: Society of Antiquaries of London 02 Mudd:Layout 1 03/07/2013 08:49 Page 86

86 excavations at netherfield farm, south petherton

Wainwright, G. J. and Smith, K. 1980. The Shaugh Moor project: second report — the enclosure, Proc. Prehist. Soc., 46, 65–122 Webster, C. J. (ed.) 2008. The Archaeology of South West England: South West Archaeological Research Framework, Resource Assessment and Research Agenda, Taunton: Somerset County Council Welch, F. B. A. and Crookall, R. 1948. British Regional Geology Bristol and Gloucester District, London: H.M.S.O. Wheeler, M. 1943. Maiden Castle, Dorset, London: Society of Antiquaries of London Wheeler, E. A., Baas, P. and Gasson, P. E. 1989. IAWA list of microscopic features for hardwood identification, IAWA Bulletin New Series 10, 219–332 Whittle, A. W. R. 1977. The Earlier Neolithic of Southern England and its Continental Background, Oxford: Brit. Archaeol. Rep. Int. Ser., 35 Whittle, A. 1997. Sacred Mound, Holy Rings. Silbury Hill and the West Kennet Palisade Enclosures: a Later Neolithic Complex in North Wiltshire, Oxford: Oxbow Whittle, A., Pollard, J. and Grigson, C. 1999. The Harmony of Symbols. The Windmill Hill causewayed enclosure, Wiltshire, Oxford: Oxbow Whittle, A., Healy, F. and Bayliss, A. 2011. Gathering Time: Dating the Early Neolithic Enclosures of Southern Britain and Ireland, Oxford: Oxbow Wilkinson, K. N. 2008. Ilchester-Barrington Pipeline (IBP 05): Geoarchaeological Assessment. Unpublished report, ARCA, Winchester: University of Winchester Wilkinson, K. N., Watson, N., Black, S., Warren, G., and Marini, N. 2011. Ilchester-Barrington Pipeline, Somerset: Analytical Report on the Geoarchaeology of Samples from Site 48, Unpublished report ARCA, Winchester/Reading: Dept of Archaeology, University of Winchester/Quest: School of Human and Environmental Sciences, University of Reading Williams, D. F. 1976. Thin-section examination of the pottery, in Coles and Orme 1976, 64 Wilson, V., Welch, F. B. A., Robbie, J. A. and Green, G. W. 1958. Geology of the Country around Bridport and Yeovil, Memoirs of the Geological Survey of , London: H.M.S.O. Wood, J. 2001. Prehistoric Cooking, Stroud: Tempus Woodward, A. 1990. The Bronze Age Pottery, in Bell 1990, Brean Down: Excavations 1983–1987, 121–45, London: English Heritage The publication of this article was supported by a grant from National Grid

Andrew Mudd and Mark Brett, Cotswold Archaeology, Building 11, Kemble Enterprise Park, Cirencester, Gloucestershire, gl7 6bq Email: [email protected] [email protected]