Burhill Fort - Alistair Marshall

Alistair Marshall 1989 Report - click here






This paper describes an important and unusually well-preserved assemblage of earlier iron age pottery recovered from the su,face at the margin of the enclosed area of the univallate hillfort at Burhi/1 (SP 08503630) and briefly discusses the significance of the site within its broader archaeological context.



The site is situated at the tip of a pronounced, level-topped spur which forms part of the  lower slope of the main Cotswold Scarp (FIG. 5). The interior of the hillfort is at approximately 167 m OD, on Middle Lias, and directly overlooks the Severn/Avon valley, which is formed by Lower Lias clay with overlying gravel patches some 90 m below. The location is an extremel y advantageous one for settle ment , since it offers a relatively sheltered, well-watered, mid-scarp position at a lower altitude within the local range, maintains the defensive advantages of a terminal location on a spur, and provides equ al access both to limestone upland areas above the site and to the valley below; areas of differing and complementary agricultural potential. The entire interior area of the hillfort is under intensive annual cultivation, ploughed to a depth of approximately 0. 25 m. Only some 80 m of the rampart remains unploughed, and about 80 per cent of the rest of the spur is under cultivation.

The defences of the hillfort survive only on the north-eas tern side as a fragment of univallate rampart about 1.5 m high, 9 m wid e, with an external ditch about 7 rn wide and Im deep in silted condition (FIG I, #1; RCHM 1976 22-23). Any remaining circuit, which if complete would have enclosed 3-4 hectares, has been removed or severely reduced by repeated ploughing. Additional rampart may be indicated by a very low bank some 0. 1-0. 2 m high, with no indication on the ground of a surviving ditch, but with an approximately central gap of some 30 m suggesting an entrance (FIG. 5, #2). There is no clear evidence for the rampart extending around the margin of the spur as a complete circuit, the only defences which are apparent are those which cross the spur and continue for a short distance around its northern Aank (FIG. 5, #3). The remain ing stretch of relatively intact rampart owes its survival entirely to its position among marginal boundary scrub at the scarp edge, where it has not been practical to plough.

The pottery described in this paper was recovered from a dense and localized surface scatter of occupation

debris, containing iron age pottery, baked clay and daub fragments , gritstone and quern fragments, fire-cracked pebbles, and fragments of animal bone. It produced nothing which could be directly assigned to a date later than the earliest iron age. The area of the scatter covered approximately 0. 5 ha and was located immediately outside the possible entrance (FIG. 5, #5). In view of its good preservation it seems likely that the Burhill sample represents more deeply-stratified material recently brought up from greater depth by changes in ploughing pract ice . In contrast, pottery previousl y recorded from the site (Saville 1984, 170) has been considerably abraded. Evidence from other sites, such as the hillfort at Dowdeswell (SO 99901890) and the settlement at Gui ting Power (Sa ville I 979), suggests that such external pits occurred elsewhere on iron age sites in the area.

Finds from the interior of the hillfort at Burhill are infrequent, and a general scatter of occupation debris  is abse nt. Surface finds of iron age sherds from the area of the rampart have been noted in RCHM 19 76, 22 and a single fragment, again a surface find from the inte rior , is illustrated in Saville 1984 (FIG . 3, 25). The latter is a rim fragm nt with finger-tip/nail impressions along the lip, and is entirely consistent with the early iron age pottery described in this paper. A well-preserved saddle quern of prehistoric type has also been recovered from the interior of the hillfort (FIG. 5, #4).

In contrast to thi dense scatter of occupation debris near the presumed entrance to the hillfort,  sampling of surface scatters over the entire spur has produced only very sporad ic finds of domestic stone-work debris (variously abraded a:nd fire-cracked pebbles , gritstone fragments, some derived from saddle querns).


Such items, although undated, are of prehistoric type and are typical of the range of material recovered from interiors and peripheries of other hillforts elsewhere in the Cotswolds (unpublished field-work). Since Roman occupation has not yet been detected in the immediate area, certainly not from the ploughed top of the spur, these finds are best seen as derived from the occupation of the hillfort itself.

Although there is no evidence from surface finds for post-iron age activity or settlement on the spur, the area of the hillfon itself has produced evidence of pre-iron age activity in the form of a grade 2 flint scatter (definition in Marshall 1985). This scatter occurs over approximately I ha centred on SP 08603625 and contrasts with the rest of the spur, over which the density of surface flint is negligible . This grade 2 scatter is almost devoid of well-produced flakes and implements and as such is probably best interpreted in terms of activity rather than settlement. Pre-iron age activity at similar hillfort sites is suggested by the occurrence of flint scatters at Roel (SP 04902440) and ?Salter's Hill (SP 04502860), but not at Shenberrow (SP 08053345) .


Details of the pottery from Burhill

The sample of pottery contained vessels of two main types: situlate jars and carinated bowls. Minor types are included as a residual group of other bowl/jar forms. The sample of pottery falls entirely within the Shenberrow-Chastleton assemblage dating from the earliest iron age, and consequently shows strong ceramic affinities with the upper Thames valley (Marshall 1978).

A representative range of examples is illustrated in FIG. 6 as follows:


1-8 Fragments of situ/ate jars. Forms with evened rims recurved at the tip are represented (1,2), including examples of considerable diameter. Other types with upright (4), or gently everted rims (3) also occur. Bands of decoration or plain/decorated cordons frequently occur at the neck (1,5,6) or shoulder (4,7,8). Decoration consists of repetitive finger-nail (1,4,7,8), or finger-tip impressions (not illustrated), diagonal slash marks (6), or applied cordons can be left undecorated (5). All fabric 2.


9-16 Fragments of carinated bowls. Forms with gently everted rims (9, 10,14,16) and sharply carinated shoulders (11,12,13,15) are represented. Several fragments (15) suggest that some  of  the  bowls  were  unusually small. Decoration, when present, consists of incised hatching (9),  or  burnished  lines  (12, 13). Several fragments (not illustrated) have haematite  slip applied  over internal and external  surfaces.  All fabric 5, except examples 12 and 16 which are of fabrics 3 and I respectively. Micaceous fabric  5,  which predominates in this group, may be non-local and represent imported wares. The occasional appearance of carinated bowl forms in other fabrics (12 and 16) may represent local copies .


17 Other bowlljar f orms. Although situlate jars and carinated bowls dominate the sample, occasional fragments of other types do occur. The example illustrated represents a rounded bowl with a short evened, but not beaded, rim. Fabric I .


All vessels are hand-mad e and the fabrics can be classified into 5 main types as follows:


Fabric I

Matrix: uniformly black in colour over internal and external surfaces and within the section. Some irregularity in firing produced occasional patches of brown within the black.

Filler: high levels of crushed white shell filler occur within the section and at the surface. Granules are generally flattened and 1-4 mm in length.

Hardness: friable, flaking along the line of granules of shell filler.

Surface: unburnished but relatively smooth, with wipe-marks from final preparation of the external and internal surfaces . Shell filler is very visible at the surface.

Thickness f fragments: 6-13 mm.

Type of v essel: large situlate jars, including examples with angled shoulders and evened necks. Decorated examples appear infrequently but one fragment of a large jar with a single, shallow, rounded groove,  5 mm wide and I mm deep, does occur in the sample.


Comments: appears to be a deliberate variant or less well-fired version of fabric 2, which is generally reddish in colour because of its higher firing-temperature. The forms of vessel appear similar in both fab rics.


Fabric 2

Matrix: exterior surfaces and sections are red/brown, light brown, or buff-yellow. Interior surfaces are usually darker than this, generally grey to black, especially when the fragment derives from the body of the vessel where increased reducing conditions pertained during firing. Patchy colouration of surfaces indicates non-uniform firing temperatures.

Filler: as fabric I.

Hardness: the redder, more oxidized fabric caused by higher firing-temperatures confers greater hardness .

Surface: as fabric I.

Thickness of fragments: as fabric I.

Type of vessel: situ.late jars, as fabric I, but the incidence of decorated fragments appears to  be  higher, although this may be fortuitous. Examples of situ.late jars occur with decoration at the shoulder/neck  in the form of a narrow band of diagonal slashing, a narrow band of upright finger-tip/nail  impressions,  or  an  applied cordon with vertical finger-nail impressions.

Comments: probably a higher-temperature variant of fabric I.


Fabric 3

Matrix: relatively uniform grey/black in section and at the surface. Some examples occur with patchy black surfaces and some with red bands within the section, indicating variable firing.

Filler: contains a lower concentration of filler than fabric I. The filler is crushed shell or limestone with smaller, more rounded granules <2 mm in mean diameter.

Hardness: higher firing-temperature and smaller size of filler confer greater hardness than fabric I.

Surface: not burnished, but wipe-marks  from final finishing are visible,  and  pock-marks occur  where grains of filler are absent, leaving holes. Some examples are smoothed, and where the level of filler is slight, are relatively fine, hard fabrics .

Thickness of fragments: ~ 7 mm .

Type of vessel: smaller situlate jars and carinated bowls.


Fabric 4

Matrix: uniformly orange in section and at the surfaces.

Filler: low levels of finely granular crushed li mestone .

Hardness: similar hardness to fabric 3, far harder than fabrics I and 2.

Surface: matt, relatively smooth but pock-marked where grains of filler have been removed.

Slip: examples occur with red haematite slip over interior and exterior surfaces.

Thickness of fragments: ~ 8 mm.

Type of vessel: carinated bowls, and smaller bowl/jar forms with evened rims.


Fabric 5

Matrix: grey/black throughout the section and at surfaces, with some tendency for patchiness of colour at the surface. There is a highly characteristic sparkle over all surfaces caused by fine grains of mica in the matrix. Filler: none is visible in some examples, but others contain low levels of crushed shell.

Hardness: similar in hardness to fabric 3, but far harder than fabrics I and 2.

Surface: matt, smoothed, with some pock-marking where grains of filler have been removed.

Slip: examples occur with red haematite slip over interior and exterior surfaces.

Thickness of fragments: ~ 7 mm .

Type of vessel: carinated bowls.