4. Material Culture

Overview | Flint | Prehistoric pottery | Briquetage | Roman and later pottery | Ceramic objects | Ceramic building material | Fired clay | Clay pipe | Metalwork | Worked stone | Glass

4.2 Prehistoric pottery

E.L. Morris

A well-preserved assemblage, primarily of later Iron Age pottery, was recovered. Analysis revealed that nearly all of the Iron Age pottery had been brought to the settlement through trading networks, which reflects the location of this site in the middle Severn valley. The significance of the assemblage lies in its date, which spans the end of the middle Iron Age up to the beginning of the latest pre-Roman Iron Age, from the 2nd century into the 1st century BC. There is no strong evidence to indicate that early 1st-century AD activity is represented in the Iron Age pottery assemblage. The variety of vessel forms and fabrics of this later Iron Age period makes the Blackstone collection an important regional, single period type-site assemblage, in particular due to the lack of redeposited pottery from any earlier Iron Age activity. Variation in distribution of pottery indicates that while much of the pottery is thinly deposited in features scattered across the site and in ditch cuts, two deposits stand out as distinctively different special events. Everyday life in the later Iron Age is, therefore, well represented. The pottery was found occasionally in association with fragments of Droitwich briquetage containers, ceramic salt packs used to dry and transport salt from their brine springs source to settlements in the region. What was special, however, was the additional presence of a single sherd of Cheshire briquetage container in the assemblage. Together, the pottery vessels and salt packs symbolise the complexity of exchange systems during this period. Earlier prehistoric pottery in the assemblage is represented by plain body-sherds in three variants of a later Neolithic-type fabric and several sherds from one highly decorated early Bronze Age collared urn.

A total of 589 sherds (12.4kg) of prehistoric pottery was identified, including later Neolithic, early Bronze Age, and Iron Age material (Table 6). The majority of the assemblage is later Iron Age in date. The early Bronze Age and Iron Age pottery was in very good condition with large mean sherd weights of 25.5g and 21.0g respectively. Most of the pottery was recovered from pits, with additional sherds from ditches, scoops and postholes.

The assemblage was first analysed in the late 1970s as part of a contribution to postgraduate research into the production and distribution of pottery and salt containers during the Iron Age in western Britain (Morris 1983), and was re-examined and recorded in 2008 according to the guidelines recommended by the Prehistoric Ceramics Research Group (PCRG 1995), with recorded data comprising count, weight in grams, fabric, rim/base/profile form, diameter and percentage of form present, surface treatment and position, decoration and position, and evidence of use, where appropriate (available in the digital archive). The fabric codes utilised in this report are those ascribed originally, which had been supported at that time by a detailed programme of fabric sampling and petrological analysis (Table 7). The pottery fabrics are correlated to the Worcestershire fabric-type series (WFS; Worcestershire On-line Ceramic Database) where possible. All sherds with diagnostic features such as rim, base or decoration, were sketched for both archival record and illustration reference at 1:1. The pottery was handmade unless otherwise indicated.

This report presents details about the fabrics, forms and decorations, correlation of surface treatment to form and fabric, vessel types and sizes with evidence of use, and discusses the date range of pottery, in particular with regard to the sequence of deposition into the two main enclosure ditches and the spatial variation of the pottery across the excavated area of the site. A focus of this assemblage is the quantity of later Iron Age and latest pre-Roman Iron Age pottery recovered, and the nature of pottery procurement during these periods at a settlement site along the middle Severn valley (modern Worcestershire).

4.2.1 Later Neolithic

A total of seven sherds of pottery (36g), all tempered with large, angular pieces of white quartz, are likely to have derived from later Neolithic vessels. Each fragment was an undecorated body-sherd from a vessel that had been oxidised on the exterior surface or fully oxidised. It appears that there were at least three different vessels represented among these sherds, judging by the variation in clay matrices into which the angular quartz temper had been added (TF8A-C); three of the sherds have been sampled for petrological analysis (Table 7). None of these correlate to any other Neolithic pottery fabrics from excavations in Worcestershire, and are likely to indicate the variety of local clays available for pot making during this period. The addition of large, white additives to create fabrics is typical of Neolithic pottery in the region (cf. Stanford 1982, 283-6), and the Peterborough or Impressed Ware tradition in particular (Gibson 1995, 24), and this characteristic may have had 'a deeper significance than the simply functional' (ibid. 1995, 29), a visual effect on the users and observers of such vessels. Five sherds, not associated with any other pottery, were recovered from pit 0079.1, and two from pit 1027 (Fig. 5), including one from layer 1027.9. This pit also contained a sherd of Iron Age pottery. Fabric types

4.2.2 Early Bronze Age

Approximately 10% of a large, fragmented collared urn (22 sherds, 561g; Fig. 36) was recovered from gully 0517 (layer 1). This vessel was made from a grog-tempered fabric (TF9), had a diameter of 380mm and was well over 400mm in height. The bevelled rim leads to an applied collar, clearly visible in cross-section, and this is attached to the slightly carinated girth section of the urn which runs into the lower, undecorated, narrower third of the vessel. No base sherds were recovered. Three zones of different designs, usually made from the same motifs in different patterns, were impressed onto this vessel which created a highly distinctive effect. On the top third of the vessel, from the rim to the lower collar edge, there are two parallel, horizontal rows of continuous-appearing whipped cord, impressed lines above two parallel, upright, zigzag 30mm long, whipped cord, impressions above another pair of parallel horizontal rows of continuous-appearing whipped cord impressed lines on the collar, which creates a bounded or closed design on this zone. Below the collar and above the carination, possibly a different whipped cord instrument had been used to create a herringbone-like pattern of at least three rows of short (15mm), diagonal lines typical of the 'maggot' style of impressions. Just beneath the soft carination is a single, horizontal row of thumb-side or edge of thumb impressions, which are distinct but shallow. The lower part of the vessel is undecorated. The vessel appears to have a very small amount of carbonised residue on the interior at the carination, which may indicate that it had been used as a cooking pot but could equally be the result of containing or being associated with cremated remains, including fragments of the pyre, for several millennia.

Figure 36

Figure 36: Bronze Age pottery.

The only collared urns found elsewhere in Worcestershire are single vessels from Holt (Hunt et al. 1986) decorated with 'deeply incised herringbone' on the exterior, and from Mathon in Herefordshire (Hamilton 1938). At least nine collared urns containing burials have been found in Gloucestershire but all were found in the southern part of the Cotswolds or are of uncertain location (Darvill 1987). These include one from Bourton-on-the Water and another from Cirencester, both decorated with twisted cord impressions; the former in herringbone style and the latter with lines of vertical zigzags. Fabric type

4.2.3 Iron Age Fabrics

The Iron Age assemblage is dominated by three distinctive fabric types:

The source for TF1A fabric is located approximately 25km to the south-west of Blackstone in the Malvern Hills area and the most likely source area for TF2 occurs in the Martley area about 17km to the south-west of the site, while that of TF3 is located in the Clee Hills district between 15-20km to the west. Together, these three fabric types account for 94% of the assemblage by weight and 89% by count of sherds (Table 6). TF1A pottery has been found on a variety of hillfort and settlement sites in the Herefordshire-Worcestershire area (Peacock 1968, fig. 2; Morris 1981, fig. 5:3), but usually in association with Group B1 Palaeozoic limestone fabric pottery (Peacock 1968, 421-2). However, at Blackstone this was not the case, and it is TF2 that dominates the assemblage. No sherds of this distinctive limestone fabric were recovered. Detailed analysis of the patterns of distribution of vesicular mudstone pottery compared to Palaeozoic limestone pottery has revealed that these are usually complementary in nature, with the Palaeozoic limestone pottery being found on sites in two zones, one to the west around Woolhope (Herefordshire) and one to the east of the Malverns between Bredon Hill and the Cotswolds, and the vesicular mudstone pottery along the middle and upper Severn valley, at Midsummer Hill, along the upper Wye valley and into Gwent (Morris 1982, figs 3.2-3.3). Petrological analysis and study of the distribution patterns revealed that the Woolhope Hills were the likely source of the Palaeozoic limestone fabric pottery, guided by the presence of over 90% of the Sutton Walls assemblage being this fabric type (Morris 1983, 116-22, figs 4.17-4.18).

A very complex and often changing network of pottery distribution occurred in this region during the Iron Age. The unifying elements of this dynamic exchange network were the vessel shapes and decorations on the pots, particularly apparent during the middle Iron Age, no matter where the vessels had been made, and this is discussed further below. The type of exchange pattern is likely to have been similar to down-the-line exchange with distance from the source directly affecting the quantity of pottery found on this site, which lies in the centre of the regional network. The vesicular mudstone fabric source is the closest one to Blackstone and is found in the greatest quantity (35-49%), while the sources of the Clee Hills and Malvernian rock fabrics are slightly further away and the quantities are significantly less (24-30% and 22-24% respectively). The regional distribution of TF1A pottery (Morris 1981, fig. 5:3) is as great as that of TF2 (Morris 1982, fig. 3.2) and they are remarkably coterminous, with outliers, while that of TF3 is also coterminous but half the area of these. The properties of these different fabrics may have satisfied the need for acquiring either a volcanic rock-derived fabric (TF1A, TF3) or a more vesicular and loosely structured fabric (TF2) vessel depending upon the purpose of the required pot, as the major form types were found in all three of these fabric types (see below). A similar combination of these three major fabric types occurred at Caynham Camp, near Ludlow in Shropshire (Gelling 1965). Here detailed fabric analysis of the entire assemblage revealed that 52% of the collection (by weight) was made from the relatively local Clee Hills dolerite fabric, 38% from Malvernian volcanic rock fabric and 10% from the Martley mudstone ware (Morris 1983).

In addition to these major fabric types, six other fabrics were defined in the assemblage. Between 3-5% of the collection had been made from a variation of the Group A Malvernian rock fabric. This type, TF1B, is characterised by a significant amount of epidote present in an otherwise quartz sand fabric. Epidote is an ever-present element of the Malvernian rock fabric (TF1A) and, therefore, this epidote-rich fabric may be a variant of that deposit. Research and petrological analysis has suggested that the likely sources for the inclusions found in this fabric are the Llandovery (Silurian) deposits found in the Malvern Hills, Woolhope, or May Hill areas (Earp and Hains 1971, 58-62; Morris 1983, 137), and that the likeliest among these is probably the Ledbury-Malvern inlier where 'Llandovery rocks crop out extensively in the Eastnor Park area and in a narrow belt from north of Herefordshire Beacon to Old Storridge Common. There is also a small inlier at Ankerdine Hill' (Earp and Hains 1971, 61). Analysis of nine Iron Age site assemblages in the Wye-Severn-Avon-Cotswold region revealed very small numbers of TF1B or Group E (Llandovery quartzose fine sandstone) fabric pottery, including sherds from Salmonsbury (Dunning 1976; six highly burnished sherds, 66g), latest pre-Roman Iron Age Guiting Power (Saville 1979, fig. 9, nos 36 and 39), Beckford (original drawing 456), Conderton Camp (Morris 1983, table 4.9; 15 sherds, 98g), Friar Street-Droitwich (Morris op. cit.; 1 sherd, 4g), Croft Ambrey (Stanford 1974, fig. 93, nos 18-19 and fig. 95, no. 30), and the Breiddin (Musson 1991, fig. 54, 104; Morris 1991), as well early Roman Collfryn (Britnell 1989, fig. 26, no. 3) and a possible candidate from Midsummer Hill, based on its similarity in form and decoration to the Beckford vessel (Stanford 1981, fig. 65, no. 20a). The vessel forms, decorations and dating of these examples are discussed below. The Blackstone assemblage contains the largest quantity of TF1B pottery (28 sherds; 334g), with Croft Ambrey having the next (6 sherds; 257g).

This same pattern of some elements of TF1A being present in a sandy clay matrix is repeated in TF5, a rare fabric in the assemblage, which consists of infrequent pieces of ferruginous, quartzose fine sandstone, epidote and mudstone in a sandy clay matrix. It is possible to see the difference between TF1B and TF5 using simple binocular microscopy owing to the size of the sandstone rock. TF5 is likely to derive from a source near the Martley area because TF2 fabric sherds have also been found to contain epidote and various types of siltstones and sandstones in very rare amounts (Morris 1982, 16). A more significant, but still minor, fabric type is TF4 representing 1-3% of the assemblage, which is simply a quartz sand fabric with no other inclusions indicative of a particularly distinctive source. It is possible that TF4 derives from a local, sandy clay source but it could equally come from any number of sources located elsewhere.

Fabrics TF7A and TF7B are represented by single sherds. Both are quartz sand fabrics but there are textural and manufacturing differences between them and also compared to TF4. All of the other fabrics are used to make handmade pottery, but fabrics TF7A and TF7B had been used to make wheel-thrown vessels. Fabric types Forms

A limited range of vessel forms was identified in the assemblage. The most common shape is an ovoid jar with convex-profile and no neck (TV2). It is the typical middle and late Iron Age shape in the region and was made by the three major fabric production outlets already identified, TF1A (Malvern Hills), TF2 (Martley vicinity) and TF3 (Clee Hills). It was identified in contexts from the very end of the early Iron Age and throughout the middle and late Iron Age at Beckford (type series form 2.1; Wills in prep. [2010]), and has been radiocarbon-dated by association to the late 5th through 1st century cal BC. In addition, there is a necked version of this barrel-shape with a short upright rim (TV1A and TV1B) which has been shown to first occur slightly later in date than the ubiquitous ovoid jar (Beckford type series form 2.2; Wills in prep. [2010]), from the end of the 4th century cal BC onwards. This type is well known elsewhere, as at Croft Ambrey (Stanford 1974, fig. 90, nos 25 and 37, fig. 91, no. 15, fig. 92, D3, fig. 93, no. 9, fig. 94, no. 15, fig. 97, no. 84), Midsummer Hill (Stanford 1981, fig. 63, nos 22-26, and 47-50, fig. 64, nos 62, and 64-65), Sutton Walls (Kenyon 1954, figs 9-14) and Guiting Power (Saville 1979, fig. 9, nos 2, 4 and 13-14). There is no clear evidence as to when these two significant jar forms ceased to be manufactured and used due to the complexities of redeposition at Beckford. Stylistic links with the Wessex area can be invoked owing to the identification of three examples of undecorated, vertical-walled, saucepan pots in the Blackstone assemblage, and similar examples have been found at Beckford (type series 5.0; Wills in prep. [2010]), Stoke Lane, Wychbold (Jones and Evans 2006, fig. 10, no. 7) and Stonebridge Cross, Westwood (Hurst 2004a, fig. 10, no. 1), both the latter being near Droitwich. Plain saucepan pots first occurred at Danebury during the second half of the 4th century cal BC but continued in use through the 1st century cal BC (ceramic phases 4-7; Cunliffe 1995, 17-18, table 3).

Other form types include an incurved and round-profile, flat-topped rim jar (TV11; Fig. 39, no. 34), another of uncertain body shape but again incurved and with a rounded rim (TV6; Fig. 40, no. 52). One example of a necked jar with everted rim was made from sandy fabric TF4 (TV9; Fig. 38, no. 13). Similar examples have been found on sites in the region, invariably from Iron Age deposits as at Beckford (type 7.0; Wills in prep. [2010]) and Evesham (Hurst 2000b, fig. 6, no. 5), but handmade examples in sandy fabrics are rare.

Two handmade, bead-rim jars, both in fabric TF2, were recovered from the top fills of features, layer 80.1 covering both pits 0080 and 0098, and inner ditch cut 0305.1 (Fig. 38, no. 7 and Fig. 39, no. 47). Handmade bead-rim jars are well-known late Iron Age forms.

What is most intriguing about the Blackstone assemblage is the variety of bowl forms present. These include two types: a high-shouldered one with a thickened, slightly pulled-bead rim and narrowed lower profile made from fabric TF2 (Fig. 38, no. 9); and one which is hemispherical in shape with a flattened rim and straight, thick walls made from TF1A (Fig. 39, no. 46). In addition, there is a beautifully finished, black and burnished, round-bodied, necked bowl with a rounded, everted, short rim and low, pedestal or flared base made from TF1B (TV7A; Fig. 39, no. 27). The two examples of this form type are presently unique in the region but the general form of this bowl has been found elsewhere in fragmented condition at Guiting Power (Saville 1979, fig. 9, no. 36), Croft Ambrey (Stanford 1974, figs 93, nos 18-19 and fig. 95, no. 30), and the Breiddin hillfort (Musson 1991, fig. 54, no. 104), all made from fabric TF1B. One of the Croft Ambrey vessels had been recovered from a deposit assigned to defences Period VIIA, late in that hillfort sequence, in association with linear-tooled decorated pottery (discussed further below).

In addition, a wheel-thrown vessel represented by a distinctive foot-ring style base sherd made from sandy fabric TF7A (Fig. 40, no. 56) was recovered from an upper fill (1548.14) of the outer enclosure ditch, in near association with a body sherd from a different wheel-thrown vessel in fabric TF7B from an overlying fill (1548.15). The base type had been created as an image of a true, applied foot-ring by incising the underside of the flat base to create a rounded ring-like appearance and also to provide stabilising support for the vessel. It is highly likely that the base derives from a squat, round-bodied jar which may have had a raised cordon at the neck which itself is upright and carries a slightly everted rim. This is similar to examples from The Ditches in central Gloucestershire, which were recognised as having been made from early Severn Valley ware fabric (Trow 1988, 64-73, fig. 31, nos 27-29, fig. 32, no. 45, fig. 33, nos 60 and 61-62, handmade or slow wheel; and fig. 34, nos 64 and 77) and found with many other imitation Gallo-Belgic vessels of 1st century BC to 1st century AD date in three major key groups at that site. Other examples of this vessel type have been recovered from the double-ditched enclosure settlement at Salmonsbury, also in association with numerous imitation Gallo-Belgic vessel types recovered from contexts stratigraphically later than Period I handmade pottery (Dunning 1976, fig. 16, no. 3, fig. 17, nos 3-4, fig. 19, no. 1, Period II, i-ii of site trench III). The form has been assigned to ceramic phases 8-9, now radiocarbon dated to 50 BC-AD 40, at Danebury (Cunliffe 1984, 248-9 and 293, fig. 6.63, no. 621; Cunliffe 1995, 17-18, table 3).

Table 8 presents the number of identifiable vessels by form-type correlated to fabric. The majority of the very common ovoid jars with no necks (TV2; 21 vessels) were made in fabrics TF2 (Martley mudstone; 7 vessels) and TF3 (Clee Hills dolerite; 12 vessels), while the ubiquitous necked, barrel-shaped jars and bowls (TV1B; 15 vessels) with upright rims were nearly all made from TF1A (Malvern Hills volcanic rock; 13 vessels). Fabric TF1B was only used to make unusual and rare vessels (TV7A, TV11, B2). This variation demonstrates the distinctiveness of each potting area which was producing vessels for exchange but also the similarities of products in some cases. For example, there are two ovoid jars (TV2) in fabric TV1A and this is the most common form made by the potters of the Malvern Hills during the middle Iron Age but it appears that a change was taking place during the later Iron Age with new barrel-shaped forms with upright rims being made in that location. Nevertheless, the Clee Hills and Martley potters continued to make ovoid jars into the late Iron Age. Form types Decoration

Two examples of decorated middle-late Iron Age pottery were recovered. The one found in the topsoil is decorated with a row of irregularly paired stamps (Fig. 40, no. 58) and the other, from inner ditch recut 313, is linear-tooled (Fig. 40, no. 53); both were fabric TF1A. Stamped decoration is extremely common in the region and first occurs at the beginning of the middle Iron Age period, while linear-tooled decoration is a much later introduction to the decorative repertoire. Stamped decoration occurs on all major and many minor fabric type vessels and is truly a pan-regional style, with both traded wares and local wares displaying this effect. Linear-tooling, however, in its classic sense of wide and shallowly impressed lines of geometric decoration that do not break through the surface of the vessel but simply push the surface inwards (i.e. not incised), is found only on fabric TF1A vessels. A case for the late development of linear-tooled pottery has been presented by Stanford (1974 and 1981) based on his observations of stratigraphic relationships at Croft Ambrey and Midsummer Hill but with very limited radiocarbon dating support. Linear-tooled pottery first occurs in gateway phase VI D at Croft, thought to be during the 2nd century BC (Stanford 1974, fig. 104), and linear-tooled pottery was recovered from occupation on Rampart B at Sutton Walls (Kenyon 1954, 34, fig. 9, no. 8). Recent work at Beckford has not contradicted this interpretation but a major programme of direct radiocarbon dating of burnt residues on vessels has not been able to provide conclusive evidence to support these observations.

In addition, there is a single example of finger-tip impressed decoration on a shapeless body sherd (Fig. 39, no. 44), the only pottery in pit 1617 (F8; Fig. 13). This type of decoration is usually located on the top of rims and the angle of shoulder sherds from jars of late Bronze Age/early Iron Age date, but this particular example from a straight wall zone suggests that the vessel was made during the end of the early Iron Age period, when distinctive shoulders on jars were in decline. It is also of some interest that the fabric of the original vessel was TF1A, normally reserved for ovoid and barrel-shaped jars, and this may therefore be a unique example. Surface treatment

One of the most distinctive aspects of the Blackstone Iron Age assemblage is the frequency and quality of burnishing on vessels. Those vessels with burnished exterior are usually fired black in colour which, when coupled with the strong burnish, results in a very striking, nearly metallic appearance. Burnishing can be interpreted as having both stylistic and functional effects, the latter resulting from the reduced porosity of the clay once burnished. The reduction in porosity can create a degree of impermeability, for example to keep moisture inside a vessel such as a liquid serving bowl, or to keep moisture out of a vessel such as a storage jar for dry foodstuffs. Two thick-walled vessels with unknown forms were found to have burnished exterior surfaces; one (9-10mm thick) distributed throughout the northern butt end of inner enclosure ditch 1012 (Fig. 9) and the other (11-14mm thick) in pit 1550 (Fig. 13). Whether another large vessel (Fig. 38, no. 9) was similarly burnished could not be established as this vessel is now missing. Burnishing was found on 50% of the 58 identifiable vessels with diagnostic forms, the majority having the technique on the exterior only (14 jars and one saucepan), 12 bowls and a saucepan have both surfaces burnished, and a single bowl was burnished on the interior only. Most significantly, however, burnishing was found mainly on fabric types TF1A and TF1B sherds, representing 54% of all the 279 burnished examples. In addition, smoothing of surfaces was found on four identifiable vessels and wiping on two.

Two vessels (Figs. 38-39, nos 2 and 45) were burnished in a special style, which is here referred to as 'HV pattern burnish' (HV: horizontal/vertical). It consists of the rim to neck zone of the vessels displaying consistent, high-intensity burnish in a horizontal direction, while below the neck zone the body of the pot is burnished in a contrasting vertical pattern. There are no gaps between the burnished vertical lines. Therefore, this is a covering technique but executed in a vertical style that deliberately emphasises the method of application rather than attempting to create a smooth, metallic appearance as found on most burnished vessels in the assemblage. HV pattern burnish is commonly found in late Iron Age ceramic phase E at Beckford but first occurs in later middle Iron Age ceramic phases C and D. One of the Blackstone examples is identical to a type 3.82 middle Iron Age form at Beckford (Fig. 38, no. 2), while the other could be called a tubby cooking pot vessel (Fig. 39, no. 45). The latter is a vessel type recognised as late Iron Age and early Roman in date (Beckford type 14; Wills in prep. [2010]) and shown by petrological analysis to have been consistently made from clays containing igneous and metamorphic rocks identical to those forming the Malvern Hills (Peacock 1965-7, 15). The Blackstone examples with HV pattern burnish were both made from TF1A. Therefore, while one vessel is acceptable as middle Iron Age in date, the other is more likely to have been made and used during the late Iron Age period, and together they present a distinctive surface treatment technique at Blackstone. Vessel sizes and evidence of use

The Blackstone Iron Age assemblage has a very distinctive range of vessel sizes which can inform on the nature of food consumption and group size, as well as the nature of pottery production. Tables 9-10 present the frequency of recognised vessels by form-type, fabric type and rim diameter. The diameter measurements are categorised by 2cm intervals, with 8cm representing 8 to 9.9cm, 10cm representing 10-11.9cm, etc. It is possible to visualise five different size groups of pottery based on the holding and lifting of vessels by their users: very small pots, less than 10cm in diameter; small pots, 10-19cm; medium-sized pots, 20-29cm; large pots, 30-39cm; and very large vessels, 40cm or larger (cf. Fig. 37). It may be appropriate to consider these as representative of user group sizes: very small pots for use by small children or for special functions; small for use by individuals as the capacity would feed an adult; medium-sized for family use; large and very large pots for food and drink preparation, serving of large groups and for storage. While it would be far more accurate to adopt such categories using the capacity of vessels if this could be calculated, the consistency of Iron Age vessel shapes during this period and in this region allows the use of rim diameters as a substitute method of measurement (cf. Woodward and Blinkhorn 1997). This is clearly the case at Blackstone, where nearly all the vessels are ovoid/barrel-shaped/round-bodied or straight-sided, with one exception, which is a high-shouldered bowl (Fig. 38, no. 9).

Figure 37

Figure 37: Prehistoric pot sizes.

Four very small vessels were identified in the assemblage. These were made in four different fabrics (TF1A, TF2, TF3, TF6) and three different vessel types (TV1B, TV3 and TV10; Figs 39-40, nos 21, 28, 55 and 62). The majority of measurable vessels (77%; 33 pots), however, were made in the small group size range, and all three of the commonest fabric production sources were making this range. Similarly, these three sources were producing vessels among the limited group of medium-sized vessels. It is very interesting to see that there is only one large vessel in the assemblage and it is made from TF2 (vesicular mudstone) in a unique vessel shape (TF4; Fig. 38, no. 9), which is a bowl form probably for preparing and serving food or drink. The most likely source for TF2 is the closest of all the regional sources to Blackstone. Such a limited range of vessel sizes, dominated by the small or personal size range, is in slight contrast to the normal distribution for middle Iron Age assemblages which usually has a more significant number of medium-sized vessels although they are never the largest group. It is also unusual to see that the only large vessel represented is a bowl rather than the more typical middle Iron Age storage jar (Kenyon 1954, fig. 10, no. 8; Dinn and Evans 1990, fig. 17, no. 2; Morris et al. 2005, fig. 44, PO81, plate 54), but late Iron Age conical bowls are known in the area (Dinn and Evans 1990, fig. 17, no. 1; Trow 1988, fig. 38, no. 133). The very small vessels are also relatively more frequent in the Blackstone assemblage than elsewhere in the region where one or two vessels would have occurred in a collection of this size (Beckford; Wills in prep. [2010]). It is highly likely that these differences are a result of both the date of the occupation of this site and site function or specific site activities (discussed further below).

The functions of some vessels can be deduced by the actual observation of evidence of use such as soot on the exterior, burnt residue on the interior, and abrasion on the interior. No instance of lime-scale was noticed, though the possibility of this depended on the source of the water used, and, due to the lack of calcareous fabrics in the assemblage, there were no examples of interior pitting from holding acidic foodstuffs. Twenty-five examples of cooking pots were found in all three main fabrics, with 13 vessels in TF1A and ten made from TF3 but there is only one occurrence of a vessel made from TF2 fabric (Fig. 39, no. 50) and also a small TV3 pot made in minor sandy fabric TF6 (Fig. 40, no. 55). The cooking pot sherds were found scattered across the site in settlement features and in both the inner and outer enclosure ditch fills. The largest concentration of 'cookpots' with sooted exteriors was found in pit 1039 (PG1; Fig. 9). All of the measurable examples were very small or small in size and vessel forms include TV1B, TV1C, TV2, TV3, and TV10 with the commonest cooking pots being TV2 vessels. What was most interesting to discover was the frequency of burnished vessels (11 examples), and in particular bowls (8 examples), which had been used as cooking containers. The only linear-tooled vessel (Fig. 40, no. 53), found in inner ditch recut 313, was a cooking pot. Only four pots displayed evidence of interior abrasion, which can be caused by over-zealous use or cleaning in antiquity or from the etching of vessel interior surfaces from contact with acidic foods. The most striking examples of this effect was found on numerous body sherds from a TF1A thick-walled vessel deposited in the inner enclosure ditch butt-end, which suggests that this undiagnostic pot may well have been a storage jar used to hold acidic liquids. The other examples include a burnished bowl or saucepan sherd from pit 0150 (PG5; Fig. 10) made from TF1A, a TF3 sherd from pit 136 and the very small TV10 vessel (Fig. 39, no. 21) found in gully 0109. Dating by association and stratigraphic relationships

All of the pottery is quantified by fabric within each context in Table 11 and by form in Table 12. Examination of key groups deposited together in features reveals three broad patterns of association. The majority of features can be characterised as containing handmade pottery dominated by variations of plain ovoid jars with occasional occurrences of one or two bowls, the final layers of two features have handmade bead-rim vessels in addition to ovoid jars and bowls, and one feature has two sherds of wheel-thrown pottery.

Pit 0098 (PG3; Fig. 10) has eight handmade vessels identifiable to form (Fig. 38, nos 12-19), including single examples of TV1A, TV1B, TV3, TV9 jars, three TV2 jars, and two TV10 bowls. The TV1B jar displays HV pattern burnish. Pit 1039 has five vessels (Fig. 39, nos 39-43), including three TV10 jars, a TV2 jar and a bowl or saucepan pot of form TV1B. Pit 0001 (PG1; Fig. 9) has four vessels (Fig. 38, nos 1-4), including the two largest examples of TV2 jars and two TV1B bowls. Pits 0136 and 0139 (PG2) each have three vessels, including the two examples of low pedestal bowls (TV7A, B2), a necked bowl (TV1B), a saucepan pot (TV1C) and two jars (TV2, TV10) (Fig. 39, nos 27-9 and 30-2). Therefore, all of these features with sizeable collections have similar types of handmade, undecorated pottery. Examination of Table 11 reveals that they all have the same fabric types present as their most common wares. It is only the presence of the low pedestal bowls, made from the epidote-rich fabric TF1B, in pits 0136 and 0139 that suggest a slightly different date from the other deposits. The only other feature containing a diagnostic vessel made from this fabric was pit 0246, which had the unique, highly burnished TV11 rim sherd (Fig. 39, no. 34) as its sole potsherd.

The northern butt-end cut 1012 of the inner enclosure ditch contained only one diagnostic vessel (Fig. 39, no. 45), a type TV1B bowl (that may be interpreted as a tubby cooking pot) that has HV pattern burnish and soot on the exterior. Cut 0305 of the southern part of this ditch held sherds from four vessels (Fig. 39, nos 46-9), including two fabric TF3 examples of TV2 jars, a TV12 jar and the only TV5 bowl in the assemblage, while the inner enclosure 1549 (south-west corner) contained two pots (Figs 39-40, nos 50-51), another TV2 jar and a TV8 jar/bowl. A possible recut (0313) of this ditch contained a fabric TV1A version of TV2 and a TF5 sandy fabric version of jar form TV6 (Fig. 40, nos 52-3). Therefore, the range of vessel forms is virtually indistinguishable between the primary infill of the ditch and the infill of the recut, and all are handmade, later Iron Age types likely to date from the 2nd to 1st century BC, the most distinctive of which is the vessel with HV pattern burnish. It is significant that none of the sherds from the inner enclosure ditch derives from a wheel-thrown vessel.

Figure 38    Figure 39

Figures 38-39: Iron Age pottery.

The southern butt end of the outer enclosure ditch (1014) contained only two jars (Fig. 40, nos 54-5), the only example of a TV7B, which is very graceful in form and highly burnished, and a sandy fabric version of form TV3, while recut 1548.14 of this ditch contained the only wheel-thrown latest pre-Roman Iron Age pottery on the site (Fig. 40, no. 56). This small quantity of material suggests that the infilling of the outer ditch, at least with regard to its recut, took place at a later date than that of the inner ditch. It is, however, curious that the primary infilling of the outer ditch did not include any of the most common vessel forms in the assemblage such as TV1B, TV2 and TV10 but rather a unique, very graceful, black and shiny, small jar and less common vessel in a minor sandy fabric (Table 8) and also eight large sherds (179g) from a thick-walled vessel with exterior burnish, interpreted as a possible storage jar in fabric TF1A. Therefore, the character of the small assemblage from the outer enclosure ditch is different from the inner enclosure ditch and the interior features. It seems that the outer enclosure ditch was recut at some time only very slightly later than the infilling and recut of the inner enclosure ditch – but this cannot be strongly substantiated with firm evidence due both to the paucity of material from this feature and our uncertainty about the social climate for acquiring Romanised pottery in this area during the late Iron Age; it is not known that the occupants of Blackstone were trying to acquire Romanised pottery at this time, but they were using a range of handmade, undecorated pots which were clearly late Iron Age types, in particular an extraordinarily large bowl with thickened interior to its rim.

Figure 40

Figure 40: Iron Age pottery.

The contemporaneity of pottery from pits 0080 (PG3) and 0116 (PG5) is established by the presence of sherds from the same Clee Hills dolerite fabric, burnished TV1B cookpot (Figs 38-39, nos 8 and 23) in both (Table 12); this is the only vessel which can be positively identified as distributed between two features. A similar effect can be suggested, on different criteria, for pits 0136 and 0139 (both PG2). The only two examples of the low pedestal, black and highly burnished, B2 base vessels (TV7A) were found in these pits (Fig. 39, nos 27 and 31). They were clearly made by the same potter, in the epidote-rich fabric, and one is slightly larger than the other.

The sealing layer over pits 0080 and 0098 (layer 0080.1) contained a bead rim from a handmade jar while the only other example in the assemblage was recovered from the top fill of inner enclosure ditch, context 0305.1 (Figs 38-39, nos 7 and 47). While this form type is accepted as typical of the late Iron Age, it is important to emphasise that these examples are late in the infilling sequence of their respective features.

It is probably quite significant that the size of sherds in this assemblage, the mean sherd weight in grams, is very large compared to most Iron Age assemblages in the region, as it is the best indicator that Blackstone does not appear to have been a location of settlement activity over a long period of time. Instead, everything seems to suggest that there was no middle Iron Age activity on the site at all and, therefore, no redeposition of middle Iron Age pottery into late Iron Age, or Roman period features.

The overall impression based on the vessel forms, methods of manufacture, fabrics, surface treatment and lack of decoration indicate that the majority of activity at the site, including the infilling of the inner ditch and all of the Iron Age pits and other interior features, took place during the later Iron Age. The similarity of the range of vessels and fabrics suggests that this phase of occupation was not more than 50-75 years in length (approximately two to three generations). The recut of the outer enclosure ditch, with its wheel-thrown vessels, and the final infilling or abandonment of those features with layers containing the only bead-rim jars in the assemblage, took place at a slightly later date and appears to have been of very short duration, as indicated by the small quantity of this material in the assemblage. Nothing in the Iron Age pottery assemblage alone would indicate that activity occurred beyond the second half of the 1st century BC. Most of the pottery belongs to the later middle Iron Age repertoire of this area, characterised by the presence of ovoid jars, jars and bowls with upright rims of various kinds and saucepan pots with both stamped decoration and linear-tooled designs, together with infrequent examples of HV pattern burnish surface treatment as well as regular burnishing, designated ceramic phase D at Beckford. However, the inclusion of one possible tubby cooking pot and the overwhelming amount of regionally made vessels in the assemblage (more than 90% by weight) pushes the date for the majority of pottery later into the transitional ceramic phase D/E at Beckford. It is most likely that the activity that included the deposition of wheel-thrown sherds from two vessels into the recut outer ditch fill belongs to the ceramic phase E, which is characterised by the first appearance of wheel-thrown pottery, among a large array of other types of vessels and fabrics that do not occur at Blackstone. Therefore, the Blackstone assemblage is later Iron Age in date, from the end of the middle Iron Age into the late Iron Age period, c. 2nd to 1st century BC. Deposition and spatial variation

There appear to have been five different kinds of pottery deposition at Blackstone: two of these may be interpreted as various evidence of everyday life, one is hard to suggest an interpretation for, and two others may have been the result of different special events. The following discussion is simply a general exploration of deposition of pottery, and should be viewed with caution, and seen as a study worthy of further analysis using a levelling factor that considers the amount of soil removed from each context.

Many excavated features (c. 92%), probably including numerous animal burrow-like scrapes, contained no prehistoric pottery at all and 38 (c. 4% of all features) contained fewer than 20 sherds or less than 250g of pottery (Deposition Category 1) (Tables 11-12). These are the two commonest forms of deposition, or lack of pottery deposition in the case of the former, and because of this must reflect everyday processes at the site. In some cases these deposits may consist of a single small sherd, or a handful of sherds from one vessel, or a few sherds from a couple of vessels, indicating that considerable fragmentation of the pots had taken place prior to their purposeful or casual deposition. It is likely that these sherds had been broken during daily use as cooking pots or serving and storage vessels, had been left lying around the activity areas and eventually became incorporated into either a midden discard zone on the surface of the site or deposited with other rubbish into these features. The features with no pottery attest to the lack of purposeful or casual deposition of durable material in these cases.

Six interior features or cuts through ditches contained variable amounts of pottery. These deposits have between 250 and 1000 grams of material recovered but the number of sherds in each varies considerably from 18 to 84 sherds (Deposition Category 2). Outer ditch cut 1014 and inner ditch cut 0305 contained the smallest numbers of sherds (18 and 21), but they are not the smallest size of sherds in this pattern group, with mean weights of 17g and 16g respectively. Pits 0001 and 1039 (PG1) have modest numbers of sherds recovered from them (24 and 26 respectively) but their mean sherd size is relatively large at 22g and 31g respectively. In complete contrast are pits 0136 (PG2) and 1550, which each had large numbers of sherds (50 and 84) but their average sizes were very small for this assemblage as a whole (9g and 8g). This suggests completely different biographies of the different vessels in each pit compared to the previous examples. At the present time, examination of the pottery alone provides no obvious interpretations to place on this variety of deposits.

In contrast are the deposits from two pits (in PG3) located together in the northern part of the excavated interior area. Stratigraphically, pit 0080 cut the top fill of pit 0098 but no sherds from the same vessels found in 0098 had been redeposited into 0080. This seems to indicate that little, if any, real disturbance of 0098 took place with the digging of 0080. Both deposits contained the remnants of fires or ashy hearths. These two pits each held the largest quantities of pottery from any other excavated feature or cutting on the site; 4.025kg recovered from pit 0080 and 2.149kg from pit 0098 (Deposition Categories 3 and 4). The character of the pottery recovered from each is, however, very different in nature. Pit 0080 retained sherds from nine vessels (cf. Fig. 38, nos 7-9); two made from TF1A (5 sherds, 44g), two from TF2 (2 sherds, 3934g), four from TF3 (5 sherds, 31g) and one from TF4 (3 sherds, 16g). One vessel, found in context 0080.3 at the bottom of this pit, was recovered in a nearly complete state; only the base was entirely missing from this large and unique bowl (Fig. 38, no. 9; currently not available). The other eight vessels found distributed in layers 0080.1, 0080.1a and 0080.3 are represented by single or very few sherds in each case and the mean sherd weight for these is modest (between 5-9g). Therefore, it is possible to conclude that the large bowl was deliberately placed into this feature, while the other sherds may represent token deposition of fragments symbolising the other pots or casual incorporation of stray sherds into the soil layers making up the deposit.

The pottery from pit 0098, which is chronologically indistinguishable from the pit 0080 collection, is quite different in detail. At first glance it seems that the number of vessels identified, the fabrics present, the range of forms with the exception of the large bowl, and the lack of decoration are similar between these two key groups. However, the quantity of each pot recovered is entirely different and the nature of the deposition layers in this pit also suggest something different. Seven vessels are represented in the pit 0098 collection (Fig. 38, nos 14-20), not including the material from context 0098.1, which was a disturbed layer sealing parts of this pit and pit 80: two each made from fabric TV1A (48 sherds, 1009g) and TF2 (33, 365g), and three made from fabric TF3 (46, 660g). One vessel is represented by the total profile of the pot and includes approximately 90% of the original pot, while a second pot is represented by nearly the total profile of the vessel and includes approximately 55% of the vessel. A third vessel is substantially represented with approximately 25% of the pot recovered. Much less of each remaining vessel in the key group was recovered but the sherds are larger in weight than those found in pit 0080. Importantly, four of the vessels had either joining sherds found in different layers in the pit or at least sherds from the same pots in different layers; this did not occur with any of the pottery from pit 0080, all of the large bowl in particular being recovered from the same lowest layer.

It is possible to interpret these two key group deposits differently based on work conducted by Cunliffe (1992) and Hill (1995). The deposit found in pit 0080 could be a vessel used to propitiate the gods of the underworld – a single large and special container filled with food, or more likely drink, to satisfy and placate the controllers of the world of agricultural uncertainty. This could have been a linked contribution, as the drink may have been beer made from corn which was to be stored in below-ground facilities, thus linking the gods of the underworld with the storage and growing of grain and the occupants' appreciation through alcoholic drink. The pottery from pit 0098, on the other hand, is represented by several different smaller pots that are not unique or special in any way, as these types and sizes of vessels are found elsewhere on site and likely to represent everyday life. But the clustering of so many pots, and such large parts of some of them, seems to indicate that a special gathering had occurred to conduct an event and that this may have involved the provision of food. The small sizes of the vessels, however, do not suggest a large gathering but rather a group of individuals performing a task and being rewarded for their work. It is tempting to suggest that this may have been the construction of the second, outer ditch. Regional identity and new cultural influences

There is good evidence to suggest that the occupants of the Blackstone site were able to obtain artefacts that demonstrated a more Romanised influence being apparent during the late Iron Age in the wider region, as at Bagendon (Clifford 1961), Salmonsbury (Dunning 1976) and The Ditches (Trow 1988) in Gloucestershire, and at the same time to maintain their roots in the very strong regional identity derived from the middle Iron Age. The presence of bead-rim jars and vessels with thickened interior edges to the rim, albeit handmade, and wheel-thrown vessels are indicators of this development in the area. Blackstone is quite a distance up the Severn from the area normally associated with these distinctive changes, and, therefore, it is not surprising that continental imported goods cannot be found in the ceramic record of this 1st-century BC activity. Nevertheless, the pottery recovered demonstrates that these influences were known by the occupants of the site, and in particular by the potters in the region. It may well be that the two wheel-thrown vessels were early examples of Severn Valley ware, which has been shown to have been made and used prior to the Roman conquest (Timby 1999, 40; 2000, 363). A simple reading of the pottery evidence suggests that they were maintaining their traditions, while at the same time being open to new influences and ideas, as part of the process of forging a new identity.


© Internet Archaeology/Author(s) URL:
Last updated: Wed July 21 2010