The criteria for defining ritual or special deposits away from religious sites and other potentially sacred locations (e.g. watery places) have been discussed, particularly from the British Iron Age (e.g. Wait 1985; Hill 1995), but less often for the Roman era (Fulford 2000). This is a problematic and complicated area, since deposits are so varied and definition will depend on interpretation and some assumption. Here we might define unusual groups of material which do not appear like the artefact debris encountered in by far the majority of settlement contexts as 'structured' if intentional selection has seemingly determined the composition of a group. In the case of samian this might be through the presence of whole or near complete vessels, unusual proportions of certain types, associations between samian vessels, and through the occurrence of samian with other finds indicative of selection. Such attributes may distinguish these deposits (to employ some short-hand) from de facto deposits and secondary rubbish (cf. Schiffer 1976). Yet individual sherds might have been regarded as sufficiently symbolic or have been defined by past actors as 'representative' and so structured deposits are not limited to the incidence of whole pots or substantial parts (cf. Chapman 2000).
Thus detecting structured deposits may not be easy; so often the archaeological record can be equivocal. There are, nonetheless, an increasing number of 'identified' examples of symbolic or ritual deposition in the Roman era away from religious foci, that is temples and shrines (Clarke 1997; Fulford 2000).
Structured deposits at settlements and elsewhere are not abnormal, but a variation (albeit uncommon) within the range of types of deposits likely to be encountered in the archaeological record. What lies behind them is less clear. Certainly they represent 'events', such as, potentially, gifts or items relating to feasting. These events may be bound up with rituals developed out of such domains as:
Samian is markedly prominent among many such deposits (e.g. Willis 1997a; 1998a). Structured deposits warrant analytical attention since they provide an opportunity to interpret how the people of the time acted and may have thought. Some examples are assessed here.
At the Roman Small Town of Alcester, Warwickshire, two pits (pits B I 3 and D II 29/29a) excavated during the 1964-6 works in the Birch Abbey area contained unusual groups of artefacts. Pit D II 29/29a was located by a main road in an apparently open area and without associated features. The main assemblage of unusual finds came from the lower secondary fill, and this was dated c. AD 150-160. It yielded sherds from around 69 samian vessels. Most of the vessels, both plain and decorated forms, were represented by substantial fragments, with many almost complete or more than half complete; the group is composed of contemporary vessels showing little ware (Hartley et al. 1994, 106-10). Decorated bowls account for 43.5% of these samian vessels, which is an exceptionally high proportion when compared with the equivalent figures for other smaller civil centres (cf. Table 35). In reporting the samian Hartley et al. postulated that this samian might be thought 'most likely to have come from a shop selling pottery' (1994, 107), though note that, 'one wonders whether the large group ... is due to something more than an accidental breakage in one pottery store' (1994, 94). The group looks even more extraordinary when the other types of remains are considered. This pit, which was perhaps roofed, also contained at least six unworn glass vessels and a complete Dressel 20 amphora, minus only its rim, a dog skeleton, oyster shells and nails (Cracknell 1994, 252). The main fill of the pit held a fragment from a copper-alloy bracelet, a copper-alloy plaque and an iron stylus (Cracknell 1994, 252). Cracknell postulates that 'duplicates' among the samian and glass vessels imply this was the discarded stock of a trader (1994, 252). Glass and samian were specialist traded items that would doubtless have been sold by the same people or in the same shops and the samian and glass might have been deposited following an accident. This remarkable group, however, seems open to interpretation as evidence of a 'special event' of ritual connotation (cf. Hill 1995). The bracelet fragment, stylus and oyster shells, to say nothing of the plaque, are items frequently found with votive deposits from elsewhere in Britain. It may be that the debris relate to a feasting event, or perhaps a roadside shrine.
A somewhat analogous deposit of essentially similar date to the Alcester group was recovered at Park St, Towcester, Northamptonshire. Likewise the exceptional group of samian is associated with a set of glass vessels (Lambrick 1980). The group derives from a pit fill in an area of pitting. Potentially significant is the fact that this particular 'pit' had been cut into the top of a filled-in well (or is perhaps this was not a pit in itself but a sinkage fill in the top of the well). The deposit was characterised in 1980 by the excavator as the rubbish from 'some wealthy establishment' (Lambrick 1980). It is now possible to see this deposit, like that at Alcester, as open to a different interpretation, as a structured deposit relating to perhaps a symbolic offering, or related to feasting.
A complete Drag. 33a and a near complete Drag. 33a were recovered from a pit (Pit 8.1) believed to be a rubbish/cess pit at Newarke St, Leicester (Cooper 1996, fig. 25, nos 14 and 15). Both vessels are from La Graufesenque and they lie within the c. AD 65-90 date range (Pollard and Dickinson 1996). The two vessels might therefore have been a pair. To encounter one complete vessel is very unusual among normal site assemblages, to find two from the same feature, of the same type, suggests that they were intentionally discarded, for an unknowable reason. Coarse pottery in a qualitatively similar condition was also recovered from this feature, while animal bone, vessel glass (represented by several fragments, being an unguentarium or small flask), small finds and plant remains were also found. The fill of Pit 8.1 is dated to c. AD 100-25 by the finds. It is suggested that the pit functioned as a rubbish pit. However, the preferred interpretation in the report is 'that the pit served as a cesspit or soakaway with the majority of the finds resulting from secondary use as a convenient rubbish pit when it became redundant ... identification as cess is confirmed by [palaeo-biological remains]' (Cooper 1996, 8).
During the 1993-95 excavations at Elms Farm two Drag. 30 decorated bowls were recovered from the same feature in Area K, specifically context 4148 (Atkinson forthcoming). The bowls were manufactured in Southern Gaul. They were buried (?upright) in a feature otherwise lacking material culture. The vessels are provisionally identified as a structured potentially votive deposit (Mark Atkinson, pers. comm.). Their find spot was not very distant from the site of the temple at Elms Farm, though in a different zone.
A pit group from Pit l 20008, fill 20009 in Area L, contained an apparent structured deposit. Samian was far from prominent among the finds which included 11 complete or near complete pottery vessels. The date range of the group is c. AD 55-80. The group size was 10.7kg, less than 1% of which is samian; the EVE total is 12.48, 1% of which is samian. In fact there were only two sherds of samian among this group.
The excavations at Baldock published by Stead and Rigby encountered a widespread incidence of unusual deposits with various artefacts found, often together in pits, indicating ritualised activities at what ostensibly appeared a normal settlement site of middle rank (Stead and Rigby 1986, 47, 257-9). A number of groups with 'complete' pots were recovered; samian vessels occurred among these groups. Groups of this type from the site are documented in the database. From the Upper Walls Common Site A, Pit A 12, fills 7 and 8 dating to the Claudio-Neronian era, 22 samian vessels were represented. Of these samian vessels, five were almost complete; several coarse ware vessels represented were also virtually complete. Some near complete vessels were also present in the group from Pit A 19, layer 7, (Baldo2).
Stead and Rigby (1986, 47, 257-9) noted that the composition of certain pit fills at Baldock included complete pots which appeared to have been deliberately selected, and noted that these resembled burial groups in the limited range of types represented. It is possible that the pits, or at least some of them, were filled water holes/pits. The creation of these groups, for whatever reason, was an endemic practice at this site through time and space.
A very rare decorated black samian beaker of Déch. form 66, was found complete in a pit within an enclosure at the rural site at Nursling (cf. Appendix 6.2; Adam et al. 1997). The enclosure contained very few other features. The vessel had figured decoration with an erotic theme. This is an altogether unusual discovery. This is a case of a highly uncommon vessel, potentially of some value, discarded or lost in an otherwise rather mundane rural setting; no direct clue as to the function/s of the pit were forthcoming. The pit was cut by a later pit and it might be that the pit was a grave, perhaps an unoccupied grave, though this was not apparent (Adam et al. 1997); alternatively the vessel may be a votive item. Other, less prosaic possibilities, are conceivable for this odd occurrence.
During the Bullock Down landscape project, by Beachy Head, East Sussex, a series of sherds from a cut-glass technique decorated samian beaker of Déch. form 72 were recovered from a 'marl pit' on a field boundary (Drewett 1982; Rudling 1982, 104; the vessel was Central Gaulish and Late Antonine). This form is comparatively rare and could have been held in some value within a rural environment such as its find location. The find-spot was apparently some distance from domestic occupation, in an agricultural landscape. It is possible that it was intentionally selected for deposition at a boundary location for symbolic reasons, though this, of course, is a matter of possibility and interpretation.
© Internet Archaeology
URL: http://intarch.ac.uk/journal/issue17/1/12.4.html
Last updated: Mon Mar 7 2005