5. Discussion

This site saw the creation of a carefully engineered well in the Late Roman period, positioned high on the hillside and using newly acquired, good-quality masonry: all in contrast to earlier wells. This occurred at a time of considerable landscape change, marking a significant alteration to the monumentalisation of this zone. Not everything about this feature was new, however. The masonry lining of the well incorporated, in an architecturally awkward way, a finial once used atop a stone roof. The original source of this architectural element is unclear, but the construction of the well signalled its move from an exposed height to a semi-concealed depth: it would have still been visible to users of the well in its new position as a noticeable intrusion into the regular coursing of its lining. This item was, seemingly, the only element from any earlier structure to be reused in this way, and its recycling is best interpreted as symbolic rather than opportunistic.

Finds from the well's fills, considered as a single assemblage, are also distinct from the rest of the site in many ways: there is very little residuality or intrusion; jars dominate the ceramic groups and the faunal material is distinct in terms of bone condition (little gnawed or burnt, much butchered), ABG numbers and taxa represented. This assemblage is thus far from being domestic rubbish, a characteristic that is mirrored in the distinctive filling of other late Roman wells in Yorkshire e.g. at Dalton Parlours (Sumpter 1990) and Rudston (Stead 1980).

Yet it would be wrong to assume that late Roman wells in the region were always filled in such an unusual way. In ceramic terms, at least, the Heslington assemblage contrasts with that from the broadly contemporaneous well at Rothwell Haigh, where a much wider range of tablewares and vessels used in food preparation and cooking were present (Leary 2011). The latter patterning is also true of the waterhole at Shiptonthorpe (Millett 2006), albeit alongside some complete vessels. Animal skulls and complete animal burials, as well as a writing tablet, from the latter feature might suggest ritual activity here. Clearly, where one artefact type has symbolic implications, there is no reason to expect that other assemblages will follow suit.

When individual fills are examined, this diversity of response becomes even clearer. Thus our lowest fill, 2109, yielded larger amphorae than elsewhere and bones suggesting low-level, unintentional discard, together implying deposition in a functioning feature. Pottery cross-joins with the overlying fill, which seems to mark the start of the well's demise, suggest that this initial deposit must belong to the very last use of the well. Thus, assuming that the feature did not have a very short life (as suggested by the investment in its construction), it must have been kept clean before this: any faunal or ceramic signatures from this first layer tell us only about a short moment in time (in contrast, for example, to 1.5m of 'usage' deposits in the base of the Dalton Parlours well: Wrathmell and Nicholson 1990, fig. 115).

Overlying fill 2093, deposited at a time when scrub, and perhaps even heathland, were evident in the vicinity and insects and frogs/toads were falling into the well, was presumably forming soon after 2109 (see cross-joins, above). 2093 seems certain to have marked the feature's demise following the deposition of two adult female pig skulls, an act that must have fouled the water source. If this was deliberate, then the choice of two mature sows, no doubt responsible for numerous litters, may link this action to fertility rites (the deposition of cattle and horse crania elsewhere in the well strengthens the notion of the symbolic significance of crania: Wilson 1999, 303). That said, other bones from this layer indicate simple carcass processing and meat consumption, so there is no proof here of symbolic intent (although this group does include butchered horse and a solitary red deer bone, both discussed in more detail below).

By the same token, fill 2093 yielded a wooden bucket and, just above this, a virtually complete Huntcliff-type jar, both of which might be taken as some form of structured deposition, especially given that this type of jar was found in wells at both Rudston (Rigby 1980) and Dalton Parlours (Sumpter 1990), along with complete buckets at the latter (Morris 1990). However, there may still be some reason to question this interpretation. Other pottery from this fill was incomplete when recovered, thus suggesting breakage elsewhere and discard, something supported by the CBM sherd size. Equally, the bucket may have been deposited with its iron hoops, something that would surely have been retained and recycled if the item was intended as rubbish, yet it lacked a mount and handle. This artefact can most reasonably be interpreted as the result of accidental breakage, presumably when in use with the well.

In terms of evidence of ritual in 2093, then, much depends on how one interprets the complete Huntcliff-type jar. Such vessels are commonly associated with wells, but they are also evident elsewhere on many sites, including Heslington East. Vessels of this type are frequently sooted from use, particularly around the neck and upper body, and sometimes scorched. In addition lime-scaling is common inside them, this combination of features leading to the suggestion that they were used to heat water (Sumpter 1990, 244). Hence this jar might be interpreted in similar vein to the bucket, as an item that could have been used with the well. It need not have been deliberately selected to be deposited there at the end of its life.

The next pair of deposits in the well, 2046 then 1979, provides perhaps the most sustained case for involving ritual elements in their formation. Forming at a time when there was wasteland vegetation in the general vicinity, they included pitfall insects and frogs/toads and seem to have been deposited as a single activity. 2046 included a very large cobblestone, which would have precluded any attempt to get at the water. Faunal material from these two fills was also unusual. 2046 suggested the rapid disposal of butchered cow and horse. Apart from two small fragments, bones from overlying 1979 comprised only immature deer, dog and calf, seemingly selected by age, alongside an adult cow and the sacrificing of a large antler, valuable raw material. The deer and dog are clearly unusual and will be discussed more fully below.

The unusual concentration of horse butchering from this horizon, in an animal thought to be rarely consumed in the Roman period (Toynbee 1973, 185), might also be considered significant. It must be remembered, however, that the notion of revulsion to horsemeat is derived from documentary sources concerned, in the main, with elite activities at the centre of empire. When one looks at the purely archaeological material, as done for the Eastern Dutch River Area in the Roman Netherlands (Lauwerier 1988), a more complex picture is evident, in which some settlements do seem to avoid such practices (marks on horse bones here being attributed to dogs) but others are consuming horses more regularly. Similar diversity could be evident in our study area, as implied by the consumption of horsemeat at Rudston (Chaplin and Barnetson 1980, 152). Thus one cannot assume that butchering evidence from Heslington East is somehow exceptional. Indeed, hippophagy must have been common enough in the Roman period for later Christian laws to attempt to forbid it, portraying it as linked to unacceptable pagan practices (Poole forthcoming).

These fills yielded far less pottery than underlying strata, but did include an unusual long-necked beaker and a complete grey ware jar, without lugs, the latter perhaps already rather old when deposited. This form sometimes has worn areas on the girth and chipped rims, perhaps partly a result of being used to draw water (an example from the Dalton Parlours well was found with a cord fragment still tied to the lug: Sumpter 1990, 244). It is also similar to a complete vessel, with lugs, found at the bottom of the Langton well. The excavators (Corder and Kirk 1932), commenting on its fragility compared to Huntcliff types, suggest that it must have been used by those 'unwilling or not permitted to use the ordinary well gear' (Corder and Kirk 1932, 52). Lugs are absent from the Heslington vessel, however, so it may not be so directly related to our feature. More generally, the lack of sooting and limescale on these grey ware vessels suggest a contrast with Huntcliff-type jars. The unusual beaker and complete grey ware jar, in a layer largely lacking other ceramics, and whose deposition made the well clearly unusable, combine to suggest the deliberate closure of the well.

All subsequent fills had similar assemblage signatures to each other, comprising small amounts of pottery of a range of forms and dates, probably derived from the sides of the well and drifting in from the surface, and few bones beyond the occasional pitfall victim. The very latest parts of the sequence almost certainly comprise post-well accumulations, including residual material, mixed CBM and some ceramics of Anglian date.

The above discussion shows, then, that in trying to distinguish any particular ritual component in this well, one must not look simply to particularly complete pots or ABGs, or even to whole faunal or ceramic assemblages. Rather these components must be considered in relation to each other and in relation to deposit information (and, even then, one cannot be completely certain, most obviously over the status of the material recovered from fill 2093). The closure of the well is marked most obviously with the deposition of 2046/1979 containing a large cobblestone, the complete grey ware jar, plus the adult cow and large antler, set beside the calf and immature red deer and dog. These last two faunal elements deserve wider consideration.

The occurrence of the sub-adult red deer is not without precedent within late Roman wells. Occurrences at Baldock, Hertfordshire, were believed to represent the natural deaths of animals trapped in the well head (Chaplin and McCormick 1986, 410-11), while two carcasses within a well from Burst, Belgium, may have been hung there but never retrieved (Ervynck et al. 1987). Other evidence might take us beyond functionalist explanations, for example with the partially articulated remains of an immature red deer in a well at Drapers' Gardens, London (Gerrard 2011, 555). Deer bones are rare in Romano-British contexts, whether locally in York (Foster 2012, 11; Foster and Jaques 2012, 6; O'Connor 1987, 7) or more generally. Cool (2006, 114) suggests that venison consumption 'could well have only taken place in special circumstances'. The act of hunting during summer or autumn, then transporting and consuming some of the Heslington animal, before final disposal in the well, must have had particular meanings for this agricultural community.

The finding of dog skeletons in wells is a 'common theme' on Roman sites (Black 1983; Fulford 2001, 214-5), local examples being Dalton Parlours (Berg 1990, 252-3), Rothwell Haigh (Cool and Richardson forthcoming), Shiptonthorpe (Millett 2006, 314) and Welton (Mackey 1999). They have been seen, in a functionalist vein, as reflecting the purge of stray dogs (Maltby 2010, 29) or the convenient disposal of (inedible) 'waste' carcasses (Seeley and Wardle 2009, 156). Yet they could also represent 'termination' deposits (Seeley and Wardle 2009, 148-9), as seems likely, for example, at Swann Street, Southwark (Beasley 2006). It is, after all, the Roman period that sees the first clear indication of pet dogs (Grant 1989) and perhaps dedicated breeding programmes (Bruch 2003). It is notable here that the 15 dogs in the well at Welton, noted above, included puppies alongside mature animals: Mackey 1999, 26). Pit 2601 at Silchester contained an articulated dog skeleton and an ivory folding knife depicting two mating dogs plus raven remains, albeit alongside otherwise nondescript artefacts, and well 5735 contained a dog whose remains were scattered over several fills (Eckardt 2011, 311). Thus the ritual deposition of an immature canid at Heslington East might be understandable at a time when the bond between dogs and humans was undergoing significant transition (Snyder and Moore 2006; see Smith 2006 for a study of domesticated dogs in both Iron Age and Roman periods).


Internet Archaeology is an open access journal. Except where otherwise noted, content from this work may be used under the terms of the Creative Commons Attribution 3.0 (CC BY) Unported licence, which permits unrestricted use, distribution, and reproduction in any medium, provided that attribution to the author(s), the title of the work, the Internet Archaeology journal and the relevant URL/DOI are given.

University of York legal statements

File last updated: Fri Jun 21 2013