The vertebrate assemblage from this site was mainly concentrated in ditch and pit fills and deposits of early 2nd to early 3rd century AD date. Identified vertebrate material consisted mainly of cattle, pig and caprovid remains (no bird bones or wild mammal species were noted) but detailed analysis was hindered by the poor preservation. From the identified remains, the relative abundance of the different species suggested that cattle were prevalent, with pig forming a quarter of the identified assemblage and caprovids being of less importance. Despite the limitations of the assemblage, it does seem to fit well with other vertebrate material of this date, particularly from similar sites which have close associations with military facilities. A predominance of cattle is characteristic of many Roman assemblages and shows a marked change from the dietary preference for lamb and mutton typically suggested by Late Iron Age assemblages.


For the vertebrate remains, data were recorded electronically directly into a series of tables using a purpose-built input system and Paradox software. For each context, subjective records were made of the state of preservation, colour of the fragments, and the appearance of broken surfaces (‘angularity’). Additionally, semi-quantitative information was recorded concerning fragment size, dog gnawing, burning, butchery and fresh breakage.

Identifications to species or species group were carried out using the PRS modern comparative reference collection. Selected elements were recorded using the diagnostic zones method described by Dobney and Rielly (1988), whilst remaining elements which could be identified to species were only counted. Other fragments (classified as ‘unidentified’) were, where possible, grouped into categories: large mammal (assumed to be horse, cow or large cervid), medium-sized mammal (assumed to be sheep, pig or small cervid) and totally unidentified.

Caprovid tooth wear stages were recorded using those outlined by Payne (1973; 1987), and those for cattle and pig followed the scheme set out by Grant (1982). Cattle, caprovid and pig mandibles and isolated teeth were assigned to the general age categories outlined by O’Connor (1989) and Payne (1973; 1987). Where present, epiphyseal fusion data was recorded. Mammal bones were described as ‘juvenile’ if the epiphyses were unfused and the associated shaft fragment appeared spongy and porous. Measurements followed von den Driesch (1976) unless otherwise specified.


In total, 1280 fragments of bone were recovered from four of the excavated trenches. The largest quantity of material came from the area excavation of Trench 12, and ditch 1221 in particular. Poor preservation was recorded throughout and the material was extremely fragmented.  Tables 1 and 2 present summary information for the vertebrate assemblage by context and by trench.

Click for Table 1

Click for Table 2


Bone, amounting to 145 fragments, was recovered from 11 deposits in Trench 1. These deposits included pit and ditch fills, surfaces and layers, which dated from throughout the Roman period. Many of these features were associated with the ditches, surfaces (1st century AD) and subsequent realignment (2nd century AD) of a road. The bones were extremely poorly preserved, which may have resulted from their secondary use as part of the hard core material forming and stabilising the road/trackway encountered in this area. The bones may have originally been discarded elsewhere and then, perhaps, deliberately brought to the site and dumped as construction material.

Six of the deposits gave identifiable remains (Contexts 102, 103, 119, 122, 126 and 183), which were primarily those of cattle, with a collection of tooth enamel fragments from Context 102 which were identified as a horse tooth. Cattle remains included a tooth, metapodials and phalanges from Contexts 103, 119, 122 and 126, with tarsals and radius fragments from Context 183. Some of the large mammal shaft fragments recorded in the ‘unidentified’ component from Context 119 were probably parts of the metapodials recovered from this deposit, as the condition of the bones was poor and the surfaces had split into layers. Tooth enamel fragments in these deposits were not identified to species but those recorded as ‘large mammal’ were probably cattle; they were too badly broken to reconstruct.


Nine of the 11 deposits from this trench which produced bone were fills of a large rectangular feature, Context 403, described by the excavator as a possible ‘tank’ which had been cut into the natural sand. No evidence of a lining for this feature was seen or recorded. Fills of two other associated, but smaller, pits gave another 11 fragments of bone. The fills of the three pits were of early 2nd to early 3rd century AD date and, in total, 294 fragments were recovered, of which just 34 were identifiable.

As with the material from Trench 1, the vertebrate remains were of extremely poor preservation, with a high degree of fragmentation. This was largely the result of fresh breakage damage which had occurred during excavation and post-excavation processes but the very poor condition of the bones in the ground rendered them easily broken when moved. Much surface erosion was apparent, suggesting that the surrounding matrix was not conducive to the survival of bone. Bones were splitting into layers and some were flakey and crumbly. Some fragments of large mammal tooth enamel were present, but these were also of rather poor preservation.

The remains were mostly identified as cattle and pig, with a very few sheep/goat bones also being present. Some fragments were recorded as possible cattle, because their poor condition made a more precise identification impossible. A range of skeletal elements were recorded for cattle, including major meat-bearing bones such as humeri and femora, and fragments typically interpreted as primary butchery waste e.g. mandibles, isolated teeth and phalanges. The pig remains were clearly biased in favour of teeth, with a mandible and scapula fragment also noted.

The unidentified fraction or fragments which could only be identified into broad categories, such as ‘large mammal’ and ‘medium-sized mammal’ consisted mainly of tooth enamel, shaft and rib fragments. These were almost certainly cattle, particularly the enamel fragments, but the presence of fragments of horse or even perhaps large deer (both of which have bones of a similar size) cannot be ruled out especially as the preservation was so poor. Two hundred fragments could not be identified; this component comprised many small fragments of bone, some of which had lost all of their surfaces and, in most cases, distinctive morphological characteristics were entirely absent. Small burnt bones, frequently white in colour were found throughout the deposits.


The vertebrate assemblage from this trench was recovered from six deposits, four of which were fills of Context 909, an oval pit which cut through a buried soil. Infilling of the pit was thought to have occurred by about 150 AD. The buried soil also produced a small collection of bones (Context 905=906). In total 156 fragments of animal bone were recovered from this trench, of which 24 were identifiable to species or family group (Table 3).

As seen from elsewhere on the site, the bone preservation was poor and the assemblage extremely fragmented. The material from Context 908, the fill from which many of the bones were recovered, did show slightly better preservation, although surface erosion and fresh breakage damage were still much in evidence. The bones were in such a poor condition that breakage on excavation was inevitable. Tooth enamel, which usually survives better in adverse conditions because of its higher mineral content, was also not particularly well preserved.

Identified remains from the earlier buried soil were restricted to two caprovid teeth. The degree of wear (attrition) of one of these (a mandibular M3) suggested the tooth came from an adult individual of approximately 3 to 4 years of age when it died (after Payne 1973; 1987).

Bones from the fills of the oval pit were dominated by the remains of cattle, in particular skeletal elements typically discarded during initial carcass preparation, e.g. horncore fragments, metapodials, calcaneii, astragalus and phalanges. There were also two scapula fragments but both of these were very eroded and poorly preserved. Given the poor preservation, it is likely that the remains are biased in favour of the smaller more robust fragments such as phalanges and calcaneii. All of the bones were from adult animals.

Pig remains from Context 908 were mainly teeth or mandible fragments and included a fragment of a canine from a male individual. The ?maxillary and mandibular third molars were unworn so may not have erupted (or at least were not in wear) and the isolated teeth appeared to be fragments of molar that may have been from the same tooth but which could not be reconstructed.


This trench formed the main intervention at the site and was an open area which encompassed some of the evaluation trenches (although not Trenches 1, 4 or 9). Features encountered in this trench included a clay floor of an industrial building, with a large circular pit (Context 1247) at its centre and a ditch (Context 1221) along its southern edge. Hand-collected bone was recovered from 27 deposits, seven of which were fills of the large pit and three represented deposits from within the ditch. The ditch fills produced just over 50% of the vertebrate remains from this trench. Most of the features probably dated to the first half of the second century, and had fallen out of use by 150 AD, at which time a substantial midden layer, Context 1217, was deposited in this area. A total of 683 fragments of bone were recovered, but of these only 64 were identifiable to species or family group.

Preservation of the remains was poor, and this was particularly noticeable for the fragments recovered from the fills of ditch 1221. A high degree of fragmentation was noted throughout, much of which resulted from fresh breakage damage; the bones were very fragile, however. Surface erosion was frequently observed and some of the bone surfaces had lost their outer ‘face’ and were splitting into layers. Fragments of tooth enamel were common and clearly were small fragments or broken into pieces, whilst burnt fragments, usually white in colour, were noted from many of the deposits. On the basis of the colour, these fragments have probably been subjected to high temperatures or have been exposed to prolonged heating.

Identified remains from this trench were restricted to the main domestic mammals, with cattle and pig remains being most numerous. A single small fragment of a dog maxilla was identified from Context 1228. A large quantity of bone (619 fragments) could not be identified other than to broad groupings and of these 226 fragments were wholly unidentifiable. The large mammal category included very many fragments of tooth enamel, most of which showed recent damage. These were probably cattle teeth but there was too much damage to make definitive identifications. Medium-sized mammal fragments were fewer in number and were mainly rib, shaft and vertebra fragments, although a few cranium fragments were also present.

The cattle bones included a range of skeletal elements with the ditch fills, in particular, dominated by head and terminal limb elements (metapodials, carpals, tarsals and phalanges). Remains of pigs showed a similar emphasis, whilst the few caprovid bones included several meat-bearing bones such as scapulae and pelves.

Age-at-death data were somewhat scarce. Caprovid remains included a mandible with some of the teeth in situ which was from an animal of approximately two years old, whilst an isolated third molar was from a slightly older individual aged three to four years when it was killed. Other isolated teeth were, for the most part, too fragmented to provide any information.


The vertebrate assemblage from this site was mainly concentrated in ditch and pit fills and deposits of early 2nd to early 3rd century date. Detailed analysis of the vertebrate remains was hindered by the poor preservation which resulted in there being few fragments that could be identified to species or that were able to provide useful biometrical and age-at-death information. The high proportion of tooth enamel fragments clearly highlighted a taphonomic bias in favour of certain skeletal elements, i.e. those which are more durable and survive better when preservational conditions are poor – teeth tend to be prevalent in such conditions because of their higher mineral content (in comparison to bone). Given the problems with preservation and limitations resulting from taphonomic factors, the apparent frequencies of different species represented and the identification of disposal patterns may not truly reflect the economic significance of the species or the original composition of the discarded remains. However, with these caveats in mind, some information was gleaned from the assemblage.

Identified vertebrate material consisted mainly of cattle, pig and caprovid remains; no bird bones or wild mammal species were noted. Taking the assemblage as a whole (it was too small to try to differentiate between date periods), the relative abundance of the different species suggested that cattle were prevalent (60%), with pig forming a quarter of the identified assemblage and caprovids being of less importance (12 %). This was supported to some extent by the unidentified component, almost 40% of which consisted of large mammal fragments, in comparison with the 5% of fragments representing medium-sized mammals. Horse and dog were the only other species to be identified but remains of these were few. From the little age-at-death data obtained, most of the animals represented were adult when they were slaughtered, although the presence of younger animals was hinted at for each of the three main species.

The higher proportion of cattle is noted at the adjacent Fairclough Homes site- 76% of the identifiable assemblage (see Viner in Williams and Reid 2008, 170). Sheep/goat and pig formed a much smaller proportion of the assemblage, while horse and dog were represented by single fragments.

At the community dig site skeletal element representation for cattle and pigs suggested that the bones probably represented primary butchery waste from the initial preparation of carcasses. Consumption refuse (represented by meat-bearing elements such as scapulae, radii, femora and pelves) was also present but only formed a minor component of the assemblage, with most of the few caprovid remains being in this category.

Much of the material was recovered from features which must have been convenient places for the disposal of rubbish, e.g. ditch 1221, pit 909 and tank 403. Material from Trench 1, associated with the road and its surface, may represent bones deliberately brought to the site to be used as hard core.

Despite the many limitations of this assemblage, it does seem to fit well with other vertebrate material of this date, particularly from similar sites which have close associations with military facilities. A predominance of cattle is characteristic of many Roman assemblages and shows a marked change from the dietary preference for lamb and mutton typically suggested by British Late Iron Age assemblages (Albarella 2007). However, this shift in emphasis did not occur everywhere, nor was it instantaneous following the Roman invasion, but, generally, the earliest occurrences of this trend have been noted from assemblages recorded from military sites and the civilian settlements which grew up around them (King 1978). It has been suggested that this change represents an adoption of ideas associated with the incoming Roman military population; the cultural and dietary traditions represented being more characteristic of the Low Countries and Germany from where most of the army based in Britain originated (Dobney 2001). Higher frequencies of pig remains have also been noted at early Roman military and urban sites (King 1978) and some researchers (King op. cit.; Dobney 2001) have suggested that higher proportions of pig remains may provide evidence of higher status occupation or could, perhaps, be related to occupancy by people of Mediterranean origin. Whether this is the case at Middlewich is unclear but the activities in the area do seem to have been closely linked to the military installation nearby.


Albarella, U. (2007). The end of the Sheep Age: people and animals in the Late Iron Age, pp. 389-402 in Haselgrove, C. and Moore, T. (eds). The Later Iron Age in Britain and beyond. Oxford: Oxbow Books.

Cappers, R. T. J., Bekker, R. and Jans J. E. A. (2006). Digitale Zadenatlas van Nederland. Gronigen Archaeological Studies 4. Gronigen: Barkhuis Publishing & Gronigen University Library.

Dobney, K. (2001). A place at the table: the role of vertebrate zooarchaeology within a Roman research agenda, pp. 36-46 in James, S. and Millett, M. (eds.), Britons and Romans: advancing an archaeological agenda. Council for British Archaeology Research Reports 125. York.

Dobney, K. and Rielly, K. (1988). A method for recording archaeological animal bones: the use of diagnostic zones. Circaea 5, 79-96.

Grant, A. (1982). The use of tooth wear as a guide to the age of domestic ungulates, pp. 91-108 in Wilson, B., Grigson, C. and Payne, S. (eds.), Ageing and sexing animal bones from archaeological sites. British Archaeological Reports, British Series 109. Oxford.

King, A. C. (1978). A comparative survey of bone assemblages from Roman sites in Britain. Bulletin of the Institute of Archaeology 15, 207-32.

O’Connor, T. P. (1989). Bones from Anglo-Scandinavian Levels at 16-22 Coppergate. The Archaeology of York 15 (3), 137-207 + plates VII-XI. London: Council for British Archaeology.

Payne, S. (1973). Kill-off patterns in sheep and goats: the mandibles from Asvan Kale. Anatolian Studies 23, 281-303.

Payne, S. (1987). Reference codes for the wear state in the mandibular cheek teeth of sheep and goats. Journal of Archaeological Science 14, 609-14.

Van der Veen, M. and O’Connor, T. (1998). The expansion of agricultural production in late Iron Age and Roman Britain, pp. 127-43 in Bayley, J. (ed.), Science in Archaeology: an agenda for the future. London: English Heritage.

von den Driesch, A. (1976). A guide to the measurement of animal bones from archaeological sites. Peabody Museum Bulletin 1. Cambridge Mass.: Harvard University.

Williams, M and Reid, M (2008). Salt: Life and Industry. Excavations at King Street, Middlewich, 2001-2002. BAR British Series 456