This paper discusses the animal bones recovered during excavation at two Roman villas on the Via Gabina, near Rome. The excavations were directed by Prof. Philip Oliver-Smith and Prof. Walter Widrig, of Rice University, Texas (U.S.A.), between 1976 and 1989. The sites take their 'names' from the numbers assigned during survey by Ward-Perkins and Kahane (1972) (Via Gabina site 10, Via Gabina site 11. The animal bones were studied in July 1990, primarily at The British School at Rome. The data from the two sites will be presented separately, although the results will be considered together, and placed within their broader context. All bones were recorded individually and assigned a reference number which may be given, in places, in the text.

Via Gabina, Site 10 Villa

A sample of 4,495 fragments was examined from this site, excavated between 1978 and 1989 (Table 1) It proved impossible to identify to species or bone 1519 (33.8%) of these. Vertebrae form 6.6% of the sample: the majority is from very small and small animals (108 and 150 fragments respectively). All parts of the vertebral column are present. Ribs represent 14.2% of the fragments recovered: those of small animals are most frequent. There are 2,042 fragments (45.4%) which are identifiable to species or group of animals. For fifty-eight fragments, the definitive species identification is still to be established. Nine fragments of fish bone and five pieces of turtle/tortoise carapace and plastron were isolated. There are 150 fragments of rodents and 213 fragments of bird.

The bones of wild animals are present in the sample although they are rather infrequent. There are three bones of roe deer (a metatarsal, a proximal phalanx and a medial phalanx)) two of badger (a lower tooth and a tibia) and forty-nine of hare. For badger and roe deer just one individual of each may be represented, whilst the hare remains belong to at least four individuals. A single bone (a humerus) was found of a cat. A discrete number of equid bones was found (48 fragments), representing at least three individuals. Fragments of canid bones are more frequent than those of cattle, although in both cases a minimum number of seven individuals is represented. The dominant animals are sheep/goats and pigs: there are 379 fragments of sheep/goat bones belonging to at least twelve individuals, and 706 fragments of pig bones belonging to at least thirteen individuals.

In Table 2 the composition of the main species identified sample is shown in terms of the frequency of representation of the different anatomical elements. For pigs there is a quite considerable number of cranial and dental elements (49.2% of the pig sample). The upper parts of the limbs (that is, excluding the carpals/tarsals and below) are reasonably well represented (36.7%) whilst the bones of the limb extremities are rather scarce especially given their numerical frequency in the skeleton (13.3%). For sheep/goat the fragments are more evenly distributed, with the bones of the upper parts of the limbs being most frequent (39.8%). In fact cranial and dental elements are, as a group of elements, least frequent (24.5%). Similarly for cattle the fragments are more evenly distributed, with fragments of the bones of the limb extremities being slightly more frequent than the other categories (35.9% cf. 29.8%). A large number of equid teeth was found, although the limbs are also represented. The canid remains present cover the whole skeleton, no part being particularly dominant (given differences in the frequency of each elements in the complete skeleton). The hare bones are primarily from the upper part of the fore limb and from the hind limb.

Information upon the ages at which animals of the main species had died is summarized in Table 3. For pig there appears to be a discrepancy between the results based upon the fusion data and those based upon the dentition evidence. The fusion data imply that a considerable number of animals had died before reaching 11-19 months, that almost all had died before reaching 23-31 months, with very few animals exceeding 31-35++ months. However, from the dental data it appears that a reasonable number of animals exceeded (even if by little) 19-23 months, the remaining animals being killed at a range of ages (a hypothesis based upon the number of jaws with teeth erupting and the variety of toothwear stages represented). However, both sets of data allow it to be suggested that most pigs had died before reaching 31 months of age. In addition, it should be noted that thirteen bones are of foetal animals and 104 are from animals which had died soon after birth. Whilst some sheep/goats died whilst very young (for example, four bones of foetal individuals, twenty-one of neonatals), most survived until at least 48 months. At least one animal exceeded 4860 months at death. Although one cow/ox died soon after birth and another prior to 1.5 years, most survived beyond 2-2.5 years. Some quite old individuals are indicated by the toothwear evidence in particular. Seven of the canid bones are from neonatal individuals, and a range of immature and mature animals seems to be represented. On the basis of the horse teeth, most of the animals died at circa eight to ten years of age, although one slightly older animal (fourteen years) was noted and one animal is represented by a particularly worn upper third premolar.

Butchery markings were quite frequent and are listed in Appendix 2. it should be noted that evidence was found for deliberate cutting on the bones of pigs, sheep/goats, cattle, equids and a canid, as well as rib, vertebra and unidentifiable fragments. It seems that the carcase was split in two along the spinal column. Particularly common division points were the distal end of the scapula; the distal humerus - proximal radius - proximal/mid ulna; the pelvis; the femur; the distal tibia and tarsals; and the metapodials.

In Appendix 1 (part i) are listed the metric data obtained from this sample.

Where possible distinction has been made between sheep and goats, using primarily criteria proposed by Boessneck (1969). The following results were obtained: horncore - sheep three, goat one; cranium - sheep one; metacarpal - sheep four, goat one; metatarsal - sheep six; astragalus - sheep one; proximal phalanx - sheep fifteen, goat two; medial phalanx - sheep seven, goat one; distal phalanx - sheep three. To summarize, of the forty-five diagnostic fragments, forty (88.9%) are of sheep.

There are sixty-six fragments of canines of pigs: forty-three are of males and twenty-three are of females. Five fragments of bone had pathological conditions: cattle femur (389), sheep/goat metacarpal (992), pig maxilla (3715), pig mandible (3729) and a canid metacarpal IV (4428).

Weathered surfaces were noted on 113 fragments. In most cases this was slight or moderate although some instances of heavy weathering were found. Forty-four fragments, from a variety of contexts, had been burnt. One fragment had been gnawed (4091).

Via Gabina, Site 11 Villa

A small sample of animal bones was recovered from excavations which took place between 1976 and 1980. Only 262 fragments were found (Table 4). Of these, 139 (53.1%) could not be identified to species or to bone type. There is one fragment of a thoracic vertebra of a sheep/goat-pig size animal, and twenty fragments of rib of a similarly sized animal. It was possible to identify 128 fragments (48.9%) to a group of animals or to species. The precise species identification of one fragment is still to be confirmed. Twelve fragments have been identified as being of rodent; four are tortoise/turtle carapace and plastron fragments; one is of fish; fourteen are of birds; and two shells were noted. The majority of the sample is composed of bones of "domestic species", that is, pig, sheep/goat, cattle, equid and canid. The relationship between these animals is shown in Table 4 both in terms of the number of fragments identified and the minimum number of individuals represented. Canid and cattle bones were infrequent. The sheep/goat and equid samples consist of a similar number of fragments from the same minimum number of individuals. The small site 11 sample is dominated by the bones of pigs.

Obviously the small size of the sample precludes the representation of all parts of the skeleton and the emergence of valid patterns. However, in Table 5 the predominance of pig and equid teeth should be noted. The data available upon the ages at which the major 'economic' animals had died are given in Table 6. For pig the sequences and estimated ages used are those proposed by Bull and Payne (1982); for sheep/goat those of Bullock and Rackham (1982); and for cattle those of Silver (1969). The toothwear has been recorded using the methodology proposed by Grant (1975, 1982). The single cow/ox definitely present had died after having reached 1.5-2.5 years. Whilst one sheep/goat had died prior to reaching 35-36 months, another survived beyond 48 months. The majority of pigs had died prior to reaching 23-31 months, although one survived beyond this. Of the two equids definitely present, on the basis of the crown height of the teeth (Levine 1982), one had died at circa 4 years whilst the other had died at circa 8~9 years.

Few measurements could be taken of the bones from site 11. Those that were possible are given in Appendix 1 (part ii). Of the pig canines, nine are of males and four are of females. Four fragments of bone had cut marks related, in all probability, to butchery: a rib of a small animal has a cut edge at one end (008); a rib of a small animal has a cut marks and a cut edge (then snapped) at one end (014); a rib of a small animal has a cut and snapped edge at the ventral end (127); a sheep/goat humerus has a sawn cut edge at the distal end of a split proximal shaft fragment. A total of twenty fragments had weathered surfaces of varying degrees of severity: slight (nine), moderate (seven), moderate-heavy (three), heavy (one). Part of one rib fragment of a small animal had been burnt (129). A thoracic vertebra of a small animal has a malformed spine (067).

Discussion

The data resulting from analysis of sample of animal bones from two villas (sites G10 and Gl1) on the Via Gabina have been presented. Before going on to interpret and discuss the evidence, certain critical points must be made. The samples are not particularly large, especially given the extensive areas excavated. However, it should be remembered that these sites are villas and, although little is known on refuse disposal practices, it might be reasonable to assume that the bulk of household refuse was removed from the living area. The rubbish might, for example, have been gathered together at some distance from the villa itself, or burnt, or scattered on cultivable ground. Therefore, it may be assumed with certainty that the animal bones recovered during excavation do not represent anything other than a part, perhaps very small, of the animal carcases (whether complete or partial) once associated with the villas.

In addition, it should be appreciated that a sample of animal bones will rarely reflect upon an economic system in its entirety, but rather upon aspects of it. For example, bones of sheep/goats, pigs and cattle may reflect directly upon the food consumption of the associated human group, but may not be at all representative of the agricultural production system of the surrounding area. Moreover, it should be recalled that the bones found are not always a result of meat consumption but may belong to animals kept as pets or for work, or may be representative of a raw material source.

In the case of the villa at Site 11, pig is the dominant animal, followed by sheep/goat and then equid. However, that we are dealing with a rather particular sample is indicated by the marked predominance of pig and equid teeth. Two main situations might be envisaged: firstly, that this is a result of a deliberate acquisition/disposal practice; secondly, that teeth have survived in greater numbers because of their more resistant physical structure. At present it is not realistic to prefer one hypothesis to the other, and the two are not mutually exclusive in fact. That at least some of the material is food refuse is implied by the presence of cut marks related, in all probability, to butchery.

Obviously there is the potential for obtaining rather more information from the site 10 sample. An initial impression is that this sample relates much more to food consumption than that from site 11, given the dominance of pig, sheep/goat and cattle bones. The major exception to this is the large number of canid bones: although one bone has a cut mark, it is likely that these animals had been working dogs or household pets. The equids present may also have been kept as working animals, whether for use in agricultural activities or in transport, although there is no reason why they should not ultimately have formed part of the meat supply. The occasional exploitation of wild animals is indicated by the presence of hare and roe deer remains.

The critical elements in diet were pork, beef and lamb/mutton. Although numerically least common, cattle may have provided a considerable portion of the meat supply (circa 49% on the basis of the number of fragments and circa 63% on the basis of the minimum number of individuals). Given the fact that many of the cattle present appear to have died after having reached maturity, it would appear that they had been raised primarily for milk and/or as working animals, although ultimately forming part of the meat supply. The contribution of sheep/goats to the meat supply appears to have been relatively slight (circa 12% on the basis of the number of fragments and circa 13% on the basis of the minimum number of individuals). However, many of the sheep/goat had in all probability been raised for milk and/or wool and only culled once their yields had begun to decline. Pigs, in addition to being the most frequently exploited animal, would have been of fundamental importance in the food supply. As mentioned above, there seems to be some discrepancies in the mortality data as a result of which it is difficult to comment upon the implied pig-raising regime with any certainty. Suffice it to say that it appears to have been at least relatively intensive.

In Appendix 1 measurements of some bones have been listed. Briefly these will be compared with those obtained for Settefinestre (King 1985) and Matrice (Clark forthcoming). The sheep/goat at the Via Gabina are rather large animals: those from Settefinestre are in the lower part of the Via Gabina ranges; those for Matrice spread over ranges which start before those for the Via Gabina and finish before. The cattle at the Via Gabina are similar in size to those at Matrice and comparable to the larger animals at Settefinestre. The pigs at Via Gabina may have been slightly smaller than those at Matrice, but broadly comparable to those at Settefinestre.

King (1984, 1985) has discussed the data available from animal bone studies in the Late Republic and Empire in west central Italy. He concludes that the relationship between the species of 60% pig, 25% sheep/goat and 15% cattle is distinctive of the dietary regime prevailing in this area at that time. The Via Gabina data fit this pattern remarkably well, being pigs 56%, sheep/goat 30% and cattle 14%). In terms of the sheep/goat raising strategy, the patterns observed in the mortality data are rather similar to those noted at Settefinestre and dissimilar to those at Matrice. The exploitation of pigs seems to have been similar at all the sites being discussed here. The exploitation of cattle at Via Gabina appears to be more similar to that at Matrice then that at Settefinestre.

To conclude, whilst not large in numerical terms, the samples obtained from the villa sites G10 and G11 on the Via Gabina, just south of Rome, have added valuable shading to our picture of life on and around such sites. As was emphasized before, no group of material is likely to provide a comprehensive image, but particular parts of it may be defined more clearly. As a result, this paper need not represent anything other than a first stage in the analysis of these data, as attention in the future may be paid, for example, to chronological developments and spatial distribution patterns.

Gillian Clark
British School at Rome
November 1990

References

• Boessneck, J. 1969: Osteological differences between sheep (Ovis aries Linne) and goats (Capra hircus Linne). In D. Brothwell and E.S. Higgs (eds), Science in Archaeology (second edition): 331-358. London: Thames and Hudson.
• Bull, G. and Payne, S. 1982: Tooth eruption and epiphysial fusion in pigs and wild boar. In B. Wilson, C. Grigson and S. Payne (eds), Ageing and Sexing Animal Bones from Archaeological Sites: 55-71. Oxford: British Archaeological Reports, British Series 109.
• Bullock, D. and Rackham, J. 1982: Epiphysial fusion and tooth eruption of feral goats from Moffatdale, Dumfries and Galloway, Scotland. In B. Wilson, C. Grigson and S. Payne (eds), Ageing and Sexing Animal Bones from Archaeological Sites: 73-80. Oxford: British Archaeological Reports, British Series 109.
• Clark, G. forthcoming: The animal bones from Matrice. In J. Lloyd (ed.), The Samnite and Roman Villa at Matrice (CB). Archaeological Monograph of the British School at Rome.
• von den Driesch, A. 1976: A Guide to the Measurement of Animal Bones from Archaeological Sites. Cambridge (Mass.): Peabody Museum Bulletin 1.
• Grant, A. 1975: The use of tooth wear as a guide to the age of domestic animals - a brief explanation. In B. Cunliffe, Excavations at Portchester Castle Vol. I. Roman: 437-450. London: Report of the Research Committee of the Society of Antiquaries of London 32.
• Grant, A. 1982: The use of tooth wear as a guide to the age of domestic ungulates. In B. Wilson, C. Grigson and S. Payne (eds), Ageing and Sexing Animal Bones from Archaeological Sites: 91-108. Oxford: British Archaeological Reports, British Series 109.
• King, A. 1984: Roman diet - the evidence of animal bones from Settefinestre and other Italian sites. Paper presented to the Third Conference of Italian Archaeology, Cambridge.
• King, A. 1985: 1 resti animali; i mammiferi, i rettili e gli anfibi. In A. Carandini (ed.), Settefinestre. Una Villa Schiavistica nell'Etruria Romana. Vol. 111: 278-300. Modena: Edizioni Panini.
• Levine, M. 1982: The use of crown height measurements and eruption-wear sequences to age horse teeth. In B. Wilson, C. Grigson and S. Payne(eds), Ageing and Sexing Animal Bones from Archaeological Sites: 223-250. Oxford: British Archaeological Reports, British Series 109.
• Silver, 1. 1969: The ageing of domestic animals. In D. Brothwell and E.S. Higgs (eds), Science in Archaeology (second edition): 283-302. London: Thames and Hudson.
• Ward-Perkins, J. and Kahane, A. 1972: The Via Gabina. Papers of the British School at Rome 40: 91-126.

Table 1. Via Gabina Site 10: number of fragments and minimum number of individuals represented

Table 2. Via Gabina Site 10: anatomical composition of the sample

Table 3. Via Gabina site 10: mortality data (For key see below)

Key:

Table 4. Via Gabina Site 11: number of fragments and minimum number of individuals represented

Table.5. Via Gabina Site 11: anatomical composition of the sample

Table 6. Via Gabina Site 11: mortality data (For key see below)

APPENDIX 1. Via Gabina metric data

The measurements are based on those proposed by von den Driesch (1976). All measurements are given in millimetres.