Ilon Gábor szerk.: Pápai Múzeumi Értesítő 6. (Pápa, 1996)

Bronzkor a Nyugat-Dunántúlon - L. Bartosiewicz: Bronze age animal keeping in northwestern Transdanubia, Hungary. Bronzkori állattartás az Északnyugat-Dunántúlon

animal remains could be used in faunal studies.

3 This number in itself is indicative of a relatively small degree of fragmentation, a taphonomic phenomenon that by definition reduces the possibilities of identification. It is noteworthy, however, that the large sample from the site of Börcs-Paphomlok-dülő was one of the least well preserved (74% identifiable). Many of the highly fragmented bone splinters came to light from features such as post-holes. On the other hand, bones found in larger features such as the Bronze Age house at the site of Abda - Hármasok (95% identifiable) are better preserved and therefore remained in a more easily identifiable state. These end points illustrate the two extremes of deposition that arc represented by the data set available for study. Bronze Age meat consumption at the sites under discussion here was characterized in terms of raw fragment counts, that is the number of identifiable bone specimens (NISP, Table 1). Although, especially in small samples, this simple way of counting bones can be a source of bias, it is still the most easily performed and rcconstructable form of presenting archaeozoological data. While the minimum number of individuals (MNI), another widely used parameter, could be relatively accurately estimated in smaller assemblages that include material from only a few features, the calculation of this value in the material from large, complex sites may become misleading: the number of individuals becomes increasingly meaningless in practical terms when long but poorly defined time intervals arc involved. On the other hand, aggregation effects, that is the multiple counting of bones from the same individual whose various body parts ..were enterred in several features, may distort the results. Consequently, as a result of sfrongly Varying sample sizes, MNI estimates may have been biased in different ways. The relationshipobcrtwecn NISP and MNI values also differs depending on the type of animal exploitation (hunting v.s. animal keeping) pursued at individual settlements.

4 jbjWI In light of the aforementioned broad variability of sampje.size^j the faunal composition of individual sites would have been difficult to compare. Taxonomic richness, that is the number of animal species identified, increases parallel with sample size because the more bones are available for study, the greater the statistical probability that remains of even relatively rare animals will occur in the sample. The relationship between the number of identifiable specimens and the number of species recognized, however, is not linear. Beyond a certain sample size it is increasingly difficult to find additional species even in large faunal assemblages. When decimal logarithms of R are plotted against those of NISP in the eleven assemblages available for study the following linear regression equation is obtained: IgR = 0.244 + 0.2771gNISP (r = 0.857***) The high, statistically significant correlation and the low regression coefficient of IgNISP are indicative of a strongly degressive tendency, which mean that even in the case of strongly increasing sample sizes it is unlikely that the number of newly occurring species would become significantly greater. This trend, which may be visually appraised in Figure 1, is usually characteristic of settlement materials tíiat show a heavy reliance of domestic animals in the exploitation for meat. The taxonomic composition of faunal assemblages usually varies depending of the degree and specialization of hunting. In the material under discussion here, it may be considered a normal phenomenon that rare wild animal species (e. g. red fox, brown hare and especially roe deer) were more likely to occur in relatively large assemblages. This order of magnitude corresponds to data published by Vörös 1978. 71.; Vörös 1979. 137. Bartoslewlcz, L.: Species interferences and the interpretation of Neolithic animal exploitation. AAH 42 (1990) 287-292. Grayson, D. K.: Quantitative zooarchaeology. New York, 1984. Bartoslewlcz, L.: Archaeozoological studies from the Hahót Basin, SW Hungary. In B. Szőke ed.: Archaeology and settlement history in the Hahót Basin, South-West Hungary. Antaeus 22 (1995) 307-367.