M. Járó - L. Költő szerk.: Archaeometrical research in Hungary (Budapest, 1988)

Analysis - BARTOSIEWICZ László: Water-sieving experiment at örménykút, site 54

number of effects which disturb the quantification of recovery data. In this experiment, more standardized parameters may be attributed to the following features of the material: 1. Bone, the only kind of find in the deposit of the butchering site, occurred in the form fragments and splinters of varying size but similar colour thus providing a highly homogeneous material for the purposes of comparison. 2. Because the butchering site was of limited extent, complete recovery and data gathering could be carried out during a workday under more or less constant light conditions (daylight). 3. Hand collecting and water sieving were carried out by the same person. Thus, aside from the periodicity in the daily cycle of individual concentration, it is not a variety of subjective elements which may have biased the compilation of the data set. Theoretically, under these circumstances it was only the individual size of each bone splinter which served as the primary variable in addition to the finite resolution of human attention. The variability and distribution of fragment size could thus be compared between the discussed techniques of recovery which represented grouping variables. Following the retrieval of large animal bones from the deposit, some 1.8 m

3 of the backdirt was removed for the purposes of the experiment. First, the remaining pieces of bone were picked up by hand. In the second step the material was washed through two screens of different mesh sizes (25 mm and 0.8 mm respectively). This procedure was preceded by the drying of material which was subsequently soaked prior to the actual sieving. Measurement of the retrieved fragments was carried out using a graph paper. The greatest length of each splinter was recorded. These data were analysed by standard statistical procedures. Results The size distribution of splinters is shown in the frequency polygon of Fig. 1. As may be seen in this picture, the refinement of recovery techniques resulted in increasingly skewed distributions. In each case, the gradual increase in the number of fragments drops suddenly following a peak corresponding to the smallest size category when 0.8 mm mesh size water-sieving is applied. Smaller size categories represent a greater investment of labour, thus they may be perceived in terms of energy expenditure, which increases as the refinement technique is improved. Consequently, the efficiency of such expenditure may be estimated on the basis of size distributions shown in Fig. 1. For the purpose of this calculation, sub-sample frequencies within the sets of bone splinters obtained by hand collection and water sieving (25 and 0.8 mm mesh size respectively) were arranged in size categories between 65 to 05 cm. There was a 05 cm constant size interval between these categories (Table 1 ). Proportions calculated between the number of fragments in each size category and the increase in these numbers per decrease in size (Oi cm) were interpreted as mean efficiency of the refinement (y/x) and efficiency by increasing refinement (Ay/Ax) respectively. When these values obtained for the three different methods of recovery are compared different tendencies occur. While, with the exception of hand collection, the mean efficiency of refinement (y/x) increases steadily, the sudden drop in the efficiency is manifested by negative values (Ay/Ax). These mark the limits beyond which the reliability of a particular method dramatically declines. Natorally, this threshold corresponds to the largest, 2 cm size category when only hand collection is practiced. Even