Janus Pannonius Múzeum Évkönyve 37 (1992) (Pécs, 1993)

Régészet - Finnegan, Michael–Szalay, Ferenc: Population Distance Between Late Roman Period to 11th Century Arpadiam Age Populations as Determined by Non-metric Trait Analysis

100 Age dependency We have not tested for age dependency in this sample inasmuch as rarely was cranium less than 18 years of age used in the sample, and independent age criteria was not available based on non-cranial material. As well, previous research has shown that while correlation between trait incidence and age may exist, the dependency is not strong when tested by chi square analysis between youngest and oldest age samples within the same population (FINNEGAN 1978). Even those who P-Ist Population Separation Trait frequencies from Table 2 and their corresponding sample size (N) taken from Table 3 were utilized in the Grewal-Smith statistic as presented by FINNEGAN (1972) and FINNEGAN and MARCSIK (1979) with the results presented in Table 7. This table shows the biological distance or mean measure of divergence (MMD) as computed, usisng the Grewal-Smith transformation withouth any corrections for frequency or sample size (for alternative statistics see FINNEGAN and COOPRIDER 1978). In Table 7 the numbers represent the mean measure of divergence between each sample pair. The underwritten italics figures represent estimates of the variance. All between sample distance measures are significant at the .05 level with one third of the distances suggest there is some age dependency do not independently analyze earlier and later age groups (see, for example, OSSENBERG 1970; BUIKSTRA 1972; KOREY 1970 and CORRUCCINI 1974). In the current analysis, we have not been concerned about age groups because almost all individuals used in the samples are adult and those who are not, are very nearly adult and we did not have independent age profiles for males and/or females. We also conclude that age dependency is not so great a problem that we have to divide our samples before analysis. P-Szé Ellen Nagyp Kékes Majs significant at the .01 level. The significance figure is important as it shows that each of these samples is derived from different exclusive biological populations. From Table 7, the MMD's in matrix form were committed to a numerical taxonometric statistic in order to reduce the seven dimensional array to a two dimensional phenogram plot (SOKAL and SNEATH 1963). This was accomplished by using the statistical programs of ROHLF et al. (1974), subjecting the MMD distances to a TAXON analysis which is a sequential agglomerative hierarchiai cluster analysis. In this analysis we specified the unweighted pair group method using arithmetic averages with lowest values considered for similarity. One output of this particular program produces a phenogram which is presented in Figure 1. A second program MXCOMP (ROHLF et al. 1974) was run which compares Table 7. Measures of divergence (biological distance) between population samples used in this study. Underwritten figures in italics are estimates of the variance. All distance measures are significant (+=p<0.05) and some are very significant (*=p<0.01). P-Ist P-Szé Ellen Nagyp Kékes Majs Pécs-Székesfehérvár u. Eilend Nagypall Kékesd Majs Zengővárkony 0.042+ 0.010 0.095* 0.033+ 0.007 0.007 0.095* 0.036+ 0.045+ 0.012 0.011 0.009 0.087* 0.048+ 0.029+ 0.041+ 0.010 0.009 0.007 0.012 0.049* 0.043* 0.021+ 0.047+ 0.036* 0.007 0.006 0.004 0.008 0.006 0.116* 0.060+ 0.053+ 0.070+ 0.060+ 0.068+ 0.016 0.015 0.013 0.017 0.015 0.012