Boros István (szerk.): A Magyar Természettudományi Múzeum évkönyve 52. (Budapest 1960)

Thoma, A.: Anthropometric characters and selective survival

actual source of error is, however, implied by the circumstance that the data of the generalizing, and those of the individual, observation are represented by samples taken from two different populations. On the other hand, our method is based on the assumption that with both populations the changes exclusively caused by aging are of an identical direction and magnitude. In our opinion it is extremely improbable that in two populations of Homo sapiens europaeus leading a chiefly peasant way of life, the changes due to aging should follow divergent trends. Real differences may, however, materialize in the intensity of the changes. This is an error we are unable to measure and to eliminate. Yet, keeping in mind its potentiality, we render our conditions even more exacting, bv accepting as real the effect of the selection only if, according to our calculations, the variance due to selective survival is more than the double of that caused bv aging, viz. R. I. ^> 0,67 . Actually, the deviations between the age groups may be influenced also by a third factor, viz. the so-called secular changes, to which B ü c h i (1) attaches great importance. As matter of fact, these changes are so intensive in the case of stature, that — as numerous observations have shown — they mav bring about deviations even between consecutive generations. The author (5) has found also in a Hungarian material that within a population the subjects aged approximately 20 were higher than the older age groups. The secular changes of the head form are much slower. Even if the head measurements do show some regression upon stature, it is highly improbable that this relation should cause significant discrepancies between consecutive generations. On the other hand, speculative regressions, like P i t t a r sd's „law", are statistically unfounded and entirely subjective hypotheses. Our attempts to demonstrate the variational-allometric interrelations of the skull measurements have led to a negative result (unpublished observation, carried out with Professor Mala n). Results In the two series at issue the number of common metric characters was 18, and naturally it was only for these that calculations could be made. The results of the latter are shown in Table 1. For the variances of the individual observation we bring both the raw values and those corrected according to class size, whereas for the special cases of the presumable negative selection we give also the values adjusted as described in the preceding chapter. Only these transformed parameters have been used in calculating the intensity. As a matter of course, our results apply only to the 20—74 year period. Concerning positive selection, the coefficients calculated according to (3) have no sign, whereas in a mixed selection they have a ^ sign. According to (4) the cases of negative selection feature a — sign. Both in the Swiss and in the Irish material stature increases moderately up to the mid-30's, and then decreases continuously. The values of the span show within the whole interval, and in both curves, a first slower and later more rapid decrease. The relative span is on the whole constant, although it shows a statistically not everywhere significant diminishing tendency. If selective survival does influence the age variations of these characters, it can only be a positive selection, since the generalizing variances are everywhere greater than those obtained by the individual method. The coefficients of relative intensity remain, however, inferior to the implicitly accepted level of significante.