On the assumption of equality of variance-covariance matrices in the sex and racial diagnosis of human skulls

Some multivariate tests for differences in sexual dimorphism between human populations

Two multivariate statistical tests are presented for examining differences in sexual dimorphism between human populations. One test refers to general differences in sexual dimorphism; the other, only to differences in size aspects of sexual dimorphism. There is no current resolution to the complementary problem of testing for differences of direction in sexual dimorphism. By applying the tests to a comparison of sexual dimorphism in 26 samples of skulls from a world-wide distribution of human populations, significant results (at alpha = 0.05) were found in the majority of cases with the general test, and in a minority of cases with the size test.

Migration models of prehistoric postmarital residence

It has previously been suggested that residential practices can be inferred from within- or between-group analysis of male and female skeletal morphology. Arguments have proceeded from intuitively derived hypotheses about the genetic consequences of drift and migration. In this paper, a formal basis is presented for these hypotheses using modified versions of Wright's island model and the migration matrix method. It is shown that the usual measures of standardized genetic variance or genetic kinship can be decomposed into male, female, and male/female components. The male and female components can in turn be used separately to assess the effect of different residential practices on the population genetic structure of the two sexes. Following upon these models, nonmetric cranial trait data from prehistoric sites in west-central Illinois are used to delineate the probable residential practices for this region.

An analysis of crania from Tell-Duweir using multiple discriminant functions

Historical and archaelogical evidence suggests that the Iron Age biblical city of Lachish had a multinational population of diverse geographical origins. A multivariate analysis of crania, using canonical discriminant functions and metric variables, tends to confirm this. The approach employed stresses that population discriminant analysis studies should be both biologically and statistically legitimate. An ecological interpretation of the data suggests a research design for analyzing the affinities of cranial series. Similarities probably should be assessed in an analytical space containing the widest possible range of variation.