An argument for the use of total side frequencies of bilateral nonmetric skeletal traits in population distance analysis: the regression of symmetry on incidence.

In population studies based on frequencies of bilateral nonmetric skeletal traits, the choice between sampling by individuals or by sides should depend less on the exigencies imposed by fragmentary remains than on fundamental assumptions about the biological meaning of symmetry/asymmetry. Though the latter has been interpreted in various ways, little attention has focused on the possibility that bilateral correlation is meaningful in quantifying genetic liability for a trait. Analysis of two independent mandibular features, mylohyoid bridge and suppressed third molar, in Indian and Eskimo population samples (total N congruent to 1,200) reveals a statistically significant pattern of increasing bilateral occurrence with increasing population incidence. This pattern is consistent with the theory that liability for a "quasi-continuous" variant is normally distributed with constant increment between thresholds on an underlying scale. According to theory, phenotypes with more pronounced expression (bilateral occurrence) have greater genetic potential than those with less pronounced expression (unilateral occurrence). Therefore, scoring traits in total left and right sides, by giving greater weight to bilaterally affected individuals, may provide a better estimate of the liability for the trait in the population. Viewed in a theoretical context broader than that of the sampling debate, this pattern of positive regression of symmetry on incidence means that prevalence of unilateral occurrence probably cannot be used to assess the relative strength of genetic versus nongenetic control of threshold variants.