QUANTITATIVE GENETICS - A CRITIQUE

The quantitative aspects of human genetics have hitherto been confined to Mendelian categorization of qualities, which is enumerative and best applied to single-locus conditions; and Galtonian analysis, which is an examination of the first and second moments of quantities and best applied to multiple-locus conditions. The latter analysis throws no light on the details of the genotype or how they shape the phenotype; and even as a descriptive device it is adequate only where narrowly restrictive assumptions (equality, additivity, and independence of the component contributions) are met. The utility of such general models as approximate describers, provided that they are tempered by explicit empirical inquiry, is illustrated by familiar analogies. Several examples are discussed where neither Mendelian nor Galtonian analysis is appropriate. Many mechanisms do not even fulfill the assumption of monotonicity. Heritability in the conventional sense is briefly discussed.