These techniques are used to analyze extremity long bone measurement data in 11 populations. Interpopulation variation of sexual dimorphism has often been employed to interpret both the interpopulation variation of coefficients of sex discriminant function and the decrease of correct diagnosis rate in cross-population application. Statistical examination shows, however, that many such phenomena are attributable to sampling errors. Although significant variation of sexual dimorphism certainly exists even among subpopulations within the modern Japanese, this does not prohibit use of the formally derived common discriminant function. Its discriminant efficiency varies among populations, but compares well with that of each population-specific function. The interpopulation variation of sexual dimorphism is related mainly to the extent of femoral pilastry, while the common function reflects primarily general "size" and secondly femoral pilastricity and relative epiphysial size. Further study of such functions may reveal the common elements in sexual dimorphism of mankind. 12 tables, 41 references. (Publisher abstract)