Comparison of the mean FST values of the two types of polymorphisms represents a modification of the Lewontin-Krakauer (L-K) test for neutrality (LEWONTIN and KRAKAUER 1973). As originally formulated, this test in- volved comparing the allele frequency variation at one set of loci (allozymes) to an expected theoretical vari- ance. Application of the L-K test to spatial variation was criticized on the basis that special patterns of migration or mutation among subsets of the populations can lead to inflated variance ratios (NEI and MARUYAMA 1975; ROBERTSON 1975). By comparing allozyme and FWLP loci in samples from similar geographic regions, how- ever, one eliminates the confounding effects of un- known population relationships. The L-K test statistic is highly significant (x2 = 24.6, d.f. = 5, P < 0.001), indicating that as a group the 10 allozyme loci are more homogeneous than expected from the variation exhib- ited by the RFLP loci. Removal of the most variable RFLP locus (GM798) does not alter this conclusion (x2 = 13.1, d.f. = 5, P= 0.023).