By using a number of different alcohols as substrates, eight alcohol dehydrogenase loci were discovered in Drosophila pseudoobscura. Each of these loci can take more than one substrate. Several of these loci differed in their tissue specificities and activity patterns during development. The genic variation in natural populations was studied at four of these loci and three of them were polymorphic. A quantitative study of substrate-specific differences among alleles of the same locus produced negative results. This result appears to be typical of most studies done on this aspect. From this it was concluded that the substrate specificity for enzymes is not an important factor in determining the greater amount of genic variation at Group II loci than at Group I loci, as proposed by Kojima, Gillespie and Tobari (1970). There are several observations which suggest a different explanation for the differences in the genic variability at Group I and Group II loci: (1) There are, on an average, more isozyme loci (loci with similar substrate specificity) for enzymes in Group II than in Group I; (2) The null alleles are far more common at Group II loci than at Group I loci; (3) There is significant heterogeneity in the number of alleles and the heterozygosities at loci within each of these two groups of enzymes; (4) Relatively higher levels of genic variation are observed at Group II loci even in populations which appear to be living in homogeneous environments; and (5) Some loci (e.g. esterases) are highly polymorphic in most species investigated by gel electrophoresis techniques. Based on these general observations, it is proposed that (1) the substrate-specific differences are between isozyme loci and not between alleles of a given locus, and (2) neutral alleles are proportionately far more common at loci at Group II than at loci in Group I, because the former is under less selection constraint than the latter.