SUMMARYGenetic divergence between two forms of Ostrinia nubilalis, which were characterized by a different composition of the sex pheromone, was studied on the basis of 30 gene-enzyme systems. Nei's measure of standard genetic distance between the forms was D = 0024, strikingly different from the value of D =0002 previously obtained by Harrison and Vawter on 10 enzyme loci between populations genetically homogeneous with those presently studied. In order to examine the discrepancy, and to test Sarich's hypothesis of a correlation between metabolic role of enzymes and their rate of evolution, the sample of 30 loci was subdivided into two classes: 14 loci coding for variable-substrate and regulatory (sensu Johnson, 1974) enzymes on the one hand and 16 loci coding for nonregulatory enzymes on the other. The mean values of genetic distance calculated separately for the two classes of enzymes were significantly different: D1 = 0056 and D, =0004 respectively. The ratio D1/D, is 14, a value of the same magnitude as found in various taxa at the beginning of their process of divergence, as indicated by recent data on both vertebrates and invertebrates.The data confirm Sarich's hypothesis of a bimodality in the rate of evolution of different classes of enzymes, and explain the difference between our value of the mean genetic distance and that obtained by Harrison and Vawter, who studied mainly non-regulatory enzymes. They confirm the importance of the choice of loci in evaluating genetic divergence. Mean values of genetic distance may be meaningless if they do not take into account the proportions of loci with different metabolic roles.