Gene arrangement frequencies were determined at two stages in the life history of Drosophila pseudoobscura taken from nature. Three po ulations in the central highlands of Mexico were each sampled twice during 1976.Gene arrangement frequencies were measured in adult males and in larvae that were the offspring of females collected at the same time. The adult males were in all likelihood a representative sample of those who fathered the larvae produced by the wild females. Differences in gene arrangement frequency between these two life stages should indicate the operation of natural selection. One-third of our comparisons of common gene arrangement frequencies in males and in larvae from the next generation were statistically significant, as were one-third of our comparisons of total frequency arrays in the two life stages. We consider the components of selection that could produce such frequency changes and reason that male mating success must be the major one. Gene arrangement frequencies in the Mexican populations fluctuate within wide bounds. Selection must act to retain the polymorphism in the face of this flux in gene arrangement frequencies, and we suggest that male mating success plays an important role. One of the early triumphs of ecological genetics was the demonstration that selection in nature could be intense-in fact, one or several orders of magnitude more powerful than the founders of population genetics imagined. No experimental system played a more important role in the analysis of selection than the chromosomal polymorphism for gene arrangements in Drosophila pseudoobscura. In this species a series of inversions on the third chromosome binds large blocks of genes together as units, just as though they were alleles of a single "supergene." Natural selection was first implicated when Dobzhansky (2) showed that the frequencies of certain gene arrangements went through seasonal cycles in two of three populations on Mt. San Jacinto in California; subsequent studies showed that these cycles were repeated in years scattered over a span of 2 decades (3, 4). The frequencies of gene arrangements in the third population on Mt. San Jacinto did not cycle, but between 1939 and 1946 they underwent a directional change that Dobzhansky (5) also ascribed to natural selection. Dobzhansky and Levene (6) then showed that karyotypic frequencies in eggs laid by wild females were generally in accord with Hardy-Weinberg expectations, but that frequencies in wild males were not. They concluded that the karyotypes suffered differential mortality during the transition from fertilized egg to adult fly. That selection on the D. pseudoobscura inversions occurred in nature seemed to be settled, and it was generally taken for granted that viability differences accounted for the major part of it.Some 20 years later, a series of papers by Prout (7-10) stimulated evolutionary biologists to pay greater attention to the various components to fitness. These components determine the separate bits of selection that operate at...