Populations of Drosophila littoralis are known to be latitudinally highly variable in photoperiodic adult diapause and pupal eclosion rhythm. Phenotypic correlations between the two time-measuring systems among the strains from different latitudes are, however, weak. In the present study, two differing strains were crossed reciprocally in order to search for causal (genetic) correlations between the two traits in the strains. Segregation in the F2 generations showed that variation in each trait was based on a few variable loci only. In the F2, flies having different eclosion times also differed in their diapause. This association was not complete and could have been due to genetic linkage between the traits. For that reason, the hybrid generations were raised for eight generations more to allow recombination between the traits. In F8, selection against diapause was started in the lines by raising them in a light-dark cycle of 15:9, where only females of the southern type reproduce. After eight selected generations, the lines were studied for the traits. Diapause was completely of the southern type, and the eclosion rhythm had also changed in parallel. The change in the phase of the free-running rhythm was not complete. From the present experiment, and from earlier knowledge of the geographical variation in D. littoralis, I conclude that the same pacemaker that is seen in the eclosion rhythm could also participate in daylength measurement for diapause. However, there are also noncorrelated variable parts in the measuring systems of both traits, which may mask the correlated variation.