Life history traits--hatchability, longevity, and egg production--of five wild-caught populations of Drosophila melanogaster were measured after these populations had been reared in constant laboratory conditions during a 4-year period. The results were analyzed together with those that had been obtained with the same populations just after capture. They are probably the first convincing results that reveal the existence of genetic variability for some life history traits measured in the laboratory. Besides, no significant phenotypic correlations, either positive or negative, between early and late components of fitness were found. Finally, the five populations showed different patterns of genetic correlation between early and late fitness traits. One of the populations showed a negative correlation, another showed a positive correlation, while the remaining three populations showed no correlation at all. This was equally observed at the within- and between-population levels. That result suggests that both the antagonistic pleiotropy hypothesis proposed by Williams and the concordant pleiotropy hypothesis suggested by Lints are not of general validity.