An important tool in the genetic analysis of longevity and aging in Drosophila melanogaster is the use of strains selected directly for late-age reproduction and indirectly for extended longevity. Following some initial failures to select for extended longevity, there are now a number of laboratories which have successfully selected for long life, using the techniques of late-age reproduction as well as selection for stress resistance. Baret and Lints [Gerontology 1993;39:252-259] have recently cast doubt on the reality of a number of these selected strains, including our own, suggesting that the difference in longevity between the long-lived and normal-lived strains disappears when the data are examined as a function of the number of days after the beginning of the selection experiment instead of as a function of the number of generations. With regard to our selected lines, they based their analysis on a subset of the published data dealing with these strains, and which covered 21 generations, or 40 months, of selection. We now present data for over 70 generations, or 155 months, of selection and maintenance. The Baret-Lints hypothesis makes two strong predictions, namely that (1) the longevity difference between the several strains should disappear when the data are replotted according to their fashion, and (2) there should be no other significant biological difference between the strains. Our data falsifies both of these predictions. The Baret-Lints hypothesis is flawed and should be disregarded.(ABSTRACT TRUNCATED AT 250 WORDS)