It has proven relatively easy to select normal-lived strains of Drosophila for extended longevity in the laboratory. Long-lived strains have not been observed in the wild as yet. Of the various life-history traits that have been investigated for their role in modulating the evolution of extended longevity, none have yet shown a consistent or convincing relationship. Other than developmental time, the traits usually investigated in this regard are those associated with the adult phase of the life cycle. We assayed developmental timing and viability in six pairs of normal- and long-lived strains, four pairs of which are from previously described strains and two pairs of which are new strains that have been independently and recently selected. We find that the life-history trait most obviously associated with all our long-lived strains is a significantly reduced developmental viability, with the long-lived strains' having as much as twice the developmental lethality as do any of the normal-lived strains. The long-lived strains also pupate closer to the food, a behavior known to decrease fitness. Thus the reduced fitness of the long-lived strains appears to be due to both physiological and behavioral factors and may well explain why long lived strains are not usually found in the wild. The extension of longevity involves costs as well as benefits that, in this case, are borne by different individuals.