LABORATORY EVOLUTION OF POSTPONED SENESCENCE IN DROSOPHILA MELANOGASTER.

Abstract

Evolutionary genetics seems to have found the fundamental cause of senescence: the decline in the sensitivity of natural selection to gene effects expressed at later ages in most populations of organisms with separate somatic and germline tissue. (Here refers to decline in age-specific fitness-components after the onset of reproductive maturity.) This idea traces back to Haldane (1941) and Medawar (1946, 1952), with considerable elaboration and elucidation since then (Williams, 1957; Hamilton, 1966; Edney and Gill, 1968; Emlen, 1970; Charlesworth and Williamson, 1975; Charlesworth, 1980; Rose, 1983a). While there are still clear limitations to the mathematical formulation of this theory (cf. Hamilton, 1966; Charlesworth, 1980), the basic formal analysis leads to a straightforward conclusion: the first partial derivative of fitness with respect to appropriately scaled changes in age-specific life-history characters usually declines in magnitude with the age of these changes. The force of natural selection thus declines with age. This overall theory and its particular subsidiary variants lead to a number of empirically testable corollaries (Rose, 1983a, 1983b). Some of these corollaries are specific to the subsidiary variants of the theory (Rose and Charlesworth, 1980, 1981a, 1981b; Rose, 1983b), sothattests of them individually do not test the theory as a whole. Fortunately, there are two corollaries which follow from the general theory itself: the reproductive schedule of an outbred population will give rise to natural selection acting to (i) accelerate senescence in populations with a relatively earlier age of reproduction and (ii) postpone senescence in populations with a relatively later age of reproduction (Edney and Gill, 1968; Rose, 1983a). The former prediction has been corroborated by Sokal (1970) using Tribolium castaneum, while the latter has been corroborated by Wattiaux (1968a, 1968b) and by Rose and Charlesworth (1980, 198 lb), using Drosophila species. Once a theory has been well-developed mathematically and then empirically corroborated, attention turns to experiments in which the theory either is not clearly corroborated or is ostensibly refuted. It would be misleading to claim that all relevant experimental results directly corroborate the evolutionary theory of senescence. Sokal (1 970) and Mertz (1975) using Tribolium castaneum and Taylor and Condra (1980) using Drosophila pseudoobscura found heterogeneity between lines in experiments with replication, such that some lines did not exhibit the predicted response to the imposed selective regime. Taylor and Condra (1980) also found a difference in the response of the sexes which was later attributed to the pattern of female mating preference (Taylor et al., 1981). More problematic still are the studies from the Lints laboratory, one of which failed to obtain a direct response to artificial selection for longevity (Lints et al., 1979), while another gave puzzling fluctuations in life-history attributes (Lints and Hoste, 1974, 1977). Lints (1978, 1983) has made a great deal of these problems, contending that they cast doubt on all proposed evolutionary theories of senescence. While it can be argued that these puzzling results are due to technical artifacts such as inbreeding, genetic disequilibrium, and inadequate controls (cf. Rose and Charlesworth, 1981b), the only ef-