| Summary: | How do environmental fluctuations affect the evolution of life history traits? Advocates of r- and K-selection claim that fluctuations select for early maturity, large reproductive efforts, and many young. Advocates of bet-hedging claim that they can select for delayed maturation, small reproductive efforts, and few young. In testing these ideas, I worked with two species of poeciliid fish introduced to reservoirs in Hawaii: Gambusia affinis and Poecilia reticulata. In some reservoirs, water levels fluctuated. In others, they did not. I concluded that (1) Gambusia and Poecilia have undergone an evolutionary experiment in Hawaii lasting 50-70 years, and that (2) the experiment was fairly well controlled.
I compared fish from stable and fluctuating reservoirs for three traits: size at maturity, number of young, and reproductive efforts. (3) Eight of ,9 intraspecific differences in reproductive traits between stable and fluctuating reservoirs were in the direction predicted by r- and K-selection, but 7 of these 9 differences were not significant. (4) There were no intraspecific differences in size at maturity, which r- and K-selection had predicted should be largest differences of all.
(5) Poecilia had more young, smaller young, and made larger reproductive efforts in the fluctuating reservoirs, as predicted by r- and K-selection. (6) Although Poecilia has reproductive traits, relative to Gambusia, which would be called r-selected, Gambusia, not Poecilia, dominated the fluctuating reservoirs, and Poecilia did better in the stable reservoirs.
I suggested that Gambusia was pre-adapted for fluctuating reservoirs. Gambusia females produce female young with a wide range in ages at maturity (2.5-7 months). Poecilia females, whose young mature at 2.3 to 3.5 months, cannot do this. By insuring that some progeny always survive long drawdowns, this trait uncouples Gambusia from the water level fluctuations. I concluded; (7) Reproductive traits need not be associated in the particular groupings that recent evolutionary theory predicts. Poecilia, for example, showed large differences in number of young and reproductive efforts, but no differences in size at maturity. (8) The trends discussed by r- and K-selectionists
may be incorporated into a reproductively polymorphic population as part of a bet-hedging adaptation.
(9) By doing multiple regression analyses with long-term measures of water level fluctuations derived from time series 16 to 27 years long, I could explain at best 36.6% of the variability in number of young. (10) When I added short-term measures, the explainable variability in number of young shot up to 61.8%. Short-term measures alone could account for 46.4% of the variability in number of young. (11) Therefore I abandoned the idea that most of the variability in life history traits observed in the field has evolved; much of it results from developmental plasticity.
By raising Gambusia in the lab under controlled density, food, and temperature, I found: (12) Size of young at birth, growth rates, and ages at maturity differed significantly among reservoirs. (13) By varying food and temperature, I could
produce in a single laboratory stock differences as large as I observed between any two reservoirs in the field. (14) The differences among stocks in reproductive traits were, with two
exceptions, in the same direction in the lab as in the field.
(15) Long periods of low water and restricted food favored large fish (>20 mm) in both lab and field. Short-term fluctuations in the field seem to favor small fish (<22 mm).
These laboratory results led me to the following conclusions: (16) The life history traits of Gambusia have evolved, since 1905, in different directions in different reservoirs. A detectable and significant portion of the variability among stocks had a genetic basis. (17) In the
fluctuating reservoirs, Gambusia are undergoing oscillating selection pressures on age at maturity and growth rate. They cope with these pressures by producing young with a great range of growth rates and ages at maturity.
Although I did not falsify the original predictions, I did demonstrate that the models leading to those predictions ignored details that were demonstrably important: (a) developmental plasticity in life history traits, (b) the range of variability of life history traits in a single brood, and (c) the differences among several different kinds of instability. === Science, Faculty of === Zoology, Department of === Graduate
|
|---|
