HABITAT DURATION,LENGTH OF LARVAL PERIOD,AND THE EVOLUTION OF A COMPLEX LIFE CYCLE OF A SALAMANDER,AMBYSTOMA TEXANUM

In many organisms, genotypic selection may be a less effective means of adapting to unpredictable environments than is selection for phenotypic plasticity. To determine whether genotypic selection is important in the evolution of complex life cycles of amphibians that breed in seasonally ephemeral habitats, we examined whether mortality risk from habitat drying in natural populations of small-mouthed salamanders (Ambystoma texanum) corresponded to length of larval period when larvae from the same populations were grown in a common laboratory environment. Comparisons were made at two levels of organization within the species: 1) among geographic races that are under strongly divergent selection regimes associated with the use of pond and stream habitats and 2) among populations within races that use the same types of breeding habitats. Morphological evidence indicates that stream-breeding A. texanum evolved from pond-breeding populations that recently colonized streams. Larvae in streams incur heavy mortality from stream drying, so the upper bound on length of larval period is currently set by the seasonal duration of breeding sites. We hypothesized that selection would reduce length of larval period of pond-breeders that colonize streams if their larval periods are inherently longer than those of stream-breeders. The results of laboratory experiments support this hypothesis. When grown individually in a common environment, larvae from stream populations had significantly shorter larval periods than larvae from pond populations. Within races, however, length of larval period did not correlate significantly with seasonal duration of breeding sites. When males of both races were crossed to a single pond female, offspring of stream males had significantly shorter larval periods than offspring of pond males. Collectively, these data suggest that differences in complex life cycles among pond and stream-breeders are due to genotypic selection related to mortality from habitat drying. Stream larvae in the common-environment experiment were significantly smaller at metamorphosis than pond larvae. Yet, the evolution of metamorphic size cannot be explained readily by direct selection: there are no intuitively obvious advantages of being relatively small at metamorphosis in streams. A positive phenotypic correlation was observed between size at metamorphosis and length of larval period in most laboratory populations. A positive additive genetic correlation between these traits was demonstrated recently in another amphibian. Thus, we suspect that metamorphic size of stream-breeders evolved indirectly as a consequence of selection to shorten length of larval period.