Larval ecology and macroevolution in marine invertebrates

In marine mollusks, planktotrophic and nonplanktotrophic modes of development generally confer different scales of larval dispersal, which in turn influence species' geographic ranges and genetic population structures. In Late Cretaceous gastropods, planktotrophic species (as inferred from larval shell morphologies) exhibit significantly broader geographic ranges, longer species durations, and lower speciation and extinction rates than nonplanktotrophic species; these effects outweigh a variety of adult traits. Nor are the effects of these species-level larval traits reducible to selection at the organismic level, but instead evidently fulfill the requirements of true species selection: differences in genetic populations structure–mediated by modes of larval development–affect the multiplication and survival of species, vary among species but rarely within species, and are heritable at the species level. The geographic and stratigraphic pattern of first appearance of species with respect to other members of their clade suggests that larval types affect modes as well as rates of evolution, with the genetically-subdivided nonplanktotrophs subjected to a greater variety of processes than the potentially more panmictic planktotrophs. Nonplanktotrophs have the appropriate population structure for the operation of shifting balance mechanisms, and their fossil record is consistent with this possibility, with first occurrences nested in, but stratigraphically abutting, the ranges of related species. However, survival of genera during the end-Cretaceous mass extinction is unrelated to larval development and several other traits (e.g., species richness) that were influential during times of normal, background levels of extinction. Instead, taxonomic survivorship is related to the magnitude of the clades' geographic ranges–regardless of the geographic range of constituent species. This change in selectivity indicates that mass extinctions are not simply intensifications or extrapolations of background processes, but may be indifferent to the quality of adaptation, fitness values, or clade configurations achieved during background times. The results presented here suggest that evolutionary theory (1) should be more explicitly hierarchical (as argued by Gould, Eldredge, Vrba, Salthe, and others), with selection, drift, and other processes occurring at a series of focal levels that are neither wholly extrapolations nor wholly independent of processes at adjacent levels; and (2) must take into account the alternation of background and mass extinction regimes.