SEX DETERMINATION IN TURTLES - DIVERSE PATTERNS AND SOME POSSIBLE ADAPTIVE VALUES

New data on the genders of young turtles from eggs incubated at controlled temperatures demonstrate temperature-dependent sex determination (TSD) in 17 species surveyed for the first time and corroborate TSD in another 11 species. The well-known pattern of males from cool temperatures and females from warm ones (= Pattern Ia) occurs in eight species and may occur in seven more. A pattern of cool females, intermediate males, and warm females ( = Pattern II) occurs in ten species. Data on three species surveyed for the first time are compatible with genetic sex determination (GSD). Within Clemmys, C. guttata has TSD; C. insculpta, GSD. Comparisons among phylogenetic sister groups suggest at least four independent losses of TSD in turtles. Large variation in extent of sexual differentiation at hatching among TSD species vs GSD species obscures any evidence that either system provides earlier or more complete differentiation. Pattern Ia occurs mainly in species in which adult females average larger than adult males; Pattern II occurs mainly in species with females smaller than males or in which body size is not dimorphic. Among TSD reptiles generally, the smaller gender typically arises at the coolest incubation temperatures. The new data, together with a literature survey, invite four possible explanations for various aspects of sex determination in reptiles: phylogenetic inertia, temperature-dependent differential fitness, sib-avoidance, and group-structured adaptation in sex ratios. Key demographic features from the literature include a high incidence of unisexuality within clutches and a predominance of female biases in sex ratios of hatchlings but not in those of adults. Each of these explanations remains partly but not fully plausible. Explanations for the patterns of sex determination may ultimately require a combination of hypotheses.