The ways in which organisms determine sex are diverse. Sex determination systems are important because of the effect they have on sex ratios within a population, which affects reproductive success, levels of inbreeding, and population viability. In environmental sex determination (ESD), sex is determined in response to immediate environmental factors after conception. A common form of ESD in non-avian reptiles and some fish is temperature-dependant sex determination (TSD). Individual sex is determined by the temperature during the middle one-third of embryonic development, and the threshold between male and female development is often over a very small (< 2yC) range. While this form of sex determination has been maintained in a species or population over millions of years, TSD species may experience drastic skews in sex ratio in response to large climatic upheavals. Theoretically, response to selection for the rarer sex may produce evolutionary change at two levels in reptiles: 1) maternal nesting behavior with respect to thermal conditions and 2) thermal sensitivity of the threshold between male and female development.

The results from this dissertation reveal that both onset of nesting and nest-site vegetation cover have low heritability and thus low potential to respond to selection. Nesting behavior is dependent on the winter preceding the nesting season, however, and environment-specific analyses suggest that additive genetic variance increases for onset of nesting after warmer winters and increases for vegetation cover over the nest after cold winters. As a result, heritability may be dependent on the winter preceding the nesting season in this system. Estimates of repeatability corroborate these results. There is a significant, genetic-based tendency for turtles to nest in areas with minimal vegetation cover after colder winters, while there is a significant, genetic basis for earlier first nesting dates after warmer winters. Genetic correlations between warmer and colder winters for vegetation cover are very high and suggest that no potential genotype-by-environment (G y E) interaction exists, while potential G y E between winter environments for first nesting date between winter environments is inconclusive.

Threshold temperatures may respond to selection for sex ratio biases as well. Maternal half-sib analysis facilitated by natural multiple paternity suggests that family effects on hatchling sex in TSD species are not exclusively driven by maternal effects and that there is a detectable genetic variance of the sire. Thus, the sex determination pathway's sensitivity to temperature (i.e. primary sex ratio) may evolve in response to sex-ratio selection. The effective heritability, which predicts the relative rate of change of threshold temperature, was estimated to be 0.11, while the effective heritability of nesting behavior was estimated to be 0.079.

Since sire genetic variance can substantially influence the sex of offspring when the clutch is incubated at temperatures that produce both sexes, mating with multiple males may have a homogenizing effect on sex ratio variance within and between nests. No significant reduction in sex ratio variance for multiple paternity clutches was observed in the study presented here, but theoretically TSD may provide an ideal situation for bet-hedging. Clutches with multiple sires had higher hatching success rate and lower variance for hatching relative to clutches with a single sire. The incidence of multiply sired clutches increased with plastron length (a proxy for age), although there was no significant interaction between fitness, plastron length, and multiple paternity.

In conclusion, these studies suggest that temperature dependent sex determination may respond to selection from sex ratio biases either through threshold temperature or nesting behavior, but both are likely to respond to selection slowly. Further, multiple paternity does not seem to homogenize sex ratios between clutches for the population examined here; however, additional studies are needed to exclude possible confounding factors.