No conclusive evidence regarding the evolutionary advantage of TSD has been discovered yet. However, several hypotheses have been investigated including sexual dimorphism and predator-avoidance strategies. The current thinking regarding the overall evolution of sex determining mechanisms in holds that the ancestors of vertebrates were hermaphroditic with environmental sex determination evolving later. Genotypic sex determination is believed to have evolved from ESD systems, with heteromorphic sex chromosomes as the most recently evolved characteristic (Bull 1980).
Some models have proposed that ESD offers species an advantage by allowing the sex of the embryo to differentiate into the sex that is most favored by the immediate environment (Bull 1980). However, in any case of TSD, one must show that both offspring sex and fitness are influenced by temperature and that the range of male- producing temperatures correlates to heightened male fitness. The same must be true for females as well. (Janzen 1995).
Sexual dimorphism hypothesis
The sexual dimorphism hypothesis holds that incubation temperatures should correspond with the production of the sex that has the greatest advantage of being large (Janzen and Paukstis 1991). That is to say, the sex that benefits from being large should be produced at temperatures that yield a larger adult. Typically, there are three recognized patterns of TSD.
Figure 2 depicts the phylogenetic tree for turtles showing sex-determining mechanisms and patterns of sexual-dimorphism in body size. E=ESD, H=homomorphic sex chromosomes, XY=male heterogamety, ZW=female heterogamety. Inequality signs represent body size of male adults in comparison to body size of female adults. Figure 2 adapted from: Janzen, F. J. and G. L. Paukstis. “A Preliminary Test of the Adaptive Significance of Environmental Sex Determination in Reptiles.” Evolution 45 (1991): 435-440. JSTOR. Davidson Coll. Lib., Davidson, NC. 15 Sept. 2004.
When this hypothesis is tested by correlating adult body size with either ESD or GSD the results are inconclusive. Only six of eleven groups of turtles with ESD are known to have large-male dimorphism and only three of five lizard assemblages with ESD exhibit large female dimorphism. In crocodiles the pattern of ESD does not correlate to any pattern of sexual dimorphism (Janzen 1991).
In another study of the pancake tortoise Malacocherus tornieri, researchers demonstrated that while the species exhibits large female sexual dimorphism, males are the faster growing sex. This finding does not support the hypothesis that the larger sex should be produced at nest temperatures that produce faster growth (Ewert et al 2004).
Another hypothesis regarding the adaptive advantage of TSD holds that offspring produced at different temperatures may be better suited to avoid predation and thus increase their fitness. One study using hatchling snapping turtles (Chelydra serpentine) found that offspring produced at intermediate temperature (one that produced both sexes) had a much higher propensity to run. Those offspring from either all male-producing temperatures or all female-producing temperatures were more inclined to remain motionless. In semi-controlled field conditions, the intermediate temperature hatchlings experienced significantly higher predation rates over a period of one year, presumably due to their higher visibility while moving (Janzen 1995). While this study involved only one species, it is important to remember that selection may act differently from species to species, leading to difficulty in finding systemic patterns.
Chelydra serpentina. Photo by Michael Dorcas, used with permission.
Obviously, the dependence on environment and temperature in particular could have detrimental effects on the sex ratio of a species with TSD. Large scale climatic changes could significantly alter the sex ratios of certain species and skew population dynamics, in turn favoring species with GSD (Bull 1980). However, the persistence of TSD in reptiles through millions of years would suggest that this scenario has not had a significant effect.