Abstract
<b><i>Introduction:</i></b> Sex is a fundamental characteristic of an individual. It is therefore puzzling why in some systems sex is precisely determined by a genotype, while in others it is influenced by the environment or even subtle, not to say random, factors. Some stochasticity in sex determination would be expected if environmental conditions did not have a large sex-specific effect on fitness. Although data are only available for a small fraction of species, geckos show exceptional variability in sex determination. <b><i>Methods:</i></b> We tested the effects of three incubation temperatures on sex ratio and adult body size in the invasive gecko <i>Phelsuma laticauda</i> and the vulnerable gecko <i>Phelsuma nigristriata</i>. <b><i>Results:</i></b> We document temperature-dependent sex determination (TSD) in both species. Only females hatched at a low temperature (24°C), whereas male production peaked at an intermediate temperature (28°C) and declined, at least in <i>P. laticauda</i>, again at the highest temperature (31°C). Interestingly, full siblings hatched from eggs glued together during the incubation at temperatures producing both sexes are often of the opposite sex. We found no significant effect of incubation temperature on adult body length. <b><i>Conclusions:</i></b> Documentation of TSD in the day geckos has implications for conservation practice in environmental management of endangered species or eradication of invasive species. However, it appears that a very subtle (random?) factor may also be involved in their sex determination. In line with this, we found no significant effect of incubation temperature on adult body length.