Search

Many turtle species have temperature-dependent sex determination (TSD), raising the prospect that climate change could impact population dynamics by altering sex ratios. Understanding how climate change will affect populations of animals with TSD requires a reliable and minimally invasive method of identifying the sexes of young individuals. This determination is challenging in many turtles, which often lack conspicuous external sexual dimorphism until years after hatching. Here, we explore four alternatives for sexing three age classes of captive-reared young gopher tortoises (Gopherus polyphemus), a terrestrial turtle of conservation concern native to the southeastern United States: (1) naive testosterone levels, (2) testosterone levels following a follicle stimulating hormone (FSH) challenge, (3) linear morphological measurements, and (4) geometric morphometrics. Unlike some other turtle species, male and female neonatal gopher tortoises have overlapping naive testosterone concentration distributions, justifying more complicated methods. We found that sex of neonates (<7 days old) is best predicted by a "random forest" machine learning model with naive testosterone levels and morphological measurements (8% out-of-bag error). Sex of hatchlings (4-8 months old) was predicted with 11% error using a simple threshold on naive testosterone levels, or with 4% error using a simple threshold on post-FSH testosterone levels. Sex of juveniles (approximately 3.5 years old) was perfectly predicted using a simple threshold on naive testosterone levels. Sexing hatchlings at >4 months of age is the easiest and most reliable non-surgical method for sex identification. Given access to a rearing facility and equipment to perform hormone assays, these methods have the potential to supplant laparoscopic surgery as the method of choice for sexing young gopher tortoises.

Nest site selection has consequences for hatching success by mediating the temperature and moisture conditions that eggs experience during the incubation period. Understanding the potentially complex pathways by which nest placement influences these abiotic mediators, and therefore hatching success, is important for predicting which nests will be successful and which may require management action. We studied the effects of loggerhead sea turtle (Caretta caretta) nest site selection on hatching success by linking nest placement characteristics to hatching success through a structural equation model. We monitored 170 nests on Ossabaw Island, Georgia, during the summers of 2017 and 2018 and tracked nest conditions throughout the incubation period. Temperature had a complex effect on hatching success-nests had higher hatching rates if they were exposed to higher mean temperatures but also if they experienced both extremely high (>34 degrees C) and extremely low (<26.5 degrees C) temperatures, suggesting that temperature variability plays a role in determining nest outcomes beyond the mean temperature. Likewise, hatching success declined with a higher incidence of nests being inundated by tides. We found that nests placed at the highest elevations had the highest hatching success rates, likely because those nests had a much lower chance of being washed over by high tides and had higher mean temperatures. Nests were also more successful when placed in greater amounts of vegetation, again because vegetated nests were generally warmer and were associated with fewer washover events. These results shed light on the mechanisms behind selection for certain nest site characteristics and can guide the relocation of nests as a conservation action.