Thermal performance and the MF and FMF patterns
The Mighty Males hypothesis, if correct, stands to provide insight into the evolution of thermal stress and thermal limits. This is because a main prediction of the hypothesis is that parameters associated with TSD (e.g., TPiv, the range of male-producing temperatures, etc) evolve so that males are produced in thermal environments that impart relatively high-quality phenotypes. Here, I summarize aspects of thermal performance theory that are relevant to Mighty Males, as well as key research on whether ecologically relevant incubation environments regularly experience heat stress.
Thermal performance curves (TPCs) characterize the relationship between performance and temperature (Huey & Stevenson 1979). Biochemical constraints dictate that the shape of TPCs are typically Gaussian and left skewed, and a key characteristic of TPCs is that they predict a rapid decrease in performance at high temperature (Schoolfield et al. 1981; Kingsolver 2009; Amarasekare & Johnson 2017). A consequences of TPC asymmetry is that temperatures higher than the optimal temperature for performance depress fitness more than an equivalent temperature displacement below the optimal temperature (Martin & Huey 2008). Thermoregulatory behaviour should therefore evolve such that mean body temperature is lower than the temperature that maximizes performance, as organisms do not thermoregulate with perfect accuracy, and overshooting the optimal temperature has relatively strong and negative fitness consequences (Martin & Huey 2008). For reptiles with TSD, capacity for embryonic thermoregulation exists (Ye et al.2019), but embryos cannot physically displace themselves, and so thermoregulation is unlikely to result in widespread avoidance thermally stressful environments (Telemeco et al. 2016). The key point here is that avoidance of heat stress has a large influence on the evolution of thermoregulation, such that avoidance of heat stress can be considered evolutionarily important (Martin & Huey 2008). Given that embryos cannot move, they are generally far more susceptible than adults to the negative fitness consequences associated with heat stress.
There are many specific examples of how exposure to hot incubation environments results in low-quality phenotypes, and through a variety of pathways. For instance, exposure to extreme heat, but not extreme cold, during natural incubation results in hatchling shell deformations in wild Chyrsemys picta (‘extreme’ defined as ±2SD from grand mean incubation temperature over two years), and the deformations themselves seem to have negative fitness consequences (Telemeco et al.2013). Similarly, warm incubation temperatures supress the innate immune response of hatchlings in two distantly related turtles species, whereas cool temperatures enhance immune response (Freedberg et al. 2008; Dang et al. 2015). It is also possible that elevated embryo metabolism may not be matched by increased oxygen supply in reptiles incubated at high temperatures (Hall & Warner 2019), such that negative fitness consequences may arise in part through oxygen deprivation. As a final example, high and constant incubation temperature is also associated with small size at hatchling (Warner et al. in press; Packard et al. 1987b , 1988; Janzen and Morjan 2002), likely because temperature has stronger effect on development than on growth at all life stages (Forster et al. 2011), and small size tends to be associated with lower fitness in juveniles and adults (Rollinson & Rowe 2015; Armstrong et al. 2017). In sum, there are a variety of ways in which hot environments can decrease phenotypic quality.
More generally, evidence of the stress imparted by high temperature arises in the existence of heat-shock proteins (HSPs). HSPs are a broadly conserved group of molecular chaperones designed to buffer the impact of heat stress on phenotypes (Sørensen et al. 2003), for which there is no known equivalent for cold stress (Sinclair & Roberts 2005). Both heat stress and/or the overexpression of heat-shock proteins have subsequent deleterious and long-term effects on performance, including development and survival, acting through a variety of phenotypic pathways (Feder & Hofmann 1999; Kingsolver & Woods 2016). Thus, one simple test of Mighty Males is to assess whether thermal stress is more likely under female-producing conditions, by testing whether the expression of heat shock proteins is positively associated with female sex under environmentally relevant conditions. Critically, testing this prediction should also be done in concert with exploring the range of incubation temperatures in wild nests in order to estimate environmentally-relevant temperatures. For instance, in some populations, embryos of FMF populations rarely experiences temperatures beyond the upper TPiv (e.g., Warner & Shine 2008a; Rollinson et al. 2018); in other FMF populations, temperatures below the lower TPiv are rare (Janzen 2008). Indeed, a broad spatial and temporal characterization of incubation environments is necessary to estimate environmentally relevant temperatures (e.g., Carter et al. 2018; Francis et al. 2019).
Although heat stress is generally expected at high temperature, embryonic thermal performance will ultimately adapt to the thermal environment, such that ecological context is required understand thermal stress and thermal limits. The Mighty Males hypothesis generates at least two predictions will arise from local adaptations of embryos to the thermal environment. The first deals specifically with FMF species. The logic of the prediction arises from the theoretical expectation of a trade-off between TPC height and breadth (Gilchrist 1995), such that in seasonal environments, local adaptation of TPCs will result in relatively platykurtotic TPCs centered on low mean temperatures (e.g., Fig. 4a,b). This reflects thermal adaptation to a relatively unpredictable environment that features both seasonal variation in thermal means, and pronounced diurnal thermal fluctuation of a magnitude that varies seasonally (Amarasekare & Johnson 2017; Francis et al. 2019). In other words, seasonality favours a form of TPC evolution where individuals are relatively tolerant of, and adapted to, a wide range of temperatures; hence “good” thermal environments for TSD species feature a wide range of incubation temperatures. This prediction is, therefore, that the range of male-producing temperatures will be positively associated with the degree of temperature variation inherent in the environment (Fig. 4a,b), or more specifically with the range of incubation temperatures experienced by the average embryo. Indirect support for this prediction is provided by Ewert et al. (2004), whose data suggest that the range of male-producing temperatures was positively associated with latitude in an FMF turtle across six populations, as is generally expected under Mighty Males (Fig. 4c). A quantitative test of this prediction is nevertheless warranted, as Ewert et al. (2004) focus on latitude and not variance in the average incubation environment. Unfortunately, it is not intuitive how thermal adaption to seasonality would influence TSD parameters in MF species under Mighty Males, precluding a similar prediction for MF species. In any event, recent evidence suggests that TSD parameters are not strongly related to latitude or longitude in at least one MF species (Carteret al. 2019a), although variance of incubation temperature was not explored in this study.
The second prediction of Mighty Males under local adaptation to the thermal environment is that females should suffer greater mortality than males, especially at the egg and hatchling life stages. The logic is that Mighty Males predicts TPiv to exhibit a correlated evolution with thermal performance and thermal tolerance, specifically so that TPiv marks the departure from favourable to unfavourable thermal environments that impart low-quality phenotypes. The prediction may be difficult to test at the adult stage, as viability selection on adults tends to be relatively weak in nature in the first place (Kingsolver et al.2001, 2012), and females that are most strongly affected by thermal stress will die either before hatching or shortly thereafter, leaving females with relatively subtle phenotypic effects to survive until adulthood. Indeed, the only study to my knowledge to examine survival differences between the sexes of adult turtles found no difference (Chaloupka & Limpus 2005). However, this prediction should be straightforward to test at the egg and hatchling stage, where viability selection tends to be stronger (Rollinson & Rowe 2015). The specific expectation under constant incubation conditions is that fitness (e.g., embryonic survival) should depreciate relatively rapidly when embryos are incubated above vs below (upper) TPiv. Under fluctuating thermal conditions, a similar prediction for MF species is that fitness (e.g., survival or embryonic deformity rate) should be positively associated with the extent of female-biased sex ratios. The recent publication of a comprehensive database on phenotypic outcomes of reptilian incubation will facilitate tests of this prediction (Noble et al. 2018b, a).