Search

1. Over half of all crocodylians are threatened with extinction, with many species considered to be functionally extinct throughout their historical ranges. How such losses impact crocodylian functional diversity (FD) and its conservation is poorly understood, but can be investigated using measurable traits. Where such information is unavailable, phylogenetic diversity (PD) has been proposed as an effective strategy to capture FD, but this assumption is not well tested. 2. We constructed a global trait database for 28 extant crocodylians, and quantified their FD, distinctiveness and groupings. Functionally distinctive crocodylians include the most and least evolutionarily distinct species, likely because unusual trait values and combinations can be produced by both evolutionary isolation and evolutionarily recent adaptive radiation. We identified four functional groups of crocodylians, within which species have similar ecological roles. Highly distinctive species, such as the gharial, fit less well into functional groups. 3. We combined functional distinctiveness (FUD) and threat ranking scores to prioritise species for conservation. The metric is termed Ecologically Distinct and Globally Endangered (EcoDGE), and we tested how well it operates to preserve crocodylian FD among extant Crocodylia. Under current extinction scenarios, 32%–38% of crocodylian FD might be lost within 100 years. Losses increase to 48% if all threatened species go extinct, with Asia identified as a hotspot of threatened FD in crocodylians. Highly threatened distinctive species lead this extinction scenario, exacerbated by clusters of species sharing traits linked to high extinction risk. These traits include habitat specialism and lower reproductive investment; in contrast, species that tolerate climate extremes are more resilient. 4. Prioritising at-risk species based on FUD and threat status appears to be an effective way to conserve present-day crocodylian diversity. Conservation prioritisation based on the EcoDGE metric performed well to minimise projected losses in FD. Among extant crocodylians, FD losses are projected when high diversity overlaps with high extrinsic threats. 5. We then examined the utility of PD-based measures as surrogates for FD conservation, to better understand the advantages and drawbacks of such substitutions. A PD-based measure such as the Evolutionarily Distinct and Globally Endangered metric provided an effective tool to capture FD in our tests.

The sex of American Alligator (Alligator mississippiensis) hatchlings is deter- mined by the egg temperature during the middle third of the 9-12 week incubation period. As a consequence, predictable sex ratios are possible for clutches incubated in constant temperatures in the laboratory, but naturally occurring sex ratios of American Alligator hatchlings from wild nests exposed to fluctuating temperatures are not well documented. Over a 5-year period (1995-1999), we determined the sex of American Alligator hatchlings from wild nests left in the field until after sex was irreversibly determined. A total of 6226 hatchlings from 232 naturally incubated wild nests showed a strong female bias (71.9% females, yearly range = 62.3-89.4% females). Most nests (64.2%) produced hatchlings of both sexes. Of the remaining clutches that produced exclusively one sex (83 nests), 78 nests produced all females, and 5 nests produced only male hatchlings. For the 2 years in which nest-cavity temperatures were known, higher temperatures led to production of significantly more male hatchlings (P < 0.001 for both 1997 and 1999). Knowledge of natural sex ratios of hatchlings can aid in the management and harvest of this commercially valuable species, and in understanding sex-ratio bias in American Alligator populations.

Sex ratios of the American snapping turtleChelydraserpentina are female-biased at cool temperatures, male-biased at moderate temperatures, and only females are produced at warm temperatures. The laboratory incubation of eggs at several constant temperatures yields patterns of skewed sex ratios. In this study, incubated eggs represented regional samples of C. serpentina from six different latitudes. The sample from the highest latitude yielded malebiased sex ratios across the broadest range of cool temperatures. This male-biased zone and those of the other five samples collectively contributed to a nested arrangement that is approximately symmetrical around a temperature of 24.5 ◦ C. Two indices of temperature-dependent sex determination (TSD) patterns (MF pivotal temperature, width of the male-biased zone) were significantly associated with latitude and with published data on several climatic variables (e.g. sunshine and rainfall). At high latitudes, females nest in exposed locations, locally the warmest places. At low latitudes females place their nests in shaded locations. Such nest site selection renders high latitude nests warmer than low latitude nests, even if only intermittently. Given that constant conditions show faster embryogenesis in high vs low latitude eggs, exposed nest sites seem to be an additional adaptation to hasten embryogenesis in high latitude nests. At low latitudes, shading prevents nests from overheating. We propose that a trade-off between evolution, (1) to balance sex ratio and (2) to maintain male-biased sex ratios in a temperature range that favours male fitness, leads to the observed arrangement of concentric TSD patterns centred around male production at 24.5 ◦ C.

Recent theoretical papers on temperature-dependent sex determination in reptiles are based on the assumption that crocodilian populations, particularly adults, are markedly female biased. While there is evidence that some crocodile populations may conform to this expectation, there is no compelling evidence that American alligator populations are female biased anywhere in the species' range. Previous data had indicated that some populations of juvenile and/or adult alligators were significantly male biased. However, these studies were criticized for sampling errors. Adult males typically occupy a different habitat from that of females, and males frequent areas where they are more likely to be caught. In contrast, in juveniles, both sexes occupy the same habitat. We determined the sex ratios of c. 3000 juvenile alligators collected from 11 sites over 6 years in south Louisiana. Our results indicate a significant sex bias (58% male), but variation was evident among samples. Sex ratios varied by year and site, and the interaction of these factors was significant. At one site in one year, there was a female bias (71%), but in no year was there an overall female bias across sites. From this study and earlier reports, we conclude that there is no evidence at present of female-biased sex ratios in the juvenile and/or adult alligator populations, and consequently, that certain models of temperature-dependent sex determination in reptiles require re-evaluation. Furthermore, our data suggest a pattern of differential mortality of females vs males during the first years of life, a pattern consistent with a key prediction of several differential fitness models for the adaptive significance of temperature-dependent sex determination in reptiles.

We assessed the ability of hatchling painted turtles (Chrysemys picta) from northern Minnesota to resist freezing when exposed to conditions like those occurring in natural hibernacula (natal nests). We placed animals individually into artificial hibernacula constructed in jars of damp, loamy sand and then lowered the temperature to approximately -0.4°C, which was below the equilibrium freezing point for water in the soil (approximately -0.1°C) but above that for body fluids of the turtles (approximately -0.7°C). Next, ice was used to initiate freezing of water in the soil, after which the substratum was allowed to freeze to an equilibrium temperature of -0.4°C. The minimum temperature then was reduced by 1°C/day to either -2.5 or -6.5°C. The minimum was maintained for 9 days in the former treatment and for 5 days in the latter, so that turtles in both treatments were exposed for the same length of time to temperatures below the equilibrium freezing point for their body fluids. Some animals in each treatment were inoculated by ice crystals that penetrated their integument, and these turtles froze and died. However, many other animals resisted inoculation, remained unfrozen, and survived the test of tolerance. We conclude that hatchling painted turtles from northern Minnesota, like those from North Dakota, north-central Nebraska, and northern Illinois, withstand the cold of midwinter by supercooling and not by tolerating freezing.

Unlike birds and mammals, in many reptiles the temperature experienced by a developing embryo determines its gonadal sex. To understand how temperature-dependent sex determination (TSD) evolves, we must first determine the nature of genetic variation for sex ratio. Here, we analyze among-family variation for sex ratio in three TSD species: the American alligator (Alligator mississipiensis), the common snapping turtle (Chelydra serpentina) and the painted turtle (Chrysemys picta). Significant family effects and significant temperature effects were detected in all three species. In addition, family-by-temperature interactions were evident in the alligator and the snapping turtle, but not in the painted turtle. Overall, the among-family variation detected in this study indicates potential for sex-ratio evolution in at least three reptiles with TSD. Consequently, climate change scenarios that are posited on the presumption that sex-ratio evolution in TSD reptiles is genetically constrained may require reevaluation.

Abstrcact.-We examined a critical component of the Charnov-Bull hypothesis of temperaturedependent sex determination (TSD) by determining the reaction norms of hatchling growth to embryonic incubation temperature in the common snapping turtle, Chelydra serpentina. Hormone manipulations of eggs produced females at male temperatures and vice versa, which thereby permitted same-sex comparisons of hatchling growth across a range of incubation temperatures. In this way, the normally confounded effects of incubation temperature and sex were dissociated experimentally. The resultant hatchlings, including controls and experimentals, exhibited normal gonadal structure and sex steroid profiles. The subsequent growth of hatchlings monitored for 6 mo was strongly affected by embryonic incubation temperature but not by sex. As predicted, growth was enhanced at incubation temperatures that produced males. Clutch effects and interaction effects (clutch by incubation temperature) on growth were significant. In addition, there was a positive genetic covariance among incubation temperatures, but incubation temperature effects varied among clutches. The variation in growth plasticity among clutches was consistent with Charnov-Bull predictions. In this TSD species, incubation temperature is likely to have differential fitness effects on the sexes mediated via differences in growth.

We examined agonistic behaviour in hatchling Australian freshwater crocodiles (Crocodylus johnstoni) at 2 weeks, 13 weeks, and 50 weeks after hatching, and between C. johnstoni and saltwater crocodiles (Crocodylus porosus) at 40–50 weeks of age. Among C. johnstoni, agonistic interactions (15–23 s duration) were well established by two weeks old and typically involved two and occasionally three individuals, mostly between 17 : 00 and 24 : 00 hours in open-water areas of enclosures. A range of discrete postures, non-contact and contact movements are described. The head is rarely targeted in contact movements with C. johnstoni because they exhibit a unique ‘head raised high’ posture, and engage in ‘push downs’. In contrast with C. porosus of a similar age, agonistic interactions between C. johnstoni were conducted with relatively low intensity and showed limited ontogenetic change; there was also no evidence of a dominance hierarchy among hatchlings by 50 weeks of age, when the frequency of agonistic interactions was lowest. Agonistic interactions between C. johnstoni and C. porosus at 40–50 weeks of age were mostly low level, with no real exclusion or dominance observed. However, smaller individuals of both species moved slowly out of the way when a larger individual of either species approached. When medium- or high-level interspecific interactions did occur, it was between similar-sized individuals, and each displayed species-specific behaviours that appeared difficult for contestants to interpret: there was no clear winner or loser. The nature of agonistic interactions between the two species suggests that dominance may be governed more strongly by size rather than by species-specific aggressiveness.

Hatchling painted turtles (Chrysemys picta), Blanding's turtles (Emydoidea blandingi), slider turtles (Trachemys scripta), and snapping turtles (Chelydra serpentina) were inoculated with ice at temperatures near the equilibrium freezing point for their body fluids (ca. -0.7 C) and then frozen at -2.0 C. All animals survived freezing for up to 30 h, but mortality among sliders and snapping turtles increased rapidly with longer exposures. Despite the fact that neonates of all four species are able to recover from freezing at a high subzero temperature, the adaptive strategy manifested by animals in the field is to avoid freezing altogether via attributes of morphology (painted turtles), behavior (Blanding's turtles, snapping turtles), or distribution (slider turtles). The discovery of a tolerance for freezing in hatchlings of species having diverse distributions and natural histories raises the possibility that such tolerance is a trait of general occurrence among neonatal turtles and that it is not an adaptation specifically enabling animals to withstand the rigors of winter. Hatchlings of the North American painted turtle (Emydidae: Chrysemys picta) typically spend their first winter inside the shallow, subterranean nest where they completed embryogenesis the preceding summer (Ernst et al., 1994). This behavior commonly causes neonates in northerly populations to be exposed during winter to ice and cold (Breitenbach et al., 1984; Storey et al., 1988; Packard et al., 1989; Costanzo et al., 1995), with temperatures in nests sometimes going below -10 C (Woolverton, 1963; DePari, 1996; Packard, 1997; Packard et al., 1997a). Hatchling painted turtles are small (3-6 g), ectothermic animals with limited heat capacity, so their deep body temperature presumably tracks the temperature in their nest quite closely (Claussen and Zani, 1991). Nevertheless, some turtles commonly survive exposure to such temperatures and emerge from their nest when the ground thaws the following spring (Woolverton, 1963; DePari, 1996; Packard, 1997; Packard et al., 1997a). We have proposed that neonatal painted turtles in northerly populations withstand exposure to ice and cold by remaining unfrozen and supercooled (Packard and Packard, 1995a). According to this theory, the skin of hatchlings is not easily penetrated by growing crystals of ice (Packard and Packard, 1993b), so turtles are not necessarily caused to freeze by inoculation when they make contact with ice in frozen soil (Packard et al., 1997b, 1999). This protection from inoculation enables the animals to exploit an innate capacity for supercooling (Packard and Packard, 1993a, 1995b, 1997; Packard and Janzen, 1996; Costanzo et al., 1998), so that they remain unfrozen over the usual range of subzero temperatures encountered in nests in the field (Packard et al., 1997b, 1999). However, on those occasions when the integument is not up to the challenge and ice crystals manage to penetrate into body compartments from the environment, the animals freeze and die (Packard et al., 1997b, 1999). Thus, turtles that emerge from nests in the spring presumably are ones that remained unfrozen and supercooled for the entire winter. However, a nagging problem with the concept that hatchling painted turtles exploit a capacity for supercooling to survive winters in the field is the indisputable evidence from experiments in the laboratory that neonates often recover from freezing at relatively high subzero temperatures (Storey et al., 1988; Churchill and Storey, 1992a; Rubinsky et al., 1994; Costanzo et al., 1995; Attaway et al., 1998). The problem that these findings create for proponents of an adaptive strategy based on supercooling can be reduced to a simple question: Why do animals tolerate freezing if such a tolerance is not important to their ecology? We believe that we have obtained a partial answer to the aforementioned question. Results from the first of two experiments reported here indicate that a tolerance for limited freezing is common to hatchling turtles generally, that is, that such tolerance probably was acquired from some common ancestor of turtles and is a shared (plesiomorphic) character in contemporary forms. The evidence was gathered by studying tolerance for freezing in hatchlings of four species of North American turtles-painted turtles, Blanding's turtles (Emydoidea blandingi), slider turtles (Trachemys scripta), and snapping This content downloaded from 207.46.13.129 on Sun, 26 Jun 2016 06:51:49 UTC All use subject to http://about.jstor.org/terms FREEZING IN HATCHLING TURTLES turtles (Chelydra serpentina)-that differ in geographic distribution and in the natural history of overwintering neonates. All the hatchlings in this experiment were frozen at -2 C and then held at that temperature for up to 48 h before their condition (i.e., alive or dead) was assessed. We also performed a second experiment, in which hatchling painted turtles were frozen at -2 C for 24 h and then exposed for varying intervals to -2 C, -3 C, or -4 C. Data from this second experiment provide new insights into the possible role of freeze-tolerance in enabling neonates of this species to survive winters in the field. MATERIALS AND METHODS Eggs of painted turtles and snapping turtles were collected in June 1996 from newly constructed nests located on the Valentine National Wildlife Refuge in north-central Nebraska, whereas eggs of Blanding's turtles were collected that same month from fresh nests located along the Mississippi River in central Minnesota. Eggs of sliders, on the other hand, were obtained by injecting gravid females from westcentral Illinois with synthetic oxytocin to induce them to oviposit. The eggs of all species were incubated to hatching on damp vermiculite. Hatchling Blanding's turtles and sliders later were transported to Colorado State University by air express. Average mass for hatchlings was 3.7 g (SD = 0.6 g) for painted turtles; 7.7 g (SD = 0.4 g) for Blanding's turtles; 6.0 g (SD = 1.1 g) for slider turtles; and 7.8 g (SD = 1.1 g) for snapping turtles. The turtles were acclimated to approximately 2 C and held in darkness until they were used to study their tolerance for freezing (Packard and Packard, 1993a). Samples of 8-12 turtles were drawn at random from the pools of available animals. Each hatchling in a sample then was prepared for study by cleaning it thoroughly with a small paint brush, after which the measuring junction of a copper/constantan thermocouple (26 gauge) was affixed to its carapace with epoxy resin. When the resin had hardened, the turtle was dipped in chilled (2 C) tap water to wet its integument-especially the skin in the axillary and inguinal pouches, because this integument in painted turtles may be more susceptible than skin elsewhere on the body to penetration by growing crystals of ice (Packard and Packard, 1995b). Each turtle was placed immediately into a pint-volume canning jar, and the sealed jar was placed in an environmental chamber set at approximately 2 C. After all jars were in the chamber, free ends of the thermocouples were attached via a multiplexor to a Campbell CR-10 datalogger so that temperature on the carapace of the animals could be measured every 30 sec. We next allowed temperature of the turtles to equilibrate with that of the environmental chamber. We then opened the chamber, removed the lid from each jar, placed several pieces of crushed ice against the turtle, and closed the jar. After ice had been added to all the jars, the door to the environmental chamber was closed and the microprocessor controlling the chamber was reset to a nominal temperature of -2 C. Temperature in the chamber declined rapidly, and so too did surface temperature of the turtles (Fig. 1). When temperature reached approximately -0.4 C, ice crystals contacting each turtle caused water on the integument to begin to freeze, as evidenced by a slight decrease in slope of the cooling curve brought about by the release of latent heat of fusion by water changing phase from liquid to solid (Fig. 1). In most cases, the turtle was inoculated shortly thereafter, as evidenced by a pronounced freezing exotherm (i.e., the sudden increase in temperature caused by the release of latent heat of fusion) that lasted for 12-15 h (Fig. 1A, C, D). No freezing exotherm was detected in the temperature profiles for three of the painted turtles (Fig. 1B), so these animals apparently avoided inoculation and remained unfrozen for the duration of their exposure. Timing of each exposure began with the appearance of the freezing exotherm for the turtle itself (Fig. 1A, C, D). Animals in the first experiment comparing hatchlings of all four species were frozen at approximately -2 C for 24-48 h. Some variation occurred in the equilibrium temperatures that actually were measured on the carapace of animals in different jars, owing to the spatial variation in temperature that commonly occurs in environmental chambers (Measures et al., 1973). The average temperature across species and exposures was -2.1 C (range -1.8 C to -2.6 C) and did not vary among species or treatment. Painted turtles in the second experiment were frozen first at -2 C for 24 h. Some of the hatchlings then were allowed to remain at this temperature for another 48 or 72 h whereas animals in other groups were subsequently cooled to minima of -3 (range -2.8 C to -3.2 C) or -4 C (range -3.8 C to -4.2 C). Rates of cooling in the latter tests were 1 C/day or 2 C/day (see Fig. 3). Temperature in the chamber was increased to 2 C at the end of each test and the animals were allowed to thaw overnight. The condition of the turtles (i.e., alive or dead) then was assessed by examining their eyes (which often are only partially open in moribund animals) and by observing their spontaneous activity and respons537 This content downloaded from 207.46.13.129 on Sun, 26 Jun 2016 06:51:49 UTC All use subject to http://about.jstor.org/terms

Temperature during embryonic development irreversibly determines gonadal sex in many oviparous reptiles. Although embryonic temperature also influences a number of other traits in these species, it is unclear whether such effects are primarily due to incubation temperature or to gonadal sex. Here we dissociated these normally confounding effects via hormonal manipulations of snapping turtle embryos (Chelydra serpentina), a species with temperature-dependent sex determination. We then ascertained temperature and sex effects on residual yolk mass, abdominal fat body mass, and total mass in neonates shortly after hatching or after one month without feeding. Yolk mass was initially affected by incubation temperature, clutch identity. and their interaction, but not gonadal sex. Yolk mass was not influenced by any of these variables 30 d after hateching. Fat mass was initially affected by temperature and clutch, but not sex. Gonadal sex did, however, affect fat mass at 30 d of age as did temperature and clutch. Incubation temperature affected total mass in a complex manner. There were significant clutch by temperature interactions initially, but no main effect of temperature. These effects changed so that the temperature effect was significant but interaction effects were not detected at 30 d of age. Repeated measures analysis of total mass after hatching and at 30 d of age indicated that turtles from lower temperatures lost more mass than turtles from higher incubation temperatures. and that females lost more mass than males. These data indicate that some hatchling traits were influenced directly by incubation temperature, whereas others were independently affected by gonadal sex. These results are consistent with a leading hypothesis for the evolution of temperature-dependent sex determination.