Journal of Herpetology, Vol. 48, No. 4, 466â€“470, 2014 Copyright 2014 Society for the Study of Amphibians and Reptiles Temperature-Dependent Sex Determination in the Diamond-backed Terrapin (Malaclemys terrapin) RUSSELL L. BURKE1 AND ARTHUR M. CALICHIO Department of Biology, Hofstra University, Hempstead, New York 11549 USA ABSTRACT.â€”Most Emydidae, including Diamond-backed Terrapins (Malaclemys terrapin), have temperature-dependent sex determination. However, the full relationship between incubation temperature and offspring sex in the Diamond-backed Terrapin has not been reported, and the pivotal temperature, range of transitional temperatures, and the temperature-sensitive period have not been investigated fully. Here we report on these parameters, comparing our data with data from other emydid turtles, and also comparing our laboratory data to field data collected previously for this population. RESUMEN.â€”La mayorÄ±Â´a de Emydidae, incluyendo a las Diamond-backed Terrapins (Malaclemys terrapin), tienen determinacioÂ´n de sexo por temperatura. Sin embargo, la relacioÂ´n entre temperatura de incubacion y sexo de crÄ±Â´as en el Diamond-backed Terrapin no ha sido reportada, ni el umbral de temperatura ni el perÄ±Â´odo sensible a la temperatura han sido investigados. AquÄ±Â´ reportamos ambos fenoÂ´menos y los compararmos con otras Emydidae, asi como con datos de campo de eÂ´sta poblacioÂ´n previamente colectados. In their landmark paper, Ewert and Nelson (1991) reported sex determination data for a wide range of turtle species, laying the groundwork for multispecies analyses of sex determination patterns in reptiles. Of particular interest was the Emydidae, wherein both temperature-dependent sex determination (TSD) and genotypic sex determination (GSD). Among TSD turtles, there are two patterns known, TSD Ia and TSD II, but their evolutionary and ecological significance remains enigmatic (Ewert et al., 2004). Although Ewert and Nelson (1991) reported detailed infor- mation for many emydids, their data for Diamond-backed Terrapins (Malaclemys terrapin) were limited. Their article showed only that Diamond-backed Terrapins have TSD, with all female hatchlings from eggs (Accomac County, Virginia; Hotaling, pers. comm.) incubated at 308C and all male hatchlings from eggs (from Bergen County, New Jersey, E. Hotaling, pers. comm.) incubated at 248C. Later, Jeyasuria et al. (1994) incubated Diamond-backed Terrapin eggs from St. Maryâ€™s County, Maryland, and found that high temperatures (31 and 328C) resulted in 100% female hatchlings and low incubation temperatures (26 and 278C) resulted in 100% male hatchlings and that eggs incubated at 29.58C resulted in 13% male hatchlings. Thus, Diamond-backed Terrapins seem to be a typical TSD Ia species (Ewert and Nelson, 1991), with a single transitional temperature range; incubation temperatures below this range result in males, whereas incubation temperatures above this range produce females. The TSD I pattern is evident in most but not all emydid turtles. However, without sex ratio data from a wide range of temperatures, it is possible that a second transitional range exists, as is seen in TSD II species (Ewert and Nelson, 1991). Pivotal (=threshold; Bull et al., 1982) temperature is the constant incubation temperature that results in 1:1 offspring sex ratios in TSD species (Mrosovsky and Pieau, 1991). This value has become an important component of models that relate incubation temperatures to offspring sex ratios (e.g., Georges et al., 1994; Doody et al., 2006) and is a common repeatable metric for intraspecific (e.g., Bull et al., 1982; Ewert et al., 2005) and interspecific (Sarre et al., 2004) comparisons. Less frequently calculated, yet perhaps as important for understanding TSD (e.g., Hulin et al., 2009), is the transitional range of temperature (TRT). The TRT is the range of constant incubation temperatures that produce both sexes and may be associated with the likelihood of evolutionary response to selective pressures because of changing thermal environments (Hulin et al., 2009). Girondot (1999) pointed out that calculation of TRT depends on statistical technique, number of incubation temperatures, and samples sizes used. There is significant genetic structuring across the Diamond- backed Terrapin range (Hart, 2005; Hauswaldt and Glen, 2005) that spans 2,000 km northâ€“south and could include substantial differences in incubation conditions. Thus, different populations may have evolved different pivotal temperatures (Bull et al., 1982; Ewert et al., 1994, 2005) and different TRTs. Unfortunately, Jeyasuria et al. (1994) did not report obtaining males and females at more than a single incubation temperature, so their data cannot be used to estimate pivotal temperature or TRT (Girondot, 1999). The temperature-sensitive period (TSP) (Bull and Vogt, 1981) is that portion of the incubation period during which incubation temperature can affect the eventual sex of hatchlings. Identify- ing the TSD for a species is critical for models that are used to explore the relationship between temperatures in natural nests and the sex of resulting offspring (e.g., Georges et al., 2004). The TSP has not been reported for Diamond-backed Terrapins, although Jeyasuria et al. (1994) and Jeyasuria and Place (1997) estimated the TSP for Diamond-backed Terrapins to start at embryonic stage 12â€“14, both without explanation. The TSP has been reported for four other emydid turtles: European Pond Turtle (Emys orbicularis) (stages 16â€“21; Pieau and Dorizzi, 1981), Red-eared Slider (Trachemys scripta scripta (stages 16â€“19 or 20; Wibbels et al., 1991), Ouachita Map Turtle (Graptemys ouachi- tensis) (stages 16â€“22; Bull and Vogt, 1981); and Painted Turtle (Chrysemys picta) (stages 16â€“22; Bull and Vogt, 1981). All five of these studies used Yntema (1968) to identify embryonic stages, so their results are comparable. Our goals were to 1) determine whether the Diamond-backed Terrapin is a TSD type Ia species as is common in emydid turtles, 2) calculate the pivotal temperature of sex determination and TRT for the northern Diamond-backed Terrapin, 3) explore geographic variation in pivotal temperature of sex determina- tion and TRT in Diamond-backed Terrapins, and 4) test the hypothesis of Jeyasuria et al. (1994) regarding the TSP and 1Corresponding Author. E-mail: email@example.com DOI: 10.1670/13-188 developmental stage for Diamond-backed Terrapins. We ex- plored these phenomena in Diamond-backed Terrapins because of our investigations into the relationship between temperatures in wild nests and offspring sex ratios (e.g., Scholz, 2007) in this species. We used constant temperature incubation temperatures to address goals 1â€“3 and shift-temperature experiments to address goal 4. MATERIALS AND METHODS We collected Diamond-backed Terrapinsâ€™ freshly laid eggs from nests in 2009 and 2010 on Rulers Bar Hassock, a 463-ha island (408570N, 738500W) in the mainly estuarine Jamaica Bay Wildlife Refuge. This site is located in Kings and Queens counties, New York, and near the center of the range of the Northern Diamond-backed Terrapin subspecies Malaclemys terrapin terrapin. In both years, we transported eggs to the laboratory; there they were marked and mass was determined. Clutches were divided so as to maximize representation of clutches among treatments. They were buried in vermiculite that had been mixed 2 parts distilled water:1 part vermiculite by weight and placed in constant temperature incubators within 48 h of oviposition. The mass of eggs and containers was determined weekly during incubation, and evaporated water was replaced. In 2009, we incubated eggs at constant temperatures (24, 25, 26, 27, 28, 29, 30, 31, and 328C) until hatching. We added data for eggs incubated at constant 348C from Giambanco (2003) where applicable; Giambanco collected eggs in 2000 from the same population and used the same methodology described here for constant-temperature experiments. After pipping, eggs were removed from the incubators and hatchlings were allowed to emerge fully from the eggs. Hatchlings were then euthanized via chemical injection 3â€“4 d postemergence. Sex was identified by gross inspection of gonads using the presence or absence of MuÂ¨llerian ducts (indicating female and male, respectively), and both microscopic and macroscopic examinations of the external appearance of the gonads. In 2010, we collected eggs and incubated them as described above, except we conducted shift experiments (Valenzuela, 2001) to identify the TSP. We used baseline temperatures of 25 and 318C. Eggs were held at one of the single baseline temperatures through most of incubation, but they were shifted to a different temperature for a specific length of time (Figs. 1 and 2). The baseline temperatures for the shift experiments (25 and 318C) were chosen as likely to be male producing and female producing, respectively, based on results from 2009 experiments. We predicted incubation duration for the shift experiments based on results from 2009 experiments, and we divided the predicted incubation duration for the 318C baseline into 5-d intervals and the predicted incubation duration for the 258C baseline into 7-d intervals (Figs. 1 and 2). At each time interval when eggs were moved from their assigned baseline incubation temperature to the shift temperature, five eggs were chosen randomly for preservation. With these eggs, we used a bright light to locate an area within the shells without blood vessels; there, we used forceps to peel back the eggshells and expose the embryos. The embryos were removed and placed in Bouinâ€™s solution for preservation. Preserved embryos were staged according to Yntema (1968) and Greenbaum (2002). We analyzed the constant temperature and sex of offspring data (including that of Giambanco ) to assess sex determination mode and estimate both TSP and TRT with the data from Diamond-back Terrapins from New York and Maryland (Jeyasuria et al., 1994) using the program TSD 4.0.3 (Girondot, 1999, 2012; Godfrey et al., 2003). Sex ratio data were fitted using several models (GSD, logistic, Hill, and Richards) using maximum likelihood with binomial distributions. The GSD model for genotypic sex determination is a model in which sex ratio has a predicted value of 0.5 and sex ratio is independent of incubation temperature. The logistic and Hill models both use a sigmoidal equation with two parameters, whereas the Richards model uses a sigmoidal equation with three parameters. Goodness of fit testing was based on likelihood ratios between the saturated model (df1 = 0) and fitted model (df2 = [number of temperatures] â€“ [number of estimated parameters]). Comparisons between different models were conducted using Akaike Information Criterion (AIC), a method to rank different models by order of merit by penalizing for too many parameters. Akaike weight is a method that converts AIC into probabilities assessing whether each model is really the best choice among the tested models (Burnham and Anderson, 2013). RESULTS The GSD model can be excluded from fitting the New York data on the basis of a very high AIC ranking and very low weight (goodness of fit < 0.001, Table 1). The logistic and Hill FIG. 1. Shift experiments with 258C baselines, indicating shift periods and resulting sex ratios (male:female). TSD IN DIAMOND-BACKED TERRAPINS 467 models both performed better than the Richards model, based on AIC ranking and Akaike weight. However, although the logistic and Hill models cannot be distinguished, we used the logistic based on parsimony because the Hill model is a special case of the logistic model with the x-axis being log transformed. Goodness of fit testing for New York data rejected the model as fitting the data correctly (p = 10-5). Observations at four incubation temperatures were particularly unlikely, such as the male observed at 348C. However, even if this observation is removed from the data set, the model is still rejected (p = 2.10-3), because there are significant variations in sex ratio among incubation temperatures that are not well represented by this model (Fig. 3). This could be due to fluctuation or nonstability of temperature within incubators or maternal effects. The pivotal temperature for the New York data was 28.168C and the TRT was 25.88â€“30.458C (Fig. 3). Incubation duration at constant 258C averaged 69.4 d (SD = 2.0); incubation duration at constant 318C averaged 43.2 d (SD = 1.3). Only two incubation temperatures are available for the New Jersey and Virginia populations, and none of these temperatures produced mixed sex ratios. With such a situation, it was impossible to fit these data to any model, and it did not make sense to add them to our other data sets. The GSD model can be excluded from fitting the Maryland data on the basis of a very high AIC ranking and very low Akaike weight (goodness of fit < 0.001; Table 1). The logistic model (goodness of fit = 1.00) performed better than the Hill or Richards models, based on AIC ranking and Akaike weight. The pivotal temperature for the Maryland data was 29.258C and the TRT was 28.85â€“29.648C. The combined data set (New York data and Maryland data) had a high AIC value and low Akaike weight (Table 1), especially in comparison with the two data sets analyzed separately (Table 1). This indicates significant differences between the data sets (goodness of fit < 0.001). Our shift experiment with the 258C baseline indicated that the TSP occurred between incubation days 28 and 35, 40.3â€“50.4% of incubation (Fig. 1). All five embryos preserved from 258C baseline after 28 days of incubation were at developmental stage 16 as described in Yntema (1968) and Greenbaum (2002). FIG. 2. Shift experiments with 318C baselines, indicating shift periods and resulting sex ratios (male:female). TABLE 1. Comparison of model fits among the populations of M. terrapin using AIC and Akaike weight. Population Model AIC Akaike weight New York GSD 300.01 2.785 Â· 10-55 New York Logistic 50.52 4.165 Â· 10-1 New York Hill 50.46 4.303 Â· 10-1 New York Richards 2.07 1.532 Â· 10-1 Maryland GSD 176.141 8.44 Â· 10-38 Maryland Logistic 7.129 4.237 Â· 10-1 Maryland Hill 7.14 4.20 Â· 10-1 Maryland Richards 9.13 1.56 Â· 10-1 New York and Maryland combined Logistic 69.18 3.13 Â· 10-3 New York and Maryland separate Logistic 57.65 9.97 Â· 10-1 FIG. 3. Relationship between constant incubation temperature and offspring sex (percentage of males) in the Diamond-backed Terrapin. Plotted points indicate actual data; curves indicate the Hill equation (dashed line) fit to data bracketed by confidence interval of 2 SDs (solid lines). 468 R. L. BURKE AND A. M. CALICHIO The shift experiment with the 318C baseline indicated that the TSP occurred between incubation days 30 and 35, 69.4â€“81.0% of incubation (Fig. 2). Four of the five embryos preserved from 318C baseline after 30 d of incubation were malformed, and we could not stage them reliably. The remaining embryo was at developmental stage 23. The embryos from the previous shift (removed at day 25) were at developmental stage 22; the embryos from the subsequent shift (removed at day 35) were at developmental stage 24. DISCUSSION We found that Diamond-backed Terrapins have TSD Ia pattern, as originally suggested, based on limited data, by Ewert and Nelson (1991). The pivotal temperature for the New York population is 1.098C cooler than Diamond-backed Terra- pins from populations from Maryland. These results span the two most northern of the six genetically based management units delineated by Hart (2005) and a range of ~400 km, which is only ~20% of the Diamond-backed Terrapin northâ€“south distribution. Because of the large fraction of the Diamond-back Terrapin range for which TSD data are unknown, there is still considerable possibility for even more variation in pivotal temperature and TRT in this species, as has been demonstrated in other turtles (Bull et al., 1982; Ewert et al., 1994, 2004, 2005). However, our results indicate that pivotal temperatures and TRT for both populations of Diamond-backed Terrapins are within the range of other emydid turtles (28â€“298C and 1â€“28C, respectively, e.g., Pieau, 1976; Bull et al., 1982; Crews et al., 1991; Ewert and Nelson, 1991; Ewert et al., 2004; Hulin et al., 2009). Our finding that the TSP for Diamond-backed Terrapins corresponds with embryonic stages 16â€“23 is not surprising because it is consistent with other emydid turtles (Bull and Vogt, 1981; Pieau and Dorizzi, 1981; Wibbels et al., 1991). Mrosovsky and Pieau (1991) suggested that the TSP in turtles occurred in the middle third of incubation. Although our findings confirm this for eggs incubated at the 258C baseline, the TSP occurred in the last third of incubation for eggs incubated at 318C. The cause of this striking variation is unclear, and we are unaware of similar results in any other species. In all cases, the same guide to staging embryos (Yntema, 1968) was used. Our results have potentially valuable applications. For example, Scholz (2007) recorded temperatures in 82 Diamond- backed Terrapin nests throughout incubation at our study site in 2004. Average temperatures for these nests varied considerably (18.9â€“27.08C, mean = 25.28C, SD = 1.6). Therefore, these nests apparently spent relatively little time near or above the pivotal temperature and were probably biased strongly toward male hatchlings. Further analysis is needed to investigate the relationships between constant temperature incubation results reported here and variable temperature incubation patterns observed in the field. Acknowledgments.â€”We are grateful to B. Clendening, J. Joseph, and C. Peterson for assistance with techniques and data collection. E. Hotaling shared unpublished data, M. Godfrey helped interpret TSD output, M. Girondot was extraordinarily helpful with analytical assistance and use of his model, and L. Rodriguez helped with the Spanish translation. This work was permitted under New York State License to Collect and Possess #383 and the Hofstra University Institutional Animal Care and Use Committee. LITERATURE CITED BULL, J. J., AND R. C. VOGT. 1981. Temperature-sensitive periods of sex determination in emydid turtles. Journal of Experimental Zoology. 218:435â€“440. BULL, J. J., R. C. VOGT, AND C. J. MCCOY. 1982. Sex determining temperatures in turtles: a geographic comparison. Evolution 36: 326â€“332. BURNHAM, K. P., AND D. R. ANDERSON. 2013. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. 2nd ed. Springer, USA. CREWS, D., J. J. BULL, AND T. WIBBELS. 1991. Estrogen and sex reversal in turtles: a dose-dependent phenomenon. General and Comparative Endocrinology 81:357â€“364. DOODY, J. S., E. GUARINO, A. GEORGES, B. COREY, G. MURRAY, AND M. EWERT. 2006. Nest site choice compensates for climate effects on sex ratios in a lizard with environmental sex determination. Evolutionary Ecology 20:307â€“330. EWERT, M., AND C. NELSON. 1991. Sex determination in turtles: diverse patterns and some possible adaptive values. Copeia 1991:50â€“69. EWERT, M. A, C. R. ETCHBERGER, AND C. E. NELSON. 2004. Turtle sex- determining modes and TSD patterns, and some TSD pattern correlates. In N. Valenzuela and V. Lance (eds.), Temperature- dependent Sex Determination in Vertebrates, pp. 21â€“32. Smithsonian Books, USA. EWERT, M. A., D. R. JACKSON, AND C. E. NELSON. 1994. Patterns of temperature-dependent sex determination in turtles. Journal of Experimental Zoology 270:3â€“15. EWERT, M. A., J. W. LANG, AND C. E. NELSON. 2005. Geographic variation in the pattern of temperature-dependent sex determination in the American snapping turtle (Chelydra serpentina). Journal of Zoology 265:81â€“95. GEORGES, A., C. LIMPUS, AND R. STOUTJESDIJK. 1994. Hatchling sex in the marine turtle Caretta caretta is determined by proportion of development at a temperature, not daily duration of exposure. Journal of Experimental Zoology 270:432â€“444. GEORGES, A., S. DOODY, K. BEGGS, AND J. YOUNG. 2004. Thermal models of TSD under laboratory and field conditions. In N. Valenzuela and V. Lance (eds.), Temperature-dependent Sex Determination in Verte- brates, pp. 79â€“89. Smithsonian Books, USA. GIAMBANCO, M. R. 2003. Comparison of Viability Rates, Hatchling Survivorship and Sex Ratios of Laboratory- and Field-incubated Nests of the Estuarine Emydid Turtle Malaclemys terrapin. Unpubl. M.S. thesis, Hofstra University, USA. GIRONDOT, M. 1999. Statistical description of temperature-dependent sex determination using maximum likelihood. Evolutionary Ecology Research 1:479â€“486. â€”â€”â€”. 2012. TSD (Version 3.0.4). CNRS et UniversiteÂ´ Paris-Sud, UMR- 8079-Ecologie, SysteÂ´matique et Evolution, Department dâ€™Ecologie des Populations et des CommunauteÂ´s, Groupe de Conservation des Populations et des communauteÂ´s, France. GODFREY, M. H., V. DELMAS, AND M. GIRONDOT. 2003. Assessment of patterns of temperature-dependent sex determination using maxi- mum likelihood model selection. Ecoscience 10:265â€“272. GREENBAUM, E. 2002. A standardized series of embryonic stages for the emydid turtle Trachemys scripta. Canadian Journal of Zoology 80: 1350â€“1370. HART, K. M. 2005. Population Biology of Diamond-backed Terrapins (Malaclemys terrapin): Defining and Reducing Threats across Their Range. Ph.D. diss., Duke University, USA. HAUSWALDT, J. S., AND T. C. GLEN. 2005. Population genetics of the diamond-backed terrapin (Malaclemys terrapin). Molecular Ecology 14:723â€“732. HULIN, V., V. DELAMAS, M. GIRONDOT, M. H. GODFREY, AND J.-M. GUILLON. 2009. Temperature-dependent sex determination and global climate change: are some species at greater risk? Oecologica 160:493â€“506. JEYASURIA, P., AND A. R. PLACE. 1997. Temperature-dependent aromatase expression in developing diamondback terrapin (Malaclemys terra- pins) embryos. Journal of Steroid Biochemistry and Molecular Biology 61:415â€“425. JEYASURIA, P., W. M. ROOSENBURG, AND A. R. PLACE. 1994. Role of P-450 aromatase in sex determination of the diamondback terrapin, Malaclemys terrapin. Journal of Experimental Zoology 270:85â€“111. MROSOVSKY, N., AND C. PIEAU. 1991. Transitional range of temperature, pivotal temperatures and thermosensitive stages for sex determina- tion in reptiles. Amphibia-Reptilia 12:169â€“179. PIEAU, C. 1976. DonneÂ´es reÂ´centes sur la diffeÂ´renciation sexuelle en function de la temperature chez les embryons dâ€™Emys orbicularis L. TSD IN DIAMOND-BACKED TERRAPINS 469 (CheÂ´ lonien). Bulletin of the Zoological Society of France 101(Suppl.)4:46â€“53. PIEAU, C., AND M. DORIZZI. 1981. Determination of temperature sensitive stages for sexual differentiation of the gonads in embryos of the turtle (Emys orbicularis). Journal of Morphology 170:373â€“382. SARRE, S. D., A. GEORGES, AND A. QUINN. 2004. The end of a continuum: genetic and temperature-dependent sex determination in reptiles. Bioessays 26:639â€“645. SCHOLZ, A. L. 2007. Impacts of Nest Site Choice and Nest Characteristics on Hatchling Success in the Diamondback Terrapins of Jamaica Bay, New York. Unpubl. M.S. thesis, Hofstra University, USA. VALENZUELA, N. 2001. Constant, shift, and natural temperature effects on sex determination in Podocnemus expansa turtles. Ecology 82:3010â€“ 3024. WIBBELS, T., J. J. BULL, AND D. CREWS. 1991. Chronology and morphology of temperature-dependent sex determination. Journal of Experimen- tal Zoology 260:371â€“381. YNTEMA, C. L. 1968. A series of stages in the embryonic development of Chelydra serpentina. Journal of Morphology 125:219â€“252. Accepted: 12 January 2014. 470 R. L. BURKE AND A. M. CALICHIO Copyright of Journal of Herpetology is the property of Society for the Study of Amphibians & Reptiles and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.