An Approximation of the Sex Determination Parameters in the Savanna Side-Necked Turtle Podocnemis vogli (Podocnemididae)
ABSTRACT
With the exception of the savanna side-necked turtle (Podocnemis vogli) and the red-headed river turtle (Podocnemis erythrocephala), the species within the pleurodire genus Podocnemis have been relatively well-characterized in terms of their temperature-dependent sex determination (TSD) parameters. In this study, we incubated 4 nests of P. vogli (46 eggs) under controlled laboratory incubation conditions to determine whether this species also presents a TSD Ia pattern where males are produced at lower temperatures, characterize the constant temperature producing a 1:1 sex ratio (Tpiv), and document the transitional range of temperatures (TRT) that produces both sexes. The incubation temperatures employed were 29°, 31°, and 33°C (± 0.15°C). Our results were consistent with the conclusion that P. vogli exhibits a TSD Ia pattern, but it was not possible to document the Tpiv or the superior TRT, because the 29° and 31°C conditions produced all males and even the 33°C condition produced male-biased sex ratios, suggesting a high pivotal temperature for this species. We also documented maternal effects in initial size of the hatchlings, with the female that laid larger eggs producing hatchlings that were larger. Incubation periods of P. vogli were notably longer than those of its congeners incubated under comparable constant temperature and soil moisture conditions. In species of this genus, the TSD Ia pattern suggests that besides the threats these species face from overexploitation and habitat degradation, they also are highly vulnerable to the effects of global warming.
Temperature-dependent sex determination (TSD) is a form of environmental sex determination that is common in many taxonomic groups, including turtles (Ewert et al. 1994). In turtle species with TSD, eggs incubated under constant temperatures exhibit one of two patterns: pattern Ia, where low temperatures produce males and high temperatures produce females, and pattern II, where extreme temperatures produce females and intermediate temperatures produce more males (Abreu-Grobois et al. 2020). Pattern Ib, where low temperatures produce females and high temperatures produce males, is known in reptiles only in lizards (Lockley and Eizaguirre 2021).
In TSD turtle species, it is important to document the pattern of TSD exhibited and to estimate the pivotal temperature (Tpiv, the constant temperature that produces a population-wide 1:1 sex ratio; Mrosovsky and Pieau 1991), and the transitional range of temperatures (TRT, those constant temperatures that produce mixed sex ratios; Mrosovsky and Pieau 1991), because TSD species with pattern Ia are more vulnerable to global warming producing skewed sex ratios of 100% females in the population than are species with pattern II or non-TSD species (Lockley and Eizaguirre 2021). Similarly, pattern Ia species with a narrower TRTs may be more vulnerable to global warming than species with wider TRTs (Hulin et al. 2009; Mitchell and Janzen 2010; Valenzuela et al. 2019). Finally, many management programs incubate turtle eggs artificially and so must be aware of the TSD pattern and the Tpiv to be able to confidently produce individuals of both sexes (Frazer 1992).
These TSD parameters have been documented using constant or semiconstant temperatures in 5 of the 6 species of the genus Podocnemis (de Souza and Vogt 1994; Valenzuela 2001a; Vogt 2001, 2008; Páez and Bock 2004; Hulin et al. 2009; Páez et al. 2009; Camillo et al. 2022). All of these species have been shown to exhibit a TSD Ia pattern and variable TRTs (Gallego-García and Páez 2016; Camillo et al. 2022). Although not as frequently quantified, in these species the hatching success rate and other aspects of the hatching phenotype, such as size, shape, body growth rates, and behavior, may be influenced by different forms of maternal effects and their interactions with incubation temperature regimes (de Souza and Vogt 1994; Valenzuela 2001b; Escalona 2003; Páez et al. 2015a; Camillo et al. 2022). Under controlled artificial incubation experiments, preovulatory maternal effects could be explained by differences in endogenous energy sources for the offspring by means of yolk allocation, including the hormones, antioxidants, and antibodies it contains (Wibbles et al. 1991; Bowden et al. 2000; Steyermark and Spotila 2000; Elf 2003). In turtles there is frequently a positive relationship between egg size and hatchling size, where larger eggs containing more yolk produce larger hatchlings (Congdon et al. 1983; Finkler and Claussen 1997; Kaplan 1998; Páez et al. 2015a).
Little is known regarding how incubation temperature and preovulatory maternal effects affect the phenotype of hatchlings of the savanna side-necked turtle, Podocnemis vogli. This species is restricted to the Orinoco drainage of Venezuela and Colombia (Ramo 1980; Portocarrero-Aya 2008; Morales-Betancourt et al. 2012), with its area of occupation in Colombia covering only 304 km2 (Forero-Medina et al. 2014). In Colombia, P. vogli is classified as Least Concern (LC) based upon its apparent abundance (Morales-Betancourt et al. 2015). However, the exploitation of P. vogli has being reported in Colombia for more than a century (MADT 2002) and every ecological study of this species both in Colombia and Venezuela have documented its consumption (both meat and eggs) by indigenous groups and other rural people living in the Llanos (Ramo 1980; Castaño-Mora and Medem 2002; Lee 2004; Rodríguez 2009; Morales-Betancourt et al. 2012).
The goals of the present study were to determine the Tpiv and TRT under constant temperatures using nests from 2 populations of P. vogli, to evaluate possible maternal effects in size and initial growth rates of the hatchlings in the laboratory, and to compare our results with those in the literature on other congener species that used similar methodologies. We predicted that P. vogli would exhibit a TSD Ia pattern (like all other members of the genus) and have a lower Tpiv in comparison with other podocnemidids, as a result of its preference for nesting in savanna habitat under vegetation in clayish soils rather than on exposed river beaches.
METHODS
Field Procedures. — In November 2021, we transported 4 clutches of P. vogli from the Colombian Llanos to the Laboratory of Herpetology at the Universidad de Antioquia in Medellín, Colombia. Two nests were collected at Hacienda La Corocora in the municipality of San Carlos de Guaroa, Meta department (3°42′44″N, 73°15′19″W) in the westernmost portion of the Colombian Llanos. Two weeks later another 2 nests were taken from the Reserva Privada Bojonawi in the municipality of Puerto Carreño, Vichada department (9°09′59″N, 74°40′43″W), 700 km from the first site, on the western border of Colombia with Venezuela. In both sites, we performed diurnal and nocturnal walks along shorelines and into the savannas to locate either freshly laid nests or nesting females. In Hacienda Corocora, we found 2 fresh nests and extracted the eggs after adding a small amount of water over the soil. Two other nests were obtained in Bojonawi by following, from a distance, adult females that we had encountered on the savanna. We transported the eggs to the Laboratory of Herpetology of the Universidad de Antioquia in Medellín, Colombia, travelling by car from the Corocora site and by boat and airplane from the Bojonawi site.
Incubation and Rearing Procedures. — We measured the length and width of each egg to the nearest 0.1 mm with a digital caliper and egg mass with a digital balance (± 0.1 g; Lexus BN12–1200, Singapore). To control for family effects, the eggs from each nest were randomly assigned in approximately equal numbers to 1 of 3 digital BINDER incubators set to 29°, 31°, and 33°C. We selected these incubation temperatures based on the published mean pivotal temperatures of congeners (Table 1) because there are no data on incubation temperatures in natural nests for P. vogli. To control the temperature settings, we placed type K thermocouples at 2 different levels within each incubator and recorded temperatures daily by connecting the thermocouples to a precision thermometer (± 0.01°C; Fluke 54, Fluke Corporation, Everett, Washington, USA). We placed the eggs from each nest in compartments within plastic containers on a bed of humid vermiculite prepared in a 1:2 weight ratio of distilled water to vermiculate and then covered with plastic wrap. In addition, we weighed each container before placing it into the incubator and we added distilled water to each container every 2 wks during the incubation period to compensate for water lost as a result of evaporation. At this time, we also rotated the containers between shelves to reduce possible effects of temperature gradients within the incubators.
We recorded the day each neonate pipped and the day it emerged from its shell. In this study we define incubation period as the number of days from placing the eggs in the incubators to the date the hatchlings were completely out of the eggshell. On the day of hatching, we marked each hatchling by making individual notches in the marginal scutes and we measured each individual for straight-line carapace length (SCL) and body mass. We reared the hatchings in a 40-gallon (151-L) tank with a constant water temperature of 28°C and a 12:12-hr day–night light cycle using UVB lights and provided food (kale, spinach, and water hyacinth) ad libitum. The tanks contained platforms for basking and rock crevices as hiding sites. We reared the juveniles until they were approximately 60 d old and we then remeasured them for the same morphometric variables as before. We then sacrificed them with an injection of 2% xylocaine in the axial region. We sexed all sacrificed neonates using 2 methods. The first entailed inspection of the gross anatomy of the gonads and oviducts (Malvasio et al. 1999). Then, each hatchling was fixed in 10% buffered formalin and after 5 d, we removed one gonad for histological preparation and staining with hematoxylin and eosin, to corroborate the gross anatomy sex determinations. We made the histological sex diagnoses based upon the criteria of Danni and Alho (1985) and Estrada and Uribe (2002). We deposited the voucher specimens in the Museo de Herpetología of the Universidad de Antioquia (MHUA-R 13975 [01-21 San Carlos de Guaroa] -13976 [01-12 Puerto Carreño]).
Statistical Analyses. — We used regression analyses to inspect for relationships between egg mass and hatchling mass at emergence and at 2 mo of age. We used analysis of variance (ANOVAs; and when significant, Student’s post hoc t-tests among groups) to compare eggs sizes and masses of the nests, incubation periods by nest, incubation periods by incubation temperature, and differences among incubation temperatures in hatchling sizes at emergence, after 2 mo of growth, and mean hatchling growth rates. The egg mass-hatchling mass regression was significant, so we also compared differences among incubation temperatures in hatchling sizes at emergences using analysis of covariance (ANCOVA) with egg mass as a covariate. All statistical analyses were conducted using the JMP software package (JMP Statistical Discovery LLC/SAS Institute, Cary, North Carolina, USA).
RESULTS
Mean clutch size was 11.5 eggs (SD = 1) and 3 of the 4 nests exhibited high hatching success rates (mean = 89%). The fourth nest (from Bojonawi) produced only 3 hatchlings (25% hatching success rate) and 2 of them died within a week, so we excluded this nest from the statistical analyses (Table 2). Mean egg length, width, and mass were significantly different among the remaining 3 clutches (ANOVAs; egg length, F2,27 = 11.34; egg width, F2,27 = 49.82; egg mass, F2,27 = 24.36; all p-values < 0.001), with Nest 3 (from Bojonawi) having significantly larger and heavier eggs than did Nests 1 and 2 (Table 2). Egg mass (all nests pooled) was positively related to hatchling mass at emergence (r2 = 0.246, p < 0.010), but not to body mass after 2 mo of growth (r2 = 0.096, p > 0.10). There were significant differences among the 3 nests in hatchling mass (F2,27 = 15.23, p < 0.001), again with Nest 3 producing hatchlings that were heavier than those from Nest 1 and Nest 2 (Table 2). However, after 2 mo of growth under laboratory conditions, mass differences among juveniles from the 3 nests were no longer apparent (F2,25 = 2.48, p > 0.10; Table 2).
Of the 33 hatchings sexed, 30 (93%) were males, with 2 nests producing female hatchlings at 33°C. Incubation periods differed among incubation temperature conditions (F2,27 = 10.40, p < 0.001) but were extremely variable (range: 82–159 d), both within nests and among incubation temperatures (mean = 110.7 d; SD = 20.8; Table 3; Fig. 1). Despite dividing eggs from all 3 nests evenly among the 3 temperature regimes, over all incubators, Nest 3 (from Bojonawi) exhibited longer incubation periods (F2,27 = 6.76, p < 0.005). Incubation periods of all 3 nests in each temperature regime were longer than those reported for other podocnemidids incubated under similar temperature and humidity regimes (Fig. 2).
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World’s Turtle and Tortoise Journal 22, 2; 10.2744/CCB-1565.1
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World’s Turtle and Tortoise Journal 22, 2; 10.2744/CCB-1565.1
At emergence, hatchlings had a mean body mass of 12.4 g (SD = 0.7) and a mean SCL of 3.8 cm (SD = 0.7). Each incubator contained roughly equal numbers of eggs from each nest, so there were no differences among incubators in egg mass (F2,27 = 0.08, p = 0.92); and at hatching, mass and SCL also did not differ significantly among the 3 temperature regimes (ANOVA, mass, F2,27 = 0.43, p > 0.10; SCL, F2,27 = 1.37, p > 0.10; Table 3). Hatchling mass also did not differ among the 3 temperature conditions after controlling for the effects of egg mass (ANCOVA, F3,26 = 3.39, temperature effect, p > 0.10, egg mass effect, p < 0.01). However, after 2 mo of growth, there were significant differences in mass (F2, 25 = 11.42, p < 0.001; Fig 3) and SCL (F2,25 = 11.94, p < 0.001), with the hatchlings obtained from the coolest temperature condition larger and heavier than those from the other 2 conditions. This result was due to significant differences in growth rates among the 3 groups for both body mass (F2,25 = 12.45, p < 0.001; Table 3) and SCL (F2,25 = 14.47, p < 0.001; Table 3).
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World’s Turtle and Tortoise Journal 22, 2; 10.2744/CCB-1565.1
DISCUSSION
Prior to this study, information on the sex determination parameters of P. vogli was lacking, but we now know that all 6 species in the genus Podocnemis exhibit TSD, apparently with a TSD Ia pattern with a high pivotal temperature in comparison with other turtle species (Valenzuela 2001a; Vogt 2001, 2008; de Souza and Vogt 1994; Páez and Bock 2004; Hulin et al. 2009; Páez et al. 2009; Gallego-García and Páez 2016; Camillo et al. 2022; present study). We could not characterize the entire TSD reaction norm for this species, but we documented a masculinizing temperature of 31°C and produced results that suggest the species has a Tpiv presumably slightly higher than 33°C, because at this regime we obtained a 7:3 male–female ratio. Two of the 3 nests with high hatching success (>91.6%) produced at least one female at 33°C, but we still do not know whether there is interpopulational variation in Tpiv or TRT in this species, or if this difference among our nests was the result of maternal manipulation of the sensitivities of the embryos via differential allocation of endogenous hormones in the yolk (Bowden et al. 2000; Elf 2003; Ewert et al. 2004). For example, in Trachemys scripta, it has been shown that nests oviposited later in the 5-wk nesting season have yolks with greater hormonal investments (Carter et al. 2017). Podocnemis vogli exhibits a 4-mo nesting season (Ramo 1980; Sepúlveda et al. 2020), so it is possible that the 2-wk difference in oviposition dates of the 4 nests from the 2 study sites also influenced the sex ratios obtained using the same constant laboratory incubation conditions. However, at both sites, we spent several weeks searching for nests before finding the two we collected, suggesting that in both sites (separated by 700 km) we obtained some of the first nests oviposited that season.
Quantifying levels of intra- and interpopulational variation in TSD parameters is important if we want to characterize them at the species level, not only because they might vary between populations (Bobyn and Brooks 1994; Ewert et al. 2005; Gallego-García and Páez 2016), but also particularly because for podocnemidid species, one of the most common management strategies is ex situ egg incubation or nest transfer (Páez et al. 2015b). Also, climate change could have a rapid and drastic negative effect on their long-term population viability (Valenzuela et al. 2019) by feminizing sex ratios. The atypically long incubation periods we documented in this study for P. vogli also mean its clutches are not only exposed to nest predators for more time, but also are more vulnerable to the effects of periodic heatwave exposures during incubation that would further skew sex ratios toward females (Carter et al. 2018, 2019; Breitenbach et al. 2020).
The temperature conditions experienced by the nests affected not only their sex ratios but also their hatching success rates, incubation periods, and fitness-related aspects of the hatchling phenotypes. Hatching success was above 60% for all incubation temperatures and highest (87.5%) at our lowest (29°C ± 0.15) incubation regime. This hatching success rate is similar to that reported for P. lewyana at 29.5°C (Páez et al. 2009) and at 29.14°C for P. sextuberculata (Camillo et al. 2022), suggesting that the temperature regimes used in our experiment were within the natural incubation temperatures experienced by the species (and supported by preliminary unpublished field data, V.P.P., pers. obs.). The mean incubation period per temperature regime documented in this study was longer than those reported for other podocnemidids incubated under similar temperature and humidity regimes (Fig. 2). In fact, all other species in the genus exhibit incubation periods at these 3 temperatures of < 90 d, while P. vogli always exhibited incubation periods > 90 d, even at 33°C. Given that the incubation season of this species is long, beginning at the end of October and extending to February or March, it is possible that this species exhibits diapause (Kuchling and Hofmeyr 2023), allowing the embryos of nest laid early in the nesting season to delay their emergence to coincide with the rainy season (and also to subsist on yolk reserves while remaining in the nest chamber after hatching for over a week; Ramo 1980). What was surprising from our results was the variation in incubation periods among siblings incubated at the same temperatures, considering most species exhibit hatching synchrony under these conditions, or even synchrony in cases where eggs incubated together are in different developmental stages at the start of incubation (Colbert et al. 2010; McGlashan 2015). Studies of the thermal and humidity variation of natural nest of this species are necessary to better understand its thermal biology, as are studies of nests laid in the latter portion of the nesting season.
Although there were no significant differences in size or mass of recently hatched neonates that were incubated under the different temperature regimes, the sizes attained after 60 d of growth showed that the juveniles from the 2 coolest incubator conditions had attained larger body sizes than individuals from the warmest treatment. In our case, the unexpected result of obtaining only males from 2 of the 3 incubation temperatures and a 7:3 male:female sex ratio at the warmest temperature allowed us to largely exclude sex as a potential confounding variable in this examination of the effects of temperature on initial hatchling size, mass, and subsequent growth rates. The direction of the effect of incubation temperature on podocnemidid hatching size and growth rates varies between species and populations, but most frequently, lower or intermediate incubation temperatures produce larger and or heavier hatchlings (Páez and Bock 2004; Páez et al. 2009; Ceballos et al. 2014; Camillo et al. 2022). Yet comparing these results between species from an ecological or evolutionary perspective could be misleading because other microclimatic factors such as the variation in incubation temperature or the humidities of the incubation substrates also could affect hatchling phenotypes (Gutzke et al. 1987; Deeming and Ferguson 1989; Packard 1989; Du and Ji 2003), as could rearing conditions (Ji et al. 2003). Regardless of the multiple potential influences affecting the documented effects on hatching size and growth, it is clear that both dependent variables probably influence important fitness components for these species (often interacting in unpredictable ways) and thus should not be ignored, or manipulated carelessly, in management programs. In the present study, we observed maternal effects on egg mass and initial hatchling mass and size, with differences highly associated with egg mass, which varied among nests. Further study is needed to determine the exact nature of more subtle additional maternal effects (Radder 2007).
Although the present study did not provide all the information that is lacking for this species in Table 1, our data suggest that P. vogli is a type Ia TSD species with a Tpiv that is probably slightly higher than 33°C. Most surprising were the long and variable incubation periods exhibited by the eggs, relative to those of congener species, presumably related to the different nesting habits of P. vogli. Future studies of TSD in P. vogli should attempt to document its TRT, examine the effects of fluctuating temperatures on sex determination patterns, and examine the factors influencing the variable incubation periods and lack of hatching synchrony in this species.
Variation in egg incubation periods among the 3 incubation temperatures for each nest.
Mean incubation periods at 29°, 31°, and 33°C for 5 species of Podocnemis turtles. Values were from this study (for P. vogli), from published accounts that used these 3 temperatures (for P. lewyana: Páez et al. 2009, 2015a; Gómez-Saldarriaga et al. 2016; for P. unifilis: Páez and Bock 1998, 2004; P. expansa: Lubiana and Ferreira-Júnior 2009; Ceballos et al. 2014), or from interpolations based on regressions of temperature vs. incubation periods using other temperature values (P. sextuberculata: Camillo et al. 2022).
Egg mass, neonate mass, and juvenile mass (at 2 mo of age) for the 3 incubation conditions.
Contributor Notes
Handling Editor: Peter V. Lindeman
