Study Area. —

The state of Quintana Roo on the eastern part of the Yucatan Peninsula (lat 17°48′–21°10′N, long 86°48′–89°10′W) (Fig. 1) is bordered by the Gulf of Mexico to the northwest and by the Caribbean Sea to the east. The climate is warm subhumid, with an average rainfall of 1100–1500 mm year⁻¹ (Rioja-Nieto et al. 2019). The rainy season lasts 5 mo, from June to October; September has the most days of rain, an average of 14.7 days (Parra et al. 2016). The temperature ranges from 10°C to 26°C, with the dry season from November to May (Rioja-Nieto et al. 2019). The Quintana Roo coast provides foraging habitats for adult and immature sea turtles in diverse neritic zones (Labrada-Martagón et al. 2017; Muñoz et al. 2017) as well as nesting beaches for 4 species: green sea turtle (C. mydas), loggerhead sea turtle (Caretta caretta), hawksbill sea turtle (Eretmochelys imbricata), and leatherback sea turtle (Dermochelys coriacea) (Perera-Valderrama et al. 2020).

View Figure

• Download as Powerpoint
• Download Figure

For the green sea turtle, breeding and nesting seasons occur mainly from May to October in Quintana Roo. The remigration period is 2–3 yr for females (Stokes et al. 2014). The growth rate has been estimated to range between 0.39 and 14.03 cm curved carapace length (CCL) yr⁻¹ at Akumal Bay, which is an important growth and foraging site for green sea turtles (Labrada-Martagón et al. 2017). An immature sea turtle recruited to the feeding ground at 28 cm of CCL would require between 13 and 14 yr (Labrada-Martagón et al. 2017) to reach the average size of nesting females in Quintana Roo (106.6 cm CCL; Zurita et al. 1993) based on growth rate predictions from a multivariate model (Labrada-Martagón et al. 2017). In feeding grounds such as Akumal, Punta Herrero, and Xcalak, the size range of immatures is between 27.8 and 81.0 cm CCL (Labrada-Martagón et al. 2017; Guevara-Meléndez et al. 2023). The central and southern region coasts of Quintana Roo (e.g., Akumal, Xcalak, Punta Herrero) are regarded as feeding and developmental habitats for immature juveniles of the green turtle, whereas adults prefer coastal regions and deep water farther north (e.g., Punta Arenas) (Muñoz et al. 2017).

Blood samples were obtained during the rainy season (June–October) from 2013 to 2019 in 7 localities on the eastern coast of Quintana Roo. Most of the samples were obtained at foraging grounds, from north to south, Punta Arenas (lat 21°30′8″N, long 86°55′40″W), Ixlache (lat 21°29′46″N, long 86°47′48″W), Akumal (lat 20°24′25″N, long 87°19′16″W), Punta Herrero (lat 19°27′40″N, long 87°26′55″W), Mahahual (lat 18°42′50″N, long 87°42′34″W), and Xcalak (lat 18°17′51.8″N, long 87°49′30.6″W) (Fig. 1). At Xcacel Beach (lat 20°20′27″N, long 87°20′37″W), blood samples were obtained from nesting females. Sample sizes were unequal across study sites and years (Table 1).

Table 1. Number of green turtles (Chelonia mydas) captured (number of serum samples) and capture per unit effort (CPUE) from the coast of Quintana Roo, Mexico.

View Fullscreen

Sea Turtle Capture and Sample Collection. —

The sea turtles (n = 197) were captured as previously described (Labrada-Martagón et al. 2017; Muñoz et al. 2022; Guevara-Meléndez et al. 2023). At the northern sites, where open and deep water are found (Ixlache and Punta Arenas), 4 monofilament fishing nets (45 meters long; 60 cm mesh) were suspended in the water and monitored every 2 hr during each sampling period of 13.5–18.5 h (Muñoz et al. 2017). In Akumal Bay (central region), capture was made by hand using snorkel equipment when the turtles were feeding or resting on the bottom (Labrada-Martagón et al. 2017). In the southern localities (Punta Herrero, Mahahual, and Xcalak), where surface and depth conditions permit and where numbers of turtles can be seen at the surface, the rodeo and sighting chase technique (Limpus 1978) was used (Guevara-Meléndez et al. 2023). Nesting females were sampled on Xcacel Beach (central region) without interfering with their nesting behavior (Schroeder and Murphy 1999), without artificial light, and only after digging had begun.

Immediately after the capture, a blood sample was obtained from the cervical venous sinus (Owens and Ruiz 1980) using double-ended Vacutainer needles (1.5 inch length, 32 mm caliper) and Vacutainer blood collection tubes without anticoagulant (7 ml; Becton Dickinson, Franklin Lakes, NJ) as described elsewhere (Keller et al. 2004; Labrada-Martagón et al. 2010, 2011, 2013, 2019). Blood samples were transported to the field camp in coolers with refrigerant gel packs (4°C), where tubes were centrifuged (890 × g) for 15 min in a field centrifuge (Mobilespin, Vulcon Technologies, Grandview, MO). The serum was transferred into labeled microtubes (1.5 ml cryovials, Eppendorf) and immediately frozen and stored in cryogenic shippers (CP 100, Taylor-Wharton, Baytown, TX) for transfer to the laboratory. Samples were stored at −80°C in the laboratory until processed.

The CCL was measured ( ± 0.1 cm) from the anterior point at the midline (nuchal scute) to the posterior tip of the supra caudals, and curved carapace width (CCW) was measured at the widest point (Bolten 1999). Secondary sexual characteristics, such as tail length, were recorded (Blanvillain et al. 2008) to assess adult status. On each turtle, a Monel tag (National Band and Tag Company, Newport, KY) was attached in the axillary position on each anterior flipper (Labrada-Martagón et al. 2017; Muñoz et al. 2022) and was released immediately after sample collection. Serum samples were analyzed at the Health Ecology Laboratory of the Faculty of Sciences, Universidad Autónoma de San Luis Potosí, Mexico.

Serum Analysis. —

Estradiol (E₂), testosterone (T), and free thyroxine (T₄) concentrations were determined in a total of 150 serum samples by direct solid-phase competitive enzyme-linked immunoassay (ELISA) as reported before for sea turtles (Ikonomopoulou et al. 2008; Labrada-Martagón et al. 2013; Allen et al. 2015). Measurements of hormone levels were not possible for all of the sea turtles captured (n = 197); serum samples from 47 individuals were not used because of insufficient blood or hemolysis of the samples. Commercial kits used in this study were previously tested for steroid analysis in C. mydas (Ikonomopolou 2008) and previously reported (Labrada-Martagón et al. 2013). ELISAs were performed in 96 microwell plates coated with streptavidin, and the conjugates supplied were anti-estradiol monoclonal antibody, anti-testosterone-biotin solution, and anti-T₄-biotin solution (Diagnostics Biochem Canada, London, Ontario, Canada). All reagents and samples were maintained at room temperature (18°–25°C) before use. Samples were analyzed in duplicate (Ikonomopoulou et al. 2008) per the manufacturer’s instructions. Absorbance was measured with a microplate reader (ELX 800 Biotek, MexLab, Vermont, USA) at 450 nm. Concentration values were derived from a 4-parameter equation using GraphPad Prism v.9 (GraphPad Software, Inc., Massachusetts, USA, 2020–2022). Levels of hormones are presented as ng ml⁻¹ for E₂ and T and as µg 100 ml⁻¹ for T₄.

For all the ELISAs, the standard curve determination coefficient values (r²) were > 0.998 for each hormone. Interassay coefficient variability was acceptable (CV < 20%; Hanneman et al. 2011); coefficients of variability intra- and interassay estimated, standard curves, and sensitivity of the kits reported by the suppliers are presented in Table 2. The assay precision was similar to previous work that followed a T extraction procedure in plasma samples of sea turtles by adding anhydrous ethyl ether (Allen et al. 2015). Most of the serum samples (96%) fell within the range of the manufacturers’ calibration curves. For E₂ and T, 3% and 4% of samples, respectively, were excluded since they were outside the detection limits of the assays (Table 3). The lack of correlation between hormone concentrations and handling time indicates that the sample processing was adequate and did not influence our data.

Table 2. Standard curves, sensitivity, and precision (CV%) of the enzyme-linked immunoassays (ELISAs) for steroid hormones and thyroxine.

View Fullscreen

Table 3. Descriptive statistics and confidence intervals estimated of the sex steroids and thyroxine levels of green turtles (Chelonia mydas) captured on the coast of Quintana Roo during 2013–2019.^a

View Fullscreen

Data Classification and Statistical Analyses. —

Capture per unit effort (CPUE) was estimated for each study site as an indicator of relative abundance (Bjorndal et al. 2000). This was done by using the number of green turtles captured per site divided by the number of hours spent during capture (Labrada-Martagón et al. 2010, 2017). Within the foraging regions (northern, central, or southern) green turtles were grouped into distinct categories to be evaluated according to their age/size classes as proposed by Aguirre and Balazs (2000) and following the adjustments suggested by Musick (2002) as follows: (a) recruits, < 38 cm CCL; (b) juveniles, 38–68 cm CCL; (c) subadults, 68–88 cm CCL; and d) adults, > 88 cm CCL. The single exception was the grouping for reproductive classes (mature or immature) according to the minimum nesting size (86 cm CCL) estimated for the green turtle population of Quintana Roo (Zurita 2015). Thus, individuals with CCL < 86 cm were classified as immature, and those with CCL > 86 cm as mature (Labrada-Martagón et al. 2017). Sex was identified only in adults (> 86 cm CCL) with evident secondary sexual characteristics such as tail length (larger in males) (Blanvillain et al. 2008).

Descriptive statistics and confidence intervals of the mean of the hormone levels are presented for all samples pooled and for transient and resident groups. Descriptive statistical information for age/size classes (Aguirre and Balazs 2000) is compared with other studies. Nonparametric tests were used due to the lack of normality in all the variables (Shapiro-Wilks, p < 0.05). A Box-Cox transformation was used to approximate data to a normal distribution to estimate confidence intervals (95%) of the mean (Box and Cox 1964; Zar 1984; Sakia 1992). Kruskal-Willis’s H-test (KW-H) was used to assess differences in variables between age/size classes and regions (northern, central, or southern), and the Mann-Whitney’s U-test (MW-U) to compare differences between sex, reproductive classes (immatures and mature), and resident and transient turtles. Contingency tables and chi-square tests (Daniel 1991) were used to evaluate the independence of the frequency of animals observed by hormone level intervals and the classifications used: reproductive state (immature and mature), sex, and habitat preference (resident and transient sea turtles). Additionally, Kolmogorov-Smirnov tests were performed to evaluate differences in frequency distributions between the categorical variables (Young 1977). For the former analyses, 4 class intervals of hormone levels were generated using Sturges’ rule (Daniels 1991). The correlations between the concentration of the hormones with the size of the individuals (CCL) and handling time used, measured from the moment of the capture to when the blood was extracted, were evaluated with Spearman’s correlation coefficients. All statistical analyses were performed using R software (v. 4.0.5, R Core Team 2021) with a statistical significance of α ≤ 0.05.

Population Structure. —

From 2013 to 2019, 197 green sea turtles were captured in 7 foraging areas along the northern, central, and southern Quintana Roo coast of Mexico. From those, 150 serum samples were obtained and assessed with ELISA for hormone determinations (Table 1); in this section information on those 150 sea turtles is presented and described. The size of the green turtles varied significantly across regions (KW-H = 95.9, p < 0.05). Northern and central turtles were larger than southern turtles (Fig. 2). Mean CCL in the northern region (Punta Arenas and Ixlache) was 89.8 ± 12.6 cm (66.5–115.0 cm CCL), in the central region (Akumal Bay) 66.2 ± 16.2 cm (47.2–80 cm CCL), and in the southern region (Punta Herrero, Mahahual, and Xcalak) 49.7 ± 11.4 cm (26.2–71.5 cm CCL). Only 4 female green turtles were sampled at the Xcacel nesting beach, 110.6 ± 5.9 cm (103.7–116.4 cm CCL).

View Figure

• Download as Powerpoint
• Download Figure

When the frequency of all green turtles was grouped by size classes 2 distinct distributions became evident (Fig. 2). Therefore, a new classification for the green turtle stock of Quintana Roo is proposed, based on habitat preference as follows: resident green turtles inhabiting the neritic zones of Akumal, Punta Herrero, Mahahual, and Xcalak (26.2–80 cm CCL) and transient green turtles found in deeper, offshore habitat in Punta Arenas and Ixlache; the former group includes nesters from Xcacel (66.5–116.4 cm CCL) (Fig. 3). The residents (smaller green turtles) represent 68% (n = 102) of the samples and transients 32% (n = 48). Transient turtles were bigger (CCL 91.6 ± 13.4 cm, 66.5–116.4 cm) than residents (CCL 54.8 ± 12.1 cm, 26.2–80.0 cm) (MW-U = 63.5, p < 0.05).

View Figure

• Download as Powerpoint
• Download Figure

Between age/size classes (Aguirre and Balazs 2000), there were significant differences in size (KW-H = 129.1, p < 0.05), as expected. Juveniles (n = 71) represented 47.3% of the individuals in this study, followed by subadults (n = 40; 26.7%), adults (n = 31; 20.7%), and recruits (< 20 cm CCL, n = 8; 5.3%). Among the reproductive classes, 119 green sea turtles (79.3%) were classified as immatures, being significantly smaller (CCL 58.0 ± 13.8 cm, 26.2–85.4 cm) than the 31 individuals (20.7%) considered as mature (CCL 99.58 ± 8.88 cm, 86.6–116.4 cm) (MW-U = 3598, p < 0.05). All the mature/adult sea turtles captured in foraging areas (n = 25) were found in the northern areas of the coast of Quintana Roo; 68% were captured in Punta Arenas (n = 17) and 32% in Ixlache (n = 8). These 2 sites exhibited a major proportion of adults at 61% and 39% immatures. According to the external morphological characteristics, 72% (n = 18) mature green turtles were classified as females (CCL 101.08 ± 8.56 cm, 87–116.4 cm), and 28% (n = 7) as males (CCL 94.4 ± 8.5 cm, 86.6–110.7 cm). The sex ratio estimated was 1:3 (M:F) (χ² = 0.12, p = 0.27) in foraging sites. The size of the green turtle adults showed significant differences between females and males (MW-U = 121, p = 0.049).

Serum Steroid Profiles. —

Mean values, descriptive statistics, and confidence intervals for the mean (95%) of pooled data are shown in Table 3; Tables 4 and 5 present the descriptive data of the hormone levels by grouping the sea turtles by age/size classes and habitat preference, respectively. The handling time used for bleeding was not correlated with any hormone concentration (E₂, r = 0.31; T, r = 0.008; T₄, r = 0.005, p > 0.3). The CCL of the green turtles was correlated with the concentration of estradiol (r = 0.25, p < 0.006), but not for T (r = 0.08, p = 0.86) and T₄ (r = 0.28, p = 0.13).

Table 4. Descriptive statistics and confidence intervals estimated of the sex steroids and thyroxine levels of green turtles (Chelonia mydas) captured on the coast of Quintana Roo during 2013–2019 grouped by age/size classes according to Aguirre and Balazs (2000).^a

View Fullscreen

Table 5. Descriptive statistics and confidence intervals estimated of the sex steroids and thyroxine levels of green turtles (Chelonia mydas) captured on the coast of Quintana Roo during 2013–2019 grouped by habitat use, residents and transients.^a

View Fullscreen

In adult females (n = 24), the mean concentrations of the hormones were 0.72 ± 1.23 ng ml⁻¹ of E₂ (0.002–1.45 ng ml⁻¹), 1.34 ± 1.39 ng ml⁻¹ of T (0.08–5.09 ng ml⁻¹), and 2.92 ± 2.19 µg 100 ml⁻¹ of T₄ (0.01–6.17 µg 100 ml⁻¹). For males (n = 7), hormone levels were 0.16 ± 0.11 ng ml⁻¹ of E₂ (0.023–0.32 ng ml⁻¹), 0.54 ± 0.44 ng ml⁻¹ of T (0.10–1.29 ng ml⁻¹), and 2.01 ± 1.40 µg 100 ml⁻¹ of T₄ (0.45–4.20 µg ml⁻¹). There were no significant differences in the concentration of the hormones when age/size classes, reproductive classes, or sex were compared. A sex ratio of 1:5 toward females was estimated in foraging grounds based on T levels, for both immatures (χ² = 0.087, p = 0.23) and adults (χ² = 0.008, p = 0.71) (Supplemental Fig. S1; all supplemental material is available at http://dx.doi.org/10.2744/CCB-1604.1.s1). The distributional patterns of the frequency of individuals, grouped by interval classes of hormone levels, did not differ between demographic categories (Supplemental Figs. S2–S4), and the frequency of animals observed by hormone intervals did not depend on the demographic classification used.

The T levels differed significantly across regions (KW-H = 6.8, p = 0.03), whereas E₂ (p = 0.2) and T₄ (p = 0.4) concentrations did not. The highest levels of T were found in the central region and the lowest in the northern region (Fig. 4). Based on habitat preferences, significant differences between residents and transient green turtles were also found in T concentrations (MW-U = 1970, p = 0.04), where resident turtles presented the highest values (Fig. 4).

View Figure

• Download as Powerpoint
• Download Figure

Population Structure. —

Sea turtle conservation in the Mexican Caribbean began in the early 1970s (Frazier 1993). However, continuous monitoring and publication of long-term studies in this region about population ecology are lacking. In this study, hormone data from 150 green sea turtles, from whom a blood sample was obtained, are presented for the first time for 7 foraging areas located along Mexico’s Quintana Roo coast, from 2013 to 2019. The relative abundance of green turtles was higher at Akumal and Xcalak and the lowest at the northern sites. In this work, the lowest capture effort was performed in the nesting beach of Xcace; thus, only 4 nester females (the largest turtles in the study) were sampled. However, some bias in the population structure reported in this study is recognized, given the limitations during fieldwork. Capture methods of green turtles depend on individuals’ size and the physical and biotic characteristics of the habitat such as depth, bottom type, presence of tidal currents, reef barrier protection, and water visibility among others (Ehrhart and Ogren 1999). In general, sampling using a tangled net soaked in the water has been considered effective in foraging habitats, by changing technical specifications according to habitat conditions (Ehrhart and Ogren 1999). In this study, using a net with 60 cm of mesh size in the northern sites of Punta Arenas and Xcalak was useful to capture adults and immature green turtles as small as 66 cm CCL. Difficulties given by fieldwork in particular habitats, when working with free-living sea turtles, could result in bias of population data (e.g., males or oceanic turtles < 25 cm CCL) (Balazs 1995) and restriction of sample sizes captured. However, the information presented may be useful to complete the missing data required to understand the population ecology of regional management units such as the Atlantic Northwest green turtle stock (Wallace et al. 2010). Future studies about dynamic populations, using base data presented in this work, will be useful to develop management and conservation strategies for sea turtles (Zurita et al. 1993; CONANP 2011; Thomassiny and Chan 2011; Zurita 2015).

Studies about population ecology in foraging habitats of green turtles from Mexico are scarce, thus comparisons are limited. The average size of green turtles in Quintana Roo (66.1 ± 21.3 cm CCL) is very similar to the foraging groups of the west coast of the East Pacific green turtle population in Baja California Sur (Bahía Magdalena) (54.6 cm straight carapace length [SCL]) (Koch et al. 2006) but smaller than the Gulf of California (Bahía de Los Angeles) foraging area (74.3 cm SCL) (Seminoff et al. 2003). Considering all the individuals captured, juvenile immatures (56.6%) were the major percentage of the sea turtles in this study, followed by subadults, adults, and recruits. This result coincided with the East Pacific green turtle population of the Baja California Peninsula, Mexico, where immatures are more frequent in the coastal foraging lagoons such as Bahía Magdalena (Koch et al. 2006; López-Castro et al. 2010; Labrada-Martagón et al. 2013), Estero Banderitas (Koch et al. 2007), Punta Abreojos (López-Castro et al. 2010; Labrada-Martagón et al. 2013), Laguna Ojo de Liebre (López-Castro et al. 2010), and Laguna San Ignacio (López-Castro et al. 2010; Labrada-Martagón et al. 2013).

All the adult sea turtles captured in foraging areas were found in the northern areas of the coast of Quintana Roo (Punta Arenas and Ixlache). The northeastern coast of Quintana Roo (where Punta Arenas and Ixlache are located) has been described as a critical foraging habitat for Chelonia mydas, where postnesting green turtles from the Gulf of Mexico present spatially restricted movements around the Yucatan Peninsula tip (Cuevas et al. 2008, 2020; Uribe-Martínez et al. 2021). We found 2 distributions in the frequency of sea turtles grouped by size classes (CCL) designated by differences in habitat conditions and in capture methods used. The shorter green turtles were found in the southern coastal bays and the bigger turtles in the northern region, proposed in this study as residents and transients, respectively. The former terms were given considering the ontogenetic change in habitat use in sea turtles (Musick and Limpus 1997; Bolten 2003). Recruited individuals (> 20 cm) move toward shore to inhabit the neritic areas of shallow bays or reefs (Musick and Limpus 1997; Bolten 2003) where they spend many years or decades before reaching sexual maturity (Seminoff et al. 2002; Labrada-Martagón et al. 2017), at which time the subadults undergo a final transition, often forming foraging aggregations (Owens 1997). The term transient, used in this study, was given considering the migratory behavior of adult sea turtles. It has been reported that green turtles nesting on the coast of Quintana Roo can travel to feeding home ranges outside of Mexican waters, including the Florida Keys, Belize, and Nicaragua (Uribe-Martínez et al. 2021). The habitat where the smaller turtles of this study were captured is characterized as being protected bays (e.g., absence of strong waves) with coral reef barriers, shallow water columns, and many hectares of seagrass beds (Tussenbroek et al. 2017). Green turtle residents living in central and southern regions on the coast of Quintana Roo such as Akumal are immature individuals that could spend many years, even decades (15–16 years), in their foraging grounds until they reach sexual maturity (Labrada-Martagón et al. 2017). These southern residents have not passed the final transition to the foraging habitat type where they will remain until they initiate their reproductive migrations (Owens 1997).

The smaller size segregation has been reported before in Bahía Magdalena, Mexico, where the size of the East Pacific green turtles captured inside the estuaries was significantly smaller than green turtles captured in the Pacific Ocean, along the west coast of the Baja California Peninsula (López-Mendilaharsu 2002; Koch et al. 2006, 2007). This size segregation of individuals also observed on the coast of Quintana Roo can be explained by differences in the physical attributes of the foraging and developmental habitats and coincides with habitat preferences and ecological factors observed previously in green turtles. Juveniles are found generally using coastal and shallow habitats (e.g., bays, reefs, coastal lagoons), while subadults and adults prefer subtidal habitats at the near-shore line (Limpus et al. 2005; López-Mendilaharsu et al. 2005; Limpus 2009). In contrast, the northern sites where aggregation of larger forager turtles can be found in feeding areas are near-shore locations, 20 km from shore, in the continental shelve, in deeper water (< 50 m), with a complex system of currents including the Gulf of Mexico current, Cozumel channel, Yucatan channel, and Yucatan current (Candela et al. 2003; Cuevas et al. 2008, 2020; Uribe-Martínez et al. 2021). Additionally, the size/habitat preferences have been explained by differences in food requirements, shelter, and protection from predators (López-Mendilaharsu 2002; Limpus 2009).

According to the external morphological characteristics of adults (> 86 cm CCL), there were more females (n = 18) than males (n = 7), showing a female-biased sex ratio. The sex ratio 1:3 (M:F) of foraging adult turtles (n = 25) in the current study was biased toward females as are many other foraging areas from the Hawaiian Islands and Australia (Wibbels et al. 1993; Limpus et al. 2005; Limpus 2009; Hof et al. 2017; Jensen et al. 2018; Bell et al. 2019). This is the first report on the sex ratio for adult green turtles captured on the coast of Quintana Roo. Even when hatchling and adult sex ratios correspond in some turtle species (Ewert and Nelson 1991), the relationship between the sex ratios between age-size classes (hatchling, immature, and adult) is not clear for sea turtle populations (Wibbels et al. 1987, 1991). Previous results suggest that temperature-dependent sex determination (TSD) in sea turtle populations is capable of producing female-biased sex ratios (Wibbels et al. 1987, 1991; Mrosovsky and Provancha 1991; Limpus et al. 2005; Hof et al. 2017; Bell et al. 2019), or male-biased sex ratios (Limpus 1985; Mrosovsky et al. 1992). However, there is a need to evaluate changes in population sex ratios relative to other factors, such as physical parameters of nesting beaches, reproductive ecology, and nesting behavior in populations, and to compare sex ratios of different age classes of sea turtles within a population as well (Wibbels et al. 1993). Bell et al. (2019) provide some interpretations as to why the female-biased sex ratio in breeding green turtle populations could be sustained over time, including aspects of the migratory and breeding behavior of adult males (Limpus et al. 2005), and the polygamous mating behavior of the species (Jessop et al. 1999). However, even when female-biased operational sex ratios may be able to sustain nesting populations, the extent to which a female-biased hatchling production can be viable requires further investigation (Bell et al. 2019). Considering the impacts of TSD, global warming temperatures are a serious threat to the persistence of sea turtle populations. The proportion of female hatchlings increases naturally with the incubation temperature, creating a lack of male turtles that will eventually impact the overall fertility rates in the population (Jensen et al. 2018).

The sex ratio of this study was generated by classic morphological criteria such as the minimum nesting size (86 cm CCL) estimated for the green turtle population of Quintana Roo (Zurita 2015) and by measuring tail length in adults. Nevertheless, using secondary sexual characteristics to determine the sex of adult turtles could potentially bias the sex ratio toward females since immature males may have delayed adult differentiation (Limpus and Read 1985; Labrada-Martagón et al. 2014). Additionally, the average and minimum nesting size of C. mydas differs between populations. Some examples are the mean nesting size estimated for the Gulf of California (74.3 cm SCL, Seminoff et al. 2003), Revillagigedo Islands (94 cm CCL, Turner et al. 2022), and the coast of Michoacán Mexico (94 cm CCL, Turner et al. 2022). Hence, the use of nesting size as the criterion to establish the population structure requires previous, calibrated data for each population (Seminoff et al. 2003; Koch et al. 2007; López-Castro et al. 2010; Labrada-Martagón et al. 2013). More reliable sex determination methods, such as laparoscopy and blood hormone concentrations (Jensen et al. 2018; Bell et al. 2019), have been highlighted (revised by Labrada-Martagón et al. 2014).

Serum Steroids Profile. —

The current study provides, for the first time, relevant and new data about hormone levels associated with reproduction (E₂, T) and metabolism (T₄) of the green turtle population from the coast of Quintana Roo. This information about reproduction is currently lacking and is of great relevance for understanding the developmental ecology of this population in the Caribbean, which could be used as basal data for monitoring and comparison with other populations of green turtles. Similar results have been previously reported about T levels of immature hawksbill turtles (E. imbricata) and nonbreeder male green turtles after being exposed to a capture stress protocol, even when corticosterone concentrations increased (Jessop et al. 2004).

The hormone levels (E₂, T) fall within ranges previously reported for the species in Cayman turtle farm (Licht et al. 1979, 1985), the coast of Michoacán, Mexico (Licht 1980), Surinam (Licht 1980), China (Gross et al. 1995), Australia (Wibbels et al. 1992; Moon et al. 1998; Jessop et al. 1999; Hamann et al. 2005), Hawaiian Islands (Wibbels et al. 1993), Malaysia (Ikonomopoulou et al. 2008), the west coast of Baja California Peninsula (Labrada-Martagón et al. 2013), San Diego, California (Allen et al. 2015), and Costa Rica (Howell et al. 2022). The variability in hormone range levels is similar to previous reports (Hamann et al. 2005; Howell et al. 2022).

Considering all green turtles of this study, the range of T₄ (0.001–9.22 µg 100 ml⁻¹) coincides with reports of free-living green turtles from Heron Island in Australia (0.14–0.31 µg 100 ml⁻¹, Moon et al. 1998), foraging coastal lagoons of Baja California Sur, Mexico (0.649– 3.399 µg 100 ml⁻¹, Labrada-Martagón et al. 2013), and Costa Rica (0.5–0.9 µg 100 ml⁻¹, Howell et al. 2002). In the wild, higher concentrations of T₄ have been reported during summer (May and June), coinciding with the beginning of the rainy season (Licht 1970). In this study, the levels found between August and September (end of summer) coincide with those reported in captive turtles of the Cayman Islands (0.89–1.05 µg 100 ml⁻¹, Licht et al. 1985). The T₄ is implicated in physiological functions associated with an elevated energetic cost in reptiles, with an increased thyroid activity associated with growth, maturity, mating behavior, yolk deposition, and ovulation (Dickhoff and Darling 1983; Licht et al. 1985; Leatherland 1987; Moon et al. 1998). Additionally, T₄ is 1 of the hormones responsible for energy regulation and nutrient assimilation; thus, this hormone is considered an indicator of metabolic activity in sea turtles (Moon et al. 1998; Ikonomopoulou et al. 2014) and has been related to glucose levels in immature East Pacific green turtles (Labrada-Martagón et al. 2013). There were no differences found in levels of T₄ between any demographic category considered in this study; our results suggest that the green turtle population of Quintana Roo is in a good nutritional state, with a high-quality diet with elevated protein levels as reported for captive individuals of C. mydas from Cayman Islands (Licht et al. 1985; Moon et al. 1998)_. Also, all green turtles from this study presented very good body condition (unpublished data) based on classification criteria from Flint and collaborators (2009).

The only difference found in this study with statistical significance was the highest levels of T found in green turtles captured in the central region (Akumal Bay), where only immature, resident individuals were captured, in comparison with northern individuals, where 61% of the turtles captured were adults. These results suggest that green turtles found in shallow and protected foraging bays covered with seagrass beds found at 1.5 to 2.5 m depth (Molina and Tussenbroek 2014), where immature green turtles (26.2–80 cm CCL) could spend more than 15 years in their developmental areas until reaching the required nesting size (Labrada-Martagón et al. 2017), are presenting higher values of T than forager subadult and adult turtles. Some of the resident immature sea turtles from the center region may be reaching sexual maturity (Supplemental Fig. S3). Increasing values in circulating T occurred along with age (Owens 1997). The concentration of T, together with cholesterol, glucose, and T₄, explained the E₂ levels in immature East Pacific green turtles (Labrada-Martagón et al 2013). T levels of immature sea turtles also vary with photoperiod and temperature; increasing levels of T have been correlated with ambient temperature, with the highest values during the warm months of spring and summer (Morris 1982 in Owens 1997).

On the other hand, the lowest levels of T observed in northern adults may reflect values of some number of postbreeding individuals during late summer (capture period in this study) and some reproductively inactive adults (Wibbels et al. 1990; Hamann et al. 2002). Significant monthly increases of circulating T have been reported in males and females when they start becoming reproductively active and during courtship in spring and in the early nesting season, respectively (Wibbels et al. 1990; Hamann et al. 2003). Also, the highest T levels are reported in late autumn (October–November) in females when approaching migration (Wibbels et al. 1990). T levels could be informative about the sexual maturity process of sea turtles and may indicate the reproductive status of individuals within a population (Blanvillain et al. 2008; Valente et al. 2011). Although the hormone variations during puberty/immaturity and the sexual maturation process are unknown for this group, broader studies are necessary to better understand the maturity process of sea turtles along all size classes.

Except for the immature male/female T differences, the reproductive hormones considered in this study (E₂ and T) did not vary predictably between reproductive classes. The results could be affected by the sample sizes; 79% of the individuals captured were immatures, and 68% of all green turtles corresponded to the smaller residents. Considering only adults (n = 31), 77% were females. Although similar results have been reported for immature East Pacific green turtles from the Baja California Peninsula, no differences were found when age/size classes were compared (Labrada-Martagón et al. 2013). The elevated range and average T levels found in adult females in this study compared to males (Supplemental Fig. S2) can be explained by the metabolism of sex steroids and the reproductive physiology of sea turtles. T is the precursor of E₂, and the presence and quantity of both hormones are closely related (Hernández-López and Cerda-Molina 2012). The physiological changes occurring during vitellogenesis and with follicular development include increased T and estrogens, among other hormones (Hamann et al. 2002, 2003). Also, individuals with different reproductive states coexist on foraging grounds (Owens 1997); it is possible to find reproductively inactive males, with lower T levels, when captured in the water during reproductive months (Wibbels et al. 1987).

Another factor that has to be considered is the misclassification generated when the demographic data (sex and reproductive state) of the green turtles are based only on morphometric criteria in sexually monomorphic species (Labrada-Martagón et al. 2014). The estimated sex ratio based on testosterone levels for both adult and immatures was 1:5 toward females, which is similar to many other green turtle populations from Australia such as the Great Barrier Reef (1:4.2) (Jensen et al. 2018), Queensland (1:4.2) (Bell et al. 2019), and Moreton Bay (1:5) (Limpus et al. 1994). Misclassification of age class and reproductive state has been reported before (Wibbels et al. 1987, 1993; Labrada-Martagón 2011; Labrada-Martagón et al. 2013). Results concerning levels of T (> 0.17 ng ml⁻¹) and E₂ (> 0.18 ng ml⁻¹) of green sea turtles from the Hawaiian Archipelago (Wibbels et al. 1993) suggested a bias in the number of juveniles, which were defined only on the absence of secondary sexual characteristics and their small size as in the present study, but were presenting already concentrations of sexual steroids within the range of nester adults from Florida and Australia (Wibbels et al. 1987, 1992). Opposite results were found in the Baja California Peninsula population, where only 28% of the East Pacific green turtles that were classified as female adults given the mean nesting size criterion (> 77.3 cm SCL), presented elevated levels of estrogenic hormones (Labrada-Martagón 2011), and only 1 individual (14%) classified as female had sex steroid concentrations within the ranges reported for adults in the nesting period (Labrada-Martagón et al. 2013). Further validation of the utility of the T for sex classification in green turtles is necessary before it can be used for sex ratio determinations in distinct foraging areas, with distinct cohorts, and different ages and maturity states (Allen et al. 2015). However, we believe this first assessment of the Mexican Caribbean population of immatures and adults in foraging habitats will prove useful.

The steroids that have received the most attention in studies with reptiles and C. mydas are T, progesterone, and E₂, but other steroids may play equally important roles in the physiology of reptiles (Licht 1979, 1982; Licht et al. 1985; Owens and Morris 1985; Wibbels et al. 1990, 1992; Owens 1996, 1997, 1999; Ikonomopoulou et al. 2014). Estrogen, T, and progesterone influence the sexual behavior of female nesting reptiles. T is elevated during courtship and mating in many female reptile species and may increase female receptivity (Norris and Lopez 2018). Elevation of T in males has physiological roles for reproduction (Randall et al. 2002) but also appears to have some behavioral roles, preparing specific brain regions (Rostal et al. 1998). In sea turtles in particular, T appears to function in regulating the seasonal reproduction in both male and female sea turtles (Owens 1996; Rostal et al. 1998; Hamann et al. 2002; Blanvillain et al. 2011; Howell et al. 2022). The T in green turtle females is relatively high when nesting season begins and slowly decreases along the season (Licht 1980). When additional nesting attempts occur in a sea turtle’s nesting sequence, more fluctuating and often decreasing concentrations of T have been observed (Hamann et al. 2002) and described in green turtles, Kemp’s ridley turtles (Lepidochelys kempii), and loggerhead sea turtles (Owens 1996). Mating usually occurs when the ovaries release estrogens and reach a peak in females (Licht et al. 1985) and then decrease, even below detection levels, while nesting (Licht et al. 1979; Licht 1980; Wibbels et al. 1990; Ikonomopoulou et al. 2014). In loggerhead sea turtles (C. caretta), E₂ remains detectable without significant changes during the reproductive season (Wibbels et al. 1992; Smelker et al. 2014), and testosterone continues to decrease as the nesting season progresses (Licht et al. 1979; Al-Habsi et al. 2006; Kakizoe et al. 2010). In this species, the minimum concentration of E₂ coincides with a general decrease in T and an increase in progesterone at the time of ovulation (Wibbels et al. 1992).

This study found no differences in any hormone when nesting females and nonreproductive green turtles were compared. The same observation has been reported in the Cayman Islands (Licht et al. 1979), but our results are inconclusive since this work considers only 4 females sampled during the nesting season. Capture months of this study corresponded to the late nesting season in the Mexican Caribbean, which may be contributing to the minimal differences in hormone levels found between reproductive and nonreproductive females. It has been suggested that sea turtles with no detectable hormone levels may correspond to individuals captured between reproductive and nesting seasons or as immature subadults (Licht 1982; Smelker et al. 2014). Only 4% of our samples were out of the detectable range of the standard curves for E₂ and T, with all but 1 corresponding to immature turtles. Similar results were reported in immature East Pacific green turtles of similar size (40–92 cm SCL) from the Baja California Peninsula and using the same analytical methodology of this study, which presented detectable ranges of E2, T, and T4 in all samples evaluated (Labrada-Martagón et al. 2013).