Methods

The sampling was carried out along a 5-km line transect in the Isla Aguada Village (lat 18°47′05″N, long 91°29′30″W) in Campeche, Mexico, which belongs to the Natural Protected Area of Terminos Lagoon. The shore of Campeche is among the most critical areas of nesting for 3 species of sea turtles: the hawksbill turtle (Eretmochelys imbricata), the green turtle, and the Kemp's ridley (Lepidochelys kempi) in the Yucatán Peninsula region. Moreover, Isla Aguada and the area of Chenkan are the nesting sites with the highest production of hatchlings of green turtle in this area (Crevenna et al. 2011; Islebe et al. 2015).

Our field efforts occurred during 18 night-sampling sessions from June to September 2016. For each encountered nesting female we measured its minimum curved carapace length (CCL-min), as there is less variation in the measures than when the curved carapace length notch to tip is used (Bolten 1999). No additional measurements were obtained.

For turtles with ectoparasites and epibionts we collected sample specimens of each species and recorded the overall parasite/epibiont load on the body of each turtle. We calculated the frequency of occurrence of each species ectoparasites and epibionts as % = (f/n) × 100, where f is the number of times either ectoparasites or epibionts appeared on the carapace, flippers, neck, or other area of the body of the total number of examined turtles, which is represented by n. According to the criteria of Mougi (2016) we defined parasites as organisms living on the outer or inner surface of a host turtle to obtain benefits while potentially causing damage or injuries, whereas epibionts are considered as being commensal organisms living attached to a host turtle without benefiting or harming it.

We collected blood samples (5 ml) by venipuncture on the dorsal cervical sinuses (Owens and Ruiz 1980), using 5-ml syringes (BD Vacutainer) and 21-gauge needles. So as not to interfere with the nesting process, blood samples were collected after oviposition and nest covering. Blood samples were stored in individual lithium heparin tubes (BD Microtainer) in a cooler for 5–6 hrs until processing in the laboratory (Flint et al. 2010). The integrity of the blood samples has been shown to not be affected by this collection-to-analysis lag time (Stacy and Boylan 2014). Samples were centrifuged (7500 × g for 10 min) to separate blood cells and plasma. Our health assessment examined the 13 most common blood analytes: albumin (ALB), alkaline phosphatase (ALKP), alanine aminotransferase (ALT), amylase (AMYL), blood urea nitrogen (BUN), calcium (Ca), cholesterol (CHOL), creatinine (CREA), globulins (GLOB), glucose (GLU), phosphorus (PHOS), total bilirubin (TBIL), and total proteins (TP) (Bolten and Bjorndal 1992; Hasbún et al. 1998; Aguirre and Balazs 2000; Montilla et al. 2008; Redondo Zúñiga 2008; Flint et al. 2010; Labrada-Martagón et al. 2010; Anderson et al. 2011; Lara-Uc et al. 2012; Prieto-Torres et al. 2013; Page-Karjian et al. 2014; Brito-Carrasco 2016). All analyses were performed with a chemistry analyzer (Vet-Test-8008, IDEXX Laboratories, Inc. Main, USA) using a general health profile kit (GHP, IDEXX Laboratories). Acceptable values for each analyte were established according to the respective range of values for sea turtles provided by the Vet-Test-8008 Analyzer (IDEXX Laboratories). We also compared our results with data presented in Bolten and Bjorndal (1992) and Flint et al. (2010), both of which reported blood chemistry values based on the analysis of samples from ≥ 90 green turtles.

The skin and carapace of each sampled female were examined for ectoparasites and epibionts. When detected, ectoparasites and/or epibionts were counted and collected with forceps and stored in 70% ethanol for later taxonomic identification. In addition, 100 of the collected ectoparasites were stained with Meyer's paracarmine for 8 min and mounted in Canada balsam (Pritchard and Kruse 1982) to facilitate species identification through morphometric analysis. Species identification of epibionts followed the keys of Naylor (1972) and Relini (1980), whereas species identification of ectoparasites was based on Davies (1991). Estimation of the prevalence and abundance of both types of organisms were performed according to Bush et al. (1997). In brief, the prevalence was estimated as (HP/HR) × 100, where HP is the number of turtles showing symbionts, and HR is the total number of turtles surveyed. The abundance was calculated as the number of symbionts of each species found in the turtles. For each analyte in our study, the means ± 1 standard deviation, as well as the minimum and maximum values, were determined. We used the Shapiro-Wilks Normality test to explore the distribution of the resulting values obtained from the analyses. The homoscedasticity of the analytes from turtles showing ectoparasites and epibionts and the turtles free of symbionts was tested through Fisher's F-tests. The data from the analytes that showed nonnormal distributions were normalized through a z-transformation. However, because the results based on original and transformed data did not differ, we used untransformed data and implemented parametric tests (Student's t-tests) for normally distributed data, and nonparametric tests (Mann-Whitney U-tests) for nonnormally distributed data, to compare the obtained values of the analytes between turtles with and without ectoparasites and epibionts. All tests were performed as implemented in the base packages of R 3.3.1 (R Development Core Team 2013).

Results and Discussion

We recorded 46 green turtles (mean size = 107.1 ± 7.1 cm CCL-min) nesting in our study area during our sampling period. Blood samples were obtained from 20 females, 10 of which showed symbionts. Additional samples from females without ectoparasites or epibionts were obtained (n = 10). However, no analysis was performed on these samples because of the insufficient quantity of blood. The CCL-min between turtles with and without symbionts was not statistically different (mean = 106 ± 5.2 cm, and mean = 108 ± 8.8 cm, respectively, t(18) = 0.60, p > 0.05).

Two groups of symbionts living on the green turtle females of Campeche were identified: ectoparasites and epibionts. Only one species of ectoparasite, the annelid Ozobranchus branchiatus was found (Fig. 1). The estimated prevalence of this organism was 21.7%. The frequency of occurrence was 70% on the neck and 30% on the flippers. We observed on average 38 ± 25.4 ectoparasites per turtle, with a size range of 8–9.8 mm. We also identified 2 species of epibionts living on the green turtle females: the barnacle Chelonibia testudinaria (diameter range = 1.2–7 cm; Fig. 1), and the amphipod Elasmopus sp. (body length range = 9.2–10.1 mm; Fig. 1). The prevalence of these species was 26% and 2%, respectively. Chelonibia testudinaria was found more frequently on the carapace (65%), than on the flippers (35%), while Elasmopus sp. was found only on the carapace. The observed average number of epibionts per turtle (including both species) was 14.9 ± 15.1 organisms.

View Figure

• Download as Powerpoint
• Download Figure

Five of 13 analytes showed values outside of the normal range. According to the reference values (IDEXX; Bolten and Bjorndal 1992; Flint et al. 2010), the analytes that showed values below the reference range were AMYL and BUN, with 85% (17/20) and 100% (20/20) of sampled turtles, respectively. About 70% of the turtles without ectoparasites or epibionts, and all the examined turtles hosting these organisms showed values of AMYL below the range reported by IDEXX. Additionally, as compared with the only existing report of AMYL in green turtles (Anderson et al. 2011), our observed values for this analyte were also low. For all the examined turtles, including those with and without ectoparasites and epibionts, the BUN values were below the range (IDEXX; Bolten and Bjorndal 1992; Flint et al. 2010). This result was similar to previous reports for green turtles (see Table 1). On the other hand, those analytes showing high abnormal values were CREA 70% (14/20), PHOS 45% (9/20) and TBIL 20% (4/20). Our results of CREA were similar to the values reported for green turtles from the Bahamas, the Emirates, Venezuela, Chile, Puerto Rico, and Baja California and Yucatán in Mexico (Table 1). However, other studies from Hawaii (Aguirre and Balazs 2000), Venezuela (Montilla et al. 2008), Australia (Flint et al. 2010), and North Carolina (Anderson et al. 2011) reported values concordant with the ranges of IDEXX. The values of PHOS and TBIL in our study area were similar to the values reported in Yucatán (Lara-Uc et al. 2012). In both cases, in Campeche (this work) and Yucatán, the values were high regarding the ranges of IDEXX, whereas previous studies are concordant with the reported ranges for sea turtles (IDEXX; Bolten and Bjorndal 1992; Flint et al. 2010). Table 1 shows the average values for the 13 analytes examined. Although we detected a decrease of AMYL and an increase of CREA as the number of ectoparasites and epibionts increased (pooled data), these relationships were not significant (adjusted r² = 0.18; F₅₁₄ = 1.84, p = 0.16; Table 2). When analyzed by species of ectoparasite, we found that only CREA had a significant positive relationship with turtle leech presence (adjusted r² = 0.29; F_17,42 = 2.43, p = 0.009; Table 2). No relationship between epibiont load and analyte concentration was detected.

Table 1. Comparison of the mean values of blood chemistry of Chelonia mydas between adult green turtles from Campeche and other populations around the world (including juveniles, sub-adults, and adults). Standard deviation (SD) is shown only for Campeche.^a

View Fullscreen

Table 2. Effect of the number of ectoparasites and epibionts (first 5 rows) on the concentration of 5 analytes in the blood of gravid green sea turtles. Although AMYL and CREA showed a weak trend, no significant effect was found. The sixth row shows the effect of turtle leech presence on creatinine levels in the study animals.

View Fullscreen

The blood chemical profile of the green turtle females nesting in Isla Aguada is concordant with the typical values (ALB, Ca, ALKP, CHOL, GLOB, GLU, TP) reported for green sea turtles foraging in the Bahamas (Bolten and Bjorndal 1992) and Australia (Flint et al. 2010). Although we observed atypical values for 5 analytes (AMYL, BUN, CREA, PHOS, TBIL), these results are likely related to factors such as environmental condition or nesting-related stress (Lutz et al. 1997; Deem et al. 2006) more than to owing to pathological circumstances. Moreover, the values could have been influenced by food quality or feeding habits (Bjorndal and Bolten 2010; Komoroske et al. 2011).

Because most of the clinical values of the tested analytes were in the normal range, and no physical damage or bodily injuries were observed in the sampled turtles, our data suggest that the presence of ectoparasites and/or epibionts is apparently harmless for the green turtle females nesting in Campeche. About the half of the examined green turtles did not possess any epibionts, although some showed a considerable number of ectoparasites (> 90). However, we did not observe any apparent injury or negative effects to the turtles. The epibionts were found in negligible quantities (max = 37 per turtle), thus not representing any apparent threat to health to the green turtles examined (Lescinsky 2001; Gámez-Vivaldo et al. 2006). The occurrence of O. branchiatus and C. testudinaria on the skin and carapace, respectively, is consistent with the findings for sea turtles from Yucatán, where Ozobranchus margoi was found in the soft anatomical regions of hawksbill turtles, and C. testudinaria on the caparace of both hawksbills and green turtles (Lara-Uc et al. 2012). However, C. testudinaria has also been observed on flippers and neck of other sea turtle species (Kitsos et al. 2005).

We found CREA values above the range provided by IDEXX Laboratories, as well as by Bolten and Bjorndal (1992) and Flint et al. (2010). Although this condition may be related to dehydration and kidney disease in reptiles, it is also a consequence of muscular stress or damage (Bar-Shai et al. 2008; Li et al. 2015), such as that which results from the high energy output during nesting activities. The tidal dynamics in the area of Isla Aguada erode the soil and form sandbanks ranging from 1 to 1.5 m in height with a slope up to 50° along the beach, which the turtles have to climb over in order to access nesting habitat that is safe from high tides (J.C.C.D., pers. obs.). Additionally, because of the local environmental conditions, the soil of Isla Aguada may be too dry (May–July) or very wet (end of September–October), at the middle and end of the egg-laying season (Guzmán and Garcia 2016), and this condition may require extra effort for nest building, causing additional muscular stress. Another probable cause of the increased CREA level could be the stress induced by venipuncture, especially when the blood sampling was difficult (Guillette et al. 1995; Snoddy et al. 2009; Stacy and Boylan 2014).

High levels of AMYL in the blood are related to pancreatic damage in clinical veterinary cases (Stacy and Boylan 2014). However, so far this has not been tested in green turtles. Nonetheless, there is evidence that the amount of plasma enzymes in the blood of ridley sea turtles living in cold marine currents might diminish (Anderson et al. 2011). To our knowledge, there is only one report of AMYL concentration in green turtles (Anderson et al. 2011), but these findings contrast with the high levels observed in our study. More samples and other clinical tests are needed to relate the results of AMYL in green turtles to pathological conditions.

We observed low levels of BUN in our samples as compared with the values reported previously for green turtles in North Carolina (Anderson et al. 2011). Our results are comparable to the findings for green turtles in Hawaii (Aguirre and Balazs 2000), Chile (Brito-Carrasco 2016), Puerto Rico (Page-Karjian et al. 2014), and Bahamas (Bolten and Bjorndal 1992), which were all below 7 mg/dl, but contrast with the values (21–33.5 mg/dl) reported for green turtles from Yucatán (Lara-Uc et al. 2012) and Venezuela (Montilla et al. 2008; Prieto-Torres et al. 2013). Although BUN deficiency in several species of reptiles indicates renal disease, low protein intake, and dehydration, in green turtles it may be because of their herbivorous feeding habits (Stacy and Boylan 2014). Thus, this analyte is not informative in assessing the health of green turtles when other clinical tests such as hematology are lacking.

About 50% of the examined animals showed hyperphosphatemia (that is, values of PHOS above the normal range), which is a relatively common condition in geriatric reptiles (Campbell 1998). However, because the age classes of green turtles are estimated by morphometrics (Moncada et al. 2006), it is difficult to know the exact age of each animal, and thus it is not possible to relate our results of PHOS to age. Another possible reason for the marked hyperphosphatemia observed is gravidity. Such an effect is present in gravid musk turtles (Sternotherus minor; Silvestre et al. 2013). Moreover, the rise of PHOS levels due to the process of yolk formation and folliculogenesis has been previously reported in several species of sea turtles (Stacy and Boylan 2014).

Similar to the findings in the shores of Yucatán (Lara-Uc et al. 2012), we observed TBIL values above the reference range. However, our results contrast the findings for green turtles in Hawaii (Aguirre and Balazs 2000), which may suggest a regional pattern in the values of this analyte. High values of TBIL could be related to hemolytic or hepatic diseases, but also to anemia, migratory status, and diet (Camacho et al. 2013). Because the turtles of Campeche and Yucatán showed similarities in TBIL, this result is likely related to the food quality and availability in the mutual foraging regions for these 2 nesting populations.

Continuous population monitoring that includes physical examination and clinical tests such as hematology and comparisons between sexes and among turtles of different age classes are needed to assess the health status of the green turtles found along the Campeche coast. It would also be of value to perform ecotoxicology and habitat quality studies that provide information about the impact of pollution (e.g., contaminants) on the health and food resources of green turtles in this region. The results presented here suggest possible factors affecting the results of BUN, PHOS, CREA, and TBIL in gravid green turtles, and are the first data describing part of the clinical profile of this population. We hope that these data are useful as a reference for future assessments in the region.