PATHOLOGY OF LEATHERBACK SEA TURTLE (DERMOCHELYS CORIACEA) EMBRYOS AND HATCHLINGS FROM NESTS IN ST. KITTS, WEST INDIES (2015–16)

Leatherback sea turtle (Dermochelys coriacea) is the largest turtle in the world, weighing up to 750 kg, and is distinguished by the lack of a bony carapace (Wyneken and Witherington 2001). Distributed worldwide, these sea turtles are pelagic and migrate vast distances, forage in cold waters, and nest in their natal beaches in tropical and subtropical regions (Hughes et al. 1998). Yet, like other sea turtle species, they are conservation dependent and have been heavily impacted by fishing bycatch, marine pollution, coastal habitat destruction, and poaching of turtles and eggs (Spotila et al. 2000; Tomillo et al. 2008; Mrosovsky et al. 2009). The global population has declined by approximately 40% over the past few generations, and the species is currently listed as vulnerable by the International Union for Conservation of Nature (Wallace et al. 2013).

Recovery of leatherback populations globally is challenged by notoriously low hatch success. The percentage of yolked eggs that hatch globally averages 50% per nest, lower than that of any other sea turtle (Eckert and Eckert 1990; Rafferty et al. 2011). Erosion, tidal inundation, predation, and plant root invasion may be problematic for leatherback nests (Wyneken et al. 1988), but they are typically excluded from estimates of hatch success (Eckert and Eckert 1990). Moreover, low hatch success is thought to be a result of a high rate of embryo mortality for unknown reasons (Bell et al. 2004). Studies have focused on physical nest properties that are not conducive to embryonal development, such as temperature, sand grain size, and moisture (Ackerman 1981; Wallace et al. 2004; Santidrian Tomillo et al. 2012). Postmortem examination with histology is a basic step to identify fatal diseases affecting sea turtles and embryos of egg-laying vertebrates (Flint et al. 2009; Rideout 2012), yet there is a paucity of literature addressing the pathology affecting leatherback embryos.

The population of nesting leatherback females in St. Kitts, West Indies, is estimated to be a minimum of 259 individuals (Stewart et al. 2016). Recent monitoring efforts showed that hatch success in St. Kitts is much lower than the global average: from 2003 to 2016, hatch success in St. Kitts averaged 18.7% (Stewart 2019). This is in addition to the 8% of nests that fail due to tidal inundation, sand mining, depredation, and poaching (Stewart 2019). Indeed, hatchling production in St. Kitts may be too low to sustain its nesting population long term. Our objective was to describe the pathology of leatherback sea turtle embryos and dead-in-nest hatchlings and to identify lesions that could explain perinatal mortality in St. Kitts.

Nest data were collected as part of the annual surveillance undertaken by St. Kitts Sea Turtle Monitoring Network (SKSTMN) since 2003 (Stewart 2019). We focused on data from 2015 to 2016, collected by morning and night patrols, undertaken from April to July on Keys (17°20′N, 62°43′′W) and North Friars (17°21′N, 62°43′′W) beaches. When a female was observed laying a nest, morphometrics, flipper tag, and passive integrated transponder tag identification numbers were recorded and full health assessments were conducted. When necessary, flipper tags were placed bilaterally in the inguinal skin flap and passive integrated transponder tags were placed in the right shoulder (Eckert and Beggs 2006). Eggs were counted during oviposition. The location of nest position on the beach was defined by distance from vegetation and high tide line, and GPS coordinates were recorded. In the event that oviposition was not observed, the disturbed area farthest from the previous night's high tide line was used for location data. Wooden stakes were placed approximately 30 cm behind the nest chamber for optimal relocation of the nest during excavation. Each stake was inscribed with deposition date, female identification numbers (or “unknown” if oviposition was not observed), and SKSTMN contact information.

Nests were monitored throughout the incubation process for signs of depredation, tidal inundation, erosion, and hatch. If there were signs of emergence, such as a depression or hatchling tracks, the excavation was performed within 0–3 d. If no signs of emergence were seen, nests were excavated at approximately days 65–75 of incubation.

Upon reaching the nest chamber, all eggs were removed and counted during excavation. Hatched and unhatched yolked eggs were separated and counted along with the yolkless eggs. An egg was categorized as hatched when 50% or more of the eggshell was present. All yolked eggs were opened and categorized by the stage of embryo development: no growth or signs of development, early, late, pipped, and hatched. No growth or signs of development were recorded for any egg where embryo development was not discerned macroscopically, typically before iris pigmentation and corresponding to less than 25% of developmental time (Miller et al. 2017). Early stage was defined as any stage of development from the time an eyespot (pigmented iris) was grossly evident to the period when embryo size was equal to or smaller than the volume of the yolk sac, corresponding to approximately 29–78% developmental time. Late-stage embryos were those whose size was larger than the volume of the yolk sac but were still within the egg (unpipped), corresponding to approximately 86–102% developmental time. Pipped was a hatchling that broke through the shell but was still within the shell, corresponding to approximately 51–62 d of gestation. Dead-in-nest hatchlings were turtles that were out of the shell but died within the nest, corresponding to approximately 53–62 d of gestation. Hatch success for each nest was determined by dividing the number of hatched eggs by the total number of yolked eggs laid.

Embryos and dead-in-nest hatchlings were sampled at excavation. In 2015, up to five each early-stage and late-stage embryos, and all dead-in-nest hatchlings, were collected from each nest. Preference for postmortem evaluations was given to individuals that had the least decomposition (absence of softening or maceration of tissues and slippage or sloughing of skin). As there were historically low numbers of nests in 2016, a census was conducted where all embryos >2 cm in length and all dead-in-nest hatchlings were sampled from each nest.

Necropsies were performed within 24 h of nest excavation, and when not performed immediately, the embryo or hatchling was stored at 5 C. During the necropsies, embryos were completely removed from the shell and the yolk sack was trimmed approximately 5 mm from the plastron and discarded. The plastron and pectoral muscles were removed to expose the coelomic cavity. Swabs were collected from the surface of the lungs for microbiologic culture of any individuals in 2015 that had gross lung lesions and of all individuals in 2016 except for those from the first nest excavated that year. Embryos and hatchlings were then fixed in 10% neutral-buffered formalin for at least 48 h.

A microbiologist with previous experience culturing leatherback hatchling tissues led the bacteriologic investigation (Miller et al. 2009). Swabs were stored at −80 C until bacterial cultures could be performed. Each swab was thawed to room temperature, placed into nutrient broth for 10 min, and briefly vortexed. Using a 10-μL loop, the sample was inoculated into culture plates in the following order: blood agar, tryptic soy agar, MacConkey agar, and Sabouraud's dextrose agar. Plates were incubated at 25 C for up to 5 d. Isolates were identified using a combination of conventional biochemical tests and a semiautomated bacterial identification system (Sensititre™, Thermo Scientific, Waltham, Massachusetts, USA). Fungal cultures on Sabouraud's dextrose agar were incubated for a minimum of 7 d. Where multiple samples came from the same nest, containing similar bacterial isolates, complete isolation and identification of colonies were performed on at least one representative hatchling in the nest.

Transverse sections of cranial thorax (including lung, liver, esophagus, vertebrae, spinal cord, skeletal muscle, axillary fat, and potentially thyroid or thymus), parasagittal sections of the caudal abdomen (including kidney, gonad, gastrointestinal tract, liver, spleen, and/or pancreas), and a sagittal section of head (including brain, salt gland, eye, bone marrow, and skeletal muscle) were selected from each formalin-fixed embryo or hatchling for histopathologic examination. Tissues were processed for histology using standard methods, sectioned at 4 μm in thickness, and stained with H&E. Ancillary histochemical stains were also performed when indicated by microscopic pathology. von Kossa's stain was used to confirm calcium deposition. Tissues affected with granulomatous inflammation were stained for infectious agents with Ziehl-Neelsen acid-fast, Brown-Brenn, and Brown-Hopps gram stains and with periodic acid-Schiff (PAS).

To evaluate the extramedullary hematopoiesis, a score of 0=none, 1=mild, 2=moderate, or 3=severe was given to each studied organ: lung, bone marrow, kidney, and heart. For bone marrow, the score was based on the percentage of bone marrow that was composed of hematopoietic cells: mild, <25%; moderate, 25–50%; and severe, >75%. For liver and heart, the score was based on the number of cell layers composing subcapsular or epicardial hematopoiesis: mild, ≤5; moderate, 6–10; and severe, >10. For kidney and lung, the score was based on the percentage of parenchyma affected in a 10× field: mild, <5%; moderate, 5–15%; and severe, >15%. The overall extramedullary hematopoiesis score was determined for each individual by generating the sum of all organ hematopoiesis scores.

Sex was histologically determined using differential features of cortical epithelial cells of gonads (Ceriani and Wyneken 2008). Where the cortex was clearly apparent as a distinct layer of simple cuboidal cells, gonads were classified as female. Where the cortex was obscure and comprised a simple layer of squamous cells, gonads were classified as male.

The prevalence of the different lesions (bacterial pneumonia, renal mineralization, and skeletal muscle necrosis) was compared between stage of development, beach, and year. To account for the nonindependence of eggs in the nest, generalized linear mixed models were used with a binomial link and the nest identification as the random variable. Organ hematopoiesis scores and the overall hematopoiesis score were compared between embryonal stages of development using a Kruskal-Wallis test followed by Wilcoxon signed-rank test in case of significance. All data analyses were carried out using R software (R Development Core Team 2017). The significance threshold was set at 0.05.

In 2015, 70 adult female emergence events by 22 females were observed; in 2016, 31 adult female emergence events by 10 females were observed. In 2015, 62 nests were identified and marked and 10 were excavated. Nests that could not be located for excavation were typically not observed during oviposition, and thus did not have accurate location data, were lost due to beach erosion or had lost or misplaced stakes. All of the 10 nests that were excavated had embryos or hatchlings and were sampled for our study (Table 1). In 2016, 29 nests were identified and marked. Nine nests were excavated in 2016 and five were sampled. All unhatched eggs in the remaining nests showed no growth or signs of development and could not be sampled. Mean hatch success in 2015 was 5.5% overall (Stewart 2019) and was 11.0% (SD: 14.7; range: 0–41.9%) for the nests that we sampled. Mean hatch success in 2016 was 7.3% overall (Stewart 2019) and 13.3% (SD: 20.7; range: 0–48.8%) when considering only the nests sampled for this study.

In total, 84 individual embryos or hatchlings from 15 nests were sampled over the study period (Table 1). In 2015, five nests were sampled on each of Keys and North Friars beaches. In 2016, three nests on Keys beach and two nests on North Friars beach were sampled. In 2015 and 2016, the mean (SD, range) percentage of eggs with no growth or signs of development per nest was 62.9% (30.8, 3.7–76.7%) and 45.4% (18.6, 10.7–50.9%), respectively; the mean percentage of early stage embryos per nest was 16.0% (12.0, 0.9–43.4%) and 31.4% (17.5, 3.5–45.1%), respectively; the mean percentage late stage embryos per nest was 8.6% (25.7, 1.1–79.8%) and 10.5% (17.1, 6.1–21.4%), respectively; and the mean percentage of dead-in-nest hatchlings per nest was 0.4% (3.0, 0–8.6%) and 0.9% (2.5, 0–6.0%), respectively. The median number (range) of individuals sampled per nest was two (2–4) in 2015 and seven (4–24) in 2016.

Histologic assessment was completed on the following tissues: skeletal muscle (n=84), skin (n=84), bone marrow (n=84), kidney (n=83), heart (n=83), esophagus (n=83), lung (n=82), gastrointestinal tract (n=82), brain (n=78), liver (n=76), gonad (n=68), thymus (n=7), spleen (n=5), and pancreas (n=2). No lesions were identified within brain, esophagus, trachea, yolk sac, spleen, pancreas, or thyroid. Pathology was identified in 38% (32 of 84, 95% confidence interval [CI]: 28–49) of individuals (Table 2), including skeletal muscle degeneration and necrosis, bronchopneumonia, and renal mineralization. In 21% (18 of 84, 95% CI: 13–31%) of individuals, the pathology we observed was of sufficient severity or physiologic significance to explain death.

Renal mineralization affected 24% (20 of 83, 95% CI: 15–34%) of individuals, including 26% (6 of 23) in 2015 and 23% (14 of 60) in 2016. It was not grossly identified in any case. Histologically, renal mineralization consisted of basophilic granular deposits in the tubular lumina and occasionally in epithelial cells (Fig. 1). The basophilic material stained black with von Kossa's stain, indicating mineral composition. There was no significant association between renal mineralization and bacterial pneumonia (P=0.621), study year (P=0.572), or beach (P=0.168). The prevalence was lower in late-stage embryos (14%, 9 of 63) relative to early-stage embryos (64%, 7 of 12; P =0.010). The nine nests in which renal mineralization was identified were from at least five different mothers (for the other four nests maternal identification was unknown).

Bacterial pneumonia was observed in 12% (10 of 82, CI: 6–21%) of samples examined, including 17% (4 of 23) in 2015 and 10% (6 of 59) in 2016. Forty percent (6 of 15) of study nests were affected. Five individuals (three in 2015 and two in 2016) showed grossly evident nodules in their lungs (Fig. 2). The six nests having bacterial pneumonia were from at least three different mothers (for the other three nests, maternal identification was unknown). We found no association between the presence of turtles with pneumonia in a nest and nesting beach (P=0.333) and study year (P=0.391). Occurrence of pneumonia was not significantly different among natural versus relocated nests (P=0.067).

Histologically, bacterial pneumonia was similar among cases, consisting of bilateral, multifocal, heterophilic, and granulomatous bronchopneumonia (Fig. 3). In all cases, the severity of the lesion was sufficient to explain death and gram-negative rods were identified within inflamed foci using Brown-Hopps stain. Acid-fast positive bacteria were not observed in any case. The PAS staining failed to demonstrate fungi in these cases.

Skeletal muscle degeneration and necrosis were observed in 7% (6 of 84, 95% CI: 3–15%) of individuals, including 0% (0 of 14) in 2015 and 10% (6 of 60) in 2016, although the difference in prevalence between study years was not statistically significant (P=0.998). There was no association found between lesion presence and nesting beach (P=0.666) or embryo stage (P=0.952 for late-stage embryo and 0.949 for hatchling). In no case was skeletal muscle degeneration and necrosis identified grossly. Histologically, affected myofibers were present as groups or individuals and had swollen, vacuolated, condensed, fragmented, or hypereosinophilic cytoplasm, and in some cases, nuclear pyknosis or karyorrhexis (Fig. 4). The lesion was mild in all the specimens examined.

Extramedullary hematopoiesis (EMH) was observed in lung, liver, kidney, or heart of all individuals, most consistently in lung and liver, but it was also marked in kidney (Table 3). In addition, bone marrow was hyperplastic in 95% (80 of 84) of samples. Microscopically, EMH consisted of granulocytic cells infiltrating the tissue. Granulocytes were in sheets in the capsular surface of the liver or epicardium (typically near the heart base) or in poorly formed aggregates in the interstitium of the lung (specifically pulmonary trabeculae) or kidney. The overall hematopoiesis score significantly decreased between late stage and hatchling stage (P<0.001). For dead-in-nest hatchlings, hematopoiesis was significantly less severe in the lung (P=0.011), the liver (P=0.012), and bone marrow (P=0.029) relative to late-stage embryos.

Other lesions occurred infrequently. Cardiac mineralization affected 2% (2 of 83, 95% CI: 0–6) of individuals, an early- and late-stage embryo from the same 2016 nest, and both embryos were concurrently affected by renal mineralization. Pustular dermatitis, associated with mixed bacteria, was seen in a single late-stage embryo from North Friars beach in 2015. Epidermal necrosis was seen in a single late-stage embryo from Keys beach in 2015, but was not associated with histologically discernible infectious agents. Gastrointestinal tract hemorrhage was seen in a single dead-in-nest hatchling from Keys beach in 2016.

Hepatic lipidosis was observed in 100% (76 of 76) of embryos and hatchlings. It was severe (n=4), moderate (n=57), or mild (n=13) and consisted of distinct vacuoles that did not stain with PAS, suggestive of lipid rather than glycogen storage.

Gonad was identified histologically in the sections prepared from the caudal abdomen in 81% (68 of 84) of individuals. All were classified as female (Fig. 5).

Rathke's gland was microscopically identified in 61% (51 of 84) of individuals in the sections prepared from the cranial thorax region and was located within axillary skeletal muscle. It was composed of plump cuboidal epithelial cells arranged into solid nests of loosely associated cells (Fig. 6). Cells were variably vacuolated and/or contained bright eosinophilic cytoplasmic globules.

We isolated multiple bacterial species from lung swabs from individuals with pneumonia (Table 4). Pseudomonas spp. were often isolated from individuals from nests with and without bacterial pneumonia. Patterns of bacteria that were isolated were not apparent when comparing beach and year. Fungal cultures were negative in all cases.

This is the first study to describe the pathology of leatherback embryos, despite embryo mortality being problematic for the species globally. We described several lesions that may have contributed to perinatal mortality in leatherbacks in St. Kitts. The most common lesions that we observed in the St. Kitts' population were renal mineralization, bacterial pneumonia, and skeletal muscle degeneration and necrosis, respectively. These lesions were described at a differing prevalence in leatherback dead-in-nest hatchlings on Florida's Atlantic coast (Miller et al. 2009). Miller et al. (2009) also described several lesions we did not observe, including nephritis, renal tubular degeneration, cardiac degeneration, and pulmonary edema and thrombosis. Pathology of perinatal leatherbacks is not homogenous across nesting locations and emphasizes the importance of conducting pathologic investigations in areas where hatch success is low.

Renal mineralization, the lesion that we most commonly observed, occurred in all developmental stages evaluated. It was associated with cardiac mineralization in two cases, suggestive of metastatic calcification associated with hypercalcemia. Calcium homeostasis is not well understood in reptiles, particularly during embryonic development. The embryo obtains most of its calcium from yolk and the eggshell, and the renal distal tubule is a primary site of calcium absorption (Bustard et al. 1969; Stewart and Ecay 2010). Hypercalcemia in nesting females may result from high dietary calcium, hypervitaminosis D₃, and can be observed during vitellogenesis (Stacy and Innis 2017). Affected nests were derived from at least five different mothers, and elevations in serum calcium and phosphorus were unusual in nesting females evaluated in St. Kitts from 2006 to 2016 (Stewart 2019); thus, maternal influence seems unlikely. Other causes of hypercalcemia in reptiles include renal failure, granulomatous disease, and osteolytic lesions (Campbell 2014). Although granulomatous disease and osteolytic lesions could be excluded based on histopathology results, primary renal tubular injury was possibly present and was potentially obscured by mineralization and decomposition. Renal tubular degeneration and necrosis were not observed in our study but are prevalent among Floridian hatchlings and have been theorized to be due to dehydration during low-humidity incubation (Miller et al. 2009). The prevalence of renal mineralization we found differed between the two nesting beaches, suggesting an environmental influence on its development. Leatherback embryos take up greater amounts of calcium in environmental conditions having lower temperatures or moisture (Bilinski et al. 2001). Future studies evaluating calcium levels in maternal serum, and eggshell and yolk, as well as nest humidity during incubation could be informative for understanding the pathogenesis of renal mineralization.

Bacterial pneumonia is prevalent among dead-in-nest hatchlings in St. Kitts and Florida's Atlantic coast (Miller et al. 2009). The lesion was grossly evident as pulmonary nodules in only 50% of the St. Kitts' cases, emphasizing the importance of conducting histopathology for accurate diagnosis. Removal of organs and careful examination before fixation for histopathology may facilitate pneumonia identification. Intralesional bacteria consisting of gram-negative rods appeared histologically similar in all instances of pneumonia in our study. Culturing samples collected under field conditions from a decomposing nest environment proved challenging, and no bacterium was consistently isolated from affected individuals. Of the bacteria that we cultured, Pseudomonas and Ralstonia are gram-negative rods and opportunists that could potentially be the causative agent, and Pseudomonas is also prevalent among Florida hatchlings (Miller et al. 2009). The pattern of bronchopneumonia we described was indicative of an inhaled route of bacterial infection. Bronchopneumonia may be a common consequence for sea turtles hatching in a decomposing bacteria-rich nest environment. Wyneken et al. (1988) suggested that a negative correlation exists between the number of bacterial species isolated from a nest and hatch success. We also observed bacterial pneumonia in late-stage embryos, which was surprising, as respiration is thought to occur endogenously in unhatched reptiles (Ackerman 1981). However, it seems bacteria are capable of penetrating sea turtle eggshells (Al-Bahry et al. 2009) and may descend airways to infect embryonic lung. All nests with bacterial pneumonia that we examined were from different mothers where maternal identity was known, making vertical transmission seem less likely. Further study is needed to determine the significance of this disease to perinatal leatherbacks globally and to identify factors that may predispose to infection, such as immunosuppression brought on by high temperature or environmental pollutants. Pneumonia associated with gram-negative bacteria is also highly prevalent in American alligator (Alligator mississippiensis) late-stage embryos and hatchlings from a region contaminated with organochlorine pesticides (Sepúlveda et al. 2006), a pollutant known to cause immunosuppression.

Degeneration and necrosis of skeletal muscle, and to a lesser extent cardiac muscle, were predominant findings in Florida hatchlings; selenium deficiency was a suspected etiology (Miller et al. 2009). A later study showed hatchling selenium levels and selenium:mercury ratios positively correlate with hatch success (Perrault et al. 2011). The prevalence of skeletal muscle degeneration and necrosis was much lower in our study, suggesting that the same etiology did not impact the nesting population of leatherbacks in St. Kitts and that the pathogenesis of this lesion in leatherback may be multifactoral. Alternative explanations for this lesion could include high nest temperature, exertion or distress in the egg or nest, and vitamin E deficiency.

Infection with Fusarium solani, a widespread saprophytic fungus, is implicated as a cause of mortality in developing loggerhead sea turtles (Caretta caretta; Sarmiento-Ramírez et al. 2014). However, we did not isolate this organism, nor did we identify gross or microscopic lesions suggestive of fungal infection.

Extramedullary hematopoiesis was ubiquitous in our study population and especially involved the lung, liver, kidney, and heart. This is a common finding in Floridian hatchlings and posthatchlings (Miller et al. 2009) and is also widespread in avian embryos (Rideout 2012). It seems that this change represents a normal component of embryonal development for leatherbacks that begins to subside at the hatchling stage. Hepatic lipidosis was also pervasive in the study population and not considered to represent pathology; rather, it is likely that the vacuoles represent energy stores required for hatching and early survival.

We found an unusual gland in the axillary musculature that was morphologically consistent with Rathke's gland (Plummer and Trauth 2009), previously described in loggerhead, Kemps Ridley (Lepidochelys kempi), green (Chelonia mydas), and hawksbill (Eretmochelys imbricata) sea turtles (Weldon and Tanner 1990; Wyneken and Witherington 2001), but not in leatherbacks. The function of Rathke's gland is unknown, but it is thought to produce a lipid exudate that may be used to repel predators when frightened (Plummer and Trauth 2009) or that may have an antifouling or antimicrobial function (Wyneken and Witherington 2001).

Comprehensive histologic evaluation allowed us to assess sex ratios in the sea turtles that we examined. Interestingly, the population was overwhelmingly female. The ratio of male-to-female turtles is determined by incubation temperature, where temperatures >30 C during the middle third of embryonic development dictate female sex (Rimblot et al. 1985; Binckley et al. 1998). Therefore, the totality of females in the study population likely reflected high nest temperature. Coincidentally, hatch success declines as mean incubation temperatures exceed 29 C (Garrett et al. 2010; Howard et al. 2014). Temperature probes placed within the clutch of the middle of 12 nests in St. Kitts in 2015 and 2016 showed mean nest temperature throughout incubation was 33.2 C (range: 19.9–44.4 C; Stewart 2019), suggesting that temperature is playing a critical role in female-skewed sex ratios and possibly the high degree of embryo mortality observed in St. Kitts.

Most mortality in St. Kitts' leatherback nests occurred at a very early stage where signs of embryo development were not discernible, consistent with the findings of others (Rafferty et al. 2011). Leatherback eggs that fail to develop past this early stage did so due to embryo mortality rather than infertility (Bell et al. 2004). Unfortunately, classical tools of pathology have little application to early embryonal death because of the degree of postmortem decomposition that affects such eggs by the time the nest is excavated and because primordial embryos lack ordered tissues with well-defined responses to injury. Moreover, whole-nest egg failure at an early stage is likely to reflect unfavorable environmental conditions, and further research is needed to determine whether derangements of nest moisture, temperature levels, or other environmental factors are the basis for such nests.

In conclusion, we identified significant pathology that may result in perinatal mortality of leatherbacks and demonstrated that pathology assessment was an important component of conservation programs monitoring hatch success. Research is needed to understand factors associated with the occurrence of these lesions, and future pathology studies could aim to pair histopathology with contaminant exposure and assessment of nest environment conditions, including temperature and humidity. Also, a comparative study of other populations with differential hatch success would help elucidate the extent to which these lesions impair hatch success, regionally or globally.

This work was funded by The Morris Animal Foundation Veterinary Student Scholars program (D17ZO-604) and an intramural grant from the Center for Conservation Medicine and Ecosystem Health at Ross University School of Veterinary Medicine. The authors thank St. Kitts Department of Marine Resources for support of this project, David Hilchie for assistance with histology and histochemical stains, and Randel Thompson and Maurice Matthew for technical assistance during postmortem examinations. We thank the volunteers and staff of the St. Kitts Sea Turtle Monitoring Network, in particular Natalia Lord, Victoria Maroun, Jenna Strapple, Joseph Keeton, Nahillyl Santiago, Kristen Mader, Melissa Martin, Katelin Radosevich, Theophilus Taylor, Kevin Fahie, and Gary Buckles.