The hawksbill turtle Eretmochelys imbricata is listed as “critically endangered” on the Red List of International Union for Conservation of Nature; it requires in situ and ex situ conservation. To establish an effective captive breeding technique for the hawksbill turtles, its reproductive parameters and behavior in captivity, and the physiological changes, including hormonal changes, that lead to egg laying must be understood. Here, we observed the mating behavior of male (n=5) and female (n=3) hawksbill turtles (housed in a tank) from 2003 to 2007. We regulated the tank water temperature and daylight hours to simulate their natural environment. We sampled their blood at regular intervals to measure plasma concentrations of minerals and sex hormones. One (No. Ei21) of the female turtles was reproductively more active than others. Further, males mounted it more frequently than others. Subsequently, the female (Ei21) laid eggs twice, on 10 May and 8 July 2003. Females including Ei21 showed annual changes in 17β-estradiol and progesterone (P4) levels, whereas males showed annual changes in testosterone levels. Plasma calcium (Ca) level peaked in only the egg laying turtle (Ei21); however, inorganic phosphorus did not show a sharp increase. In 2005, as in 2003, Ei21 underwent similar physiological changes. We conclude that this phenomenon—of rapid rise in plasma P4, a hormone elevated after ovulation, followed immediately by a rise in plasma Ca level—is likely involved in the eggshell formation after ovulation. In future, we will further investigate these phenomena in hawksbill turtles.
Introduction
Of the current seven extant species of sea turtles in the world, three species lay eggs in Japan (Uchida and Nishiwaki, 1982). Among them, the hawksbill turtle, Eretmochelys imbricata (Linnaeus, 1766), has declined significantly due to impacts associated with human activities and is the less abundant than the green turtle, Chelonia mydas, and the loggerhead turtle, Caretta caretta. Therefore, in 1986, the hawksbill turtle was designated “critically endangered” on the Red List of the International Union for the Conservation of Nature (IUCN) (Mortimer and Donnelly, 2008). Research on breeding is necessary to protect endangered species (Kawazu et al., 2014, 2015); moreover, emphasis should be placed on establishing breeding techniques to advance further research. The breeding environments of endangered sea turtles, including temperature and daylight conditions, influence reproductive parameters such as clutch frequency and size and internesting interval. Providing an indoor breeding environment will greatly contribute to research on the reproductive physiology of sea turtles. The Port of Nagoya Public Aquarium (PNPA, Nagoya, Japan), which is equipped with well-stocked breeding tanks, is one aquarium that has successfully bred sea turtles (Kakizoe et al., 2013). In this study, we used PNPA's facilities to investigate the reproductive behavior and physiology of the hawksbill turtle.
Both male and female reptiles undergo physiological changes which help them lay eggs during the reproductive phase (Krohmer et al., 1987; Rostal et al., 2001; Sacchi et al., 2017; Zena et al., 2019). Increased concentration of the sex hormone testosterone (T) induces mating behavior in males (Sereau et al., 2010; Thompson et al., 2023). Meanwhile, changes in sex hormones, such as 17β-estradiol (E2) and progesterone (P4), induce egg laying in females (Guillette et al., 1991; Rostal et al., 2001; Al-Habsi et al., 2006; Shankey and Cohen, 2024). Like in other vertebrates, E2 stimulates synthesis of the calcium (Ca) binding protein vitellogenin in the liver of reptiles (Tinsley, 1985; Heck et al., 1997; Currylow et al., 2013). Additionally, in reptiles, E2 elevates plasma Ca and inorganic phosphorus (iP) levels by bone resorption (Magliola, 1984; Sheehan et al., 1999). Taken together, Ca metabolism is closely associated with reptilian reproductive physiology.
To maintain a constant blood pH, turtles store divalent cations such as Ca and magnesium in their carapace (Warburton and Jackson, 1995; Jackson, 1997). The turtle embryos use Ca from the eggshell and egg yolk for skeletal and carapace development (Simkiss, 1962; Sahoo et al., 1998). Thus, Ca metabolism likely plays a more significant role in reproductive physiology of turtles than in other reptiles. However, the relationship between Ca metabolism and reproduction in turtles has not been fully studied. Unravelling these relationships could help improve breeding of endangered turtles.
Using PNPA's facilities, here, we aimed to establish a captive breeding method for the hawksbill turtle by examining their reproductive behavior in captivity while simultaneously examining the physiological changes (including hormones and minerals) that lead to egg laying.
Table 1.
Body size and mass of turtles used in the experiment
Materials and Methods
Animals
Hawksbill turtles housed in the PNPA were studied here. Hawksbill turtle reproduction was studied using three female (tag-numbers: Ei15, Ei21, Ei27) and five male (tag-numbers: Ei16, Ei17, Ei20, Ei25, Ei26) turtles. Size and weight of these turtles is shown in Table 1.
We ensured that the turtles were not traumatized by the PNPA veterinarians and staff during blood sampling. This study followed the ethical guidelines for animal exhibition and research issued by Japanese Association of Zoos and Aquariums, and was approved by the board of directors of the PNPA.
Experimental conditions
We observed the mating behavior of three females and five males housed together in a tank (Fig. 1). The tank is connected to a sandy beach, where the turtles can come ashore to lay eggs. The turtles were housed under conditions similar to natural environmental temperatures and daylight hours during all of experimental period (2002 to 2007). The tank water temperature was maintained to mimic the annual temperature variations as flows: (1) the low water temperature period from 5 January to 3 March 2003 was simulated by maintaining the tank water temperature at 22.0°C; and (2) the high water temperature period from 13 May to 10 June, and from 11 June to 9 August 2003 was simulated by maintaining that tank water temperature at 28.0 and 29.0°C, respectively. Further, the tank water temperature during the rest of the year was maintained to mimic the gradually decreasing natural temperature variations in the rest of 2003.
Fig. 1.
Tank used in this study. (A) schematic representation; and (B) cross sectional views; The dark and light-gray areas indicate the tank and sandy beach, respectively. The tank is connected to the sandy beach, where the turtles can come ashore to lay eggs. Two high-sensitivity cameras were placed over the tank. Using these cameras, the behavior of turtles was recorded throughout the day (24 hours) using a time-lapse method at 12-second intervals. The coming ashore of turtles was observed using infrared sensors.
Two high-sensitivity cameras were installed in the tank at locations where they could cover certain areas from the top. The activity of turtles was recorded throughout the day (24 h) through the entire period of the study, using the time-lapse method (where each frame was 12 sec long). This section of work was conducted for approximately nine months from 1 January to 5 October 2003. Subsequently, recordings of Camera No. 2 were played back at 12x speed, and their mating behavior was checked in detail and summarized in chronological order; the occurrences (start and termination times) of the mounting behavior and the tag-number of the individuals involved was recorded. During the other periods, only reproductive parameters were observed.
Measurement of plasma hormones and minerals
Blood was sampled regularly once a month from the housed hawksbill turtles. Blood (10 ml) was collected from the jugular vein using a heparinized syringe (syringe with 18G needle, Terumo Corporation, Tokyo, Japan) to prevent clotting at around the same time of the day (between 13:00–15:00 h) every month. The blood plasma was immediately isolated using centrifugation (2,000 x g, 5 min) and stored frozen at –20°C. Subsequently, we measured plasma concentrations of hormones (E2, P4, and T) and minerals (Ca and iP). The concentrations of E2, P4, and T in the plasma samples were analyzed using chemiluminescent enzyme immunoassay kit by a vendor (Nagoya Rinsho Kensa Center, Nagoya, Japan). The detection limit of each hormone is 7.33 pg/ml for E2, 0.1 ng/ml for P4 and 12.4 ng/ml for T, respectively. The plasma Ca and iP concentrations were measured using various kit (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) with Fuji Drychem-3000 (Kakizoe et al., 2013).
Results
Behavioral analysis
The tank water temperature was maintained in the low water temperature period from January to March at 22.0°C and then in the high water temperature period from May to June at 28.0°C and from June 11 to August 9 at 29.0°C (Fig. 2). By controlling the temperature, we were able to induce reproductive behavior as follows. The results are shown in Fig. 3. The female Ei15 was tracked 41 times and mounted 41 times during the period, with the male Ei17 as its primary partner. The seasonal peak of mating behavior was in late March under the low temperature period. In particular, mating behavior by the male Ei17 was observed 15 times in March and 18 times in April, with a maximum of nine times per day. Thereafter, male Ei17 showed a smaller peak of tracking and mounting activity in late May (less frequent than in late March). Although the number and duration of being mounted during the period was higher for Ei15 female than those for other females, every mating event did not lead to successful copulation. For instance, video footage showed the refusal behavior, in which the Ei15 female blocked the insertion of the male genitalia with her hind legs (refusal position).
Fig. 2.
Annual variations in water temperature and daylight hours in 2003. The turtles were maintained in the tank under conditions similar to natural environment temperatures and daylight hours during the entire experimental period.
The female Ei21 was mated the highest number of times during the study period, totaling 769 times among the three individuals. Ei26 was the most frequent male partner, with a total of 571 mating events, followed by Ei16 with 96 mating events and Ei25 with 75 mating events, while Ei17 and Ei20 were less frequent, with 16 and 11 mating events, respectively. The seasonal peak of mating behavior was observed at the beginning of May, with a total of 218 mating events in May, with a daily maximum of 32 mating events. Thereafter, the frequency declined slightly, but peaked regularly, about every other month. Although mating behavior was initiated often, mounting occurred only five times, with only one mounting activity between Ei21 female and Ei26 male, lasting for >30 min. This event occurred on 20 March 2003 under the low-temperature period with Ei26's mounting was of longer duration, lasting approximately 140 min, and genitalia insertion was visually confirmed. The most focused site of Ei21 being bitten by Ei26 was the “neck” (146 times).
Fig. 3.
Mating of three female hawksbill turtles (A: Ei15; B: Ei21; C: Ei27) in 2003. Video recorded by camera was used to count the number of mating behaviors of male and female hawksbill turtles. Mounting times are indicated by diamond shapes. Mounting in which mating was successful is indicated by arrow.
The Ei27 female was observed being tracked at 55 times and mounted at five times during the study period. Some mountings lasted as long as 200 min. However, this individual was mated the least compared to the other two females.
Reproductive parameters of hawksbill turtles (Ei21) in the PNPA (2003–2007)
Table 2 shows the reproductive parameters of Ei21 from 2003 to 2007. In 2003, egg laying took place twice, with 59 days of the internesting interval. In the first nesting, 26 eggs were produced and did not hatch. However, in the next nesting, 90 eggs were produced, and the hatching success rate was 34.4%. In 2005, eggs were laid four times, with an internesting interval of 16 or 17 days. There were 112, 93, 154, and 154 eggs, and their hatching success rates were 0, 28, 29.9, and 22.1%, respectively. In 2007, eggs were laid twice, with an internesting interval of 28 days. There were 118 and 58 eggs laid, and their hatching success rates were 5.1 and 17.2%, respectively.
Changes in plasma T, E2, and P4 levels
Male plasma T concentrations decreased from low to high water temperatures, then increased during the winter low water temperature period, peaking around March (Fig. 4).
The E2 concentrations varied between the females. In 2002, E2 concentrations increased, starting from the low water temperate period to the nesting season (Fig. 5). In 2004, the E2 concentration began to increase in the Ei15 female at the beginning of the low temperature period, but remained low in the Ei21 and Ei27 females (Fig. 5).
The variation in P4 concentrations in the three females is shown in Fig. 6. All three individuals showed a change in P4 concentration from the low water temperature period to the end of the nesting season. In particular, a rapid surge-like increase in progesterone concentration (9 and 15 April 2003) was observed in Ei21, followed by egg laying on 10 May 2003. Thereafter, the second peak of plasma P4 in Ei21 was observed on June 14, which was followed by egg laying on 8 July 2003. The number of eggs laid on the 10 May and 8 July 2003 was 26 and 90, respectively.
In 2002, an increase in T concentrations in females was observed as the tank water temperature changed from that at the low water temperature period to that at the end of the nesting season (Fig. 7). In the winter of 2003, the concentrations of T in Ei15 female began to increase at low water temperatures but remained low in the Ei21 and Ei27 females (Fig. 7).
Table 2.
Reproductive parameters of hawksbill turtle (Ei21) in the Port of Nagoya Public Aquarium (2003–2007)
Fig. 4.
Changes in plasma testosterone (T) concentration in the five male hawksbill turtles in 2002 to 2004.
Fig. 5.
Changes in plasma 17β-estradiol (E2) concentration in the three female hawksbill turtles in 2002 to 2004.
Noting the nesting in Ei21, changes in the plasma concentrations of T, P4, and E2 are summarized in Fig. 8A. In addition, changes in plasma concentrations of T, P4 and E2 from 2004 to 2005 are shown in Fig. 8B. In 2003, T, P4, and E2 were highly variable and peaked at several times. In 2005, on the other hand, the major peaks of plasma T and P4 levels occurred only from April to May, when water temperature began to rise.
Changes in plasma Ca and iP levels
The changes in plasma Ca concentrations are shown in Fig. 9. Plasma Ca concentrations were almost constant throughout the year in males. In females, high plasma Ca concentrations were maintained from autumn until the end of the nesting season. Interestingly, the Ei21 female (she laid eggs) had markedly elevated plasma Ca concentrations on 15 January 2003. On the other hand, plasma iP levels in both male and female turtles were variable, but did not show periodicity (Fig. 10).
Fig. 6.
Changes in plasma progesterone (P4) concentration in the three female hawksbill turtles in 2002 to 2004.
Fig. 7.
Changes in plasma testosterone (T) concentration in the three female hawksbill turtles in 2002 to 2004.
We focused on the spawning Ei21, and the changes in plasma Ca and iP concentrations are summarized in Fig. 11A. Changes in plasma Ca and iP concentrations from 2004 to 2005 are also shown in Fig. 11B. In 2003, plasma Ca concentrations had a pronounced peak, but no peak was observed in the iP. In 2005, on the other hand, plasma Ca concentrations increased gradually and showed high values from October to February.
Discussion
Here, we succeeded in the captive breeding of the endangered hawksbill turtle by mimicking natural environmental conditions (in terms of the water temperature and daylight hours) in an artificial rearing tank. Male plasma T levels peaked around March, encouraging male mating behavior, and mounting was observed. Mating with Ei26 male on 20 March 2003 allowed the Ei21 female to successfully lay eggs twice, on 10 May and 8 July 2003, on the sandy beach. In 2005, egg laying also occurred four times, with internesting intervals of 16 or 17 days. In 2007, eggs were laid twice, with an internesting interval of 28 days. The internesting interval of nesting hawksbill turtles recorded in the wild was 11–28 days (Miller, 1997); thus, the internesting interval of our turtles is solidly within this range. Therefore, we can conclude that the experimental conditions created by us in terms of water temperature and daylight hours (regulated to mimic the natural environment) were suitable for the reproductive physiology of the hawksbill turtle.
Fig. 10.
Changes in plasma inorganic phosphorus (iP) concentration in the three female and five male hawksbill turtles in 2002 to 2004. Male plasma iP concentrations are shown as averages of five individuals in 2002 to 2004.
Fig. 11.
Changes in the plasma calcium (Ca) and inorganic phosphorus (iP) concentrations in the spawning hawksbill turtle (Ei21) during 2002 to 2004 (A) and 2004 to 2005 (B).
Kobayashi (2012) reported that plasma T levels in hawksbill turtle reared in a tank on Ishigakijima Island (Okinawa, Japan) increased between March–May. Consistently, we found that male plasma T peaked in March, which possibly induced the male mating behavior. Plasma T and E2 concentrations in females showed similar changes because E2 is metabolized and biosynthesized from T (Shankey and Cohen, 2024), like in other turtles such as yellow-blotched map turtle, Graptemys flavimaculata (Shelby et al., 2000) and common snapping turtle, Chelydra serpentina (Mahmoud and Licht, 1997). On the other hand, on 3 May 2003, plasma E2 levels were elevated in the Ei21 female. The phenomenon of elevated plasma E2 concentrations prior to mating has been observed in mature female wild loggerhead turtle (Wibbels et al., 1990) and in Kemp's ridley turtle, Lepidochelys kempii under captive conditions (Rostal, 2004), consistent with the results of this study. Furthermore, P4, a hormone that rises after ovulation, was elevated in the present study. To the best of our knowledge, our study is the first observation of the annual cycle of P4 in captive nesting hawksbill turtles. P4 elevation observed here was similar to that seen in other turtles such as the loggerhead turtles (Guillette et al., 1991) and the green turtles (Al-Habsi et al., 2006).
A unique feature of this study is that we also focused on Ca metabolism. E2 causes bone resorption, and E2 administration increases iP levels in the plasma along with Ca increase, in the three-toed box turtle, Terrapene triunguis (Magliola, 1984) and the red-eared slider, Trachemys scripta elegans (Sheehan et al., 1999). Here, the plasma Ca levels were markedly elevated in the nesting Ei21 female, but the iP levels were not. Thus, it is likely that the increase in plasma Ca concentration in Ei21 is not an E2-induced effect. Moreover, Ca concentration increases in the uterus of female snapping turtles (C. serpentina) due to eggshell formation after ovulation (AlKindi et al., 2006). Indeed, we observed that subsequent to P4 rise, the plasma Ca concentration increased at the first time of egg laying in both 2003 and 2005 (Fig. 12). Particularly in 2003, we were able to detect a peak in plasma Ca levels. However, plasma Ca concentrations did not increase noticeably during the second P4 peak in 2003 (Fig. 12). The Ca concentration in the uterus may have lasted long, since there was no remarkable increase in plasma Ca before the second nesting. In 2005, the second Ca peak (May to July) coincides with laying eggs. We surmise that Ca for the elevated plasma Ca levels may be sourced from other supply sources. We plan to investigate this hypothesis in more detail in the future.
Fig. 12.
Changes in plasma progesterone (P4) and calcium (Ca) concentration in the spawning hawksbill turtle (Ei21) during 2002 to 2004 (A) and 2004 to 2005 (B).
The Ca metabolism related findings in this study are interesting. Therefore, we would like to focus on Ca metabolism in future studies, especially on the Ca-regulating hormone calcitonin. In fish, calcitonin is involved in reproductive physiology (Suzuki et al., 2004; Kuroda et al., 2023), and is elevated just before ovulation (Norberg et al., 1989). We have previously determined the structure of reptilian calcitonin and reported that its sequence is similar to that of fish (Suzuki et al., 1997). Further, we have developed an enzyme-linked immunosorbent assay (ELISA) to measure the plasma calcitonin level in fish (Suzuki et al., 2004; Kuroda et al., 2023). Preliminary experiments revealed that our ELISA can measure calcitonin in the blood of loggerhead turtle. In future, we would like to investigate the role of calcitonin in the reproductive physiology of the hawksbill turtles.
Here, plasma hormone levels were altered in male and female hawksbill turtles under captive conditions, leading to successful mating, and nesting. Further, the understanding of the reproductive biology of hawksbill turtles has improved due to data accumulating from other recent studies. Head-starting—in which hatch-lings are raised in captivity for a certain period of time (few days to several years) before being released into the wild—has been attempted for green turtles, Kemp's ridley sea turtles, and hawksbill turtles, as a means of restoring sea turtle stocks (Sato and Madriasau, 2001; Fontaine and Shaver, 2005; Jualaong et al., 2021). We would like to continue our basic research for head-starting to increase the population of endangered sea turtles, including the hawksbill turtle, which are listed on the IUCN Red List.
Acknowledgments
This study was partially supported by grants to M.K. (Japan Bekko Association) and to M.U. (Nippon foundation).
© 2025 by The Herpetological Society of Japan
