Many turtles and tortoises have temperature-dependent sex determination and the nonlethal sexing of juveniles becomes increasingly important in ecological studies and in conservation and recovery projects of threatened species. Endoscopy is currently the only 100% accurate nonlethal method available that can be used to sex juvenile turtles that do not show external sexual dimorphism (Rostal et al. 1994; Kuchling 1999; Wibbels et al. 2000).

Endoscopy is an invasive procedure that requires training and, in many countries, specific permits from relevant authorities. Its use in turtle studies is thus not widespread. A variety of noninvasive sexing methods for juvenile turtles have also been tested (karyotype, H-Y antigen, DNA fingerprinting, serum testosterone levels), but it is still essential to validate those methods for each species and population with samples from turtles of known sex (Wibbels et al. 2000). This makes endoscopy particularly important. Despite this central role of endoscopy for the development of other sexing techniques in turtles, no detailed description of the endoscopic imaging of the various developmental stages of juvenile turtle gonads and accessory ducts is currently available.

A conservation program for the Madagascar freshwater turtle Erymnochelys madagascariensis has included a head-starting component of hatchlings collected from wild nests of which the temperature was monitored (García 1999; Kuchling 1999, 2000). In order to investigate the effect of incubation temperature on sex determination in the wild nests (García et al., unpubl. data, 1999, 2000), all head-started E. madagascariensis juveniles in this program were sexed by endoscopy. The present paper describes the endoscopic images of the reproductive tracts of these juvenile E. madagascariensis of various size and age.

METHODS

Hatchlings of E. madagascariensis were collected from monitored wild nests in Ankarafantsika National Park and were raised in outdoor ponds of the Durrell Wildlife Conservation Trust's Turtle and Tortoise Breeding Facility at Ampijoroa in northwestern Madagascar. Pond design and captive management were outlined by Kuchling (2000). All captive-raised juvenile turtles with an age of 3 years (n = 22, hatched January–February 1999, 92–375 g body mass) and 2 years (n = 113, hatched January–February 2000, 72–192 g body mass) and some 1-year-old juveniles (n = 5, hatched January–February 2001, 98–202 g) were examined endoscopically during early February 2002. In addition, some larger wild turtles that were caught as part of a population and reproductive study in Lake Antsilomba at Ankarafantsika National Park were also examined.

Body mass and standard morphological parameters, including carapace length (CL, to the nearest mm), were measured for all turtles. The turtles were anesthetized by intravenous injection (carpal sinus) of ketamine hydrochloride (30–40 mg/kg body mass). Optimum anesthetic depth was achieved after about 20–30 minutes. Both hind legs were pulled backwards and tied together. The left inguinal pocket and neighboring skin, shell, and leg were scrubbed with antiseptic soap and povidone-iodine. A 2.7-mm-diameter rigid Storz Hopkins 30° endoscope was inserted into the abdominal cavity through a stab incision in the lower anterior part of the inguinal pocket. In order to minimize postoperative buoyancy problems, the abdominal cavity was not insufflated. A battery-operated Storz cold-light projector was used as light source.

Gonads and accessory ducts were visualized, usually behind intestinal loops, and their appearance, color, and texture were noted. The wide-angle lens of the endoscope did not allow measurement or accurate estimation of the size of observed structures. All size estimates are crude approximations only. A digital camera with a macro function (Nikon Coolpix 995) was used for photo documentation. The eyepiece of the endoscope was custom-adapted to fit into the protective ring of the camera lens and photos were taken by holding the camera against the eyepiece.

The surgical procedure lasted 2–10 minutes (depending on whether photos were taken or not). After completion of endoscopy, the skin wound was sutured using 2 stitches of 4–0 vicryl. The turtles recovered from anesthesia 1–2 hours after the surgical procedure. The turtles were kept overnight individually in dry buckets after the endoscopic examination and the following day their diving behavior was observed at release into their ponds. The turtles generally swam and dove immediately without problems and were typically feeding shortly after their return to the ponds. Only 1 of the 140 returned juveniles floated tilted to the left side at the surface and therefore received a 10-day course of enrofloxacin (Baytril).

Turtles that died during captive rearing were preserved in formalin, but their condition was not suitable for histology because none were preserved immediately after death and their internal organs were already in various stages of decomposition. The preservation condition of the gonads of 6 turtles allowed their external measurement with calipers to the nearest 0.1 mm and, after squashing the gonads on glass slides, the measurement of the diameter of ovarian follicles and of tubuli seminiferi with a micrometer in the eyepiece of a microscope. These measurements were used to validate the size estimates determined during endoscopic examination.

RESULTS

The gonads of juvenile E. madagascariensis are thin and elongate and fixed to the dorsal part of the body cavity, very close to the kidneys, adrenal glands, and lungs. The gonads and other organs (epididymis, oviduct, kidney, adrenal, lung) can generally be viewed relatively clearly through translucent peritoneal membranes such as the mesentery and the coelomic serosa. Despite being attached to the dorsal coelomic wall by various membranes, gonads and reproductive tracts move and change their position easily relative to the kidneys, adrenals, and lungs (which have more or less fixed positions) when, for example, turtles are tilted from one side to the other during endoscopy. The gonads also move with the breathing movements of the lungs. Of the 140 examined captive raised juveniles, 75% were females and 25% males.

Females

The ovaries are enveloped by a peritoneum and attached to the dorsal wall of the coelomic cavity or to a membrane that separates them from the lungs (which are adherent to the carapace). The oviducts extend farther cranially than the ovaries and are ventral or lateral to the ovaries. They often cross ventrally over the posterior part of the ovary on the way to the cloaca.

The ovaries of 1-year-old juveniles are translucent, thin, elongate sheets with small transparent-whitish follicles (estimated diameter about 0.1–0.15 mm); their oviducts are relatively thin, straight and transparent-whitish (Fig. 1A). Small 2-year-old turtles (body mass 72–80 g) have ovaries of similar size, but the follicles are slightly larger and the oviducts are wider, although still straight and transparent-whitish (Fig. 1B). In slightly larger 2-year-old females the density and probably number of follicles increases (Fig. 1C, D) and continues to increase progressively without a change in the appearance of the follicles or oviducts (Fig. 1E, F) up to a body mass of about 180 g (which were the largest 2-year-old females examined). Until that stage the follicles appear as small, transparent, whitish flat discs of probably 0.2–0.25 mm diameter (Fig. 1G). In 3-year-old females, the oviduct is wider and thicker and more whitish, but still mainly straight (Fig. 1H), although the prospective glandular segment of the oviduct may sometimes show undulations; the ovaries further increase in size and the follicles are dense, slightly larger (probably up to 0.3 mm) and probably thicker and of whitish-yellowish color (Fig. 1H, I).

View Figure

• Download as Powerpoint
• Download Figure

Postmortem measurements of ovaries and follicles of E. madagascariensis that died during captive rearing demonstrate an increase in the size of ovarian follicles from 1 to 2 years of age (Table 1). The measurements largely agree with the size estimates determined during endoscopic examination of the captive-reared turtles and support their validity.

Table 1. Postmortem measurements of ovaries and ovarian follicles of female E. madagascariensis that died during captive rearing.

View Fullscreen

Larger (> 0.5 kg) and presumably older wild juveniles show further enlarged ovaries with larger, yellowish follicles (up to about 0.35 mm) and thicker, slightly convoluted glandular segments of the oviducts that are now white (Fig. 1J). The prospective uterine segments of the oviducts remain straight in juveniles of all sizes, but are thicker than the glandular portions in larger juveniles (> 0.5 kg). With increasing size of juvenile females, the size of the ovaries and oviducts increases further and the follicles become larger (up to about 1 mm) and brighter yellow (Fig. 1K), although they are still flat discs and previtellogenic.

The ovaries of adult females during the nonbreeding year of their biennial reproductive cycle (Kuchling 1993), when no vitellogenesis takes place (Fig. 1L), appear very similar to those of the larger juveniles, but with many of the yellow previtellogenic follicles having diameters of about 1–2 mm; the oviduct of nonbreeding adult females has a convoluted, white glandular segment (Fig. 1L).

Table 2 summarizes the developmental traits observed in the female reproductive system.

Table 2. Traits of female reproductive organs according to age and/or body mass.

View Fullscreen

Males

The testes are bound to the kidneys by the mesorchium, which extends mediolaterally over the epididymis and kidney and is continuous with the body wall and the dorsal mesentery. Testes of the only examined large 1-year-old male (202 g) and of some small 2-year-old turtles (< 125 g) are small, elongate, half-roundish (in cross section) and of rose pink-whitish color; the epididymis is transparent, and the vas deferens transparent and barely visible (Fig. 2A). At close distance it is possible to see the small, whitish tubuli seminiferi and the slightly reddish interstitium of the testis (Fig. 2B) that gives the testis the overall rose pink color. In the other 2-year-old males, the testes are yellowish-orange (Fig. 2C) because the tubuli seminiferi increase in size and turn to a yellowish color while the interstitium remains reddish (Fig. 2D). In 2-year-old turtles > 130 g, the epididymis turns whitish-transparent (Fig. 2E) and with increasing body size the testes become progressively more roundish (Fig. 2F). In 3-year-old males, the testis is roundish in cross section and looks like an elongate sausage with pointed ends and the epididymis is thicker and the whitish, meandering ductus epididymidis is visible (Fig. 2G). In large juveniles the epididymis becomes bulky because of the coiled ductus epididymidis, and the vas deferens is thick and whitish (Fig. 2H). During the nonbreeding season, the epididymis of adult turtles is relatively large and whitish even though there are few or no sperm in the ductus epididymidis, the vas deferens is thick and white, and the testis in its regressive state is orange (Fig. 2I).

View Figure

• Download as Powerpoint
• Download Figure

Postmortem measurements of testes and tubuli seminiferi of E. madagascariensis that died during captive rearing are only available for a 2.5-month-old turtle and a 22-month-old turtle (Table 3). A pronounced increase (doubling) in the diameter of the tubuli seminiferi occurs between these ages. This supports the endoscopic observation of a size increase of the tubuli seminiferi, correlated to a change of their coloration from whitish to yellowish, in the 2-year-old age group.

Table 3. Postmortem measurements of testes and tubuli seminiferi of male E. madagascariensis that died during captive rearing.

View Fullscreen

DISCUSSION

Endoscopy is a well-established technique to assess the reproductive condition of turtles and tortoises (Wood et al. 1983; Limpus and Reed 1985; Kuchling 1989, 1998, 1999; Owens 1999; Schildger et al. 1999). The endoscopic examination of the body cavity of turtles is often referred to as laparoscopy (e.g., Wood et al. 1983; Limpus and Reed 1985; Rostal et al. 1994; Wibbels et al. 2000), but that term is more appropriate for mammals and humans.

Endoscopy has been successfully used to sex captive-raised juvenile desert tortoises, Gopherus agassizii (Rostal et al. 1994), captive-bred juveniles of the highly endangered ploughshare tortoises, Geochelone yniphora (Kuchling and López 2000), and captive-bred juvenile Aldabra tortoises, Geochelone gigantea (Kuchling, unpub. data, 2000, 2004). Endoscopy has long been used to study the sexes of juvenile sea turtles (Wood et al. 1983; Limpus and Reed 1985; Owens 1999), but these studies mainly concerned larger juveniles, which were being recruited into the adult population, rather than small young turtles. I also found endoscopy useful to sex 37 juvenile Caretta caretta (440–1040 g body mass) that were washed up at Perth (Western Australia) beaches during winter storms in 1992 (Kuchling 2004). In the present study, the gonads of all examined E. madagascariensis down to a body mass of 72 g could be visualized by endoscopy and unambiguously identified as testes or ovaries. Endoscopy is clearly a valuable technique in threatened species conservation programs when accurate nonlethal sexing of juvenile tortoises and turtles is required.

Although Rostal et al. (1994) used endoscopy to sex G. agassizii with a size range of 28–1250 g body mass, they did not report any changes in the appearance of gonads and accessory ducts with changing age and/or body size. In the present study it was possible to establish progressive developmental changes of female and male reproductive organs with age in E. madagascariensis throughout the juvenile phase. These changes, apart from the challenges of surgery in smaller turtles, make the endoscopic sexing of older juveniles easier than the sexing of younger individuals. In E. madagascariensis, developmental steps of juvenile reproductive tracts seem to depend on age rather than size. For example, the reproductive tract of the smallest of the 3-year-old females was at the developmental stage of the other, larger 3-year-old females and not at the stage of 2-year-old or 1-year-old females, even though those were often double the body mass of the former. And the reproductive tract of the 1-year-old male of 202 g body mass still resembled that of some smaller 2-year-old turtles, but not that of 2-year-old turtles of similar size.

Female E. madagascariensis reach sexual maturity at about 260 mm CL (Kuchling 1988) and males at about 250 mm CL (Kuchling, unpub. data, 2002). The endoscopic images of the 228-mm female (Fig. 1K) and the 212-mm male (Fig. 2H) represent large juveniles before they enter the subadult stage (puberty). The gonads and accessory ducts of large juveniles already resemble very much those of adult turtles in their nonbreeding state (Figs. 1L, 2I). Female E. madagascariensis typically show a biennial reproductive cycle (Kuchling 1993) and their ovaries regress more dramatically during their nonbreeding period than do the ovaries of most freshwater turtles that show annual reproductive cycles (Kuchling 1999). During vitellogenesis and ovulation (and spermatogenesis and spermiation) the gonads and accessory ducts of adult E. madagascariensis show pronounced increases in size and changes in appearance, but for the comparative purposes of this study it seemed more appropriate to present adult reproductive organs in the state of regression.

Interestingly, in small and young E. madagascariensis up to 2 years of age, the female gonads and accessory ducts appear to be relatively small and poorly developed compared to males, and during endoscopy it often takes longer to clearly identify female reproductive tracts than male reproductive tracts. I found the opposite to be the case in G. yniphora, G. gigantea, and C. caretta; in small specimens of those species it is often more difficult to clearly establish male reproductive tracts than female reproductive tracts (Kuchling, unpub. data, 1992, 2000, 2004). The incidence of persisting oviducts in males (e.g., 56% in juvenile G. agassizii; Rostal et al. 1994) also seems to be lower in juvenile male E. madagascariensis. (I noted obvious persisting oviducts in only 3 out of the 35 juvenile captive-raised 2-year-old males [90–128 g], or in 8.6% of cases.) However, the oviducts in juvenile males turn progressively into less and less obvious thread-like ligaments and it is a question of interpretation how long these structures are considered to represent “persistent oviducts,” and they may have partly escaped my attention. In any case, the residual oviducts in juvenile males were never cause of any ambiguities in determining the sex.

With some experience, the endoscopic sexing of juvenile chelonians is accurate and straightforward. Rostal et al. (1994) endoscopically sexed G. agassizii as small as 28.3 g and found that small tortoises appeared more stressed during the endoscopic procedure than did larger ones and also took longer (up to 48 hours) to recover. In contrast to the present study, Rostal et al. (1994) did not anesthetize the tortoises and did not routinely close the incision by suturing. It may be more important in aquatic turtles to properly seal the incision site in order to avoid postoperative infections; however, because Rostal et al. (1994) reported that in 6% of cases the urinary bladder prolapsed through the incision site after the endoscopic sexing of G. agassizii, the routine closure of the incision site by suturing or other closure methodology seems also indicated in terrestrial tortoises. With careful aseptic techniques as described here, the routine prophylactic administration of antibiotics as described by Rostal et al. (1994) may be unnecessary. This is supported by the fact that only 1 of the 140 captive E. madagascariensis died during the 2 years following endoscopy and that all other 139 showed perfect health and growth rates. This very low mortality rate does not differ from that of captive juveniles that were not subjected to endoscopic examination.

Until this time, and throughout this study, I did not perform endoscopy in turtles smaller than 72 g in order to avoid exposing them to the stress and negative impacts of endoscopic surgery described by Rostal et al. (1994) for small tortoises, but any impact of the procedural protocol used in this study may be minimal and may also make the sexing of smaller individuals feasible and ethically justifiable. However, the small hatchling size of most freshwater turtles would still make the endoscopic sexing of freshly hatched turtles quite a challenge.