The use of oxytocin to induce egg laying (oviposition) in turtles has gained popularity since reports by Yntema (1964), Ewert and Legler (1978), and Ewert (1979). Today oxytocin is widely used by veterinarians to treat “egg bound” turtles, by breeders, and by researchers to obtain eggs for experimental purposes and headstart programs. Despite this widespread usage, the dosage recommendations still vary from 5 to 40 units/kg (DeNardo 1996; Ewert and Legler 1978). Since the article by Ewert and Legler (1978), there have been no additional published studies that evaluated the efficacy and safety of oxytocin induction in turtles.

This study was begun in 1978 as an attempt to narrow the range of these dosage recommendations and to see if sensitivity to oxytocin was species specific. Because of the limited and unpredictable numbers and species of animals available each season, I used an empirical approach over a 28-year period to gradually clarify what the most effective dose would be for each species evaluated.

More recently, Tucker et al. (2007), working with large numbers of wild Trachemys scripta elegans and following an experimental protocol, determined that a dose of 10 units/kg produced the best results with the first injection. There has only been 1 study that compared the viability of naturally occurring nest eggs and eggs obtained by oxytocin injection (Wilgenbusch and Gantz 2000) and that showed comparable hatching rates for Chelydra serpentina.

Oxytocin is a pituitary hormone of mammals. There are numerous preparations available for human and veterinary use, and it is available worldwide. Oxytocin is stable for years at room temperature and is very inexpensive. The reptilian and avian equivalents of oxytocin are arginine vasotocin (AVT) and mesotocin (Archer et al. 1972). AVT is 10 times more potent than oxytocin and 16 times more potent than mesotocin in isolated oviducts (LaPointe 1977). However, AVT is very difficult to work with because it breaks down rapidly at room temperature, it must be kept frozen until use, and it retains its potency only if diluted just before injection (increasing the potential for dosage errors). AVT is also expensive and more difficult to obtain than oxytocin because of legal requirements. Dosage recommendations vary widely in the literature: Lloyd (1990) suggested 0.01–1.0 ng/g, and Mahmoud et al. (1987) used 4.6 ng/kg with C. serpentina.

A complex interaction of steroid hormones, neurohypophysial peptides, prostaglandins, and direct neural modulation is necessary for coordinated oviposition (Guillette et al. 1990b). Because of this, use of exogenous AVT or oxytocin at the incorrect point in the reproductive cycle is relatively ineffective. We still do not have a complete understanding of how hormones, peptides, prostaglandins, and neural modulation interact, but some elements have been defined.

After ovulation, the corpus luteum secretes progesterone. The elevated progesterone levels have been shown to inhibit oviductal contractions in vivo in C. serpentina (Mahmoud et al. 1987). Similarly, exogenous progesterone blocks AVT activity in vitro in Chrysemys picta bellii (Callard and Hirsch 1976). Progesterone is also a potent inhibitor of prostaglandin F_2α (PGF) synthesis from the oviduct (Guillette 1990). The elevated progesterone levels characteristic of the early postovulatory phase allows the oviducts to remain quiescent while albumin and shell components are being secreted.

In C. serpentina, natural oviposition does not occur until progesterone levels drop after luteolysis (Mahmoud et al. 1988). A single injection of PGF causes luteolysis and a fall in progesterone levels within 24–30 hours but does not lead to oviposition, so some other factor is probably involved in AVT regulation as well (Mahmoud et al. 1988).

Oviductal tissues of C. picta, Sternotherus oderatus, and some other reptiles synthesize PGF. This synthesis of PGF in the oviducts of reptiles and birds is stimulated by AVT (Guillette 1990). AVT, PGF, and prostaglandin E₂ levels rise abruptly just before oviposition in 2 species of sea turtles (Figler et al. 1989; Gross et al. 1992).

In some lizards, the administration of a prostaglandin inhibitor can slow or prevent complete oviposition (Guillette et al. 1990a), further supporting the idea that oviposition is the result of a combined direct effect by AVT to increase oviductal tone and a second, indirect, effect produced via PGF that strengthens peristaltic contractions. A similar pattern occurs in mammals, with 2 distinct oxytocin receptors present in the uterus, one causing increasing tone and the other PGF secretion (Graves 1996). It has also been observed in a variety of species that AVT induced oviductal contractions operate through a different mechanism than PGF-induced activity (Cree and Guillette 1991).

Snapping turtles stressed by capture had elevated levels of estrogen and progesterone for a week after capture (Mahmoud et al. 1989). Because progesterone blocks the production of PGF and the activity of AVT and oxytocin, the stress of the early stages of captivity can block oviposition via progesterone secretion and/or direct beta-adrenergic stimulation.

It has been shown in some lizards, birds, and mammals that beta-adrenergic stimulation inhibits uterine/oviductal contractions by reducing both PGF production and response. But, in other lizards, beta-adrenergic stimulation inhibits the effects of AVT but not PGF (Cree and Guillette 1991). These effects can be prevented by using a beta-adrenergic blocker (Jones and Guillette 1982), and, in some reptiles, a beta-adrenergic blocker alone will induce oviposition (Gross et al. 1992). In the lizard Sceloporus virgatus, the combination of propranolol and PGF induced normal nesting behaviors (oviposition and nest guarding) after oviposition (Gross et al. 1992). A similar process occurs in humans (Zahradnik 1985). One of the treatment options for women with ineffective labor is to use a beta-adrenergic blocker to reduce the peripheral beta-adrenergic response to the stress reaction, thus enhancing the uterine contractions (Sanchez-Ramos et al. 1996). Therefore, blocking the effects of beta-adrenergic stimulation with propranolol (a commonly used beta-adrenergic blocker) could theoretically enhance the actions of AVT, PGF, or oxytocin in turtles.

Another approach to blocking peripheral beta-adrenergic activity is to use an agent that blocks anxiety centrally. Ketamine is a dissociative anaesthetic that is very popular in human pediatric medicine because of its safety, anxiolytic action, and sedative effects. If the activity of oxytocin or AVT is being blocked by centrally mediated anxiety, ketamine could hypothetically produce anxiolysis and enhance oviposition. For these reasons propranolol or ketamine was used in conjunction with AVT or oxytocin in parts of this study.

Methods

From 1978 to 2006, oviposition was induced a total of 245 times in 13 North American turtle species. Of the 245 inductions, 195 involved the use of oxytocin alone, 22 used AVT alone, 13 used a combination of oxytocin and ketamine, 8 combined propranolol and oxytocin, and 7 were given propranolol plus AVT. Calcium was not administered in any of the cases.

When ketamine or propranolol was used, it was administered intramuscularly (IM) at the same time as the oxytocin or AVT was injected intraperitoneally (IP). Because ketamine is cleared by the kidneys (Frye 1991), IM injections were administered into the muscles of the shoulder girdle. This avoids the first-pass effect through the kidneys that is characteristic of injections into the pelvic muscles in some reptiles.

Propranolol doses varied between 11 and 40 mcg/kg and were based on the recommended human dose at the time of 14 mcg/kg. There was no cardiac monitoring equipment available to determine the effectiveness of the beta-adrenergic blockade. Without monitoring equipment, it is difficult to know if the beta-adrenergic blockade was adequate to inhibit beta-adrenergic stimulation of the oviducts.

Animals for use in the study came from a variety of sources over the years. Most were obtained from people living in south-central Pennsylvania who found turtles crossing roads or wandering around their properties or golf courses. Others came from private collections, known nesting areas, or the edges of waterways in late May to early July. In the last decade of the study, all the subjects came from my own or other captive colonies of T. scripta elegans maintained in New Zealand.

Most of the wild animals were released back to the area of capture. In cases where that was unknown or the area was in jeopardy (new road development or construction), they were released in the nearest area judged to be relatively safe. Hatchlings were kept until the yolk sac was completely absorbed and then they were released to the area the female came from or her relocation site.

A 1-mL tuberculin syringe combined with a 27-gauge, 1 1/4-inch needle was used for all IP injections. IM injections were done with a 26-gauge, 3/8-inch needle. Before injection, the turtle was palpated for eggs. The technique used was to hold the carapace up and palpate both inguinal fossae simultaneously by placing the right middle finger (or the middle, fourth, and fifth fingers in larger turtles) in the right inguinal fossa while using the index finger and the thumb to support the shell. At the same time, the left hand was positioned in the same fashion in the left inguinal fossa. By rocking the turtle slowly, and pushing the middle fingers gently in and out, the eggs could be palpated easily (Fig. 1).

View Figure

• Download as Powerpoint
• Download Figure

When injecting, the animals were turned plastron up and head down at a 45-degree angle, with the carapace resting between the operator's thighs. While holding the back leg in full abduction and directing the inguinal pocket upward with one hand, the free index finger of the other hand was used to palpate the nearest egg. The needle was then directed slowly toward the nearest egg. Once the needle tip touched the egg, the needle was slowly withdrawn while the drug was injected (Fig. 2). The dose was divided into 2 injections, one given on each side. These adaptations reduced the chance of injecting the entire dose into the bowel or an egg.

View Figure

• Download as Powerpoint
• Download Figure

Some researchers (Tucker et al. 2007) prefer IM injections into the thigh because of concerns about the safety of IP injections. However, the dangers of delayed and immediate muscle contractures and neurologic damage from IM injections are well known in humans (Bergeson et al. 1982; Greenblatt and Koch-Wiser 1976; Villarejo and Pascual 1993). The neurologic damage is related to the volume, location, and drug injected. It can be very difficult to avoid the sciatic nerve when injecting into the thigh of a small turtle, because the nerve lies adjacent to the femur. The volumes of oxytocin used, even at concentrations of 20 units/mL, proportionally far exceed volumes used in any modern human IM injection procedures, so there is considerable theoretical risk of neurologic injury. However, the volumes injected could be reduced significantly by using the 222 units/mL concentration (Sigma) of aqueous oxytocin (Wilgenbusch and Gantz 2000).

There is a long history of using IP injections as the preferred technique in small animal research with mice and rats. Studies of IP injection in mice show a 10%–20% incidence of injection into the bowel (Arioli et al. 1970), but IP injection in mice is regarded as a safe technique (Steward et al. 1968), because membranes quickly seal over after a puncture wound from a 27-gauge needle. Injection into the bowel does not reduce the effectiveness of oxytocin, because it is rapidly absorbed through a variety of membranes into the blood (Graves 1996).

Ewert (1985) suggested subcutaneous injection as the preferred method. Unfortunately there are no studies available that compared the risks of subcutaneous, IM, and IP injections in turtles (Tucker et al. 2007), so, for now, the choice of delivery method will need to be based on personal preference.

Four different approaches were used to protect the oviposited eggs after induction.

1. Sling. Just before the injection, a sling was made of 2-inch-wide duct tape and attached to the turtle. After the injection, the turtle was suspended at a 45-degree angle (head up, tail down) over a moistened towel, as shown in Fig. 3. The turtle was kept high enough above the towel to prevent the claws of the fully extended back leg from puncturing the eggs. Some T. scripta elegans became agitated in the sling and did not oviposit until placed in a tub of water.

2. Tub. In cases where the animal could be continuously observed, the turtle was allowed to expel its eggs in water approximately as deep as 3 times the carapace height (to prevent damage to the egg).

3. Grid. When there was no concern about damage to the eggs from prolonged immersion, a wide-spaced grid was used, as suggested by Tucker et al. (2007). The grid was set 15 cm under the water but allowed the eggs to fall through, beyond the reach of the turtle's claws.

4. Ramp. Turtles that were injected with ketamine were laid prone on a board slanted at about 20 degrees with their heads up. From this position, the oviposited eggs rolled down the board onto a wet foam pad.

View Figure

• Download as Powerpoint
• Download Figure

Records were kept of each animal's species, location and time found, straight line carapace length (CL), drugs used, dosages, time to egg deposition (most cases), number of eggs laid, number of eggs retained, number of injections required, and, occasionally, weight. CL was measured with a dial caliper. During the first decade, length was used as a basis for determining dosage rather than weight. This is because weight was influenced by a highly variable egg mass and bladder fluid load, but length was stable. For this reason weight data are unavailable in some species.

The relative viability of eggs obtained from oxytocin induction vs. natural nests was also evaluated. For this study, 60 natural nest eggs were obtained at dawn, in the same area, from 8 Chrysemys picta picta nests less than 12 hours old (as judged by the residual moisture in the nest plug from the night before). The superior pole of each egg was marked before removal from the nest to avoid rotation, although that was probably unnecessary (Feldman 1983). In the evening of the same day, in the same location, 14 C. picta were collected on land while in the act of finding a nesting location. They were injected with IP oxytocin (without calcium or other agents) within 6 hours. In 4 cases, more than 1 injection was required 12 hours later. Sixty-two eggs were obtained by injection. The eggs collected were not defined by clutch but were simply divided into 2 groups by source (either oxytocin-induced or natural nest). All the eggs were incubated in a 50:50 mix by weight of vermiculite and water. Incubating temperature ranged from 20°C to 38°C as the room temperature varied.

Results

Average hatching time for C. picta picta was 58 days for both oxytocin-induced and natural nest eggs; 58 (97%) of the 60 nest eggs and 57 (92%) of the 62 oxytocin-induced eggs hatched. There was no statistically significant difference (Z = –1.137, p = 0.256) detected in the hatching rates of the oxytocin-induced vs. natural nest eggs.

Terrapene carolina carolina

Forty-two different animals were induced; 9 were captive specimens. CLs varied from 9.5–15 cm and weights from 450–700 g. Thirty of the 42 animals (71%) laid all their eggs after the first injection of oxytocin. The most consistently effective first dose was 8–10 units, regardless of size. Time to finish egg laying after injection varied from 30 minutes to 7 hours. Six turtles required a second dose of 4–10 units the same day or the next day before they would lay all their eggs. Three turtles (7%) required a total of 3–5 doses before they laid all their eggs, and 3 (7%) never laid all their eggs, despite a total of 2–4 doses over several days.

Chrysemys picta picta

Thirty-four different animals were induced; all were wild. CLs varied from 11.7 to 15.9 cm. Twenty two of the 34 animals (65%) laid all their eggs after the first injection of oxytocin. The most consistently effective dose was 7–8 units, regardless of size. Three turtles required a second dose of 3–10 units the same day or the next day before they would lay all their eggs. Five turtles (15%) required a total of 3–5 doses before they laid all their eggs; 4 turtles (12%) never laid all their eggs, despite a total of 2–4 doses over several days.

Clemmys guttata

Thirty-seven different animals were injected; all were wild. Two were injected in successive years for a total of 39 inductions. CLs varied from 9.0 to 11.6 cm and weights from 150 to 290 g. During 32 of the 39 inductions (82%), the animals laid all their eggs after the first injection of oxytocin. The most consistently effective dose was 4–6 units, regardless of size. Time to finish egg laying after injection varied from 2.5 to 7 hours. Five turtles (13%) required a second injection of 2.5–8 units the same day or the next day before they would lay all their eggs. Two turtles (5%) required a total of 3 doses each over 2 days before they laid all their eggs. All turtles laid all their eggs.

Trachemys scripta elegans

Twenty-one different captive animals were induced. There was a total of 44 inductions. Fourteen turtles had 2 or more clutches induced with oxytocin; either successive clutches in 1 year and/or over several years. CLs varied from 16.2 to 21.9 cm and weights from 745 to 2000 g. During 29 of the 44 inductions (66%), the animals laid all their eggs after the first injection of oxytocin. The most effective dose for smaller animals was 10–12 units. For larger animals, 14–16 units worked best. Time to finish egg laying after injection varied from 1.5 to 6.5 hours. Five turtles (11%) required a second injection of 10–15 units the same day or the next day before they would lay all their eggs. Seven (16%) required a total of 3 doses over 2–3 days before they laid all their eggs. Three turtles (7%) never laid all their eggs, despite a total of 3–4 doses. Two captive specimens that had been kept in aquariums laid double clutches made up of 5–8 hypercalcified eggs and 5–8 normally shelled eggs.

Sternotherus odoratus

Fourteen different animals were induced. CLs varied from 7.0 to 12.0 cm. Nine of the 14 animals (64%) laid their eggs after the first injection of oxytocin. The most consistently effective dose was 5–8 units, regardless of size. Time to finish laying after injection was 3–4 hours. Two turtles (14%) required a second dose of 3 units (same as initial dose) the same day or the following day before they would lay all their eggs. Three turtles (21%) never laid all their eggs, despite a total of 3–4 doses.

Other species

Smaller numbers (6 or less) of Glyptemys insculpta, Kinosternon subrubrum subrubrum, Actinemys marmorata, Terrapene carolina major, Terrapene carolina triunguis, Pseudemys rubriventris, and C. serpentina were also induced with oxytocin. Details of the results are in Table 1. No dosage suggestions were possible for G. insculpta and C. serpentina because of the poor outcomes.

Table 1. Outcome of induction using oxytocin alone.

View Fullscreen

Other Drug Combinations

The following drugs and drug combinations were also evaluated with smaller numbers of turtles: oxytocin (suggested dose) + ketamine (< 25 mg/kg), oxytocin (suggested dose) + ketamine (35 mg/kg), oxytocin + propranolol (14–38 mcg/kg), AVT (50 ng/g) + propranolol (11–14 mcg/kg), AVT (5 ng/g), AVT (25 ng/g), and AVT (50 ng/g). Some details and outcomes are shown in Table 2. The “suggested dose” of oxytocin used in Table 2 was the suggested dose from Table 1 for that species. Ketamine at a dose of 35 mg/kg produced sedation for 3–5 hours, so animals must be kept out of the water to avoid drowning until the drug's effects have completely resolved.

Table 2. Outcome of induction by using arginine vasotocin or combination therapy.

View Fullscreen

Safety and Adverse Effects

Most turtles that were injected were observed and fed for several days or more before release. There was 1 death noted among the AVT inductions. That specimen was a young G. insculpta with an old amputation of a rear leg. At necropsy, an egg was found wedged in a malformed pelvic os. No other deaths were observed after IP injection. No morbidity was observed among the 2 C. guttata, 9 T. c. carolina, and 21 T. s. elegans (14 injected more than once) that were observed for a year or more after IP injection.

Normally, T. s. elegans in my collection lay eggs every 3 weeks for a total of 5 clutches per season. When oxytocin was used for the last clutch of a season the animal induced produced its first clutch of the following year at the appropriate time. However, when oxytocin induction was used on any given animal earlier in the season, it always delayed the development of the following clutch and usually reduced the total number of clutches that season. The confounding fact was that the only captive T. s. elegans that were injected were ones that had retained their eggs at least 2 weeks after the expected nesting date and/or demonstrated repeated abnormalities in nesting behavior. However, similar effects were noted with wild C. picta in Nebraska (Iverson and Smith 1993).

A significant adverse effect was observed among a few individuals among the population of captive T. s. elegans. Three days to 2 weeks after laying all their eggs after oxytocin induction, the turtles displayed nesting behavior by digging completed nests and then remaining there for 2 to 3 hours before abandoning the site. This behavior would only happen once after an induction. The same animals tended to demonstrate it repeatedly. Again, the confounding factor was that the only T. s. elegans that were injected in the collection were specimens that had retained their eggs at least 2 weeks after the expected nesting date and/or demonstrated repeated abnormalities in nesting behavior.

Tucker et al. (1995) observed similar behavior in wild T. scripta, but it only occurred within 12 or 24 hours after induction with oxytocin. McCosker (2002) observed the same response in 2 species (3 animals) of wild Australian freshwater turtles that were injected with oxytocin. Despite laying all their eggs after oxytocin induction, the turtles emerged again 15–21 days later to dig nests, despite having no eggs to lay. McCosker (2002) concluded that this behavior “may be attributed to the failure of a normal nesting-related hormonal sequence to occur during artificially induced oviposition.”

This adverse effect could pose a potential risk to wild animals treated with oxytocin, because they are exposed to multiple dangers while nesting. Turtles that nest near roadways would be especially threatened.

Discussion

Eggs obtained by oxytocin induction are as viable as eggs obtained from natural nests, thus confirming that eggs obtained by induction are useful for a wide range of experimental purposes. However, this experiment was carried out with oxytocin-induced eggs from wild nesting females. In the 20 years since that work was done, I have observed numerous instances where retained eggs obtained by induction from captive animals had hypercalcific shells and never developed.

Dosage recommendations for specific species with oxytocin alone are shown in Table 1. Recommended dosage ranges for oxytocin alone vary from 0.7 to 4.0 units per 100 g body weight, with considerable variation between species. With species where 28 or more animals were treated, the success rates with the first injection (by using suggested doses) varied between 74% and 82%. With a second injection that increased to 83%–94%. Tucker et al. (2007) had a 91.8% success rate with a first injection of 10 units/kg in wild T. s. elegans, but their definition of success was that the turtle “retained 2 or fewer eggs”, whereas my definition of success was no retained eggs after the first induction.

It should be noted that 5 of 6 G. insculpta responded very poorly to oxytocin. Other researchers have noted resistance to oxytocin among C. serpentina and Amyda spinifera (J.T. Tucker, pers. comm., January 2005), in certain Asian genera like Cuora, Heosemys, Leucocephalan, and Manouria (C. Tabaka, pers. comm., January 2005), and in Chinemys, Heosemys, Melanochelys, Chelonia mydas, and Batagur baska (Ewert 1985).

The effectiveness of a single dose of oxytocin for many species may not be acceptable, especially for veterinarians, with the goal of removing all the eggs with the initial dose. The fact that some species do not seem to respond to oxytocin and the unfortunate adverse effect of repeated nesting attempts after successful oxytocin induction highlight the need to find a better method of induction.

Unfortunately, I found AVT alone to be no more effective than oxytocin alone and very difficult to work with. There is a need to find a combination of easy-to-use drugs that yields a higher success rate with the initial injection for a wider variety of species. My results with propranolol combined with AVT or oxytocin were no better than oxytocin alone, but I had no way of confirming the effectiveness of the dosage of propranolol chosen (11–40 mcg/kg). To properly evaluate the potential of beta-adrenergic blockers combined with oxytocin in turtles, an experiment would first need to be done to determine the dosage required to reduce the heart rate (it would be reasonable to assume that a dose high enough to reduce the heart rate would also block the adrenergic receptors in the oviducts). There have been a wide range of beta-adrenergic blockers developed since this work was done, so there are now many options other than propranolol, each with varying specificities and durations of activity.

Although only small numbers (13 animals) were involved, there was a suggestion that the combination of ketamine (at a dosage of 35 mg/kg) and oxytocin may be more effective than oxytocin alone. The sedating effects of ketamine also make it physically easy to position the animal for egg laying. A larger study with this combination could properly evaluate this possibility.

Natural oviposition is complex and, at least, involves the interaction of peripheral beta-adrenergic neurons, AVT, and PGF. Beta-adrenergic activity blocks PGF production and function in some reptiles and AVT activity in others. Some other potential approaches that might take advantage of this relationship would be to use a beta-adrenergic blocker with oxytocin or PGF, or ketamine with oxytocin or PGF. During natural oviposition AVT increases oviductal tone directly and indirectly strengthens peristaltic contractions via PGF secretion. This suggests that another treatment option might be to combine oxytocin with PGF. Any of these 5 treatment combinations might prove to be more physiologic than oxytocin alone. A more physiologic induction method has the potential to eliminate the adverse effects that result from using oxytocin alone and might provide an induction technique that is effective for all turtles.