Proceedings of the First International Symposium on Turtles & Tortoises: Conservation and Captive Husbandry, 127-137, 1991.

Captive Husbandry and Conservation of the Madagascar Ploughshare Tortoise, Geochelone yniphora

by James O. Juvik, Duane E. Meier and Sean McKeown

A captive group of Madagascar ploughshare tortoises or angonoka (Geochelone yniphora) has been maintained at the Honolulu Zoo since April 1971 (Juvik & Blan, 1974). There was little reproductive activity in this group until the late 1970's when an additional adult female G. yniphora was secured on breeding loan from the San Antonio Zoo in Texas. Between September 1879 and September 1982 this tortoise laid a total of 25 eggs in 6 clutches. However, none of these eggs proved to be fertile. Also, over this period, the zoo herpetology staff developed techniques to encourage successful natural mating by periodically isolating the males and females and then reintroducing them to one another for short periods. More natural mating attempts followed the adoption of this management approach. McKeown (1982) and McKeown et al. (1982) have described general captive husbandry, courtship behavior and egg laying for this species.

Captive Husbandry

Reproduction

Figure 1. Hatchling Geochelone yniphora; age 1 month (October 1983); carapace length 50 mm; weight 32 g.

During the summer of 1982, several apparently successful courtship and mating sequences were observed involving the male and egg laying female G. yniphora. Between September 1982 and February 1983, the female laid a total of five egg clutches at approximately one month intervals. The egg laying dates and clutch sizes were as follows: 22 September 1982, four eggs; 25 October 1982, seven eggs; 1 December 1982, five eggs; 6 January 1983, six eggs; and 19 February 1983, three eggs. All eggs were incubated in covered plastic containers with moist vermiculite at 28.9° C (+/- 1.0° C). On 12 March 1983, rotting eggs of the 25 October clutch were opened and one revealed a dead embryo (approximately 19 mm in length). This was the first evidence of egg fertility among the many egg clutches laid by the female since 1979.

Other eggs of the 1982-83 clutches were examined and candled periodically during the summer of 1983 (two additional dead embryos were discovered). One egg of the 2 February 1983 clutch, that appeared to be developing during periodic candling, hatched successfully on 16 September 1983 after an incubation period of 226 days. At birth, the baby G. yniphora was 42 mm long and weighed 24.6 g.

Artificial Insemination

Figure 2. Electroejaculation of G. yniphora at the Honolulu Zoo (13 October 1982). Left: tortoise secured upright in working position. Right: penis relaxed and protruded, 2.5 cm diameter rectal probe (with electrodes) inserted dorsal to exposed penis into the cloaca a depth of 10-12 cm.

Because of egg infertility in G. yniphora clutches laid in Honolulu between 1979 and 1982, Zoo staff became interested in the possible application of semen collection/evaluation and artificial insemination (AI) procedures then being developed for turtles and tortoises (Platz et al., 1980). With technical expertise and equipment provided by Dr. C. Platz, Jr. (Texas A&M University) and Dr. Greg Mengden (Australian Museum-Sydney), an AI program was initiated at the Honolulu Zoo. In October 1982, Mengden visited Honolulu, and began initial experimental work with the Zoo's G. radiata (radiated tortoise) group. The electroejaculator consisted of a variable voltage power source and a plastic rectal probe 2.6 cm in diameter with three 3.0 mm electrodes arrayed longitudinally. After first fully testing semen collection procedures on a male G. radiata, a small quantity of motile sperm was successfully collected from the male G. yniphora on 13 October, 1982 and immediately injected into the egg-laying female. This was the only AI procedure with the female prior to September 1983. The female G. yniphora had laid a clutch of 4 eggs on 23 September 1982 (all proved infertile). Twelve days after the first AI procedure, on 25 October 1982, the female laid another clutch of 7 eggs. It was from this clutch that the first evidence of egg fertility was recorded when a single dead embryo was discovered in March 1983. We conclude that the interval between AI and egg laying (12 days) for this 25 October 1982 clutch provided insufficient time to credit the AI procedure for the recorded egg fertility; fertilization was probably produced by natural inseminations from active mating episodes that occurred during the summer and fall of 1982.

Evident fertility associated with additional dead embryos recorded in egg clutches laid on 1 December 1982 and 19 February 1983, and the live hatchling (also from the 19 February 1983 clutch), is thus equivocal with respect to the possible role of the October 1982 AI procedure or concurrent natural fertility. In July 1983, one of us (McKeown) left the Honolulu Zoo, and the zoo's herpetologist position remained unfilled for almost a year. During this period AI work with G. yniphora was under the direction of the zoo veterinarian. After a further series of electroejaculation procedures were undertaken in September 1983, the male G. yniphora stopped eating and died 30 days later (necropsy was inconclusive but suggested kidney failure).

Captive Growth

Growth data for the 1983 G. yniphora hatchling are presented in table 1.

Figure 3 illustrates a normal growth curve (made linear by log transformation of the weight values), with weight and age highly correlated (R=0.997; p<0.001) at age 2 years and above.

Figure 3. Relationship between age and weight (log scale) of the captive born G. yniphora in the Honolulu Zoo. Regression line fitted to data points beginning at two years of age.	Figure 4. Growth rates for the Honolulu captive bred G. yniphora in comparison with wild tortoises (age estimated from annual growth rings). Wild tortoise data from Juvik et al., 1981 and Curl et al., 1984.

If this growth curve was maintained over the next few years, the regression equation (log Y = 2.045 + 0.213X) predicts that typical adult weight (10-15 kg) would be achieved at between 9 and 10 years of age. Figure 4 illustrates the relationship between age and carapace length for the Honolulu captive-raised G. yniphora in comparison with growth of wild tortoises in the Baly Bay region (age estimated from annual growth rings). The linear regression equations yield annual growth increments of 25.77 mm/yr for the captive tortoise compared with 14.96 mm/yr in the wild. At the captive tortoise's current age (6 years, 10 months) its shell length (225 mm) is 57% larger than that predicted (143 mm, from regression) for a wild tortoise of comparable age.

Figure 5. Honolulu born G. yniphora at age 2.5 years (carapace length 106 mm). Abnormal "doming" of shell evident.

Gaymer (1968) in an analysis of comparative growth in Geochelone gigantea (Aldabra tortoise) found an even larger difference in growth rates between captive-raised and wild populations. At age 7 years, the average carapace length of captive-raised G. gigantea was 94% larger than in wild populations. This extreme difference may be explained by generally suppressed growth in high density wild populations experiencing intense intraspecific competition for food and shade (see Merton et al., 1976). Such competition-related growth suppression should not be a significant factor in wild G. yniphora which now exist at very low densities.

Zovickian (1973) reported that sexual maturity in the closely related G. radiata is achieved at carapace straight length of 330 mm (male) to 340 mm (female). If these values are roughly transferable to G. yniphora (with a slight upward adjustment to compensate for the overall larger size of this species, i.e., 350 mm), then, based on the above regression equations, the Honolulu captive should reach maturity at year 12, compared to year 20 for wild tortoises.

In retrospect, the growth rate achieved in the Honolulu Zoo, captive-raised G. yniphora was too fast, and led to distortions in allometric growth. At one year of age the captive G. yniphora exhibited abnormal shell "doming". This species naturally possesses a high domed shell, particularly in adult females, where shell height averages 62% of carapace straight length (Juvik et al., 1981). By age 3 years, the captive G. yniphora shell height reached 76% of carapace length, although by age 6 this value had retreated to 73%. This shell deformity is not only of aesthetic concern, particularly if carried into adulthood. Excessively domed shell geometry in G. yniphora could be expected to lead to reduced reproductive success.

Our captive growth results in G. yniphora indicate that allometric relationships must be closely monitored during the first few years. Diet and environmental factors should be manipulated to ensure that captive growth rates do not get too far ahead of "natural" (wild) rates so as to produce shell distortions.

Figure 6. Ratio of shell height to carapace length in the Honolulu captive born G. yniphora and wild tortoises (data from Juvik et al., 18981 and Curl et al., 1984)

Shortly after the departure of the supervising herpetologist (McKeown) to the Fresno Zoo in 1983, the diet for juvenile tortoises which had consisted of chopped butter lettuce, romaine, kale, grated carrots, tomato and papaya was modified by the Honolulu Zoo veterinarian. Additional fruit in the form of banana and apple were added along with moistened (dry) dog food and later, ZU/PREEM Marmoset diet. We believe that the abnormal growth of the captive raised G. yniphora may be related to both the additional fruit, as well as the dog chow and marmoset diet provided during the first two years of its life.

Captive Breeding of Geochelone yniphora in Madagascar

Throughout the 1970's and early 1980's, the Malagasy Government (Direction des Eaux et Forets: DEF) maintained a captive breeding facility for confiscated native land tortoises at the Ivoloina Forestry station in Toamasina (Tamatave) on the island's humid (windward) east coast.

On a visit to the Ivoloina facility in January 1984, the senior author found the tortoise collection to include 43 adult G. radiata (21 male, 22 female) and 7 adult G. yniphora (5 male, 2 female), all confined together in a small thatched shed (area 22.5 m, or 2.2 tortoises/sq. m). In spite of space limitations the tortoises appeared well cared for and in good health. The G. radiata had been breeding successfully for several years. About 50-60 eggs were laid annually and incubated in closed wooden boxes of sand under ambient temperature conditions. Each year, 20-30 hatchlings emerged after an incubation period of 8-12 months (A. Befeno, personal communication, 1984). In January 1984, there were 144 captive bred young G. radiata at the Ivoloina facility, ranging in age from recent hatchlings to 5 year olds. The largest captive bred animals (1979 hatch) had carapace straight lengths of 82-125 mm (X=105 mm; N=6), and noticeable pyramiding of the carapace scutes.

In contrast to the prolific reproduction of G. radiata, no egg laying was reported for the G. yniphora as Ivoloina. Although confined in cramped quarters with adult G. radiata, it was observed that G. yniphora males made clear interspecific discrimination as both episodes of male combat and courtship in G. yniphora were noted during the brief visit.

Because the Ivoloina site did not possess climatic conditions comparable (i.e., distinct wet-dry season contrasts) to that found in the natural range of G. yniphora, it was earlier recommended in the Species Recovery Plan (Juvik et al., 1982) that the Ivoloina animals be relocated to a west coast forestry facility in the Mahajanga area (some 100 km east of Baly Bay) where general environmental conditions were more appropriate.

In 1986, with support from the World Wildlife Fund, the Jersey Wildlife Preservation Trust (JWPT) in cooperation with the Malagasy DEF, established a G. yniphora captive breeding facility at the Ampijoroa Forestry station near Mahajanga. In September 1986, the Ivoloina G. yniphora (5 male, 2 female) were transferred to Ampijoroa. More recently, additional captive held G. yniphora from villages in the Soalala area have been donated to the Ampijoroa breeding program, bringing the adult tortoise population at the facility to 20 animals (8 males, 12 females) in October, 1991. The first captive breeding success at Ampijoroa was achieved in 1987 with the emergence of a single hatchling G. yniphora. By the end of 1989 a total of 21 captive hatchlings had been produced, including 12 in 1989 alone, from 7 clutches totaling 25 eggs (Reid et al., 1989). In 1990 there were a further 10 hatchlings, and for 1991, there are 55 eggs under incubation (Reid, pers. comm. November, 1991). Eggs have been left to incubate in situ in outdoor tortoise pens, with wire screen covers placed over nest sites to provide protection from predators. Natural incubation periods have ranged from 168-266 days and are apparently correlated with wet-dry season onset and duration (Reid, 1990a).

Wild Populations of G. yniphora: Status and Prospects

Figure 7. Location map, northwestern Madagascar

Based on field surveys during the early and mid 1970's, Juvik et al. (1981) documented the presence of wild G. yniphora populations at four different dry forest locations in the Baly Bay area:

1. North Ankoro (tortoise seen in wild, see Andrianarivo, 1977)
2. Cape Sada (several tortoises seen in wild)
3. East of Andanivato (fresh tortoise dung)
4. Forest patches between Anky and Antanandava (fresh tortoise dung)

More recent surveys in 1983 and 1986 (Curl et al., 1984, 1985 and Curl, 1986a) collected considerable anecdotal information (and documented a substantial number of captive tortoises in villages) but failed to uncover or confirm any new wild populations beyond those listed above.

In January 1990, Don Reid, the Madagascar based conservation officer with JWPT, spent two weeks in the Baly Bay region surveying G. yniphora populations and potential protected area sites. Reid was unable to visit the forest locality north of Ankoro (Anjaha) on the west side of Baly Bay but visits to this area in April, 1991 by local Malagasy nature protection officials confirmed the continued presence of a significant G. yniphora population in the Belambo forest to the north and east of Ankoro. This habitat area may well prove to be as important as Cape Sada for G. yniphora conservation and is currently the focus of ongoing field work.

Reid (1990b) also interviewed villages at Antanandava concerning the status of G. yniphora populations inhabiting remaining forest patches to the north. These dry forest areas were reported to have been heavily degraded by fire and cattle grazing over the past few years, with villagers convinced that no tortoises remained in the area. Reid (1990b) did find one adult G. yniphora in the wild, in a small (30 hectare) dry forest "island" (surrounded by palm/grass savanna) in the Betheta/Anamboho area. Reid (1990b) also visited Cape Sada and confirmed the earlier recommendations of Juvik et al. (1982) and Curl et al. (1985) that Cape Sada is the most appropriate site for initial protected area status based on the following factors:

a) 200-250 hectares of suitable tortoise habitat (little affected by burning) which could potentially support up to several thousand tortoises;
b) the absence of a permanent human population or intensive human utilization of the area;
c) ease and economy of constructing a pig/cattle barrier fence across the Cape Sada peninsula (about 2 km); and
d) the convenience of access (by boat) from the administrative center of Soalala.

Juvik et al. (1982) gave specific design specifications on a mesh and barbed wire pig fence for Cape Sada, based on successful fencing techniques used to control feral pigs in Hawaii Volcanoes National Park. According to Reid (1990b) local Malagasy authorities expressed some concern that villagers near Cape Sada might well find alternative uses for valuable fencing materials. This potential problem can possibly be overcome by employing local villages in fence constructing and maintenance and providing each nearby village with some fencing material for pig traps and garden protection.

In a recent uncritical review of the literature on G. yniphora conservation, Burke (1990) draws several unsupported and errant conclusions which only detract from ongoing conservation efforts for the species. Perhaps confused by contradictory assertions on the persistence of vegetation boundaries around Baly Bay (Curl et al., 1984, 1985 and Curl 1986b), Burke (1990) concludes there is little evidence for ongoing tortoise habitat degradation, and that "the situation of the wild angonoka (G. yniphora) may be somewhat less precarious than previously thought." Such contentions are merely "wishful thinking" in view of recent field work results (Reid, 1990b), and our own (Kiester and Juvik, in preparation) assessment of regional habitat change based on comparative analysis of early aerial photography (1949) and recent hand-held Space Shuttle photography and Landsat thematic mapping images (both from December, 1990). Although preliminary remote sensing analysis indicates that some forest fragments have disappeared over the 41 year interval, other forest patches have retained their boundaries virtually unchanged (e.g., the Antanandava Forest). Ground surveys, however, indicate that even in these persistent forest fragments human induced habitat degradation is diminishing the quality if not the quantity of the areas for native wildlife. These conclusions are consistent with the recent work of Andrianarivo (1990) who studied dry forest cover changes from satellite data for a nearby area (Mahajamba Bay) in western Madagascar. Although Andrianarivo found little net change in total forest cover between 1973-84, ground surveys revealed qualitative forest ecosystem decline in the form of reduced tree basal area and lower densities for arboreal lemurs.

Burke (1990) also asserts that Cape Sada is unsuitable for a nature reserve because its small size (200-250 hectares of tortoise habitat) would probably not support a viable tortoise population. If pigs and cattle were effectively excluded (by fencing) from Cape Sada, G. yniphora would remain the only significant terrestrial herbivore in an ecosystem which could easily support several tortoises per hectare, or a total population exceeding 1,000 animals. It should be noted that much larger tortoises (G. gigantea) exist at average densities of 17/hectare on the south coast of Aldabra Atoll, some 600 km to the north of Baly Bay (Hnatiuk et al., 1976). We agree that Cape Sada comprises less than ideal habitat size, however there is simply no realistic alternative at the present time given current land use trends and cultural practices in the region.

Acknowledgements

The authors wish to thank Dr. Lee Durrell and Mr. John Hartley of the Jersey Wildlife Preservation Trust for access to unpublished reports and photographs.

Literature Cited

Andrianarivo, A. J. 1977. L'Angonoka, survivra-t-elle? These, Department Eaux et Forets, Universite de Madagascar, Antananarivo.
Andrianarivo, A. J. 1990. Analysis of forest cover changes and estimation of lemur populations in Northwestern Madagascar using satellite digital data. Ph.D. thesis (unpublished), School of Forestry and Environmental Studies, Duke University, North Carolina, 144 pp.
Burke, R. L. 1990. Conservation of the world's rarest tortoise. Conserv. Biol. 4:122-124.
Curl, D. A. 1986a. The rarest tortoise on earth. Oryx 20(1):35-39.
Curl, D. A. 1986b. A recovery programme for the Madagascar plowshare tortoise, Astrochelys (=Geochelone) yniphora. Dodo, J. Jersey Wildlife Preserv. Trust 23:68-79.
Curl, D. A., Scoones, I. C., Guy, M. K. and Rakotoarisoa, G. 1984. The angulated tortoise of Madagascar: Current status and distribution. Zool. Dept., Oxford Univ. Unpubl. ms.
Curl, D. A., Scoones, I. C., Guy, M. K. and Rakotoarisoa, G. 1985. The Madagascar tortoise Geochelone yniphora: Current status and distribution. Biol. Conserv. 34:35-54.
Gaymer, R. 1968. The Indian Ocean giant tortoise Testudo gigantea on Aldabra. J. Zool. 154:341-363.
Hnatiuk, R. J., Woodell, S. R. J. and Bourn, D. M. 1976. Giant tortoise and vegetation interactions on Aldabra Atoll, Part 2: Coastal. Biol. Conserv. 9:305-316.
Juvik, J. O., Andrianarivo, A. J. and Blanc, C. P. 1981. The ecology and status of Geochelone yniphora: A critically endangered tortoise in north-western Madagascar. Biol. Conserv. 19:297-316.
Juvik, J. O., Andrianarivo, A. J., Blanc, C. P. and Bour, R. 1982. Revised species recovery plan for the endangered Madagascar angulated tortoise (Geochelone=Astrochelys yniphora). IUCN/SSC Tortoise Specialist Group. Unpubl. ms.
Juvik, J. O. and Blanc, C. P. 1974. The angonoka of Cape Sada. Animals 16:148-153.
McKeown, S. 1982. Priorities and techniques of captive breeding reptiles at the Honolulu Zoo, with special emphasis on Testudinidae, Iguanidae, Scincidae and Gekkonidae. pp. 93-102. In: Fifth Annual Reptile Symposium on Captive Propagation and Husbandry. Thumont, Maryland.
McKeown, S., Juvik, J. O. and Meier, D. E. 1982. Observations on the reproductive biology of the land tortoises Geochelone emys and Geochelone yniphora in the Honolulu Zoo. Zoo Biol. 1:223-235.
Merton, L. F. H., Bourn, D. M. and Hnatiuk, R. J. 1976. Giant tortoise and vegetation interactions on Aldabra Atoll, Part 1: Inland. Biol. Conserv. 9:293-304.
Platz, C. C., Jr., Mengden, G. and Quinn, H. 1980. Semen collection and analysis in endangered chelonians. Unpubl. Ms. Reid, D. 1990a. Project Angonoka/Kapidolo: Report for 1989. Jersey Wildlife Preserv. Trust. Unpubl. ms.
Reid, D. 1990b. Report of a field study of Geochelone yniphora in January 1990. Jersey Wildlife Preservation Trust. Unpubl. Ms.
Reid, D., Durrell, L. and Rakotobearison, G. 1989. The captive breeding project for the angonoka Geochelone yniphora in Madagascar. Dodo, J. Jersey Wildlife Preserv. Trust 26:34-48.
Zovickian, W.H. 1973. Captive reproduction of the radiated tortoise. HISS News Journal 1:115-118.