Field Surveys, Natural History Observations, and Comments on the Exploitation and Conservation of Indotestudo forstenii, Leucocephalon yuwonoi, and Cuora amboinensis in Sulawesi, Indonesia
Abstract
We investigated the distribution and conservation status of Indotestudo forstenii, Leucocephalon yuwonoi, and Cuora amboinensis in northern Sulawesi, Indonesia during 2005–06. Village interviews suggest that I. forstenii and L. yuwonoi are more widely distributed than available records indicate. We verified the occurrence of I. forstenii at five heretofore unreported localities in Central and North Sulawesi, including xeric hills above the Palu Valley and sites adjacent to Lore Lindu National Park, and local villagers reported encounters in Panua Nature Reserve. Sexual size dimorphism was not evident in our sample (n = 103) of I. forstenii. We were unable to locate wild populations of L. yuwonoi, but interview data suggest that this endemic species might occur in Bogani Nani Wartabone National Park and Panua Nature Reserve. Among the L. yuwonoi we measured (n = 150), males were larger than females; large body size may give males an advantage in intraspecific dominance contests. Cuora amboinensis remains common in northern Sulawesi where it occurs in a variety of wetland habitats. Subsistence harvesting of turtles appears minimal in northern Sulawesi owing to a cultural bias against consuming turtle meat. We found little evidence of commercial exploitation of turtles in North Sulawesi and Gorontalo; however, a modest but increasing number of turtles are being harvested in Central Sulawesi to supply local ethnic Chinese, and international pet, food, and medicinal markets. Future conservation efforts should seek to verify the occurrence of endemic chelonians in protected areas and develop management plans to insure the survival of these populations.
The fauna of Sulawesi (formerly Celebes) is considered among the most distinctive in Indonesia, with high levels of endemism among terrestrial mammals, reptiles, amphibians, and invertebrates (Whitten et al. 1987; Evans et al. 2003; Gillespie et al. 2005). At least 4 and perhaps 5 species of nonmarine chelonians are known from Sulawesi, including the geographically wide-ranging Malayan box turtle (Cuora amboinensis) and Asiatic softshell turtle (Amyda cartilaginea; Koch et al. in press) and 2 endemic species, the Sulawesi tortoise (Indotestudo forstenii) and Sulawesi forest turtle (Leucocephalon yuwonoi; Iverson 1992; Samedi and Iskandar 2000; Platt et al. 2001; Platt 2006). A single specimen of Pelochelys sp. found in a market in Ujung Pandang (Webb 2002) suggests that a fifth species of chelonian may also be present on the island.
Cuora amboinensis ranges throughout much of Southeast Asia (Iverson 1992); in Sulawesi the species probably occurs island-wide, including many of the smaller satellite islands (Whitten et al. 1987; Platt et al. 2001; Gillespie et al. 2005). According to Rummler and Fritz (1991), Sulawesi populations of C. amboinensis belong to the nominotypic subspecies C. a. amboinensis. Large numbers of C. amboinensis are currently being exported from Sulawesi to Chinese food and medicinal markets at levels that are almost certainly unsustainable (Platt et al. 2001; Schoppe in press). Although apparently common in many areas of Sulawesi, little is known concerning the ecology and natural history of C. amboinensis on the island (Platt et al. 2001). Amyda cartilaginea was recently added to the list of Sulawesi's cheloniafauna on the basis of several specimens found near Palu (Koch et al. in press). The distribution and conservation status of both A. cartilaginea and Pelochelys sp. in Sulawesi remain unknown.
Indotestudo forstenii, regarded as one of the world's rarest and least studied testudinids (Platt et al. 2001), is the only tortoise known to occur east of Wallace's Line (Hoogmoed and Crumly 1984). Earlier reports that populations of I. forstenii on Sulawesi originated from the introduction of Indotestudo travancorica by early seafarers (Hoogmoed and Crumly 1984) are now considered erroneous, and morphological and genetic data indicate the 2 species are quite distinct (Pritchard 2000). It has also been suggested that I. forstenii consists of several cryptic species owing to discrete individual variation in the presence or absence of a nuchal scute (reviewed by Ives 2006). However, Ives et al. (in press) found that the presence or absence of a nuchal scute is not correlated with genetic divergence and concluded that 2 genetically similar but morphologically distinct forms of I. forstenii exist. Whether differences in the appearance of the nuchal scute represent recently differentiated populations or environmentally induced variation has yet to be resolved (Ives et al. in press).
The historic and current distribution of I. forstenii on Sulawesi remains ill-defined; 14 museum specimens exist, but only 5 of these are accompanied by meaningful locality data. Four specimens collected prior to 1900 originated from Sulawesi, 2 lacking specific locality data, and 2 others collected from the northern Minahasa Peninsula (Platt et al. 2001). More recently, Platt et al. (2001) obtained 3 shells from the Cape Santigi region of Central Sulawesi and 6 preserved juveniles of unknown provenance from a reptile dealer in Jakarta. A population also reportedly occurs near the Morowali Reserve in Central Sulawesi (Groombridge 1982), but this remains to be verified.
Leucocephalon yuwonoi was initially described in 1995 as Geoemyda yuwonoi but subsequently reassigned to the monotypic genus Leucocephalon by McCord et al. (2000). The original description was based on 8 specimens purchased at markets in Gorontalo and Poso (McCord et al. 1995). Biologists have rarely observed L. yuwonoi in the wild, and very little is known about its distribution or ecology. Platt et al. (2001) and Hagen and Ching (2005) found L. yuwonoi in small forest streams on Cape Santigi in Gorontalo Province. Riyanto (2006) searched several locations in Central Sulawesi and found L. yuwonoi inhabiting small, heavily vegetated lowland creeks and marshes adjacent to agricultural lands and second-growth forest. McCord et al. (1995), Innis et al. (2002), Innis (2003), and Ives (2006) have reported various aspects of captive husbandry.
Despite the paucity of data on wild populations and the limited geographic distribution and endemic status of I. forstenii and L. yuwonoi, large numbers of both are collected annually for the international pet trade, with lesser numbers channeled into food and medicinal markets (Compton 2000; Samedi and Iskandar 2000). This is particularly alarming with regards to L. yuwonoi because the species is extremely difficult to maintain in captivity and even well-equipped zoos and knowledgeable breeders have experienced only limited success at long-term propagation (Innis et al. 2002; Innis 2003; Ives 2006). Indotestudo forstenii and L. yuwonoi are classified as Endangered and Critically Endangered, respectively, on the International Union for Conservation of Nature and Natural Resources (IUCN) Red List (IUCN 2007) and listed in Appendix II of Convention on International Trade in Endangered Species (CITES 2007). In Indonesia, both species are protected and managed through export quotas (Samedi and Iskandar 2000; Platt et al. 2001; Hagen and Ching 2005). However, indications are that for a variety of reasons, these quotas are regularly exceeded and an extensive illegal trade in endemic Sulawesi turtles exists (Hagen and Ching 2005).
Given the lack of data on the distribution, conservation status, and natural history of endemic turtles in Sulawesi, the recommendations of previous workers for further investigation (Platt et al. 2001; Hagen and Ching 2005; Kuncoro 2005; Riyanto 2006), and the recognition that urgent action is required in the face of threats posed by overharvesting and habitat destruction, we conducted additional surveys for I. forstenii and L. yuwonoi during 2005–2006. Incidental to this investigation we also collected data on occurrence, habitat use, and harvest of C. amboinensis in Sulawesi. Herein we present the results of these surveys and provide conservation recommendations based on our findings.
STUDY AREA AND METHODS
Geographically, our study area was defined by the political boundaries of 3 provinces: Central Sulawesi, Gorontalo, and North Sulawesi (Figs. 1, 2, and 3). The Minahasa Peninsula encompasses all of the provinces of North Sulawesi and Gorontalo and part of Central Sulawesi. Major population centers in our study area include Gorontalo, Manado, Kotamobagu, and Palu. Fieldwork in Central Sulawesi was conducted in the Palu Valley and areas adjacent to Lore Lindu National Park (LLNP) from 19 to 27 February 2005. The Palu Valley is one of the driest regions of Sulawesi owing to its position with respect to the local rain shadow. Annual rainfall in the valley is less than 600 mm and extended dry periods of up to 6 consecutive months are common (Whitten et al. 1987). The characteristic vegetation is open scrub dominated by drought-tolerant woody species, cacti, and grasses, including the invasive alang-alang grass (Imperata cylindrica; Whitten et al. 1987). Lore Lindu National Park is a 231,000-ha World Heritage Site. Annual rainfall in this region of Central Sulawesi averages 2000–3000 mm. The region we surveyed adjacent to LLNP is approximately 600 m asl (above sea level) and characterized by primary and secondary lowland forest (Whitten et al. 1987). Land-use activities in this area include agriculture and small-scale harvesting of timber and nontimber forest products (Ives 2006).
Citation: Chelonian Conservation and Biology 7, 2; 10.2744/CCB-0718.1
Citation: Chelonian Conservation and Biology 7, 2; 10.2744/CCB-0718.1
Citation: Chelonian Conservation and Biology 7, 2; 10.2744/CCB-0718.1
We conducted fieldwork on the Minahasa Peninsula from 15 June to 30 July 2006. This region is characterized by steep mountains rising to over 1000 m in some places, with deep valleys and high plateaus. Rainfall is highly variable owing to the perpendicular orientation of the steep mountains to prevailing winds, which results in localized rain shadows. In general, the northern coast of the peninsula receives the greatest annual rainfall; in most areas monthly precipitation averages about 150 mm (Whitten et al. 1987). The region experiences a dry season from June to late October; although, this varies depending on local topography. The vegetation is classified as lowland and hill forest with scattered areas of other forest types growing on ultrabasic and volcanic soils (Whitten et al. 1987). Many areas are dominated by secondary vegetation or alang-alang grassland as a result of long-term human disturbance. Protected areas on the Minahasa Peninsula include Bogani Nani Wartabone National Park, and Tangkoko, Dua Saudara, Gunung Lokon, Panua, and Tanjung Panjang Nature Reserves (Colijn 2001; Platt 2006 and references therein).
During this investigation, we visited villages and agricultural hamlets and conducted open-ended interviews (sensu Martin 1995) with farmers, fishermen, and other knowledgeable individuals regarding the local occurrence of turtles, levels of exploitation, collecting methods, traditional use, and general knowledge of turtles. Such individuals are generally recognized as an excellent source of information on the local cheloniafauna (Thirakhupt and van Dijk 1994). In accordance with the protocol of an open-ended interview, we asked each informant a series of questions that included standard questions prepared in advance and others that arose during the course of conversation (Martin 1995). Interviews also included a photo array showing 3 species of turtles found in Sulawesi (I. forstenii, L. yuwonoi, and C. amboinensis). Individuals were presented with the array and asked to indicate which species occurred locally. Using calipers (± 1.0 mm), we measured turtles and shells found in villages, and deposited photographic vouchers in the Campbell Museum (CUSC), Clemson University, Clemson, South Carolina, United States.
In addition to gathering interview data, we accompanied villagers to search for turtles, inspect habitat at reported capture sites, and observe capture methods. We searched for I. forstenii during the morning (0730 to 1100 hours) and late afternoon (1500 hours to dusk). If available, we used local hunting dogs because our previous work in Myanmar indicated that dogs are significantly more efficient at locating concealed tortoises than humans (Platt et al. 2003). We searched for L. yuwonoi at night in low- to high-gradient rocky streams using flashlights and headlamps. Particular attention was devoted to searching deep pools and streamside vegetation. This method proved effective during previous studies of L. yuwonoi (Platt et al. 2001; Hagen and Ching 2005).
One of us (IEI) visited the compound of a commercial turtle trader in Palu and obtained morphometric data from I. forstenii and L. yuwonoi; straight-line carapace length (CL) of each turtle was measured with calipers (± 1.0 mm). Sex of the adults was determined on the basis of plastral morphology; males exhibit a plastral concavity that is lacking in females. The inguinal region of each female was palpated for the presence of shelled eggs. The turtle trader was also queried regarding provenance and harvest levels of I. forstenii and L. yuwonoi.
We used a 2-tailed Student's t-test (Zar 1996) to test the null hypothesis that there was no significant difference between the mean CL of male and female turtles (I. forstenii and L. yuwonoi). The degree of size dimorphism (defined as a statistically significant difference in mean length of sexually mature individuals) between the sexes was quantified with a compressed sexual size dimorphism index (SDI; Lovich and Gibbons 1992). SDI is a dimensionless number calculated by dividing the mean size of the larger sex by the mean size of the smaller sex and then adding or subtracting 1 from this value depending on whether males or females, respectively, are the larger sex (Lovich and Gibbons 1992). Although SDI may be based on mass or some measure of body length (CL or plastron length), we selected CL as the appropriate variable because body mass often exhibits considerable variation among animals of similar length owing to the presence of eggs in gravid females, recent ingestion of large meals, and overall body condition (Lovich and Gibbons 1992). Mean values are presented as ± 1SD, and results considered significant at p ≤ 0.05.
RESULTS AND DISCUSSION
We received numerous reports of turtles other than C. amboinensis during interviews with rural inhabitants in northern Sulawesi, but due to the linguistic diversity of the region (e.g., 10 distinct languages are spoken in the provinces of North Sulawesi and Gorontalo in addition to Bahasa Indonesia) and the plethora of unique and sometimes ambiguous names used in local folk taxonomies, it was often difficult to be certain which species our informants were describing. Even using photo arrays, it frequently proved difficult to resolve the confusion over species identifications. However, almost without exception our informants were familiar with C. amboinensis and insisted that at least 1 additional species of turtle occurred in their area. Although local folk taxonomies must be interpreted with caution (Berlin et al. 1966; Platt et al. 2004) and species identifications can only be confirmed by the collection of voucher specimens or photographs, collectively these reports suggest that I. forstenii and L. yuwonoi are more widely distributed in northern Sulawesi than previously recognized.
Interestingly, a number of informants in northern Sulawesi mentioned the local occurrence of a “white turtle”, which is said to inhabit small wetlands, grass swamps, and springheads. Initially we assumed this was a reference to L. yuwonoi, possibly based on the white head coloration of the male, but informants maintained that the white turtle was not among those pictured in the photo array. Although recognizing the pitfalls of uncritically assuming that vernacular names can always be equated with scientifically recognized taxa (e.g., Platt et al. 2004), these interviews raise the possibility that a hitherto undescribed species of turtle exists on the Minahasa Peninsula.
Indotestudo forstenii. — We verified the occurrence of I. forstenii at 5 heretofore unreported localities in Central and North Sulawesi (Fig. 1). Two I. forstenii captured in hills ca. 5 km east of the Palu Valley were being held by a hunter in Bora Village (km 25 on Palu to Gimpu Highway). Another hunter in Watunonju Village had a small I. forstenii captured in hills about 10 km east of the Palu Valley. Both hunters reportedly capture tortoises regularly from this area using hunting dogs. The hills where these tortoises were collected represent a transition zone between the xeric Palu Valley and the cooler, mesic environment surrounding LLNP. The hills are heavily eroded and support drought-tolerant vegetation such as Acacia farnesiana, Eucalyptus deglupta, Kalanchoe pinnata, and Opuntia nigricans.
We found 2 additional I. forstenii held in villages along the western border of Lore Lindu National Park. A tortoise in Kulawi Village was reportedly captured nearby in an agricultural clearing along a stream (01°26′S; 119°59′E). Another tortoise (CUSC 2361) kept as a family pet in Lempelero Village was found on a stream bank within the village (01°39′S; 120°02′E). The vegetation at both sites was a mosaic of secondary forest and tree plantations of sugar palm (Arenga sp.), cacao (Theobroma sp.), and coconut (Cocos nucifera). Significantly, these tortoises were collected within 1 km of the western boundary of LLNP, strongly suggesting that I. forstenii occurs within the park.
We obtained a sixth tortoise (CUSC 1087) from a hunter in Wonggarasi Timur Village. The tortoise was captured on a steep slope (ca. 60°) above the village (00°32.637′N; 121°40.588′E; altitude = 130 m asl). Villagers found the tortoise in secondary forest bordering an agricultural field after being alerted by their dog. The forest had a dense understory beneath a relatively open canopy. This record represents an eastward range extension of approximately 100 km from previously reported populations on Cape Santigi in Central Sulawesi (Platt et al. 2001). Finally, bird trappers reported observing, capturing, and releasing several I. forstenii from coastal scrub habitat in Panua Nature Reserve.
In addition to the 6 I. forstenii examined in the field, we obtained morphometric data from 36 tortoises at the compound of a turtle trader in Palu, 38 captives from The Turtle Bank (Upton, Massachusetts, United States), 12 specimens in the holdings of the Chelonian Research Institute (Oviedo, Florida, United States), and 11 turtles measured during an earlier survey (Platt et al. 2001). The CL of adult males and females averaged 202 ± 4.8 mm (range = 123 to 272 mm; n = 56) and 199 ± 4.1 mm (range = 115 to 254 mm; n = 47), respectively. Sexual size dimorphism was not evident in our sample, and mean CL of male and female I. forstenii was not significantly different (t = 0.5; df = 101; p > 0.05). Shelled eggs were detected in 1 of 9 (11%) female I. forstenii palpated at the trader's facility in Palu during February 2005. The gravid female (CL = 226 mm) was reportedly collected from the Minahasa Peninsula.
According to hunters in Wonggarasi Timur, I. forstenii has been observed consuming monkey feces and grubs from rotting logs. The former undoubtedly contains undigested vegetation and fruit pulp of nutritional value to tortoises. Although rarely observed in the wild, coprophagy has been reported among other turtles (Goodman and Stewart 1998; Platt et al. in press). Two bird trappers in Panua Nature Reserve reportedly observed I. forstenii grazing on grass and sedges on the forest floor. Such observations are noteworthy given that virtually nothing is known regarding the ecology of I. forstenii in the wild (Platt et al. 2001).
Leucocephalon yuwonoi. — We were unable to find L. yuwonoi during nocturnal surveys of seemingly suitable habitat. However, interviews with rural inhabitants suggested that L. yuwonoi is more widespread on the Minahasa Peninsula than the few available field records would indicate (Fig. 2). Villagers also stated that L. yuwonoi occurs in a wide variety of habitats, including coastal palm swamps, grass swamps, and hillside bamboo forests, and is not restricted to riparian forests and small creeks as previously thought (Platt et al. 2001; Hagen and Ching 2005; Hagen et al. in press). These reports are consistent with the observations of Riyanto (2006), who found a number of L. yuwonoi in a densely vegetated marsh amidst cacao and coconut plantations.
Leucocephalon yuwonoi has yet to be found in any protected area of Sulawesi; however, rattan cutters described occasional encounters with L. yuwonoi in Bogani Nani Wartabone National Park, and bird trappers reported snaring L. yuwonoi in Panua Nature Reserve. Owing to the aforementioned confusion surrounding local folk taxonomies, these records cannot be accepted until verified with specimens or photographs. If correct, Bogani Nani Wartabone National Park and Panua Nature Reserve could play a pivotal role in future efforts to conserve this species. Leucocephalon yuwonoi is included on a checklist of reptiles occurring in Lore Lindu National Park (Colijn 2001); although, the basis for its inclusion is unclear (Hagen et al. in press), and the individuals we interviewed in villages adjacent to the park had never encountered the species. In addition, a wildlife trader in Palu reportedly obtains L. yuwonoi from Bangkir (ca. 295 km north of Palu), Ongka (ca. 285 km north of Palu), Tate (ca. 85 km north of Palu), and Tompe (ca. 80 km north of Palu) villages on the Minahasa Peninsula in Central Sulawesi (Fig. 2). Clearly the distribution of L. yuwonoi in Sulawesi warrants further investigation and must be resolved before effective conservation efforts can be initiated.
Forty-six L. yuwonoi were measured at the compound of a wildlife trader in Palu; additional morphometric data were obtained from 62 L. yuwonoi at The Turtle Bank, 38 specimens in the holdings of the Chelonian Research Institute, and 4 specimens measured during an earlier survey (Platt et al. 2001). The CL of males and females averaged 240 ± 26 mm (range = 180 to 278 mm; n = 73) and 198 ± 19 mm (range = 142 to 240 mm; n = 77), respectively (Fig. 4). The mean CL of males was significantly greater than that of females (t = 13.8; df = 148; p < 0.001). We calculated a SDI of þ 2.21 for our sample of L. yuwonoi. The factors that ultimately select for large body size in males must remain speculative until more is known about the life history of L. yuwonoi. Among other chelonians in which males are the larger sex, male–male combat is common (Berry and Shine 1980; Fitch 1981). Large body size is undoubtedly an advantage in these contests, and presumably the mating success of males increases with body size (Berry and Shine 1980; Howard 1981). Notably, intraspecific aggression is common among captive male L. yuwonoi housed in the same enclosure (Charles Innis to SGP, in litt. 2006). Large body size among males may also evolve as an adaptation to enable forcible insemination of females (Berry and Shine 1980).
Citation: Chelonian Conservation and Biology 7, 2; 10.2744/CCB-0718.1
Shelled eggs were detected in 3 (20%) of 15 wild-caught females examined in February 2005 at the trader's facility in Palu. Little is known regarding the reproductive phenology of L. yuwonoi in the wild. In captivity, females have deposited eggs during every month except June and the mean internesting interval is 4.6 months (Innis 2003; Hagen et al. in press).
Cuora amboinensis. — We examined 39 C. amboinensis in northern Sulawesi. Fifteen turtles were captured during nocturnal searches, and villagers provided another 24 specimens (Fig. 3). Many of the latter came from Hindu villagers in the Dumoga River Valley. These transmigrants from Bali maintain small walled shrines where turtles are kept for religious reasons. Our sample of C. amboinensis included 22 females, 10 males, and 7 juveniles, with CL ranging from 55 to 200 mm. There was no difference between the mean CL of males (167 ± 13 mm; range = 150 to 200 mm) and females (168 ± 13 mm; range = 136 to 191 mm). The mean CL of the combined sample of adult male and female C. amboinensis was 167 ± 13 mm (n = 32; range = 136 to 200 mm).
Our field observations and interview data indicate that C. amboinensis are habitat generalists, occurring in a variety of wetland types in northern Sulawesi. We found C. amboinensis in swift-flowing rocky streams where swimming turtles could make little headway against the strong current, small spring-fed rivulets on forested hillsides, Nypa swamps, village ponds, rice-field ditches, and grass-dominated wetlands. Villagers stated that large numbers of C. amboinensis are sometimes encountered when clearing grass swamps to plant rice. Based on our field experience, C. amboinensis is cryptic and even where common can be difficult to locate. Most of the specimens found during our nocturnal searches were concealed beneath cover of some sort (e.g., mud, leaves, submerged palm fronds, etc., or in very dense vegetation).
The feces of 2 C. amboinensis that we collected in a Nypa swamp contained vegetation, pieces of fruit, and seeds. Patterns of carapacial scarring observed on 5 (12.8%) C. amboinensis in our sample were consistent with descriptions of fire damage to Terrapene carolina (Dodd et al. 1997). Fire scars were likewise noted on C. amboinensis during an earlier investigation (Platt et al. 2001) and probably result from burns sustained when fast-moving wildfires sweep over turtles that are buried beneath leaf litter or grass. There is a widespread belief among rural villagers that C. amboinensis produces nocturnal vocalizations audible to the human ear at a distance of several meters. Some villagers even claimed to locate turtles by these vocalizations, but we were unable to verify this in the field. Vocalizations have been noted among other species of turtles (Campbell and Evans 1972; Pritchard 1979), including Cuora galbinifrons (Price 1992). Although their biological significance remains unclear, vocalizations are often associated with copulation and feeding (Campbell and Evans 1972; Jackson and Awbrey 1978; Price 1992).
In spite of its cryptic behavior, the relative ease with which we found C. amboinensis even in densely populated agricultural areas, the large number of specimens we obtained from villagers and local perceptions of abundance collectively indicate that C. amboinensis remains common throughout the area we surveyed in northern Sulawesi. However, we caution against complacency because C. amboinensis is heavily exploited in Central and South Sulawesi at levels that are certainly unsustainable (Platt et al. 2001; Schoppe in press). Once turtle stocks in those regions become depleted, it is likely that wildlife traders will turn their attention to northern Sulawesi.
Exploitation and Conservation
We found little evidence of subsistence harvesting of chelonians in northern Sulawesi. With the exception of ethnic Chinese living in larger urban centers, few people in northern Sulawesi consume turtle meat. This is thought to be due to a strong cultural bias and religious proscriptions against consuming turtle flesh among both Christians and Muslims in the region.
The situation regarding commercial harvesting of turtles in northern Sulawesi is less clear. We found nothing to suggest that widespread commercial exploitation was underway in the provinces of Gorontalo or North Sulawesi. This contrasts markedly with the commercial demand for wild mammals and birds in these provinces (Lee 1999; Lee et al. 2005). Several people at scattered localities in Gorontalo Province indicated that wildlife traders from Central Sulawesi had visited in the past, but these visits ceased and villagers no longer had any economic incentive to collect turtles. On the basis of our interviews, it appeared that only small numbers of turtles (< 100/village) had been harvested during the past 2 years to supply these periodically visiting traders. This limited harvest seemed to result from opportunistic encounters (i.e., people collect turtles when they chance upon them while engaged in other pursuits) rather than the efforts of professional collectors. Finally, the number of C. amboinensis we found in anthropogenic habitats near villages, supports our conclusion that turtles in this region are not subjected to intensive harvesting. Cuora amboinensis would be unlikely to persist in these accessible locations if such harvesting were taking place.
Unlike the provinces of Gorontalo and North Sulawesi, commercial exploitation of turtles is occurring in the parts of Central Sulawesi that we visited. Although the extent of this harvest is difficult to quantify, our interviews suggest a modest but increasing number of turtles are being removed from the wild to supply ethnic Chinese living in urban centers, as well as international pet, food, and medicinal markets. A single hunter in Watunonju Village claimed to collect 1 I. forstenii every week to sell to a Chinese restaurant in Palu. According to the reptile dealer in Palu, approximately 50 L. yuwonoi are obtained each month from hunters at various locations in Central Sulawesi. Hunters receive 35,000 Rupiah (ca. US$ 3.50 in February 2005) for each adult turtle, which are later sold to an exporter in Jakarta for 150,000 Rupiah (ca. US$ 15.00). Significantly, the value of wild-caught L. yuwonoi has changed little over the past 4 years. Hagen and Ching (2005) reported the same dealer paying hunters about the same price for L. yuwonoi and in turn receiving the same price from the Jakarta-based exporter in 2002. An increase in value might be expected if turtle populations were perceived to be declining or if L. yuwonoi was becoming more difficult to collect. It is important to note, however, that this situation may change in the near future as stocks of wild turtles in southern Sulawesi become depleted by the intensive harvesting currently underway there (Schoppe in press) and collectors turn elsewhere to meet market demand.
Owing to the paucity of baseline data on wild populations, conclusions regarding the impact of the current harvest on I. forstenii and L. yuwonoi are somewhat tentative. The fact that the value of L. yuwonoi has remained constant over the past 4 years suggests that harvesting has yet to seriously impact source populations. In contrast, reports of declining catch-per-effort by hunters (Platt et al. 2001) indicate that chelonian populations are being overharvested, at least in some areas. The knowledge that few chelonian populations can sustain even minimal levels of exploitation (Congdon et al. 1993; Thorbjarnarson et al. 2000) calls into question the biological basis of the export quotas established by Indonesian regulatory authorities (Samedi and Iskandar 2000). Furthermore, it appears likely that these export quotas are regularly exceeded and large numbers of turtles are being illegally exported from Sulawesi (Hagen and Ching 2005). In light of the biological uncertainty surrounding export quotas and an apparently substantial illegal harvest, we urge Indonesian regulatory authorities to implement an immediate moratorium on turtle exports until the status of wild populations can be assessed, and it is determined what level of harvest, if any, turtle populations can sustain.
In addition to overharvesting, habitat loss probably threatens populations of I. forstenii and L. yuwonoi in many areas. Deforestation rates in Sulawesi are among the highest in the world as a result of commercial logging, small-scale shifting cultivation, large plantation projects, and transmigration schemes designed to resettle people from Java to other less populated islands of the Indonesian archipelago (Whitten et al. 1987; Myers 1992). It is encouraging to note that neither I. forstenii nor L. yuwonoi appear dependent on undisturbed primary forest (Platt et al. 2001; Hagen and Ching 2005; Riyanto 2006). Thus, if effectively protected from overharvesting, populations of endemic chelonians could likely survive in anthropogenically modified habitats.
To date, the occurrence of either endemic chelonian has not been verified in any protected area of Sulawesi, but our interview data strongly suggest that I. forstenii occurs in LLNP, L. yuwonoi is likely found in Bogani Nani Wartabone National Park, and both species may be present in Panua Nature Reserve. Earlier reports of I. forstenii in the Morowali Reserve of Central Sulawesi (Groombridge 1982) await investigation. Given the threats posed by overharvesting and habitat destruction, it is imperative to determine if the protected areas of Sulawesi harbor populations of endemic chelonians and future fieldwork should attempt to answer this question. Finally, should endemic chelonians be found in protected areas, it is essential to implement effective protection and develop species-specific conservation plans to ensure the long-term viability of these populations.
Localities where Indotestudo forstenii is known or thought to occur in northern Sulawesi, Indonesia. ▪ = localities where we verified the occurrence of I. forstenii; ▴ = localities where I. forstenii may occur based on our interviews of villagers and wildlife traders; • = literature records from Platt et al. (2001). Stars denote major population centers and names of provinces are italicized. Lower map shows relation of Sulawesi to other islands of the Indonesian Archipelago.
Localities mentioned in the text where Leucocephalon yuwonoi is known or thought to occur in northern Sulawesi, Indonesia. ▴ = localities where L. yuwonoi may occur based on our interviews of villagers and wildlife traders; ▪ = localities where L. yuwonoi was found by Platt et al. (2001), Hagen and Ching (2005), and Riyanto (2006). Stars denote major population centers and names of provinces are italicized.
Localities where Cuora amboinensis was recorded during our survey in northern Sulawesi, Indonesia. ▴ = localities where C. amboinensis was captured during nocturnal surveys; ▪ = localities where C. amboinensis were obtained from villagers. Stars denote major population centers and names of provinces are italicized.
Size-class distribution of female (n = 77) and male (n = 73) Leucocephalon yuwonoi.
