Introduction
Many freshwater turtles and tortoises in Asia are threatened because of a burgeoning demand for them as food and for use in traditional medicines (van Dijk et al., Reference van Dijk, Stuart and Rhodin2000). Of the 87 non-marine chelonian species recorded in Asia, 47 are categorized as Endangered or Critically Endangered on the IUCN Red List (IUCN, 2008). The demand for turtles in China has severely affected the turtle fauna in adjacent countries (Cheung & Dudgeon, Reference Cheung and Dudgeon2006). In Taiwan there are four native freshwater turtles: the Chinese stripe-necked turtle Mauremys sinensis, Asian yellow pond turtle Mauremys mutica, Chinese three-keeled pond turtle Mauremys reevesii and Chinese softshell turtle Pelodiscus sinensis (Mao, Reference Mao1971; Iverson, Reference Iverson1992). With both a high commercial value and high demand the four species have suffered intensive harvesting pressure (Chen et al., Reference Chen, Lin and Chang2000). Because these species have been commonly consumed and farmed in China (Cheung & Dudgeon, Reference Cheung and Dudgeon2006; Shi et al., Reference Shi, Parham, Fan, Hong and Yin2008) they are susceptible to illegal cross-border trade, especially smuggling. Furthermore, while M. sinensis and P. sinensis have also been farmed (or ranched) and are released as prayer animals in traditional rituals in Taiwan, the wild populations are threatened by extensive collection for replenishment of breeding stock and ritual release (Chen et al., Reference Chen, Lin and Chang2000).
These species were formerly widely distributed in lowland aquatic habitats (Lue et al., Reference Lue, Tu and Shiang2002). However, M. mutica, M. reevesii and M. sinensis are now categorized as Endangered and P. sinensis as Vulnerable on the IUCN Red List (IUCN, 2008; we use the genus names provided by Turtle Taxonomy Working Group, 2009: Mauremys reevesii = Chinemys reevesii and Mauremys sinensis = Ocadia sinensis). Extensive habitat loss and commercial exploitation are the main threats. The Taiwanese populations of M. mutica and M. reevesii are protected under the Wildlife Conservation Act. With little information available on the current status and distribution of the freshwater turtles of Taiwan it is difficult to plan appropriate conservation measures. The objective of the study reported here was, therefore, to obtain this information.
Methods
We conducted an island-wide trapping programme (including Kinmen Island near mainland China) of the freshwater turtles of Taiwan from February 2001 to December 2007. Based on 1 : 25,000 maps published by the Ministry of Interior, Republic of China, we identified 182 potentially suitable habitats for freshwater turtles (Fig. 1), i.e. permanent and temporary ponds, slow-moving rivers and streams, agricultural ditches, wet meadows and wetlands. Turtles were collected with home-made funnel traps (with 2 × 2 cm mesh) baited with canned cat food. We have successfully used this trap in previous studies (Chen & Lue, Reference Chen and Lue1998a,Reference Chen and Lueb, Reference Chen and Lue2009; Chen, Reference Chen and Lue2006). Usually 3–5 traps were set at each site along riverbanks or at the margins of other water bodies and were checked once or twice per week. If no turtles were captured after 2 weeks of trapping at a site we assumed that turtles were not present or that their density was very low. As this study was carried out in conjunction with a trapping programme for the analysis of the population genetics of native turtle species we tried to collect at least 10 individuals of each native species at each site. We moved the trapping to a new site after 10 individuals were captured or after 1–2 months of trapping, except at six sites where we trapped for > 1 year in conjunction with population studies. Because trapping efficiency is related to season and weather condition we did not estimate trapping efficiency. Trapped turtles were marked with individual marginal notches, with a hand-saw or triangular file, for further identification (Cagle, Reference Cagle1939), sexed using external secondary sexual characteristics, measured (straight carapace length) to the nearest 0.1 mm with vernier callipers, and then released at point of capture. In the population analyses we pooled samples of sites < 1 km apart.
Results
A total of 2,109 individual freshwater turtles were trapped from 111 sites, including 281 individuals of four non-native species (Trachemys scripta elegans, Pseudemys nelsoni and Graptemys pseudogeographica from North America, and Cuora trifasciata from South-east Asia), and 1,824 individuals of four native species and four hybrids at 103 of the 111 sites (Table 1, Appendix). M. sinensis was the most abundant and widely distributed species, comprising 78.6% of the native turtles captured (Fig. 2a). M. mutica comprised 17.8% of the turtles caught (Fig. 2b). M. reevesii was trapped at only seven sites, on Kinmen Island (Fig. 2c). P. sinensis, the most important commercial species, comprised only 2.3% of the total trapped (Fig. 2d). The numbers of turtles captured at most sites were low (Appendix); > 20 individuals of M. sinensis and M. mutica were trapped at only nine and six sites, respectively. The sex ratio of all native species was significantly male-biased (Table 1).
* All sex ratios were significantly male-biased: M. sinensis, G = 414.81, P < 0.001; M. mutica, G = 14.24, P < 0.001; M. reevesii, G = 7.72, P < 0.01; P. sinensis, G = 8.15, P < 0.01
M. sinensis was mostly trapped in northern, central and southern Taiwan, with only two individuals trapped in the east, in slow-moving rivers and ditches at low elevations, and in wetlands, agricultural reservoirs and ponds on plains, plateaus and in montane areas. It was also trapped in brackish water near estuaries such as the lower section of the Houlong River.
M. mutica was most commonly trapped in the northern region, in small water bodies in montane or plateau areas, including agricultural and temporary forested ponds, irrigation ditches and wet meadows, and none were trapped in the southern region. A few individuals were captured in riverine habitats or larger-sized water bodies in lowland areas. The localities in which the species was trapped were mostly associated with forest.
M. reevesii was not trapped on the main island of Taiwan. However, four M. reevesii-related hybrids were captured at three sites in northern Taiwan, indicating that this species has either declined to an extremely low density there or disappeared.
P. sinensis was trapped at 19 sites in rivers, streams, wetlands and ponds. None were captured from southern Taiwan. On Kinmen Island only one hatchling P. sinensis was caught.
Discussion
M. sinensis was formerly a common species in southern and northern Taiwan (Stejneger, Reference Stejneger1907; Pope, Reference Pope1935), inhabiting slow-moving water of open country at low elevations, and was regarded as distributed island-wide (Takahashi, Reference Takahashi1934). Although it was the most widely distributed turtle in our survey, capture success at most sites was low, suggesting that population sizes are small.
M. mutica is regarded as a semi-aquatic species (Ernst & Barbour, Reference Ernst and Barbour1989), occurring in water bodies with slow currents in mountain basins or river valleys and making occasional terrestrial excursions (Pope, Reference Pope1935; Mao, Reference Mao1971; Wang et al., Reference Wang, Chen, Hsia and Chen1984; Ernst & Barbour, Reference Ernst and Barbour1989). We trapped M. mutica in the vicinity of broad-leaf woodland or secondary forest. Similar to our findings, earlier surveys in central and southern Taiwan detected only a few individuals in the central region (Nantou County) and none in the southern region (Lin, Reference Lin1996; Peng, Reference Peng1997a,Reference Pengb, Reference Peng2000a,Reference Pengb, Reference Peng2001).
Although M. reevesii occurs throughout China, Korea and Japan we located only a few individuals, on Kinmen Island. Compared to other species this turtle was recorded relatively late in Taiwan, suggesting that, even formerly, it was less abundant or had a narrow geographical range. It was first reported on the basis of a specimen collected near Taipei, in 1931 (Horikawa, Reference Horikawa1934). Kuntz & Dien (Reference Kuntz and Dien1970) collected some specimens from the Taipei basin (detailed localities are accessible in the collection database of the Field Museum of Natural History, Chicago, USA). Although Mao (Reference Mao1971) collected a few specimens near Taipei and in central Taiwan (Nantou) the species has not been found in central Taiwan in more recent surveys (Lin, Reference Lin1996; Peng, Reference Peng1997a,Reference Pengb). However, although we did not locate M. reevesii on mainland Taiwan the finding of M. reevesii-related hybrids in northern Taiwan indicates that this turtle may exist at a low density or may have been recently extirpated. Analyses of mtDNA of these hybrids indicate that they are closely related to M. reevesii (T.-H. Chen, unpubl. data).
P. sinensis was rarely captured in this study although it was earlier reported as a widespread species (Chen, Reference Chen1969). This species is highly aquatic and carnivorous, preferring slow-moving habitats with a bottom substrate of soft mud or sand (Wang et al., Reference Wang, Chen, Hsia and Chen1984; Ernst & Barbour, Reference Ernst and Barbour1989). It has been commonly found in central and southern Taiwan, although at low densities (Lin, Reference Lin1996; Peng, Reference Peng1997a,Reference Pengb, Reference Peng2000a,Reference Pengb, Reference Peng2001). P. sinensis may be underrepresented in our survey because of its specialized food habits and habitat preference. Wild-caught P. sinensis are regarded as a valuable food and have been harvested extensively. Collection pressure may be higher in southern Taiwan, where most of the turtle farms are located, than in other regions, and we did not locate any individuals in this region. The single P. sinensis that we located on Kinmen Island may be an accidental escapee or intentional release from experimental turtle farming conducted by the local government.
Although we did not evaluate trapping efficiency in this study the highly male-biased sex ratios of captured turtles and the generally low capture success raise concerns regarding the population stability and persistence of the four native species. Unbalanced sex ratios seem to be common in natural populations of turtles (Bury, Reference Bury, Harless and Morlock1979). This could be explained by sampling biases, differential mortality of the sexes or differences between the sexes in age at maturity (Gibbons, Reference Gibbons and Gibbons1990). In northern Taiwan the male bias of the turtle sex ratio increased dramatically following a flood control project (Chen & Lue, Reference Chen and Lue2009), suggesting that male-biased sex ratios are unlikely to be due only to sampling bias.
The structure of freshwater turtle populations may be affected by disturbances such as road mortality, habitat fragmentation and recreational activities (Dodd, Reference Dodd1990; Garber & Burger, Reference Garber and Burger1995; Rizkalla & Swihart, Reference Rizkalla and Swihart2006). In disturbed habitats the population structure of freshwater turtles may become male-biased and adult-dominated because of differential mortality between sexes and decrease in recruitment (Marchand & Litvaitis, Reference Marchand and Litvaitis2004; Gibbs & Steen, Reference Gibbs and Steen2005), and the increasingly male-biased sex ratios are associated with differential road mortality between the sexes (Gibbs & Shriver, Reference Gibbs and Shriver2002; Marchand & Litvaitis, Reference Marchand and Litvaitis2004; Steen & Gibbs, Reference Steen and Gibbs2004). The unbalanced sex ratios found in our surveys were probably caused by a combination of habitat alteration and sex-biased road mortality.
Habitat destruction and loss, particularly in riparian and aquatic habitats, may be the greatest threat to freshwater turtles in Taiwan. In most river management projects the requirements of aquatic fauna are not considered and rapid habitat alteration in such projects has had negative impacts on turtle populations (Bodie, Reference Bodie2001; Chen & Lue, Reference Chen and Lue2009). Taiwan has experienced rapid human population expansion in the past 50 years and this has resulted in extensive urban and suburban development and consequent loss of many important aquatic habitats for freshwater turtles. In addition, the extension of the road network has had localized impacts on turtle populations.
Although the Taiwanese rarely consume the meat of hard-shelled turtles because of religious beliefs and cultural factors, many wild turtles (mainly M. sinensis) have been collected for the pet trade, replenishment of turtle farm breeding stocks, and illegal trade to China (Chen et al., Reference Chen, Lin and Chang2000). The harvest of turtle species that are not on the protected wildlife list (M. sinensis and P. sinensis) has never been closely regulated in Taiwan. The sustainability of unmanaged freshwater turtle resources has therefore been questioned (Thorbjarnarson et al., Reference Thorbjarnarson, Lagueux, Bolze, Klemens, Meylan and Klemens2000; Cheung & Dudgeon, 2006). The regional commercial trade of freshwater turtles is a large-scale business in Asia, mostly associated with mainland China (van Dijk et al., Reference van Dijk, Stuart and Rhodin2000). The native freshwater turtles in Taiwan, both the species with legal conservation status and the unlisted species, are also at risk of being traded illegally to China. Without effective management of commercial harvests, extirpation of populations of freshwater turtles in Taiwan is likely unless conservation efforts are undertaken soon.
Based on our results the conservation status of the freshwater turtles of Taiwan was re-evaluated in 2008 and M. reevesii was upgraded under the Wildlife Conservation Act from Rare and Valuable to Endangered because of the species’ limited distribution range and low population density. More effective management plans are still nevertheless needed for the long-term survival of all of the freshwater turtle species of Taiwan.
Acknowledgements
We thank E.Y. Tsoa and two anonymous reviewers for their valuable comments on this manuscript. We greatly appreciate the help of S.-L. Liu, Y.-D. Lue and T.-W. Tzeng, who assisted in fieldwork. This study was supported by the National Science Council and Council of Agriculture, Republic of China. Studies were conducted under the permits of the Wildlife Conservation Law from the Council of Agriculture (Permits to T.-H. Chen: 0920127668, 0931614045, 0941611580).
Appendix
The appendix for this article is available online at http://journals.cambridge.org
Biographical sketches
Tien-Hsi Chen is a member of the IUCN Tortoise and Freshwater Turtle Specialist Group and has worked mainly on the ecology, status and conservation of the freshwater turtles and marine turtles of Taiwan. He is also concerned about the potential impacts of introduced red-eared sliders on the native fauna and has studied the turtle shell trade in the traditional Chinese medicine market. Kuang-Yang Lue’s research focuses on the biogeography and phylogeny of amphibians and reptiles in East Asia, life histories of amphibians in Taiwan, and expansion of introduced amphibians and reptiles. He has described several new species of frogs, salamanders and lizards from Taiwan and has also been widely involved in biodiversity conservation education.