Effective conservation plans require up-to-date taxonomy to delimit species and accurate information about their geographic distributions to know where those species live. For Chinese turtles, research and conservation are hindered by errors in taxonomy and geographic distribution. Taxonomy, the discipline of recognizing and delimiting species, endlessly reevaluates boundaries between species, resulting in name changes (Thomson et al. 2018). The taxonomy used for Chinese turtle taxonomy in research papers (especially non-taxonomic work) and legislation can be outdated and incorrect (e.g., Chen et al. [2022]use the old name Chinemys reevesii for Mauremys reevesii, a name suggested in Feldman and Parham [2004]). Knowledge about the distributions of species is critical for their effective conservation, yet few data are available for Chinese turtles (e.g., Fong et al. 2002). These information gaps are due largely to the relatively low number of Asian turtle scientists in Asia and the inability of non–Chinese-speaking scientists to make use of the Chinese literature (Parham and Wang 2000; Fong et al. 2002; Fong and Qiao 2010). The wildlife trade that transports turtles around the world has—for decades to centuries—taken thousands of turtles from their native habitats and shipped them around China to local markets (van Dijk et al. 2000). As a result, the natural distribution of turtles has been distorted through the unnatural dispersal and introduction of turtle species to non-native habitats. As it stands, most of our knowledge of Asian turtles is based on purchased specimens, and this has led to substantial confusion regarding their distributions (de Bruin and Artner 1999; Fong and Qiao 2010).

The number of documented localities for turtles on Hainan is sparse (Iverson 1992). Most species are known from only 1 locality, reported to have been collected from “Hainan Island” with no detailed locality data, or were collected by local people or hunters (Schmidt 1927; McCord and Iverson 1992), with the true source likely from local markets (de Bruin and Artner 1999; Zhao 2005). Locality data from specimens bought in markets are unreliable since these turtles, as a result of the trade, could have been transported over large distances to their point of discovery in the market (Lau and Shi 2000).

Hainan Island has rich turtle diversity for China (Zhang et al. 1998; Zhao et al. 2000; Fellowes et al. 2003) and has been designated as a priority area for turtle conservation (Myers et al. 2000; Gong et al. 2003). However, the species of nonmarine turtles recorded from Hainan Island are inconsistent and confusing, as there is a lack of concordance in local, domestic, and international references (Iverson 1992; Zhang et al. 1998; Zhao 1998; Zhao et al. 2000). Natural presence of some species is uncertain, such as Geoemyda spengleri, until Shi (2005). Other species, such as Sacalia bealei, have no clear evidence of being distributed on Hainan Island, but old and incorrect records have been continually cited, perpetuating the belief that this species is found here. These errors are even found in some of the important files about the biodiversity of the provincial government and the state and provincial laws for animal protection in China (Zhao 1998; People's Government of Hainan Province 2006). Although some errors have been corrected, they may persist, such as the literature citing Mauremys iversoni and Cuora serrata as valid species (Endangered Species Import and Export Management Office of the People's Republic of China 2002; Zhou 2004) despite the fact that they were shown to be invalid hybrids (Parham et al. 2001). Often, our authorship team reviews and suggests corrections for submitted papers that erroneously report species to occur on Hainan; some authors accept the suggestions, but others insist on citing previous records even though these species do not exist on Hainan (Turtle Taxonomy Working Group [TTWG] 2021), such as Manouria impressa (Forestry Administration of Hainan Province 2021). Such insistence may be due to the authors referencing earlier records that were wrong, and it is thus necessary to clear up this confusion. In this article, we describe the species of nonmarine turtles on Hainan Island based on literature reviews, museum specimens, and field surveys during the past 27 yrs.

METHODS

Literature Review. — Nonmarine turtle data records from Hainan Island are from the published literature. In the online literature survey, “turtle” and “Hainan” are used as the key words for searching and collecting related literature. Then “reptile or herpetology” and “Hainan” are used as the key words for further searches. The offline literature survey is based on our research group's long-term collection of turtle research results on Hainan Island, such as from books, newspapers, and policy papers. The information collected includes the recording scientist, the recording time, the location, and relevant descriptions of the species.

Museum Surveys. — We conducted surveys at more than 30 domestic and international museums, including Chengdu Institute of Biology (CIB), Beijing Institute of Zoology (BIZ), Kunming Institute of Zoology of the Chinese Academy of Sciences, Shanghai Natural History Museum, Zoological Museum of Wuhan University, South China Institute for Endangered Animals in Guangdong Province (SCIEA), Smithsonian Natural History Museum (Washington, DC), American Museum of Natural History in New York, Chelonian Research Institute in Florida, British Natural History Museum, French Natural History Museum in Paris, and the Museu de Zoologia da Universidade de São Paulo. The original records, if available, were checked carefully. The collections of CIB, BIZ, and SCIEA had the most nonmarine turtle specimens from Hainan Island, but the specimens in BIZ and SCIEA have no labels of collection site and date, undermining the value of these precious collection resources.

Current distributional data must be scrutinized owing to the turtle trade, and therefore Chinese museums and literature hold valuable historical data since many of the specimens predate the mass transport of chelonians in the 1980s (Fong and Qiao 2010). We include specimen data only if local data are more specific than province-level data and the date of collection was before 1980. Although it is useful to know if a species is present in a province, this information is too general for locating persisting wild populations. For date of capture, because the turtle trade dramatically increased in the 1980s, specimens collected before this time were tentatively considered reliable, while specimens collected after this time had a higher probability of being imported into the area (Fong and Qiao 2010).

Integrative Field Surveys. — We conducted 2 yrs (1997–1998) of field surveys on Hainan Island through a nationwide survey of terrestrial wildlife. This work was done by a survey team called the Hainan Wildlife Resource Survey Team (HWRST) composed of 20 people from the South China Institute for Endangered Animals, Hainan Normal University, the Hainan Forestry Bureau, and various nature reserves of Hainan Province. Surveys were done in agricultural fields (77 survey lines, intensity 0.03 km/km²), forest (379 survey lines, intensity 0.11 km/km²), and a specially selected intensive survey area (296 survey lines, intensity 0.97 km/km²) (Song et al. 2002). Afterward, HWRST conducted a 6-mo island-wide survey of wetland resources, a 2-yr comprehensive survey that resulted in 6 new proposals for provincial nature reserves, a 1-yr survey of biodiversity resources to justify upgrading Jianfeng Mountain and Wuzhi Mountain from provincial to national nature reserves, and a 2-wk Global Environment Fund project surveying wildlife resources in the Jianfengling Nature Reserve.

Hainan Normal University has been collecting and researching amphibians and reptiles in the recent half century. In the past 27 yrs, together with zoological field practice with an annual average of 80 undergraduate and 20 graduate students, we surveyed the north (Haikou), west (Bawangling Nature Reserve, lat 18°57′–19°11′N, long 109°03′–109°17′E), southeast (Jianfengling Nature Reserve, lat 18°23′–18°52′N, long 108°44′–109°02′E; Diaoluoshan Nature Reserve, lat 18°43′–18°58′N, long 109°43′–110°03′E), and central (Wuzhishan Nature Reserve, lat 18°48′–18°59′N, long 109°32′–109°43′E; Limushan Nature Reserve, lat 19°07′–19°14′N, long 109°39′–109°48′E) parts of the island (Fig. 1) (Shi et al. 2001, 2011). Additionally, many local-scale surveys have been conducted for implementing ecological, species conservation, and biodiversity survey projects, both domestic and international.

View Figure

• Download as Powerpoint
• Download Figure

Kadoorie Farm and Botanic Garden (KFBG) surveyed Hainan biodiversity from 1998 to 2006. Its Rapid Biodiversity Survey Team, including members from the Beijing Institute of Zoology, South China Normal University, and Hainan Normal University, conducted a wide range of field surveys across Hainan Island (KFBG 2002, 2003a, 2003b, 2003c, 2005).

Specialized Field and Interview Surveys for Turtles. — Our turtle research team began turtle studies on Hainan Island in 1996 and has established 6 long-term research sites. To choose ideal research sites, we conducted integrative turtle surveys all over the island and established population density studies for targeted turtle species among these research sites. The distribution, trade, and captive breeding of targeted species was also implemented at least 1 time around the island. These specialized surveys are as follows:

1. Wanling Town, northern Wuzhi Mountain, Qiongzhong County, for 5 yrs (1997–2002) of research on Sacalia insulensis and Mauremys sinensis (Shi et al. 2005)

2. Nanmao Town, central Wuzhi Mountain, Qiongzhong County, for 19 yrs (2002–2020) of research on Cuora galbinifrons, Cuora mouhotii, and G. spengleri (Shi et al. 2005; Gong et al. 2006b)

3. Limushan Nature Reserve for 14 yrs (2004–2017) of research on S. insulensis (Gong et al. 2006a; Gaillard et al. 2017)

4. Diaoluo Mountain, Lingshui County, for 15 yrs (2006–2020) of research on the nonmarine turtles (Wang et al. 2011a, 2011b; Xiao et al. 2021)

5. Wanquan River in Qionghai City for 3 yrs (2009–2011) of research on the exotic Trachemys scripta elegans (Ma and Shi 2017)

6. Yinggeling for 2 yrs (2019–2020) of research on nonmarine turtles

During each of our specialized field surveys above, we brought a turtle atlas and turtle specimens to serve as examples when speaking with hunters, fishers, farmers, and villagers about which species occur locally (Gong et al. 2017).

Market and Turtle Farm Surveys. — Trade in wild, nonmarine turtles occurs in all cities and counties of Hainan. We conducted intermittent surveys of the markets and restaurants in cities and the countryside across the island starting in 2002 (Wang et al. 2005). We also conducted market surveys in other provinces to determine the spread of the nonmarine turtle trade from Hainan Island. There are 6 major turtle farms on Hainan Island, which are the most important hubs for trading with wild nonmarine turtles, and we have surveyed each at least once (Shi et al. 2001, 2008).

RESULTS

Literature Review. — The results of the literature review show that 23 species of nonmarine chelonians are recorded to be distributed on Hainan Island, including Cuora flavomarginata (Li 1958), C. galbinifrons (Zhao 2005), C. mouhotii (Schmidt 1927), C. serrata (Fritz and Obst 1997), Cuora trifasciata (Schmidt 1927), G. spengleri (Zhao and Adler 1993; Shi 2005), M. impressa (Zong and Ma 1974), Mauremys megalocephala (Zhao and Adler 1993), Mauremys nigricans (Iverson 1992), Mauremys mutica (Schmidt 1927; Zhao 2005), M. reevesii (Zhao and Adler 1993), M. sinensis (Schmidt 1927; Iverson 1992; Gong et al. 2005a), Ocadia philippeni (McCord and Iverson 1992), Palea steindachneri (Schmidt 1927; Li 1958; Gong et al. 2005a), Pelochelys cantorii (Schmidt 1927), Pelodiscus parviformis (Gong et al. 2018), Pelodiscus sinensis (Schmidt 1927), Pelodiscus variegatus (Farkas et al. 2019), Platysternon megacephalum (Schmidt 1927), S. bealei (Schmidt 1927; Zhao 1990), S. insulensis (Lin et al. 2018), Sacalia pseudocellata (Iverson and McCord 1992), and Sacalia quadriocellata (Pope 1935). Among them are 3 invalid species that are invalid hybrids: C. serrata (Stuart and Parham 2004), O. philippeni (Stuart and Parham 2007), and S. pseudocellata (Stuart and Parham 2007). M. megalocephala was demonstrated to be a synonym of Mauremys reveesii, with diet variation leading to head variation (Iverson et al. 1989; Zhang et al. 1998; Barth et al. 2003). Three new species were recorded from Hainan Island: S. insulensis, P. parviformis, and P. variegatus (Gong et al. 2018; Lin et al. 2018, 2022; Farkas et al. 2019). An earlier record of S. insulensis was found under the name S. quadriocellata (Lin et al. 2018, 2022; Shi et al. 2008).

Museum Surveys. — The CIB is recognized as the Chinese center for herpetological research. We checked the original records, with collection dates from 1957 to 1990. Here we list each of the 11 nonmarine turtle species encountered, with the number of specimens in parentheses: P. megacephalum (2), C. galbinifrons (1), C. trifasciata (6), C. mouhotii (1), C. flavomarginata (2), S. insulensis (6), G. spengleri (1), M. mutica (1), M. sinensis (1), P. sinensis (5), and P. steindachneri (4). Of these, 9 specimens were recorded to have been purchased from the market of Haikou, 15 were recorded to have been bought from the local villagers, and 2 had no locality data. Two specimens (1 S. insulensis and 1 C. trifasciata) had specific collection records, including the name of the village, altitude, temperature of weather and water, and even water pH. Owing to their recording date, these 2 specimens likely represent valid specimens collected from the wild. There was important information found for a S. bealei specimen; on the specimen card of specimen no. 105094, the original locality of “Diaoluoshan, Hainan” was crossed out and replaced with “Longnan County, Jiangxi Province”. This is a valuable clue to how S. bealei was believed to have been erroneously reported to have a distribution on Hainan Island.

For the Zoological Museum of Wuhan University, there were 2 records for S. bealei in the original record cards, with the collection dates being 5 April 1957 and 20 October 1972. However, labels on the specimen jars were both 5 April 1957. Additionally, one of the specimens is S. quadriocellata rather than S. bealei. It is clear that both records and identification were errors. On further questioning, the S. bealei specimen kept by SCIEA was renamed as S. quadriocellata.

Market and Turtle Farm Surveys. — A total of 22 markets (19.6% of total markets surveyed) sold nonmarine turtles. Nonmarine turtles were found at markets in 64 towns, which was 65.3% of total towns investigated. A total of 103 wild turtle purchasing stations were discovered, comprising 489 wild turtle individuals and 10 local species: P. megacephalum, C. mouhotii, C. galbinifrons, C. trifasciata, S. quadriocellata, M. sinensis, M. mutica, G. spengleri, P. sinensis, and P. steindachneri.

Field Surveys. — The HWRST survey recorded 3 species of turtle (S. insulensis, C. galbinifrons, and C. mouhotii). From our own surveys, approximately 120 individuals of 6 species (P. megacephalum, C. galbinifrons, C. mouhotii, M. mutica, M. sinensis, and S. insulensis) were found and marked for ecological studies. In 2005, we confirmed the presence of G. spengleri on Hainan Island, and we recorded a wild hybrid individual of C. serrata. In addition, we confirmed the distribution of P. cantorii on Hainan Island through a long-term investigation. Based on these findings, we verify that there are 12 species of nonmarine turtles on Hainan Island (Table 1).

Table 1. Distribution of 12 native nonmarine turtle species of Hainan.

View Fullscreen

DISCUSSION

The combined results of multiple surveys show that 10 of the 22 species reported previously should be excluded from the list of nonmarine turtles distributed on Hainan Island. Among them, 3 invalid species are invalid hybrids: C. serrata (Fritz and Obst 1997; Stuart and Parham 2004), O. philippeni (McCord and Iverson 1992; Stuart and Parham 2007), and S. pseudocellata (Iverson and McCord 1992; Stuart and Parham 2007). Additionally, in our long-term survey, we never encountered 5 species that had been recorded as naturally occurring on Hainan Island, including S. bealei (Zhao 1990), M. impressa (Zong and Ma 1974), M. reevesii (Zhao and Adler 1993), M. nigricans (Iverson 1992), and C. flavomarginata (Li 1958). Moreover, based on our analysis, S. insulensis replaces S. quadriocellata, P. variegatus is valid, and the natural presence of P. parviformis on Hainan Island seems questionable, even though this species is now likely bred on the island in local farms.

In our museum surveys, a total of 169 nonmarine turtle specimens consisting of 12 species from 15 provinces in China were present in the BIZ, and 110 of the turtle specimens were problematic: 1 was purchased from a market, 1 had a problem with the locality (data to the province level only), 52 had problems with the date (missing data or post-1980), and 56 had problems with both locality and date. The main reason for such a high proportion of problematic turtle specimen records is the lack of turtle experts and the limitations of research methods, which result in errors of taxonomy and distribution. In addition, turtle specimens are obtained mainly from markets and local villagers, which is one of the reasons for many errors in turtle distribution records. The same reasons have led to many erroneous records on Hainan Island.

Market Specimens. — The reason for so many invalid species is that many specimens are collected from the market, and these are unreliable for establishing locality of origin. In the museum surveys, 9 specimens were reportedly purchased from the market of Haikou, the capital city of Hainan, and 15 from local villagers. From the 1980s to the 1990s, more than 10 new species of nonmarine turtles were described (Buskirk 1989), most of these being purchased from markets where artificial hybridization likely occurred (Shi and Parham 2001; Dalton 2003; Stuart and Parham 2007). Fong and Qiao (2010) believe that many specimens are unreliable due to the turtle trade. This is supported by van Dijk et al (2000), who found that since 1992, an extensive trade network has expanded to local food markets throughout Asia. In addition, the ritual of releasing nonmarine turtles by some religious groups has been a common practice, with nonmarine turtles supplied by pet stores or animal vendors near the temples; both local and exotic species are included in such releases (Severinghaus and Chi 1999; Chen et al. 2000). Such releases of diverse nonmarine turtle specimens likely result in genetic mixing of closely related taxa (Shi et al. 2007), and in turn, these hybrids are often captured and traded. When scientists buy them, they are often described as new species (Dalton 2003).

Nonmarine turtles from all over the world are heavily traded across Hainan (Gressitt 1940; Lau and Shi 2000; Gong et al. 2005a). Zhao (2005) reported that 6 of 7 recorded turtle species on Hainan Island were bought from Haikou. However, we conducted a 10-day trip and found more than 400 nonmarine turtles at the Dongmen Market in Haikou in 2006. These turtles most definitely came from beyond Hainan Island rather than within a small radius of the place where they were shown. Samples from markets are not reliable, so previous records need to be carefully scrutinized. For example, Zong and Ma (1974) bought a M. impressa from the market in Haikou, and the erroneous distribution record of this species on Hainan Island persists in the literature (TTWG 2011; Forestry Administration of Hainan Province 2021). Fu and Zhao (1990) justified the separation of a third species of the genus Sacalia on morphological grounds, and this appeared reasonable based on limited genetic evidence from 2 unauthorized pet trade samples (Spinks et al. 2004). The more immediate evidence was that when we were surveying nonmarine turtles in Wanning County—in the southeast of the island—in 1999, locals sold us 3 Cuora amboinensis, suggesting that records of turtles would not be credible if they were not collected by the researchers in the field.

Limitations of Research Methods. — We acknowledge that errors are often the result of incorrect record keeping by scientists due to the limitations of research methods. Many species are identified only by morphological differences, including S. bealei (Smith 1931; Zhao 1990) and C. flavomarginata (Li 1958). Smith (1931) did not notice that S. bealei and S. quadriocellata are 2 different species with distinctive morphological characteristics. In the museum survey, the collection site of the S. bealei (specimen no. 105094) was initially incorrect, which is a very important and valuable clue that helped reveal what led to the erroneous report of S. bealei occurring on Hainan Island. These kinds of reports not only led to the wrong records of S. bealei on Hainan Island but also contributed to the chaotic distribution data for this species in China, which still remains unresolved. For example, the aforementioned specimen of a labeled S. bealei in the Natural History Museum of France is instead S. insulensis (specimen no. 126-2; the origin place is Southeast Asia, 1986); the same problem occurred in the Smithsonian's Natural History Museum (H.-T. Shi, pers. obs.).

Lack of Turtle Experts. — Another major reason for misidentifications and incorrect or incomplete species distributions is the lack of scientific expertise. Li (1958) reported a new subspecies Cuora flavomarginata hainanensis on Hainan Island; this is the first time that C. flavomarginata had been recorded there. However, the photo of the species provided in the article is of C. galbinifrons. Similarly, Li (1958) reported collecting Sacalia bealei bealei and a new subspecies, Sacalia bealei quadriocellata, in the same town (Dali Town, Hainan). Yet based on his description, it is apparent that he made the mistakes of identifying the female S. quadriocellata as S. bealei bealei and the male S. quadriocellata as a new subspecies, S. bealei quadriocellata. Indeed, despite extensive fieldwork on Hainan Island, we never found S. bealei (Shi 2000, 2004; Xiao et al. 2021). The observation of behavior and sexual dimorphism provides further proof that S. bealei quadriocellata are the female of S. bealei (Sachsse 1984; Rodel 1985). The situation was eventually clarified to what we currently accept, which is supported by Fu and Zhao (1990) and Iverson and McCord (1992), recognizing S. bealei and S. quadriocellata as distinct species. Lack of taxonomic expertise also likely resulted in the specimen of S. bealei, collected in Guangdong Province and donated to Wuhan University, being mistakenly identified as S. quadriocellata. Not surprisingly, the SCIEA—which led the research—did not employ any herpetologists at the time. Finally, as an example of an incorrect species distribution, TTWG (2021) states that M. reevesii does not occur on Hainan Island, yet the distribution map in this report nevertheless includes Hainan Island as part of its distribution; we hope this error is addressed in subsequent TTWG publications.

CONCLUSIONS

Accurate distribution records are important not only for scientific research but also for law enforcement, management, advocacy, and education in biodiversity conservation. This article clarifies the distribution of nonmarine turtle species on Hainan Island, providing a basis for protection, management, and law enforcement. From our work exploring museum collections, conducting literature reviews, and carrying out extensive field surveys, we verify that there are 12 species of nonmarine turtles on Hainan Island: P. megacephalum, C. galbinifrons, C. mouhotii, C. trifasciata, G. spengleri, M. mutica, M. sinensis, S. insulensis, P. steindachneri, P. cantorii, P. variegatus, and P. sinensis.

Through developing this well-vetted list of Hainan Island's nonmarine turtle species, we hope to put an end to law enforcement errors, management confusion, and waste of investigation and research resources. As our investigation illustrates, the incorrect distribution records of nonmarine turtles of Hainan Island are due mainly to trade turtle specimens bought from the market, taxonomic errors, incorrect location records, and records before Hainan became an independent province. The types of errors reported here are not confined to Hainan Island, and it is likely that similar incorrect species identifications and distributions are present for other regions and will be repeated again. This warrants similar efforts throughout the region to clarify species presence and distributions.