Site fidelity refers to the restriction of dispersal distance of an animal and its tendency to return to a stationary site. To our knowledge, the homing ability of freshwater turtles and their fidelity is reportedly very low in Asia. We examined mark-recapture data spanning a 4-year period in Diaoluoshan National Nature Reserve, Hainan Province, China, to investigate the site fidelity and homing behavior of big-headed turtles Platysternon megacephalum. A total of 11 big-headed turtles were captured, and all individuals were used in this mark-recapture study. The site fidelity results showed that the adult big-headed turtles (n = 4) had a 71.43% recapture rate in the original site after their release at the same site, whereas the juveniles (n = 1) showed lower recapture rates (0%). Moreover, the homing behavior results showed that the adults (n = 5) had an 83.33% homing rate after displacement. Adult big-headed turtles were able to return to their initial capture sites (home) from 150–2400 m away and precisely to their home sites from either upstream or downstream of their capture sites or even from other streams. However, none of the juveniles (n = 4) returned home, despite only being displaced 25–150 m away. These results indicated that the adult big-headed turtles showed high fidelity to their home site and strong homing ability. In contrast, the juvenile turtles may show an opposite trend but further research is needed.

Field studies were conducted in the Diaoluoshan National Nature Reserve, Hainan Province, China, within an altitude range of 500–1050 m a.s.l. from 2015−2019 (except 2017). Population density of the big-headed turtle was 1.65 individuals per km in our study area, and they have been less affected by hunting in recent years under the strict management of the nature reserve (Xiao et al. 2021). Long cylindrical nylon cages (length: 60 cm, diameter: 33 cm; baited with dried fish) were arranged along the stream at an interval of ~30 m. The traps were set for a total of 3 days and checked every morning. After capture, the carapace length of each individual was recorded, and their sex and age group were determined. Females and males with mean carapace lengths of more than 100 mm and 130 mm, respectively, were regarded as adults (Sung et al. 2015a). For identification, captured turtles were uniquely marked on their marginal scutes, with care to avoid injury. Marked individuals were released at their capture sites (original or home sites) or a translocation site for the site fidelity experiment and then recaptured at their original site or release site throughout the year, as well as during subsequent years. Translocation sites were not randomly chosen; they were, instead, the original sites of another turtle (Figure 2) under two presumptions: 1) that different individuals of the same species would prefer similar habitats, and 2) that the habitats of the translocation sites would be similar in habitat types and characteristics of the original sites. We calculated the recapture rate as the ratio of the times the turtles were recaptured in the original sites and the total time of recapture. This recapture rate was used as an indicator of site fidelity, where a high recapture rate in the original site would indicate a high site fidelity. In the homing behavior experiment, the recapture rate in the original site was used as the homing rate. In addition, the geographic coordinates of the original and translocation sites were determined using a hand-held global positioning system unit and the trajectory distance along the streams to represent the distance between the original and translocation sites. For adults, the translocation sites were 150–2400 m away (except 35 m for one individual), which is beyond an adult home range distance (97 m; Sung et al. 2015b), whereas juveniles were released 25−150 m away. At the end of the experiment, all cages were removed to avoid harming the local wildlife, and all recaptured individuals were released at their original sites.