Research on nest site selection and nesting behavior is essential for species conservation. Research on the reproductive ecology of Reeves' turtles is relatively scarce. Therefore, from April to September 2022, we conducted a study in Qichun County, China, on the field nest site selection of 11 female Reeves' turtles using tracking line boxes. We also investigated the nesting behavior of 20 female turtles using high-definition surveillance cameras placed within an enclosure and recorded the clutch size. The results showed that slope, canopy cover, and distance from the water source significantly influenced nest site selection. Canopy cover significantly differentiated selected quadrats from random quadrats and the distance from the water source affected nest predation risk. The nesting behavior was divided into five stages: landing, excavation, egg laying, covering, and return, with a nesting season from early May to the end of July and nesting most occurring during dusk and dawn. The average clutch size was 7.33 ± 0.33 eggs. Reproduction was affected by high nest predation rates and extreme weather conditions. This study provides valuable information on this endangered species and a foundation for the development of conservation plans.

Nest site selection

This study was conducted in Qichun County, Hubei Province, China (29°59′–30°40′ N, 115°12′–115°56′ E), from April to August 2022. We captured 11gravid female wild Reeves' turtles (using a portable X-ray machine to scan the turtles' lower abdomen to confirm the presence of hard-shelled eggs ) using cage traps , which were dismantled after capture to avoid harm to local wildlife. Then, a tracking line box was attached to the rear of the turtle's carapace. A fishing line was wound around the pulley to a length of approximately 70 m. The total weight of the tracking line box and adhesive did not exceed 40 g. Because the experimental turtles were large (≥ 400 g), the tracking line boxes did not affect their normal movement. The experimental Reeves' turtles were then released where they were captured. To prevent the fishing line from running out and affecting the turtle's normal activity, the specific locations of the turtle were determined at 06:00, 14:00, and 20:00 and the fishing line was reeled in. Signs of nest construction were checked along their path to identify the nests. After the experiment was completed, all tracking line boxes were carefully removed and the experimental turtles were released into the aquatic habitats nearest to their nesting sites.

After locating a nest site, two quadrats of different sizes centered on the nest were established. In the small quadrat (1 × 1 m), we quantified soil hardness, leaf litter thickness, shrub and herb cover, and height. In the large quadrat (5 × 5 m), we quantified the slope, slope position, slope direction, distance from human disturbance, canopy cover, vegetation type, distance from the water source, and height above the water source.

Nest characteristics, clutch size, and egg characteristics

After locating the nest, the eggs were gently excavated while minimizing damage to the nest structure. The side of the egg with the fertilization spot facing upward was placed and the nest depth, inner width, entrance diameter, egg length, and egg width were measured using Vernier calipers. Egg weight was measured using an electronic balance and the number of eggs in the clutch was recorded. After measurements were taken, the eggs were placed back into the nest and the nest was restored to its original condition.

Nesting behavior

At the study site, we constructed a 15-m² enclosure simulating a natural habitat. Surveillance cameras and temperature loggers were installed in the enclosure. We introduced 20 sexually mature wild female Reeves' turtles captured using cage traps into the enclosure. To cover the entire nesting season, the monitoring system was turned on for 24-h recording from May 25th to September 1st. During the experiment, a video from the monitoring system was downloaded and watched daily to observe nesting behavior. Data such as the time of occurrence, total duration, duration of each behavioral phase, and weather were recorded. After completion of the experiment, all experimental turtles were released into several aquatic habitats where we had previously captured individuals of this species.

Statistical analysis

All statistical analyses were conducted using SPSS 19.0 (SPSS Inc., Chicago, IL, USA). For quantitative data in the study of nest site selection, the normality of the data was assessed using the Kolmogorov-Smirnov test before further analysis. Paired-sample t-tests were used to analyze significant differences in ecological factors between the selected and random quadrats if the data were normally distributed and had homogeneous variances. Descriptive statistics were reported as mean ± standard error (SE). If the data were not normally distributed, the Wilcoxon signed-rank test was used and the results are presented as medians (minimum, maximum). For qualitative data, significant differences were determined using the chi-squared test after calculating the frequency distribution. Discriminant function analysis was used to identify the ecological factors that could most significantly differentiate between the selected and random quadrats, as well as the main factors affecting nest predation. For studies that did not require hypothesis testing, descriptive statistics were presented as mean ± SE. The significance level was set at p < 0.05.