Population Status and Ecological Features of the Endemic and Critically Endangered Ta Kou Bent-Toed Gecko (Cyrtodactylus takouensis) in Vietnam

Abstract

Population estimates and microhabitat characteristics are widely used to support conservation decisions. However, there had been no surveys focusing on the population status of the endemic and Critically Endangered Ta Kou bent-toed gecko to inform conservation actions across its distribution range. In this study, we conducted the first field surveys to assess its population status using the mark–capture–recapture method, determine microhabitat characteristics, and identify anthropogenic threats to the species’ survival in Binh Thuan Province, Vietnam. Based on our study results, Cyrtodactylus takouensis was only recorded on granitic rocks at various elevations from 265 to 694 m a.s.l. In total, 148 individuals of C. takouensis were detected in the dry season, and 95 individuals of C. takouensis were encountered in the rainy season. Of these, 73 and 51 adults were documented during the two seasons, respectively. The estimated total population size of C. takouensis was 315 individuals in the dry season (95% confidence intervals ranging between 189 and 581 individuals), whereas it comprised 149 individuals, calculated using the Petersen–Lincoln and Schnabel formula, in the rainy season (95% confidence intervals ranging between 108 and 361 individuals). The estimated difference in total population size was probably due to several factors, such as the rapid growth of interlaced vines making parts of the surveyed transects inaccessible during the rainy season, weather variations, and differences in survey effort and detection probability. Additionally, several microhabitat variables and species behaviors were investigated in both seasons. However, humidity was the only significant environmental variable when compared between the two seasons. Moreover, we found that tourism activities and parasites could pose threats to C. takouensis on Ta Kou Mountain. However, no structured or quantitative framework was employed to assess these risks in this study. Further research is needed to quantify factors affecting the species’ survival.

1. Introduction

The bent-toed gecko lizard genus Cyrtodactylus represents one of the most diverse vertebrate genera in the world, with 381 recognized species and still many undescribed forms [1,2,3,4]. However, this genus is one of the most neglected vertebrate groups in terms of conservation attention since only a few studies have been conducted to provide information on its population status and the main anthropogenic threats to threatened species [5,6,7]. To date, although 381 species of this genus have been discovered, as many as 142 species, approximately 37.3%, have not been assessed by the International Union for Conservation of Nature’s (IUCN) Red List of Threatened species [4,5]. This is the highest number of unassessed species compared to several other genera in the Gekkonidae family, such as Hemidactylus (31.66%), Lepidodactylus (14.89%), Lygodactylus (29.47%), Pachydactylus (1.72%), and Phelsuma (1.87%) [4,5]. Consequently, there is an urgent need to evaluate the population status and identify threats to these species in order to support necessary conservation programs.

Additionally, among those assessed, 13.39% (51 species) of species are categorized as Data Deficient, 33.07% (126 species) as Least Concern, 7.87% (30 species) as Near Threatened and Vulnerable, 4.99% (19 species) as Endangered, and 2.89% (11 species) as Critically Endangered [5]. Nevertheless, current assessments have largely been based on general surveys, observations related to their relative commonness or rarity, and the coverage of their habitats by protected areas [5]. Only few of the assessed species have been evaluated for the population status, distribution range, and habitat suitability [5,6]. Furthermore, threats have mostly been recorded by observation during general herpetological surveys, and interviews have barely been undertaken to better understand their extinction risks [5]. Therefore, there is a significant gap in our knowledge of the population status of and main anthropogenic threats to many threatened species of Cyrtodactylus.

Located in the Indo-Burma hotspot, Vietnam harbors an unprecedented diversity of Cyrtodactylus, with 56 species, and many new ones wait to be discovered [4,8,9]. At the same time, several bent-toed geckos in Vietnam are suffering from anthropogenic threats, such as habitat loss and degradation and pollution due to infrastructure and tourism development, as well as agricultural expansion [5,6,10,11,12,13]. These impacts might negatively affect their survival, especially for threatened endemic species, even when their habitats are covered by protected areas [3,5]. According to the IUCN Red List, there are three species listed as Critically Endangered, three as Endangered, and five as Vulnerable, all of which are endemic to the country [5] (Figure 1). However, only one of them (C. gialaiensis) has been preliminarily surveyed to estimate its population size [6]. As a result, additional information on population status and main anthropogenic threats should be collected for the remaining species.

The Ta Kou bent-toed gecko (Cyrtodactylus takouensis [14]) is locally endemic to Ta Kou Mountain in Binh Thuan Province, Vietnam. This species inhabits the lowland deciduous forests and is often found in association with granite rocks and deep caves [14]. Until now, C. takouensis has only been known from its type locality—Hang To Cave, Ta Kou Mountain [14,15]. Consequently, the species has been classified as Critically Endangered (CR) by the IUCN Red List since 2017 due to its extremely restricted range characterized by an extent of occurrence (EOO) of just 16 km² [15]. Nonetheless, like other bent-toed geckos in the country, C. takouensis has been largely neglected by conservation initiatives. To date, no studies have been conducted to provide information about its population status and microhabitat preferences. Furthermore, all identified threats have been based solely on herpetofauna surveys rather than targeted surveys.

In this study, we carried out field surveys at the type locality of C. takouensis and its surrounding areas to answer four key questions: (1) Can C. takouensis be found elsewhere within Ta Kou Mountain or in nearby regions? (2) Is the natural population of C. takouensis small and fragmented? (3) Does C. takouensis prefer specific microhabitats, and does its habitat preference change between different seasons? (4) What potential threats are adversely affecting the C. takouensis population? Based on the study results, we provide recommendations to mitigate risks associated with a declining population size.

2. Materials and Methods

2.1. Field Surveys

Surveys were conducted at all potential localities of C. takouensis in the southern central coastal region of Vietnam based on our review of previous herpetofauna surveys from 2010 to 2024 and the literature [14,16,17,18,19], as well as interviews with local people and rangers. All the surveys were conducted in both the dry (April 2022) and the rainy seasons (October 2022, January 2024, and November 2024) on Ta Kou Mountain, on Ta Dang Mountain, and in Nui Ong NR, Binh Thuan Province. A total of twenty-one transects were set up, consisting of nine on Ta Kou Mountain, six in Nui Ong Nature Reserve, and five on Ta Dang Mountain. The survey area covered approximately 92.28 km², at elevations ranging from 265 m above sea level (a.s.l.) to 694 m a.s.l.

The transects where C. takouensis was found spanned a total area of approximately 11.42 km² in the dry season and a smaller area of approximately 7.84 km² in the rainy season, which was measured by Quantum GIS software (QGIS Version 3.12.0, Development Team. 2020; available online at http://qgis.osgeo.org [accessed on 8 October 2024]). A team of three researchers and one local ranger participated in each survey, which was carried out after sunset between 18:30 and 00:30 to guarantee the highest detection probability. Although two researchers and one local ranger participated, and the time frame was consistent in both seasons to reduce bias, differences in detection probability and observer effects could still influence the seasonal results (please see more detail in the Discussion Section). Geckos were captured by hand and subsequently released at the collection site after being photographed and measured for their SVL (snout–vent length, from tip to snout to vent) with a digital caliper to the nearest 0.1 mm. Coordinates were recorded using a GPSmap 62s (Garmin, Olathe, KS, USA) in a WGS84 datum. Coordinate data can be shared by the authors upon request.

2.2. Population Estimation

To estimate the population size of C. takouensis, a “capture–mark–recapture” method, which has been successfully used for many lizard species in Vietnam [6,20,21,22,23,24], was employed. In particular, each individual was marked only once on the head or middle of the body with a correction fluid pen (Tipp-Ex, Eltville, Germany) using a series of dots (for juveniles—SVL < 67 mm [14]) or numbers (for adults—SVL ≥ 67 mm [14]) to identify and record the recaptured ones. A series of dots and numbers was assigned continuously each day. For example, on the first day on transect 1, four individuals were captured and labeled as 1 to 4 (adults represented by 1, 2, 3, and 4, and juveniles shown as dots from one to four). The next day, four newly caught adult individuals were labeled from 5 to 8.

Each transect was surveyed at least twice, with an interval of 1–2 days between visits to minimize potential population changes (such as births and deaths) during the survey period. Since there had been no studies on the home range of the species, we considered the home range of the Ta Kou bent-toed gecko to be very small (approximately 0.0001 km²), based on data of another gekkonid species, Gekko japonicus [25]. We therefore assumed that the populations remained closed with no immigration or emigration during the 5–7 days of each survey. In case of two sampling periods, the “Petersen–Lincoln Index” was used following the formula P = (n₁ × n₂)/m₂, where P is the estimated population size; n₁ is the number of marked individuals released in the first sample; n₂ is the size of the second sample; and m₂ is the number of recaptured marked animals. Since m₂/n₂ > 0.1, a binomial confidence interval was applied to obtain confidence intervals for the Petersen–Lincoln estimates [26]. In case of at least three sampling times (one capture and two recaptures), the “Schnabel Index” was applied following the formula below to estimate the population size ($\widehat{\mathrm{N}}$) [26,27]:

$$\widehat{\mathrm{N}}={\displaystyle \frac{{\sum }_{\mathrm{i}=1}^{\mathrm{t}}\left({\mathrm{C}}_{\mathrm{i}}{\mathrm{M}}_{\mathrm{i}}\right)}{{\sum }_{\mathrm{i}=1}^{\mathrm{t}}{\mathrm{R}}_{\mathrm{i}}}}$$

$$\mathrm{Variance} {\mathrm{s}}^2\left({\displaystyle \frac{1}{\widehat{\mathrm{N}}}}\right)={\displaystyle \frac{{\sum }_{\mathrm{i}=1}^{\mathrm{t}}{\mathrm{R}}_{\mathrm{i}}}{{\left({\sum }_{\mathrm{i}=1}^{\mathrm{t}}\left({\mathrm{C}}_{\mathrm{i}}{\mathrm{M}}_{\mathrm{i}}\right)\right)}^2}}$$

$$\mathrm{Standard} \mathrm{error} {\mathrm{s}}_{\overline{\mathrm{x}}}\left({\displaystyle \frac{1}{\widehat{\mathrm{N}}}}\right)=\sqrt{{\mathrm{s}}^2\left({\displaystyle \frac{1}{\widehat{\mathrm{N}}}}\right)}$$

where M_i is the total number of previously marked animals at time i; C_i is the number of animals caught at time i; and R_i is the number of marked animals caught at time i. Confidence intervals for the Schnabel population estimate were obtained from the Poisson distribution because the total number of recaptures (∑R_i) was less than 50. In particular, a 95% confidence interval on ΣR_i was calculated as follows [28,29]:

$$\mathrm{Lower} 95\% \mathrm{confidence} \mathrm{limit} = {\displaystyle \frac{\sum \left({\mathrm{C}}_{\mathrm{i}}{\mathrm{M}}_{\mathrm{i}}\right)}{\sum {\mathrm{R}}_{\mathrm{i}}}}$$

$$\mathrm{Upper} 95\% \mathrm{confidence} \mathrm{limit}={\displaystyle \frac{\sum \left({\mathrm{C}}_{\mathrm{i}}{\mathrm{M}}_{\mathrm{i}}\right)}{\sum {\mathrm{R}}_{\mathrm{i}}}}$$

Regarding population structure, individuals of C. takouensis were classified into two different age groups, including juveniles (SVL < 67 mm) and adults (SVL ≥ 67 mm) [14]. The sex of adults was determined based on the presence of large swollen hemipenal bulges in males and un-swollen ones in females [30,31]. Population densities of all individuals and only adults were further calculated per square kilometer (indiv./km²) with reference to each surveyed transect and day (indiv./km²/day).

2.3. Microhabitat Characterization and Behaviors

To record microclimatic parameters, the air temperature (°C) and relative air humidity (%) were measured with a digital thermometer (TFA Dostmann/Wertheim Kat. No. 30.5015, Wertheim, Germany) at each location where animals were captured. A digital infrared thermometer (Mestek, Shenzhen, China) was used to measure temperatures (°C) at the substrate surface and at the ventral body surface of the animals. Moreover, two HOBO pendant temperature data loggers (HOBO, Eichstetten am Kaiserstuhl, Germany) and one Elitech RC-51H USB temperature and humidity data logger, 3200 (Elitech, London, UK), were also employed to record the air temperature (°C) twice per day between April and October 2022 at three transects where animals were found.

Other microhabitat characteristics were also documented: the substrate type (classified as dead leaves, branches, dead wood, rock, roots, soil, trunk), rocky surface (classified as bare or covered with moss and lichen, tree roots), position (outside or inside a rocky cave/crevice), canopy (percentage of vegetation coverage above each animal—estimated by direct observation), substrate condition (dry or wet), substrate angle (between the substrate surface axis and the horizontal axis, ranging from 0° to 180°), and elevation of capture locations using the GPSmap 62s (Garmin, Olathe, KS, USA). The animal posture (hanging—identified as the animals being attached to the underside of rocks; standing—identified as the animals being positioned on top of rocks), activity (resting—identified as the animals standing when encountered; feeding—identified as the animals eating when detected; or moving—identified as animals moving when encountered), and encounter time were further recorded.

2.4. Statistical Analysis

Statistical analysis was performed using Rv4.4.3 [32]. A Chi² test was applied to test the difference in population structure between the dry and rainy seasons. For microhabitat characteristics, six categorical variables (i.e., substrate type, rocky surface, location, substrate condition, animal posture, and activity) were also seasonally compared using Chi² tests. Other microhabitat traits (i.e., canopy, air temperature, substrate temperature, animal temperature, and humidity) were estimated for normal distribution using Shapiro–Wilk tests and tested for differences by Wilcoxon tests. For all these tests, a significant difference was determined using the p-value (p < 0.05) and effect sizes (r ≥ 0.5).

2.5. Threat Identification

To detect impacts of human activities on the species in Vietnam, two large local markets near Ta Kou NR were visited, and ten local people, each having lived near Ta Kou NR for at least two generations and knowledgeable of the area, and one ranger were also interviewed to determine the local use of the species (Appendix A). Nocturnal and diurnal surveys were also carried out to obtain evidence of human disturbances and direct threats to the species, such as poaching, deforestation, direct-burning incenses in some caves, and littering on Ta Kou Mountain and along transects in the surrounding areas. Additionally, data related to direct effects on the species were collected during the field surveys.

3. Results

3.1. Population Status

The Ta Kou bent-toed gecko was not found on Ta Dang Mountain (an isolated mountain belonging to Ta Kou NR) or in Nui Ong Nature Reserve, and it was only detected along three transects (transects 1, 2, and 3) on Ta Kou Mountain. A total of 243 individuals of C. takouensis were encountered along the transects, including 118 captured individuals (73 adults—62%) and 30 escaped ones during the dry season and 76 captured individuals (51 adults—67%) and 19 escaped ones during the rainy season (

Table 1. Total observed numbers and estimated population size of Cyrtodactylus takouensis on Ta Kou Mountain. D_A = density estimation of adult individuals per km²; D_AD = density estimation of adult individuals per km² per day; D = overall density estimation; D_D = overall density estimation per km² per day; N = estimated total individuals. * During the capture–recapture event, this transect was surveyed two times. Therefore, Nicole Peterson’s formula was applied for estimation.

Transects	Transect 1 (T1)	Transect 2 (T2)	Transect 3 (T3)	Total
Dry Season—April 2022
Total Adults	14	21	38	73
Total Obs	19	42	87	148 (NA = 30)
Area (km2)	2.36	2.97	6.09	11.42
DA (Adults/km2)	5.93	7.07	6.24	6.39
DAD (Adults/km2/day)	1.48	1.77	1.56	1.60
D (ind./km2)	8.05	14.14	14.29	12.96
DD (ind./km2/day)	2.01	3.54	3.57	3.24
N–Total Schnabel	27	109	179	315
Variance	1.48 × 10−6	1.19 × 10−5	0.0156	0.015638
Standard error	0.00122	0.00345	0.125	0.1297
Lower 95% confidence limits	119	56	14	189
Upper 95% confidence limits	294	233	54	581
Rainy season—September 2022
Total Adults	8	7	36	51
Total Obs	11	12	72	95 (NA = 19)
Area (km2)	1.23	0.52	6.09	7.84
DA (Adults/km2)	6.50	13.46	5.91	6.51
DAD (Adults/km2/day)	2.17	4.49	1.97	2.17
D (ind./km2)	8.94	23.08	11.82	12.12
DD (ind./km2/day)	2.98	7.69	2.96	3.03
N–Total Schnabel	14	24 *	142	180
Variance	0.00108	-	2.93 × 10−6	-
Standard error	0.00171	-	0.0329	-
Lower 95% confidence limits	6	10	92	108
Upper 95% confidence limits	35	75	251	361
Human impacts	Severely disturbed	Intact	Disturbed

). We also documented 61 recaptures in both seasons (37 and 24 individuals in the dry and rainy seasons, respectively). In terms of the sex ratio among adult captured individuals, the number of females was higher than that of males in both seasons (Figure 2, F:M = 42:31 in the dry season and 29:22 in the rainy season). However, there were no significant differences in sex ratio between the dry and rainy seasons (Chi² = 0.995, effective size = 0.072, CI lower = 0.00, CI upper = 0.19, df = 2, p = 0.608 > 0.05). According to the Petersen–Lincoln and Schnabel indices, the total population size of C. takouensis was estimated at about 315 individuals in the dry season (95% confidence intervals ranging between 189 and 581 individuals) and 180 individuals in the rainy season (95% confidence intervals ranging between 108 and 361 individuals) (Table 1).

In general, the average densities of C. takouensis were similar between the two seasons, around 12–13 individuals/km² (6–7 adults/km²). When recaptures on surveyed days were incorporated, the average density in the rainy season (7.5 individuals/km²/day) was higher than that in the dry season (3.8 individuals/km²/day) (Table 1). There was substantial variation among the surveyed transects in terms of the population density (minimum: 2.97 individuals/km²/day along T1 during the dry season; maximum: 13.47 individuals/km²/day along T2 during the rainy season).

3.2. Microhabitat Selection and Behaviors

Cyrtodactylus takouensis was only found on granitic rocks, both with and without the coverage of evergreen broad-leaved forest intermixed with ferns, shrubs, and vines on Ta Kou Mountain, at elevations from 265 m to 694 m a.s.l. (Figure 3). The gecko species was not found at dragon fruit plantation sites, nor in some small bamboo areas. A majority of individuals were sighted on granite rocks in both seasons (83% and 92% in the dry and rainy seasons, respectively) (Figure 4A–D). Consequently, there were no significant differences in substrate type between the two seasons (Chi² = 11.01, effective size = 0.17, CI lower = 0.00, CI upper = 0.28, df = 6, p = 0.088 > 0.05; Figure 4A,B). Concerning rocky substrates only, the species was mostly found on the bare surface of granite rocks in both seasons (65% and 57% in the dry and rainy seasons, respectively), followed by rocks covered by moss/lichen (32.7% and 41.9% in the dry and rainy seasons, respectively) or small tree roots (2% and 1% in the dry and rainy seasons, respectively). However, there were no significant differences in the surfaces of rocky substrates between the two seasons (Chi² = 2.31, effective size = 0.10, CI lower = 0.00, CI upper = 0.21, df = 2, p = 0.315 > 0.05; Figure 4B). Therefore, substrate types and rocky surfaces may not be important environmental factors when comparing between the two seasons.

Additionally, 80% of captured individuals of C. takouensis were observed on dry substrates in the dry season, while 69.9% of individuals were on wet substrates in the rainy season, with a significant difference between the two seasons (Chi² = 59.02, effective size = 0.50, CI lower = 0.37, CI upper = 0.63, df = 1, p < 0.0001; Figure 4D). In terms of behaviors, the species was mainly encountered hanging (dry: 56.6%, rainy: 74.7%) (Chi² = 7.75, effective size = 0.18, CI lower = 0.05, CI upper = 0.31, df = 1, p = 0.005 < 0.05), outside granite caves (78% and 51.6%, respectively) (Chi² = 16.65, effective size = 0.28, CI lower = 0.14, CI upper = 0.41, df = 1, p = 4.48 × 10⁻⁵ < 0.05), but mostly resting (more than 80%) (Chi² = 3.02, df = 2, p > 0.05) in both seasons, with no difference between the two seasons (Figure 4C,D,F).

The canopy coverage was much less dense during the dry season (86.71 ± 2.3%; n = 123) compared to the rainy season (90.54 ± 2.57, n = 93). However, there were no differences in canopy coverage between the dry and rainy seasons (W = 4757.5, r = 0.18 < 0.5, CI lower = 0.05, CI upper = 0.32, p = 0.007 < 0.05; Figure 5;

Table 2. Environmental parameters characterizing the microhabitat of Cyrtodactylus takouensis.

Parameters	Dry Season	Rainy Season	Wilcoxon Test
Canopy cover [%]	0–100 (86.71 ± 2.3) (n = 123)	0–100 (90.54 ± 2.57) (n = 93)	W = 4757.5, r = 0.18 < 0.5, CI lower = 0.05, CI upper = 0.32, p-value = 0.007 < 0.05 -> no significant seasonal differences
Elevation [m]	267–700 (434.25 ± 14.15) (n = 146)	265–678 (358.14 ± 0.05) (n = 93)
Animal Temp. [°C]	20.3–34.2 (24.08 ± 0.22) (n = 118)	20.8–25.8 (23.72 ± 0.12) (n = 76)	W = 4450, r = 0.006 < 0.5, CI lower = 0.002, CI upper = 0.16, p-value = 0.93 > 0.05 -> no significant seasonal differences
Substrate Temp. [°C]	19.1–27.3 (22.57 ± 0.12) (n = 144)	20.8–24.4 (22.82 ± 0.08) (n = 93)	W = 6034.5, r = 0.08 < 0.5, CI lower = 0.01, CI upper = 0.20, p-value = 0.20 > 0.05 -> no significant seasonal differences
Air Temp. [°C]	23.3–30.5 (26.76 ± 0.13) (n = 145)	23.8–30.9 (26.75 ± 0.19) (n = 93)	W = 6958, r = 0.03 < 0.5, CI lower = 0.002, CI upper = 0.16, p-value = 0.68 > 0.05 -> no significant seasonal differences
Air Humidity [%]	40–85 (62.16 ± 0.87) (n = 145)	69–88 (81.77 ± 0.58) (n = 93)	W = 644, r = 0.76 > 0.5, CI lower = 0.72, CI upper = 0.80, p-value < 2.2 × 10−16 < 0.05 -> significant seasonal differences
Slope [°]	0–180 (83.12 ± 3.51) (n = 136)	0–180 (97.03 ± 4.04) (n = 91)	W = 4682.5, r = 0.21 < 0.5, CI lower = 0.09, CI upper = 0.34, p-value = 0.002 < 0.05 -> no significant seasonal differences

). The microclimatic niche of C. takouensis was characterized by overlapping ranges of air temperatures (23.3–30.9 °C, n = 238) and substrate temperatures (19.1–27.3 °C, n = 237) across both seasons, with no statistically significant differences between the dry and rainy seasons (p > 0.05; Figure 5; Table 2). In contrast, the mean relative humidity at microsites was 62.16 ± 0.87% (40–85%, n = 145) during the dry season, significantly lower than that of 81.77 ± 0.58% (69–88%, n = 93) during the rainy season (W = 644; p < 0.001; Figure 5; Table 2). Therefore, canopy, air temperature, substrate temperature, and animal temperature might not be significant determining environmental variables when comparing between the two seasons.

Three temperature dataloggers at different times (dry season: one month; rainy season: five months) recorded air micro-temperatures ranging from 22.43 °C to 28.95 °C (Figure 6).

We further measured the body temperature of C. takouensis individuals (dry: 24.08 ± 0.22 °C, rainy: 23.72 ± 0.12 °C, W = 4374, p > 0.05; Table 2). We also noted that the body temperature of C. takouensis was correlated with the substrate temperature (r = 0.4, t = 6.06, df = 192, p = 7.14 × 10⁻⁹ < 0.05) and air temperature (r = 0.19, t = 2.63, df = 192, p = 0.009 < 0.05). Regarding its active time, C. takouensis was found at all night survey times, with the main peak occurring between 19:00 and 20:00 in both seasons (Figure 7).

3.3. Main Threats

Even though individuals of C. takouensis are neither collected for food nor poached for the pet trade, tourism activities and parasites are considered important threats to its wild populations. In particular, infrastructure developments associated with the cable car built a decade ago are still expanding to accommodate the growing tourism around transect 1, since Ta Kou NR has become a popular tourist destination in southern Vietnam in recent years. Additionally, transect 1 was visited every day and flashlit every night by tourists in both seasons, while another surveyed transect (transect 3) was visited at least twice or three times a month. During surveys along transects in both seasons, a large amount of plastic waste left by tourists was observed at several locations where C. takouensis was found (Figure 8A,B). Furthermore, many individuals (dry: 22; rainy: 26) were detected with ticks as ectoparasites and an unidentified parasite as endoparasites on the body surface (Figure 8C,D). Although we could not identify the parasites, they were common in adult individuals of C. takouensis (23% during the dry season and nearly 43% during the rainy season). Further studies are therefore needed to provide more information on the parasites and how they affect the fitness of C. takouensis in the future.

4. Discussion

4.1. Population Status

Using capture–mark–recapture data, we provided the first estimate of the Ta Kou bent-toed gecko population. Our results show that this endemic and Critically Endangered species possesses a very small population size in our study area. The effective population size only reflects the assessment at the three known sites of the species on Ta Kou Mountain. Potential observer bias limits the ability to generalize our findings to other areas where the species may exist. The total population size of the species is assumed to be limited and does not exceed the size of a minimum viable population required for long-term stability, which is estimated to be at least 3000 to 7000 individuals [33,34], because this species almost exclusively occupies granite habitats. Therefore, this range-restricted species is especially imperiled by additional anthropogenic threats.

This study also revealed wide seasonal fluctuation in population estimates, which was probably caused by several factors. Firstly, the rapid growth of interlaced vines made parts of surveyed transects 1 and 2 inaccessible in the rainy season. In the dry season, 2.36 km² and 2.97 km² were assessed in transects 1 and 2, respectively, while only 1.23 km² and 0.52 km², respectively, were evaluated during the rainy season (Table 2). Secondly, variations in survey effort and detection probability may account for fluctuations in population estimates. Throughout both seasons, the number of researchers and rangers involved in each survey remained unchanged, with three-quarters of the team members consisting of the same individuals to minimize bias. However, differences in detection probability still occurred among survey members due to personal health or individual circumstances. Moreover, weather conditions such as wind speed and humidity, microhabitat surface wetness, and gecko behavior may also have impacted the detection probability of species in the two seasons. Therefore, population estimates across seasons may not be ecologically meaningful. Future studies should incorporate detectability-corrected frameworks (e.g., N-mixture or occupancy models) to provide more accurate seasonal estimates of the total population. Our results suggest that the mean density of C. takouensis, at 12–13 individual/km², is significantly lower than that of other Critically Endangered gecko species (

Table 3. Comparison of density estimations between C. takouensis and other species.

Species	Locality	Density Estimation	Method Notes	References
Cnemaspis thackerayi	Salam District, India	1,000,000–2,000,000 individual/km2 (10–15 individuals/7.5 m2)	Herpetofauna surveys	[35]
Gonatodes daudini	Grenadines and Saint Vincent	8700–21,900 individuals/km2 (87–218 individuals/ha)	These figures resulted from surveys targeting four cryptic reptile species. The process involved sifting through litter and lifting various types of cover, such as rocks, logs, and deadfall, by two researchers along the transects, while a third person monitored the edges to observe any escaping animals	[36,37]
Lygodactylus williamsi	Tanzania	35,300 individuals/km2 (353 individuals/ha)	Observations by three researchers in the mornings and afternoons	[38]

).

4.2. Microhabitat Selection and Behaviors

Cyrtodactylus takouensis is considered a microhabitat specialist only occupying bare granitic boulders or those covered with moss, lichen, and small tree roots in evergreen forest intermixed with ferns, shrubs, and vines. As it is an ectothermic gecko, the basic physiological functions of C. takouensis, such as locomotion, growth, and reproduction, and its microhabitat selection are assumed to be influenced by environmental conditions [39]. Resembling another insular gecko, Cnemaspis psychedelica, in southern Vietnam, C. takouensis also exhibits seasonal variation in its microhabitat selection, including the substrate condition (dry or wet), canopy coverage, and humidity [40]. Using the ordination test, we demonstrated that canopy and humidity are the most important parameters explaining the seasonal variation in the microhabitat selection of C. takouensis.

4.3. Main Threats

Tourism development has not only substantially fragmented but also resulted in the loss and degradation of animal habitats worldwide [41]. For C. takouensis, due to its very small population size, annual tourism growth could threaten the species’ long-term survival, particularly through habitat disturbance, if not properly managed. Although the impact of plastic trash was not assessed quantitatively, the presence of plastic waste implies potential habitat degradation, which could negatively impact the quality of the limited microhabitat of this species. Furthermore, the prevalent parasites detected in C. takouensis in this study might also adversely affect the population of this species. However, no structured or quantitative framework was employed to assess these risks in this study. Further research is needed to quantify the factors affecting the species’ survival.

4.4. Conservation

In the Vietnam Red Data Book, Nguyen et al. [42] reassessed the conservation status of the species and classified it as Endangered (B1ab(iii)), based on an estimated extent of occurrence (EOO) of approximately 120 km². However, our results suggest that C. takouensis as currently known only occurs on Ta Kou Mountain, where the species occupies a specific microhabitat and is restricted to a small area of the mountain. The population size of C. takouensis is also quite small, with an estimation of 315 individuals, with the number of adult individuals ranging from 51 to 73. Using localities where the species is found, we recalculated the EOO utilizing the GeoCAT tool (available at “https://geocat.iucnredlist.org/editor (accessed on 12 May 2025)”), which was introduced by the IUCN Red List, based solely on the GPS data from our surveys [43]. The EOO of C. takouensis was estimated to be approximately 2 km², which falls within the Critically Endangered category (B2ab(iii)) [44].

Besides its very small range, this species might be particularly vulnerable to infrastructure development, which often results in habitat loss and fragmentation. We therefore recommend that Ta Kou NR closely monitor the construction works along transect 1 and enforce a higher protection level within the granite habitat of C. takouensis, such as limiting visitors from entering the granite areas of transect 1 and transect 3. Further studies should be undertaken to better understand the severity of the parasites found on the species. To save this species in the long term, it is essential to raise awareness among the local community and visitors with regard to littering and habitat protection. Under supported projects, we trained the protected-area staff in using the Spatial Monitoring and Report Tool (SMART) for patrolling and designed and printed signboards and leaflets for educational purposes, but additional long-term conservation measures need to be implemented in the nature reserve.