Population-level information is essential for the effective development of management and conservation plans (Gibbs and Amato 2000; López-León 2008; Primack 2012). The collection of basic ecological and demographic data provides essential information (i.e., population size, recruitment, and population structure) for the development of population demography models (Iverson 1982, 1991a; Macip-Ríos et al. 2011; Bernhard et al. 2017). These models are critical in understanding long-term population dynamics, identifying abiotic and anthropogenic factors influencing population trends, and establishing effective conservation plans and strategies (Congdon et al. 1993, 1994; López-León 2008; Bernhard et al. 2017).

Other relevant information, such as morphological variation between age, size, and sex are also important for prioritizing conservation actions (Thomassen et al. 2010) and resolving taxonomic uncertainties. A morphological description of age (or size category) and sex is critical to understand morphological variation within a species populations (Bookstein 1982; Martínez-Freiría et al. 2008; Arendt 2010; Kaliontzopoulou 2011); also, the description of multiple morphometric variables could help identify local adaptations (Stearns 1992). Böhm et al. (2016) suggested that understanding how biological traits and environmental factors interact with threats may help to predict the population stability (current or future) of a species. Furthermore, basic morphological traits can be used in meta-analyses to assess the extinction risk of reptiles (Böhm et al. 2016; Miles 2020). Finally, morphological descriptions which include basic descriptive statistics of meristic and diagnostic characters (Bookstein 1982; Sokal and Rohlf 1995) could add data to resolve taxonomic controversies at the species level.

In Mexico, the genus Kinosternon is represented by about 12 species (López-Luna et al. 2018; Rhodin et al. 2018); however, taxonomic uncertainties remain (Legler and Vogt 2013; Flores-Villela and García Vázquez 2014). One controversial taxon is the Central Chiapas Mud Turtle (Kinosternon abaxillare), known only from the Central Valley of Chiapas and northern Guatemala. It was considered a subspecies of Kinosternon scorpioides (Berry 1978; Berry and Iverson 2001; Legler and Vogt 2013); nevertheless, Iverson et al. (2013) elevated it to full species status based on molecular evidence.

Kinosternon abaxillare is considered as Data Deficient on the International Union for Conservation of Nature (IUCN) Red List (Turtle Taxonomy Working Group [TTWG] 2017; Rhodin et al. 2018). Basic ecological and demographic information for K. abaxillare is restricted to a few anecdotal data on reproduction, growth, and natural history (Stejneger 1925; Álvarez del Toro 1982; Sánchez-Montero et al. 2000; Berry and Iverson 2001; Iverson 2008). However, most of these studies were limited by small sample size, short sampling periods, and meager descriptions of morphological characteristics. Therefore, the aims of this article are to describe the basic population ecology of the species and provide an updated review of its morphology within a population in central Chiapas, México. Specifically, our goals were to estimate population size and density, determine the population's structure, and describe the morphological characteristics of each size class and compare those morphological traits by sex.

METHODS

Study Site. — We studied the population ecology of K. abaxillare in three different ponds (designated A, B, and C) in Villa Hidalgo, Chiapas, Mexico (16°18′13.66″N, 93°9′60″W, WGS 84; Fig. 1). Each pond is located on private land and surrounded by a matrix of pasture for cattle, mango, and corn fields. The ponds are within a 10-km² area and located within the Chiapas Central Depression physiographic region (National Institute of Statistics and Geography [INEGI] 2017). Pond A has an oval shape with an approximate surface area of 0.12 ha, pond B has a rectangular shape and an approximate surface area of 0.11 ha, and pond C has an oval shape and an approximate surface area of 0.01 ha and is used as a wastewater disposal reservoir. Because the ponds were in close proximity (i.e., ∼1.28 km), we pooled and reported the data as a single population.

View Figure

• Download as Powerpoint
• Download Figure

The climate at the study site is warm subhumid (Thornwaite 1931) with abundant rains during summer (Chanona 2013). The rainy season extends from May to October and the dry season from November to April (Servicio Meteorológico Nacional 2019). The highest temperatures occur in May (mean 27.6°C) and the lowest during January (mean 22.1°C). Average annual temperature is 25.1°C. The highest rainfall (mean = 249.9 mm) occurs in July and the lowest (mean = 0.7 mm) during January. The original vegetation of the study site was characterized as subdeciduous low forest (Miranda 2015).

Trapping Protocol. — This study was conducted from March 2018 to February 2019. Fieldwork was carried out for 3 d once a month. Our sampling was standardized among ponds. We used 1 fyke net (with 3 rings with a diameter of 80 cm, a length of 220 cm, and a mesh opening of 1.5 cm) and 2 folding box traps (approximately 70 cm long, 40 cm wide, 30 cm high, and a mesh opening of 1 cm) per pond. Traps were baited with fresh fish or chicken liver and intestines and were replaced every morning. Traps were placed in each pond from 0600 to 1900 hrs and checked every 3 hrs. Sampling effort for the season was 1404 trap/hours.

Turtle Processing Protocol. — All captured turtles were marked using the shell-notch code system from Cagle (1939). We measured body mass (BM) to the nearest gram using a spring scale (Eisco PH0036C, Victor, NY). We recorded the following morphological characters: straight-line midline carapace length (CL), maximum carapace width (CW), maximum carapace height (CH), straight-line midline plastron length (PL), maximum plastron width (PW), right bridge length (RBL), maximum length of the anterior plastral lobe (Lant), maximum width of the anterior plastral lobe (Want), maximum length of the posterior plastral lobe (Lpost), maximum width of the posterior plastral lobe (Wpost), gular scute length (Gu), interhumeral seam length (Hu), interpectoral seam length (Pe), interabdominal seam length (Ab), interfemoral seam length (Fe), interanal seam length (An), and head width (HW). All measurements were recorded to the nearest 0.01 mm using a dial caliper (Neiko 01407A, China).

Size classes were defined using intervals of 10 mm CL following Macip-Ríos (2005), Macip-Ríos et al. (2009), and the results of a pilot study (Reyes-Grajales 2019). We decided to use this categorization to compare with other studies conducted on kinosternids of the same size (Forero-Medina et al. 2007; Vázquez-Gómez et al. 2016; Macip-Ríos et al. 2018). In addition, we divided the CL data into 5 size class categories: yearlings (soft shell, little development of plastron, lacked growth rings in their plastral scutes, and < 49.9 mm of CL); juveniles (secondary sexual characters absent or barely visible and 50–99.9 mm of CL); subadults (secondary sexual characters were partly visible [tails were not fully differentiated, but concave plastron possibly present], and 100–119.9 mm of CL); small adults (full secondary sexual characters visible and 120–139.9 mm of CL; this range was made considering the minimum size reported with enlarged follicles in females following Iverson [2008]); and large adults (> 140 mm of CL). Individuals were sexed by tail size (greater in males than females), concavity of plastron (more evident in males than females), and other secondary sexual characters described by Iverson (1991a) and Legler and Vogt (2013) for kinosternids.

Statistical Analyses. — Considering the number of individuals captured (n = 168) and the percentage of recaptures (> 40%) in our data set, a Jolly-Seber model for open population was conducted to infer population size (Schwartz 2001). This model was implemented with the Rcapture package (Louis-Paul and Baillargeon 2014) in R (R Core Team 2013) using the Chao Mh Model (Chao and Yang 2003). To calculate density, we used the calculated population size and the sum of available water surface of each pond in the study area. To estimate standing crop biomass, we multiplied the population density by the mean body mass of the turtles we captured. To determine if the sex ratio was significantly different from 1:1, we used a chi-square (χ²) test (Zar 1999). The plastral formula was calculated based on the average values of each midline plastral seam length, for each size category, and by sex (adults only). Differences in body size and other morphological measurements between males and females were tested with a Student t-test for the variables that attained parametric assumptions. For those variables which did not attain the parametric assumptions, we used a Mann-Whitney U-test. In addition, comparison of the PL/CL ratio was also conducted with a Student t-test. Parametric assumptions for normality and homogeneity of variance were tested using the Shapiro-Wilk test for normality and Bartlett test for homoscedasticity (Zar 1999). All statistical analyses were implemented in R (R Core Team 2013), with an α = 0.05.

RESULTS

Demography. — A total of 168 K. abaxillare turtles was marked during the sampling period. Sixty-four turtles were recaptured, with 7 individuals recaptured multiple times. Captured turtles included 2 yearlings (1.1%), 21 juveniles (12.5%), 29 subadults (17.2%), 36 small adult males (21.4%), 52 small adult females (30.9%), 11 large adult males (6.5%), and 17 large adult females (10.1%). The number of captures per month was highest in January 2019, with 32 individuals captured, and lowest in November, with only 9 individuals captured (Fig. 2).

View Figure

• Download as Powerpoint
• Download Figure

Capture–Mark–Recapture Model. — A population size of 231 (± SE 38) turtles (95% CI, 194.1–265.9) was estimated for the Villa Hidalgo locality in Chiapas. Based on the sampled area (surface water coverage), the total turtle density was estimated at 700 turtles/ha and the standing crop biomass was estimated to be 189 kg/ha. The smallest individual captured had a CL of 36.7 mm and the largest was a male with a CL of 158 mm. Most of the individuals were in the larger size classes and few appeared in the smaller or middle size classes (Fig. 3): 11.3% belonged to the 110–119.9-mm class, 19% to the 120–129.9-mm class, 32.1% to the 130–139.9-mm class, and 13% to the 140–149.9-mm class. The other size classes (8 size categories from 30–39.9 to 100–109.9) had abundances representing less than 6% of captured turtles (Fig. 3). The sex ratio was significantly (χ² = 4.17, p = 0.04) skewed toward females (1:1.5).

View Figure

• Download as Powerpoint
• Download Figure

Morphology. — A summary of body mass and all morphometric measurements of immature individuals is presented in Table 1. The comparison of males and females (adults) is shown in Table 2. We found 1 immature individual (0.5%), 3 males (1.7%), and 3 females (1.7%) that had partial or complete axillary scute seams. The plastral formula was identical in all size classes and sex categories: Ab > An > Gu > Hu > Fe > Pe. Males and females did not differ in body size (CL) (t = 0.1, p = 0.91) and body mass (U = 1338.5, p = 0.07); however, females had a higher carapace (CH) (U = 1101.5, p < 0.05), a wider plastron (PW) (t = 3.03, p < 0.05), a longer bridge (RBL) (U = 779.5, p < 0.05), and a longer interabdominal seam (Ab) (t = 5.04, p < 0.05). Males had wider heads than did females (U = 1132.5, p = 0.02). The other analyzed characteristics did not show statistical differences between sexes (Table 2). Finally, the PL/CL ratio of males (0.90 ± 0.03, range of ratios = 0.80–0.98) was statistically different (t = 3.17, p = 0.001) from females (0.92 ± 0.03, range of ratios = 0.81–0.99).

Table 1. Summary of body mass and morphometric measurements of immature Kinosternon abaxillare captured in Villa Hidalgo, Chiapas. Data presented as mean ± 1 SD (range). BM = body mass (g); CL = carapace length (mm); CW = maximum carapace width (mm); CH = maximum carapace height (mm); PL = straight-line midline plastron length (mm); PW = maximum plastron width (mm); RBL = right bridge length (mm); Lant = maximum length of the anterior plastral lobe (mm); Want = maximum width of the anterior plastral lobe (mm); Lpost = maximum length of the posterior plastral lobe (mm); Wpost = maximum width of the posterior plastral lobe (mm); Gu = gular scute length (mm); Hu = interhumeral seam length (mm); Pe = interpectoral seam length (mm); Ab = interabdominal seam length (mm); Fe = interfemoral seam length (mm); An = interanal seam length (mm); and HW = head width (mm).

View Fullscreen

Table 2. Summary of body mass and morphometric measurements and assessment of sexual dimorphism of adult males and females of Kinosternon abaxillare captured in Villa Hidalgo, Chiapas. Data presented as mean ± 1 SD (range). — = not applicable (see Table 1 for definition of abbreviations).

View Fullscreen

DISCUSSION

This is the first study providing ecological and morphological data (of all size classes and sexes) for K. abaxillare (microendemic categorized as Data Deficient [Rhodin et al. 2018]). Our results are limited to a single location. We report data of monthly activity, population size and density, population structure, sex ratio, and morphological difference among sexes. Most of our ecological data were like other kinosternids of the region. For example, captures of K. abaxillare peaked during the early rainy season (i.e., May and June) when individuals emerge from estivation, returning to their aquatic habitats to feed, reproduce, and restore water balance (Mahmoud 1969; Moll and Moll 2004; Buhlmann et al. 2009). Activity peaks from December and January can be attributed to the recruitment of some turtles from seasonal ponds to permanent ponds, where some individuals begin to estivate due to lack of food or the lowering of water level. Monthly fluctuations of highest occurrence during rainy season have been reported in other turtles from Mexico such as Kinosternon integrum (Macip-Ríos 2005), Kinosternon scorpioides cruentatum, Staurotypus salvinii, and Trachemys grayi (López-León 2008). Estimated population density (700/ha) and biomass (189 kg/ha) were similar to other population estimates for other kinosternids such as Kinosternon herrerai (411/ha, 109 kg/ha; Iverson 1982), K. integrum (1196/ha, 341 kg/ha; Iverson 1982), and Kinosternon sonoriense (815/ha, 100.3 kg/ha; Hulse 1974).

The population structure of K. abaxillare was skewed toward larger individuals. A skewed population structure toward old and larger individuals might reflect an unstable population (Gibbs and Amato 2000). However, the larger number of adults and immatures near reproductive size suggest that this population could persist for many decades, as long as anthropogenic threats (e.g., climate change, subsidized predators, programmed burning of land around ponds, collection for pet trade, habitat loss, and the diversion of water for irrigation) are minimized (Spencer and Thompson 2004; Macip-Ríos 2005; Macip-Ríos et al. 2009). Alternatively, our sampling method could have been ineffective at capturing hatchlings and small juveniles. It is notoriously difficulty to trap small individuals of freshwater turtle species, and many studies (e.g., Chelydra serpentina [Rose and Small 2014], Macrochelys suwanniensis [Johnston et al. 2015], Glyptemys insculpta [Brooks et al. 1992 and Curtis and Vila 2015], and Clemmys guttata [Breisch 2006]) report skewed size distributions similar to our population.

We report a female-biased sex ratio, which is not uncommon for other kinosternids such as Kinosternon oaxacae (Vázquez-Gómez et al. 2016), K. scorpioides (Acuña-Mesén 1990; Barreto et al. 2009), Kinosternon scorpioides albogulare (Forero-Medina et al. 2007), K. s. scorpioides (Vogt et al. 2009), and K. sonoriense (Douglas 2009). This female-biased sex ratio could be related to our sampling protocol or a combination of a land-use and climate change. We fully acknowledge that our sampling duration limits the ability to extrapolate, and long-term monitoring of this population is needed. The site has several unnatural canopy openings caused by anthropogenic activities that could increase nest temperatures (see Janzen 1994). In addition, populations with temperature-dependent sex determination may be unable to evolve rapidly enough to counteract the negative consequences of global temperature change (Ewert and Nelson 1991; Janzen 1994; Heppell 1998). Fortunately, a female-biased sex ratio is not as harmful because 1 male can inseminate several females, which increases recruitment (Heppell 1998).

The morphological characters of K. abaxillare measured and registered in this study agree partially (due to the low sample sizes and details used in published literature) with the qualitative descriptions provided by Álvarez del Toro (1982), Berry and Iverson (2001), Iverson (2008, 2010), and Legler and Vogt (2013). This study is the first to report qualitative and quantitative characteristics of immature size classes (yearlings and juveniles). The absence of axillary scutes in K. abaxillare is the main diagnostic character of this taxon. Iverson (2008) reported less than 10% of K. abaxillare individuals had partial and/or complete axillary scutes (on one or both sides). This finding was consistent with our study, but we found that roughly 4% of individuals had an axillary scute seam on either side of the shell.

The morphological measurements of the shell and scutes of immature individuals of K. abaxillare have similarities to those reported for K. scorpioides (Berry 1978; Berry and Iverson 2001; Legler and Vogt 2013). However, measurement from adults showed similarity with Kinosternon chimalhuaca, K. integrum, and K. scorpioides (Iverson and Berry 1998; Iverson et al. 1998; Iverson 1991b, 2010; Legler and Vogt 2013). The plastral formula of K. abaxillare is identical across all size and sex classes. The lack of ontogenetic variation in the plastral formula has been reported in other kinosternids (Legler and Vogt 2013), such as Kinosternon hirtipes (Iverson 1991b), K. integrum (Iverson et al. 1998), and K. sonoriense (Iverson 1991b). However, K. abaxillare plastral formula differs from that reported for K. chimalhuaca (Iverson and Berry 1998), K. oaxacae (Iverson 1991b), K. scorpioides (Iverson 1991b), and K. s. cruentatum (Legler and Vogt 2013) (Table 3).

Table 3. Plastral formula reported for various kinosternid species (see Table 1 for definition of abbreviations).

View Fullscreen

Our results suggest no difference in average shell length between males and females, but there is a tendency for heavier females (Table 2), presumably due to their higher carapace and larger plastron. Acuña-Mesén and Cruz-Márquez (1993) mentioned that in some species of turtles the females have larger body size because this allows them to house a greater muscular and visceral volume, increasing metabolic efficiency, in addition to providing more space to hold eggs. However, in other studies (Ceballos et al. 2013; Agha et al. 2017), sexual size dimorphisms have been associated with habitat type, with larger females in more-aquatic turtles, and with larger males in semi-terrestrial turtles. However, female-biased sexual size dimorphism has been reported in other kinosternids such as Kinosternon acutum (Iverson 1991b), Kinosternon angustipons (Iverson 1991b), Kinosternon baurii (Iverson 1991b), K. hirtipes (Iverson 1991b), K. scorpioides (some populations: Berry 1978; Iverson 1991b; Berry and Iverson 2001; Bedoya-Cañón et al. 2018), K. s. cruentatum (López-León 2008; Legler and Vogt 2013), K. sonoriense (Iverson 1976, 1991; Hensley et al. 2010), and Kinosternon subrubrum (Iverson 1991b). Iverson (2008) also found no detectable sexual size dimorphism in K. abaxillare, although the plastron was relatively shorter in males than in females. In addition, we detect a statistical difference in the PL/CL ratio in males and females in the studied population; this agrees with what was reported by Iverson (2008).

Kinosternon abaxillare is a poorly studied species from southern Mexico. Although the duration of our study is limited (i.e., 1 yr, March 2018–February 2019) and from only 3 adjacent permanent ponds, this study provides ecological and morphological data for all size classes of this species. More important, our study establishes fundamental baseline data for this species; however, long-term monitoring for this and other populations is needed to reliably describe basic population trends. More research is needed throughout the distribution of K. abaxillare to better understand variation in demographic characteristics, dynamics, and the influence of anthropogenic activities (such as pollution, agriculture, livestock, etc.) and climate change on this microendemic lineage. In addition, an evaluation of its range-wide conservation status is also needed.