Brophy (2004) recently reviewed the systematics of the genus Malayemys (Testudines: Geoemydidae [Bataguridae]) and argued for the presence of two taxonomically distinct species. Analyses of head-stripe and shell characters revealed a clear pattern of geographic variation that was consistent with the topography of Southeast Asia and the poor dispersal abilities of these turtles. Turtles from the Mekong River Basin retained the name M. subtrijuga (Schlegel and Müller 1844), whereas those from the Chao Phraya and Mae Klong river basins, coastal areas of southeastern Thailand, and the Malay Peninsula were assigned the name M. macrocephala (Gray 1859).

Malayemys macrocephala is a small geoemydid turtle reaching maximum sizes of 22 cm carapace length (Srinarumol 1995). This species has pronounced sexual dimorphism, with females exhibiting larger overall body sizes, proportionally wider carapaces, and shorter, narrower tails (Ernst and Barbour 1989; Srinarumol 1995; van Dijk and Thirakhupt, in press). Populations of M. macrocephala can be found in virtually all lowland areas of the Chao Phraya River Basin in central Thailand, where it is the most common turtle (van Dijk and Thirakhupt, in press).

Sexual dimorphism and allometry of the turtle shell have been studied extensively (reviews in Mosimann 1956; Berry and Shine 1980; Ernst and Lovich 1986; Gibbons and Lovich 1990). My research interest focused on geographic variation and the possibility of regional differentiation and speciation in M. subtrijuga (sensu lato). Studies of regional variation require the recognition and elimination of character variation due to factors such as sex, age, and ecology. Without such considerations, critical errors in taxonomic judgement are likely to occur.

Although M. macrocephala is a common turtle with high popularity in the pet trade, its biology is known only through an assortment of anecdotal reports. I discovered that despite the seeming abundance of M. macrocephala voucher specimens, few had precise locality data. I was able, however, to assemble a moderately large sample from the Chao Phraya River Basin. This sample permits the first published study to quantify allometry and sexual dimorphism in this species.

Methods

I examined 97 museum specimens of M. macrocephala from the Chao Phraya River Basin of central Thailand. The geographic origin of each specimen was based on museum records, and the sample was divided by sex and life stage. Dial calipers (accurate to 0.1 mm) were used to take the following 29 straight-line measurements on the shell of each specimen: maximum carapace length (CL); carapace width at the level of the seam separating vertebral scutes 2 and 3 (CW); shell height at the level of the seam separating vertebral scutes 2 and 3 (SH); maximum plastron length (PL); maximum width (APLW and PPLW) and length (APLL and PPLL) of both plastral lobes; minimum bridge length (BrL); maximum width and length of vertebral scutes 1, 2, 3, and 5 (Vert1, 2, 3, 5W and L); maximum width and length of pleural scute 1 (Pleu1W and L); medial seam length of gular (GulL), humeral (HumL), pectoral (PecL), abdominal (AbdL), femoral (FemL), and anal (AnL) scutes; and maximum width of gular (GulW), humeral (HumW), femoral (FemW), and anal (AnW) scutes. One meristic character, RLatK, recorded the position (as a proportion) of the right lateral keel as it bisected pleural scute 2. Larger RLatK values corresponded to relatively greater distances from the median keel. The condition of bilateral characters was recorded from the right side of the carapace and the left side of the plastron unless damaged.

Tail morphology was the primary characteristic used for sexual identification in this study. Sexual dimorphism of this character is pronounced in both subadults and adults, with males having much longer and thicker tails (Ernst and Barbour 1989; Srinarumol 1995; van Dijk and Thirakhupt, in press). When tail morphology was not available (shell and skeletal material; some dried specimens), information from museum records formed the basis of sexual identification. Assignment of specimens to appropriate life stages (juvenile, subadult, adult) was based primarily on Srinarumol (1995), who distinguished adults from subadults based on the complete development of testes and ovaries, and subadults from juveniles based on tail morphology.

To test for allometric variation, CL was used as the independent variable for regression analyses (least squares method) of other shell characters. Nontransformed data (mm) were utilized for all specimens that had a determinable sex (juveniles, subadults, adults), and males and females were analyzed separately. The slope and intercept of each regression equation were tested for differences from zero using Student t-tests. Intercepts that were significantly different from zero (α = 0.05) indicated differential growth (i.e., allometry) of the characters involved (Mosimann 1958; Stickel and Bunck 1989).

Sexual dimorphism of shell characters was examined using the regression analyses detailed above. The regression slopes of each bivariate relationship were compared for males and females using analysis of covariance (ANCOVA), with CL as covariate and sex as factor. Significantly different slopes (α = 0.05) indicated sexual dimorphism in the characters regressed against CL (Mosimann and Bider 1960; Mouton et al. 2000). In addition, sexual differences in CL were tested using Student t-test and expressed by the sexual dimorphism index (SDI) proposed by Gibbons and Lovich (1990) and modified by Lovich and Gibbons (1992). Kolmogorov-Smirnov and F-tests were used to verify normality and homogeneity of variances, respectively.

Sexual dimorphism of shell characters was also examined using multivariate techniques. Twenty-eight mensural shell characters were divided by CL, and the resulting ratios comprised most of the data set. RLatK was not divided by CL because it was standardized upon measurement (expressed as a proportion). Using all 29 shell variables, stepwise selection (PROC STEPDISC; SAS, 1989; significance level for entry and removal = 0.30) was used to obtain a set of potential models that would classify turtles relative to their predetermined sex. Final selection of the best model was based on model size and classification accuracy. The best model gave the most accurate cross-validation results (PROC DISCRIM; SAS 1989) and had no more variables than the number of individuals in the smallest sample. This protocol was designed to select conservative models that had a low number of variables and a high level of classification accuracy. Using the best model as defined above, the probability of correctly classifying each turtle relative to its predetermined sex was calculated using the cross-validation results of linear discriminant function analysis (PROC DISCRIM; SAS, 1989). To minimize the effects of allometric variation, only adult and larger subadult turtles of each sex (males ≥ 80 mm CL; females ≥ 100 mm CL) were compared.

Results and Discussion

A frequency distribution of CL (Fig. 1) indicated that females were larger than males. Adult females averaged 148.60 ± 20.23 (mean ± 1 SD) mm CL (114.4–187.0 mm; n = 21), whereas adult males were considerably smaller and averaged 117.21 ± 9.55 mm CL (100.3–130.7 mm; n = 15). This difference in mean CL was statistically significant (t = 5.6, df = 34, p < 0.0001). Subadult females and males averaged 94.64 ± 9.56 (85.3–113.2 mm; n = 11) and 85.74 ± 7.68 mm CL (69.7–95.4 mm; n = 24), respectively. Juvenile females and juveniles of indeterminate sex averaged 75.75 ± 4.63 (68.1–83.4 mm; n = 18) and 57.33 ± 9.33 mm CL (42.7–67.9 mm; n = 8), respectively. Juvenile males could not be distinguished because all individuals < 68 mm CL lacked sexual dimorphism of tail morphology.

View Figure

• Download as Powerpoint
• Download Figure

Srinarumol (1995) reported that adult females and males from his study area averaged 155.48 ± 27.91 mm CL (116.5–220.0 mm; n = 25) and 112.20 ± 9.83 mm CL (100.8–133.0 mm; n = 14), respectively. Srinarumol (1995) also distinguished between subadults and juveniles and found that males could be identified at CL ≥ 80 mm and females at CL ≥ 86 mm.

Allometric growth of the shell was evident (Table 1). Among males, shell shape changed as CL increased proportionally more than shell width (CW, APLW, PPLW), shell height (SH), plastral length (PL and APLL), several scute widths (Pleu1W, Vert1W, Vert2W, Vert3W, HumW, FemW, and AnW), and a few scute lengths (Vert1L, BL, and AnL). For females, shell shape did not change as much because CL did not increase proportionally more than shell width or shell height but did increase proportionally more than plastral length (PL and PPLL) and a few scute widths (Vert1W, Vert3W, FemW, AnW) and lengths (BrL, AbdL, AnL).

Table 1. Allometric relationships of shell characters to carapace length for Malayemys macrocephala from the Chao Phraya River Basin.

View Fullscreen

Allometry of shell characters is a widespread phenomenon among turtles. Srinarumol (1995) performed regression analyses similar to those presented here, but he did not test for differential growth of shell characters. The allometric pattern that emerges for M. macrocephala is one where males grow proportionally longer than wider or higher, whereas females show proportional growth for most characters. This allometry yields adult males with relatively narrower, flatter shells and adult females with relatively wider and higher shells. It is critical to emphasize the interrelatedness of allometric growth and sexual dimorphism. The differences in allometric growth between male and female M. macrocephala produce the sexually dimorphic adults. Such a connection has been demonstrated by other authors working with a variety of turtle species (Mosimann 1956, 1958; Mosimann and Bider 1960; Stickel and Bunck 1989; Ernst et al. 1998).

Sexual dimorphism of the shell was also evident. ANCOVA indicated that the regression slopes of males and females differed significantly (p <  0.05) in 22 of the 28 characters examined (Table 2). Among these, differences in relative shell width, shell height, and plastral length were most significant (p <  0.0001). Females had relatively wider carapaces (CW, Vert1W, Vert2W, Vert3W), higher shells (SH), and wider (PPLW, FemW, AnW) and longer (PL, PPLL, BL, AnL) plastra (Fig. 2). Srinarumol (1995), using a similar method, found females to have relatively wider carapaces, longer plastra, and longer midline gular and pectoral lengths. The SDI for M. macrocephala from this study was calculated as +0.27. This is comparable to the SDI of +0.39 derived from Srinarumol's (1995) data.

Table 2. Comparison of regression slopes (ANCOVA) of shell characters vs. carapace length among male and female Malayemys macrocephala from the Chao Phraya River Basin.^a

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

One character of particular interest was anal scute length (AnL). The present data showed that males had relatively shorter AnL than females (Table 2). Van Dijk and Thirakhupt (in press) stated that males are distinguished from females by the shape of their anal notches. Males have deeper, V-shaped notches whereas females have shallower, round ones. A deeper anal notch would correspond to a shorter AnL. The V-shaped anal notch and relatively shorter AnL allow for a longer precloacal distance in males (Mosimann and Bider 1960). This is significant because the precloacal region of the tail accommodates the male's penis (Mosimann and Bider 1960).

Sexual dimorphism of the shell was also demonstrated by multivariate techniques. The best model to classify turtles according to predetermined sex contained 6 of the original 29 shell character ratios. These were AnL/CL, PPLL/CL, RLatK, Vert3W/CL, FemL/CL, and PecL/CL. Mean values for these 6 shell character ratios are presented in Table 3. Using the 6-variable model, cross-validation results of linear discriminant function analysis correctly classified 93.1% of males and 89.5% of females (Table 4).

Table 3. Shell character ratios—mean ± 1 SE, (range), and [n]—used in discriminant function analysis to classify male and female Malayemys macrocephala from the Chao Phraya River Basin.

View Fullscreen

Table 4. Cross-validation results for male and female Malayemys macrocephala from the Chao Phraya River Basin based on linear discriminant function analysis of shell character ratios (percentages in parentheses).

View Fullscreen

Based on the preceding analyses, a clear pattern of sexual dimorphism emerges for M. macrocephala. Females attain larger sizes (Fig. 1) and have relatively wider and higher shells (carapace and plastron) and longer plastra than males (Fig. 2; Tables 2–4).

According to Gibbons and Lovich (1990), sexual size dimorphism (SSD) may be the result of a trade-off between the benefits of early maturity (increased matings leading to increased reproductive output) and the negative environmental consequences of small body size (increased risk of predation, desiccation, and thermal stress). Small body size may be favored in male M. macrocephala because the benefits of early maturity outweigh the risks of small body size.

Both Berry and Shine (1980) and Gibbons and Lovich (1990) recognized the importance of fecundity as a factor influencing body size in female turtles. Darwin's “fecundity advantage” hypothesis states that natural selection should favor large body size in females because this would allow them to produce more offspring. For turtles in general, larger female size generally results in more or larger eggs (Gibbons et al. 1982). Such a relationship has also been suggested for M. macrocephala specifically (van Dijk and Thirakhupt, in press). Although fecundity selection could induce an increase in overall female size, it should primarily act on the relative size of the abdominal cavity (Mouton et al. 2000). This helps to explain the many relatively wider, higher, and longer shell characters exhibited by female M. macrocephala.

Some authors (review in Gibbons and Lovich 1990) have suggested that SSD is a result of ecological forces or natural selection. The most frequently cited ecological cause is probably competitive displacement (Brown and Wilson 1956; Dunham et al. 1979). In this model, the sexes evolve to exploit different resources in the environment, thereby reducing competition between them. Large females of M. macrocephala consume freshwater mussels, whereas males and other small individuals feed almost exclusively on aquatic snails (Srinarumol 1995; van Dijk and Thirakhupt, in press). The weakness of the displacement model in explaining this situation is that it cannot predict which sex should be larger (Gibbons and Lovich 1990). Rather than ecological factors being the cause of SSD in M. macrocephala, it may be that ecological differences between the sexes are simply consequences of sexually selected dimorphism (Shine 1986).

Malayemys macrocephala has SDI values ranging from +0.27 to +0.39 (Srinarumol 1995 and current study). SDI values for turtles range from −0.45 to +1.75 (Gibbons and Lovich 1990). When compared to other species that have females as the larger sex (mean SDI = +0.36; median SDI = +0.23), M. macrocephala displays average SDI values (Gibbons and Lovich 1990). In summary, the SSD pattern exhibited by M. macrocephala may be the result of a combination of selective pressures. Selection for increased fecundity may produce larger females (Berry and Shine 1980; Gibbons and Lovich 1990), whereas selection for early maturity may result in smaller males (Gibbons and Lovich 1990).