Sexual size and shape dimorphism in the Wushan salamander, Liua shihi (Liu, 1950) (Urodela: Hynobiidae)

Abstract

Sexual dimorphism in size or shape is widespread in the animal world. Many studies have explored this topic, but few have focused on Asian salamanders. In this study, we analyzed morphometric data of the Wushan salamander Liua shihi, an endemic Chinese salamander, to examine sexual dimorphism in size and shape. We evaluated data sets that included 17 morphometric characteristics of 61 females and 55 males using univariate and multivariate methods. Results show that sexual dimorphism in this species includes not only body size (males have a longer snout-vent length than females), but also shape (females have a longer space between axilla and groin and head size than males, while males have greater limb size). This article discusses the evolution of intersexual dimorphism according to models of ecology, fecundity and sexual selection. We propose that sexual dimorphism of body size can be attributed to sexual selection and local climates, that AGS may contribute to fecundity selection, that head size may be attributed to reproductive roles and ecology, and that limb size may be beneficial for reproductive success. However, most aspects of the reproductive biology, ecology and life history of this species remain unknown. This article fills a gap in the literature on this species and builds a foundation for future research. Research on such species is increasingly important as many species of salamander are becoming threatened or endangered throughout the world.

Keywords:

• Sexual dimorphism
• morphometrics
• salamanders
• hynobiids

Introduction

Sexual dimorphism is common throughout the animal kingdom (e.g., Darwin 1871; Shine 1979; Andersson 1994; Butler et al. 2000; Herler et al. 2010) and involves a phenotypic difference between males and females within species. These differences may reflect the divergent selective pressures operating on each sex (Malmgren & Thollesson 1999; Herrel et al. 2012). Different evolutionary mechanisms have been proposed as the cause, including sexual and ecological selection, as well as fecundity (Hedrick & Temeles 1989). Sexual selection is driven by individual variation in reproductive success caused by competition over mates (Andersson 1994), which includes competition between males (intrasexual selection) and female mate choice (intersexual selection) (Bovero et al. 2003; Bakkegard & Guyer 2004; Bakkegard & Rhea 2012).

Fecundity selection favors larger female body size related to larger abdominal volume that improves fecundity (Griffith 1990; Jockusch 1997). Ecological selection contributes to ecological-niche divergence, and occurs when each sex has adapted to different ecological niches, resulting in a different diet (Slatkin 1984; Shine 1989; Fontenot & Seigel 2008).

In amphibians, sexual dimorphism is present in many species (Shine 1979), and dimorphic traits include body size, glandular development, skin texture, dermal ornamentation, vocal sacs, colouration and many other traits (Duellman & Trueb 1994). The sexual dimorphism of salamanders has attracted considerable interest in evolutionary biology (e.g., Halliday & Arano 1991; Kalezić et al. 1992; Andersson 1994; Malmgren & Thollesson 1999; Serra-Cobo et al. 2000; Ivanović et al. 2008), but few studies have focused on Asian salamanders (Kuzmin 1995; Hasumi 2001, 2010; Poyarkov et al. 2012).

The Wushan salamander, Liua shihi (Liu 1950) is an endemic Chinese salamander. Including type locality (Wushan County, now belongs to Chongqing Municipality), L. shihi occurs in the Wuxi and Chengkou in Chongqing Municipality, and Shenglongjia in Hubei Province. This salamander is aquatic year-round, inhabiting mountain streams with rapidly flowing water. Eggs are deposited in water and attach to surfaces under stones (Liu et al. 1960; Fei et al. 2006). Liu et al. (1960) point out that the cloaca (males have a carnose ruga at the front of the cloacal slit) and tail (males have longer and width tail) are the main sexually dimorphic characteristics, while other sexually dimorphic characteristics have not been reported.

In this paper, we explore sexual size and shape dimorphism in the Wushan salamander based on the specimens stored in the Museum of Chengdu Institute of Biology (CIB) at the Chinese Academy of Sciences. Because sexual dimorphism in salamanders can vary among populations and geographic areas (Kalezić et al. 2000; Serra-Cobo et al. 2000; Ivanović et al. 2008; Denoel et al. 2009; Ivanović & Kalezić 2012), individual specimens were collected only from the type locality in order to avoid the potential influence of geographic variation on sexual dimorphism. The purpose of this study is to describe the expression of intersexual differences in this species and interpret the results in light of existing theories, filling an important gap in the literature.

Materials and methods

A total of 116 (55 males, 61 females; Appendix I) preserved individuals of adult Liua shihi were collected from the type locality and stored in 10% formalin. These specimens from the CIB at the Chinese Academy of Sciences were measured to analyze sexual dimorphism. To standardize the shrinkage caused by preservation in formalin, all specimens had been preserved for more than 5 years (Malmgren & Thollesson 1999). To quantify intersexual differences in morphology, the following 17 variables were measured with dial calipers to the nearest 0.1 mm from the right side of each individual: snout-vent length (SVL, from the tip of the snout to the posterior margin of the cloaca); head length (HL, from the tip of the snout to the gular fold); head width (HW, width of the head at its widest point); head height (HH, height of the head at its highest point); snout length (SL, from the anterior border of the eye to the tip of the snout); tail length (TL, from the posterior margin of the cloaca to the tip of the tail); tail height (TH, the height of the tail at its highest point); tail width (TW, the width of the tail at its widest point); interorbital space (IOS, the minimum space between the upper eyelids); eye diameter (ED, the maximum diameter of the eye on the horizontal axis); internarial space (INS, the space between the nares); intercanthal space (IC, the minimum space between the anterior corners of the eyes); length of forelimb (FLL, from the base of the forelimb to the tip of the longest finger); length of hindlimb (HLL, from the base of the hindlimb to the tip of the longest toe); space between axilla and groin (AGS, the space between the posterior base of the forelimb and the anterior base of the hindlimb); forelimb width (FLW, the maximum width of the forelimb); hindlimb width (HLW, the maximum width of the hindlimb).

Because SVL is highly collinear with other variables (e.g., Romano et al. 2009), it was excluded from the subsequent analyses. The sexual dimorphism of SVL was analyzed with the independent samples t-test. All original data of the other 16 variables were log10-transformed and tested for normality (Kolmogorov-Smirnovh test) and homogeneity of variances (Levene’s test). Since variances were homogeneous, a principal component analysis (PCA) was performed to investigate sexual size and shape dimorphism.

This analysis was performed using the correlation matrix for a pooled data set. The first principal component (PC) calculated from a set of morphometric measurements is generally interpreted as an axis of body size variation when all traits load heavily and in the same direction (Reyment et al. 1984; Bookstein 1985), and the remaining variance describing relative shape differences is expressed in subsequent PCs (Schäuble 2004). Next, a univariate analysis of covariance (ANCOVA) was conducted, with sex as a factor and PC1 score as a covariate (Guillaumet et al. 2005; Romano et al. 2009) for each morphological variable independently. This allowed us to determine which variables differed between males and females. Sexual size dimorphism (SSD) was calculated using the size dimorphism index (SDI) of Lovich and Gibbons (1992), in which SSD = (size of the larger sex/size of the smaller sex) – 1, made positive when males are the larger sex and negative when females are the larger sex. Using this index, SSD = 0 when the sexes are equal in size, and SSD > 0 or SSD < 0 indicates that males or females are larger, respectively. Data analysis was carried out with SPSS software, version 17.0 (SPSS Inc., Chicago). Values are presented as mean ± standard deviation, and the significance level is set at α = 0.05.

Results

Table I presents mean values and ranges for original morphological measurements. Findings show that males had larger mean values in all original measurements. SVL ranged from 63.13 to 91.34 mm in males, and from 57.89 to 90.72 mm in females. Two-tailed t-tests clearly showed significant differences in SVL. The mean SVL was greater in males (76.46 ± 6.07, N = 55) than in females (71.18 ± 7.69, N = 61; t = −4.124, df = 112.084, P < 0.001). Thus, males are larger than females, at SDI = 0.07.

Table I. Descriptive statistics of original morphometric characters (mm) in males and females of Liua shihi.

Variables	Female (n = 61)		Male (n = 55)
	mean ± SD	range	mean ± SD	range
snout-vent length	71.18 ± 7.69	57.89–90.72	76.46 ± 6.07	63.13–91.34
head length	21.14 ± 2.00	17.04–27.81	22.75 ± 1.97	18.34–28.30
head width	14.69 ± 1.38	11.20–19.70	16.10 ± 1.60	13.10–20.20
head height	8.89 ± 1.24	5.90–11.50	10.32 ± 1.83	7.20–18.30
snout length	4.63 ± 0.88	3.54–10.21	4.91 ± 0.50	3.83–6.42
tail length	66.42 ± 7.97	49.87–87.30	74.40 ± 8.01	59.54–94.96
tail height	11.04 ± 1.25	8.86–14.27	12.06 ± 1.64	9.02–16.30
tail width	8.37 ± 1.20	6.64–11.10	9.60 ± 1.58	7.37–13.41
interorbital space	3.43 ± 0.42	2.69–4.51	3.78 ± 0.55	2.78–5.27
eye diameter	4.26 ± 0.32	3.35–5.35	4.47 ± 0.28	3.92–5.54
internarial space	5.57 ± 0.52	4.35–6.87	5.94 ± 0.64	3.81–7.57
intercanthal space	5.60 ± 0.47	4.86–6.70	6.04 ± 0.57	4.90–7.48
length of forelimb	19.07 ± 1.43	15.81–23.07	21.04 ± 1.26	18.51–23.34
length of hindlimb	21.79 ± 1.78	17.90–26.30	23.45 ± 1.74	19.90–28.00
space between axilla and groin	33.32 ± 4.07	25.86–44.38	34.62 ± 3.67	26.80–44.53
forelimb width	3.68 ± 0.49	2.90–5.75	4.51 ± 0.58	3.34–6.18
hindlimb width	5.20 ± 0.71	3.85–8.43	6.25 ± 0.81	4.57–8.10

The first principal component (PC1) explained the largest proportion of overall variation (63.31%). All original variables loaded heavily and in the same direction (positively, r > 0.60 for all variables) onto this component (Table II). Therefore, the individual scores on PC1 were used to estimate the differences in overall body size. Other PCs explained 36.39% of the variance, and factor scores for these components were retained as the variable body shape. PC1 differed significantly between females and males [analysis of variance (ANOVA), F_{1, 114} = 37.864, P < 0.001], and females had lower scores than males.

Table II. Factor loadings for the principal components (PC; eigenvectors), eigenvalues and proportion of total variance described by the first two components obtained from PCA on a correlation matrix. Results of analysis of covariance (ANCOVA) with PC1 scores as covariate tests for differences in morphological variables. All variables are log-transformed; df = 113.

Variables	PC1	PC2	F	P-value
head length	0.926	0.027	7.100	0.009
head width	0.744	−0.170	4.401	0.038
head height	0.810	−0.267	0.273	0.602
snout length	0.658	0.062	1.696	0.195
tail length	0.925	−0.087	0.021	0.886
tail height	0.701	−0.428	0.114	0.736
tail width	0.879	−0.361	0.491	0.485
interorbital space	0.793	0.063	1.401	0.239
eye diameter	0.608	0.449	0.173	0.678
internarial space	0.739	0.184	1.895	0.171
intercanthal space	0.864	0.099	0.989	0.322
length of forelimb	0.827	0.261	16.479	0.000
length of hindlimb	0.757	0.397	1.004	0.319
space between axilla and groin	0.736	−0.031	13.516	0.000
forelimb width	0.862	−0.081	29.004	0.000
hindlimb width	0.824	0.008	14.734	0.000
Eigenvalue	10.130	0.909
Proportion	63.310	5.685

Differences in shape between the sexes were further revealed by the ANCOVA. The interaction term was nonsignificant (P > 0.5) and could thus be removed from the model for all variables. A univariate ANCOVA revealed significant differences in body shape in six morphological variables (Table II): HL, HW, FLL, AGS, FLW and HLW. Females have larger values for HL, HW and AGS, as well as smaller values for FLL, FLW and HLW.

Discussion

Our study demonstrates that sexual dimorphism occurs not only in body size in this species (males larger than females, as shown in Table I), but also in body shape (females had proportionally longer HL, HW and AGS, while males had longer FLL, FLW and HLW, as shown in Table II). In contrast to findings by Liu et al. (1960) that sexual dimorphism is present in the tail of this species, our findings did not support this notion. This may be because Liu et al. (1960) did not exclude the influence of body size and used results from original data. Since little research has been conducted on this species, our findings build a foundation for future research as salamanders become increasingly threatened throughout the world.

SSD, the most conspicuous difference between sexes in many species (see e.g., Shine 1979; Andersson 1994; Fairbairn 1997; Malmgren & Thollesson 1999; Ivanović et al. 2008; Romano et al. 2009; Rastegar-Pouyani et al. 2013), is a difference in the average body size of males and females (Bakkegard & Rhea 2012). There are three patterns of SSD in mature amphibians according to body size: (1) females larger than males: female-biased SSD; (2) males larger than females: male-biased SSD; (3) females equal to males: unbiased SSD (Shine 1979; Bruce 1993, 2000).

In salamander species, the first pattern in SSD is the most common (e.g., Shine 1979; Salvidio & Bruce 2006; Ivanović et al. 2008; Marvin 2009; Romano et al. 2009; Seglie et al. 2010), with a few species exhibiting the second pattern, e.g., Euprocuts platycephalus (Bovero et al. 2003) and Phaeognathus hubrichti (Bakkegard & Guyer 2004). Body size is influenced by sexual selection, life history and ecological differences between sexes (Bakkegard & Guyer 2004). Male competition (aggressive behavior and male fighting) is the major factor in male body size evolution and consequent SSD (Shine 1979), because larger males have higher fighting ability and mating success (Shine 1979; Howard 1981; Raxworthy 1989; Bruce 1993; Andersson 1994; Bakkegard & Guyer 2004).

Seglie et al. (2010) point out that the intensity of selective pressure on male size mainly depends on the size advantage gained in intrasexual competition. Thus, sexual dimorphism in the body size of L. shihi may be attributed to sexual selection (male competition). However, there are no reports on the social behavior and reproductive biology of L. shihi. Shine (1979) reported that 86.7% of the urodele species, in which males equal or exceed females in size, engage in male combat.

Therefore, we speculate that the behavior of male combat is present in males of the L. shihi species. However, this conjecture requires validation in future reproductive biology research. L. shihi inhabits mountain streams (altitude 900–2350 m), and the ecological conditions fit a moderate mountain climate. The sexual dimorphism of SSD is consistent with findings by Serra-Cobo et al. (2000) that male Euproctus asper is larger than female in moderate mountain climates.

Head size (HL and HW) is relatively greater in females than males of Liua shihi. In many salamander species, males have a larger head size than females (e.g., E. platycephalus, Bovero et al. 2003; Ph. hubrichti, Bakkegard & Guyer 2004; Amphiuma tridactylum, Fontenot & Seigel 2008; Plethodon kentucki, Marvin 2009; Salamandrella keyserlingii, Hasumi 2010). However, in a few species of salamander, females have greater head size (e.g., Neurergus microspilotus, Rastegar-Pouyani et al. 2013).

Sexual dimorphism in head size between sexes may result from sexual selection, reproductive roles and ecology. The first of these alternatives is characterized by better fighting ability during male-male competition for access to females and mating opportunities (Fauth & Resetarits 1999; Hasumi 2001; Bovero et al. 2003; Bakkegard & Guyer 2004; Marvin 2009). The second is reflected in the differences in reproductive investment between sexes. For example, females are responsible for a larger reproductive investment (Duellman & Trueb 1994). Therefore, a larger head should maximize energy intake (Selander 1972; Shine 1979, 1989; Malmgren & Thollesson 1999). Romano et al. (2012) reported that the diet of male and female Triturus carnifex was similar, but found that the number of taxa in the female stomachs was twice that of the males. The third alternative is reflected in different feeding strategies. Individuals of both sexes have adapted to different resource utilization patterns through niche partitioning (Malmgren & Thollesson 1999). These result in differences of diet and prey size (e.g., Godley 1983; Cooper & Vitt 1989; Fauth & Resetarits 1999; Bovero et al. 2003; Fontenot & Seigel 2008; Seglie et al. 2010). Different diets and prey sizes between males and females reduce intersexual food competition (Serra-Cobo et al. 2000). Thus, the sexual dimorphism in head sizes of L. shihi may be due to their reproductive roles and ecology.

Our results show that trunk length (AGS) is greater for females than males of L. shihi. Greater elongation of females than males has been documented in several salamanders (e.g., Batrachoseps attenuatus, Jockusch 1997; Pl. kentucki, Marvin 2009; Salamandrina perspicillata, Romano et al. 2009). In most salamander species, sexual dimorphism of AGS may be related to the fecundity advantage (Romano et al. 2009), because females with larger AGS have a larger abdominal volume, and can produce larger clutches and/or a larger number of eggs (Shine 1988) and thus have an increased reproductive capacity (Griffith 1990; Jockusch 1997; Marvin 2009).

Our results also show that limb size (FLL, FLW and HLW) of L. shihi shows obvious sexual dimorphism, with males having larger limbs. This difference in limb size is also found in other salamanders, e.g., T. montandoni (Dandová et al. 1998); T. cristatus and T. vulgaris (Malmgren & Thollesson 1999); Tylototriton verrucosus (Serra-Cobo et al. 2000); E. asper (Seglie et al. 2010). The fertilization mode of these salamanders is internal (Duellman & Trueb 1994). Longer and robust limbs in males may be beneficial for courtship performance (Rehak 1983; Malmgren & Thollesson 1999). This may influence reproductive success (Lee & Price 2001) because males with long forelimbs are better equipped to retain their grip on females during amplexus and to resist attempted take-over by competing males (Howard & Kluge 1985). However, hynobiids as well as cryptobranchids and sirenids have external fertilization, and there is no known courtship (Duellman & Trueb 1994). Thorn (1963) reported that the male of Hynobius nebulosus grasps the egg sac in the fore- and hindlimb and crawls along the sac. Therefore, sexual dimorphism in limb size of L. shihi may be attributed to reproductive success.

Conclusions

This study examined morphometric data of Liua shihi to determine sexual size and shape dimorphism. The results show that sexual dimorphism in this species includes not only body size, but also shape. The sexual dimorphism of body size can be attributed to sexual selection and local climates, that of AGS may contribute to fecundity selection, that of head size may be attributed to reproductive roles and ecology, and that of limb size may be beneficial for reproductive success. Because few studies have explored this topic in this species, this study lays the groundwork for future studies. Further research may explain the cause of the sexual dimorphism in body size and body shape exhibited by the Wushan salamander, which may be linked to sexual selection, ecological differences and life history adaptations.

Acknowledgements

We thank Yuezhao Wang, Xiaomao Zeng, Jiatang Li, Shengqun Li and Yueying Cheng of the Chengdu Institute of Biology (CIB), Chinese Academy of Sciences, for assistance in examining and measuring specimens. This work was supported by grants from the National Natural Science Foundation of China (No. NSFC 30900138), and the Science and Technology Program of Henan Province (No. 122102110035). We also thank two anonymous reviewers for providing helpful comments that improved this paper.