Sand lizards Lacerta agilis with higher digit ratios are more likely to autotomy

Abstract

Digit ratio is a morphological feature regarded as a biomarker of the balance of sex hormones during early development. The exposure of embryos to a set of sex hormones and the mutual relations between those hormones cause the emergence of individual morphological and/or behavioural characteristics as well as differences between sexes. We have thus hypothesised that differences in one of these morphological traits—digit ratio—may be a proxy representing a tendency towards tail autotomy. The aim of this study is to investigate the digit ratio (2D:3D, 2D:4D, 3D:4D) of the sand lizard, Lacerta agilis, Lacertidae, a species characterised by well‐developed sexual dimorphism, whereby females are larger than males. We also tested associations between patterns in digit ratio and caudal autotomy, a common defensive mechanism among lizards. To our knowledge, the relationship between a tendency towards autotomy and digit ratio pattern has never been researched. To date, studies on autotomy have mainly focused on the consequences, costs or evolutionary background of tail loss. Hence, researchers examined mostly the frequency of autotomy in the context of predatory pressure or habitat conditions, omitting an individual's behavioural tendency to shed its tail. However, behavioural traits can affect an individual's exposure to predator attack and consequently the need to use an anti‐predator strategy. Thus, following this logic, dropping the tail may be the result of the lizard's intraspecific personality characteristics, resulting from the effect of hormones on behaviour or innate traits. Therefore, we suggest that the inclusion of autotomy as a factor explaining observed digit ratio patterns and their variability between taxa has great potential. We used computerised measurements of photographed limbs to determine the length of digits. We found that the digit ratios for all four limbs were significantly lower in females than in males, excluding the 3D:4D ratio for the right hindlimbs. Therefore, the results confirmed the pattern already observed for most lizards. The novel element in our study is the detection of the relationship between a tendency towards caudal autotomy and digit ratio. Individuals with a tendency towards autotomy have a higher 2D:4D ratio in the right forelimbs and a lower 2D:3D ratio in the right hindlimbs. Obtained results suggest that these morphological characteristics are most likely related to intraspecific differences (between bold and shy individuals) which consequently may determine an individual's reaction or susceptibility to be a prey and escape behaviour. Thus, our results are probably the first attempt to link digit ratio to the susceptibility of lizards to tail autotomy.

In this study, we conducted a detailed analysis of the digit ratio patterns (2D:3D, 2D:4D and 3D:4D) in Lacerta agilis, Lacertidae and Squamata. Obtained results indicated a clear dimorphic pattern for most digit ratios, whereby the digit ratios of all four limbs were significantly lower in females than in males. Furthermore, a clear relationship between caudal autotomy and right forelimb 2D:4D ratio has been shown, such that individuals with autotomy are characterized by higher 2D:4D ratios.

1. INTRODUCTION

Digit ratio, defined as the ratio of the lengths of two different digits, is a morphological feature considered as a biomarker of the balance of sex hormones during early development. Some studies have already proven that digit ratios are controlled by the balance of androgen to oestrogen signalling during digit development (Zheng and Cohn, 2011). According to Manning (2002), the exposure of embryos to a set of sex hormones and the ratio between those hormones' levels may cause the emergence of individual morphological and/or behavioural characteristics as well as differences between sex. Consequently, the digit ratio, recognised as an indicator of behavioural and/or fitness‐related traits, is attracting the attention of many researchers working in different fields of biology, where humans are the main object of research (Voracek and Loibl, 2009). Although the majority of research on the digit ratio in humans are aimed at finding some correlation, most of published results led to the assumptions about the linkage between the digit ratio and prenatal exposure to sex hormones (Voracek and Loibl, 2009). What is important, experimental approaches have already been successfully incorporated in some studies looking for the pattern explaining the observed variation in the digit ratio, thus offering much greater opportunities to determine the mechanisms of developing this feature than simple correlation studies. Those experimental studies are only possible for non‐human vertebrates, that is why this type of research is very important (Zheng and Cohn, 2011; Nagy et al., 2016; Lofeu et al., 2017). However, it is still not very clear, if digit ratio is really conserved in all Tetrapoda and there are several studies which did not confirm the presence of this phenomenon in some groups/species, for examples in mammals (Yan et al., 2008; Lilley et al., 2009; Nelson and Shultz, 2010; but see: Brown et al., 2002; Leoni et al., 2005; McIntyre et al., 2009; Fuse and Sawada, 2019). To date, digit ratios are determined in an early stage of prenatal development (narrow window) exactly when digits grow and when sex steroids masculinize or feminize the brain (Zheng and Cohn, 2011).

In comparison with studies on humans, significantly fewer published studies concern the examination of other vertebrates in either laboratory or field conditions (Lombardo et al., 2008; Nelson and Shultz, 2010; Gomes and Kohlsdorf, 2011; Kaczmarski et al., 2015; Kazimirski et al., 2020). Digit ratio studies initially concerned the second and fourth digits, that is 2D:4D (Voracek and Loibl, 2009); subsequently, in further studies of some amphibians, sauropsids (birds and reptiles) or mammals, other ratios, such as 2D:3D (Romano et al., 2005; Leoni et al., 2008; Woodhead et al., 2018), 2D:5D (Leoni et al., 2008), 3D:4D (Tobler et al., 2011; Tobler et al., 2012), or all three variants (Rubolini et al., 2006; Saino et al., 2007; Ruuskanen et al., 2011; Balogová et al., 2015; Van Damme et al., 2015; Kazimirski et al., 2020) were also tested. Generally, females tend to have higher 2D:4D digit ratios than males in most mammals (Brown et al., 2002; Leoni et al., 2005; McIntyre et al., 2009; Fuse and Sawada, 2019) as well as in Caudata/tailed amphibians (Kaczmarski et al., 2015). On the other hand, investigated species of Anura amphibians, reptiles and birds are mostly characterised by the opposite pattern (Burley and Foster, 2004; Chang, 2008; Leoni et al., 2008; Van Damme et al., 2015). However, among species for which the digit ratio pattern has been already described, the results are not consistent and fail to present a common picture. Presumably, this is due to the factors such as differences between measuring methods, genetic differences between populations in response to prenatal exposure to hormones (Lombardo and Thorpe, 2008) or not big enough sample size used in the study. In addition, in sexually monomorphic species sexual dimorphism in digit ratio does not usually occur (Gooderham and Schulte‐Hostedde, 2012; Balogová et al., 2015, Lofeu et al., 2017, Kazimirski et al., 2020; but see Leoni et al., 2008), which makes it possible to explain, at least partly, that there is an inconsistency in results obtained for the Tetrapod so far. This is also in line with the conclusion of Rubolini et al. (2006) that interspecific variation of sex differences in digit ratios could be associated with sex‐specific growth trajectories. Nevertheless, general digit ratio pattern within Tetrapod is still not very clear and conclusions of some previous studies seem to be very variable and sometimes even controversial.

In the first studies of lizards, variation in the digit ratio was investigated mostly in the context of sex differences (Chang et al., 2006; Rubolini et al., 2006). To date, patterns observed in lizards range from no sex differences in 2D:4D (Anolis carolinensis: Lombardo and Thorpe, 2008; Gonatodes albogularis: Peñuela, 2011) and differences in ratio in some single limbs only (Anolis carolinensis: Chang et al., 2006) to sex differences in both fore‐ and hindlimbs in Podarcis melisellensis (Van Damme et al., 2015). This variability seems to be associated with the effects of prenatal steroid's exposure on digit ratios which may differ between fore‐ and hindlimbs due to fingers and toes development at different times and therefore under various hormonal milieus (Van Damme et al., 2015). Gomes and Kohlsdorf (2011) made a significant contribution to the identification of the digit ratio (2D:4D) pattern by examining 25 species of Iguania lizards and presenting results on the phylogenetic background of particular species in the context of their use of microhabitats. Generally, in Iguania, males exhibited higher digit ratios than females in fore‐ and hindlimbs. Moreover, arboreal species exhibit more readily distinguishable sex differences in digit ratio due to females' tendency towards a low ratio in forelimbs (Gomes and Kohlsdorf, 2011). However, the dimorphism pattern in digit ratios in Squamata is not as unambiguous as initially assumed. Pronounced sexual dimorphism was detected in all limbs in 2D:3D and 2D:4D ratios in P.melisellensis, but only in both hindlimbs in 2D:4D ratio in P.siculus (Van Damme et al., 2015). In the case of 2D:3D, detected differences between sexes were characterised by a high degree of variability among species. For example, in adult Otago/Southland geckoes Woodworthia, significant sexual dimorphism in 2D:3D (with this ratio being higher in males) was found only in the right hindlimb (Woodhead et al., 2018), in Podarcis siculus only in the left hindlimb (Van Damme et al., 2015). In Podarcis muralis male had larger 2D:3D ratio on the left side than females, whereas in Mabuya planifrons males had lower 2D:3D ratios than females (Rubolini et al., 2006). In painted dragon lizards (Ctenophorus pictus), the 3D:4D ratio was sexually dimorphic in both hindlimbs, but the pattern was reversed, that is females were characterised by higher digit ratios than males (Tobler et al., 2011).

The experimental study on painted dragon lizards involving the addition of testosterone directly to eggs did not result in the predicted masculinisation effect, but in turn changed the digit ratio (3D:4D) to the opposite direction. The authors speculate that this may have been the result of the aromatisation of testosterone into oestradiol; however, this does not change the fact that a certain level of testosterone in eggs affecting the digit ratio is therefore a feature that can be used as a phenotypic indicator for prenatal steroid exposure (Tobler et al., 2011). The experiment conducted by Braña (2008) in oviparous population of the common lizards Zootoca vivipara, indirectly corroborates the impact of exogenous hormones during embryonic development on differences in dimorphism, including the length of 4D digit in the right hindlimb in females. In this study, female hatchlings tended to have longer digits when incubated in eggs adjacent to the eggs with male siblings. Hence, even a small amount of steroid diffusing from adjacent eggs from the opposite sex embryo have developmental effects not only on 4D digits length but also on body mass. Such transfer caused masculinisation, increase in digit length in females and thus affected the innate digit ratio pattern. Hence, more studies examining such relationships are needed with especially a broad taxonomic approach required (Tobler et al., 2011; Tobler et al., 2012).

The relationship between digit ratios and phenotypic traits has rarely been the subject of research on non‐human vertebrates (see Table 1 in Tobler et al., 2012, for details). What is more, the evidence indicating such relationships is inconsistent. For instance, in lizards (C.pictus) the 3D:4D ratio is associated with the presence of the gular bib (granular fold on the ventral throat). Males with bibs are characterised by higher digit ratios than those without it; it emerged that this trait was also related to greater reproductive success (Tobler et al., 2011). In the same species, males with different head colour morphs exhibited differences in digit ratios: males that did not express head colour had lower 3D:4D ratios than two other morphs, but these differences were not significant, possibly due to insufficient sample size (Tobler et al., 2011). Among yellow morph specimens of P.melisellensis, greater sexual dimorphism in average digit length ratio in hindlimbs was proved in comparison with two other morphs (Van Damme et al., 2015). Generally, exposure to steroid hormones during early development is correlated with the colour morph expression in adult lizards (Hews et al., 1994), but may also change other characteristics, including some behavioural or performance traits. In the context of the latter, human individuals with lower 2D:4D ratios have a predisposition to achieve better results in endurance‐related sports (Manning and Hill, 2009). The relationship between digit ratio and physical performance has already been a subject of several research projects on vertebrates; for example, in the case of Squamata some attempts have been made to examine the relationship between digit ratio and sprint speed in Podarcis lizards as well as in Woodworthia gecko, but so far, no clear correlation was found (Van Damme et al., 2015; Woodhead et al., 2018). Furthermore, no association between digit ratio and endurance was detected in C.pictus, although individuals with higher 3D:4D ratios are characterised by shorter escape response times. Thus, such individuals can be considered less inclined to ‘take risks’ or just more alert. Furthermore, males with a gular bib exhibited shorter escape response times and lower endurance than those without this feature. This was the first observed anti‐predator behaviour associated with the phenotypic trait correlated with the digit ratio in lizards (Tobler et al., 2012). Referring to Homo sapiens, individuals with lower digit ratios are more likely to exhibit risky behaviour (Ronay and von Hippel, 2010; Garbarino et al., 2011). Taking all these examples into consideration, we believe that it is worth developing research on digit ratios in the context of other anti‐predator‐related traits and their consequences on survival ability in lizards.

Caudal autotomy is a common anti‐predator behaviour considered to be conditioned by a set of environmental features as well as individual and species‐specific traits (Formanowicz et al., 1988; Zani, 1996; Clause and Capaldi, 2006; Pafilis et al., 2017). Numerous factors have been reported to affect the use of this defensive mechanism among lizards, including predatory pressure, age (adults vs juveniles), temperature and intraspecific competition (Pafilis et al., 2009; Fleming et al., 2013). However, these papers mainly concern the direct consequences of autotomy for specimens. Apart from the obvious costs of regeneration of the lost body part, numerous changes have been described in the behaviour of specimens after dropping their tails (Formanowicz et al., 1988; Zani, 1996; Clause and Capaldi, 2006; Pafilis et al., 2017). More precisely, tail loss affects endurance and changes locomotor abilities (locomotor capability or sprint speed) (Formanowicz et al., 1988; Ekner‐Grzyb et al., 2013; McElroy and Bergmann, 2013). A recent study shows the lack of sexual differences in the incidence of autotomy in Lacertidae lizards (Pafilis et al., 2017), although males seem more vulnerable to tail loss due to different life strategies, for example active searches for a partner or male‐male combat during mating season. Most researchers agree that this feature is context‐related, depending mainly on predator pressure and habitat traits, but it is important to consider also innate disposition towards autotomy (Hare and Miller, 2010). However, the innate tendency to use this strategy may have an immeasurable effect on its frequency in the population. Behavioural traits can affect an individual's exposure to a predator attack and consequently the need to use an anti‐predator strategy. In the Namibian rock agama (Agama planiceps), individual behaviour traits were significantly consistent during life, with bolder males suffering more frequent incidences of tail loss than shyer males, probably due to more frequent predation (Carter et al., 2010). Van Damme et al. (2015) hypothesised that differences in digit ratio between two species of lizards can be considered at the level of their ‘temperament’. They expected ‘shyer’ P.melisellensis specimens to have higher digit ratios than those of ‘bolder’ P.siculus. Hence, probably the variation in the risk‐taking tendency and/or temperament of individuals is the result of hormonal activities at the embryonic stage, which may be reflected in the digit ratio, which thus could serve as a biomarker. Therefore, following this logic, discarding the tail may be an effect of the intraspecific personality characteristics of the lizard, resulting from the effect of hormones on behaviour or innate traits. In our opinion, the inclusion of autotomy as a factor explaining the observed digit ratio patterns and their variability between taxa has great potential.

According to most researchers investigating digit ratio, further studies on nonhuman vertebrates are necessary. Therefore, the aim of this study was to investigate digit ratios (2D:3D, 2D:4D, 3D:4D) in the sand lizard, Lacerta agilis, Lacertidae, a species with well‐developed sexual dimorphism, whereby females are larger than males. We also tested our results for a potential relationship between any digit ratio pattern and caudal autotomy occurrence, as a feature partially related to the tendency to ‘risk taking’ or simply related to the level of lizards' ‘temperament’. However, if only present, we made no specific predictions about the direction of such relationship between digit ratio and autotomy.

2. METHODS

2.1. Ethics statement

This research complies with the current laws of Poland and was performed with appropriate collection and research permits (under permit from the Regional Director of Environmental Protection, WPN‐II.6401.139.2018.KL and WPN‐II.6401.142.2018.AG). We followed all applicable institutional and national guidelines for the care and use of animals. Moreover, the main investigator (MK) was trained by the Polish Laboratory Animal Science Association. All animals were released at the site of capture in accordance with the permission granted.

2.2. Study sites and research object

Fieldwork was conducted from April to May 2018 and the time of capture sessions was adapted to the lizards' activity. We analysed a total of 137 sand lizard Lacerta agilis (Linnaeus, 1758) (61 females, 76 males) collected from three separate sites in eastern Poland, at two sites located within the city of Poznań (Morasko – site 1: 52.46°N, 16.92°E, Lutycka – site 2: 52.44°N, 16.86°E;) and one in Potrzebowice forest (site 3: 52.84°N, 16.22°E). At site 3, lizards were much less numerous and much more scattered—only one of 18 captured individuals was female. Hence, only one individual of this sex was included in the further analyses. In Table 1, the number of individuals caught in the other two locations is presented separately for both sexes.

TABLE 1.

Frequency of autotomy in three examined sand lizard (Lacerta agilis) populations (sorted by sex and site)

Sex/Site	Sample size	Individuals with autotomy	Frequency of autotomy (%)
Males, total	76	44	57.9
Site 1	29	21	72.4
Site 2	30	18	60.0
Site 3	17	5	29.4
Females, total	61	32	52.5
Site 1	30	17	56.7
Site 2	30	15	50.0
Site 3	1	–	–

The animals were captured manually or with the use of special herpetology nets. In L.agilis, strong female‐biased sexual dimorphism is clearly expressed in size (but see Majláth et al., 1997), with the exception of head size (Ekner et al., 2008). Females have less bulky heads than males. Determination of sex in juveniles or sub‐adults is possible if the number of ventral scales' rows and the shape of the anal plate exhibit pronounced sexual dimorphism (Eplanova and Roitberg, 2015). It was decided to measure adult specimens only. In order to avoid repeated measurement of recaptured individuals, the lizards were marked by cauterisation on ventral scales (Ekner et al., 2011). Five characteristics were measure, using a calliper (accuracy ‐ 0.01 mm): tail length (TL), SVL (total body length without tail—the measure taken from the tip of an animal's nose to the cloaca opening at the tail base), head width (HW), head length (HL) and head height (HH). To minimise additional human‐induced systematic error (experimenter effect), all specimens were measured by only one researcher. Following measurements, lizards were released into the environment at the same location where they were captured.

Tail status was also noted in both recent and previous life stages, regardless of whether individuals had already regenerated or freshly discarded their tails, and these traits were coded as follows: (b) no autotomy—intact tail, or (b) visible tail autotomy/autotomised tail (individuals who had discarded their tails). Once lost, the tail never returns to its original form. It is worth mentioning that tail discards are likely be repeated throughout the life of a specimen (Hare and Miller, 2010), and therefore we usually do not know how many times an individual has lost its tail.

2.3. Digit ratio measurements

The digital measurements were done according to Kaczmarski et al. (2015), with some modifications in the protocol as described below. In the investigated species, the basal joint of the finger is clearly visible, similarly to the previously studied species, for example the common wall lizard Podarcis muralis (Rubolini et al., 2006). Therefore, a proximal landmark was marked as close as possible to the basal joint of the finger and a measurement was taken from the basal crease to the tip of the digit. Thus, according to the modified protocol, we performed one measurement for each digit. Digit lengths were calculated using digital images obtained with a Samsung Galaxy J3 (6) smartphone (8‐megapixel primary camera) as source material. A special measurement platform was created in order to take all photographs from the same distance, depicting the same position of the digits. The platform consisted of a wooden frame with two glass surfaces placed at a distance of 6 cm from each other. Each time, the limb and camera were installed in the frame; the limb was placed on the upper surface (lined with millimetre paper serving as a scale) and the camera on the lower surface (the opposite side). Separate pictures were taken for each fore‐ and hindlimb. We adopted digits numbering from previous publications (Chang et al., 2006). In order to calculate intra‐observer error, the digit measurements were carried out twice for 3‐ randomly selected individuals. The level of technical measurement error was calculated using intraclass correlation coefficient (ICC).

2.4. Data processing and statistical analysis

In the following analyses, we used a general linear mixed model with normal error distribution and identity link function. Since the lizards were captured and measured at three different locations, the study site was treated as a random effect (random intercept). The size of lizards (PC1), autotomy (IT—intact tail or AT—autotomised tail) and sex (F, M) were tested here. We performed modelling on the following response variables: forelimb left and right 2D:4D, hindlimb left and right 2D:4D, forelimb left and right 2D:3D, hindlimb left and right 2D:3D, forelimb left and right 3D:4D, hindlimb left and right 3D:4D. In accordance with the above approach, we have also tested the length of individual digits (second ‐ 2D, third ‐ 3D and fourth ‐ 4D).

We compared each digit ratio (2D:4D, 2D:3D, 3D:4D) of fore‐ and hindlimbs between the right and left side using the t test for dependent variables. For each response variable, we started with the global model with all explanatory variables and then dropped all non‐significant effects (p > 0.05). The sexual dimorphism index (SDI) for each digit's length and the digit ratio for each limb was calculated based on the approach adopted by Gomes and Kohlsdorf, 2011. All statistical tests were performed in R software (R Core Team, 2014, 2014) with the ‘lme4’ package (Bates, 2010).

3. RESULTS

Seeing that we found strong positive correlations between each of the following variables: SVL, head length, head width and head height, we then used principal component analysis (PCA) to reduce the number of variables to one of these four size characteristics. The first principal component (PC1) was positively correlated with SVL (r = 0.52, p < 0.001), head length (r = 0.93, p < 0.001), head width (r = 0.91, p < 0.001) and head height (r = 0.82, p < 0.001). Therefore, this variable (PC1) was used as the indicator of the lizard's size (the ‘size variable’), as it explained 66% of detected variance. All obtained values of ICC were significant and acceptable (ICC = 0.971, p < 0.05).

Sixty‐one captured individuals had intact tails (44.53%), while in the other 76 cases the tails had been autotomised (55.47%) (Table 1).

Based on the general linear mixed model, we found that the sex of investigated lizards' individuals was the only significant variable, differentiating 2D:4D of the left forelimb, 2D:4D of both hindlimbs, 2D:3D of both forelimbs, 2D:3D of the left hindlimb and 3D:4D of both forelimbs, with lower digit ratios in females (Tables 2, 3, Figure 1, Table S1); tail autotomy and the chosen size variable (PC1) were both insignificant and were thus dropped from the final model. Generally, the size variable (PC1), being insignificant, was dropped from all final models, with the exception of 3D:4D of left and right hindlimbs (see below).

TABLE 2.

Summary of general linear mixed model results with all non‐significant effects dropped, along with t tests for dependent variables to compare each digit ratio (2D:4D, 2D:3D, 3D:4D) of fore‐ and hindlimbs between right and left sides in sand lizards (Lacerta agilis)

Digit Ratio	Limb	Side	Sex	Autotomy	PC1 (size)	T test
						t	p
2D:4D	Fore	L	***	–	–	−0.93	0.352
		R	***	*	–
2D:4D	Hind	L	***	–	–	−1.45	0.150
		R	***	–	–
2D:3D	Fore	L	***	–	–	0.40	0.692
		R	***	–	–
2D:3D	Hind	L	***	–	–	0.02	0.988
		R	***	*	–
3D:4D	Fore	L	*	–	–	−1.39	0.165
		R	*	–	–
3D:4D	Hind	L	*	–	**	−2.45	0.016
		R	–	(*)	(*)

Levels of significance: ***p < 0.001, **p < 0.01, *p < 0.05, *, marginally non‐significant, –, no effect.

TABLE 3.

Comparison of 2D:3D, 2D:4D, and 3D:4D digit ratios between adult male and female sand lizards (Lacerta agilis)

Digit Ratio	Limb	Side	N	Males	N	Females	SDI
				Mean ± SE		Mean ± SE
2D:4D	Fore	L	75	0.844 ± 0.062	61	0.773 ± 0.053	1.092
		R	75	0.855 ± 0.072	61	0.771 ± 0.048	1.110
2D:4D	Hind	L	75	0.565 ± 0.055	60	0.525 ± 0.369	1.076
		R	76	0.564 ± 0.043	60	0.543 ± 0.387	1.039
2D:3D	Fore	L	76	0.839 ± 0.049	61	0.783 ± 0.043	1.072
		R	76	0.840 ± 0.055	61	0.777 ± 0.040	1.081
2D:3D	Hind	L	75	0.726 ± 0.058	59	0.684 ± 0.038	1.061
		R	75	0.722 ± 0.051	58	0.690 ± 0.042	1.046
3D:4D	Fore	L	75	1.009 ± 0.063	61	0.988 ± 0.052	1.021
		R	75	1.019 ± 0.070	60	0.996 ± 0.055	1.023
3D:4D	Hind	L	75	0.777 ± 0.043	58	0.768 ± 0.033	1.012
		R	75	0.781 ± 0.039	59	0.785 ± 0.034	0.995
	SVL		76	65.738 ± 5.427	61	72.414 ± 7.276	0.998

The SDI values higher than 1 and lower than 0 indicate a male‐ and female‐biased sexual dimorphism, respectively.

Abbreviation: SDI, sexual dimorphism index.

FIGURE 1.

Mean (±SE) 2D:4D, 2D:3D, 3D:4D digit ratios of adult male (dark grey) and female (light grey boxes) sand lizards (Lacerta agilis). Asterisks indicate the statistical significance level of the difference in digit ratios between sexes: ***p < 0.001, *p < 0.05. FL = left forelimb, FR = right forelimb, BL = left hindlimb, BR = right hindlimb

In the right forelimb, sex was similarly significant variable, affecting 2D:4D, which was also lower in females (Tables 2, 3, Figure 1, Table S1); however, in this case, autotomy status turned out to be significant as well—the 2D:4D digit ratio of individuals with autotomy was higher (Table S1, Figure 2a,b).

FIGURE 2.

(a) Values of right forelimbs (FR) 2D:4D ratio showed the differences within tail status: IT—‘intact tail’ (black box) and AT—‘autotomised tail’ (white box). (b) Mean (±SE) 2D:4D digit ratio of right forelimbs (FR) of adult male (dark gray box) and female (light gray box) sand lizards (Lacerta agilis). (c) Values of right hindlimbs (BR) 2D:3D ratio showed the differences within tail status. (d) Mean (±SE) 2D:3D digit ratio in right hindlimbs (BR). Asterisks indicate the statistical significance level of the difference in digit ratios between tail status (p < 0.05)

In the right hindlimb, we found a significant difference in 2D:3D associated with the sex—the digit ratio in females was lower (Tables 2, 3, Figure 1, Table S1). In terms of autotomy status difference, the 2D:3D digit ratio of individuals with autotomy was lower (Table S1, Figure 2c,d). Surprisingly, this pattern is opposite to that of the right forelimb 2D:4D.

In the left hindlimb, we found a significant difference in 3D:4D ratio related to the sex, with the digit ratio in females being lower again (Tables 2, 3, Figure 1, Table S1). It is also noteworthy that we detected a negative relationship with lizard size (PC1; Table S1, Figure 3a). Tail autotomy, since it appeared to be insignificant, was dropped from the final model. In the right hindlimb, 3D:4D digit ratio we found some autotomy status differences: albeit marginally non‐significant, the 3D:4D digit ratio of individuals with autotomy was higher (Table S1). Similarly, we found a marginally non‐significant negative relationship with lizard size (PC1; Table S1, Figure 3b).

FIGURE 3.

Results of the predictions of the two models fitted to: (a) the relationship between the size of an individual (first principal component variable, PC1) and left hindlimb 3D:4D digit ratio. (b) the relationship between the size of an individual (PC1) and right hindlimb 3D:4D digit ratio. Dashed lines indicate 95% confidence intervals

In general, males are characterised by higher digit ratio values than females, except for 3D:4D in the right hindlimb (Table 3, Figure 1). Pearson's correlations between different digit ratios are shown in Figure S1 for males and Figure S2 for females. Excluding 3D:4D of the hindlimb, we found no significant differences between body sides and any of the three studied digit ratios (Table 2).

Based on the general linear mixed model, we found that only the size (PC1) of investigated lizards' individuals was a significant variable, differentiating digits lengths (Table S2, S3). In the hindlimbs, we found significant differences in 2D between the body site associated with the sex—females exhibited shorter digits than male. Opposite tendency occurred in the forelimb 3D, 4D of both body sites and in the left hindlimb 4D with female‐biased pattern (Table S2, S3). Interestingly, we found an autotomy status difference only in the relation to the left forelimb 3D lengths.

4. DISCUSSION

In this study, we have conducted a detailed analysis of the digit ratio patterns in L.agilis. The main question addressed here was whether there were any differences between males and females in three chosen digit ratios, 2D:3D, 2D:4D and 3D:4D. Excluding 3D:4D of the right hindlimb, the digit ratios of all four limbs were significantly smaller in females than in males (Table 3, Figure 1). These results showing clearly the existence of male‐biased dimorphic digit ratios patterns, support findings obtained for other members of the Lacertidae family (Rubolini et al., 2006; Van Damme et al., 2015) as well as most investigated iguanian lizards (Gomes and Kohlsdorf, 2011), but are in contradiction to some other findings (Chang et al., 2006; Tobler et al., 2011; Direnzo and Stynoski, 2012).

Generally, sexual dimorphism in sand lizards is reflected in differences in many morphometric traits. Heretofore, female‐biased sexual dimorphism has been observed in body traits (SVL, TL, body mass); at the same time, male‐biased sexual dimorphism has been expressed in head size (pileus length and width) (Ekner et al., 2008) and length of both fore‐ and hind legs (Majláth et al., 1997). Furthermore, skin hyperaemia behind the cloaca enables determination of sex in this species even in hatchlings (Eplanova and Roitberg, 2015). However, the sand lizard is characterised by great intraspecific variation, including pronounced geographic variation in, inter alia, the magnitude of sexual dimorphism (Roitberg et al., 2015). Therefore, it is necessary to confirm the results describing digit ratio patterns in other locations within the range as well as among other subspecies.

However, the detected sex differences regarding digit's length are not so unambiguous (Table S3). Described in the sand lizard male‐biased hindlimbs 2D:4D ratio seems to result from the 4D length which is steroid‐responsive feature (according to Braña, 2008), while the length of the 2D remains constant, which is consistent with the ‘hypothesis of sex‐specific digit responses‘ (Lofeu et al., 2017). Such observation stays in contrast to the male‐biased pattern confirmed in Leptodactylus frogs, where the 2D is sexually dimorphic and is sensitive to increase testosterone concentrations (Lofeu et al., 2017). Therefore, in our case, sexually dimorphic left hindlimb 4D (as well as in both forelimbs) are significantly shorter in males than in females, and its sensitivity to hormones probably ‘could have evolve through changes in steroid receptors concentrations in specific phalanges of that digit’ (see: Lofeu et al., 2017). However, we must remain aware of the lack of clear dimorphic pattern detected in the right hindlimb 4D or in both side in hindlimbs 3D (Table S2, S3). Thus, this between‐sex differences in some digits length and digit ratios often vary and each feature is dependent of sex‐ and species‐specific growth trajectories and require further detailed experimental study.

Secondly, a clear but contrary relationship has been found between caudal autotomy and specific digits ratio in the sand lizard population, such that individuals with autotomy are characterised by higher 2D:4D ratios in right forelimbs and lower 2D:3D ratios in right hindlimbs (Figure 2, Table S1). For both digit ratios, respectively, the length of fore right 4D and hind right 2D show significant differences between sexes (Tables S2, S3). It seems that the difference in 2D:4D between lizards with autotomised and intact tail, could be explained by potential sensitivity to sex steroids of 4D as in experiment conducting on Z.vivipara (Braña, 2008). However, in light of the current state of knowledge simultaneous occurrence of inverse associations between autotomy and the right hindlimb 2D:3D is difficult to explain. For instance, testosterone's pronounced effect on 2D were reported in the collared flycatcher Ficedula albicollis or in the rufous frog Leptodactylus fuscus in experimental testosterone‐exposed group (Nagy et al., 2016; Lofeu et al., 2017).

The results presented in this paper are probably the first which link digit ratios (suggested to be hormone‐dependent traits) to susceptibility to autotomy in lizards. Although, determining the reasons of occurrence of differences observed only in some limbs and digits is beyond the scope of our work (side‐dependent effects of steroid hormones cf. Nagy et al., 2016). However, it should be noted that our knowledge about the digit ratio of lizards is still limited, and thus such results should be discussed with a great caution. In the context of earlier articles concerning anti‐predator behaviours, Tobler et al. (2012) found that C.pictus it took longer for males with lower digit ratios to initiate an escape response. This behaviour was associated with 3D:4D on the right hindlimbs. At the same time, on average, males with bibs were characterised by shorter latencies in fleeing. In fact, delay in escape from predators may be an adaptive feature which may persist in a population depending on the context (availability of hiding places, the composition of predators) (cf. Tobler et al., 2012). Potential higher levels of circulating testosterone concentration affected behavioural traits (anti‐predator response) as well as the digit ratio, which is higher in individuals who have shed the tail (Figure 2a). So, inter‐individual differences in the amount of circulating steroid levels (testosterone concentration) can be interpreted as a factor establishing variation in digit ratios and another dimorphic trait (Lofeu et al., 2017). The relationship between fitness and digit ratio was found pronounced, for example in the red squirrels Tamiasciurus hudsonicus, where individuals with lower digit ratios had higher reproductive success and lower endoparasite occurrence intensity (Gooderham and Schulte‐Hostedde, 2012). Hence, the ability to avoid predators and tail injuries may directly affect future reproductive success, for example clutch size or success in mating in individuals with original tails (Fox et al., 1990; Martin and Salvador, 1993; Bernardo and Agosta, 2005). For an individual which has shed its tail, the amount of resources available for use in reproduction is consequently reduced, regardless of the benefit (survival) (Clause and Capaldi, 2006). Therefore, this feature (ability to avoid tail loss) may be preserved by natural selection in a given population, along with digit ratio pattern, which may be related to this behaviour (according to model proposed by: Lofeu et al., (2017), see Fig. 4 c.f.). Further research is also needed in this area, since the consequences of tail loss for a lizard vary likely depending on species (cf. Fleming et al., 2013). We assume that the tendency to lose tail results from differences in the expression of sex hormones as the determinants of the digit ratio and may also be related to personality (differences between bold and shy individuals). As we mentioned, sex hormones may partly determine anti‐predator reaction or susceptibility to predators, and consequently the ability to avoid being hunted down. Interspecies variability in digit ratio patterns in Podarcis lizards might be mediated by testosterone and reflected in differences occurring in ‘temperament’ between species (Van Damme et al., 2015; see Introduction). However, it should be mentioned that differences in the behaviour of individuals are observed not only between species but also within species or even on the population level (Pačuta et al., 2018). Finally, in most cases, it is impossible to determine whether a given individual has experienced autotomy more than once. Therefore, for instance, studies on autotomy in individuals of different age should be interpreted with caution. However, in our opinion, the simple approach we used, that is the division between individuals with either an intact or autotomised tail, enabled us to draw correct conclusions. Interestingly, in previous studies carried out in the Wielkopolska region, the share of individuals with caudal autotomy (those having lost a tail at least once), without considering their sex, ranged from 30% to 42.9% (Ekner et al., 2008; Kurczewski, 2014; Dudek et al., 2016). Usually, this phenomenon was noted more often in adult lizards than in juveniles (Ekner et al., 2008; Kurczewski, 2014). The frequency of autotomy observed in the present study exceeded the upper range limit cited above, reaching over 50% of the total number of adults, which may indicate a high degree of predatory pressure at the surveyed region; however, we have no additional data confirming this assumption.

5. CONCLUSION

The use of various limbs digit ratios (2D:3D, 2D:4D and 3D:4D) in presented research poses a greater challenge in comparing the results (see in Introduction); this issue should be kept in mind for the purpose of designing future experiments. In addition to contributing to the advance of knowledge concerning digit ratios in Squamata, the necessity of which has been raised by numerous authors (Rubolini et al., 2006; Gomes and Kohlsdorf, 2011; Tobler et al., 2011; Van Damme et al., 2015; Woodhead et al., 2018), our results also provide new insights into the relationship between digit ratios and caudal autotomy. However, it should be emphasised that autotomy is a complex feature and tail damage is likely to be repeated throughout the life of a specimen, and the probability of keeping the tail intact decreases with age (Bustard and Hughes, 1966). Theoretically, shyer individuals which are more cautious and prepared to conceal themselves quickly can avoid attacks and keep their original tails (Carter et al., 2010). This positively influences the use of available resources for activities related to reproduction (Dial and Fitzpatrick, 1981; Clause and Capaldi, 2006). Moreover, an original tail serves as a storage for surplus substances, or an attribute conferring a higher social position and grater matting success (Fox et al., 1990; Martin and Salvador, 1993; Bernardo and Agosta, 2005). Therefore, despite the fact that sexually dimorphic patterns in digit ratios would initially be hidden from selection (Lofeu et al., 2017), as presented above, possible relationship with tendency to shed tail mediated by testosterone seems to show adaptive features, which may have been evolutionarily preserved. Therefore, in our opinion the inclusion of autotomy as a factor explaining observed digit ratio patterns and their variability between taxa has a great potential.

Our results are novel in a way that we have shown a relationship between autotomy and digit ratio. Furthermore, digit ratio may reflect effect of hormones on behaviour or/and and innate factors, which, to our knowledge, have not been sufficiently considered in the context of tail loss causes (Fleming et al., 2013). Clause and Capaldi (2006) indicated a gap in the existing knowledge concerning autotomy and formulated further research questions, one of which was as follows: ‘Are there hormonal components to this behaviour?’. Even though we did not answer these questions directly, in our opinion the results presented here offer a premise for further research in this area, with the investigated digit ratios serving as a promising biomarker. Thus, we hope that our findings will open up new opportunities to investigate the innate disposition to autotomy.

AUTHOR CONTRIBUTIONS

MK involved in contributions to concept/design, acquisition of data, data analysis/interpretation, drafting of the manuscript, critical revision of the manuscript and approval of the article. KZ involved in critical revision of the manuscript and approval of the article. PT involved in contributions to concept/design, critical revision of the manuscript and approval of the article.

Supporting information

Supplementary Material

Kaczmarski M., Ziemblińska K., Tryjanowski P. Sand lizards Lacerta agilis with higher digit ratios are more likely to autotomy. J. Anat. 2020;237:1103–1113. 10.1111/joa.13277