Melanism, body size, and sex ratio in snakes—new data on the grass snake (Natrix natrix) and synthesis

It is postulated that melanism in ectotherms is adaptive by enhancing thermoregulation, subsequent resource acquisition, and growth. Such effects may differ between the sexes as a result of the differential costs of self-maintenance and reproduction, but empirical support for the sex-specific consequences of melanism remains inconsistent. We studied the effects of melanism on body size and sex ratio in a population of the European grass snake (Natrix natrix) in SE Poland and also carried out a systematic review of the literature on the consequences of melanism in terrestrial snakes. Melanistic grass snakes of both sexes appeared to be smaller than the typical phenotype, which indicates higher predation pressure and minimal thermal benefits for black individuals. A female-biased sex ratio was observed in the typical phenotype, but not in melanistic snakes, suggesting that the costs for females and/or benefits for males are higher in melanistic individuals. In conjunction with earlier studies, our data indicate that the consequences of melanism may be related to the reproductive mode of species. In viviparous species, melanism tends to improve growth and/or body size and is more frequent in females, whereas the opposite holds for oviparous snakes. Further studies on melanism should examine a wider array of species with different reproductive strategies and traits beyond the usual thermal benefits.


Introduction
Body coloration is a multifunctional trait often characterized by sophisticated variation (Kemp et al. 2005). Therefore, discontinuous phenotypes are generally thought to bear fitness costs as a result of the primary functions of a given color variation having been lost. The maintenance of such phenotypes within populations, i.e., color polymorphism, thus represents an interesting evolutionary phenomenon (Forsman 1995a;Forsman et al. 2008).
Melanism is an example of color polymorphism in which a phenotype is characterized by an overconcentration of melanin compared with the typical color morph (Trullas et al. 2007). In small vertebrates, melanistic individuals are known to bear an elevated risk of predation . However, there are positive consequences of a melanistic phenotype, including a better capacity to cope with parasites (Roulin et al. 2001), an improved anti-UV protection (Dubey and Roulin 2014), and a higher concentration of sex hormones (Ducrest et al. 2008) which prevents its complete removal from the population. The occurrence of melanism in ectotherms is particularly interesting, owing to its predicted positive effect on thermoregulation, including increased heating rate and higher achievable body temperature (Forsman 1995b;Trullas et al. 2007). Such thermoregulatory consequences are predicted to have a positive impact on the life-history traits of black individuals, including their body size-a prerequisite of survival and fecundity (Olsson 1993;Civantos et al. 1999). The improved thermoregulation anticipated for the melanistic phenotype is also likely to have sexspecific consequences, because of the sex-specific variation in the costs of reproduction and self-maintenance (e.g., Madsen and Shine 1993). In other words, the sex that bears the highest costs should benefit more from being black-colored, and, consequently, should grow larger and/or include a higher proportion of black individuals or be more frequent in the melanistic fraction (Forsman and Ås 1987;Luiselli 1992).
Snakes are frequently studied in the context of the consequences of melanism because this phenotype is common in many snake species. However, available data do not seem to provide a consistent picture of the sex specificity of the effects of melanism. A number of studies of viviparous species, including European vipers (Vipera berus and Vipera aspis), show a tendency towards greater size in melanistic individuals of both sexes and an elevated proportion of black females, which are the sex that bears the highest costs of reproduction (e.g., Monney et al. 1995Monney et al. , 1996Madsen and Stille 1988). Available data on oviparous species are less conclusive, but the opposite effects of melanism have been observed concerning both body size and sex-specific phenotype proportion or sex ratio (e.g., Luiselli 1995;Zuffi 2008). Further research is needed to understand and clarify the direction of the effects of melanism on body size and sex ratio in these species.
Our aim in this study was to investigate the association of melanistic phenotype with body size and sex structure in a widespread oviparous reptile, the European grass snake (Natrix natrix). In this species, there is a pronounced sexual size dimorphism, i.e., females are larger than males, and the numbers of each sex in a population are roughly equal (sex ratio 1:1) (Borczyk 2007). The occurrence of a melanistic phenotype, though widely documented in this species (e.g., Nilson and Andren 1981;Böhme and Wiedl 1994), appears to be discontinuous. Moreover, little is known about the effects of melanism on body size and sex ratio. Although increased predation pressure has been confirmed in melanistic grass snakes (Madsen 1987), the widespread occurrence and persistence of the black phenotype indicate a beneficial side. In line with previous studies, we predict that the melanistic phenotype will achieve a larger size (Madsen and Stille 1988). With regard to the sex ratio, we expect females to be the sex that benefits most from being black-colored and to outnumber males in the melanistic fraction of the population. This is because male combat behavior is negligible in the grass snake (Borczyk 2004), so the costs of reproduction are clearly lower for males than for females. In addition, we reviewed the published data on sex-specific correlates of melanism in terrestrial snakes to find out whether any general patterns could be discerned. For this, we took the reproductive mode into account, owing to the fundamentally different costs of reproduction between viviparous and oviparous species.

Methods
The study was conducted in the Bieszczady Mountains (SE Poland; 49°14′ 5.28″ N 22°33′ 30.91″ E). The presence of the melanistic phenotype in grass snakes in this area was described back in the 1970s and repeatedly confirmed in subsequent reports (Błażuk 2007). Here, the snakes were surveyed monthly from April to September and searched for in a wide range of different habitats, including but not restricted to forests, meadows, riverbanks, ecotonal zones, and anthropogenic sites. Individual grass snakes were identified based on external features, such as ventral color pattern and/or scale clipping. For each snake captured in the study, we determined the sex and measured the snout-vent length (SVL). Only data from sub-adult and adult snakes for which sex could be reliably identified were considered in this study (≥~30 cm SVL; Bury and Zając 2019). Fieldwork was carried out between 1981 and 2013. Before 2009, measurements were made only on melanistic snakes, but in subsequent years included individuals of both phenotypes. Altogether, our data included 103 specimens, out of which 67 were melanistic. None of the snakes was ever recaptured, probably because of the large extent of the study area.
To get insight into the possible effect of color phenotype and sex on SVL, we analyzed SVL data using a general mixed model, with phenotype, sex, and interaction of both as fixed factors; the source of the data was a random factor. To account for the possible effect of the data source (i.e., the period of sampling-before and after 2009), we included it as a random factor in the analysis. To explore the possible association between phenotype and sex ratio, the sex ratio within each phenotype (i.e., melanistic and typical color morph) was analyzed using the chi  test. All the analyses Fig. 1 The impact of phenotype on sex-specific body size (SVLsnoutvent length) in the grass snake (Natrix natrix) (least square means ± standard error). There are significant differences between the sexes in both phenotypes (p < 0.001), as well as between both phenotypes (p < 0.001). Interaction term was non-significant (p = 0.24) were performed in the Statistica software (version 13.1; StatSoft Poland).
We carried out a systematic review of the literature by scanning the Web of Science, Scopus, and Google Scholar using the keywords: "snake," "reptile," "melanism," and "coloration". We extracted papers that reported on the association of melanism with at least one of the following variables: body size, mass, condition, growth, sex ratio, and sex-specific melanism prevalence. For each species, we also noted the reproductive mode based on the reptile database (http://www.reptile-database.org/).

Results
The general mixed model revealed significant effects of phenotype (F 1,98 = 16.61; p < 0.001) and sex (F 1,98 = 68.02; p < 0.001) on SVL, whereas the interaction of both factors appeared to be non-significant (F 1,98 = 1.38; p = 0.24). The random effect of data source was not significant (F 1,98 = 0.578; p = 0.45). Females appeared to be larger than males by ca 30% in the typical phenotype and 56% in the melanistic one ( Fig. 1). Melanistic males were ca 29% smaller than typical ones, whereas melanistic females were ca 14% smaller than typically colored ones (Fig. 1). The descriptive statistics based on the raw data (mean ± SD; minmax; sample sizes, and sex ratio) are summarized in Table 1.
Our literature review unearthed reports on just eight terrestrial species where melanism was studied in relation to sizerelated variables, sex ratio, or sex-specific prevalence. These studies included four oviparous species (Natrix natrix, Hierophis viridiflavus, Heterodon platirhinos, and Elaphe quadrivirgata), as well as four viviparous species (Vipera berus, V. aspis, Vipera renardi, and Thamnophis sirtalis). For other species, no further data were available, despite melanism being reported. For example, the occurrence of melanism in dice snakes (Natrix tessellata) was frequently described, but we failed to find any report of its effects on body size or sex ratio (Mebert 2011; Ajtić et al. 2013). Given the limited number of species for which we found data on the effects of melanism, we could not perform formal statistical tests and our analysis is limited to descriptive statistics. The studies we have summarized show that melanism in viviparous species tends to have a positive impact on body size or related variables (8/17 studies report a positive effect, 8/17 report no effect, and 1/17 report a negative effect) in both males (3 reports) and females (4 reports) or without sex distinction (3 reports). Moreover, in most cases, melanism in viviparous species is more common in females than in males (7/15 studies report a female bias, 3/15 report a male bias, and    Strugariu and Zamfirescu (2011) 5/15 report no effect). Less clear is the picture for oviparous snakes. Excluding our study, 2/5 studies reported a negative effect of melanism on size-related traits, 2/5 reported a positive effect, and 1/5 reported no effect. The prevalence of melanism does not seem to show a clear pattern either, with 2/6 studies indicating a higher frequency of melanism in males, 1/ 6 in females, and 3/6 with no tendency. The addition of our study, however, tilts the balance towards a negative effect of melanism on body size (3/6 reports of a negative effect of melanism on size and 3/7 reports indicating a tendency towards increased prevalence in males rather than females).

Discussion
Our study is one of the few to clearly show a negative association between melanistic phenotype and body size in an oviparous species, the European grass snake. We have ruled out a potential bias towards capturing individuals of a given size, because the snakes were captured randomly in a wide spectrum of habitats, throughout the season, though mostly in the post-mating period, and juveniles were not incorporated in the analysis. The documented pattern stands contrary to our predictions and the generally reported or anticipated patterns (e.g., Andren and Nilson 1981;Madsen and Stille 1988;Forsman 1995a). The magnitude of size differences between phenotypes remains similar in both sexes, whereas sexspecific effects of the black phenotype appear to occur at the sex ratio level. In the typical phenotype, we observed a clear female-biased sex ratio, in line with other data from southern Poland (Juszczyk 1987), but contrary to that generally expected for species with a genetically determined sex ratio of 1:1 (Fisher 1999;Shine and Bull 1977). Such female bias could be the outcome of a high local productivity possibly underlying a high abundance of reptiles in general (Błażuk 2007) and be beneficial for females, the sex that bears much higher costs of reproduction. In the melanistic fraction, however, we did not observe such a pattern, as the sex ratio did not differ from 1:1, which may be due to the smaller proportion of females and/or the greater proportion of males compared with the typical phenotype. Although the effects of the melanistic phenotype on both sexes seem to differ between the individual and the population levels, we consider the responses on both scales to be complementary and driven by common environmental factors.
The smaller sizes of melanistic snakes may represent an outcome of at least two non-mutually exclusive factors: the pleiotropic effect of black phenotype expression (Ducrest et al. 2008) and the high removal rate of larger melanistic individuals by predators. Whereas we cannot exclude the former factor as being responsible for our results, the latter one is partially corroborated by the pattern observed at the sex ratio level. In the grass snake, females are well known for achieving larger sizes, a pattern confirmed in this study, and known to promote predation risk (Wellborn 1994;Niskanen and Mappes 2005). Such size bias, combined with the impaired antipredatory function of the black phenotype (Madsen 1987), may impose stronger predation pressure on females. This, in turn, may lead to a reduced proportion of females in the melanistic population fraction compared with the typical one.
A smaller size is generally assumed to have a negative impact on reptiles' survival (Civantos et al. 1999) and fecundity (Olsson 1993). Therefore, the persistence of the melanistic phenotype in the studied population (Błażuk 2007), despite its negative effect on body size, suggests that black coloration must have beneficial consequences for other features. Anti-UV protection driven by high-melanin concentration seems to be of minor importance because the keratinized outermost layer is itself a good anti-UV barrier (Tercafs 1963;Chang and Zheng 2003). Also unlikely is the role of melanism in mate selection, since snakes do not appear to rely on visual cues . Melanism expression is, however, suggested as improving defense against pathogens (Roulin 2014). This effect has already been demonstrated in endotherms, but in ectotherms, the elevated immune response has only recently been demonstrated in melanistic lizards (Vroonen et al. 2013;Seddon and Hews 2016;Baeckens and Van Damme 2018). The positive impact of melanism on immunity could be more pronounced in males, which are generally known to exhibit a lower immune response compared with females (Saad and Shoukrey 1988). Such an asymmetric effect in favor of males could also explain the higher proportion of this sex in the melanistic phenotype compared with the typical one.
Our systematic review of the available data points to an emerging pattern of the negative impact of the melanistic phenotype on body size in oviparous snakes (Table 2). Conversely, in  Újvári et al. 2001 viviparous species, there is a discernible tendency towards a positive effect of melanism on size-related traits and a clear female-biased prevalence of the melanistic phenotype. Such a discrepancy between reproductive modes can be attributed to the different costs of reproduction borne by females. In general, one can expect the energetic costs of reproduction for females to be lower in oviparous species than in viviparous species, owing to the shorter time of embryo retention in the body cavity and the longer foraging time window before the onset of the winter (Madsen and Shine 1992;Gregory et al. 1999). This may weaken the impact of thermal benefits and lead to the suggested reduction in larger individuals and females as a consequence of predation Madsen 1987), indicated not only here in grass snakes, but in other oviparous species as well. The universality of this pattern requires studies of other species, preferably in conjunction with data on individuals' ages, in order to rule out the possibility of the negative effect of melanism on growth/body size instead of large-specimen removal. Such an anticipated specificity of melanism effects relative to reproductive mode may further drive the diversity of population responses towards environmental change. Future studies on the consequences of melanism should include a wider array of features in a sex-specific context if we are to broaden our understanding of the mechanisms underlying the persistence of polymorphism beyond thermal benefits. Such data could hint at a link between melanism prevalence and population viability, an aspect especially relevant in the light of climate change.
Funding information This study was funded by the Institute of Environmental Sciences, Jagiellonian University, Kraków (N18/DBS/ 000003).

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.