Diet , sexual dimorphism and reproduction of sympatric racers Philodryas aestiva and Philodryas patagoniensis from the coastal Brazilian Pampa

Herein we investigated diet, sexual dimorphism and reproductive biology of two sympatric congeneric species at the coastal Brazilian Pampa, Philodryas aestiva and P. patagoniensis. Analysis of the stomach content of the scansorial Philodryas aestiva revealed that it feeds mainly on Anura and small mammals, while it occasionally consumes Insecta, Squamata and Aves. The more terrestrial Philodryas patagoniensis presented a broader diet, composed mainly of Anura, Squamata and small mammals. Insecta and Actinopterygii were found in lower frequencies. In females of both species, snout-vent length was significantly larger than in males, while tail length/snout-vent length ratio was significantly higher in males. Males reach sexual maturity at smaller sizes than females in both species. Fecundity was higher in P. patagoniensis (3 to 24 eggs; 15 ± 8.15) than in P. aestiva (10 to 20 eggs; X = 14.50 ± 3.53). Females of P. aestiva showed secondary follicles in all seasons while eggs occurred from middle-winter to early-summer. Secondary follicles occurred throughout the year in P. patagoniensis, but concentrated in spring. eggs occurred in late-summer, middle-autumn and spring. Both species presented positive correlations between SVL and clutch size. Differences in analyzed traits seem to reflect divergences in species morphology and use of habitat/microhabitat.


INTRODUCTION
Data on natural history of snakes are crucial to understand ecological processes at community level and for species conservation (greene 1993(greene , Shine 1995)).Particularly, studies concerning closely related (e.g.congeneric species) in sympatry have a high potential to unveil mechanisms of resource partitioning and distinct life strategies adopted by the species (Carpenter 1952, goodyear andPianka 2008).For instance, studies on dietary data of closely related sympatric species could shed light on their foraging strategies, and therefore on how they partition their resources.Also, as reproduction is essential for the survival of the species, data on reproductive season and fecundity are crucial to implement management and conservation strategies (Manlik et al. 2016, Velasco et al. 2017).
The Philodryas genus comprises 23 species widely distributed in South America (Uetz et al. 2018), which presents both terrestrial and arboreal An Acad Bras Cienc (2019) 91(1) e20180296 2 | 12 habits (Cacciali et al. 2016).They are commonly known as "racers" due to their capacity for rapid displacement and slender body.In the southernmost Brazilian coast, which belongs to the Pampa biome, three species of Philodryas occur sympatrically in open habitats of grasslands and savannas: the green Racer Philodryas aestiva (Duméril, Bibron & Duméril, 1854), the Patagonian Racer Philodryas patagoniensis (girard, 1857) and the lichtenstein's green Racer Philodryas olfersii (lichtenstein, 1823).The first two species are commonly found in open environments while the latter is mostly associated with forest formations (Quintela and loebmann 2009).Comparatively, P. patagoniensis is more robust, terrestrial and attains greater sizes than P. aestiva, which in turn is a typical scansorial species, with a slender body and long and prehensile tail (lema 2002) (Fig. 1).In the southernmost Brazilian coast, Philodryas patagoniensis also occupies a greater variety of habitats, including flooded and xerophytic environments (such as the interior of marshes, saltmarshes and coastal dunes) where P. aestiva is apparently absent (Quintela and loebmann 2009).Aspects of the natural history of P. patagoniensis have been examined in several areas along its distribution (Vaz-Ferreira et al. 1970, Carreira-Vidal 2002, Pontes et al. 2003, Hartmann and Marques 2005, lópez and giraudo 2008, Sawaya et al. 2008, Marques et al. 2012, Neto-Silva et al. 2012, loebens et al. 2016, 2017, Ucha and Santos 2017) but data are still lacking from the southernmost Brazilian coast.As for P. aestiva, dietary data is scarce (Carreira-Vidal 2002, França et al. 2008) and there are no studies addressing reproduction and sexual dimorphism in the species.
given the distinct morphology and habits of the sympatric P. patagoniensis and P. aestiva, the aim of this study was to investigate and compare the feeding habits, sexual dimorphism and reproductive biology from a sample of specimens from the open areas of coastal Brazilian Pampa.In view of previous studies (Hartmann and Marques 2005, Pizzatto et al. 2007, 2008), we expect that differences in size and habits will reflect in divergences on the aspects of feeding and reproductive biology between the species.It is assumed that P. aestiva explores more scansorial prey and presents lower fecundity than P. patagoniensis.

STUDY AReA
We examined specimens collected along a stretch of ca.500 km of the coastal plain of the state of Rio grande do Sul, southernmost Brazil, between the municipalities of Cidreira (30°07'30"S; 50º11'10"W) and Chuí (33°41'04"S, 53°31'18"W) (Appendix; Fig. 2).The study area is nested in the Pampa biome and presents a temperate humid climate according to the regional classification by Maluf (2000).Seasons are well-defined and rainfall (1,600 mm in average) is concentrated mainly in the winter and spring (Vieira 1984, Maluf 2000).Specimens were collected in grasslands  Cienc (2019) 91(1)  e20180296 3 | 12 (predominantly gramineous/herbaceous) and humid savannas (predominantly herbaceous/ arbustive, with sparse trees), the most representative vegetation types in the region.Specimens are housed in the herpetological collection of Universidade Federal do Rio grande (CHFURg), Rio grande municipality, Rio grande do Sul, Brazil.All procedures were carried out according to the international practices for animal use and care under the control of an internal committee of the Universidade Federal do Rio grande, Brazil.

DIeT
We examined a total of 79 specimens of P. aestiva (45 females and 34 males) and 126 specimens of P. patagoniensis (75 females and 51 males).The digestive and reproductive tracts were externalized for analysis through a ventral incision made from the esophageal region to around 5 mm above the cloaca.All stomach and gut contents were removed and identified to the lowest possible taxonomic category.The frequency of occurrence of each identified taxon (FO%) was determined as the ratio between the number of tracts containing one specific taxon and the total number of analyzed tracts.The numeric abundance index (N%) of each identified taxon was calculated as the absolute number of prey items from one specific taxon divided by the sum of prey items of all identified taxa (Corrêa et al. 2016).Stomach and gut content was also classified in prey categories according to broad taxonomic categories, as follows: 1) Arthropoda, 2) Actinopterygii, 3) Amphibia, 4) Squamata, 5) Aves, 6) Mammalia.FO% of each prey type was determined as the ratio between the number of tracts containing one specific prey type and the total number of analyzed tracts.N% of each prey type was calculated as the absolute number of prey items from one specific prey type divided by the sum of prey items of all prey types.
SeXUAl DIMORPHISM AND RePRODUCTION The following measures were taken from each specimen: snout-vent length (SVl), tail length (Tl), tail proportion in relation to SVl (ratio Tl/ SVl).The sex was determined by the presence or absence of the hemipenis via subcaudal incision.
To determine the presence of sexual dimorphism, for each species we examined differences in SVL and in ratio Tl/SVl between males and females with a t test (Quintela et al. 2017).For both species, the Sexual Size Dimorphism (SSD) index was calculated according to gibbons and lovich (1990), where SSD = mean SVl of the larger sex divided by the mean SVl of the smaller sex.
Negative values of SSD indicate that males are larger than females.
The following data were taken from females: total number of ovarian follicles; number of secondary follicles (follicles in secondary vitellogenesis); largest diameter of the largest secondary follicle; total number of eggs; largest diameter of the largest egg.The initial diameter of secondary follicles was determined based on coloration; primary follicles differ conspicuously from secondary follicles due to their darker yellow coloration (see Almeida-Santos et al. 2014).Females were considered as sexually mature when showing at least one of the following characteristics: secondary follicles; oviductal eggs; folded oviducts, indicating recent oviposition (Mesquita et al. 2013).Males were considered sexually mature when presenting coiled and opaque ductus deferens, indicating the presence of sperm (Shine 1977, Almeida-Santos et al. 2014).The female reproductive cycle was evaluated based on the the annual profile resultant from the plot of the largest follicles and eggs of each female (Mesquita et al. 2013, Almeida-Santos et al. 2014).Fecundity was determined by the number of eggs in the oviduct (real fecundity; clutch size) and number of secondary follicles (potential fecundity) (Mesquita et al. 2013).The existence of significant differences in real fecundity between the two species was examined through a t test.The relationship between female SVl and real fecundity (clutch size) was examined through a simple linear regression analysis (Siqueira et al. 2013).
SeXUAl DIMORPHISM AND RePRODUCTION Females of P. aestiva (n=45) presented significantly larger SVl than males (n=34) (t = -4.197;P<0.0001) while males presented significantly higher Tl/ SVl ratios than females (t =13.527;P<0.0001).Similarly, females of P. patagoniensis (n=75) presented significantly larger SVl than males (n=51) (t =5.150; P<0.0001) while males showed significantly higher TL/SVL ratios than females (t =-12.180;P<0.0001).However, an overlap in Tl/SVl ratios between sexes was observed in both species, this being more pronounced in P. patagoniensis (Table II).Mean, standard deviation and range of SVl and Tl/SVl for males and females of both species are shown in Table II.The SSD index was 1.269 for P. aestiva and 1.374 for P. patagoniensis, confirming that females are larger than males in both species.
The SVl of P. aestiva mature females varied from 630 to 835 mm (X = 731.36± 76.67 mm) while SVl of mature males ranged from 383 to 572 mm (X = 487.61± 44.71 mm).Mature females of P. patagoniensis presented SVl varying from 719 to 1076 mm (X = 894.56± 121.10 mm) while SVl in mature males varied from 453 to 786 mm (X = 637.93± 97.80 mm).Thus, females reach sexual maturity at greater sizes than males in both species.
Females of P. aestiva showed primary and secondary follicles in all seasons while oviductal  eggs were found from middle-winter to earlysummer.Females of P. patagoniensis presented primary and secondary follicles in all seasons, with a higher concentration of the later during spring.
Oviductal eggs in P. patagoniensis were found sparsely in late-summer, middle-autumn and spring (Fig. 3).In both species, follicles that presented the characteristics of secondary vitellogenesis had approximately 9 mm of diameter.

DIeT
The present study adds to previous evidence that implies that both P. patagoniensis and P. aestiva are generalists, although the former seems to feed on a broader range of prey (greater variety of species within a category).The diet of P. aestiva consisted basically of small mammals and anurans, with occasional occurrences of Coleoptera, Orthoptera and birds.Insects, however, occurred in association with the remains of one unidentified Squamata and may represent secondary items (items consumed by the squamate prey).Philodryas patagoniensis, in turn, feeds on a wide spectrum which includes anurans, lizards, snakes (including cannibalism), small mammals, and occasionally insects (belostomatid Hemiptera and adult Diptera) and the fish Hoplias malabaricus.Amphibians, however, were the most frequently consumed items, as well as in other populations studied in southern Brazil (Hartmann and Marques 2005)  There still are controversies about the factors shaping divergences between dietary patterns of closely related snake species.Colston et al. (2010) demonstrated that a deep divergence in the evolutionary history of Serpentes (Scolecophidia/ Alethinophidia divergence, occurred during the early Cretaceous, approximately 110 Mya [Hsiang et al. 2015]) was responsible for most of the dietary shift in the group while divergences within the advanced snake's clade (Alethinophidia), occurred during late Cretaceous and eocene (approximately 81-50 Mya) (Hsiang et al. 2015), had few effects in the diversification of diet as a whole.Thus, divergences in dietary patterns of closely-related species can be attributed to effects of habitat, microhabitat and behavior (Colston et al. 2010).The expected differences in diet composition of the two sympatric Philodryas in the present study may be a result of divergence in their use of habitat and microhabitat by the distinct species.So, this can consist in an example of niche evolution, a condition when closely related species adopt distinct ecological strategies which permit its coexistence (Ackerly et al. 2006).Hartmann and Marques (2005) Mesquita et al. 2013, Panzera and Maneyro 2013, Quintela et al. 2017).Males of both Philodryas species presented significantly higher Tl/SVl ratios than females, but due to the overlapped range observed between sexes within both species, the character can not be considered absolutely dimorphic.The tail in males accommodates the hemipenis and retracting muscles (King 1989), thus, higher Tl/SVl proportions are usually expected for males.However, the fact that P. patagoniensis showed more range overlap when compared to P. aestiva was not expected, considering that dimorphism in tail size is discrete or even absent in species with a certain degree of arboreality (Pizzatto et al. 2008), which is the case of the latter.Females of both Philodryas species presented eggs and secondary follicles in both the warmer and colder seasons of the year.In P. aestiva, eggs were restricted to the period from middle-winter to early summer, which indicates that the female reproductive cycle in the species may follow a seasonal pattern.However, analyzes of a larger sample are needed to draw further conclusions.The occurrence of eggs in P. patagoniensis euthanized in late-summer and middle-autumn contradicts the previous data on the female reproductive cycle in the species, considered seasonal.lópez and giraudo (2008) found eggs from October (earlyspring) to January (early-summer) in samples from southern Argentina. loebens et al. (2017), analyzing populations from a large portion of southern Brazil (states of Paraná, Santa Catarina and north and central Rio grande do Sul), reported the presence of eggs from August (middle-winter) to February (middle-summer).In our P. patagoniensis sample, higher concentrations of secondary follicles and eggs in the spring supports a seasonal trend, but the occurrence of eggs in middle-autumn dissociates from the pattern.Thus, P. patagoniensis females may exhibit plasticity in reproductive cycles possibly conditioned by local environmental factors, which has already been verified in other widespread xenodontine species (see Pizzatto andMarques 2006, Quintela et al. 2017).
In accordance with our expectations, due to P. patagoniensis larger size, this species presented greater fecundity when compared to P. aestiva.Morphological adaptations to arboreality such as slender bodies and reduction of the body cavity, as a result of improved proportionality of the tail in relation to body length, act as limiting factors in the development of great clutch sizes (Pizzatto et al. 2007).However, real fecundity in our P. aestiva sample was greater than that recorded in arboreal P. olfersii (4 to 11 eggs; Mesquita et al. 2013), considering that the later reaches considerably larger body lengths.This fact may be due to the conspicuous smaller size of P. aestiva eggs in relation to P. olfersii eggs (Mesquita et al. 2013;present study), so that a larger number of eggs can develop in the former.The positive correlation between female SVl and fecundity in P. aestiva was also expected, since larger females are able to develop larger clutches (Shine 1994).In P. patagoniensis, the observed fecundity is similar to that found in other studied regions (lópez and giraudo 2008, loebens et al. 2017).The positive correlation between female SVl and fecundity in our analyzed sample also corroborates the pattern detected in other investigated populations (lópez and giraudo 2008, loebens et al. 2017).However, unlike the results of lópez and giraudo (2008) and loebens et al. ( 2017), such correlation was not significant in our study.
In conclusion, the two sympatric Philodryas species herein analyzed diverge in traits of their natural history.Such divergence can be related to differences in morphology and use of habitat and microhabitat.Thus, more comparative studies are encouraged, aiming to clarify distinct aspects of the biology and ecology of sympatric congeneric/ closely related species.

Figure 2 -
Figure 2 -location of the sampled sites at southernmost Brazilian coast, Rio grande do Sul state.

Figure 3 -
Figure 3 -Monthly variation in diameter (mm) of the largest follicles (black circles) and eggs (white circles) of Philodryas aestiva (upper image) and Philodryas patagoniensis (bottom image) from the Brazilian coastal Pampa, Rio grande do Sul state.The horizontal line represents the length of the largest axis from which the follicles are considered in secondary vitellogenesis (9 mm).

TABLE II Values of means (X), standard deviation (sd) and range of the variables analyzed in Philodryas aestiva and Philodryas patagoniensis from the Brazilian coastal Pampa, Rio Grande do Sul state (SVL = snout-vent length; TL/SVL = tail proportion in relation to SVL).
also found differences in diet composition between scansorial P. olfersii and P. patagoniensis from the northern Pampa in Rio grande do Sul and attributed such differences to divergences in foraging microhabitat use by the species.The occurrence of hylids and one bird in