Prey availability influences the diet of Scinax fuscomarginatus in a Cerrado area , Central Brazil

Prey availability influences the diet of Scinax fuscomarginatus in a Cerrado area, Central Brazil. Prey availability in an environment may change seasonally and these changes should be considered as determinant factors for the diets of anurans. Scinax species are generalist predators that feed on arthropods, but data concerning their diet in relation to prey availability are lacking. In this study, we describe the diet of Scinax fuscomarginatus by evaluating its possibly generalist diet related to prey availability in its environment. We studied the diet of Scinax fuscomarginatus by analysing the stomach contents of 48 individuals captured in the Campo Grande municipality, Mato Grosso do Sul state, Brazil. We found eight prey categories, the most common and most representative being Hemiptera. The selectivity index of the two most important prey varied inversely between dry and wet seasons. Prey availability also varied between seasons. These results suggest a temporal pattern in prey composition and in the diet of Scinax fuscomarginatus.


Introduction
Understanding the diets of anurans is paramount to unveil their natural history and ecological impacts in both terrestrial and aquatic environments as amphibians help control the populations of many organisms and represent a link between terrestrial and aquatic environments (Toft, 1980(Toft, , 1981Duellman and Trueb, 1986;López et al., 2009). Despite great efforts to analyse the diets of anuran species, few approaches have considered prey availability in environments (Sabagh and Carvalho-e-Silva, 2008;Caldart et al., 2012;Araujo-Vieira et al., 2018). Some studies have already shown that prey availability can change seasonally, and these changes should be considered as a determinant factor for anuran diets (Cogălniceanu et al., 2001) as prey availability is intrinsically related to the trophic ecology of anurans (Kikuchi and Ueida, 1998;Ott and Carvalho, 2001;López et al., 2009). Amphibians are considered opportunistic and generalist predators (Vaz-Silva et al., 2005;Toledo et al., 2007;Neves et al., 2014), but some families (e.g., Microhylidae and Dendrobatidae) have specialized diets, feeding mostly on ants (Hymenoptera: Formicidae) and termites (Isoptera) (Attademo et al., 2007;Berazategui et al., 2007;Santana and Juncá, 2007;Forti et al., 2011).
Tree frogs of the genus Scinax are considered generalist predators, commonly feeding on arthropods such as Araneae, Hemiptera, and Orthoptera, with the most commonly preyed order varying according to the species (Solé and Pelz, 2007;Teixeira and Rödder, 2007;Kittel and Solé, 2015;Blanco Torres et al., 2017). The snouted tree frog Scinax fuscomarginatus (A. Lutz, 1925) is characterized by its elongated body, small size (average snout-vent length: 23 mm), and yellowish back with wide dark brown sidebands (Brusquetti et al., 2014). It occurs in southern, central, and eastern Brazil (as far north as Piauí and Ceará, west to southern Amazonas), eastern Bolivia, Paraguay, and north-western Argentina (Leite Jr et al., 2008;Lima et al., 2017). Studies performed with the species have focused on taxonomy (Brusquetti et al., 2014), reproduction Haddad, 2005a, 2005b) and geographic distribution (Leite Jr et al., 2008;Lima et al., 2017), with no data about the diet of the species. Given the lack of information about this species' natural history, and the importance of prey availability for anuran diets, the aim of this study was to describe the diet of Scinax fuscomarginatus, to verify whether the species has a generalist diet, as suggested for other species in the genus (Blanco Torres et al., 2017), and to evaluate whether this generalist diet is related to prey availability in its environment. We predicted that the diet of Scinax fuscomarginatus would be generalist without prey selection, as seen in most anurans.

Field survey
We studied the diet of Scinax fuscomarginatus by analysing the stomach contents of 48 individuals (27 males and 21 females) captured at the Reserva Particular do Patrimônio Natural (RPPN) Fazenda Santa Fé (-20.5131 ºS, -54.7277 ºW, 500 m a.s.l.) in the Campo Grande municipality, Mato Grosso do Sul state. We collected anurans and invertebrate specimens during the dry and wet season (July 2017 and December 2017, respectively). We sampled anurans by active search (Scott and Woodward, 1994) and visual and acoustic encounters (Zimmerman, 1994) along a swamp in the area. To estimate relative abundance of prey, we installed 20 pitfall traps (300 ml plastic cups) at the soil level in random locations around the swamp. The 70 % alcohol was used as a preservative in the traps, adding drops of detergent to break the surface tension of the solution. Pitfall traps were opened at sunset and removed at sunrise. These traps may underestimate some groups such as sedentary prey, flying insects, or Orthoptera. To avoid sampling bias, we used an entomological umbrella method for 30' to capture arboreal prey. Since invertebrates obtained in the diet are partially digested, we categorized the food items into operational taxonomic units (OTUs; Sneath and Sokal, 1962), usually to Order, except for the Formicidae family, which was separated from other Hymenoptera because of its unique morphological and ecological characteristics, and larvae, which are usually included in the same OTU (e.g., Lepidoptera larvae).
The anuran specimens were killed using topical anaesthetic (Xylocaine 5 %) and then fixed with 10 % formaldehyde before analysing the stomach contents. We removed stomachs through a small abdominal incision and stored the contents in separate vials. To access the diet of frogs, the samples were killed as part of a large research project that studies anuran trophic networks (Fundect #71/700.146/2017). We preserved the frog individuals in 70 % alcohol. The captured specimens were housed at Coleção Zoológica de Referência da Universidade Federal de Mato Grosso do Sul (ZUFMS AMP 07800-07829).

Diet composition
We analysed the stomach content under a stereomicroscope and identified each item to order level. After measuring the length and width of each prey, we estimated their volumes using the formula of an ellipsoid: where, V is the volume, W the width, and L the length (Griffiths and Mylotte, 1987). For each item (prey category), we calculated the number, volume, and frequency of occurrence in both absolute and percentage values. We then calculated the Index of Relative Importance (IRI) to determine the relative importance of each prey item in the diet using the following formula, according to (Pinkas et al., 1971): where % N is equal to the relative number of each prey item per sample set, % P is equal to the mass percentage of each prey item in the sample set, and % FO represents the relative frequency of occurrence on the entire samples (Krebs, 1999). Higher IRI values indicate greater importance of the prey category in the diet. We performed a PERMANOVA analysis to test whether diet composition varies between sexes, with euclidian distance, using the prey frequency. Statistical analyses were conducted in the R software v.3.4.2 (R Core Team, 2017) using the 'vegan' package (Oksanen et al., 2015) Prey selectivity With regard to prey selection in the environment, we compared the relative abundance of each prey category in the diet with the relative abundance of the same prey sampled in the environment. For this purpose, we used Vanderploeg and Scavia's Relativized Electivity Index (Vanderploeg and Scavia, 1979). This index is calculated by first finding the selectivity coefficient for diet item i, Wi: where r i is the proportion of bites taken in each category i and p i is the proportional cover of each category i. The index W i ranges from 0 (total avoidance) to 1 (total preference). The relativized index is: where n represents the number of diet categories available. This index ranges from -1 to 1, with 0 indicating random selection, negative values indicating avoidance or inaccessibility of the prey item and positive values showing active selection.  (total), and in the dry and wet season: N, number of prey items found in total and as a percentage, where some individuals hold multiple items; F, number of individuals containing these prey items (frequency of prey items found) in total and as a percentage; V, volume in mm³ and in percentage; IRI, index of relative importance (important prey items are in bold).

Diet composition
Diet composition did not differ between sexes (males = 27, females = 21, p = 0.1). We found eight prey categories, of which Hemiptera was the most representative group (N = 51.72 %), the most important prey category (IRI = 740.09) and the most volumetric prey (V = 61.67 %) (table 1). Blattaria and Lepidoptera larvae were the least frequent items, with a single occurrence. We found stomach contents in 48.93 % of all analysed stomachs. A total of 23 stomachs were empty.

Prey selectivity
Environmental sampling provided a total of 746 inver-tebrates (484 in dry and 262 in wet seasons, respectively), representing 16 taxa in the dry season and 13 in the wet season (table 2). All prey categories found in Scinax fuscomarginatus stomach contents were also found in the sampled environment, suggesting that food resource estimations were represented by the use of pitfall traps and entomological umbrella.
In contrast, 11 prey categories recorded in prey availability were not consumed by Scinax fuscomarginatus: Acari, Dermaptera, Diptera, Hymenoptera, Lepidoptera, Hymenoptera Larvae, Orthoptera Larvae, Mantodea, Myriapoda, Opiliones, and Orthoptera. In the dry season, Scinax fuscomarginatus positively selected only Hemiptera and negatively selected Araneae and Blattaria. However, in the wet season, Scinax fuscomarginatus showed a weakly positive selection for Isoptera and Araneae, but a weakly negative selection for Hemiptera ( fig. 1).

Discussion
The diet of Scinax fuscomarginatus was composed of eight prey categories, with Hemiptera being the most important category, followed by Araneae and Isoptera. The selectivity index of the two most important prey items varied inversely between the dry and rainy seasons, as did prey availability, suggesting a temporal pattern in prey composition and consumption. We found stomach contents in 48.93 % of all analysed stomachs, a higher percentage than that in the diet of some other previously studied Scinax: S. squalirostris (40 %; Kittel and Solé, 2015), S. rostratus (≅ 29 %) and S. ruber (40 %;Blanco Torres et al., 2017), S. granulatus (≅ 36 %) and S. perereca (≅ 34 %; Solé and Pelz, 2007), but lower than that in S. argyreornatus (≅ 83 %; Teixeira and Rödder, 2007). We did not find dietary differences between sexes, possibly due to the opportunistic and generalist habits of anurans (Toft, 1980). Hemiptera was the most important order for the diet of Scinax fuscomarginatus, in contrast with other Scinax species. Orthoptera was observed to be the most important order for Scinax rostratus and S. ruber (Solé and Pelz, 2007) and Arachnida have been found to be the most important prey for S. squalirostris (Kittel and Solé, 2015). Most anurans are considered generalist predators of arthropods (Toledo et al., 2007). However, the diets of Scinax species studied thus far exhibit different categories of important prey, frequency and prey size (Blanco Torres et al., 2017). Scinax species are scansorial, like most hylids, perching on herbaceous plants or shrubs at swamp edges during the breeding season (Teixeira and Vrcibradic, 2004). Furthermore, they are considered active searchers (Teixeira and Rödder, 2007). We therefore consider microhabitat segregation is responsible for the differences among diets. These differences could be related to prey availability in the environment. The irregular conditions of the studied areas certainly influence the composition of invertebrates, changing prey availability in each habitat and leading to differences in anuran diets (López et al., 2009). However, none of the compared species showed prey availability in the environment, which made comparison among species difficult, since we could not draw conclusions about how species select these prey or whether they only feed on what is available in the environment.
In our study, prey composition showed a seasonal pattern, reflected in prey selectivity varying during wet and dry seasons. The composition of invertebrates in an environment can change throughout a year in Fig. 1. Vanderploeg and Scavia's (1979) electivity index E* for prey categories of Scinax fuscomarginatus diet.  de Vanderploeg y Scavia (1979), E*, para las categorías de presas de la dieta de Scinax fuscomarginatus. Prey category relation to climatic variations, different requirements among species, and life history stages (Kikuchi and Ueida, 1998;Ott and Carvalho, 2001;Santana et al., 2015). During the dry season, S. fuscomarginatus showed no reproductive activity, which allowed it to hunt by the active search method and select prey items. In the dry season, S. fuscomarginatus positively selected Hemiptera, a common anuran prey. On the other hand, in the wet season, when individuals are focused on reproduction (males calling and females choosing males), predation was opportunistic. In the wet season, S. fuscomarginatus positively selected Isoptera (winged individuals), an easier prey as it lives in colonies and forms termite flocks (Rafael et al., 2012). In addition, Isoptera are a greater source of protein for anurans than other sclerotized prey such as Hemiptera and Coleoptera (Biavati et al., 2004). Proteins are essential in this season because anuran reproduction demands a greater amount of energy (Taigen and Wells, 1985). There is therefore a temporal pattern in prey availability and in anuran reproduction. As emphasized by Pough et al. (1992), besides the morphology and physiology of Scinax fuscomarginatus, prey availability and reproductive activity lead to the seasonal differences in hunting approaches and thus in diet.