1 Introduction

For ecological purposes it is important to evaluate which environmental conditions present limitations to communities (^{Cetra et al., 2011}). According ^{Rivas (1964)} the term syntopic to be used in reference to two or more related species which occupy the same locality (macrohabitat). A diet analysis may reveal important information about the trophic dynamics and resource partitioning among fish species (^{Ross, 1986}). The resource utilization patterns constitute a fundamental mark of the ecological systems (^{Winemiller and Pianka, 1990}), in which the food resources partitioning among coexistent species plays a more important role than that described even for the habitat partitioning inside aquatic environments (^{Schoener, 1974}; ^{Ross, 1986}). Such a division may occur both in an intraspecific level as in an interspecific one (^{Gerking, 1994}), and is seen as one the basic causes of population and community structurations (^{Tofoli et al., 2010}). On the other hand, coexisting species may present a niche division based on spatial or temporal scales (^{Correa et al., 2011}). Mechanisms of living and interactions among species remain as a central theme for the populations ecology and niche differences have been evoked as fundamental to keep biodiversity in distinct scales (^{Leibold and McPeek, 2006}).

An elevated trophic plasticity may permit fishes to fit their feeding habits to fluctuations of food availability (^{Balassa et al., 2004}). The two stochastic processes and the abundance of some kinds of food may reduce an intraspecific competition, which may permit a better coexistence of species in a same place (^{Dias and Fialho, 2011}; ^{Uieda and Pinto, 2011}). The species which are phylogenetically related may display ecologic similarities in diverse aspects; so, in this way, feeding data may subsidies factors that segregate sympatric species, and may inform some single ecologic questions, such as competition and food resource partitioning (^{Tofoli et al., 2010}).

Other factors such as the trophic morphology, the use of distinct microhabitats, activity periods and feeding strategies may minimize overlapping effects (^{Casatti, 2002}; ^{Brazil-Sousa et al., 2009}; ^{Cetra et al., 2011}). The use of morphologic attributes may reflect an integration of a broad variety of ecosystem influences to the organisms which live inside those systems, mainly for aquatic organisms and are suited to define morphologic and ecologic similar patterns (^{Winston, 1995}). Such morphologic variations may have favored the coexistence of fish groups in diverse regions where they live. This may be due to spatial and temporal differences of their distributions, which may appear as distinct adaptations for the habitat placement and resource exploration, which may minimize some kinds of competition (^{Lagler et al., 1977}). In fact, the ralationship between diet, body form and taxonomic aspects is commonly predicted for different levels in a community, which may adjust morphologic and ecologic configurations (^{Winston, 1995}). ^{Catella and Petrere Júnior (1998)} believe that body form and feeding habits may reflect the general characteristics of the chemical and physical factors of the environment.

The Loricariidae represents a large group of fishes found primarily in freshwater ecosystems of the Neotropical Region (^{Nelson, 1994}), and are known to constitute the world’s most diverse family of catfishes. They are represented by six subfamilies, about 90 genera and a few more than 680 described species (^{Menezes et al., 2007}). Individuals of this family present morphologic differences, such as dermal plates, mouth presenting sucking ventral lips, with were certainly important role for the evolutionary success of the Loricariidae (^{Schaefer and Lauder, 1986}; ^{Schaefer and Stewart, 1993}). The Loricariidae display typical benthonic habits, living permanently near to the bottom where they use to grasp substrate algae or search for invertebrates (^{Britski et al., 1999}). Notwithstanding this fact, ecologic studies concerning loricariids are important to better understand their singular radiation, and also to figure out the dynamics of nutritional patterns in distinct tropical systems (^{Flecker et al., 2002}; ^{Nonogaki et al., 2007}).

Hypostomus represents the most diverse genus within this family (^{Weber, 2003}; ^{Hollanda Carvalho et al., 2010}; ^{Borba et al., 2013}). The species of this genus live frequently in rapid flowing rivers and display a benthonic habit near to stony bottoms (^{Garavello and Garavello, 2004}). Another aspect to distinguish Hypostomus from other Loricariidae is its distinguished ability to breathe atmospheric air. Their stomach walls are plenty of vessels which permit them to gas changing from the blood to the stomach internal space, when they swallow directly air from the surface (^{Armbruster, 1998}). According to ^{Val and Almeida-Val (1995)}, they may be classified as species to present an accessory breathing, and present this behavior specially when hypoxic conditions are happening.

There are relatively few feeding studies related to species of this genus. In spite of their large distribution in almost all Brazilian rivers there are important gaps in the comprehension of their ecologic interactions and also on morphological feeding aspects. Therefore, such ecological attributes should be better known to evaluate how they may explore feeding resources (^{Garavello and Garavello, 2004}). In the present study we examine diet and trophic interactions of four sympatric species of Hypostomus from Corumbataí river in the São Paulo state in the southwestern of the Brazil. The analysis focuses in niche breadth, niche overlap between species and correlation of the diet with morphological features.

2 Material and Methods

The Corumbataí river basin is inserted in the Medium Tietê Peripheric Depression of the São Paulo State, at the northeastern portion of the sedimentary Paraná basin (^{Cetra, 2005}), in which the Corumbataí is one of the most important affluent of the Piracicaba river. Collections were made monthly from June 2002 to May 2003. The collection site (22° 12' 47” S and 47° 37' 40” W) locate near to an urban area of Corumbataí, at 25km from the water spring (Figure 1). This river place locates up to the waste deposit coming from Corumbataí, which maintains the water condition almost unchanged (^{Lima Júnior, 2004}; ^{Cardone et al., 2006}).

Figure 1 Partial map of South America showing the Alto Paraná river basin and location of Corumbataí river basin with localization of the collection site. 

For the collections of specimens of Hypostomus sp. the employment of cast nets was used. This device measures 2 meters height and 1.2, 2.0 and 3.5 cm between the opposite meshes. It is a well suited method, as such fish use to move slowly. The caught individuals were separated in boxes containing ice, to be analyzed at the Zoology Department of UNESP - Rio Claro, SP. Identifications were based on Zoology Museum Prof. Heraldo Bristki and confirmed using ^{Graça and Pavanelli (2007)}. So, four main species were identified: H. albopunctatus (Regan, 1908), H. ancistroides (Ihering, 1911), H. regani (Ihering, 1905) and H. strigaticeps (Regan, 1908). These were the most common species found in the system.

The diet analysis was made using stomach contents from the anterior third part of the gut, including the stomach, which contains normally the less digested material. The whole material for analysis was introduced in a graduated cylinder containing 70% ethanol filled until 50 mL. From the homogenized contents standardized sub-samples were disposed on lamina for identification and counting of food items, which were identified to the lowest possible taxonomic level. The food importance based on those analyses was established based on the volumetric method (^{Hyslop, 1980}). For the evaluation of the similarity of feeding patterns among the species a comparison of the volumetric frequencies was made by the one-way test ANOSIM (5%) using the software PAST 2.15 (^{Hammer et al., 2001}).

The volumetric proportions of respective data for each food item were joined for the calculation of the niche amplitude and feeding overlap. The niche amplitude has been calculated following the Lewis Index (^{Krebs, 1999}), by using the Formula 1:

where: B represents the niche breadth, Pi the volumetric proportion of the food item i compounding the diet and n is the total number of food items of the diet.

The diet similarity was calculated based on the niche overlap index of ^{Pianka (1974)} Formula 2:

where: Ojk is niche overlap between species j and k, Pij is the proportion of the resource i used by species j and Pik is the proportion of resource i used by species k and n is the total number of resources categories.

There were measured nine morphologic attributes related to feeding: intestine length, jaw size, mouth width, head length, head height, body height, eye diameter, teeth number and gill rakers number. By using these morphometric data it was possible to build matrices using the Euclidian distance and a Mantel test of correlation was used to verify the distance of the morphology with the similar use of feeding resources.

3 Results

A total number of 582 individuals was collected for this analysis, with the great majority being represented by H. strigaticeps. Within the stomach contents of the four armored catfishes were found nine food categories: coarse sediment, fine sediment mixed with organic matter, fungi hifa, diatom algae (Navicula, Pimularia, Nitzschia, Melosira and Synedra), green algae (Closterium, Cosmarium, Ankistrodesmus, Micrasterias, Scenedesmus, Oedogonium and Clorela), blue algae (Oscilatoria and Lyngbia), Tecamoeba shells; vegetal remains (parts of autochthonous and alochthonous plants) and invertebrates. The ANOSIM test did not reveal significant differences for the feeding pattern of individuals belonging to the four species during the period (R = 1.167, p > 0.05).

The niche amplitude analyses for the four Hypostomus species showed the index to be larger for H. albopunctatus (Table 1) influenced by a higher consumption of plant material and invertebrate parts (Table 2).

During this period the Index of Pianka analyses presented niche high values of overlapping (up to 0.50) among those species, as the higher observed overlapping values occurred between H. regani and H. strigaticeps and between H. ancistroides and H. strigaticeps. The lowest overlapping value was, otherwise, observed between H. albopunctatus and H. ancistroides (Table 3).

The comparison result among the trophic similarity matrices and the trophic morphology, the Mantel test, indicated the existence of an interspecific significant correlation (r = 0.6532; p < 0.05) among the diets and the morphometric attributes (Table 4) related to the feeding of the four Hypostomus species.

4 Discussion

The four studied species show that their food items are predominantly constituted by particulate organic material, and may suggest that all of them can be classified as detritivores. A similar pattern has been discussed to occur by ^{Uieda (1984)}; ^{Arcifa and Meschiatti (1993)}; ^{Castro and Caramaschi (2003)}; ^{Hahn et al. (2004)}; ^{Casatti et al. (2005)}; ^{Brandão-Gonçalves et al. (2010)}; ^{Mazzoni et al. (2010)}; ^{Silva et al. (2012)}, as an elevated consume of sediment and alloctonous and autochthonous organic matter mainly of vegetal material predominated. Instead of this, ^{Delariva and Agostinho (2001)} found a more varied diet, which could be classified as omnivore. This may reveal the existence of a relative plasticity for the feeding behavior of Hypostomus species (^{Buck and Sazima, 1995}). Most species of this genus, however, do not use to select their food, by ingesting what is more available when they grasp the substrate (^{Delariva and Agostinho, 2001}), which depends on ingesting a relative great amount of items to get a satisfactory energetic amount.

The lithologic constitution of the river may be related to this elevated consume of sediment, as the Corumbataí river uses to get an consistent deposition of sediment (^{Viadana, 1993}). In an association to this fact, loricariids use to grasp the substrate, which contains the organic matter mixed to the sediment. In the South American systems there happens a clear dominance of fish species, which use to ingest a great amount of sediments and play the role of the basic consumers (^{Bowen, 1983}; ^{Fugi, 1993}; ^{Gerking, 1994}). Detritivory is known to be a common feeding habit for Loricariidae species and relates to a morphologic adaptation of their digestive system, such as the mouth form and position and the length of the intestine, as well (^{Agostinho et al., 1997}; ^{Hahn et al., 1997}). Their general aspect permits them to dwell and explore benthonic microhabitats, by grazing on the algae and the associated microfauna (^{Garavello and Garavello, 2004}). Iliophagy is a similar way of feeding upon the substrate, but is commonly accepted for only some species, such as the Prochilodontidae and Curimatidae species, which use to ingest sediment to absorb the organic material from them. So, in spite of ingesting sediments, Loricariidae differ from other substrate feeders in neotropical waters. In any case, such as iliophagous species, the Loricariidae also constitute and important part of the trophic web in these systems (^{Alvin and Peret, 2004}), in respect of the whole fish fauna.

By statistically comparing the food items volumetric frequencies added to the elevated values for the feeding niche overlapping it may be observed homogeneity related to the whole consummation of food organisms. Some fish display definite diets in result of their anatomic and physiologic adaptations. On the other hand, the feeding efficiency may regard the subtle differences among species when they, in fact, exist. Fish relative sizes may also play some role to these aspects, but could not be perceived by these results. Loricariidae are known by their highly specialized way of living in rapid water streams, presenting depressed bodies, staying almost immobilized and using their suctorial discs around their mouths (^{Casatti and Castro, 2006}; ^{Ferreira, 2007}).

The ingestion of Tecamoeba may also be related to a sediment swallowing, as such organisms use to be associated to such kind of sediment. This grazing and grasping of the substrate confirms that such species find their food on or inside the substrate (^{Ross, 1986}). So, a similar resource may be available and partitioned by several species, though they may be able to explore other ones, as well (^{Hahn et al., 2004}; ^{Merona and Rankin-de-Merona, 2004}; ^{Novakowski et al., 2008}).

Overlapping measurements may determine the resources segregation degree among the species, but may also furnish a descriptive measure of the population’s organization (^{Correa et al., 2011}). The higher abundance of H. strigaticeps in this system may also be another fact to explain the similarity between the diets of the Hypostomus species, in the high Corumbataí, as one may think that there may not exist any system, which should permit the abundance of several species belonging to a same genus, presenting such overlapping diets. This may suggest that some species may display a better ability in relation to the other ones by performing a broader distribution in a same area.

Though the existence of an elevated diet similarity among the species, a peculiarity presented by H. albopunctatus, which ingested more insects and vegetal material, represented mainly by dipterans (Simulidae), which use to occur in rapid streams fixed to rocks (^{Borror and Delong, 1964}) very near to the margins (personal observation). This species may differ from the other ones by using its feeding site in distinct places, which is common to happen (^{Cabral, 2000}). Distinct microhabitats may permit an easier coexistence between species (^{Declerky et al., 2002}). In places where food overlapping uses to occur kinds of food and space may be recognized as the main axes as similar foods may occur in distinct microhabitats. Food and space are recognized as the main axes of resources to a niche division in vertebrates (^{Schoener, 1974}), however ^{Chase and Leibold (2003)} state that related species living in a same community may present some type of differentiation (temporal or spatial) for the occupation of microhabitats. Some cases of spatial segregation of fishes presenting high overlapping indices for their feeding may occur as a consequence of historical differences between them in innate ability for habitat use and the absence of a species in a microhabitat (^{Edds et al., 2002}). Differences of feeding tactics may lead during the evolutionary history of the species to a reduction of the effect for alimentary overlapping (^{Sabino and Correa e Castro, 1990}). By considering that species belonging to a same genus use to show similarities in their general characters, ^{Mazzoni et al. (2012)} suppose that other characteristics which were not measured or perceived in this study may help to explain this feeding patterns and species coexistence.

The significant correlation between these similarities, both to the diets and the morphologic aspects, may play a particular role when such species are compared to the other ones belonging to the community (^{Labropoulou and Eleftheriou, 1997}). So, these species do not present distinct specializations for their diets and trophic morphology. However, the higher abundance of H. strigaticeps suggest that this specie may display a better ability in relation to the other ones by performing a broader distribution in a same area and a different intake insect larvae and plant material in H. albopunctatus diet indicate differences in local and how this species may be exploring their food compared to the others species of Hypostomus in this lotic system.