The effects of season and ontogeny in the diet of Piabarchus stramineus

1908


Introduction
Fish developed diversified foraging tactics and strategies, allowing them to use a wide range of food resources available in their environments and its surroundings (Brandão-Gonçalves et al., 2009), i.e., autochthonous and allochthonous resources.They can occupy all trophic levels in the food web (Wootton, 1992), so knowing about the diet of a species may allow recognizing different trophic guilds, make inferences on its structure, importance of the different trophic levels and interrelationships amongst its components.Knowing the diet range of fish through the stomach content analysis may also help the interpretation of its dynamics and habitat occupation.
Shifts in the diet of fish may be a result of temporal and/or seasonal resource availability (Abelha et al., 2006;Wolff et al., 2009), but also of natural changes in foraging areas due to flood and drought cycles (Wootton, 1999).Also, shifts in the diet of fish might be related to ontogenetic variations, which have been the object of broad discussion in the literature (Machado-Allison & Garcia, 1986;Winemiller, 1989;Hahn et al., 2000).In tropical environments, most fish species show broad flexibility in feeding habits (Lowe-McConnell, 1999), which may change during development according to morphological and behavioral changes (Villares-Junior, 2009;Baldasso et al., 2019).
Palavras-chave: insetos alóctones; classes de tamanho; alimentação; amplitude de nicho; sobreposição alimentar.and that such resources are strongly related to the presence of marginal riparian forests (Costa, 1987;Sabino & Castro, 1990;Graciolli et al., 2003;Lampert et al., 2003).Thus, the present work aims to describe and analyze the diet of P. stramineus from the Ibicuí River in southern Brazil, identifying 1) the origin of food resources (if allochthonous or autochthonous); 2) the diet composition, niche breadth and food overlap; 3) possible differences in the diet amongst the different seasons of the year and amongst different size classes (ontogeny).Also, a discussion on the importance of ciliary forest in the diet of this species and other fish is provided.

Data collection
Monthly samplings were conducted from April 2001 to February 2002 in the Ibicuí River (29° 50'14" S; 54° 47'53" W), Uruguay River Basin, in the city of Cacequi, RS, Brazil (Figure 1).The Ibicuí River is a large river with brown water due to high concentration of suspended sediment.Marginal riparian forest forms a narrow line on the left margin, while in the right margin the vegetation is composed essentially by grass and only a few bushes and trees.This vegetation is described as a savannah and a steppe savannah whose arboreal components' distribution follows the drainage system (Teixeira et al., 1986).Depth close to the left margin can reach over 2 m deep, while in the right margin there are shallows due to sandbanks, especially during the low water season.The surrounding areas are covered by pasture and agriculture.
Sampling was performed using seine net (three hauls; 10x2 m net size; 5 mm mesh size), close to the margins.The specimens were anesthetized with Eugenol (clove oil; 70 mg/L) and then euthanized and fixed in 10% formalin solution.In the laboratory, fish were transferred to 70% ethanol and identified to the species level.Voucher specimens were cataloged in the ichthyological collection of Universidade Federal do Rio Grande do Sul (UFRGS 6693).
Specimens were measured (standard length -SL; mm), weighed (total weight; g), and dissected for stomach extrusion and weighing.
Comparative analysis of P. stramineus's diet, according to standard length classes (SLC), was carried out to determine the influence of body size on feeding habits.The SLC were defined through the Sturges method (Vieira, 1991), with the specimens being divided into three SLC, as follows: SLC1 (n = 113) less than 28 mm, SLC2 (n = 136) 28.01-37 mm, SLC3 (n = 81) larger than 37 mm.
Stomach content analysis were conducted through the methods of frequency of occurrence (Fo, the number of stomachs with a particular food category in relation to the total number of stomachs) (Hynes, 1950); and volumetric frequency (Vo, estimated according to the quantitative contribution of each item as determined by the number of squares occupied by each item on a scaled paper in relation to the total number of squares occupied by all items) (Hyslop, 1980;Dias et al., 2017).
The importance of each food category was defined through the alimentary index (IAi) (Kawakami & Vazzoler, 1980), calculated according to the formula: IAi = [(Fo i x Vo i ) / Σ(Fo i x Vo i )] x 100, where i = food item; Fo = frequency of occurrence; and Vo = volumetric frequency.
The degree of diet specialization between the different SLCs was determined by calculating Levin's index (Krebs, 1989) for trophic niche breadth using Vo values.To standardize the measurement of trophic niche, Hurlbert's formula (Hurlbert, 1978) was calculated, which can range from 0 to 1, as: ) -1]}, where: Ba i = standardized Levin's index for predator i; p ij = proportion of predator i diet that consists of prey j; and n = total number of alimentary items.Trophic niche breadths were then classified as high (> 0.61), intermediary (0.41 to 0.6), or low (< 0.4) (Novakowski et al., 2008).
Diet overlap among SLCs was determined through pairwise-comparisons using the Morisita Index (Krebs, 1989), which can range from 0 to 1, with 0 indicating a lack of diet overlap and 1 indicating complete overlap.According to Amundsen et al. (2003), values equal to or higher than 0.6 indicate a significant diet overlap.
To assess the proportion of food items in the diet, Fo and Vo methods were combined in the graphical analysis proposed by Costello (1990) with Amundsen et al. (1996) modifications.This method consists of a two-dimensional representation of prey abundance values in the y axis, and frequency of occurrence in the x axis.Dots around 100% Fo and Vo indicates dominant items.Dots around 100% Fo and 1% Vo indicates a small consumption of different types of prey, being classified as a generalist.The opposite, with dots around 1% Fo and 100% Vo indicates a specialization towards a given item (Corrêa & Silva, 2010).
Differences in the diet composition among different SLCs and seasons were tested through Permutational Multivariate Analysis of Variance (PERMANOVA) with 999 permutations (Anderson, 2001;Anderson et al., 2008) based on Euclidean dissimilarity (Legendre & Legendre, 1998).This analysis was based on the Vo values of food items and performed with the vegan package (Oksanen et al., 2019) in R software (R Core Team, 2019).A non-metric multidimensional scaling (NMDS) (Borcard et al., 2011) was also performed to reduce the diet composition multidimensionality of all fish and visualize the dispersion between seasons.The NMDS was performed using the Euclidean distance coefficient (Legendre & Legendre, 1998), through the metaMDS function with the vegan package (Oksanen et al., 2019).
High values for allochthonous insects were observed in both Autumn/01 and Winter/01 in all analyzes (FO, Vo, IAi).In Spring/01 and Summer/01, high values for microcrustaceans, vegetal matter, and autochthonous items were more often observed.These results are shown in the Table 2.
The graphic method of Costello calculated for each SLC showed that Crustacea and autochthonous insects were the main food category for fish from SLC1, followed closely by allochthonous insects (Figure 2a).An increase in the importance of allochthonous insects can be observed from SLC2 to SLC3.Autochthonous insects and vegetal matter were important categories for both SLCs, while a decrease in the importance of Crustacea can be observed from SLC2 to SLC3 (Figure 2b-2c).
Diet overlap was observed among all size classes when Vo values were considered (SLC1 x SLC2 = 0.9629; SLC1 x SLC3 = 0.6444; SLC2 x SLC3 = 0.7834).Diet overlap was lower among SLC1 and SLC3 and among SLC2 and SLC3, respectively.The effects of season and ontogeny… Acta Limnologica Brasiliensia, 2022, vol. 34, e31 The two-way Permanova identified a significant variation (F = 3.7; p = 0.001) in diet composition when the interaction between seasons and SLCs was analyzed.Thus, the temporal variation in the diet is not the same in the different SLCs.
In addition, the NMDS showed that the two axes (extracted after 20 interactions with a stress of 0.14 and a linear fit of R 2 = 0.97) summarized the variation of the diet composition between the individual fish.The ordination evidenced distinct diets among seasons, with autumn and winter positioned in the extreme ends of the ordination space, followed by spring (Figure 3).

Discussion
The results of the present study allowed the classification of the studied population of P. stramineus as insectivorous, which feed on insects from both allochthonous and autochthonous origins, but also including microcrustacea and vegetal matter.Different sized fish showed slight differences in their diets as evidenced by the low niche breadth and low food overlap when comparing smaller and larger individuals.
According to Casatti & Castro (1998), P. stramineus is a diurnal, nektonic species which occupy marginal riffle areas during the day, swimming in the middle water and surface (Uieda, 1984), and preying on drifting items in the water surface [drift-feeding, cf.Grant & Noakes (1987)], including terrestrial insects [surface picking, cf.Sazima (1986)].Casatti & Castro (1998) defined P. stramineus as an omnivorous species in the headwater riffles of São Francisco River, due to the consumption of autochthonous and allochthonous insects, vegetal matter, algae, and microcrustaceans.Luiz et al. (1998) has also classified this species as an omnivore in two streams of Paraná River Basin, feeding on Chironomidae, vegetal remains and detritus in similar proportions.In Corumbá Reservoir, Goiás, Brazil, P. stramineus was also classified as omnivore, consuming autochthonous (67%) and allochthonous (28%) insects, and small amounts of vegetal matter (Luz-Agostinho et al., 2006).Such studies corroborate at least in part the findings of the present study and differences may be associated to the lower proportion of microcrustacea and vegetal matter found for P. stramineus from the Ibicuí River.In the Upper Paraná River floodplains, this species was considered as insectivorous, feeding mainly on Chironomidae (Luz et al., 2001).Casatti et al. (2003) recorded a predominance of autochthonous insects (54.2%), represented mainly by larvae of Diptera; allochthonous items (43.2%) were also present in the diet of the species, being adult Diptera the most frequent.These results are in accordance with the present study, especially in respect to the predominance of Diptera (allochthonous and autochthonous) in the diet of the species.As can be observed from the different works previously presented here, the feeding habits of this small fish encompass allochthonous and autochthonous items, either from animal and vegetal origins, thus evidencing the plasticity found in the diet of the same species in different environments.Such plasticity sometimes makes difficult the definition of trophic groups.Feeding plasticity results from an interaction between the quality and quantity of available food in the environment (Luz et al., 2001).This is a remarkable feature in tropical freshwater fish fauna (Goulding, 1980;Lowe-McConnell, 1999) once species can change food consumption according to changes in food resource availability (Goulding, 1980).In tropical regions, despite some fish species can be classified as specialists in a particular food type, most species exhibit trophic plasticity in their diets (Lowe-McConnell, 1999), making the definition of trophic patterns difficult.
Differences in the diet were found in the present study, and such differences can be observed both among SLCs as well as among seasons and they indicate that food consumption change diversely over time for each size class.Hence, the temporal variation is not maintained throughout its ontogeny.Such finding suggests that each size class may have unique morphological and/or behavioral traits to minimize competition, which favor specific temporal changes for each class.Seasonal changes in the diet of fish may be attributed to food resource availability (Abelha et al., 2006;Wolff et al., 2009), and/or to natural changes in the habitat/foraging areas (Wootton, 1999).The feeding habits of P. stramineus was described in different streams of Guiraí River Sub-basin in Mato Grosso do Sul (Brandão-Gonçalves et al., 2009) and differed between seasons and sampling sites.However, in all sites the diet was basically composed by terrestrial insects.
Other authors recorded high consumption of allochthonous insects especially during the rainy season (Brandão-Gonçalves et al., 2009).According to these authors, the flood season is the main period of feeding, growth, and fat reserve accumulation.
The flood invades the riverbanks and fertilizes the area, which becomes nutrient-rich due to organic decomposition, stimulating the growth of microorganisms and macroinvertebrates (insects, crustaceans, mollusks) used as food by fish (Lowe-McConnel, 1999).
Another study corroborates our findings regarding diet versus fish size.Suiberto et al. (2009) observed that small P. stramineus (larvae) fed mainly on microcrustaceans (Cladocera and Copepoda) while juvenile showed a transition to larger prey such as larvae of Diptera, suggesting a tendency towards invertivory.All main resources in the diet of larvae and juvenile P. stramineus were autochthonous, similar to the present study.This is in accordance with the pattern that in their early life stages, many fish species prey upon phytoplankton, zooplankton or small macroinvertebrates, but may switch to larger macroinvertebrates, fish, plants or detritus later in their development (Nunn et al., 2012;Huss et al., 2013).
As discussed above, depending on the study, the diet of adults may vary between autochthonous and allochthonous resources, often including both, but usually composed of terrestrial insects, with which our results agree.In many fish species, the prey size usually increases with fish size (Morinière et al., 2003;Sánchez-Hernández & Cobo, 2012), and different size classes typically consume different prey types as a result of, for example, differences in foraging abilities or habitat use (Mittelbach & Persson, 1998;Nunn et al., 2012).Fish undergo ontogenetic dietary shifts during their development (Blanco-Garrido et al., 2003;Fochetti et al., 2008).These shifts during life stage transitions may be accompanied by a marked reduction in intraspecific competition within the fish population, facilitating resource partitioning (Amundsen et al., 2003;Oscoz et al., 2006).Coexistence among fish species can be better understood through studies on resource partitioning.Eventually, even when shared food resources are abundant, small differences in diet may facilitate species coexistence (Knickle & Rose, 2014).The same may occur to different life cycle phases within the same species, leading to profound ecological consequences for fish; for example, enhancing individual growth and reproductive output or reducing the risk of mortality (Sánchez-Hernández et al., 2019).Sometimes, even small variations in diet can facilitate coexistence among species (Hynes, 1970;Gatz Junior, 1979).Therefore, differences in item consumption found between the SLCs of P. stramineus are not sufficient to change their general diet pattern.On the other hand, they may represent a contribution in resource partitioning and species coexistence in their different life cycle phases.
Allochthonous food items are strongly related to marginal forests (Alvim & Peret, 2004), whose importance for fish feeding is well documented in the literature (Angermeier & Karr, 1983;Sabino & Castro, 1990;Sabino & Zuanon, 1998;Castro, 1999;Lowe-McConnell, 1999).The conservation of such environments may guarantee a broad variety of food items supply to the fish fauna (Lowe-McConnell, 1999).Despite some studies have found a higher contribution of autochthonous items in the diet of fish (Moyle & Senayake, 1984;Uieda et al., 1997;Casatti, 2002), these items are also dependent on organic matter and nutrients inputs from riparian vegetation, the basis of trophic chain in stream ecosystems (Gregory et al., 1991), which increase the importance of riparian areas and their conservation for aquatic communities (Angermeier & Karr, 1983).
Considering the results of our study together with data found in the literature, we believe that P. stramineus shows high plasticity of food item consumption allowing resource partitioning within the studied population, especially between the different size classes.The high frequencies of allochthonous resources identified in the stomach content analysis corroborates the importance of riparian vegetation as source of food items, as evidenced by many previous studies.

Figure 1 .
Figure 1.Sampling site location (star) in the Ibicuí River, city of Cacequi, Rio Grande do Sul state, southern Brazil.

Figure 3 .
Figure 3. Graphical representation of the two first NMDS axis showing seasonal variation in the diet of Piabarchus stramineus in the Ibicuí River, southern Brazil, between April 2001 and February 2002.Aut = autumn; Win = winter; Spr = spring; Sum = summer.

Table 1 .
Frequency of occurrence (Fo%), Volumetric frequency (Vo%), Alimentary index (Iai%), of the food items and categories found in the diet of Piabarchus stramineus from the Ibicuí River, southern Brazil, between April 2001 and February 2002.A = adult; L = larvae; N = nymph; Allo Ins = Allochthonous insect; Au Ins = Autochthonous insect; VM = vegetal matter.Values of 50% and over are highlighted (bold).

Table 2 .
Frequency of occurrence (Fo%), Volumetric frequency (Vo%), Alimentary index (IAi%), by seasons of the food items and categories found in the diet of different size classes (SLC) of Piabarchus stramineus from the Ibicuí River, southern Brazil, between April 2001 and March 2002.Aut = autumn; Win = winter; Spr = spring; Sum = summer; A = adult; L = larvae; N = nymph; Allo Ins = Allochthonous insect; Au Ins = Autochthonous insect; VM = vegetal matter.Values of 50% and over are highlighted (bold).