Differences in ecomorphology and trophic niche segregation of two sympatric heptapterids (Teleostei: Siluriformes)

Morphological similarity, resource sharing, and differences in habitat use by species are factors that favor their coexistence. The objective of this study was to test possible differences in ecomorphology and diet composition of two Heptapterids (Imparfinis mirini and Cetopsorhamdia iherengi) to identify patterns related to resource use. Samplings were carried out in ten streams in Southern Brazil and 123 individuals were caught. A total of 21 ecomorphological indices were calculated for each individual and the volumetric and occurrence methods were used to quantify stomach contents. Both species presented significant differences in some ecomorphological traits, mainly related to foraging behavior. Even though both species were considered insectivorous, the consumed feeding resources differed between them. Cetopsorhamdia iheringi diet was dominated by Simuliidae and terrestrial Coleoptera larvae, while Imparfinis mirini, presented a diet dominated mainly by Trichoptera larvae, Ephemeroptera nymph, and Annelida. Although ecomorphological patterns cannot be used as an absolute factor to explain diet variations, they provide relevant information about how species share resources. These mechanisms allow us to obtain important subsidies for the conservation and management of freshwater ecosystems since they provide an effective understanding on the interactions that occur between the species.

The Heptapteridae fishes are small to medium-sized species and represent one of the largest radiations of Neotropical catfish, being endemic to this region (Bockmann & Guazzelli, 2003). Such family stands out within the Siluriformes because they occupy a wide range of habitats, from small streams to large rivers, of clear or dark waters, cold and with medium -to fast -flowing waters (Ota, Deprá, Graça, & Pavanelli, 2018). Due to similar morphology, related habits are expected for heptapterids (Bockmann & Guazzelli, 2003). The closely related species Imparfinis mirini Haseman 1991 and Cetopsorhamdia iheringi Schubart, Gomes 1959 are abundant species in streams of the Pirapó River Basin (Upper Paraná River basin) (Ota et al., 2018). Both species have an elongated body with a terminal mouth, inhabit the bottom under substrates and their diets are composed of aquatic insects (Casatti, 2002;Bonato, Delariva, & Silva, 2012;Ota et al., 2018).
Although closely related species often have similar feeding habits, niche partitioning has been inferred as a valuable mechanism to facilitate their coexistence in the Neotropical region (Prado et al., 2016;Pease et al., 2018). This mechanism involves feeding on distinct items and consequently reduction of niche overlap to avoid competitive exclusion (Schoener, 1974;Ross, 1986;Portella, Lobón-Cerviá, Manna, Bergallo, & Mazzoni, 2017). In this context, considering that dissimilarities about morphology and habitat use can cause a small niche overlap (Hutchinson, 1959;Kokkoris, Jansen, Loreau, & Troumbis, 2002;Fowler, Lessard, & Sanders, 2014), we hypothesized that Imparfinis mirini and Cetopsorhamdia iheringi exhibit ecomorphological differences that are linked to differences in habitat use and food behavior. Thus, it is also expected a low niche overlap between these species, allowing the coexistence between them.

Study area
The study was performed in ten streams (1 st to 3 rd order), belonging to the Pirapó River Basin located in the Northern region of Paraná State, Southern Brazil (between 22º30' and 23º30'S; 51º15' and 52º 15'W). The Pirapó River Basin is one of the main tributaries at the left margin of the Paranapanema River, belonging to the Upper Paraná River Basin (Cunico, Ferreira, Agostinho, Beaumord, & Fernandes, 2012) ( Figure 1). The dominant landscape in the basin is composed of agricultural activities and urban development, with the city of Maringá being the most important urban center in the region, with approximately 423,666 people (Cunico et al., 2012;Instituto Brasileiro de Geografia e Estatística [IBGE], 2019). The rainfall levels recorded annually are over 1,000 mm, with average annual temperatures between 16 and 20°C, with January being the hottest and most humid month and July the coldest and driest (Passos, 2007). Acta Scientiarum. Biological Sciences, v. 42, e49835, 2020

Sampling design
Sampling was carried out in April and May 2017 (dry season), using electrofishing in an area of approximately 30 meters. To maximize the sampling effort, each stream was blocked downstream with a 5 mm mesh net. The captured specimens were anesthetized with 100 mg L -1 Benzocaine solution until loss of equilibrium and remained immersed in the solution for at least 10 minutes after stopping opercular movements. Subsequently, they were fixed in 10% formalin and labeled for posterior analysis and measurements. Sampled fish were taken to the Núcleo de Pesquisa em Limnologia, Ictiologia e Aquicultura (Nupélia), of Universidade Estadual de Maringá (UEM), and specimens were deposited in the Ichthyological collection of Nupelia/UEM. Vouchers: Cetopsorhamdia iheringi (NUP 20045, 20057, 20097, 20101, 20109 and 20119) and Imparfinis mirini (NUP 20075, 20089, 20095, 20105, 20109 and 20123). The catches were taken under the Instituto Chico Mendes de Conservação da Biodiversidade (ICMBIO) License nº. 25560-1.

Morphological measures
Twenty-two morphological measurements were performed related to the body, fins, head, and mouth of each individual. These morphological measures are involved in habitat use and feeding (Gatz Jr., 1979;Watson & Balon, 1984;Wikramanayake, 1990;Oliveira et al., 2010). Morphological measurements were taken on the left side of each individual, using a digital caliper (accuracy of 0.01 mm). Measurements of fins and eye areas were obtained from the outline drawings of the structures, which were later digitized to calculate the internal area using the software ImageJ (Rueden et al., 2017). Only adult individuals were measured to eliminate the effects of ontogenetic change caused by allometric growth. Possible variations generated by differences in individual length were solved with the use of ecomorphological indices, since these indices minimize the effect of individual length, thus negating the likelihood of the analyzes being dominated by the individual's length (Winemiller, 1991;Prado et al., 2016). For each species, the mean of the linear morphological measure and area were calculated so that the ecomorphological indices were later obtained. After the measurements, 21 indices were calculated to determine the ecological attributes (Table 1) (Wikramanayake, 1990;Breda, Oliveira, & Goulart, 2005;Oliveira et al., 2010;Prado et al., 2016).  (Gatz Jr., 1979;Watson & Balon, 1984).

Body depression index BDI= BMH MBH -1
Low values are associated with fish that exploit habitats with a rapid flow of water because the depressed body helps the fish to remain in the water column without having to swim (Watson & Balon, 1984).
Relative length of the head RLH = HL SL -1 Higher relative head length values are found in fish that feed on larger prey, thus higher rates are expected for piscivorous species (Watson & Balon, 1984;Winemiller 1991;Willis et al., 2005). Relative height of the head RHH = HH MBH -1 Higher relative values of head height are found in fish that feed on larger prey, thus higher rates are expected for piscivorous species (Winemiller, 1991;Willis et al., 2005). Relative width of the head RWH = HW MBW -1 Higher relative values of head width are found in fish that feed on larger prey, thus higher rates are expected for piscivorous species (Winemiller, 1991;Willis et al., 2005). Relative height of the mouth RHM = MH MBH -1 The relative height of the mouth allows inferring about the relative size of the prey (Gatz Jr., 1979). Relative width of the mouth RWM = MW MBW -1 Index related to mouth size, suggesting relatively large prey for high indexes (Gatz Jr., 1979, Winemiller, 1991.
Vertical position of the eye VPE = EHRH HH -1 This index is associated with the foraging position of the species in the water column.
Relative area of the eye RAE = EA (SL) -2 This index is related to food detection and provides information on the visual acuity of the species. It can indicate the position of the species in the water column, as species that inhabit deeper areas have smaller eyes (Gatz Jr., 1979).
Ratio-aspect of the caudal fin RACF = (CFH) 2 CFA -1 Fishes with high aspect ratio caudal fins are more active and continuous swimmers in which there is a tendency to bifurcate of caudal and reduce their area (Gatz Jr., 1979) Relative area of the anal fin RAAFi = AFA (SL) -2 Larger relative areas imply greater maneuverability and stabilization of movement (Breda et al., 2005) Ratio-aspect of the anal fin RAAF = (AFL) 2 AFA -1 Larger aspect-ratios imply a greater ability to perform faster progressive and retrograde movements (Breda et al., 2005) Relative area of the pectoral fin RAPeFi = PFAr (SL) -2 Relatively larger pectoral fin areas are found in slow-swimming species, which use it for maneuverability (some Characidae) and may also be high among fishes that inhabit high current regions such as Siluriformes. Smaller areas are found in pelagic fish (Watson & Balon, 1984).
Reason-aspect of the pectoral fin RAPeF = (PFLe) 2 PFAr -1 A high ratio indicates long and narrow pectoral fin, which is expected in fish that swim continuously and reach high speed, and consequently prefer pelagic regions (Oliveira et al., 2010). Relative area of the pelvic fin RAPFi = PFA (SL) -2 Relatively larger pelvic fin areas are found in benthic fish and smaller areas in pelagic fish (Breda et al., 2005) Ratio-aspect of the pelvic fin RAPF = (PFL) 2 PFA -1 The largest values of pelvic fin aspect ratio are found in pelagic fish and are related to the ability to balance. The lowest values are associated with fish that prefer rocky habitats to support the body to the substrate (Gatz Jr., 1979).

Diet analysis
In the laboratory, stomachs were removed and only those with content were used in the diet analysis. To determine the diet composition, stomach contents were analyzed under a stereomicroscope and food resources were identified to the lowest practical taxonomic level using standard taxonomic keys. To quantify the diet, food items were quantified by the frequency of occurrence and by volumetric method, in which the volume of each food resource was measured using a millimeter Petri dish, obtained in mm³ and later transformed into mL (Hellawell & Abel, 1971;Hyslop, 1980).

Data analysis
Differences in ecomorphology between species were evaluated by a multivariate permutation analysis of variance (PERMANOVA; Anderson, 2001). The pseudo-F statistic resulting from this analysis was tested by the Monte Carlo method using 999 randomizations. The ordering of the species according to their ecomorphological characteristics was evaluated by a Principal Component Analysis (PCA) performed using the correlation matrix of the ecomorphological proportions. This analysis indicates the grouping pattern of individuals with similar morphological traits. The axes were selected according to the Broken-Stick criterion, in which those with eigenvalues greater than the axes generated by the model were used for interpretation (McCune & Mefford, 1999).
Differences in diet between species were also evaluated with PERMANOVA. The pseudo-F statistic resulting from this analysis was tested by the Monte Carlo method using 999 randomizations. The ordering of the species according to their diet composition was visualized using a principal coordinate analysis (PCoA) based on a Bray-Curtis dissimilarity matrix (Bray & Curtis, 1957). Niche overlap between species was calculated using the Pianka index: where Ojk = measure of Pianka food overlap between species j and species k; pij = proportion of food item i in the total of items used by species j; pik = proportion of food item i in the total of items used by species k, n = total number of food items. Pxi and Pyi are the proportions (volume) of the food resource i used by species x and y, and n is the total number of feeding resources. The results of interspecific niche overlap were arbitrarily considered: high (> 0.6), intermediate (0.4-0.6) or low (<0.4) (Grossman, 1986). All analyses were performed in R (R Core Team, 2017) using the Vegan package 1.17 (Oksanen et al., 2018), except for the niche overlap analysis which was performed in EcoSim version 6.18 (Gotelli & Entsminger, 2001). Graphs were built in the software Statistica 12.5 (Statsoft Inc., 2011).

Results
A total of 123 individuals were measured, but only 58 stomachs of Imparfinis mirini and 29 stomachs of Cetopsorhamdia iheringi contained food and were analyzed.
The PERMANOVA showed significant difference in ecomorphological attributes between the species (pseudo-F(1.121) = 15.18; p > 0.001). The PCA revealed the significance of the first two axes according to the Broken-Stick criterion. Together, axes 1 and 2 (Table 2) explained 38.68% and were used to characterize the morphological diversity between species (Figure 2). When analyzing the first axis of PCA, it is observed that C. iheringi is located on the negative side presenting lower values for relative area of the pectoral fin (RAPeFi), relative area of the anal fin (RAAFi), relative area of the pelvic fin (RAPFi), reason-aspect of the pectoral fin, relative length of the head (RLH) and relative length of the caudal peduncle (RLCP). Imparfinis mirini is located on the positive side presenting higher values for relative height of the mouth (RHM), relative width of the mouth (RWM) and ratio-aspect of the anal fin (RAAF). When evaluating the dietary composition of C. iheringi and I. mirini, it is observed that both species consumed predominantly aquatic food resources and Trichoptera larvae, Ephemeroptera nymph and Chironomidae larvae occurred in most of the analyzed stomachs ( Figure 3A and B). However, some preferences were observed. Out of 11 food resources recorded in the diet of C. iheringi, Simuliidae larvae was highly consumed (22.6% of volume and 27.6% of occurrence) and it was an exclusive food resource to this species. Although the allochthonous feeding resources were less representative ( Figure 3C and D), terrestrial Coleoptera larvae were the second most consumed by C. iheringi (21.84% of volume and 27.6% of occurrence), followed by the Trichoptera larvae (18, 2% of volume and 41.4% of occurrence). Out of 17 feeding resources consumed by of I. mirini, Trichoptera larvae dominated the diet, composing 54.9% of total volume and occurring in 76.8% of the analyzed stomachs, followed by Ephemeroptera nymph (9.9 % of the volume and 30.3% of occurrence), and Annelida (8.6% of the volume and 5.2% of occurrence). This latter and six other resources were consumed exclusively by I. mirini ( Figure 3A and 3B).   PERMANOVA also showed a significant difference in diet composition between the two species (pseudo-F(1.83) = 4.31; p < 0.05) (Figure 4) and ordination analysis evidenced separation between the species, mainly Acta Scientiarum. Biological Sciences, v. 42, e49835, 2020 on axis 1: C. iheringi located on the negative values and I. mirini located in the positive values. Together, axes 1 and 2 explained 30.2% of the variation (Figure 4). The niche overlap was considered intermediate (0.54).

Discussion
The results showed that despite Imparfinis mirini and Cetopsorhamdia iheringi are phylogenetically close and occupy the same stratum in the environment and are considered insectivorous species, they present differences in morphology and the use of food resources. Also, there is a low overlap, indicating low competition for food resources. Ecomorphological differences, mainly related to foraging, swimming, and habitat occupation favor the coexistence between these species. Correlations between diet and ecomorphology in fish are constantly observed (Neves, Delariva, & Wolff, 2015;Pease et al., 2018;Baldasso et al., 2019), which supports the idea that morphology acts as one of the mechanisms responsible for the variation of diet. In fact, sympatric species tend to differ in feeding preferences, segregating resources to avoid competitive exclusion (Hardin, 1960;Schoener, 1974).
In an attempt to explain the preference of fish for a particular resource, some factors may be directly related, such as abundance and nutritional quality, as well as their morphology and behavior. However, when we observe the diet of benthic fish, these factors in the predator-prey relationship are not clear, since the heterogeneity and complexity of habitat in the benthic region can act as a physical barrier during the foraging of benthic, preventing many preys from being consumed by fish (Hill & Grossman, 1987;Gillette, 2012, Guglielmetti, Silva, Higuti, & Fugi, 2019.' Several studies have shown that fish belonging to benthic invertivorous guild have a very specific diet and that the use of a determinate resource is mainly attributed to trophic morphology and predation foraging (Rolla, Esteves, & Silva, 2009;Brejão, Gerhard, & Zuanon, 2013;Neves et al., 2015). In our study, morphological characteristics such as head, mouth, and fins sizes allowed both species to feed on benthic organisms, while the terminal mouth and the laterally compressed body possibly allowed the expansion of the consumption of other resources available in the environment (Pease et al., 2018), as the allochthonous items present in the diet of both species such as terrestrial Coleoptera larvae, plants, adult Coleoptera, Isoptera, adult Trichoptera, and adult Diptera.
Despite the similarities, our results showed differences in morphology between the species that are probably related to different feeding preferences. The indices that influenced Cetopsorhamdia iheringi dispersal in ordination are associated with fish that prefer benthic and high water flow regions, presenting slow swimming and movement stability (Watson & Balon, 1984;Breda et al., 2005). In addition, the smaller head length indicates that C. iheringi feeds on smaller prey (Pouilly, Lino, Bretenoux, & Rosales, 2003;Oliveira et al., 2010), as is the case of Simuliidae larvae that attach to the substrates or plant surfaces (Kiel, 2001) and was found in abundance and exclusivity in their diet. However, the indices that influenced the dispersion of Imparfinis mirini in the ordination are associated with fish that have faster progressive and retrograde movements (Breda et al., 2005) and consequently act more efficiently in the exploration of complex structurally habitats (Prado et al., 2016). In addition, I. mirini presented a greater length and width of the mouth, suggesting that species with these characteristics feed on relatively large prey (Gatz Jr., 1979;Winemiller, 1991), such as Trichoptera larvae consumption, item of great abundance in the diet this species. This high abundance in the diet can be explained because this aquatic larvae has low mobility, which makes its escape capacity lower and the larvae became more vulnerable to predation (Tófoli, Alves, Higuti, Cunico, & Hahn, 2013). In addition, since it is a relatively large invertebrate, it makes the larvae more visible to the predator (Tófoli et al., 2013).
The niche overlap value indicates a low potential for competitive interaction (Zaret & Rand, 1971). Although we have not measured food availability in this study, several studies point out that diet divergences may reduce competition when resources are limited (Hutchinson, 1957;Colloca, Carpentieri, Balestri, & Ardizzone, 2010;Guo et al., 2014;Carniatto, Fugi, & Thomaz, 2017). Besides, it should be considered that the variety and abundance of resources consumed by the species must be mediated by their morphological characteristics (Prado et al., 2016), where one species is able to explore resources that other is not (Casatti, 2002;Freitas, Montag, & Barthem, 2017;Pagotto et al., 2011). In this case, a strong competitive potential would not be expected between C. iheringi and I. mirini, since both species occupy different morphological niches.
Another important ecological factor that can promote coexistence between species is the preference for some resources (Novakowski, Fugi, & Hahn, 2004;Tófoli et al., 2013;Walker, Kluender, Inebnit, & Adams, 2013;Bison et al., 2015). The differential use of resources, even partially, is an essential component of the trophic interactions within a community (Schoener, 1974;Portella et al., 2017;Gračan, Zavodnik, Krstinić, Dragičević, & Lazar, 2016). Trophic varieties observed in species analyzed here highlight interspecific differences in the use of food items. This specialization in different resources probably allows segregation of the niche (Prado et al., 2016;Portella et al., 2017). Trichoptera was the most abundant item in I. mirini diet and the third most abundant in C. iheringi diet, which reflects the great abundance of these invertebrates in the Pirapó basin streams (Guglielmetti et al., 2019) and makes them easily predated mainly by I. mirini, which forages the substrate since it uses the head barbs to find and capture the prey (Casatti & Castro, 2006). This feature may also explain the large number of unique items found in I. mirini diet. On the other hand, the high abundance of terrestrial Coleoptera suggests that C. iheringi uses both benthic and pelagic region, consuming drifting resources (Guglielmetti et al., 2019). Thus, the prey origin and the different use of fine-scale habitat contribute to the trophic segregation between these species (Barros, Zuanon, & Deus, 2017;Richardson, Lyons, Roby, Cushing, & Lerczak, 2018).

Conclusion
In short, our results showed ecomorphological differences between the sympatric species, and that such variations may be determinant in diet differences. Although both species are insectivorous, they present particular preferences, resulting in the low trophic overlap. The ecomorphological and diet relationship of species enable us to understand how they perform their ecological niches (Peres-Neto, 1999;Severo-Neto, Teresa, & Froehlich, 2015). Thus, these factors seem to be essential to coexistence between these species in these environments. Studying the ecology of these species provides subsidies for the management and conservation of these ecosystems that are under constant anthropogenic effects.