Fish As Ecological Tools to Complement Biodiversity Inventories of Benthic Macroinvertebrates

Hydrobiologia DOI 10.1007/s10750-011-0747-8

PRIMARY RESEARCH PAPER

Fish as ecological tools to complement biodiversity inventories of benthic macroinvertebrates

Daniel M. Maroneze • Taynan H. Tupinamba´s • Carlos B. M. Alves • Fa´bio Vieira • Paulo S. Pompeu • Marcos Callisto

Received: 3 October 2010 / Revised: 2 May 2011 / Accepted: 4 May 2011 Ó Springer Science+Business Media B.V. 2011

Abstract Sampling benthic macroinvertebrates in macroinvertebrates as a food resource for fishes, the large rivers has several limitations, arising not sampling restrictions in sediment collection done by only from the selectivity of traditional sampling dredges, and the importance of surveying benthos gears but also from difficulty in capturing organisms biodiversity, the objective of this study was to that inhabit the deeper zones and high current evaluate the stomach contents of five commonly- velocities. Considering the importance of benthic occurring invertivorous fish species as a means of complementing a benthic macroinvertebrate inventory. Three sampling campaigns (fish and benthic Handling editor: Luiz Carlos Gomes macroinvertebrate) were conducted in a reach of the Electronic supplementary material The online version of Araguari River (Minas Gerais, Brazil), approximately this article (doi:10.1007/s10750-011-0747-8) contains 9 km long and 90 m wide. Astyanax altiparanae supplementary material, which is available to authorized users. Garutti & Britski 2000, Leporinus friderici (Bloch 1794), Leporinus amblyrhynchus Garavello & Britski D. M. Maroneze T. H. Tupinamba´s M. Callisto (&) Laborato´rio de Ecologia de Bentos, Departamento de 1987, Iheringichthys labrosus (Lu¨tken 1874) and Biologia Geral, Instituto de Cieˆncias Biolo´gicas, Pimelodus maculatus Lacepe`de 1803 were the fish Universidade Federal de Minas Gerais, species collected and studied. To determine benthic Av. Antoˆnio Carlos, 6627, C.P. 486, Pampulha, macroinvertebrate taxonomic richness, a total of 54 CEP 30161-970 Belo Horizonte, MG, Brazil e-mail: [email protected] Van Veen sediment samples were obtained. We URL: http://www.biogeral.icb.ufmg.br/benthos compared lists of the benthic taxa found in fish stomachs with those from the sediment samples. C. B. M. Alves P. S. Pompeu M. Callisto The differences in the taxonomic composition of the Nuvelhas/Projeto Manuelzaõ, Universidade Federal de Minas Gerais, Av. Antoˆnio Carlos, 6627, Pampulha, benthic macroinvertebrate communities between the CEP 31270-901 Belo Horizonte, MG, Brazil sediment samples and each fish species stomachs contents were assessed through NMDS and ANOSIM F. Vieira analyses, using a Sorensen similarity index with the Centro de Transposiçaõ de Peixes, Universidade Federal de Minas Gerais, Ageˆncia Shopping Del Rey, C. P. 4011, presence/absence of taxa data. Independent of sam- CEP 31250-970 Belo Horizonte, MG, Brazil pling period, additional benthic macroinvertebrate families or classes were provided by identifying fish P. S. Pompeu stomach contents. We found a total of 30 taxa in this Departamento de Biologia, Universidade Federal de Lavras, Campus Universita´rio, CEP 37200-000 Lavras, study, including 5 unique taxa (or 17% of the total) in MG, Brazil the sediment samples, 9 unique taxa (30%) in the 123 Hydrobiologia stomach samples, and 16 taxa (53%) common to condition at a site (Junqueira et al., 2000; Karr & both. The NMDS and ANOSIM analyses showed a Chu, 2000; Buss & Borges, 2008) or for taxonomic significant separation between Van Veen sediment inventories that seek to provide a list of common samples and two fish species stomach contents— species in a specific area (Mackey et al., 1984). L. amblyrhynchus and P. maculatus. These results Recent studies have shown that analysis of stomach indicate that fish can be used as additional samplers contents of benthophagous fishes can be an important and are an efficient method to complement the strategy in taxonomic surveys (Russo et al., 2002; benthic taxonomic inventory obtained through tradi- Tupinamba´s et al., 2007; Santos et al., 2009) because tional sediment sampling techniques in large areas, as fish exploit a wide variety of microhabitats and river segments and catchments. capture macroinvertebrates in areas that are difficult to sample with conventional gears. Their feeding Keywords Stomach contents Samplers behaviors may make fish useful complementary Benthos Large rivers Trophic interactions benthos samplers (Callisto et al., 2002; Galina & Hahn, 2004). To evaluate the importance of fish stomach Introduction analyses for taxonomic inventories of benthic macroinvertebrate communities, we analyzed stomach Benthic macroinvertebrate sampling techniques are contents of five common invertivorous species highly varied. In each type of aquatic habitat, differing (Dura˜es et al., 2001; Callisto et al., 2002): Astyanax methods are used because of differing depths, sediment altiparanae Garutti & Britski 2000, Leporinus frid- textures, currents, and spatial heterogeneities of the area erici (Bloch 1794), Leporinus amblyrhynchus Garav- (Alves & Strixino, 2003; Blockson & Flotemersch, ello & Britski 1987, Iheringichthys labrosus (Lu¨tken 2005; Kikuchi et al., 2006). In large rivers, Surber 1874) and Pimelodus maculatus Lacepe`de 1803. We samplers or dredges (i.e., Eckman-Birge, Van Veen, hypothesized that benthic macroinvertebrate families Petersen, etc.) are commonly used (Bournaud et al., and classes (operational taxonomic units, OTU) 1998; Rempel et al., 2000; Gayraud et al., 2003;Moreno found in fish stomachs would differ from those found et al., 2009), and because the sampling area is known, in sediment samples because of fish foraging strate- such methods aid quantitative analysis of benthic gies. In addition, we evaluated whether fish stomach- macroinvertebrate communities (Bady et al., 2005). content analysis could effectively complement taxo- Comprehensive quantitative samples require nomic surveys conducted with conventional benthos substantial time and funds to obtain and identify sampling gears. (Rosenberg et al., 1997; Bartsch et al., 1998). Also, such samples are performed in specific habitats and as a consequence, some of the benthos unique to other Materials and methods habitats are not adequately sampled (Lenat, 1988). These limitations often occur in large rivers where We conducted the study in a reach of the Araguari sampling of benthic macroinvertebrates is concen- River (Brazil) between 18°200S–46°000W and trated in shallow areas near river margins (Reece & 20°10’S–48°50’W. The reach was 9 km long and Richardson, 2000). Sampling in the deeper zones or approximately 90 m wide, drains mainly sandstone areas of strong current normally is not conducted and basalt lithology, and in its deepest valley erodes because of access and safety concerns for the sampling granite and gneiss (Rodrigues, 2002). The local crew, and the need to use equipment such as suction climate is characterized by a rainy period from pumps and SCUBA divers (Rosenberg et al., 1997)or October to March and a dry period from May to artificial substrates that require multiple site visits September. The average annual temperature is 22°C (Humphries et al., 1998; Collier et al., 2009). with a total rainfall of 1,555 mm/year (Rosa et al., Depending on the purpose of the research, qual- 2006). In general, the river water is well oxygenated, itative samples are sufficient. This is the case for slightly acidic, and low in electrical conductivity, water quality biomonitoring programs that convert turbidity, and total dissolved nutrients (Maroneze data to a biotic index score to determine assemblage et al., 2011). 123 Hydrobiologia

We sampled in July 2005, October 2005, and June We used the same procedure to evaluate and identify 2008 at six sites along the 9-km reach averaging the organisms present in fish stomach contents. We 1.5 km apart. Fishes were caught in 10 m 9 1.7 m compared lists of the benthic OTU found in fish gill nets with mesh sizes 3, 4, 5, 6, 7, 8, 10, 12, 14, stomachs with those from the sediment samples. To and 16 cm between opposite knots. At each site, we eliminate the influence of the unequal number of fished a total of 170 m2 of gill nets from dusk to the stomachs analyzed and to determine which fish following morning for approximately 14 h. The species contributed most to the inventory, we also collected fish were fixed with 10% formalin. In the calculated the average number of exclusive taxa per laboratory, the fish were soaked in water, transferred stomach (NT) through use of the following formula: o nx to alcohol 70 GL, identified, weighed, and measured. NT ¼ Nx, where nx is the number of taxa found in the The number of stomachs removed and evaluated stomach contents of species x and not found in the varied between species, size, and month (Table 1). sediment samples, and Nx is the number of stomachs The standard length (cm) of the smallest sampled fish of species x analyzed. The greater the value of NT, in advanced maturation was used to separate the the greater the contribution of that fish species toward juveniles from adults (Vono et al., 2002; Froese & adding benthic OTU to the inventory. Pauly, 2010). Empty stomachs were excluded from The differences in the taxonomic composition of analysis. the benthic macroinvertebrate communities between To evaluate the composition of benthic macroin- the sediment samples and each fish species stomachs vertebrate communities, we took three Van Veen contents were assessed through a non-metric dimen- (0.45 m2) sediment samples from the river margins in sional scaling (NMDS) ordination technique, using a each of the fish sampling stations. The material was Sorensen similarity index with the presence/absence only collected along the margins because of the rocky of taxa data. The statistical significance of the group- nature of the central axis of the river. A total of 54 ings (sediment samples vs. fish stomachs contents) Van Veen sample units were obtained (6 sites 9 3 were tested through similarity analysis (ANOSIM) samples/site 9 3 visits/site). In the laboratory, we (a = 0.05). The ANOSIM provides a value for washed the sediment samples through sieves (1.00- interpretation where R [ 0.25 indicates separated and 0.50-mm mesh) and identified all specimens to groups (Clarke & Green, 1988; Michelland et al. family or class (OTU) through the use of a stereo- 2010). A randomization process using Monte Carlo microscope, and according to Perez (1988) and Costa testing with 10,000 interactions was conducted to et al. (2006). The individuals identified were arthro- validate the R values observed. A value of P \ 0.05 pods and annelids with a head or soft-part mollusks. indicated that the R value observed was not randomly Empty shells, eggs, exoskeletons, pupae, and other obtained. Analyses of NMDS and ANOSIM were fragments of macroinvertebrates were not considered. performed using PRIMER 6.0 software.

Table 1 Number of stomachs analyzed for each ﬁsh species sampled downstream of the Amador Aguiar I Dam (Araguari River, Minas Gerais, Brazil) Family/species Jul/05 Oct/05 Jun/08 Total Juveniles Adults Juveniles Adults Juveniles Adults

Characidae Astyanax altiparanae –15–1153061 Anostomidae Leporinus amblyrhynchus 32391119 Leporinus friderici 3 3 1 7 14 1 29 Pimelodidae Iheringichthys labrosus –4–3–815 Pimelodus maculatus 1212 78 131 Total 43 43 69 155

123 Hydrobiologia

Results L. friderici (Table 1); however, a few contained at least a single benthic organism (5 and 27%, respectively), In total, 30 operational taxonomic units (OTU) of precluding statistical analysis. Seeds and fruits were benthic macroinvertebrates were identified in this the main food items identified in the stomachs of those study. The organisms included Mollusca (Bivalvia, two species. Gastropoda), Annelida (Oligochaeta, Hirudinea), and Considering the total of 30 OTU collected in the Arthropoda (Collembola, Hydracarina, Insecta), three sampling campaigns (July 2005, October 2005, including 24 Insecta families (Table 2). We found and June 2008), we found 5 unique OTU (or 17% of 21 OTU in the sediment samples: 15 in July 2005, 11 the total) in the sediment samples (Hirudinea, Coen- in October 2005, and 15 in June 2008. In the 155 fish agrionidae, Veliidae, Odontoceridae, and Dolichopo- stomachs examined, we found 26 OTU: 14 in July didae), 9 unique OTU (30%) in the stomach s 2005, 19 in October 2005, and 14 in June 2008. The amples (Collembola, Polymitarcyidae, Glossosomati- fish stomach contents added to the richness of the dae, Helicopsychidae, Leptoceridae, Philopotamidae, sediment samples by 5 OTU in July 2005, 10 OTU in Corydalidae, Empididae, and Simuliidae), and 16 October 2005, and 5 OTU in June 2008. OTU (53%) common to both (Bivalvia, Gastropoda, Leporinus amblyrhynchus, Iheringichthys labro- Oligochaeta, Hydracarina, Baetidae, Leptohyphidae, sus, and Pimelodus maculatus consumed the greatest Leptophlebiidae, Gomphidae, Libellulidae, Elmidae, variety of benthic macroinvertebrates, with 8, 7 and 6 Hydropsychidae, Hydroptilidae, Polycentropodidae, OTU identified in their stomachs in July 2005; 15, 7 Ceratopogonidae, Chironomidae, and Tipulidae). and 10 OTU in October 2005; and 6, 6 and 6 OTU in June 2008, respectively (Table 2). In terms of adding OTU to those already found in the sediment samples, Discussion L. amblyrhynchus was the most important, adding 7 OTU to the October 2005 samples. Both I. labrosus Benthic macroinvertebrates are common in the diet of and P. maculatus added at least 1 OTU per sampling many freshwater fish species (Gerking, 1994; Russo period. In July 2005, no benthic macroinvertebrates et al., 2002; Hahn & Fugi, 2007; Pinto & Uieda, 2007). were identified in Astyanax altiparanae and Lepori- In the present study, Leporinus amblyrhynchus, nus friderici stomachs. However, in October 2005, I. labrosus, and P. maculatus stomachs contained the Simuliidae were found only in A. altiparanae, and in greatest OTU richness, and their diverse diets based June 2008, Leptoceridae were found only in L. mainly on insect larvae have been reported by others friderici. A similar pattern was observed when the (Dura˜es et al., 2001; Callisto et al., 2002; Andrade & effects of uneven samples, resulting from different Braga, 2005; Fagundes et al., 2008). Chironomidae numbers of analyzed stomachs, were removed. In this larvae were the most commonly occurring benthic case, L. amblyrhynchus, I. labrosus, and P. maculatus macroinvertebrates in the fish stomachs, especially in contained the greatest average number of exclusive L. amblyrhynchus and I. labrosus—in which they were OTU per stomach (NT) (Table 3) and, therefore, present in 100% of stomachs analyzed. The high contributed most to the taxonomic survey. consumption of Chironomidae is common among fish The NMDS and similarity analyses showed a because these insects are very abundant and available significant separation between Van Veen sediment in freshwater (Mendonça et al., 2004). Moreover, samples and two fish species stomachs contents. This Armitage (1995) suggested that the preference of distinction was observed for L. amblyrhynchus bottom-feeding fish for Chironomidae larvae and (ANOSIM; R = 0.61; P \ 0.01) (Fig. 1A) and pupae as a food source may be related to their high P. maculatus (ANOSIM; R = 0.47; P \ 0.01) protein content and high digestibility. (Fig. 1B). Although I. labrosus added at least one Similar to our results, other taxonomic groups (no OTU per sampling period, in general, the composition Chironomidae) have also been identified as food of benthic macroinvertebrates communities found in items. Ceratopogonidae, Simuliidae, and Empididae their stomachs was quite similar to that identified in are abundant in L. amblyrhynchus stomachs as sediment samples (ANOSIM; P [ 0.05) (Fig. 1C). We reported by Dura˜es et al. (2001) and Mendonça evaluated the most stomachs from A. altiparanae and et al. (2004). Gastropoda, Gomphidae, Libellulidae, 123 Hydrobiologia Table 2 Taxa of benthic macroinvertebrates found in the sediment samples and in the stomachs of five fish species

Taxa July 2005 October 2005 June 2008

SE Ap Lf La Il Pm ES EX FI SE Ap Lf La Il Pm ES EX FI SE Ap Lf La Il Pm ES EX FI

Mollusca Bivalvia x x x x x x x x x x x x x x x Gastropoda x x x x x x x x Annelida Oligochaeta x x x x x x x x x x x x Hirudinea x x x x Arthropoda Arachnida Hydracarina x x x x x x x x x x x x x x x Entognatha Collembola xxxxx Insecta Ephemeroptera Baetidae x x x x x x x x x x x x x x Leptohyphidae x x x x x x x x x x x x x x x Leptophlebiidae x x x x x x x x Polymitarcyidae x x x x xxxx Odonata Coenagrionidae xx Gomphidae x x x x x x x x Libellulidae x x x x x x x x x Heteroptera Veliidae xx Coleoptera Elmidae x x x x x x Trichoptera Glossosomatidae xxxx Helicopsychidae xxxxxxxx Hydropsychidae x x x x x x x x x x x x x Hydroptilidae x x x x x x x x x x x x x Leptoceridae xx xx x x x x x 123 Odontoceridae xx Philopotamidae x x x x x Polycentropodidae xxxx Hydrobiologia

Polymitarcyidae, Leptohyphidae, Odontoceridae, and Polycentropodidae have been observed in the stom-

taxa found achs of I. labrosus (Abes et al., 2001; Teixeira & Bennemann, 2007; Fagundes et al., 2008). Bivalvia, ,EX Hirudinea, Oligochaeta, Culicidae, and Chaoboridae have been observed in the stomachs of P. maculatus (Lolis & Andrian, 1996; Lima-Juńior & Goitein, 2004; Andrade & Braga, 2005; Silva et al., 2007) (Appendix A—Supplementary Material). Three of the sampled fish species have morphological characteristics that aid obtaining food from the taxa found in fish stomachs bottom. The relatively large eyes of L. amblyrhyn- ES , chus aid its perception of aquatic larval movement (Mendonça et al., 2004). The gill raker spacing of I. labrosus facilitate separating prey from detritus (Fugi et al., 2001). The maxillary barbells of P. maculatus have a great number of cutaneous taste buds (Rotta,

Pimelodus maculatus 2003) which aid it in detecting prey in the sediment (Lima-Ju´nior & Goitein, 2003). Neither A. altiparanae ,Pm nor L. friderici are specialized to prey on benthic macroinvertebrates (Gomiero & Braga, 2003; Melo & Ro¨pke, 2004; Bennemann et al., 2005), instead they appear to be generalist and opportunistic feeders, consuming several items from the water column (Casatti, 2002; Melo & Ro¨pke, 2004). Therefore, Il Iheringichthys labrosus

, drifting organisms may be captured. Although an abundant food item found in fish stomachs may merely reflect prey availability (Gee, 1989), several fish species consume insects groups that are not sampled by gears or are less frequent in the substrate. This pattern was observed in temperate (Howe et al., 2008) and tropical (Pinto & Uieda, 2007) ecosystems. Theoretically, if a fish species shows a high selectivity for some benthic macroinvertebrate taxa (e.g. Ephemeroptera, Diptera, or Odonata), then it could be used as a specific-taxa sampler. However, further investigations are necessary to demonstrate such selectivity. Strauss (1979)

Lf Leporinus friderici, La Leporinus amblyrhynchus mentioned two serious problems that should be , overcome in food selection studies: 1. obtain an

ﬁnal inventory considering stomach contents and sediment samples unbiased sample from the habitat that accurately

,FI represents the relative abundances of the various SE Ap Lf La Il Pm ES EX FI SE Ap Lf La Il Pm ES EX FIpotential SE Ap Lf prey La Ilas Pm ES they EX FI are encountered by the predator, and 2. obtain an unbiased sample that accurately represents the relative abundances of the Ap Astyanax altiparanae prey as they are consumed; otherwise, the results are continued not reliable. CorydalidaeCeratopogonidaeChironomidae xDolichopodidaeEmpididae x xSimuliidaeTipulidae x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x

sediment, We found 9 OTU (30% of the total) exclusively in Megaloptera Diptera Total 15 0 0 8 7 6 14 5 20 11 2 2 15 7 10 19 10 21 15 2 2 6 6 6 14 5 20 Table 2 Taxa July 2005 October 2005 June 2008 SE exclusively in ﬁsh stomachs ﬁsh stomachs. Van Veen sediment samples and two 123 Hydrobiologia

Table 3 Average number of exclusive operational taxonomic units (OTU) per stomach for each ﬁsh species sampled downstream of the Amador Aguiar I Dam (Araguari River, Minas Gerais, Brazil) Sampling set Fish species A. altiparanae L. friderici L. amblyrhynchus I. labrosus P. maculatus

July 2005 0 0 0.60 0.25 0.15 October 2005 0.18 0 0.87 0.66 0.55 June 2008 0 0.06 1.50 0.12 0.22

fish species stomach contents represented distinct Vidotto-Magnoni & Carvalho (2009) reported that groups, indicating that taxonomic composition of the surface feeders capture large amounts of Epheme- benthic macroinvertebrate communities in the sam- roptera sub-imagos, probably during emergence. ples collected by these two methods differed. This Other benthophagous taxa, such as Trichomycterus, distinction was observed for L. amblyrhynchus and live in interstices between coarse gravel (Chara et al., P. maculatus indicating that both species demonstrate 2006), rocks (Barreto & Aranha, 2005), or leaf packs high potential to be used as complementary samplers (Abilhoa et al., 2008). Such locations are also the of benthic macroinvertebrates. Similar to this study, habitats of most benthic macroinvertebrates (Russo Callisto et al. (2002) identified 21 genera of Chiro- et al., 2002). In taxonomic inventories, it is advan- nomidae in the stomachs of L. amblyrhynchus and tageous to use sampling techniques which provide a P. maculatus near Miranda Reservoir, also Rio larger list of benthic macroinvertebrate taxa. The data Araguari, noting that only 5 of them (25% of the presented in this study suggest that the stomach total) were present in sediment samples. Many large analysis of fish species with different foraging rivers have abundant habitats that support distinct strategies is a proper way to achieve this aim. benthic macroinvertebrate assemblages (Collier et al., In England, Webster & Hart (2004) demonstrated 2009). High-velocity areas with rocky substrates are that after a 24-h period without food, Gasterosteus colonized by organisms with morphological adapta- aculeatus individuals foraged preferentially in com- tions (i.e., dorsally flattened bodies) for attachment plex substrates (5–20 mm gravel) versus simple ones and current resistance (Merritt & Cummins, 1996). (sandy particles smaller than 2 mm). According to Low-velocity areas with sandy substrates are colo- those authors, a predator optimizes its energy gains nized by burrowing species (Baptista et al., 2001). by foraging in patches where there are the maximum For these heterogeneous rivers, with non-uniform of predator–prey encounter rates. This species pre- distribution of species among sites, an intensive sumably associates complex substrates with poten- sampling effort that covers all habitat diversity (e.g., tially higher prey density and diversity. Based on riffles, pools, woody debris, and macrophytes) is previous foraging experience, individuals have the required to accurately measure species richness ability to learn and to discriminate foraging habitats (Flotemersch et al., 2010). with different prey densities (Webster & Hart, 2006). Despite the scarcity of data concerning the forag- Considering that riverine habitat is heterogeneous, ing strategies of L. amblyrhynchus, P. maculatus, and with discontinuously distributed benthic organisms I. labrosus, it is reasonable to assume that commonly among substrate types (e.g., sand, clay, gravel, occurring benthophagous fishes feed in different stones, and litter patches) (Ligeiro et al., 2010), the microhabitats (Casatti, 2002). In other words, they ability to recall experiences and recognize areas present distinct feeding tactics (Vidotto-Magnoni & based on their food productivity is an important Carvalho, 2009). In a Neotropical stream, Casatti evolutionary advantage to a foraging fish (Webster & (2002) observed that Rhamdia quelen searches for Hart, 2004). Thus, fish may explore biologically invertebrates in deep crevices between rocks without richest habitats that a priori are unknown by the stirring the substrate, whereas Corydoras aeneus human sampling crew. forages in shallow margins where it stirs the sandy In Brazil, experimental tests also revealed that sediment and captures prey organisms by suction. species of Anostomidae (Leporinus reinhardti) and 123 Hydrobiologia

benthophagous fishes (e.g., trophic morphology, foraging modes, feeding habitat preferences, learn- ing ability, and swimming performance) aid invertebrate capture in areas that are more difficult than others for humans to sample in large rivers (e.g., deep and swiftly-moving waters snags, and aquatic macrophytes). Commonly, such environments are characterized by distinct benthos assemblages (Cremona et al., 2008, 2009). Considering that lotic ecosystems can be studied on several spatial scales, from centimeters to kilometers, we emphasize that this sampling technique is particularly useful to complement biodiversity inventories in large areas as river segments and catchments. At smaller spatial scales (single habitats), punctual Van Veen sediment samples can be more appropriate since benthofagous fishes search for preys in a wide number of habitats which can be different from the areas where they are captured by gill nets. It is noteworthy that this study used two levels of taxonomic resolution (Class or Family) and is a qualitative survey (presence/absence). Family identification appears to be adequate for bio-assessments and reduces taxonomic identification effort (Melo, 2005; Corbi & Trivinho-Strixino, 2006). In Brazil, species-level identification is difficult because the taxonomic knowledge of several benthic macroinvertebrate groups is still incomplete (Buss & Vitorino, 2010). We did not account for and did not compare the number of benthic individuals in both types of samples because our purpose was to evaluate stomach contents of fish species as a means of complementing a traditional, qualitative benthic macroinvertebrate survey (obtained by Van Veen grab). However, we recognize that when comparing two sampling methods, it is important to assess which is the most efficient. To do this, future studies must evaluate richness relative to sampling effort. Various researchers have proved that species richness is strongly dependent on sample size and, therefore, approaches Fig. 1 Two-dimensional NMDS ordination based on the which compare different sample units may produce presence/absence of benthic macroinvertebrate taxa identified in Van Veen sediment samples and fish species stomach unreliable conclusions (Li et al., 2001; Cao et al., contents. (A) Leporinus amblyrhynchus.(B) Pimelodus mac- 2002; Melo et al., 2003). As reported by Buss & ulatus.(C) Iheringichthys labrosus Borges (2008), an alternative is the use of rarefaction, calculated as the number of species expected to be Pimelodidae (P. maculatus) are strong swimmers found in different samples as if they had the same and, therefore, in a position to occupy habitats number of individuals. However, Cao et al. (2007) with high current velocities (Santos et al., 2007, reported that rarefaction can be highly biased 2008). Theoretically, all these characteristics of when assemblages differed in species distribution 123 Hydrobiologia

(as suggested by results of this study), and they balancing the increased sampling effort against logis- recommended using mean taxonomic similarity tical and funding limitations (Hughes & Peck, 2008). scores calculated from replicate resampling of the According to Flotemersch et al. (2010), this is data. Thus, care should be taken in the extrapolation achieved by employing multiple gear types, targeting of our results considering its qualitative nature. specific macrohabitats, increasing the number of In addition, we have three suggestions for future samples, or increasing site extent. We found that fish investigators: (1) Compare the amount of time stomach contents represent an alternative sampling required for collecting, processing, and identifying method. In the Araguari River, the stomachs’ contents both types of samples (cost/effectiveness); (2) Cap- of fish collected through use of gill nets (especially, ture fish species according to their feeding activity L. amblyrhynchus and P. maculatus) proved comple- which can be estimated from the quantity of food in mentary to the taxonomic inventory obtained through stomachs during a 24-h cycle. For L. amblyrhynchus, sediment sampling by Van Veen grabs. In this data from Rio Corumba´ (Goia´s, Brazil) indicated that context, we emphasize that our results were positive mean stomach fullness was maximum (25–75%) using a high taxonomic level identification (class and during twilight periods (Mendonça et al., 2004). family). Probably, benthic organisms identified to Intensifying the sampling effort at such times may species (currently, not possible in Brazil) would produce more benthic macroinvertebrate prey in fish provide a significantly more informative description. stomachs; and (3) Evaluate the adequacy of sediment In other words, the differences in benthic macroin- sampling (number of sites and Van Veen grabs per vertebrate community composition between sediment site) and stomach-content sampling (e.g., number of samples and fish stomachs contents could be even sites, number of nets per site, number, and species of more pronounced, reinforcing our hypotheses that fish per site). some fish species are really good complementary In this survey, our sediment sampling regime was samplers of benthic macroinvertebrates. similar to that which we employed in other ecological Despite the limitations reported, we believe that studies (we took three sediment samples per site in this qualitative method is a valid contribution, shallow areas near shore). However, Flotemersch particularly in Brazilian large rivers. The benthic et al. (2006) reported that a greater level of sampling macroinvertebrate communities of these ecosystems effort is necessary for adequately assessing benthic are poorly studied and have been continuously macroinvertebrate taxa richness in large rivers. subjected to human pressures (e.g., dams, navigation, Unlike wadeable streams, where all the habitats are water pollution, and siltation) for centuries. Cur- generally accessible to sampling, as system get wider, rently, such activities present themselves as major its physical area and habitat diversity increase and, threats to freshwater biodiversity (Agostinho et al., therefore, more sampling effort (site length, number 2005). Dudgeon et al. (2006) reported that there are of samples, and alternate gears) is required to capture no available data for most invertebrate groups in rare species (Flotemersch et al., 2010). It is known tropical freshwaters; however, they believe that these that rare species have a positive influence on richness ecosystems support highly endemic and diverse estimates (Gotelli & Colwell, 2001). In practical fauna, including different species of crustaceans, terms, this means that field-sampling designs could mollusks, and insects. If trends in human demands for also consider fishes as possible samplers of rare water remain unaltered, then the loss of yet unknown macroinvertebrate taxa. taxa will become a reality. In this context, this In conclusion, benthic macroinvertebrate sampling qualitative method improves the lists of taxa, con- is selective; each method is best in a specific habitat, tributing to the knowledge of benthic macroinverte- and consequently captures only the organisms directly brate biodiversity. associated with that habitat (Humphries et al., 1998). This renders comprehensive surveys of river benthos Acknowledgements The authors thank the Conso´rcio Capim difficult, limiting the inventory of benthic biodiver- Branco Energia for financing the project, the Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico (CNPq) for the sity. To overcome this limitation, researchers are fellowship to the first author, and the U.S. Fish & Wildlife advised to maximize the diversity and distribution Service, CAPES, and FAPEMIG for funding this study. The of microhabitats sampled at multiple sites while authors also thank Clarissa Chalub for the help in dissecting fish; 123 Hydrobiologia all their colleagues at the Laborato´rio de Ecologia de Bentos Astyanax (Characidae) em diferentes rios da bacia do rio (UFMG)—especially Rener Grego´rio and Juliana S. França— Tibagi, Parana´, Brasil. Iheringia: Se´rie Zoologia 95: for sample processing; Prof. Gilmar B. Santos (PUC-MG), Prof. 247–254. Jose´ F. Bezerra-Neto (UFMG), and Prof. Jose´ Francisco Blockson, K. A. & J. E. Flotemersch, 2005. Comparison of Gonçalves-Juńior (UNB) for numerous suggestions; and Prof. macroinvertebrate sampling methods for nonwadeable Robert M. Hughes (Amnis Opes Institute and Oregon State streams. Environmental Monitoring and Assessment 102: University) for his valuable comments and English editing, and 243–262. review. This article was written while MC was a sabbatical Bournaud, M., H. Tachet, A. Berly & B. Cellot, 1998. visitor (CAPES fellowship No. 4959/09-4) at the IMAR, Importance of microhabitat characteristics in the macro- Universidade de Coimbra, Portugal. benthos microdistribution of a large river reach. Annales de Limnologie 34: 83–98. Buss, D. F. & E. L. Borges, 2008. Application of rapid bioassessment protocols (RBP) for benthic macroinvertebrates in Brazil: comparison between sampling techniques and mesh sizes. Neotropical Entomology 37: 288–295. References Buss, D. F. & A. Vitorino, 2010. Rapid bioassessment protocols using benthic macroinvertebrates in Brazil: evalua- Abes, S. S., A. A. Agostinho, E. K. Okada & L. C. Gomes, tion of taxonomic sufficiency. Journal of the North 2001. Diet of Iheringichthys labrosus (Pimelodidae, American Benthological Society 29: 562–571. Siluriformes) in the Itaipu Reservoir, Parana´ River, Bra- Callisto, M., V. Vono, F. A. R. Barbosa & S. M. Santeiro, zil-Paraguay. Brazilian Archives of Biology and Tech- 2002. Chironomidae as a food source resource for Lepo- nology 44: 101–105. rinus amblyrhynchus (Teleostei: Characiformes) and Pi- Abilhoa, V., L. F. Duboc & D. P. A. Filho, 2008. A comun- melodus maculatus (Teleostei: Siluriformes) in a Brazilian idade de peixes de um riacho de Floresta com Arauca´ria, reservoir. Lundiana 3: 67–73. alto rio Iguaçu, sul do Brasil. Revista Brasileira de Zoo- Cao, Y., D. P. Larsen, R. M. Hughes, P. L. Angermeier & T. logia 25: 238–246. M. Patton, 2002. Sampling effort affects multivariate Agostinho, A. A., S. M. Thomaz & L. C. Gomes, 2005. Con- comparisons of stream assemblages. Journal of the North servation of the biodiversity of Brazil’s Inland Waters. American Benthological Society 21: 701–714. Conservation Biology 19: 646–652. Cao, Y., C. P. Hawkins, D. P. Larsen & J. Y. Sickle, 2007. Alves, R. G. & G. Strixino, 2003. The sampling of benthic Effects of sample standardization on mean species macroinvertebrates using two different methods: waiting detectabilities and estimates of relative differences in trays and Ekman collector. Acta Limnologica Brasiliensia species richness among assemblages. The American 15: 1–6. Naturalist 170: 381–395. Andrade, P. M. & F. M. S. Braga, 2005. Diet and feeding of Casatti, L., 2002. Alimentaçaõ dos peixes em um riacho do fish from Grande River, located below the Volta Grande Parque Estadual Morro do Diabo, bacia do alto Parana´, reservoir, MG-SP. Brazilian Journal of Biology 65: sudeste do Brasil. Biota Neotropica 2: 1–14. 377–385. Cassemiro, F. A. S., N. S. Hahn & R. Fugi, 2002. Avaliaçaõda Armitage, P. D., 1995. Chironomidae as food. In Armitage, dieta de Astyanax altiparanae Garutti & Britski, 2000 P. D., P. S. Cranston & L. C. V. Pinder (eds), The Chi- (Osteichthyes, Tetragonopterinae) antes e apo´s a for- ronomidae: the biology and ecology of non-biting midges. maçaõ do reservato´rio de Salto Caxias, Estado do Parana´, Chapman & Hall, New York: 423–435. Brasil. Acta Scientiarum Biological Sciences 24: Bady, P., S. Dodeleć, C. Fesl, S. Gayraud, M. Bacchi & F. 419–425. Scho¨ll, 2005. Use of invertebrate traits for biomonitoring Chara, J. D., D. J. Baird, T. C. Telfer & E. A. Rubio, 2006. of European large rivers: the effects of sampling effort on Feeding ecology and habitat preferences of the catfish genus richness and functional diversity. Freshwater Biol- genus Trichomycterus in low-order streams of the ogy 50: 159–173. Colombian Andes. Journal of Fish Biology 68: Baptista, D. F., D. F. Buss, L. F. M. Droville´ & J. L. Nessi- 1026–1040. mian, 2001. Diversity and habitat preference of aquatic Clarke, K. R. & R. H. Green, 1988. Statistical design and insects along the longitudinal gradient of the Macae´ river analysis for a ‘‘biological effects’’ study. Marine Ecology basin, Rio de Janeiro, Brazil. Revista Brasileira de Bio- Progress Series 46: 213–226. logia 61: 249–258. Collier, K. J., M. Hamer & W. L. Chadderton, 2009. A new Barreto, A. P. & J. M. Aranha, 2005. Assembleía de peixes de substrate for sampling deep river macroinvertebrates. um riacho da Floresta Atlaˆntica: composiçaõ e distribu- New Zealand Natural Sciences 34: 49–61. içaõ espacial (Guaraqueçaba, Parana´, Brasil). Acta Sci- Corbi, J. J. & S. Trivinho-Strixino, 2006. Influence of entiarum Biological Sciences 27: 153–160. taxonomic resolution of stream macroinvertebrate com- Bartsch, L. A., W. B. Richardson & T. J. Naimo, 1998. Sam- munities on the evaluation of different land uses. Acta pling benthic macroinvertebrates in a large floodplain Limnologica Brasiliensia 18: 469–475. river: considerations of study design, sample size and cost. Costa, C., S. Ide & C. E. S. Somonka, 2006. Insetos Imaturos: Environmental Monitoring and Assessment 52: 425–439. metaformose e identificaçaõ. Holos, Ribeiraõ Preto. Bennemann, S. T., A. M. Gealh, M. L. Orsi & L. M. Souza, Cremona, F., D. Planas & M. Lucotte, 2008. Biomass and 2005. Ocorreˆncia e ecologia tro´fica de quatro espećies de composition of macroinvertebrate communities associated 123 Hydrobiologia

with different types of macrophyte architectures and Howe, D. V., R. C. Jordan & F. Juanes, 2008. Selective feeding habitats in a large fluvial lake. Archiv fu¨r Hydrobiologie in a generalist invertivore, age-0 striped bass. Ecology of 171: 119–130. Freshwater Fish 17: 495–501. Dudgeon, D., A. H. Arthington, M. O. Gessner, Z. I. Kawabata, Hughes, R. M. & D. V. Peck, 2008. Acquiring data for large D. J. Knowler, C. Le´veˆque, R. J. Naiman, A. H. Prieur- aquatic resource surveys: the art of compromise among Richard, D. Soto, M. L. J. Stiassny & C. A. Sullivan, science, logistics, and reality. Journal of the North 2006. Freshwater biodiversity: importance, threats, status American Benthological Society 27: 837–859. and conservation challenges. Biological Reviews 81: Humphries, P., J. E. Growns, L. G. Serafini, J. H. Hawking, A. 163–182. J. Chick & P. S. Lake, 1998. Macroinvertebrate sampling Dura˜es, R., P. S. Pompeu & A. L. Godinho, 2001. Alimentaçaõ methods for lowland Australian rivers. Hydrobiologia de quatro espećies de Leporinus (Characiformes, Anos- 364: 209–218. tomidae) durante a formaçaõ de um reservato´rio no Junqueira, V. M., M. C. Amarante, C. F. S. Dias & E. sudeste do Brasil. Iheringia Se´rie Zoologia 90: 183–191. S. França, 2000. Biomonitoramento da qualidade das Fagundes, C. K., E. R. Behr & C. B. Kotzian, 2008. Diet of a´guas da bacia do Alto Rio das Velhas (MG/Brasil) at- Iheringichthys labrosus (Siluriformes, Pimelodidae) in the rave´s de macroinvertebrados. Acta Limnologica Brasil- Ibicuı´ river, southern Brazil. Iheringia Se´rie Zoologia 98: iensia 12: 73–87. 60–65. Karr, J. R. & E. W. Chu, 2000. Sustaining living rivers. Flotemersch, J. E., J. B. Stribling, R. M. Hughes, L. Reynolds, Hydrobiologia 423: 1–14. M. J. Paul & C. Wolter, 2010. Site length for biological Kikuchi, R. M., A. A. Fonseca-Gessner & G. Y. Shimizu, assessments of boatable rivers. River Research and 2006. Suction sampler for collection of benthic macroin- Applications (in press). doi:10.1002/rra.1367. vertebrates in several continental aquatic environments: a Flotemersch, J. E., K. Blocksom, J. J. Jr Hutchens & B. comparative study with the Hess and Surber samplers. C. Autrey, 2006. Development of a standardized large Acta Limnologica Brasiliensia 18: 29–37. river bioassessment protocol (LR-BP) for macroinverte- Lenat, D. R., 1988. Water quality assessment of streams using brate assemblages. River Research and Applications 22: a qualitative collection method for benthic macroinver- 775–790. tebrates. Journal of the North American Benthological Froese, R. & D. Pauly. 2010. FishBase – Word Wide Web Society 7: 222–223. electronic publication. Available in: www.fishbase.org. Li, J., A. T. Herlihy, W. Gerth, P. R. Kaufmann, S. V. Gregory, Accessed 27 August 2010. S. Urquhart & D. P. Larsen, 2001. Variability in stream Fugi, R., A. A. Agostinho & N. S. Hahn, 2001. Trophic mor- macroinvertebrates at multiple spatial scales. Freshwater phology of five benthic-feeding fish species of a tropical Biology 46: 87–97. floodplain. Brazilian Journal of Biology 61: 27–33. Ligeiro, R., A. S. Melo & M. Callisto, 2010. Spatial scale and Galina, A. B. & N. S. Hahn, 2004. Atividade de forrageamento the diversity of macroinvertebrates in a Neotropical de Triportheus spp. (Characidae, Triportheinae) utilizada catchment. Freshwater Biology 55: 424–435. como ferramenta de amostragem da entomofauna, na a´rea Lima-Juńior, S. E. & R. Goitein, 2003. Ontogenetic shifts of a do reservato´rio de Manso, MT. Revista Brasileira de Neotropical catfish, Pimelodus maculatus (Siluriformes, Zoocieˆncias 6: 81–92. Pimelodidae): an ecomorphologic approach. Environ- Gayraud, S., B. Statzner, P. Bady, A. Haybachp, F. Sho¨ll, P. mental Biology of Fishes 68: 73–79. Usseglio-Polatera & M. Bacchi, 2003. Invertebrate traits Lima-Juńior, S. E. & R. Goitein, 2004. Diet and feeding for the biomonitoring of large European rivers: an initial activity of Pimelodus maculatus (Osteichthyes: Pimelod- assessment of alternative metrics. Freshwater Biology 48: idae) in the Piracicaba river (State of Saõ Paulo, Brazil) – 2045–2064. the effect of seasonality. Boletim do Instituto de Pesca 30: Gee, J. M., 1989. An ecological and economic review of 134–140. meiofauna as food for fish. Zoological Journal of Linnean Lolis, A. A. & I. F. Andrian, 1996. AlimentaçaõdePimelodus Society 93: 243–261. maculatus Lace´pede, 1803 (Siluriformes, Pimelodidae) na Gerking, S. D., 1994. Feeding Ecology of Fish. Academic planıćie de inundaçaõ do alto rio Parana´, Brasil. Boletim Press, California. do Instituto de Pesca 23: 187–202. Gomiero, L. M. & F. M. S. Braga, 2003. O lambari Astyanax Luz, K. D. G., F. Abujanra, A. A. Agostinho & L. C. Gomes, altiparanae (Characidae) pode ser um dispersor de semen- 2001. Caracterizaçaõ tro´fica da ictiofauna de treˆs lagoas tes? Acta Scientiarum Biological Sciences 25: 353–360. da planıćie aluvial do alto rio Parana´, Brasil. Acta Sci- Gomiero, L. M. & F. M. S. Braga, 2008. Feeding habitats of entiarum Biological Sciences 23: 401–407. the ichthyofauna in a protected area in the state of Saõ Mackey, A. P., D. A. Cooling & A. D. Berrie, 1984. An Paulo, southeastern Brazil. Biota Neotropica 8: 41–47. evaluation of sampling strategies for qualitative surveys of Gotelli, N. J. & R. K. Colwell, 2001. Quantifying biodiversity: macroinvertebrates in rivers, using pond nets. Journal of procedures and pitfalls in the measurement and compari- Applied Ecology 21: 515–534. son of species richness. Ecology Letters 4: 379–391. Maroneze, D. M., T. H. Tupinamba´s, J. S. França & M. Cal- Hahn, N. S. & R. Fugi, 2007. Alimentaçaõ de peixes em res- listo, 2011. Effects of flow reduction and spillways on the ervato´rios brasileiros: alteraço˜es e consequëˆncias nos composition and structure of benthic macroinvertebrate esta´gios iniciais do represamento. Oecologia Brasiliensis communities in a Brazilian river reach. Brazilian Journal 11: 469–480. of Biology 71(3).

123 Hydrobiologia

Melo, A. S., 2005. Effects of taxonomic and numeric resolution Rosenberg, D. M., I. J. Daves, D. G. Cobb & A. P. Wiens, on the ability to detect ecological patterns at a local scale 1997. Protocols for Measuring Biodiversity: benthic using stream macroinvertebrates. Archiv fu¨r Hydrobiolo- Macroinvertebrates in Fresh Waters. University Crescent, gie 164: 309–323. Manitoba. Melo, C. H. & C. P. Ro¨pke, 2004. Alimentaçaõ e distribuiçaõ Rotta, M. A., 2003. Aspectos gerais da fisiologia e estrutura do de piaus (Pisces, Anostomidae) na planıćie do Bananal, sistema digestivo dos peixes relacionados a` piscicultura. Mato Grosso, Brasil. Revista Brasileira de Zoologia 21: Embrapa Pantanal, Corumba´. 51–56. Russo, M. R., A. Ferreira & R. M. Dias, 2002. Disponibilidade Melo, A. S., R. A. S. Pereira, A. J. Santos, G. J. Shepherd, de invertebrados aqua´ticos para peixes bento´fagos de dois G. Machado, H. F. Medeiros & R. J. Sawaya, 2003. riachos da bacia do rio Iguaçu, estado do Parana´, Brasil. Comparing species richness among assemblages using Acta Scientiarium Biological Sciences 2: 411–417. sample units: why not use extrapolation methods to Santos, H. A., P. S. Pompeu & C. B. Martinez, 2007. Swim- standardize different sample sizes? Oikos 101: 398–410. ming performance of the migratory Neotropical fish Le- Mendonça, F. P., N. S. Hahn & V. E. Loureiro-Crippa, 2004. porinus reinhardti (Characiformes: Anostomidae). Feeding aspects of Leporinus amblyrhynchus (Characi- Neotropical Ichthyology 5: 139–146. formes: Anostomidae) in the first stages of formation of a Santos, H. A., P. S. Pompeu & C. B. Martinez, 2008. Swim- Neotropical reservoir, Brazil. Neotropical Ichthyology 2: ming performance of the freshwater neotropical fish Pi- 145–150. melodus maculatus Lace`pede, 1803. Brazilian Journal of Merritt, R. W. & K. W. Cummins, 1996. An Introduction to the Biology 68: 433–439. Aquatic Insects of North America. Kendall Hunt, Iowa. Santos, A. F. G. N., F. Racca-Filho, L. N. Santos & F. Michelland, R. J., S. Combes, V. Monteils, L. Cauquil, T. G. Arau´jo, 2009. El pez Trachelyopterus striatulus (Si- Gidenne & L. Fortun-Lamothe, 2010. Molecular analysis luriformes: Auchenipteridae) como herramienta de mu- of the bacterial community in digestive tract of rabbit. estreo de la entomofauna en un embalse tropical. Revista Anaerobe 16: 61–65. de Biologia Tropical 57: 1081–1091. Moreno, P., J. S. França, W. R. Ferreira, A. D. Paz, I. Silva, E. L., R. Fugi & N. S. Hahn, 2007. Variaço˜es temporais M. Monteiro & M. Callisto, 2009. Use of the BEAST e ontogene´ticas na dieta de um peixe onı´voro em ambiente model for biomonitoring water quality in a neotropical impactado (reservato´rio) e ambiente natural (baıá) da basin. Hydrobiologia 630: 231–242. bacia do rio Cuiaba´. Acta Scientiarum Biological Sciences Peretti, D. & I. F. Andrian, 2008. Feeding and morphological 29: 387–394. analysis of the digestive tract of four species of fish Strauss, R. E., 1979. Reliability estimates for Ivlev’s Electivity (Astyanax altiparanae, Parauchenipterus galeatus, Ser- Index, the forage ratio, and a proposed linear index of rasalmus marginatus and Hoplias aff. malabaricus) from food selection. Transaction of the American Fisheries the upper Parana´ river floodplain, Brazil. Brazilian Journal Society 108: 344–352. of Biology 68: 671–679. Teixeira, I. & S. T. Bennemann, 2007. Ecomorfologia reflet- Perez, G. H., 1988. Guıá para el estudio de los macroinverte- indo a dieta dos peixes em um reservato´rio no sul do brados acua´ticos del Departamento de Antioquia. Uni- Brasil. Biota Neotropica 7: 67–76. versidad de Antioquia, Bogota´. Tupinamba´s, T. H., M. Callisto & G. B. Santos, 2007. Benthic Pinto, T. L. F. & V. S. Uieda, 2007. Aquatic insects selected as macroinvertebrates assemblages structure in two head- food for fishes of a tropical stream: are there spatial and water streams, south-eastern Brazil. Revista Brasileira de season differences in their selectivity? Acta Limnologica Zoologia 24: 887–897. Brasiliensia 19: 67–78. Vidotto-Magnoni, A. P. & E. D. Carvalho, 2009. Aquatic Reece, P. F. & J. S. Richardson, 2000. Benthic macroinverte- insects as the main food resource of fish the community in brates assemblages of coastal and continental streams and a Neotropical reservoir. Neotropical Ichthyology 7: large rivers of southwestern British Columbia, Canada. 701–708. Hydrobiologia 439: 77–89. Vono, V., L. G. M. Silva, B. P. Maia & H. P. Godinho, 2002. Rempel, L. L., J. S. Richardson & M. C. Healey, 2000. Biologia reprodutiva de treˆs espećies simpa´tricas de pei- Macroinvertebrate community structure along gradients of xes neotropicais: Pimelodus maculatus Lace´pe`de (Siluri- hydraulic and sedimentary conditions in a large gravel- formes, Pimeloidae), Leporinus amblyrhynchus Garavello bed river. Freshwater Biology 45: 57–73. & Britski e Schizodon nasutus Kner (Characiformes, Rocha, F. C., L. Casatti & D. C. Pereira, 2009. Structure and Anostomidae) do rece´m-formado Reservato´rio de Mir- feeding of a stream fish assemblage in Southeastern anda, Alto Parana´. Revista Brasileira de Zoologia 19: Brazil: evidence of low seasonal influences. Acta Lim- 819–826. nologica Brasiliensia 21: 123–134. Webster, M. M. & P. J. B. Hart, 2004. Substrate discrimination Rodrigues, S. C., 2002. Mudanças ambientais na regiaõdo and preference in foraging fish. Animal Behavior 68: Cerrado. Ana´lise das causas e efeitos da ocupaçaõ do solo 1071–1077. sobre o relevo. O caso da bacia hidrogra´fica do rio Webster, M. M. & P. J. B. Hart, 2006. Subhabitat selection by Araguari–MG. Geousp 12: 105–124. foraging threespine stickleback (Gasterosteus aculeatus): Rosa, R., J. L. S. Brito & S. C. Lima, 2006. Uso do solo e previous experience and social conformity. Behavioral cobertura vegetal na a´rea de influeˆncia do AHE Capim Ecology and Sociobiology 60: 77–86. Branco I. Sociedade e Natureza 18: 133–150.

123