Feeding Habits of Four Sympatric Fish Species in the Iberian Peninsula

Hydrobiologia (2011) 667:119–132 DOI 10.1007/s10750-011-0643-2

PRIMARY RESEARCH PAPER

Feeding habits of four sympatric ﬁsh species in the Iberian Peninsula: keys to understanding coexistence using prey traits

Javier Sańchez-Hernańdez • Rufino Vieira-Lanero • Marıá J. Servia • Fernando Cobo

Received: 11 October 2010 / Revised: 27 January 2011 / Accepted: 12 February 2011 / Published online: 4 March 2011 Ó Springer Science+Business Media B.V. 2011

Abstract Trophic interactions are important factors 81.3 up to 99.2%). In order to get a deeper insight structuring animal communities. We assessed the into mechanisms of fish species coexistence, we used trophic relations of four fish species that live in ten biological and ecological traits of macroinverte- sympatry in the River Ladra (NW Spain), and cluster brate prey to discriminate feeding preferences. As a analysis differentiated two feeding strategies: (1) result, despite the high similarity among the diets, our species with omnivorous feeding habits, feeding analyses suggest that differences in diel activity mainly on detritus and plant material but with aquatic patterns and drift behaviour of preys, as well as macroinvertebrates as an important complement differences in the prey size, are important adaptive (Achondrostoma arcasii and Pseudochondrostoma features that may reduce the inter-specific competi- duriense) and (2) species feeding mainly (Salmo tion in the fish community and permit the partitioning trutta) or exclusively (Gasterosteus gymnurus)on of food that allows coexistence. aquatic macroinvertebrates. Concerning ingested macroinvertebrates, the trophic overlap was quanti- Keywords Diet Macroinvertebrate traits fied using Schoener’s index and the results obtained Coexistence Iberian Peninsula revealed a high diet overlap among the species (from

Introduction Handling editor: M. Power Knowledge of feeding habits is essential to under- J. Sańchez-Hernańdez (&) F. Cobo stand the ecological role and the productive capacity Departamento de Zoologıá y AntropologıáFı´sica, of fish populations, and the understanding of these Universidad de Santiago de Compostela, Campus Sur s/n, 15782 Santiago de Compostela, Spain mechanisms is critical to the development of conser- e-mail: [email protected] vation and management plans (Teixeira & Cortes, 2006). Although many studies on trophic ecology J. Sańchez-Hernańdez R. Vieira-Lanero F. Cobo infer mechanisms for fish coexistence using diet Estacioń de Hidrobiologıá ‘‘Encoro do Con’’, Castroagudıń s/n, 36617 Vilagarcıá de Arousa, overlap data, the increasing amount of information on Pontevedra, Spain diverse habitat, behavioural and morphological characteristics of freshwater macroinvertebrates included M. J. Servia in fish diets may as well shed new light on the feeding Departamento de Biologıá Animal, Biologıá Vegetal y Ecologıá, Facultad de Ciencias, Universidad de A Corunã, strategy of fishes. Thus, prey trait analysis has been Campus da Zapateira s/n, 15008 A Corunã, Spain proposed as a functional approach to understand 123 120 Hydrobiologia (2011) 667:119–132 mechanisms involved in predator–prey relationships (Steindachner, 1866) and Pseudochondrostoma dur- (de Crespin de Billy & Usseglio-Polatera, 2002;de iense (Coelho, 1985). These four species are common Crespin de Billy et al., 2002), and consequently it in the Minõ basin (Hernando & Soriguer, 1992), and may be useful for understanding inter-species inter- in spite of the endangered status of G. gymnurus, actions and the mechanisms that determine food A. arcasii and P. duriense in this area (Doadrio, partitioning between them. 2001) information about their feeding behaviour and Competition for food can affect patterns of habitat trophic ecology is still poor. In this context, the aim selection, niche overlap and diel activity (Haury of our work is to improve the existing knowledge on et al., 1991; Alana¨ra¨ et al., 2001; Hilderbrand & the trophic ecology of these fish species and on the Kershner, 2004; David et al., 2007). Studies on similarity of their feeding resource niches. In addi- trophic interactions, notably in tropical areas, have tion, we want to check whether information on found differences in diet composition among species habitat, behaviour and morphology of prey provides and even within species among seasons in fish useful information on resource partitioning, thus communities (see for e.g., Magalha˜es, 1993a or helping to explain the coexistence of species. Encina et al., 2004 in temperate areas and de Me´rona & Rankin-De-Me´rona, 2004 or Novakowski et al., 2008 in tropical areas), but many others conclude that Materials and methods the same food resource can be shared by several species (Hesthagen et al., 2004; Gabler & Amundsen, The study area (altitude 395 m) was located in the 2010; Museth et al., 2010; Sandlund et al., 2010). River Ladra (Lugo, NW Spain), a tributary of the Thus, when food resources are shared, coexistence of river Minõ (308 km total length). River Ladra has a fish species has been suggested to be related to catchment area of 889 km2 and a total length of selection for one or more variables, including differ- 50 km (Rıó-Barja & Rodrı´guez-Lestega´s, 1992). ent activity patterns (Hesthagen et al., 2004)or Geologically, the study basin is characterised by an differential use of space (Amarasekare, 2003; Sandl- accumulation of alluvial gravels or tertiary lacustrine und et al., 2010). Recently, Gabler & Amundsen sediments on top of the flattened granite substratum (2010) provided evidences for the theory that com- (Vidal Romani, 1989). The Ladra basin includes a petitive coexistence in fishes with similar ecological mixture of agricultural and relatively undisturbed niches can be expected to occur when competition is areas, with small rural areas interspersed. The strong and/or the competing species are similar in vegetation structure comprises a series of extended their skills to compete (A˚ gren & Fagerstrøm, 1984; grazing lands with Monterey pine (Pinus radiata) and Schoener, 1989; Wiens, 1993). These authors found eucalypt (Eucalyptus globulus) forests. Thus, agri- extensive niche overlap between Atlantic salmon culture, stockbreeding and domestic sewage effluents (Salmo salar) parr and alpine bullhead (Cottus are the primary human impacts on the catchment. The poecilopus) in a sub-Artic river (Gabler & Amund- climate is typically Atlantic, with higher differences sen, 1999), and provided empirical support for the between extreme temperatures in summer and winter. existence of resource limitation and food competition The studied stream does not have any significant between both species (Amundsen & Gabler, 2008). flow regulation structure, and during the sampling Thus, the absence of phenotypical selection for the year (1996), the mean monthly flow varied from use of different prey resources or the failure of a 2.7 m3/s in August to 75.4 m3/s in January (mean specific resource for satisfying the complete feeding annual flow = 26.9 m3/s) (data from gauging station demand of a species might force an increase in niche number 1619, Minõ-Sil Basin Water Authority-C. width. As a consequence, generalisation in resource H. Minõ-Sil). use may be the mechanism that allows such The sampling site was located in an angling fish coexistence. area called ‘‘Coto de Bergonte’’ (UTM: 29T 606483E In this article, the feeding habits of four co- 4778014N), which is an oligotrophic watercourse occurring species have been studied in NW Spain: well-vegetated with abundant aquatic plants. At the Salmo trutta (Linnaeus, 1758), Gasterosteus gymnu- moment of survey, water temperature was circa rus (Cuvier, 1829), Achondrostoma arcasii 11.6°C, conductivity 67.1 lS/cm and pH 7.81. 123 Hydrobiologia (2011) 667:119–132 121

Dissolved oxygen levels were high (88% and 9.1 mg/l). range: 6.7–8.8 cm), G. gymnurus (n = 6, mean total Deciduous riparian vegetation was principally com- length ± SE = 4.7 cm ± 0.15; range: 4.1–5.2 cm), posed of alder (Alnus glutinosa) and willow (Salix A. arcasii (n = 12, mean fork length ± SE = spp.), and substrate consisted of boulders, gravel and 6.2 cm ± 0.18; range: 5.3–7.5 cm) and P. duriense sand. (n = 19, mean fork length ± SE = 6.5 cm ± 0.25; Samples were collected in October 1996. Protocols range: 3.9–8.2 cm). used in this study conform to the ethical laws of the Diet composition and feeding strategy of the four country and have been reviewed by the ethics species were compared by the analysis of their committee of the University of Santiago de Compos- stomach contents. Specimens were dissected and tela and the regional government (Xunta de Galicia). either stomach contents (Salmonidae and Gastero- Prior to fish electrofishing, samples of potential preys steidae) or contents of the gastrointestinal tract (benthic invertebrates) were collected from riffles (Cyprinidae) were removed. For the description of using a 0.1 m2 Surber sampler (n = 3). After collec- the diet, data are offered on frequency of occurrence tion, we fixed samples using 4% formalin and stored of preys (Fi = (Ni/N) 9 100, where Ni is the number them for later processing. In the laboratory, macro- of fishes with prey i in their stomach and N is the invertebrates were identified to the family level and total number of fishes with stomach contents of any were preserved in Eppendorf tubes using 70% kind) and relative abundance of preys (Ai = (RSi/ ethanol. In order to study prey selection, we quan- RSt) 9 100, where Si is the stomach content (num- tified the abundance and composition of aquatic ber) composed by prey i, and St the total stomach invertebrates. The feeding selectivity of fish species content of all stomachs in the entire sample). The was measured using Ivlev’s selectivity index (Ivlev, abundance of detritus and vegetal rests was not 1961). Values of this index range from -1to?1, quantified because it was impossible to count indi- with negative values indicating rejection or inacces- vidual items, and only the number of stomachs in sibility of the prey, zero indicating random feeding, which they appeared was noted. Nematoda were not and positive values indicating active selection. quantified because they might be parasites and not Fishes were collected during routine fish sampling benthic fauna, as stated in other works (Oscoz et al., using pulsed D.C. in a stretch of 110 m of the river. 2005). Prey items were identified to the family level All fishes captured were killed immediately by an when possible, and prey size was measured with a overdose of anaesthetic (benzocaine), except for digital micrometer to determine differences on size brown trout (55.3% of the individuals captured were consumption between species. When fragmented or returned to the stream). Fishes were transported in partially digested, the number of items was estimated coolboxes (*4°C) to the laboratory, where they were by counting body parts resistant to digestion. In those frozen at -30°C until processing. Four fish species cases, prey length was estimated from the width of were recorded in this study, two of which are the cephalic capsule (see Rincoń & Loboń-Cervia´, endemic species in the Iberian Peninsula (A. arcasii 1999), which was normally the best preserved and P. duriense). The fish community was dominated part. Moreover, the niche breadth of the individuals by S. trutta (0.077 fish/m2), being the density of the was estimated using the Shannon diversity index 0 rest of fish species very low (G. gymnurus = 0.007 (H =-RPilog10Pi, where Pi is the proportion of the fish/m2, A. arcasii = 0.013 fish/m2 and P. dur- prey item i among the total number of preys). In order iense = 0.021 fish/m2). to evaluate diet specialisation, evenness index (E = In total, 68 individuals belonging to these four H0/H0max) was used considering that values close to different species have been analysed. In the case of zero mean a stenophagous diet (i.e., individuals eat a S. trutta only age-0 individuals were selected, as they limited range of prey) and those closer to one a more were the most abundant and they were similar in size euryphagous diet (i.e., individuals eat a diverse range to individuals of the other species. Moreover, we of prey) (Oscoz et al., 2005). wanted to analyse individuals normally feeding on Food overlap among the four species was assessed macroinvertebrates, and avoid piscivority in older with Schoener’s overlap index (Schoener, 1970). The ones. Thus, individuals selected were: age-0 S. trutta overlap index has a minimum of 0 (no prey overlap), (n = 31, mean fork length ± SE = 7.7 cm ± 0.96; and a maximum of 1 (all items in equal proportions), 123 122 Hydrobiologia (2011) 667:119–132 and diet overlap is usually considered significant not available. Based on de Crespin de Billy & Us- when value of the index exceeds 60% (Wallace, seglio-Polatera (2002). 1981). The abundance of detritus, vegetal rests and In this study, ten macroinvertebrate biological Nematoda was not quantified, and only invertebrates traits were chosen for the analysis of trophic ecology were used for analysis of diet overlap. of co-occurring fish species. Thus, invertebrate Statistical analyses were conducted using the preferences (‘macrohabitat trait’ and ‘current velocity programme SPSS 16.0. All tests were considered trait’) were defined at the macrohabitat scale and statistically significant at P level \ 0.05. The simi- were used in this article to obtain information on the larity index and cluster analysis were conducted using preferential use of feeding habitat of the four fish the programme PRIMER statistical package version species (see trait categories in Table 1). The tendency 5.0 (Clarke & Gorley, 2001). By means of the of different invertebrate taxa to utilise different types software R (version 2.11.1), a fuzzy principal corre- of substratum (‘substratum trait’) was used to spondence analysis (FPCA) was used to analyse estimate their conspicuousness and accessibility to macrohabitat, behavioural feeding and handling effi- fishes at the meso- and micro-scales. Meso-scale ciency according to the prey items consumed by applies to channel features within a section, and fishes. FPCA is a method for robust estimation of micro-scale is related to characteristics within fea- principal components that has been described with tures, here related to the aquatic habitat of macroin- detail, for e.g., by Cundari et al. (2002), who found vertebrates (Newson & Newson, 2000). For instance, that this method diminishes the influence of outliers. Heptageniidae nymphs that use exposed microhabi- The ADE4 library for analysis in R can be freely tats are more likely to be dislodged from the obtained at http://cran.es.r-projet.org/. substratum and enter the drift than Orthocladiinae We used the same trait database and trait analyses larvae that use protected microhabitats (Rader, 1997). as de Crespin de Billy (2001) and de Crespin de Billy The ‘tendency to drift in the water column trait’, & Usseglio-Polatera (2002). To evaluate the potential ‘tendency to drift at the water surface trait’ and ‘diel vulnerability of invertebrates to fish predation, de drift behaviour trait’ were used to describe the Crespin de Billy & Usseglio-Polatera (2002) created behavioural feeding habits of the fish species, and a total of 71 different categories for 17 invertebrate finally, ‘potential size trait’, ‘concealment trait’, traits (see trait categories used in this study in ‘body shape trait’ and ‘body flexibility trait’ were Table 1). Information was structured using a ‘fuzzy used to analyse handling efficiency of fishes. coding’ procedure (Chevenet et al., 1994). Thus, a score was assigned to each taxon describing its affinity for each category of each trait, with ‘0’ indicating ‘no affinity’ to ‘5’ indicating ‘high affin- Results ity’. The taxonomic resolution (order, family, genus) used in the classification process corresponded to the The composition of benthic invertebrates included lowest possible level of determination of taxa in fish 28 taxa (density = 2030 ind/m2 and biomass = gut contents. When identification to genus was not 164.32 g/m2). The beetle Elmidae was the most possible or in the case of missing information for a abundant taxon and represented 17.56% of the total certain genus, the value assigned for a trait was that number of individuals. Chironomidae and Simuliidae of the family level, using the average profile of all contributed 16.59 and 15.61%, respectively to the other genus of the same family, as recommended by total abundance (Table 2). de Crespin de Billy & Usseglio-Polatera (2002) and Stomach contents in age-0 S. trutta showed that Rodrı´guez-Capı´tulo et al. (2009). All the taxa and diet included a wide variety of prey species, with their assigned scores for each category can be found Simuliidae (25.9% of total prey) and Baetidae (22.3% at http://www.aix.cemagref.fr/htmlpub/divisions/Hyax/ of total prey) dominating. In total, 305 preys titres-publication.htm (de Crespin de Billy & Usse- belonging to 26 taxa were identified, and the diets glio-Polatera, 2002). Ostracoda, Calopterygidae, were dominated by aquatic invertebrates (96.4% of Gyrinidae, Ecnomidae and Leptoceridae were not total prey) (Table 2). Piscivorous behaviour was included in the analysis because trait values are still found only in one brown trout (8.5 cm fork length). 123 Hydrobiologia (2011) 667:119–132 123

Table 1 Traits, categories and codes used in analyses and Table 1 continued graphics Trait Categories Code Trait Categories Code (9) Body shape Cylindrical cyl (1) Macrohabitat Hyporheic ‘burrower’ hypo-b (including cases/tubes) Spherical sph Hyporheic ‘interstitial’ hypo-i Conical con Epibenthic depositional epi-d Flattened flat Epibenthic erosional epi-e Streamlined strl Water column water (10) Body flexibility None (\10°) none (2) Current Still (0–5) 0–5 (including cases/tubes) Weak (10–45°) weak velocity (cm/s) Slow (5–25) 5–25 High ([45°) high Moderate (25–75) 25–75 Fast ([75) [75 (3) Substratum Mud mud The diet of G. gymnurus was dominated by (mm) Silt (0.001–0.2) silt Diptera (61.9% of total prey) and Ephemeroptera Sand (0.2–2) sand (19.05% of total prey). A total of 21 preys were Fine gravel (2–8) gravel identified in the stomach contents and the most Gravel-cobble (8–256) cobble abundant taxa in the diet were chironomid larvae Blocks ([256) bloc (57.14% of the total number). No terrestrial inverte- Bryophytes bryo brates were found in the diet (Table 2). Other macrophytes bryo-o The diet composition of A. arcasii was constituted Roots root by eight types of prey, and a total of 57 preys were Litter, organic detritus detr identified in the stomach contents. A. arcasii con- (4) Tendency to None none sumed both plant and detritus in a similar proportion drift in the water Weak weak (58.33 and 50% of occurrence, respectively). Diptera column Medium med (Chironomidae) were the most abundant prey High high (46.67%) and were identified in the 33.33% of the (5) Tendency to drift None none stomachs. No terrestrial invertebrates were found in at the water surface Weak weak the diet (Table 2). Medium med A total of 15 preys and 14 items were identified in High high the stomach contents of P. duriense. Detritus was the (6) Diel drift None none most important dietary item, and occurred in 84.21% of guts. In relation with benthic fauna, Diptera behaviour Nocturnal Noct (36.8% of total prey) and specifically chironomid Dawn dawn larvae were the most prevalent food, accounting for Daylight d-light 26.32% of animal prey numbers and being ingested Twilight t-light by 47.37% of the fishes (Table 2). Terrestrial inver- (7) Potential B2 \2 tebrates were also present (8.77% of total prey). size (mm) [2–4 2–4 Diet comparison among species shows that plant [4–8 4–8 material and detritus are present in A. arcasii and [8–16 8–16 P. duriense, although in different proportions, show- [16–32 16–32 ing that both species present typically omnivorous [32 [32 feeding habits (Table 2). The similarity index (8) Concealment Fixed accessory (nets, retreats) net (Table 3) and cluster analysis carried out on fre- Movable accessory case quency of occurrence data (Fig. 1) differentiated two (cases/tubes) feeding groups: (1) species feeding mainly on detritus Solidly coloured sold-c and plant material but with aquatic macroinverte- Variable var-c brates as an important complementary food category Patterned patt-c (A. arcasii and P. duriense), and (2) species feeding 123 124 Hydrobiologia (2011) 667:119–132

Table 2 Diet composition for each ﬁsh species and benthos fauna in the Ladra River Benthos G. gymnurus A. arcasii S. trutta P. duriense

Ai (%) Ai (%) Fi (%) Ivlev Ai (%) Fi (%) Ivlev Ai (%) Fi (%) Ivlev Ai (%) Fi (%) Ivlev

Oligochaeta* 2.44 – – – 6.67 8.33 0.46 0.33 3.23 -0.76 – – – Erpobdellidae 3.41 – – – – – – – – – – – – Ancylidae 0.49 – – – 20 16.67 0.95 0.66 6.45 0.15 7.02 15.79 0.87 Lymnaeidae 0.98 – – – – – – 1.31 6.45 0.14 – – – Hydrobiidae 0.49 – – – – – – – – – – – – Ostracoda* – 4.76 16.67 1 6.67 8.33 1 – – – – – – Asellidae – – – – – – – 1.64 12.90 1 – – – Baetidae 6.34 19.05 50 0.50 – – – 22.30 64.52 0.55 10.53 31.58 0.24 Ephemerellidae 2.93 – – – – – – – – – – – – Heptageniidae – – – – – – – 0.66 6.45 1 – – – Leptophlebiidae 0.98 – – – – – – – – – – – – Leuctridae 1.95 – – – – – – 5.57 22.58 0.48 5.26 10.53 0.45 Nemouridae 0.98 – – – – – – 0.98 6.45 -0.34 – – – Calopterygidae 0.49 – – – – – – 1.31 12.90 0.45 – – – Cordulegasteridae 0.49 – – – – – – – – – – – – Aphelocheiridae 0.98 – – – – – – – – – – – – Sialidae 0.98 – – – – – – – – – – – – Gyrinidae 2.44 – – – – – – 1.64 12.90 -0.20 – – – Hydraenidae – – – – – – – – – – 1.75 5.26 1 Elmidae 17.56 4.76 16.67 -0.58 13.33 8.33 -0.14 0.98 9.68 -0.90 17.54 15.79 -0.01 Brachycentridae 7.32 – – – – – – 0.98 6.45 -0.77 – – – Ecnomidae – 9.52 16.67 1 – – – 0.33 3.23 1 – – – Goeridae 0.49 – – – – – – – – – – – – Leptoceridae – – – – 6.67 8.33 1 1.97 19.35 1 3.51 10.53 1 Hydropsychidae 6.34 – – – – – – 13.44 54.84 0.35 8.77 26.32 0.16 Limnephilidae 0.49 – – – – – – 0.66 6.45 0.15 – – – Rhyacophilidae – – – – – – – 0.66 6.45 1 – – – Sericostomatidae 5.37 – – – – – – – – – – – – Polycentropodidae 1.46 – – – – – – 0.66 6.45 -0.38 – – – Athericidae 0.98 – – – – – – – – – – – – Chironomidae 16.59 57.14 83.33 0.55 46.67 33.33 0.47 14.75 45.16 -0.06 26.32 47.37 0.22 Empididae 0.49 – – – – – – – – – – – – Simuliidae 15.61 4.76 16.67 -0.54 – – – 25.90 41.94 0.24 10.53 26.32 -0.20 Tabanidae 0.49 – – – – – – – – – – – – Tipulidae 0.49 – – – – – – – – – – – – Terrestrial invertebrates Chironomidae – – – – – – 0.66 6.45 – 8.77 15.79 – Trichoptera* – – – – – – 0.98 9.68 – – – – Juliidae – – – – – – 0.33 3.23 – – – – Non-identiﬁed Insecta – – – – – – 1.31 12.90 – – – – Other prey items P. duriense – – – – – – 0.33 3.23 – – – – Nematoda* – – – – – – – 9.68 – – 5.26 –

123 Hydrobiologia (2011) 667:119–132 125

Table 2 continued Benthos G. gymnurus A. arcasii S. trutta P. duriense

Ai (%) Ai (%) Fi (%) Ivlev Ai (%) Fi (%) Ivlev Ai (%) Fi (%) Ivlev Ai (%) Fi (%) Ivlev

Detritus – – – – 50 – – – – – 84.21 – Plant material – – – – 58.3 – – – – – 36.84 – Filamentous algae – – – – – – – – – – 42.11 –

Abundance (Ai), frequency of occurrence (Fi) and Ivlev’s index. (*) Not identified. The abundance of detritus, vegetal rests and Nematoda was not quantified, and only the number of stomachs in which they appeared was noted exclusively (G. gymnurus) or mainly (S. trutta)on A. arcasii (H0 = 0.45 ± 0.249 SE, range = 0–2.25). aquatic macroinvertebrates. Differences among species were statistically signifi- A total of 27 invertebrate families were identified in cant (ANOVA test: F = 10.12, P \ 0.001). Similarly, the stomach contents, being Insecta the most common the evenness index was significantly higher in age-0 S. prey group. In all species, benthic prey constituted the trutta (0.57 ± 0.046 SE, range = 0–1) than in P. dur- most important prey in abundance terms, and terrestrial iense (0.32 ± 0.066 SE, range = 0–0.8), G. gymnurus invertebrates were only used by age-0 S. trutta and (0.29 ± 0.102 SE, range = 0–0.6) or A. arcasii P. duriense (3.3 and 8.8%, respectively, Fig. 2). The (0.1 ± 0.076 SE, range = 0–0.9) (ANOVA test: analysis of the feeding strategy of the species using F = 10.08, P \ 0.001), as these three species showed Shannon diversity index indicated that juvenile brown a stenophagous diet, especially G. gymnurus and trout had the largest niche breadth (H0 = 1.42 ± 0.128 A. arcasii. The percentage of benthic prey, terrestrial SE, range = 0–2.6), with lower values for P. duriense invertebrates and other prey items are given in (H0 = 0.88 ± 0.175 SE, range = 0–2), for G. gymnu- Table 4. rus (H0 = 0.65 ± 0.251 SE, range = 0–1.5) and for The dietary analyses, based on benthic invertebrates, showed high values of diet overlap using the Schoener index ([81.3%, even up to 99.2%) Table 3 Values of the Bray-Curtis index of diet similarity between fish species (Table 5), and all the species showed a remarkable similarity in their prey utilisation patterns. Thus, G. gymnurus A. arcasii S. trutta Chironomidae constituted a significant proportion of A. arcasii 22.2 the diet for G. gymnurus, A. arcasii and P. duriense S. trutta 38.3 21.8 (57.14, 46.67 and 26.32%, respectively). In contrast, P. duriense 30.7 58.4 46.9 age-0 brown trout fed chiefly on Simuliidae and Baetidae (25.9 and 22.3% respectively).

Fig. 1 Dendrogram resulting from the cluster analysis per- formed on stomach content data, in terms of prey occurrence, Fig. 2 Diet composition in abundance terms. Data are of the ﬁsh species presented for each ﬁsh species 123 126 Hydrobiologia (2011) 667:119–132

Table 4 Diet composition, Shannon diversity index (H0) and evenness index (E) for each ﬁsh species (mean ± SE) G. gymnurus A. arcasii S. trutta P. duriense

Benthic prey (%) 100 100 92 ± 3.56 89.4 ± 6.79 Terrestrial invertebrates (%) 0 0 6.9 ± 3.44 10.6 ± 6.79 Other prey items 0 0 1.1 ± 1.07 0 Shannon diversity index (H0) 0.65 ± 0.251 0.45 ± 0.249 1.42 ± 0.128 0.88 ± 0.175 Evenness index (E) 0.29 ± 0.102 0.1 ± 0.076 0.57 ± 0.046 0.32 ± 0.066

Table 5 Diet overlap (Schoener’s index) among different fish cases. A general tendency that can be observed is the species in River Ladra wider diversity of characteristics of age-0 S. trutta Schoener’s index (%) and P. duriense preys in comparison with those of G. gymnurus and A. arcasii. ‘Macrohabitat’, ‘current G. gymnurus–A. arcasii 99.2 velocity’ and ‘substratum’ traits showed no clear G. gymnurus–S. trutta 82.1 differences for preys of the four fish species (Fig. 3a–c), G. gymnurus–P. duriense 95.5 although overlap was higher between age-0 S. trutta A. arcasii–S. trutta 81.3 and P. duriense. Concerning handling efficiency A. arcasii–P. duriense 94.7 habits, ‘concealment’ and ‘body shape’ traits were S. trutta–P. duriense 86.6 similar for preys of the four species (Fig. 3d, f). Differences were found for ‘potential size’ (Fig. 3e) and ‘body flexibility’ (Fig. 3g), as age-0 S. trutta and In relation to prey size, there were differences in P. duriense were able to feed on preys showing weak the average prey size consumption among species body flexibility, whereas G. gymnurus and A. arcasii (Kruskal–Wallis test; P \ 0.001). Age-0 S. trutta fed preferred to feed only on prey with high flexible on mean size 5.5 mm ± 0.47 SE, being this size bodies. In relation to ‘potential size’, A. arcasii higher than that for G. gymnurus (3.7 mm ± 0.42 showed a surprising but clear tendency to feed on SE), P. duriense (2.3 mm ± 0.28 SE) or A. arcasii potentially bigger preys than the rest of species. (1.7 mm ± 0.17 SE). However, a careful revision of the trait values for its A comparison of benthic macroinvertebrate avail- preys showed that this result was due to the high ability and prey selectivity using Ivlev’s selectivity values assigned to few families of Oligochaeta (de index shows, in general, that fishes selected nega- Crespin de Billy & Usseglio-Polatera, 2002; Tachet tively for Elmidae. A. arcasii, P. duriense and et al., 2002). Finally, traits that best separated fish gut G. gymnurus had a similar pattern of selection for samples were those related to behavioural feeding the three more abundant items in the benthos habits: ‘tendency to drift in the water column’ (Elmidae, Chironomidae and Simuliidae). Thus, (Fig. 3h), ‘tendency to drift at the water surface’ whereas Elmidae and Simuliidae were negatively (Fig. 3i) and ‘diel drift behaviour’ (Fig. 3j). As it can selected, Chironomidae were positively selected, and be seen in the graphics, A. arcasii is clearly separated remained a large component of the diet of the three from the other three species. This is due to its species. In contrast, age-0 S. trutta selected nega- preference to prey on organisms with weak or no tively for Chironomidae and Elmidae, but positively tendency to drift, while G. gymnurus preys show a for Simuliidae. Interestingly, Ancylidae were highly medium tendency to drift. Thus, although both selected by A. arcasii and P. duriense (0.95 and 0.87, P. duriense and A. arcasii are omnivorous cyprinids, respectively) (Table 2). A. arcasii seems to feed exclusively on sediments, Concerning macroinvertebrate trait analyses, the eating detritus and preys with almost no ability to two first axes were sufficient to illustrate the drift. In contrast, P. duriense, as it is the case of age-0 relationships among faunal groups according to their S. trutta, shows a higher spectrum of prey, which combinations of traits (‘eigenvalues’ of Fig. 3), and reveals a greater ability to prey on different substrata. accounted for [70% of the total variability in all Moreover, P. duriense and age-0 S. trutta can include 123 Hydrobiologia (2011) 667:119–132 127

Fig. 3 Biplot of gut contents obtained from a fuzzy principal c correspondence analysis (FPCA). Distribution of each trait according to the gut contents (1) and histogram of eigenvalues (2, the first two values are in black). Data are presented for each fish species. A, A. arcasii;G,G. gymnurus;P,P. duriense and S, S. trutta. Ellipses envelop weighted average of prey taxa positions consumed by fish species: Labels (A, G, P and S) indicate the gravity centre of the ellipses. Filled lines link prey families (represented by a point) to their corresponding predator but are only 80% of their total length for readability. Dotted lines represent the width and height of ellipses. Details and data needed for the elaboration of ‘‘a’’ to ‘‘j’’ graphics can be found in the ‘‘Material and methods’’ section and Table 1 in their diets macroinvertebrates that drift in the water column or in the surface. Finally, G. gymnurus and P. duriense showed a slight preference for preys that drift during the day, but age-0 S. trutta seems to prefer to feed at dusk.

Discussion

Most Iberian fish species are insectivorous (see for e.g., Magalha˜es, 1993a; Valladolid & Przybylski, 1996), while omnivory is a less common feeding strategy (Doadrio, 2001). In this study, as expected, our results segregated captured species into two feeding strategies: (1) species with omnivorous feeding habits as A. arcasii and P. duriense and (2) species feeding mainly (S. trutta) or exclusively (G. gymnurus) on aquatic macroinvertebrates. Concerning diet composition, results are broadly in accordance with previous similar studies. Thus, detritus and plant material were important food resources for A. arcasii and P. duriense, presenting both species omnivorous feeding habits in which aquatic macroinvertebrates were an important food supply (Magalha˜es, 1993a; Encina & Granado-Lorencio, 1994;Loboń-Cervia & Rincoń, 1994). In contrast with Loboń-Cervia & Rincoń(1994), who studied the feeding habits of A. arcasii in a Mediterranean stream in central Spain, in our case they did not feed on drift, while those authors consider this feeding habit is important for the species. Although we did not quantify drift abundance, availability was expected to be low in our sampling date (September). Moreover, besides the important differences in habitat conditions between our study site and theirs that might explain this result (for presented more bottom feeders such as Squalius example river flow or substrate characteristics), fish carolitertii (Doadrio, 1988), Barbus bocagei (Steind- communities differ also on their composition, as the achner, 1865), Cobitis calderoni (Ba˘cescu, 1962) and location sampled by Loboń-Cervia & Rincoń(1994) Gobio lozanoi (Doadrio & Madeira, 2004) (Rincoń& 123 128 Hydrobiologia (2011) 667:119–132

Loboń-Cervia´, 1995). Indeed, it is well known that since ontogenetic changes in the diet may reduce brown trout is a territorial drift feeder (Elliott, 1994; competition facilitating the partitioning of resources Friberg et al., 1994), and several authors have reported (Elliott, 1967). the behavioural dominance of trout over cyprinids in Fishes do not always consume the most abundant streams (Grossman & Boule´, 1991; Facey & Gross- taxa available in the benthos (Sagar & Glova, 1995; man, 1992; Hill & Grossman, 1993). MacNeil et al., 2000; de Crespin de Billy & Usseglio- Results of diet composition and trait analyses Polatera, 2002). In our case, our results show that differ also with previous findings on P. duriense Elmidae, abundant in the benthos, were negatively feeding habits. Thus, in our case P. duriense shows a selected by all the species. This rejection may be due high spectrum of prey, which reveals a great ability to to their low energetic value, as they have an intense prey on different substrata and different macroinver- sclerotisation, but it may also be due to their bad taste tebrates, both aquatic and terrestrial. Contrastingly, (Ochs, 1969; Power, 1992; Oscoz et al., 2000). Rodrı´guez-Jimeńez (1987) and Encina et al. (2004) Indeed, as predicted by the optimal foraging theory suggest that this species presents a narrow dietary (OFT), fishes should select those prey items that spectrum, as they found that diet was composed maximize their net rate of energy gain (Gerking, almost exclusively by detritus and plant material. 1994; Gill, 2004), but a wide range of factors are However, Magalha˜es (1993a) underlines the impor- acknowledged to influence prey selection. According tance of chironomid larvae in the feeding of this to several authors, food abundance and prey charac- species, accounting for more than 70% of the total teristics (accessibility and body shape) are the prey number. principal variables that determine food selection in Despite the fact that the food habits of A. arcasii fishes [see for e.g., de Crespin de Billy et al., 2002 or and P. duriense seem to change among locations (see Johnson et al., 2007). Moreover, it is also interesting above), G. gymnurus apparently shows a more to note that Ancylidae were positively selected by constant diet. Thus, this study confirms previous A. arcasii and P. duriense, as they are obtained by research indicating that three-spine stickleback feeds scraping on the surface of rocks and other substrata almost exclusively on aquatic invertebrates (Sańchez- (Magalha˜es, 1993a; Encina & Granado-Lorencio, Gonza´les et al., 2001; Copp & Kovać, 2003). 1994). Actually, results of trait analysis in this study are in A noteworthy result of our study is that although accordance with the bottom-feeder behaviour found two of the four species analysed showed omnivorous in previous studies (Ranta & Kaitala, 1991; Hart, feeding habits, there is a remarkably high overlap 2003). among all of them concerning ingested macroinver- In relation with age-0 brown trout, simulid larvae, tebrates. These are the prey items that present higher baetid nymphs and chironomid larvae seem to be the energy values, as detritus has been considered of low most important food items for juvenile trout, as found nutritional and energetic value for omnivorous fishes by many researchers (Degerman et al., 2000; Hest- (Bowen, 1979, 1987). Indeed, analysis of macroin- hagen et al., 2004;Sańchez, 2009). Besides, results of vertebrate traits showed also a high overlap among trait analyses are in accordance also with previous the fish species, especially age-0 S. trutta and findings concerning for example prey diversity (de P. duriense. However, high overlap values may not Crespin de Billy & Usseglio-Polatera, 2002)or indicate competition, since species can adopt differ- preference to feed at dusk (Neveu, 1980; Bjo¨rnsson, ent strategies to overcome competence. 2001). However, a striking result was the finding of First, different species may specialise on distinct piscivorous behaviour in an age-0 trout. This behav- resources (Amarasekare, 2003). The ability of cyp- iour is frequent in large brown trout, and most studies rinids to feed on detritus and plant material may have show that it occurs in individuals with a size of a high competitive value in environments with severe 20–30 cm (Kahilainen & Lehtonen, 2002; Jensen competition (Magalhaes, 1992) and may reduce the et al., 2004), but it is unusual in trout of small size inter-specific competition in the area studied. More- (the individual we found was 8.5 cm in fork length) over, resource partitioning may also occur at the level (L’Abeé-Lund et al., 1992). Thus, this foraging habit of prey size, although it is not clear whether this size might be adopted in highly competitive communities, selective strategy is adopted to reduce interspecific 123 Hydrobiologia (2011) 667:119–132 129 competition or it is the result of foraging behaviour that complete the original data base elaborated by de and/or morphological constraints such as gape size Crespin de Billy (2001) and de Crespin de Billy & (Stevens et al., 2006), as it may be the case in our Usseglio-Polatera (2002). study. Another limiting factor for application of trait- Second, segregation of microhabitat is an impor- based metrics in diet analysis is related to the tant factor for reducing the effects of the competition taxonomic level that can be achieved in the identi- by the trophic resource (Baker & Ross, 1981; Yant fication of prey in stomach contents. Thus, the high et al., 1984; Haury et al., 1991). Indeed, the use of degree of digestion of most of the individuals microhabitats is often distinguishable between spe- impedes its identification below the family level, cies (Grossman et al., 1987a, b; Rincoń & Loboń- and although this drawback can be overcome (see de Cervia´, 1993; Clavero et al., 2009), and in our case, Crespin de Billy & Usseglio-Polatera, 2002 and for e.g., differences were found among species on the Rodrı´guez-Capı´tulo et al., 2009 for the use of values ability to feed at different depths of the water column. at family level), important ecological information Third, competition might get reduced also by might be lost (Dole´dec et al., 2000). differences in the diel activity patterns of fishes, as it However, despite the above-mentioned problems, is the case of species with dominance hierarchy trait analysis provides extremely valuable ecological (Alana¨ra¨ et al., 2001; David et al., 2007). Thus, in our information on the mechanisms involved in predator– case, except for A. arcasii, the rest of the species prey relationships, and complements traditional diet showed slight preferences to feed on certain periods analysis. Indeed, trait-based metrics provide a con- of the day. sistent method for fish diet analysis across local, Finally, terrestrial prey may constitute an impor- regional and continental scales, as species traits tend tant food resource, and they are present primarily on to be less constrained by biogeography than species the stream surface. They tend to be absent from the composition (Vieira et al., 2006). In addition, it is a diets of benthic feeders (Garman, 1991, but see robust statistical methodology, and free software is Coelho et al., 1997), although the intake of terrestrial available for the analyses. Finally, traits can be linked insects by cyprinids may contribute greatly to the to food web dynamics, thus reflecting not only outcome of fish interactions (Magalha˜es, 1993b). community structure but also ecosystem function In conclusion, our results demonstrate that parti- (Heino, 2005) and providing invaluable information tioning of food is possible at different levels, and for conservation and management. analysis of prey traits provided us with few important clues for understanding the coexistence of fish Acknowledgements Marıá Teresa Couto (University of species. Thus, feeding strategies concerning drift Santiago of Compostela) gave advice for trait analysis. Part of this work has been carried out in the laboratories of the behaviour of prey, diel activity and prey size seem to Station of Hydrobiology of the USC ‘‘Encoro do Con’’ at be important factors that explain the coexistence in Vilagarcıá de Arousa. This work has been partially supported this fish community at the time of sampling, as by the project INCITE09203072PR of the Xunta de Galicia. seasonal variations in feeding strategies might occur. The authors are also grateful to two anonymous referees for their helpful comments. Promising results of this study encourage its extension using data from different stations, different seasons and different ontogenic stages of the studied species. However, there are still gaps to be filled References before macroinvertebrate trait analysis can be fully adopted in fish diet studies. In fact, one of the main A˚ gren, G. I. & T. Fagerstrøm, 1984. Limiting similarity in disadvantages of this approach is the lack of plants: randomness prevents exclusion of species with similar competitive abilities. Oikos 43: 369–375. comprehensive trait databases (Vieira et al., 2006; Alana¨ra¨, A., M. D. Burns & N. B. Metcalfe, 2001. Intraspecific Statzner & Beˆche, 2010), and this has been precisely resource partitioning in brown trout: the temporal distri- one of the handicaps found in our analysis, as traits bution of foraging is determined by social rank. Journal of data for common taxa such as Ostracoda, Calo- Animal Ecology 70: 980–986. Amarasekare, P., 2003. Competitive coexistence in spatially pterygidae, Gyrinidae, Ecnomidae or Leptoceridae structured environments: a synthesis. Ecology Letters 6: are still nor available. Therefore, works are needed 1109–1122. 123 130 Hydrobiologia (2011) 667:119–132

Amundsen, P.-A. & H.-M. Gabler, 2008. Food consumption Degerman, E., I. Na¨slund & B. Sers, 2000. Stream habitat use and growth of Atlantic salmon parr in subarctic rivers: and diet of juvenile (0?) brown trout and grayling in empirical support for food limitation and competition. sympatry. Ecology of Freshwater Fish 9: 191–201. Journal of Fish Biology 73: 250–261. Doadrio, I., 2001. Atlas y libro rojo de los peces continentales Baker, J. A. & S. T. Ross, 1981. Spatial and temporal resource de Espanã. Ministerio de Medio Ambiente y Consejo utilization by southeastern cyprinids. Copeia 1981: 178–189. Superior de Investigaciones Cientı´ficas, Madrid. Bjo¨rnsson, B., 2001. Diel changes in the feeding behaviour of Dole´dec, S., J. M. Olivier & B. Statzner, 2000. Accurate Arctic char (Salvelinus alpinus) and brown trout (Salmo description of the abundance of taxa and their biological trutta) in Ellidavatn, a small lake in southwest Iceland. traits in stream invertebrate communities: effects of tax- Limnologica-Ecology and Management of Inland Waters onomic and spatial resolution. Archiv fu¨r Hydrobiologie 31: 281–288. 148: 25–43. Bowen, S. H., 1979. A nutritional constraint in detritivory by Elliott, J. M., 1967. The food of trout (Salmo trutta)ina fishes: the stunted population of Sarotherodon mossam- Dartmoor stream. Journal of Applied Ecology 4: 60–71. bicus in Lake Sibaya, South Africa. Ecological Mono- Elliott, J. M., 1994. Quantitative Ecology and the Brown Trout. graphs 49: 17–31. Oxford University Press, Oxford. Bowen, S. H., 1987. Composition and nutritional value of Encina, L. & C. Granado-Lorencio, 1994. Gut evacuation in detritus. In Moriarty, D. J. W. & R. S. V. Pullin (eds), barbel (Barbus sclateri G. 1868), nase (Chondrostoma Detritus and Microbial Ecology in Aquaculture. ICL- willkommi S., 1866). Ecology of Freshwater Fish 23: 1–8. ARM, Manila: 192–216. Encina, L., A. Rodriguez-Ruiz & C. Granado-Lorencio, 2004. Chevenet, F., S. Dole´dec & D. Chessel, 1994. A fuzzy coding Trophic habits of the fish assemblage in an artificial approach for the analysis of long-term ecological data. freshwater ecosystem: the Joaquin Costa reservoir, Spain. Freshwater Biology 31: 295–309. Folia Zoologica 53: 437–449. Clarke, K. R. & R. N. Gorley, 2001. PRIMER v5: User Facey, D. E. & G. D. Grossman, 1992. The relationship Manual/Tutorial. PRIMER-E, Plymouth. between water velocity, energetic costs and microhabitat Clavero, M., Q. Pou-Rovira & L. Zamora, 2009. Biology and use in four North American stream fishes. Hydrobiologia habitat use of three-spined stickleback (Gasterosteus 239: 1–6. aculeatus) in intermittent Mediterranean streams. Ecology Friberg, N., T. H. Andersen, H. O. Hansen, T. M. Iversen, D. of Freshwater Fish 18: 550–559. Jacobsen, L. Krojgaard & S. E. Larsen, 1994. The effect Coelho, M. M., M. J. Martins, M. J. Collares-Pereira, A. of brown trout (Salmo trutta L.) on stream invertebrate M. Pires & I. G. Cowx, 1997. Diet and feeding relation- drift, with special reference to Gammarus pulex L. Hyd- ships of two Iberian cyprinids. Fisheries Management and robiologia 294: 105–110. Ecology 4: 83–92. Gabler, H.-M. & P.-A. Amundsen, 1999. Resource partitioning Copp, G. H. & V. Kovać, 2003. Sympatry between threespined between Siberian bullhead (Cottus poecilopus Heckel) Gasterosteus aculeatus and nine-spined Pungitius pungi- and Atlantic salmon parr (Salmo salar L.) in a sub-Arctic tius sticklebacks in English lowland streams. Annales river, northern Norway. Ecology of Freshwater Fish 8: Zoologici Fennici 40: 341–355. 201–208. Cundari, T. R., C. Saˆrbu & H. F. Pop, 2002. Robust fuzzy Gabler, H.-M. & P.-A. Amundsen, 2010. Feeding strategies, principal component analysis (FPCA). A comparative resource utilisation and potential mechanisms for compet- study concerning interaction of carbon-hydrogen bonds itive coexistence of Atlantic salmon and alpine bullhead in a with molybdenum-oxo bonds. Journal of Chemical sub-Arctic river. Aquatic Ecology 44: 325–336. Information and Computer Sciences 42: 1363–1369. Garman, G. C., 1991. Use of terrestrial arthropod prey by a David, B. O., G. P. Closs, S. K. Crow & E. A. Hansen, 2007. Is stream-dwelling cyprinid fish. Environmental Biology of diel activity determined by social rank in a drift-feeding Fishes 30: 325–331. stream fish dominance hierarchy? Animal Behaviour 74: Gerking, S. D., 1994. Feeding Ecology of Fish. Academic 259–263. Press, San Diego. de Crespin de Billy, V., 2001. Re´gime alimentaire de la truite Gill, A. B., 2004. The dynamics of prey choice in fish: the (Salmo trutta L.) en eaux courantes: roˆles de l0habitat importance of prey size and satiation. Journal of Fish physique des traits des macroinverte´bre´s. Thesis. L0uni- Biology 63: 105–116. versite´ Claude Bernard, Lyon: 84 pp. Grossman, G. D. & V. Boule´, 1991. An experimental study of de Crespin de Billy, V. & P. Usseglio-Polatera, 2002. Traits of competition for space between rainbow trout (Oncorhyn- brown trout prey in relation to habitat characteristics and chus mykiss) and rosyside dace (Clinostomus fundulo- benthic invertebrate communities. Journal of Fish Biology ides). Canadian Journal of Fisheries and Aquatic Sciences 60: 687–714. 48: 1235–1243. de Crespin de Billy, V., B. Dumont, T. Lagarrigue, P. Baran & Grossman, G. D., A. de Sostoa, M. C. Freeman & J. Loboń- B. Statzner, 2002. Invertebrate accessibility and vulnera- Cervia´, 1987a. Microhabitat use in a Mediterranean riv- bility in the analysis of brown trout (Salmo trutta L.) erine fish assemblage. Fishes of the upper Matarranã. summer habitat suitability. River Research and Applica- Oecologia (Berlin) 73: 490–500. tions 18: 533–553. Grossman, G. D., A. de Sostoa, M. C. Freeman & J. Loboń- de Me´rona, B. & J. Rankin-De-Me´rona, 2004. Food resource Cervia´, 1987b. Microhabitat use in a Mediterranean riv- partitioning in a fish community of the central Amazon erine fish assemblage. Fishes of the upper Matarranã. floodplain. Neotropical Ichthyology 2: 75–84. Oecologia (Berlin) 73: 501–512. 123 Hydrobiologia (2011) 667:119–132 131

Hart, P. J. B., 2003. Habitat use and feeding behaviour in two Magalha˜es, M. F., 1993b. Effects of season and body size on closely related fish species, the three-spined and nine- the distribution and diet of the Iberian chub Leuciscus spined stickleback: an experimental analysis. Journal of pyrenaicus in a lowland catchment. Journal of Fish Animal Ecology 72: 777–783. Biology 42: 875–888. Haury, J., D. Ombredane & J. L. Banglinie´re, 1991. L0habitat Museth, J., R. Borgstrøm & J. E. Brittain, 2010. Diet overlap de la truite commune (Salmo trutta L.) en cours d0eau. In between introduced European minnow (Phoxinus phoxi- Baglinie`re, J. L. & G. Maisse (eds), La Truite: Biologie et nus) and young brown trout (Salmo trutta) in the lake, E´ cologie. INRA Editions, Paris: 121–149. Øvre Heimdalsvatn: a result of abundant resources or Heino, J., 2005. Functional biodiversity of macroinvertebrate forced niche overlap? Hydrobiologia 642: 93–100. assemblages along major ecological gradients of boreal Neveu, A., 1980. Relations entre le benthos, la de´rive, le rythme headwater streams. Freshwater Biology 50: 1578–1587. alimentaire et le taux de consommation des truites com- Hernando, J. A. & M. C. Soriguer, 1992. Biogeography of munes (Salmo trutta L.) en canal expe´rimental. Hydrobi- freshwater fish of the Iberian Peninsula. Limnetica 8: ologia 76: 217–228. 243–253. Newson, M. D. & C. L. Newson, 2000. Geomorphology, Hesthagen, T., R. Saksga˚rd, O. Hegge, B. K. Dervo & ecology and river channel habitat: mesoscale approaches J. Skurdal, 2004. Niche overlap between young brown to basin-scale challenges. Progress in Physical Geography trout (Salmo trutta) and Siberian sculpin (Cottus poecil- 24: 195–217. opus) in a subalpine Norwegian river. Hydrobiologia 521: Novakowski, G. C., N. S. Hahn & R. Fugi, 2008. Diet sea- 117–125. sonality and food overlap of the fish assemblage in a Hilderbrand, R. H. & J. L. Kershner, 2004. Influence of habitat pantanal pond. Neotropical Ichthyology 6: 567–576. type on food supply, selectivity, and diet overlap of Ochs, G., 1969. The ecology and ethology of whirligig beetles. Bonneville Cutthroat Trout and Nonnative Brook Trout in Archiv fu¨r Hydrobiologie 37: 375–404. Beaver Creek, Idaho. North American Journal of Fisheries Oscoz, J., M. C. Escala & F. Campos, 2000. La alimentacioń Management 24: 33–40. de la trucha comuń(Salmo trutta L., 1758) en un rıóde Hill, J. & G. D. Grossman, 1993. An energetic model of Navarra (N. Espanã). Limnetica 18: 29–35. microhabitat use for rainbow trout and rosyside dace. Oscoz, J., P. M. Leunda, F. Campos, M. C. Escala & Ecology 74: 685–698. R. Miranda, 2005. Diet of 0? brown trout (Salmo trutta Ivlev, V. S., 1961. Experimental ecology of the feeding of L., 1758) from the river Erro (Navarra, North of Spain). fishes. Translated from the Russian by Douglas Scott. Limnetica 24: 319–326. Yale University Press, New Haven. Power, G., 1992. Seasonal growth and diet of juvenile Chinook Jensen, H., T. Bøhn, P.-A. Amundsen & P. E. Aspholm, 2004. salmon (Oncorhynchus tshawytscha) in demonstration Feeding ecology of piscivorous brown trout (Salmo trutta channels and the main channel of the Waitaki river, New L.) in a subarctic watercourse. Annales Zoologici Fennici Zealand, 1982–1983. Ecology of Freshwater Fish 1: 41: 319–328. 12–25. Johnson, R. L., S. M. Coghlan & T. Harmon, 2007. Spatial and Rader, R. B., 1997. A functional classification of the drift: traits temporal variation in prey selection of brown trout in a that influence invertebrate availability to salmonids. cold Arkansas tailwater. Ecology of Freshwater Fish 16: Canadian Journal of Fisheries and Aquatic Sciences 54: 373–384. 1211–1234. Kahilainen, K. & H. Lehtonen, 2002. Brown trout (Salmo Ranta, E. & V. Kaitala, 1991. School size affects individual trutta L.) and Arctic charr (Salvelinus alpinus (L.)) as feeding success in three-spined sticklebacks (Gasterosteus predators on three sympatric whitefish (Coregonus lav- aculeatus L.). Journal of Fish Biology 5: 733–737. aretus (L.)) forms in the subarctic Lake Muddusja¨rvi. Rincoń, P. A. & J. Loboń-Cervia´, 1993. Microhabitat use by Ecology of Freshwater Fish 11: 158–167. stream-resident brown trout: bioenergetic consequences. L’Abeé-Lund, J. H., A. Langeland & H. Sægrov, 1992. Pisci- Transactions of the American Fisheries Society 122: vory by brown trout Salmo trutta L. and Arctic charr 575–587. Salvelinus alpinus (L.) in Norwegian lakes. Journal of Rincoń, P. A. & J. Loboń-Cervia´, 1995. Use o fan encounter Fish Biology 41: 91–101. model to predict size-selective predation by a stream- Loboń-Cervia, J. & P. A. Rincoń, 1994. Trophic ecology of red dwelling cyprinid. Freshwater Biology 33: 181–191. roach (Rutilus arcasii) in a seasonal stream; an example of Rincoń, P. A. & J. Loboń-Cervia´, 1999. Prey size selection by detritivory as a feeding tactic. Freshwater Biology 32: brown trout (Salmo trutta L.) in a stream in northern 123–132. Spain. Canadian Journal of Zoology 77: 755–765. MacNeil, C., R. W. Elwood & J. T. A. Dick, 2000. Factors Rıó-Barja, F. J. & F. Rodrı´guez-Lestega´s, 1992. Os Rıós influencing the importance of Gammarus spp. (Crustacea: Galegos. MorfoloxıáeRe´xime. Concello da Cultura Amphipoda) in riverine salmonid diets. Archivfu¨r Galega, Santiago de Compostela. Hydrobiologie 149: 87–107. Rodrı´guez-Capı´tulo, A., I. Munõz, N. Bonada, A. Gaude´s&S. Magalhaes, M. F., 1992. Feeding ecology of the Iberian Cyp- Tomanova, 2009. La biota de los rıós: los invertebrados. rinid Barbus bocagei Steindachner, 1865 in a lowland In Elosegi, A. & S. Sabater (eds), Conceptos y tećnicas en river. Journal of Fish Biology 40: 123–133. ecologıá fluvial. Fundacioń BBVA, Bilbao: 253–270. Magalha˜es, M. F., 1993a. Feeding of an Iberian stream cypri- Rodrı´guez-Jimeńez, A. J., 1987. Relaciones tro´ficas de una nid assemblage: seasonality of resource use in a highly comunidad´ ıctica, durante el estıóenelrıó Aljuceń variable environment. Oecologia 96: 253–260. (Extremadura, Espanã). Miscel la`nia Zoolo`gica 11: 249–256. 123 132 Hydrobiologia (2011) 667:119–132

Sagar, P. M. & G. J. Glova, 1995. Prey availability and diet of fish assemblage. In Stevens, M. (ed), Intertidal and Basin- juvenile brown trout (Salmo trutta) in relation to riparian Wide Habitat Use of Fishes in the Scheldt Estuary: 37–59. willows (Salix spp.) in three New Zealand streams. New Tachet, H., P. Richoux, M. Bournaud & P. Usseglio-Polatera, Zealand Journal of Marine & Freshwater Research 29: 2002. Inverte´bre´s d’eau douce (2nd Corrected Impres- 527–537. sion). CNRS editions, Parı´s. Sańchez, J., 2009. Biologıá de la alimentacioń de la trucha Teixeira, A. & R. M. V. Cortes, 2006. Diet of stocked and wild comuń(Salmo trutta Linne´, 1758) en los rıós de Galicia. trout, Salmo trutta: is there competition for resources? Thesis. Universidad de Santiago de Compostela. Folia Zoologica 55: 61–73. Sańchez-Gonza´les, S., G. Ruiz-Campos & S. Contreras-Bal- Valladolid, M. & M. Przybylski, 1996. Feeding relations deras, 2001. Feeding ecology and habitat of the threespine among cyprinids in the Lozoya River (Madrid, central stickleback, Gasterosteus aculeatus microcephalus,ina Spain). Polskie Archiwum Hydrobiologii 43: 213–223. remnant population of northwestern Baja California, Vidal Romani, J. R., 1989. Granite geomorphology in Galicia Me´xico. Ecology of Freshwater Fish 10: 191–197. (NW Espanã). Cuadernos Laboratorio Xeolo´xico de Laxe Sandlund, O. T., J. Museth, T. F. Næsje, S. Rognerud, R. 13: 89–163. Saksga˚rd, T. Hesthagen & R. Borgstrøm, 2010. Habitat Vieira, N. K. M., N. L. Poff, D. M. Carlisle, S. R. Moulton, M. use and diet of sympatric Arctic charr (Salvelinus alpinus) L. Koski & B. C. Kondratieff, 2006. A database of lotic and whitefish (Coregonus lavaretus) in five lakes in invertebrate traits for North America. U.S. Geological southern Norway: not only interspecific population dom- Survey Data Series 187 [available on internet at http://pubs. inance? Hydrobiologia 650: 27–41. water.usgs.gov/ds187]. Schoener, T. W., 1970. Nonsynchronous spatial overlap of Wallace Jr., R. K., 1981. An assessment of diet overlap lizards in patchy habitats. Ecology 51: 408–418. indexes. Transactions of the American Fisheries Society Schoener, T. W., 1989. The ecological niche. In Cherrett, J. M. 110: 72–76. (ed), Ecological Concepts. Blackwell, Oxford: 79–107. Wiens, J. A., 1993. Fat times, lean times and competition Statzner, B. & L. A. Beˆche, 2010. Can biological invertebrate among predators. Trends in Ecology & Evolution 8: traits resolve effects of multiple stressors on running water 348–349. ecosystems? Freshwater Biology 55: 80–119. Yant, P. R., J. R. Karr & P. Angermeier, 1984. Stochasticity in Stevens, M., J. Maes & F. Ollevier, 2006. Taking potluck: stream fish communities: an alternative interpretation. trophic guild structure and feeding strategy of an intertidal American Naturalist 124: 573–582.

123