**4. Diet changes with age: Food resource partitioning and change in piscivorous behaviour**

In Salmonid populations, dominant fishes may exclude less aggressive individuals, limiting their access to resources within patches. For example, dominant Atlantic salmon *Salmo salar* Linnaeus, 1758 may exclude subordinates from high-quality patches by intimidation or direct aggression [38,39]. However, subordinate fish may gain access to food by using high-quality patches when dominants are absent [40] or may be constrained to foraging in marginal areas [41]. When brown trout and Atlantic salmon co-occurred, trout has been observed to be dominant over salmon, holding feeding stations by swimming actively in the central regions of food patches, whereas salmon occupied the margins, generally remaining stationary on the stream bed [41]. Moreover, in habitats in which food is patchily distributed in time and space, fishes can benefit by moving between patches [42,43], with subordinate animals moving little in comparison with dominant fishes [39].

During their life history brown trout undergo ontogenetic habitat shifts [44 and references therein] due to changes in habitat selection operating at multiple spatial scales [44]. These shifts during fish life stage transitions may be accompanied by a marked reduction in intra-specific competition in the fish population, facilitating the partitioning of resources [e.g. 45,46]. Moreover, dietary analyses usually show high values of diet overlap among age classes, but the differences in the use of feeding habitat and behavioural feeding habits are important adaptive features that may reduce the intra-specific competition in the population [23]. Thus, although the diet comparison among age classes can show a remarkable similarity in their prey utilization patterns, sometimes the high overlap values may not indicate competition, since fishes can adopt different strategies to overcome competence, i.e. resource partitioning among age classes can occur at five different levels: (1) diet composition; (2) prey selection; (3) prey size; (4) habitat utilization for feeding; and (5) niche breadth. Also stomach fullness can vary among age classes as shown in section 4.6.

#### **4.1. Changes in diet composition with age**

**Figure 2.** Prey selectivity according to Ivlev's selectivity index of newly emerged brown trout fry in the River Iso (NW

The feeding diversity of juvenile fishes is generally greater than during the larval period, and there is often an increase in the importance of species-specific dietary traits [4]. However, recent studies have demonstrated that at the moment of complete yolk absorption, the fry shows a dramatic shift in niche breadth, which might be related to the improvement of swimming and handling ability of fry for capturing and ingesting both aquatic invertebrates and aerial

A common practice in many countries associated with river restoration is the rehabilitation of spawning sites with different techniques [37], but recently different authors have emphasized the importance of the complete recolonization of spawning grounds by benthic macroinver‐ tebrates, including first instars, in order to assure the presence of the required amount of prey for the feeding of young fry after restoration works [13]. Hence, at the moment of hatching, a

Spain) (modified from [13]).

276 New Advances and Contributions to Fish Biology

imagoes [13].

In brown trout, as in many other fish species, there is normally a change in the diet composition during the life of the fish. Thus, juveniles mainly consume prey items linked to the bottom of the river, many of them interstitial, i.e. living among grains of sand or gravel. Opposite, terrestrial invertebrates and fishes are important resources for large trouts. The contribution of these food items to fish diets increases with predator size or age because larger fish can feed on a wider range of preys as shown in Figures 3 and 4. Within a population, the percentage of the most important prey items change with age. In one study of a river in Italy the percentage of plecopteran nymphs in the diet tended to increase with the individual's age [17]. In another study, *Baetis* spp. dominated in all age classes in different proportions, whereas the percentage of caddisflies with cases, (*Allogamus* sp.) and mayflies (*Ecdyonurus* spp.) tended to increase with age [23]. Thus, as shown in Table 1, each age class consumes significantly different prey items, Chironomidae being the most frequently consumed prey item in age 0+ and age 3+ (44.41% and 29.84% respectively) and aerial imagoes of Ephemeroptera in age 1+ and age 2+ (77.49% and 49.08% respectively).

**Age 0+ Age 1+ Age 2+ Age 3+**

**(%)**

**Benthos Drift Benthos Drift Benthos Drift Benthos Drift**

**Ivlev Diet**

**(%)**

http://dx.doi.org/10.5772/54133

**Ivlev**

279

**Ivlev Diet**

**Diet (%)**

**Aquatic invertebrates**

**Terrestrial invertebrates**

**Ivlev Diet**

**(%)**

Oligochaeta gen. sp. 0.14 -0.68 1 0 -1 — 0 -1 — 0 -1 — Ancylidae 2.07 0.04 1 0.15 -0.85 1 0 -1 — 0 -1 — Hydrobiidae 2.48 0.90 0.82 2.11 0.88 0.79 0.92 0.74 0.58 4.84 0.95 0.90 Lymnaeidae 1.66 -0.18 -0.01 1.96 -0.10 0.07 20.64 0.79 0.85 29.03 0.85 0.89 Sphaeriidae 0 -1 — 0.53 0.59 1 1.38 0.82 1 0 -1 — Hydracarina gen. sp. 0.41 0.64 -0.61 0.23 0.43 -0.76 0 -1 -1 0 -1 -1 Ostracoda gen. sp. 0.55 1 1 0.23 1 1 0 — — 0 — — Baetidae 1.93 0.49 -0.06 0.15 -0.63 -0.87 8.72 0.86 0.60 0 -1 -1 Caenidae 5.10 0.94 0.50 0.60 0.59 -0.48 0.46 0.49 -0.57 0 -1 -1 Ephemerellidae 0 -1 -1 0.23 -0.77 -0.86 2.29 0.14 -0.12 0 -1 -1 Ephemeridae 0 — — 0.08 1 1 0 — — 0 — — Leuctridae 0.55 0.85 1 0.08 0.25 1 0.46 0.82 1 0 -1 — Aeshnidae 0 — — 0.08 1 1 0 — — 0 — — Calopterygidae 6.07 1 1 3.78 0.45 0.17 8.72 0.72 0.53 0 -1 -1 Coenagrionidae 0.28 0.61 1 0 -1 — 0 -1 — 0 -1 — Gomphidae 0.14 -0.13 1 0 -1 — 0 -1 — 0 -1 — Aphelocheiridae 0.41 -0.27 -0.28 0.08 -0.81 -0.81 0 -1 -1 0.81 0.05 0.05 Gerridae 0.14 1 -0.28 0 — -1 0 — -1 0.81 1 0.54 Sialidae 0.14 1 1 0.23 1 1 0 — — 0 — — Dytiscidae 0 — — 0 — — 0.46 1 1 0 — — Elmidae 0.41 -0.58 -0.56 0.15 -0.83 -0.81 0.46 -0.55 -0.52 0 -1 -1 Brachycentridae 0 -1 — 0.08 -0.64 1 0 -1 — 0 -1 — Hydropsychidae 2.76 -0.61 0.17 0.45 -0.92 -0.62 0.92 -0.85 -0.36 0.81 -0.87 -0.41 Leptoceridae 0.28 1 1 0 — — 0 — — 0 — — Limnephilidae 0.55 0.92 1 0.45 0.91 1 0.46 0.91 1 0.81 0.95 1 Philopotamidae 0.14 0.51 1 0 -1 — 0 -1 — 0 -1 — Polycentropodidae 0 -1 — 0.30 0.25 1 0 -1 — 0 -1 — Rhyacophilidae 0 -1 — 0 -1 — 0.46 0.49 1 0 -1 — Sericostomatidae 0.14 1 1 0.15 1 1 0.46 1 1 0 — — Chironomidae 44.41 0.31 0.18 4.08 -0.71 -0.76 1.83 -0.86 -0.89 29.84 0.12 -0.01 Simuliidae 15.17 -0.51 -0.18 1.89 -0.92 -0.84 0.92 -0.96 -0.92 12.90 -0.56 -0.26 Tipulidae 0 -1 — 0.08 0.54 1 0 -1 — 0 -1 —

Ontogenetic Dietary Shifts in a Predatory Freshwater Fish Species: The Brown Trout as an Example of a Dynamic...

Ephemeroptera gen. sp. 12.14 — 1 77.49 — 1 49.08 — 1 2.42 — 1 Trichoptera gen. sp. 0.14 — -0.28 0.45 — 0.30 0 — -1 0 — -1 Chironomidae 0.69 — -0.93 2.04 — -0.82 0 — -1 13.71 — -0.20 Simuliidae 0.69 — 1 0.68 — 1 0 — — 2.42 — 1

Regarding the changes in the proportion of terrestrial invertebrates consumed by trouts during the ontogeny, previous studies have demonstrated that aquatic invertebrates dominated the diet in all age classes [23,47], but it seems clear that terrestrial invertebrates were more frequently consumed by older trout [17,23,48], terrestrial organisms being important prey during warmer seasons in salmonids [49,50].

**Figure 3.** Percentage of terrestrial invertebrates (in terms of relative abundance) consumed by each age class of *Salmo trutta* in the River Anllóns (NW Spain) during summer. Error bars represent 95% confidence intervals.

**Figure 4.** Diet composition consumed by each age class of *Salmo trutta* in the River Lengüelle (NW Spain) during summer.

Ontogenetic Dietary Shifts in a Predatory Freshwater Fish Species: The Brown Trout as an Example of a Dynamic... http://dx.doi.org/10.5772/54133 279

of caddisflies with cases, (*Allogamus* sp.) and mayflies (*Ecdyonurus* spp.) tended to increase with age [23]. Thus, as shown in Table 1, each age class consumes significantly different prey items, Chironomidae being the most frequently consumed prey item in age 0+ and age 3+ (44.41% and 29.84% respectively) and aerial imagoes of Ephemeroptera in age 1+ and age 2+

Regarding the changes in the proportion of terrestrial invertebrates consumed by trouts during the ontogeny, previous studies have demonstrated that aquatic invertebrates dominated the diet in all age classes [23,47], but it seems clear that terrestrial invertebrates were more frequently consumed by older trout [17,23,48], terrestrial organisms being important prey

**Figure 3.** Percentage of terrestrial invertebrates (in terms of relative abundance) consumed by each age class of *Salmo*

**Figure 4.** Diet composition consumed by each age class of *Salmo trutta* in the River Lengüelle (NW Spain) during

*trutta* in the River Anllóns (NW Spain) during summer. Error bars represent 95% confidence intervals.

(77.49% and 49.08% respectively).

278 New Advances and Contributions to Fish Biology

summer.

during warmer seasons in salmonids [49,50].



age classes (Figure 5), displaying evidence of a generalist strategy for these prey (e.g. *Chimarra*

Ontogenetic Dietary Shifts in a Predatory Freshwater Fish Species: The Brown Trout as an Example of a Dynamic...

http://dx.doi.org/10.5772/54133

281

**Figure 5.** Feeding strategy diagram. (1) Explanatory diagram of the modified Costello method according to Amund‐

Finally, although in some occasions the diet comparison among age classes can show a remarkable similarity in their prey utilization patterns, sometimes the high overlap values may not indicate competition, since the differences in the behavioural feeding habits are important adaptive features that may reduce the intra-specific competition in the population [23]. Previously, the differences in the behavioural feeding habits among age classes had been studied using prey trait analysis [23], obtaining important advances to disentangle food resource partitioning among cohorts. Details and information needed for the elaboration of prey trait analysis can be found in the introduction section and bibliography [27-29]. Thus, in the reference [23] 'diel drift behaviour' trait showed that age-2+ is clearly separated from the other age classes. Age-2+ tended to feed on prey with no tendency in diel drift behaviour as shown by the presence of *Ancylus fluviatilis*, Gerridae, Coleoptera and Formicidae in their stomachs. On the contrary, age-0+ preferred to feed on prey with nocturnal diel drift behaviour tendency due to the presence of Philopotamidae (*C. marginata*). Age-0+ showed a wider

sen and collaborators [20]. Data are presented for each age class.

*marginata*, Simuliidae (adult), Chironomidae (adult) and *Ophiogomphus* sp.).

**Table 1.** Diet composition and prey selection according to Ivlev's selectivity index in each age class of *Salmo trutta* in the River Furelos (NW Spain) during summer.

The maximum and mean prey size eaten generally increases with size in predatory fish species [51]. Piscivorous behaviour is most frequent in large brown trout, and studies show that it occurs in older individuals with a size of 20–30 cm. Trout in smaller size classes have rarely/ never been recorded eating other fish [23,26,52,53]. In contrast, Sánchez-Hernández and collaborators have recorded piscivorous behaviour in an age-0 trout (the individual found was 8.5 cm in fork length) [29]. This behaviour could be related with the hypothesis of Mittelbach and Persson, who stated that fish species that had larger mouth gapes became piscivorous at younger ages and at smaller sizes [51], and demonstrating that mouth gape is not a limitation to use fishes by small trouts [29]. One possible advantage for small trout in diversifying into eating fish is a reduction in competition with other individuals in the same size classes.

The feeding strategy among age classes may be illustrated with graphical methods, such as the modified Costello graphical method [20] based on the relative importance of prey species. Figure 5, shows plots of prey abundance (*Ai*) against frequency of occurrence of prey in the diet (*Fi*) for three different age classes. The plots show feeding strategy differs between age classes, varying in degrees of specialization and generalization on different prey types. For all age classes, *Baetis* spp. were the most important prey, always being eaten by more than eightyfive percent of the individuals (*Fi* = 87.5% in age-1+ to 100% in age-2+) and represented a high contribution in specific abundance (*Ai* = 23.93% in age-1+ to 68.95% in age-0+). However, the majority of the prey items presented low values for both *Fi* and *Ai* (lower left quadrant) for all age classes (Figure 5), displaying evidence of a generalist strategy for these prey (e.g. *Chimarra marginata*, Simuliidae (adult), Chironomidae (adult) and *Ophiogomphus* sp.).

**Age 0+ Age 1+ Age 2+ Age 3+**

**(%)**

**Benthos Drift Benthos Drift Benthos Drift Benthos Drift**

**Ivlev Diet**

**(%)**

**Ivlev**

**Ivlev Diet**

**Diet (%)**

280 New Advances and Contributions to Fish Biology

**Other prey items**

the River Furelos (NW Spain) during summer.

**Ivlev Diet**

**(%)**

Arachnida gen. sp. 0 — — 0.08 — — 0 — — 0 — — Acanthosomatidae 0 — — 0.08 — 1 0.46 — 1 0 — — Psyllidae 0 — — 0.15 — 1 0 — — 0 — — Chloropidae 0.14 — -0.28 0 — -1 0 — -1 0 — -1 Diptera gen. sp. 0 — -1 0.08 — -1 0 — -1 0 — -1 Syrphidae 0 — -1 0 — -1 0 — -1 0 — -1 Xylomyidae 0 — — 0.15 — 1 0 — — 0 — — Cynipidae 0 — -1 0 — -1 0 — -1 0.81 — 0.25 Formicidae 0 — -1 0 — -1 0 — -1 0.81 — 0.54 Carabidae 0 — — 0.08 — 1 0 — — 0 — — Chrysomelidae 0.28 — -0.28 0 — -1 0 — -1 0 — -1 Coleoptera gen. sp. 0 — — 0 — — 0.46 — 1 0 — —

*Pseudochondrostoma duriense* 0 — — 0.45 — — 0.46 — — 0 — — Eggs 0 — — 0.15 — 1 0 — — 0 — —

**Table 1.** Diet composition and prey selection according to Ivlev's selectivity index in each age class of *Salmo trutta* in

The maximum and mean prey size eaten generally increases with size in predatory fish species [51]. Piscivorous behaviour is most frequent in large brown trout, and studies show that it occurs in older individuals with a size of 20–30 cm. Trout in smaller size classes have rarely/ never been recorded eating other fish [23,26,52,53]. In contrast, Sánchez-Hernández and collaborators have recorded piscivorous behaviour in an age-0 trout (the individual found was 8.5 cm in fork length) [29]. This behaviour could be related with the hypothesis of Mittelbach and Persson, who stated that fish species that had larger mouth gapes became piscivorous at younger ages and at smaller sizes [51], and demonstrating that mouth gape is not a limitation to use fishes by small trouts [29]. One possible advantage for small trout in diversifying into eating fish is a reduction in competition with other individuals in the same size classes.

The feeding strategy among age classes may be illustrated with graphical methods, such as the modified Costello graphical method [20] based on the relative importance of prey species. Figure 5, shows plots of prey abundance (*Ai*) against frequency of occurrence of prey in the diet (*Fi*) for three different age classes. The plots show feeding strategy differs between age classes, varying in degrees of specialization and generalization on different prey types. For all age classes, *Baetis* spp. were the most important prey, always being eaten by more than eightyfive percent of the individuals (*Fi* = 87.5% in age-1+ to 100% in age-2+) and represented a high contribution in specific abundance (*Ai* = 23.93% in age-1+ to 68.95% in age-0+). However, the majority of the prey items presented low values for both *Fi* and *Ai* (lower left quadrant) for all

**Figure 5.** Feeding strategy diagram. (1) Explanatory diagram of the modified Costello method according to Amund‐ sen and collaborators [20]. Data are presented for each age class.

Finally, although in some occasions the diet comparison among age classes can show a remarkable similarity in their prey utilization patterns, sometimes the high overlap values may not indicate competition, since the differences in the behavioural feeding habits are important adaptive features that may reduce the intra-specific competition in the population [23]. Previously, the differences in the behavioural feeding habits among age classes had been studied using prey trait analysis [23], obtaining important advances to disentangle food resource partitioning among cohorts. Details and information needed for the elaboration of prey trait analysis can be found in the introduction section and bibliography [27-29]. Thus, in the reference [23] 'diel drift behaviour' trait showed that age-2+ is clearly separated from the other age classes. Age-2+ tended to feed on prey with no tendency in diel drift behaviour as shown by the presence of *Ancylus fluviatilis*, Gerridae, Coleoptera and Formicidae in their stomachs. On the contrary, age-0+ preferred to feed on prey with nocturnal diel drift behaviour tendency due to the presence of Philopotamidae (*C. marginata*). Age-0+ showed a wider distribution of values in the fuzzy principal component analysis (FPCA) of the 'trajectory' trait; age-0+ tended to feed on prey with oscillatory and by random trajectory (Lumbriculidae, Simuliidae and Chironomidae), whereas age-2+ tended to feed on prey with lineal trajectory due to the presence of Gerridae [23]. On the other hand, no clear differences have been found in the traits 'tendency to drift in the water column' and 'tendency to drift at the water surface' for prey among age classes [23], however as shown in Figure 6 (fuzzy principal component analysis calculated with the values reported in Table 1), age-3+ is clearly separated from the other age classes, showing that the differences ability to feed at different depths of the water column is possible among cohorts.

abundance and biomass of invertebrates drifting during the day describe the potential prey available to juvenile brown trout better than abundance and biomass of benthic invertebrates do [55]. Different authors have shown that active choice guided by energetic optimization criteria appeared to be of limited importance in determining the size composition of prey eaten by trouts [11]. These authors also stated that the operating mechanisms of prey-size selection are probably not independent of the characteristics of the size-frequency distribution of the available prey. Moreover, in brown trout spatial and temporal variations in prey selection are possible, due to the preference for the different prey items related to the site-specific prey accessibility [12]. As shown in section 3, the size-frequency distribution of potential prey items in the benthos was different to that of prey in the stomachs of newly emerged brown trout [13], a result that agrees with observations made in other fish species [e.g. 54] and observations made in age-0+ individuals of *S. trutta* during the autumn [29]. In spite of the high abundance of Elmidae in the benthos, this prey item was negatively selected [29]. The rejection of the elmid beetle may be due to their low energetic value, as they have an intense sclerotisation, but it may also be due to their bad taste [56-58]. Hence, other factors, besides prey abundance, including site-specific prey accessibility, prey size, energetic selection criteria and prey preference of fishes, play an important role in the feeding behaviour of freshwater fishes [54].

Ontogenetic Dietary Shifts in a Predatory Freshwater Fish Species: The Brown Trout as an Example of a Dynamic...

http://dx.doi.org/10.5772/54133

283

Fochetti and collaborators found a high preference for species of Trichoptera by trout younger than 3+, a preference for plecopteran species by those older than three years, and a general negative avoidance for species of Ephemeroptera by all age classes [17]. As shown in Table 1, prey selection is clearly different among age classes and important prey items such as Chiro‐ nomidae in age 0+ and age 3+ and aerial imagoes of Ephemeroptera in age-1+ and age-2+ are positively selected in the benthos and drift respectively. Overall, as different researchers have demonstrated, the mechanisms involved in prey selection among age classes in the same population are complex and may be related at different levels: (1) prey selection in fishes is related to prey characteristics, such as size, locomotor skills, accessibility or anti-predator behaviour, (2) prey selection in fishes is related to fish characteristics, like prior experience, locomotor skills, stomach fullness, mouth gape, sensory capabilities and fish size and (3) prey selection in fishes is related to physical habitat characteristics, as flow patterns and structural

Ontogenetic dietary shifts may also occur at the level of prey size [23,54]. Steingrímsson and Gíslason showed a consistent, but moderate, shift towards larger prey with increased body size in brown trout [59]. Several researchers have found that mean prey size increases as predator size increases [e.g. 23,48,54,60]. The size-frequency distributions of the available terrestrial prey were always greatly dominated (75–90%) by the two smallest size classes (1–2 and 2–3 mm long), prey over 4 mm long being extremely scarce, while size distributions of aquatic prey were less skewed [11]. Thus, in general it is correct to state that trout fed mainly on prey within 1–4 mm size range [11,48], with 2–3 mm prey being the most commonly consumed [48]. On the contrary, Rincón and Lobón-Cerviá showed that organisms 1–2 mm long were generally the most numerous [11]. By age classes the average prey size consumption

complexity of habitat [12,13,34,54].

**4.3. Changes in prey size with age**

**Figure 6.** Biplot of gut contents obtained from a fuzzy principal component analysis (FPCA) based on behavioural feeding habits of the four age classes. (1) Similarity results among age classes according to the gut contents of Table 1. Data are presented for each age class. 0+: age-0+, 1+: age-1+, 2+: age-2+ and 3+: age-3+. Ellipses envelop weighted average of prey taxa positions consumed by age classes: Labels (0+, 1+, 2+ and 3+) indicate the gravity centre of the ellipses (2) Factorial correspondence analysis between both traits and prey items and their spatial distribution with histogram of eigenvalues. Details needed for the elaboration of these graphics can be found in the introduction sec‐ tion and bibliography [27-29].

#### **4.2. Changes in prey selection with age**

Prey abundance should be an important factor involved in prey choice. However, fishes do not always consume the most abundant taxa available in the environment [13,29,54]. The total abundance and biomass of invertebrates drifting during the day describe the potential prey available to juvenile brown trout better than abundance and biomass of benthic invertebrates do [55]. Different authors have shown that active choice guided by energetic optimization criteria appeared to be of limited importance in determining the size composition of prey eaten by trouts [11]. These authors also stated that the operating mechanisms of prey-size selection are probably not independent of the characteristics of the size-frequency distribution of the available prey. Moreover, in brown trout spatial and temporal variations in prey selection are possible, due to the preference for the different prey items related to the site-specific prey accessibility [12]. As shown in section 3, the size-frequency distribution of potential prey items in the benthos was different to that of prey in the stomachs of newly emerged brown trout [13], a result that agrees with observations made in other fish species [e.g. 54] and observations made in age-0+ individuals of *S. trutta* during the autumn [29]. In spite of the high abundance of Elmidae in the benthos, this prey item was negatively selected [29]. The rejection of the elmid beetle may be due to their low energetic value, as they have an intense sclerotisation, but it may also be due to their bad taste [56-58]. Hence, other factors, besides prey abundance, including site-specific prey accessibility, prey size, energetic selection criteria and prey preference of fishes, play an important role in the feeding behaviour of freshwater fishes [54].

Fochetti and collaborators found a high preference for species of Trichoptera by trout younger than 3+, a preference for plecopteran species by those older than three years, and a general negative avoidance for species of Ephemeroptera by all age classes [17]. As shown in Table 1, prey selection is clearly different among age classes and important prey items such as Chiro‐ nomidae in age 0+ and age 3+ and aerial imagoes of Ephemeroptera in age-1+ and age-2+ are positively selected in the benthos and drift respectively. Overall, as different researchers have demonstrated, the mechanisms involved in prey selection among age classes in the same population are complex and may be related at different levels: (1) prey selection in fishes is related to prey characteristics, such as size, locomotor skills, accessibility or anti-predator behaviour, (2) prey selection in fishes is related to fish characteristics, like prior experience, locomotor skills, stomach fullness, mouth gape, sensory capabilities and fish size and (3) prey selection in fishes is related to physical habitat characteristics, as flow patterns and structural complexity of habitat [12,13,34,54].

#### **4.3. Changes in prey size with age**

distribution of values in the fuzzy principal component analysis (FPCA) of the 'trajectory' trait; age-0+ tended to feed on prey with oscillatory and by random trajectory (Lumbriculidae, Simuliidae and Chironomidae), whereas age-2+ tended to feed on prey with lineal trajectory due to the presence of Gerridae [23]. On the other hand, no clear differences have been found in the traits 'tendency to drift in the water column' and 'tendency to drift at the water surface' for prey among age classes [23], however as shown in Figure 6 (fuzzy principal component analysis calculated with the values reported in Table 1), age-3+ is clearly separated from the other age classes, showing that the differences ability to feed at different depths of the water

**Figure 6.** Biplot of gut contents obtained from a fuzzy principal component analysis (FPCA) based on behavioural feeding habits of the four age classes. (1) Similarity results among age classes according to the gut contents of Table 1. Data are presented for each age class. 0+: age-0+, 1+: age-1+, 2+: age-2+ and 3+: age-3+. Ellipses envelop weighted average of prey taxa positions consumed by age classes: Labels (0+, 1+, 2+ and 3+) indicate the gravity centre of the ellipses (2) Factorial correspondence analysis between both traits and prey items and their spatial distribution with histogram of eigenvalues. Details needed for the elaboration of these graphics can be found in the introduction sec‐

Prey abundance should be an important factor involved in prey choice. However, fishes do not always consume the most abundant taxa available in the environment [13,29,54]. The total

column is possible among cohorts.

282 New Advances and Contributions to Fish Biology

tion and bibliography [27-29].

**4.2. Changes in prey selection with age**

Ontogenetic dietary shifts may also occur at the level of prey size [23,54]. Steingrímsson and Gíslason showed a consistent, but moderate, shift towards larger prey with increased body size in brown trout [59]. Several researchers have found that mean prey size increases as predator size increases [e.g. 23,48,54,60]. The size-frequency distributions of the available terrestrial prey were always greatly dominated (75–90%) by the two smallest size classes (1–2 and 2–3 mm long), prey over 4 mm long being extremely scarce, while size distributions of aquatic prey were less skewed [11]. Thus, in general it is correct to state that trout fed mainly on prey within 1–4 mm size range [11,48], with 2–3 mm prey being the most commonly consumed [48]. On the contrary, Rincón and Lobón-Cerviá showed that organisms 1–2 mm long were generally the most numerous [11]. By age classes the average prey size consumption is different with higher values in age-2+ (8.4 mm ± 1.62) than age-0+ (4.2 mm ± 0.25) and age-1+ (5.9 mm ± 0.51), but there are no significant differences between age-0+ and age-1+ [23]. Figure 7 illustrates the age-related variation in prey size, showing that mean prey size tends to increase with age.

ities [64]. Special attention has been given to ontogenetic shift in habitat preference in brown trout populations [e.g. 44,65-67]. It is well-known that in brown trout populations habitat use changes during ontogeny, preferring deeper and slower flowing water as they increased in

Ontogenetic Dietary Shifts in a Predatory Freshwater Fish Species: The Brown Trout as an Example of a Dynamic...

http://dx.doi.org/10.5772/54133

285

Habitat patches used by brown trout can be monitored by radio telemetry [e.g. 68] and although these studies have shown that brown trout feed on young white suckers *Catostomus commersonii* (Lacepède, 1803) at night in shallow habitats, little information was obtained about the habitat used for feeding. Also, microhabitat use of freshwater fishes has been studied by snorkel observations in previous studies [e.g. 69]. This methodology could be used to study feeding habitat requirements of fish species. However, there are still gaps to be filled before snorkel surveys can be fully adopted in fish diet studies. In fact, one of the main disadvantages of this approach is the need for good visibility. This has been one of the main handicaps because, although brown trout normally yields a bimodal (crepuscular) pattern of activity with a major peak at dawn and a lower one around dusk [70], brown trout can also feed at night [71]. Another limiting factor for the application of snorkel surveys to study feeding habitat require‐ ments is related to the physical characteristic of the river such as current, depth or turbidity. It is well-known that different macroinvertebrates have different preferences for habitats [72] and so prey trait analysis has been proposed as a functional approach to understand mecha‐ nisms involved in predator-prey relationships [27-29]. Consequently it may be useful for understanding inter-species interactions and the mechanisms that determine food partitioning between them [29,30]. Nowadays, they have been recently used to provide interesting results about differences in feeding habitat requirements among age classes in brown trout [23], for example, young of the year (0+) tend to capture prey living in moderate current velocities, whereas other age classes (1+ and 2+) tend to feed on prey living in fast current velocities [23].

A previous study on habitat choice in a littoral zone of Lake Tesse (Norway) showed that small trout had a strong association with the bottom and larger trout occurred more frequently higher up in the water column, and suggested that this difference in vertical distribution was also reflected in food choice [73]. In streams, competition among fish species may also be reduced by vertical segregation [e.g. 41,69,74]. In spite that terrestrial invertebrates, as component of the diet, are more important in adults than juveniles of brown trout [23], these same authors found that the ability to feed at different depths of the water is similar among age classes. Nevertheless, it is possible that different age classes of fish may become vertically segregated by concentrating on different prey types living in different parts of the water column as shown in section 4.1. Hence, additional studies are needed in order to clarify whether vertical segregation among cohorts is related to the ability to feed at different depths

Finally differences in the use of feeding habitat are important adaptive features that may reduce the intra-specific competition in the population. In this context, fuzzy principal component analysis (FPCA) has shown that age-0+ tended to feed on prey living in moderate current velocities, although overlap was higher between age-1+ and age-2+, preferring to feed on prey living in fast current velocities [23]. Moreover, these researchers have found that age-0+ showed a higher spectrum of prey, which revealed a greater ability to prey on different

size [e.g. 44,65].

of the water column.

In conclusion, prey-size selection is probably dependent on the characteristics of the sizefrequency distribution of the available prey [11], and the size-related differences in the diet of trout can be related to gape-limitations, increasing mean prey size and maximum prey-size with trout size [48].

**Figure 7.** Box plots of the age-related variation in prey size of *Salmo trutta* in the River Furelos (NW Spain) during summer. The solid line within each box represents the median, the bottom and top borders indicate the 25th and 75th percentiles, the notches represent the 95% confidence intervals.

#### **4.4. Changes in the habitat used for feeding with age**

In fishes, patches used for feeding and refuges are normally different, as shown by several researchers [61-63], due to brown trout being a habitat generalist. Also, patterns in habitat selection have been shown to be driven by physical and environmental factors operating at multiple spatial scales [44].

Many organisms exhibit ontogenetic shifts in their diet and habitat use, which often exert a large influence on the structure and expected dynamics of food webs and ecological commun‐ ities [64]. Special attention has been given to ontogenetic shift in habitat preference in brown trout populations [e.g. 44,65-67]. It is well-known that in brown trout populations habitat use changes during ontogeny, preferring deeper and slower flowing water as they increased in size [e.g. 44,65].

is different with higher values in age-2+ (8.4 mm ± 1.62) than age-0+ (4.2 mm ± 0.25) and age-1+ (5.9 mm ± 0.51), but there are no significant differences between age-0+ and age-1+ [23]. Figure 7 illustrates the age-related variation in prey size, showing that mean prey size tends to increase

In conclusion, prey-size selection is probably dependent on the characteristics of the sizefrequency distribution of the available prey [11], and the size-related differences in the diet of trout can be related to gape-limitations, increasing mean prey size and maximum prey-size

**Figure 7.** Box plots of the age-related variation in prey size of *Salmo trutta* in the River Furelos (NW Spain) during summer. The solid line within each box represents the median, the bottom and top borders indicate the 25th and

In fishes, patches used for feeding and refuges are normally different, as shown by several researchers [61-63], due to brown trout being a habitat generalist. Also, patterns in habitat selection have been shown to be driven by physical and environmental factors operating at

Many organisms exhibit ontogenetic shifts in their diet and habitat use, which often exert a large influence on the structure and expected dynamics of food webs and ecological commun‐

75th percentiles, the notches represent the 95% confidence intervals.

**4.4. Changes in the habitat used for feeding with age**

multiple spatial scales [44].

with age.

with trout size [48].

284 New Advances and Contributions to Fish Biology

Habitat patches used by brown trout can be monitored by radio telemetry [e.g. 68] and although these studies have shown that brown trout feed on young white suckers *Catostomus commersonii* (Lacepède, 1803) at night in shallow habitats, little information was obtained about the habitat used for feeding. Also, microhabitat use of freshwater fishes has been studied by snorkel observations in previous studies [e.g. 69]. This methodology could be used to study feeding habitat requirements of fish species. However, there are still gaps to be filled before snorkel surveys can be fully adopted in fish diet studies. In fact, one of the main disadvantages of this approach is the need for good visibility. This has been one of the main handicaps because, although brown trout normally yields a bimodal (crepuscular) pattern of activity with a major peak at dawn and a lower one around dusk [70], brown trout can also feed at night [71]. Another limiting factor for the application of snorkel surveys to study feeding habitat require‐ ments is related to the physical characteristic of the river such as current, depth or turbidity. It is well-known that different macroinvertebrates have different preferences for habitats [72] and so prey trait analysis has been proposed as a functional approach to understand mecha‐ nisms involved in predator-prey relationships [27-29]. Consequently it may be useful for understanding inter-species interactions and the mechanisms that determine food partitioning between them [29,30]. Nowadays, they have been recently used to provide interesting results about differences in feeding habitat requirements among age classes in brown trout [23], for example, young of the year (0+) tend to capture prey living in moderate current velocities, whereas other age classes (1+ and 2+) tend to feed on prey living in fast current velocities [23].

A previous study on habitat choice in a littoral zone of Lake Tesse (Norway) showed that small trout had a strong association with the bottom and larger trout occurred more frequently higher up in the water column, and suggested that this difference in vertical distribution was also reflected in food choice [73]. In streams, competition among fish species may also be reduced by vertical segregation [e.g. 41,69,74]. In spite that terrestrial invertebrates, as component of the diet, are more important in adults than juveniles of brown trout [23], these same authors found that the ability to feed at different depths of the water is similar among age classes. Nevertheless, it is possible that different age classes of fish may become vertically segregated by concentrating on different prey types living in different parts of the water column as shown in section 4.1. Hence, additional studies are needed in order to clarify whether vertical segregation among cohorts is related to the ability to feed at different depths of the water column.

Finally differences in the use of feeding habitat are important adaptive features that may reduce the intra-specific competition in the population. In this context, fuzzy principal component analysis (FPCA) has shown that age-0+ tended to feed on prey living in moderate current velocities, although overlap was higher between age-1+ and age-2+, preferring to feed on prey living in fast current velocities [23]. Moreover, these researchers have found that age-0+ showed a higher spectrum of prey, which revealed a greater ability to prey on different macrohabitats, whereas age-1+ and age-2+ preferred to feed on epibenthic prey living in erosional macrohabitats. In our case with the values reported in Table 1 and as shown in Figure 8, 'current velocity' trait shows no clear differences for prey of the four age classes. On the contrary, 'macrohabitat' trait shows that age-1+ has the most ample spectrum for this trait, tending to feed on epibenthic prey living in erosional macrohabitats, whereas age-3+ tends to feed on prey items available in the water column (Figure 8).

shifts from soft-bodied to hard-shelled prey and decreased animal prey breadth [76]. In contrast, several other researchers have found that niche breadth increases with body size [e.g. 23,77]. Oscoz and collaborators found that larger fish have a higher number of potential prey items available and a wider niche breadth, as indicated by their higher trophic diversity index values [77]. As mentioned in section 3, analysis of diet changes on newly emerged brown trout fry suggests a dramatic shift in niche breadth at the moment of complete yolk absorption, which might be related to the improvement the fry's swimming and handling ability in capturing and ingesting prey [13]. In a recent study on brown trout populations, it has been demonstrated that niche breadth, measured as Levin's index, increases with fish length in *Salmo trutta*. However, no differences were found in the Shannon diversity and evenness

Ontogenetic Dietary Shifts in a Predatory Freshwater Fish Species: The Brown Trout as an Example of a Dynamic...

http://dx.doi.org/10.5772/54133

287

As can be seen in Figure 9; the number of different prey types consumed by S. trutta increases during ontogeny. Age-0+ shows the smallest prey spectrum, whereas in age-1+ a significant increase in the prey types consumed by juveniles is observed. However, no significant differences in the dietary niche among ages-1+, 2+ and 3+ are observed in Figure 9. Hence, although no clear results have been observed in the variation of the diversity indices of prey among age classes [23], it could be argued that there is a tendency to increase dietary niche with increasing length or age, at least when the niche breadth of juveniles (0+), subadults (1+)

**Figure 9.** Age-related variation in the number of different prey types consumed by *Salmo trutta* in the River Lengüelle

Several researchers have found in different fish species that the stomach fullness index (defined as the weight of the stomach contents (in grams) divided by the weight of the predator (in

(NW Spain) during summer. Error bars represent the 95% confidence intervals.

**4.6. Changes in the fullness index with age**

indices of prey eaten among age classes [23].

and adults (≥2+) is compared.

**Figure 8.** Biplot of gut contents obtained from a fuzzy principal component analysis (FPCA) based on preferential hab‐ itat utilization for feeding of the four age classes. (1) Similarity results among age classes according to the gut con‐ tents of Table 1. Data are presented for each age class. 0+: age-0+, 1+: age-1+, 2+: age-2+ and 3+: age-3+. Ellipses envelop weighted average of prey taxa positions consumed by age classes: Labels (0+, 1+, 2+ and 3+) indicate the gravity centre of the ellipses. (2) Factorial correspondence analysis between both traits and prey items and their spatial distribution with histogram of eigenvalues. Details needed for the elaboration of these graphics can be found in the introduction section and bibliography [27-29].

#### **4.5. Changes in the niche breadth with age**

Deady and Fives showed that niche breadth decreases with fish length in corkwing wrasse, *Symphodus (Crenilabrus) melops* (Linnaeus, 1758), indicating an increase in dietary specializa‐ tion with increasing length [75]. Magalhães found dietary shifts throughout the ontogeny in an endemic cyprinid of the Iberian Peninsula (*Squalius pyrenaicus* (Günther, 1868)), including shifts from soft-bodied to hard-shelled prey and decreased animal prey breadth [76]. In contrast, several other researchers have found that niche breadth increases with body size [e.g. 23,77]. Oscoz and collaborators found that larger fish have a higher number of potential prey items available and a wider niche breadth, as indicated by their higher trophic diversity index values [77]. As mentioned in section 3, analysis of diet changes on newly emerged brown trout fry suggests a dramatic shift in niche breadth at the moment of complete yolk absorption, which might be related to the improvement the fry's swimming and handling ability in capturing and ingesting prey [13]. In a recent study on brown trout populations, it has been demonstrated that niche breadth, measured as Levin's index, increases with fish length in *Salmo trutta*. However, no differences were found in the Shannon diversity and evenness indices of prey eaten among age classes [23].

As can be seen in Figure 9; the number of different prey types consumed by S. trutta increases during ontogeny. Age-0+ shows the smallest prey spectrum, whereas in age-1+ a significant increase in the prey types consumed by juveniles is observed. However, no significant differences in the dietary niche among ages-1+, 2+ and 3+ are observed in Figure 9. Hence, although no clear results have been observed in the variation of the diversity indices of prey among age classes [23], it could be argued that there is a tendency to increase dietary niche with increasing length or age, at least when the niche breadth of juveniles (0+), subadults (1+) and adults (≥2+) is compared.

**Figure 9.** Age-related variation in the number of different prey types consumed by *Salmo trutta* in the River Lengüelle (NW Spain) during summer. Error bars represent the 95% confidence intervals.

#### **4.6. Changes in the fullness index with age**

macrohabitats, whereas age-1+ and age-2+ preferred to feed on epibenthic prey living in erosional macrohabitats. In our case with the values reported in Table 1 and as shown in Figure 8, 'current velocity' trait shows no clear differences for prey of the four age classes. On the contrary, 'macrohabitat' trait shows that age-1+ has the most ample spectrum for this trait, tending to feed on epibenthic prey living in erosional macrohabitats, whereas age-3+ tends to

**Figure 8.** Biplot of gut contents obtained from a fuzzy principal component analysis (FPCA) based on preferential hab‐ itat utilization for feeding of the four age classes. (1) Similarity results among age classes according to the gut con‐ tents of Table 1. Data are presented for each age class. 0+: age-0+, 1+: age-1+, 2+: age-2+ and 3+: age-3+. Ellipses envelop weighted average of prey taxa positions consumed by age classes: Labels (0+, 1+, 2+ and 3+) indicate the gravity centre of the ellipses. (2) Factorial correspondence analysis between both traits and prey items and their spatial distribution with histogram of eigenvalues. Details needed for the elaboration of these graphics can be found in the

Deady and Fives showed that niche breadth decreases with fish length in corkwing wrasse, *Symphodus (Crenilabrus) melops* (Linnaeus, 1758), indicating an increase in dietary specializa‐ tion with increasing length [75]. Magalhães found dietary shifts throughout the ontogeny in an endemic cyprinid of the Iberian Peninsula (*Squalius pyrenaicus* (Günther, 1868)), including

feed on prey items available in the water column (Figure 8).

286 New Advances and Contributions to Fish Biology

introduction section and bibliography [27-29].

**4.5. Changes in the niche breadth with age**

Several researchers have found in different fish species that the stomach fullness index (defined as the weight of the stomach contents (in grams) divided by the weight of the predator (in grams) and multiplied by 100) varies during ontogeny [78,79]. In salmonids the results are contradictory: in brook charr *Salvelinus fontinalis* (Mitchill, 1814) no differences have been found in the stomach fullness [80], whilst other researchers have demonstrated that the stomach fullness of brown trout varies among size classes [47]. Brown trout between the size of 40 mm and 320 mm fed more intensively, whilst the intensity declined above 320 mm length [47]. In the River Furelos (NW Spain), we have found that stomach fullness during the summer is different among age classes (Kruskal-Wallis test; *p* < 0.001), being higher in age-0+ (9% ± 0.64) than age-1+ (1.1% ± 0.14), age-2+ (1% ± 0.25) and age-3+ (1.1% ± 0.24) (all Mann–Whitney U test, *p* < 0.001) but no differences have been found between ages-1+, 2+ and 3+ (all Mann– Whitney U test, *p* > 0.05). Moreover, stomach fullness decreases with fish size (*r* = -0.72; *p* < 0.001) (unpublished data). Hence, stomach fullness can vary among age classes; however additional studies are needed in order to clarify whether stomach fullness varies during ontogeny in brown trout.
