Introduction

Foraging theory has received considerable attention across a diverse range of fauna over the last four decades (see Pyke 1984), including some classic work on fishes (e.g. Werner and Hall 1974). Knowledge of the feeding ecology of fishes is fundamental both to understanding the processes that function at the individual, population and community levels, and for the management and conservation of their populations and habitats. Furthermore, the factors that influence the acquisition and assimilation of food can have significant consequences for the condition, growth, survival and recruitment of fishes. This is of particular importance during early development, when fishes are invariably most vulnerable to predation, competition, disease and environmental perturbations (Mills and Mann 1985; Myers 1995; Houde 1997; Ondračková et al. 2002; Nunn et al. 2003, 2007a, 2010a; Longshaw et al. 2005, 2010). Indeed, important factors regulating the growth and survival of young fishes include food availability and their ability to capture and ingest food items, with a lack of suitable prey potentially causing either reduced growth or starvation, both of which influence survival (Cushing 1990; Mayer and Wahl 1997; Beaugrand et al. 2003; Dickmann et al. 2007; Burrow et al. 2011; Stige et al. 2011). Notwithstanding, the feeding ecology of young fishes, particularly larvae and 0+ juveniles, is poorly understood compared with older individuals. Indeed, there are a large number of species for which there is no information on their foraging ecology during the larval and juvenile periods. For example, there appears to be no primary literature for more than 90% of the freshwater fish species recorded in Europe and information is even more sparse for other regions, particularly for deep-water marine fishes (Kottelat and Freyhof 2007; Froese and Pauly 2011). This paper synthesises the current knowledge of the feeding ecology of larval and 0+ juvenile teleost fishes, using pertinent examples mainly from temperate latitudes in the northern and southern hemispheres, and identifies areas where further research is required.

The diets of larval and juvenile fishes

The ecology of yolk-sac feeding in fishes has been thoroughly reviewed by Kamler (2002) and is not repeated here. Moreover, the nutrition of yolk-feeding fishes is, necessarily, determined by the quantity and quality of yolk reserves, which are intrinsically linked to various maternal attributes (e.g. body size). Most researchers concur that the switch from purely endogenous to exogenous nutrition is a, if not the, key event in the survival of fishes. The majority of marine and freshwater fish species are planktivorous at the onset of exogenous feeding, and many exhibit mixed feeding, with exogenous nutrition commencing before complete absorption of the yolk reserves. Zooplankton serve as a vital food resource for the larvae of most fish species (Tables 1, 2; limited to European species for brevity). Rotifers are frequent prey of the majority of cyprinids, especially in the early larval period, with copepods, cladocerans, insects and phytoplankton also important in the diet (Table 1; Fig. 1a). By contrast, most centrarchids and percids prey upon cladocerans, copepods and insect larvae, while salmonids mainly consume cladocerans and copepods (in lentic environments) or insect larvae (in lotic environments), and the diets of clupeid, gadid, gobiid and osmerid larvae are characterised mainly by various developmental stages of copepods (eggs, nauplii, copepodites, adults) (Tables 1, 2; Fig. 1a). Protozoans may also be important prey for fish larvae, especially in marine ecosystems (Montagnes et al. 2010).

Table 1 The diets of the larvae and juveniles of key fish species in European fresh waters
Full size table
Table 2 The diets of the larvae and juveniles of key fish species in European inshore marine and transitional waters
Full size table
Fig. 1
figure 1

Correspondence analysis plots comparing the diet composition of a larval and b juvenile marine and freshwater fishes (from Tables 1, 2)

Full size image

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 (Tables 1, 2). The majority of juvenile cyprinids consume cladocerans, copepods and insect larvae, with aufwuchs (the periphyton and associated microfauna that grow on underwater surfaces) and adult dipterans also important in some species, while centrarchids and percids prey mainly upon cladocerans, copepods, insect larvae, isopods, amphipods and fishes (Table 1; Fig. 1b). The diets of juvenile salmonids are invariably dominated by larval, pupal or adult insects, although cladocerans and copepods are important in coregonids, while clupeids, gadids and osmerids consume copepods, isopods, amphipods, mysids, euphausiids, decapods and fishes (Tables 1, 2; Fig. 1b).

Even fish species that are piscivorous (e.g. European seabass (Dicentrarchus labrax (L.), Moronidae), northern pike (Esox lucius L., Esocidae)), herbivorous (e.g. common nase (Chondrostoma nasus (L.), Cyprinidae), butterfish (Odax pullus (Forster), Odacidae)) or detritivorous (e.g. thin-lipped grey mullet (Liza ramada (Risso), Mugilidae), fathead minnow (Pimephales promelas Rafinesque, Cyprinidae)) as juveniles or adults may initially be zooplanktivorous (Ferrari and Chieregato 1981; Aprahamian and Barr 1985; Clements and Choat 1993; Bry et al. 1995; Reckendorfer et al. 2001; Herwig and Zimmer 2007). Notable exceptions include fishes inhabiting the upper, fast-flowing reaches of rivers, where plankton is usually rare, such as many salmonid species. In such habitats, diets are invariably dominated by insects and crustaceans captured from the river bed, drift or water surface (Allen 1941; Frost 1950; Wańkowski and Thorpe 1979; Williams 1981; Zimmerman and Mosegaard 1992; Keeley and Grant 1997; Rincón and Lobón-Cerviá 1999). Calanoid copepods, mysids, euphausiids and decapods tend to be more abundant in the diets of marine than of freshwater species, largely because they are usually most abundant or only present in marine environments, whereas rotifers, cyclopoid copepods, cladocerans and insects are most important in the diets of freshwater species (Table 3). The larvae and juveniles of fishes inhabiting tidal rivers or upper estuaries (e.g. twaite shad (Alosa fallax (Lacépède), Clupeidae), Atlantic herring (Clupea harengus L., Clupeidae), European smelt (Osmerus eperlanus (L.), Osmeridae), European flounder (Platichthys flesus (L.), Pleuronectidae), common goby (Pomatoschistus microps (Krøyer), Gobiidae), three-spined stickleback (Gasterosteus aculeatus L., Gasterosteidae)) frequently have diets containing both marine and freshwater taxa (Beaumont and Mann 1984; Aprahamian 1989; Weatherley 1989; Thiel et al. 1996; Nunn et al. 2008a). Fishes can also be important prey, especially during the juvenile period and in species with large mouths (e.g. gadids, percids, centrarchids; Tables 1, 2, 3; Mittelbach and Persson 1998).

Table 3 Similarity percentages (SIMPER) analysis (Clarke and Warwick 2001) of the mean abundances of key prey taxa and their contributions (%) to dissimilarities in the diets of freshwater and marine larval and juvenile fishes (from Tables 1, 2)
Full size table

Prey selection

Optimal foraging theory predicts that predators should select prey that maximise the energetic gains available in relation to the energetic costs of capturing, ingesting and digesting the prey (Pyke 1984). For a fish predator, the energetic content of a given prey type increases with prey size, but there is also an associated increase in handling time (Lazzaro 1987). Handling times also vary between prey taxa, although few studies have succeeded in isolating the effects of taxa (e.g. morphology, behaviour) and size in prey selection by larval and juvenile fishes (Ghan and Sprules 1993; Anderson 1994). Prey selection is thus determined by the relative profitabilities of particular types and sizes of prey, with maximal profitabilities conferred by prey types and sizes that provide the maximum energy gain per unit handling time (Pyke 1984).

A key factor in prey selection is the relative availability (≈ search time) of particular prey. Townsend et al. (1986), for example, observed that juvenile roach (Rutilus rutilus (L.), Cyprinidae) switched from planktonic to non-planktonic cladocerans with a decline in the availability of planktonic prey, and similar responses have been observed for other fish species, including common bream (Abramis brama (L.), Cyprinidae) (Persson and Brönmark 2002a), Eurasian perch (Perca fluviatilis L., Percidae) (Persson and Greenberg 1990), yellow perch (Perca flavescens (Mitchill), Percidae) (Wu and Culver 1992), silver perch (Bidyanus bidyanus (Mitchell), Terapontidae) (Warburton et al. 1998), walleye (Sander vitreus (Mitchill), Percidae) (Hoxmeier et al. 2004), European flounder (Aarnio et al. 1996) and redfish (Sebastes spp., Sebastidae) (Anderson 1994), as well as for roach switching from zooplankton to aufwuchs (Garner 1996a; Mann et al. 1997; Vašek et al. 2006; Nunn et al. 2007b) and for Eurasian perch, pikeperch (Sander luciperca (L.), Percidae) and walleye becoming piscivorous (Mehner et al. 1996; Galarowicz et al. 2006). Indeed, Hoogenboezem et al. (1992) used zooplankton density to predict a switch between particulate and filter feeding in common bream. Conversely, switches in prey selection may not occur if zooplankton densities remain high.

The diets of larval and juvenile fishes can include both an increase in prey size and changes in prey species as they grow and develop (Ghan and Sprules 1993; Arrhenius 1996; Dickmann et al. 2007; Nunn et al. 2007b, c). Gape has often been identified as a key morphological trait limiting the prey size of fish that swallow prey whole (e.g. Lazzaro 1987; DeVries et al. 1998; Mehner et al. 1998; Sabatés and Saiz 2000). The small gape of larval and juvenile fishes limits their prey size, yet within constraints set by gape, food eaten influences growth and, ultimately, survival (Bremigan and Stein 1994). Indeed, Mayer and Wahl (1997) demonstrated that young walleye fed selectively and that, for at least some size classes, the selected prey increased growth or survival. Similarly, the switch of some species to piscivory is related to the relative sizes of the predator and their potential prey, and may have a direct positive effect on growth and survival, with piscivorous individuals being larger and in superior nutritional condition than non-piscivorous individuals at the end of the first year of life (Buijse and Houthuijzen 1992; Mehner et al. 1996; Bromley et al. 1997; Frankiewicz et al. 1999; Brabrand 2001; Pelham et al. 2001; Persson and Brönmark 2002b; Klemetsen et al. 2003; Galarowicz et al. 2006; Urbatzka et al. 2008). For example, Olson (1996) observed that rapid growth while feeding on invertebrates allowed 0+ largemouth bass (Micropterus salmoides (Lacépède), Centrarchidae) to gain a size advantage over 0+ bluegill (Lepomis macrochirus Rafinesque, Centrarchidae), facilitating a shift to piscivory. By contrast, slow growth while feeding on invertebrates reduced the size advantage of largemouth bass over bluegill and delayed or even prevented the shift to piscivory in the first year of life.

Fishes do not necessarily consume the largest prey possible, with Scott (1987) calculating that juvenile roach and common dace (Leuciscus leuciscus (L.), Cyprinidae) favour prey approximately 60% of their maximum gape, probably related to the increases in handling times associated with large prey (Wanzenböck 1995). Furthermore, there can be substantial interspecific variations, with Scharf et al. (2000) observing that the prey of 18 fish species off the north-east coast of the USA ranged from 10 to >50% of their body size. Gape size is less restrictive for larval and juvenile fishes that do not swallow prey whole or those, such as some cannibalistic species, that are able to partially ingest large prey (sometimes larger than themselves) and then regurgitate the remains (Baras 1999; Baras et al. 2000). Another factor influencing prey selection is the change in the effective profitability of particular types and sizes of prey in relation to fish satiation (Fig. 2; Turesson et al. 2006). For instance, three-spined stickleback have been shown to select specific sizes of prey until the stomach is full, whereafter smaller prey are selected (Gill and Hart 1998). Similarly, juvenile roach have been observed to initially feed upon large Daphnia sp. before switching to smaller individuals, with the switch occurring sooner in small than in large fish (Mikheev and Wanzenböck 1999).

Fig. 2
figure 2

Changes in the relative value of three sizes (small (solid line), optimum (stippled line), large (dashed line)) of prey with respect to fish satiation and motivation to feed. Changes in selectivity occur where the prey value lines intersect (modified from Gill 2003)

Full size image

Certain taxa may be consistently selected over others, irrespective of size, suggesting that taxa-specific characteristics are also important in the prey selection process. Nunn et al. (2007c), for example, observed that the young larvae of a number of cyprinid species selected rotifers over similar-sized copepod nauplii (Fig. 3), and Gliwicz et al. (2004) found that European smelt favoured elongated- (Daphnia spp., Diaphanosoma spp.) over compact-bodied (Bosmina spp., Chydorus spp.) cladocerans. Ghan and Sprules (1993) observed a similar phenomenon in young burbot (Lota lota (L.), Gadidae) larvae, with the globular rotifer Asplanchna sp. selected over spined rotifers and copepod nauplii, perhaps because of differences in morphology or behaviour between taxa; copepod nauplii are faster moving than rotifers, while the protuberances of spined rotifers may hinder fish attempting to prey upon them and increase handling times. Similarly, the extensive consumption of cladocerans by many freshwater fishes during the late larval and early juvenile periods is probably because of their ease of capture compared to copepods, with the latter being faster moving than cladocerans and able to detect the shock waves of approaching fishes (Winfield et al. 1983), although visibility may also be important (Mayer and Wahl 1997). The larvae of a number of marine fish species have been found to select the nauplii and copepodites of calanoid copepods over those of cyclopoid copepods, probably as calanoids swim more consistently and predictably than cyclopoids (Pepin and Penney 1997), and redfish larvae have been observed to select Calanus sp. eggs and nauplii over Oithona spp. copepodites, even though the latter were within the preferred size range (Anderson 1994). Furthermore, fishes may select egg-bearing over non-egg-bearing zooplankters, or even females with the largest clutches, probably because the opaque eggs increase their visibility (Winfield and Townsend 1983; Ghan and Sprules 1993; Gliwicz et al. 2004). Piscivorous individuals may also select particular species over others, most likely related to interspecific differences in prey morphology and the costs associated with their capture and ingestion (Nilsson and Brönmark 2000; Borcherding 2006).

Fig. 3
figure 3

Selection by seven developmental steps (six larval, one juvenile) of 0+ common dace for two size classes of rotifers (cross), copepod nauplii (triangle) and non-planktonic crustaceans (mostly Chydoridae; square) in the River Trent, England. The selection index ranges from −1 to +1, with negative values indicating avoidance, positive values indicating selection, and 0 representing no preference (modified from Nunn et al. 2007c)

Full size image

Ontogenetic shifts

There are two opposing schools-of-thought regarding the development of fishes. Subscribers to the theory of saltatory ontogeny (Balon 1979) assert that fishes with indirect development (i.e. those with a larval period) pass through a sequence of developmental steps characterised by biochemical, physiological, morphological, ethological and ecological traits that result in important shifts in resource use. By contrast, ‘gradualists’ contest that development occurs steadily, with no dramatic shifts in form and function. A comprehensive review of the debate is presented by Peňáz (2001). Irrespective of the process, the changes that occur during ontogeny, whether gradual or step-wise, often coincide with shifts in the diets of young fishes (Clements and Choat 1993; Vilizzi 1998; García-Berthou 2002; King 2005; Tonkin et al. 2006; Nunn et al. 2007b, 2008a). With respect to foraging ability, two of the key attributes are those pertaining to enhanced swimming performance, namely development of the fins, and improved vision (Lazzaro 1987; Schiemer et al. 1989). Once fins are fully developed, associated improvements in swimming performance allow a diversification of diet spectra, effectively increasing prey-capture efficiency. Similarly, improved vision increases the search volumes of larval and juvenile fishes (Fig. 4), effectively increasing prey availability (Wanzenböck and Scheimer 1989; Miller et al. 1993; Carvalho et al. 2002).

Fig. 4
figure 4

Differences in a the visual field volume of larval (solid line) and juvenile (dashed line) fishes, and b the reaction distance (the distance over which fish can see their prey) for small (solid line) and large (dashed line) prey. The faint small prey are beyond the visual field (for small prey), so the densities of small and large prey would be assessed as equal even though the real densities differ three-fold (modified from Gliwicz 2001)

Full size image

Newly hatched individuals of many fish species prey mainly upon small zooplankters and phytoplankton, partly because of their inability to handle larger prey (Whiteside et al. 1985; Krebs and Turingan 2003). In addition, larvae of the majority of fish species initially have poorly developed alimentary tracts, typically characterised by short length, narrow width, simple structure, weak digestive enzymes and, thus, limited digestive capacity (Kolkovski 2001; Makrakis et al. 2005). As the larvae develop, so their alimentary tracts develop, frequently characterised by an increase in length and width, differentiation of the gut into distinct regions and the production of potent digestive enzymes (Hofer and Uddin 1985; Junger et al. 1989). Development of the alimentary canal, concurrent with other changes in morphology, function and behaviour (Balon 1979; Werner and Gilliam 1984; Peňáz 2001), frequently coincides with shifts in the diet composition of young fishes.

A general sequence of shifts in resource use often occurs in the weeks after hatching. Moreover, distinct shifts in diet may occur that coincide with key intervals in ontogeny. Nunn et al. (2007b) differentiated two main ontogenetic shifts in the diets of a number of cyprinid and percid fishes in the lower River Trent, England, corresponding with the transition from apterolarvae (‘finfold’ larvae) to pterolarvae (‘finformed’ larvae), and from the larval to the juvenile period, probably linked to improvements in swimming performance and vision increasing the efficiency of prey capture. Similarly, King (2005) observed major dietary shifts in the larval, juvenile and adult periods of a number of floodplain fishes in Australia, and Clements and Choat (1993) reported differences in the diets of juvenile and adult butterfish off the coast of New Zealand. Such changes frequently coincide with shifts in habitat use or behaviour with development. Sunbleak (Leucaspius delineatus (Heckel), Cyprinidae), for example, exhibit a clear shift in diet at the transition from the larval to the juvenile period after moving from shallow river margins into open, deeper water (Pinder et al. 2005), and common carp (Cyprinus carpio L., Cyprinidae) switch from a planktonic to a benthic feeding mode at the onset of the juvenile period (Vilizzi 1998). Similarly, Demain et al. (2011) observed that juvenile Atlantic cod (Gadus morhua L., Gadidae) and haddock (Melanogrammus aeglefinus L., Gadidae) switched from mostly pelagic to benthic prey with a shift from shallow to deeper water. Ontogenetic shifts in the diets of smooth blenny (Lipophrys pholis (L.), Blenniidae) and Spanish toothcarp (Aphanius iberus (Valenciennes), Cyprinodontidae) have also been linked to changes in microhabitat use (Monteiro et al. 2005; Alcaraz and García-Berthou 2007).

Two of the best-studied and most widely distributed freshwater fish species in Europe are the roach and Eurasian perch. Roach larvae initially feed on pelagic rotifers, copepod nauplii and phytoplankton, with microcrustaceans and early instar chironomid larvae also eaten occasionally (Mark et al. 1987; Weatherley 1987; Matĕna 1995; Garner 1996a; Mann et al. 1997; Nunn et al. 2007b). Cladocerans (e.g. Bosminidae, Chydoridae, Daphniidae), copepods and periphytic invertebrates increase in importance in the diet as the larvae develop, although phytoplankton may be important if the abundance of other prey is low (Weatherley 1987). As 0+ juveniles, roach feed primarily upon zooplankton before switching to either benthic invertebrates or aufwuchs (Garner 1996a; Mann et al. 1997; Vašek et al. 2006; Nunn et al. 2007b, 2008b). Rotifers and copepod nauplii are the initial food items of perch larvae (Spanovskaya and Grygorash 1977; Mehner et al. 1997), with ciliates and algae also eaten occasionally (Guma’a 1978). Small cladocerans are eaten as the larvae increase in size, while Daphnia spp. become important after perch reach 10–15 mm, and Leptodora sp. and Bythotrephes spp. are consumed by >15-mm larvae (Guma’a 1978; Mehner et al. 1997). Copepods may also be common in the diet (Matĕna 1995; Nunn et al. 2007b; Kratochvíl et al. 2008), and are sometimes favoured over cladocerans (Guma’a 1978). In still waters, juvenile perch consume mainly planktonic cladocerans, especially Daphnia spp., Bosmina spp. and Leptodora sp., and copepods (Spanovskaya and Grygorash 1977; Vašek et al. 2006; Kratochvíl et al. 2008). By contrast, phytophilic and benthic invertebrates appear to be more important in rivers (Nunn et al. 2007b). In addition, some 0+ juvenile perch may become piscivorous and even cannibalistic (Brabrand 1995, 2001; Mehner et al. 1996; Mélard et al. 1996; van Densen et al. 1996; Borcherding et al. 2000; Beeck et al. 2002; Urbatzka et al. 2008).

The Atlantic cod is one of the most-studied fish species in the marine environment, and numerous investigations have demonstrated that cod larvae feed mainly on copepods (see Heath and Lough 2007). Small copepod nauplii are usually the main prey of the youngest cod larvae, with larger nauplii, copepodites and adult copepods (particularly Calanus spp., Paracalanus spp., Pseudocalanus spp., Temora spp., Acartia spp. and Oithona spp.) increasing in importance as the fish grow and develop (Voss et al. 2003; Rowlands et al. 2008). Juvenile cod feed upon a wide range of macroplankton, particularly copepods and euphausiids, and may also consume benthic invertebrates and fishes (Bromley et al. 1997; Pedersen and Fossheim 2008). Similarly, the diets of Atlantic herring larvae are dominated by zooplankton, particularly cladocerans (e.g. Bosmina sp., Pleopsis sp.) and copepods (e.g. Eurytemora sp., Acartia spp.), with copepods and euphausiids the main prey of 0+ juveniles (Arrhenius 1996; Pedersen and Fossheim 2008).

Diel variations

Larval and juvenile fishes may exhibit diel variations in behaviour (Gliwicz and Jachner 1992; Bromley et al. 1997; MacKenzie et al. 1999; Copp et al. 2005; Gliwicz et al. 2006; Nunn et al. 2010b). Mills et al. (1985), for example, observed that common dace larvae consumed up to four times more prey during daylight than at night, whereas Weatherley (1987) found that peak gut fullness occurred at night, while juvenile Eurasian perch, roach, chub (Squalius (syn. Leuciscus) cephalus (L.), Cyprinidae) and gizzard shad (Dorosoma cepedianum (L.), Clupeidae) are mostly diurnal and common bream are nocturnal (Winfield and Townsend 1988; Garner 1996b; Shepherd and Mills 1996; Okun et al. 2005; Kratochvíl et al. 2008). Alternatively, some species, such as haddock, adopt a crepuscular feeding regime, with activity being greatest in the morning and/or evening (MacKenzie et al. 1999). Notwithstanding, there may be ontogenetic shifts in diel feeding activity. For example, Bromley et al. (1997) observed that peak feeding activity in small juvenile Atlantic cod was from 12.00 to 20.00, whereas larger individuals foraged mainly from 00.00 to 08.00, with small and larger individuals, respectively, consuming invertebrates and fishes. There may also be diel variations in diet composition. Juvenile roach and chub, for instance, have been observed to consume fewer planktonic prey at night than in daylight, whereas aufwuchs increased in importance at night, suggesting a possible diel shift between particulate and benthic feeding modes (Garner 1996b). Similarly, although the diets of juvenile barbel (Barbus barbus (L.), Cyprinidae) and gudgeon (Gobio gobio (L.), Cyprinidae) are often dominated by chironomid larvae, the selection of some prey, such as simuliid larvae, may increase at night (Copp et al. 2005). Significantly, resource use may be substantially greater over the diel cycle than during daylight or darkness alone, suggesting that estimates of niche breadth based solely upon diurnal (or nocturnal) studies are potentially inaccurate (Copp 2008).

Diel shifts in diet composition are often linked to changes in habitat use. Nunn et al. (2010b), for example, observed that juvenile European bullhead (Cottus gobio L., Cottidae), stone loach (Barbatula barbatula (L.), Balitoridae) and gudgeon were mostly nocturnal, whereas European minnow (Phoxinus phoxinus (L.), Cyprinidae) and three-spined stickleback were mainly diurnal, possibly associated with directed movements to foraging habitats. Indeed, juvenile European hake (Merluccius merluccius (L.), Gadidae) inhabit the pelagial at night and the demersal during the day, which is reflected by diel qualitative and quantitative changes in prey consumption (Bozzano et al. 2005). Similarly, in deep lakes, larval and juvenile European smelt, Eurasian perch and roach often migrate into the epilimnion at dusk and the hypolimnion (smelt) or littoral (perch and roach) at dawn (Gliwicz and Jachner 1992). Another key factor influencing habitat use is the risk of predation, with diurnal species seeking refuge from predators at night and moving to profitable foraging habitats during daylight, while the opposite behaviour is characteristic of nocturnal species. Jacobsen and Berg (1998), for example, found that, under predation risk, there was significant diel variation in habitat use by juvenile Eurasian perch, suggesting a migration from open-water habitats at night into macrophytes in daylight. Similarly, shallow river margins are important night-time refuges from predation for small fishes (Baras and Nindaba 1999; Copp and Jurajda 1999), and it is possible that predator-induced shifts in habitat use may lead to diel changes in diet composition. Changes in habitat use by potential prey can also lead to diel variations in diet composition. In deep waterbodies, for example, many zooplankton species exhibit diel vertical migrations between the hypolimnion and epilimnion, seeking refuge in deep water from fish predation during daylight and grazing upon phytoplankton near the water surface at night (Gliwicz 1986). By contrast, zooplankton sometimes undertake horizontal migrations in shallow waterbodies (Lauridsen et al. 1996), which can also affect the prey availability to, and predation efficiency of, larval and juvenile fishes.

Seasonal variations

Populations of many zooplankton species vary on a seasonal basis, frequently related to fluctuations in food availability, predation pressure and environmental conditions (Sommer et al. 1986; Soetaert and Van Rijswijk 1993). Well-documented examples include planktonic cladocerans, such as Daphnia spp., which usually increase in abundance during the spring before experiencing population ‘crashes’ in mid-summer (Sommer et al. 1986). Seasonal successions in composition, abundance and biomass therefore occur in zooplankton communities, and comparable fluctuations are also a feature of many benthic or periphytic invertebrate communities, as well as phytoplankton, periphyton and macrophyte communities (Colebrook 1979; Angermeier 1982; Marker and Collett 1997a, b).

Seasonal fluctuations in prey abundance can influence the diet composition of 0+ fishes. This is particularly the case for juvenile fishes, which generally experience more than one season in their first year of life, whereas the larvae of most species experience only one season (exceptions are species that adopt multiple/fractional or protracted spawning strategies: Lowerre-Barbieri et al. 1998; Begg and Marteinsdottir 2000; Fuiman et al. 2005; Nunn et al. 2007d). Variations in the prey of juvenile European flounder in the Baltic Sea were related to seasonal changes in prey abundance, with macrofauna (oligochaetes, amphipods, chironomids) dominating the diet in spring and meiofauna (copepods) dominating in summer and autumn (Aarnio et al. 1996). Similarly, seasonal changes in the diet of pumpkinseed (Lepomis gibbosus (L.), Centrarchidae) have been linked to resource availability, with fish eggs and plant debris consumed in spring and summer and zooplankton eaten in the autumn (García-Berthou and Moreno-Amich 2000), while the diet of butterfish can be influenced by changes in algal composition (Clements and Choat 1993). Seasonal variations in diet composition imply that many fish species are able to feed opportunistically. Indeed, Townsend et al. (1986) observed that juvenile roach fed predominantly on planktonic cladocerans in spring and summer but non-planktonic cladocerans for the remainder of the year, with the switch coinciding with low availabilities of planktonic prey, and Anderson (1994) noted that the consumption of Oithona spp. copepodites by juvenile redfish increased as Calanus finmarchicus (Gunnerus) nauplii, the preferred prey, declined in abundance.

Dietary switches linked to seasonal variations in prey abundance have been reported for a range of 0+ fishes (e.g. Allen 1941; Persson 1983; Persson and Greenberg 1990; Wu and Culver 1992; Aarnio et al. 1996; Mehner et al. 1996; Mann et al. 1997; Warburton et al. 1998; Galarowicz et al. 2006) and may have implications for their growth and survival, especially if forced to switch to poor-quality food resources. At temperate latitudes, overwinter survival of many juvenile fish species is positively related to condition (a function of fish length), with smaller individuals less likely to survive than larger conspecifics, and numerous studies have reported a reduction in condition over the winter due to the utilisation of lipid reserves (see Hurst 2007). Food ration also has an influence on survival, with feeding fish more likely to surviving than starved individuals (Pangle et al. 2004). The extensive consumption of aufwuchs by many juvenile fishes in the autumn and winter is probably linked to a low availability of suitable animal prey (Nunn et al. 2008b). Indeed, while zooplankton populations in temperate fresh waters are invariably low throughout the autumn and winter (Sommer et al. 1986; Nunn et al. 2007e), aufwuchs is frequently still abundant after zooplankton populations have declined (Marker and Collett 1997b). Furthermore, although the juveniles of some species consume aufwuchs in mid- to late summer, the switch may be delayed or not occur if sufficient invertebrates are available (Weatherley 1987; Garner 1996a; Mann et al. 1997; Nunn et al. 2007b). As aufwuchs is considered a poor food resource because of its low digestibility and nutritive value (Bowen 1979; Persson 1983; Lemke and Bowen 1998), the combination of reduced feeding activity and poor food quality could have implications for the overwinter survival of juvenile fishes, particularly in areas or years that experience long or severe winters. This may also be an issue for other poor-quality food resources.

Annual variations

Larval and juvenile fishes can be highly susceptible to fluctuations in prey availability, with interannual variations in zooplankton abundance having potentially important consequences for their growth and survival (Mills and Forney 1988; Mayer and Wahl 1997; Dickmann et al. 2007; Burrow et al. 2011; Stige et al. 2011). The feeding success and growth of Arctic cod (Boreogadus saida (Lepechin), Gadidae) and sand lance (Ammodytes spp., Ammodytidae) larvae, for example, have been positively correlated with interannual differences in the density of copepod nauplii (Fortier et al. 1995). Furthermore, fluctuations in zooplankton dynamics have been implicated with long-term changes in Atlantic cod recruitment in the North Sea (Beaugrand et al. 2003). Indeed, the match/mismatch hypothesis (Cushing 1990) postulates that the survival of fish larvae will be highest in years when hatching coincides with peaks in plankton production (‘match’ scenario), with a mismatch in the timing of hatching and plankton blooms leading to high mortality (‘mismatch’ scenario) (Fig. 5), and a number of empirical studies of both marine (e.g. Atlantic cod, Arctic cod, sand lance) and freshwater (e.g. striped bass (Morone saxatilis (Walbaum), Moronidae)) fish species appear to support the theory (Fortier et al. 1995; Gotceitas et al. 1996; Chick and van den Avyle 1999). Food must therefore be available to larval and juvenile fishes in sufficient quantity, and of the correct size, with a lack of appropriately sized prey potentially causing either reduced growth or starvation (Beaugrand et al. 2003; Graeb et al. 2004; Dickmann et al. 2007).

Fig. 5
figure 5

The match/mismatch hypothesis (Cushing 1990), where a depicts a ‘match’ scenario and b and c depict ‘mismatch’ scenarios. The curves represent the abundance of fish larvae (solid line) and prey (dashed line)

Full size image

Numerous studies have reported interannual differences in zooplankton composition and abundance. Bass et al. (1997), for example, found that maximum rotifer densities in a regulated lowland river were usually around 2,000–3,000 L−1, but observed a peak of over 15,000 L−1 in one  year, and comparable interannual variations have been reported for a range of other taxa, both in marine and fresh waters (e.g. George and Taylor 1995; Reid et al. 1998; Hays et al. 2005; Lees et al. 2006). The factors causing interannual differences in zooplankton composition and abundance are generally the same as for seasonal variations, namely food availability, predation pressure and environmental conditions. Hydrological regime is a key factor governing interannual variations in riverine plankton populations, with abundance and biomass generally lowest in years with high river discharge (Bass et al. 1997; Marker and Collett 1997a). The coincidence of elevated river discharges with the larval period may thus negatively influence fish growth and survival either directly (through discharge-induced mortality) or indirectly (through reduced growth at lower water temperatures, discharge-associated increases in energy expenditure or reduced food availability) (Nunn et al. 2003, 2007a, 2010a). Conversely, a number of studies have demonstrated a negative influence of drought on benthic invertebrate populations, which may also affect the foraging ecology of larval and juvenile fishes, especially salmonids (e.g. Extence 1981; Cowx et al. 1984; Wood and Armitage 2004; Wood et al. 2010). Interannual differences in the composition and abundance of marine and lacustrine freshwater plankton have been linked to a range of complex and interacting factors, including regional changes in climate and local weather anomalies, making it difficult to identify the most important factors (George and Taylor 1995; Reid et al. 1998; Hays et al. 2005; Lees et al. 2006).

Interannual differences in the diet composition of larval and juvenile fishes have been related to variations in prey abundance. Variations in the diet of Murray cod (Maccullochella peelii peelii (Mitchell), Percichthyidae) larvae have been observed between years of high and low river discharge, with chironomid larvae dominating the diet when discharge was high and macrothricid cladocerans dominating when discharge was low (Kaminskas and Humphries 2009). Similarly, redfish larvae consumed a lower quantity and quality of prey in warm than in cool springs, probably because of a mismatch in the timing of fish hatching and the development of their preferred prey (Anderson 1994). Moreover, the availability of preferred, but not necessarily all, prey may influence fish growth, condition and survival. The growth and survival of small (<12 mm) yellow perch larvae, for example, was found to be greatest when feeding on adult copepods and nauplii, whereas adult copepods and small cladocerans were more important for larger larvae (Graeb et al. 2004), and the condition of redfish larvae was higher for individuals feeding on C. finmarchicus than for those feeding on Oithona spp. (Anderson 1994). The abundance, timing, duration and geographical distribution of marine plankton populations in particular can vary substantially on an annual basis, whereas the spawning periods of most fish species are less variable, meaning that there is a high risk of a mismatch in the timing of fish hatching and plankton blooms (Cushing 1990).

Spatial variations

The habitats used by fishes can have an important influence on prey availability and, therefore, their diet composition. The diets of marine and freshwater fishes differ largely because the majority of prey taxa are confined to either marine or freshwater environments, although the larvae and juveniles of euryhaline species may consume both marine and freshwater prey (Beaumont and Mann 1984; Aprahamian 1989; Weatherley 1989; Thiel et al. 1996; Nunn et al. 2008a). Mysids, euphausiids, tunicates, cirripedians, cumaceans, decapods and polychaetes tend to be more abundant in marine than in freshwater environments, whereas rotifers, cladocerans, oligochaetes and insects are most important in fresh water. Notwithstanding, there may be variations in diet composition within marine and freshwater ecosystems.

The distributions of plankton are intrinsically linked to the movement of water, from microscale turbulence to oceanic currents, although a range physicochemical factors, such as temperature, dissolved oxygen concentration and salinity, are also influential. In addition, zooplankton may undertake diel migrations between the epilimnion and hypolimnion or the littoral and pelagial, and often form swarms, resulting in a contagious distribution. Heath and Lough (2007) observed latitudinal patterns in the diet composition of Atlantic cod larvae, with fish at the northern edge of the species’ range consuming mainly C. finmarchicus, whereas those at the southern edge preyed upon Paracalanus spp. and Pseudocalanus spp. Furthermore, the diet composition and feeding activity of Atlantic cod and blue whiting (Micromesistius poutassou (Risso), Gadidae) larvae have been found to vary between water masses (e.g. warm vs. cool temperatures, high vs. low salinities) at a regional scale, and the direction and strength of the wind can also be important (McLaren et al. 1997; Hillgruber and Kloppmann 1999). Indeed, shifts in the diet composition of juvenile Arctic char (Salvelinus alpinus (L.), Salmonidae) have been observed to coincide with wind-induced changes in the distribution of zooplankton in Loch Ness, Scotland (Winfield et al. 2002).

The dominance of rotifers in the diets of many riverine fish larvae can be explained by their small size, poor mobility, ubiquity and high abundance compared with other zooplankton (Bass et al. 1997; Viroux 1997; Nunn et al. 2007e). Riverine zooplankton communities are invariably dominated by rotifers and small crustaceans throughout the year, with no marked development of large-bodied cladoceran and copepod populations; maximum rotifer densities in the regulated Great Ouse, England, were usually 2,000–3,000 L−1, whereas copepods (<125 L−1) and cladocerans (<10 L−1) were recorded at consistently lower densities (Bass et al. 1997). The diets of old larvae and juvenile fishes in riverine environments thus tend to be dominated by non-planktonic invertebrates, such as benthic crustaceans and insect larvae, or aufwuchs. By contrast, areas of reduced flow facilitate plankton development because of increased water-residence times. For example, the mean density of daphnids in the main channel of the lower River Trent was 0.03 L−1 compared with 76 L−1 in a connected pond (maximum >900 L−1) (Nunn et al. 2007e), and comparable results have been obtained elsewhere and for other taxa (e.g. Bass et al. 1997; Spaink et al. 1998; Tans et al. 1998). Areas such as backwaters and oxbow lakes thus provide superior feeding opportunities for planktivorous fishes compared with main river channels. Indeed, whereas rotifers and aufwuchs, respectively, were the main prey of ‘finformed’ larvae and juvenile roach in the main channel of the lower River Trent, cladocerans and cyclopoid copepods were dominant in a connected pond (Nunn et al. 2007b). Similarly, plankton densities are generally higher in river margins than in mid-channel because of increased water retention near the shore (Reckendorfer et al. 1999; Schiemer et al. 2001).

There may also be variations in diet composition at smaller spatial scales, both in marine and freshwater ecosystems. Significant differences in the diets of 0+ roach, chub, bleak (Alburnus alburnus (L.), Cyprinidae), common bream and gudgeon were observed between habitats in a lowland river (Garner 1998), and spatial variations in prey consumption have been documented in coastal populations of 0+ European flounder (Andersen et al. 2005; Grønkjaer et al. 2007). Similarly, the diets of eastern blue groper (Achoerodus viridis (Steindachner), Labridae) larvae in seagrass beds were dominated by tanaids, whereas fish from rocky reefs consumed harpacticoid copepods (Gillanders 1995), and newly hatched brook trout (Salvelinus fontinalis (Mitchill), Salmonidae) ate more dipteran larvae but fewer crustaceans and insect pupae/adults in fast- than in slow-flowing habitats (McLaughlin and Grant 1994). One of the key factors influencing the composition of benthic communities is sediment particle size, whereas physical structures are more important for epifauna and zooplankton. Aquatic macrophytes and macroalgae are particularly important for many larval and juvenile fishes as they provide refuge from predators and water currents, and habitat for phytophilic and planktonic invertebrates (Garner et al. 1996; Nagelkerken et al. 2000; Grenouillet et al. 2001a). Indeed, macrophytes invariably support larger numbers of invertebrates, greater ranges of body size and greater numbers of species than open water (Whiteside et al. 1985), and species composition and abundance may differ between macrophyte forms (e.g. submerged, emergent, floating-leaved) or species (Cyr and Downing 1988; Grenouillet et al. 2001a).

The habitats used by fishes can also have an important influence on their foraging efficiency. For example, Winfield (1986) demonstrated that the foraging efficiencies of juvenile roach and rudd (Scardinius erythrophthalmus (L.), Cyprinidae) decreased with increasing habitat complexity, largely because swimming speeds were lower in complex habitats, and Diehl (1988) found that prey consumption by Eurasian perch, common bream and roach was reduced in complex habitats. Furthermore, Diehl and Eklöv (1995) observed that perch in vegetated habitats had slow growth, partly because the physical complexity of submerged macrophytes reduced their foraging efficiency.

Inter- and intraspecific interactions

Interspecific interactions can have important implications for the growth, survival and recruitment success of fishes. Competitive effects occur when behavioural interactions cause an unequal distribution of a resource that is directly or indirectly related to growth, survival or recruitment (Wootton 1990). For example, fishes may alter their diets, and have lower growth rates, in the presence of competing species. Persson and Greenberg (1990) demonstrated that roach had a negative impact on the growth of juvenile Eurasian perch, with individual growth rates of perch decreasing with increasing roach density, which was related to competition for food resources; in the absence of roach, perch fed mainly upon planktonic cladocerans, whereas in the presence of roach they consumed copepods and macroinvertebrates. Similarly, Amundsen and Gabler (2008) found empirical evidence for food limitation and competition between juvenile Atlantic salmon (Salmo salar L., Salmonidae) and Alpine bullhead (Cottus poecilopus Heckel, Cottidae), resulting in reduced food acquisition and growth rates in salmon, and Scharf et al. (2009) observed that, when prey was limited, bluefish (Pomatomus saltatrix (L.), Pomatomidae) out-competed striped bass. Furthermore, Ward et al. (2002) demonstrated that chub preferred conspecific over heterospecific shoals, which was probably driven by interspecific competition with European minnow and the oddity effect. It should be noted, however, that the relative competitive abilities of fishes may vary between habitats (Winfield 1986; Diehl 1988).

Dietary overlap, which can be indicative of competition, is often greatest in the early larval period, probably as fishes are less morphologically and behaviourally differentiated than later in ontogeny (Garner 1996a). Indeed, a number of studies have found a high degree of overlap in the diets of young fish larvae in marine and fresh waters. The diets of many young cyprinid larvae, for example, are dominated by rotifers, while young clupeid, gadid and osmerid larvae prey mainly upon copepods (Garner 1996a; Thiel et al. 1996; Bromley et al. 1997; Nunn et al. 2007b; Pedersen and Fossheim 2008). Notwithstanding, overlap can also be substantial in the juvenile period. For example, the diets of juvenile Eurasian perch, pikeperch, ruffe (Gymnocephalus cernua (L.), Percidae), roach and common bream overlapped significantly in a lowland reservoir, suggesting strong interspecific competition for resources (Matĕna 1998), and juvenile Atlantic cod and whiting (Merlangius merlangus (L.), Gadidae) have been reported to compete for zooplankton in the North Sea (Shaw et al. 2008).

Although the diets of the larvae of many species are similar, they do not necessarily compete for resources. Early spawning species, for example, may already be juveniles by the time late-spawning species hatch: the larvae of early- and late-spawning species, therefore, are unlikely to compete for resources. Similarly, the larvae of rheophilic fish species rarely occur in the same habitats as limnophilic larvae, and competition between benthic and pelagic species is likely to be limited. Moreover, dietary overlap does not necessarily imply competition, as competition only occurs when demand for a resource exceeds the immediate supply (Zaret and Rand 1971; Wootton 1990). Overlap is often greatest among young larvae feeding upon the most abundant food resources (Garner 1996a; Thiel et al. 1996; Bromley et al. 1997; Nunn et al. 2007b; Pedersen and Fossheim 2008), leading some authors to suggest that competition is more likely to affect older larvae or juveniles consuming less abundant prey (Cowan et al. 2000). Indeed, resource partitioning is a characteristic feature of many juvenile fish assemblages, often coinciding with an increase in the importance of species-specific prey selection, with competition most likely among ecologically similar species or trophic guilds (Mark et al. 1987; Matĕna 1995; Garner 1996a; Demain et al. 2011).

Intraspecific interactions can also be important and, indeed, are inevitable as conspecifics invariably occupy identical niches (Ward et al. 2006). A number of studies have demonstrated a negative relationship between the density and growth of larval and juvenile fishes (e.g. Byström and García-Berthou 1999; Romare 2000), and the high densities of fishes inhabiting many nursery areas may increase the effects of intraspecific (and interspecific) interactions on feeding behaviour, which could have implications for the growth and survival of fishes. Grenouillet et al. (2001b), for example, stated that the survival of roach in the first year of life was density-dependent, and that intraspecific competition could negatively affect recruitment, while Cryer et al. (1986) and Perrow and Irvine (1992) demonstrated that the suppression of prey populations by juvenile roach can have a negative impact on their growth. Similarly, Welker et al. (1994) suggested that the growth and survival of gizzard shad larvae and the growth of bluegill larvae were affected by the availability of zooplankton, which may become limiting when densities of larvae are high. Conversely, assuming resources are not limiting, high densities may allow fish to increase their foraging rates, as shoals increase the probability of patchily distributed food being located (Pitcher 1986). At the individual level, the impacts of intraspecific interactions are governed by a range of complex and interacting factors, including social status and hunger (Pitcher 1986; Gotceitas and Godin 1991; Krause 1994). A thorough review of intraspecific food competition in fishes is provided by Ward et al. (2006).

Predation risk, both inter- and intraspecific, can also influence the foraging behaviour of larval and juvenile fishes. For example, food consumption by juvenile Eurasian perch is often reduced in the presence of piscivorous northern pike, and there may be an increase in the relative importance of small prey (Mikheev et al. 2006; Wanzenböck et al. 2006). In addition, juvenile perch may increase their use of complex refuge habitats in the presence of piscivores, which invariably reduces their foraging efficiency (Diehl and Eklöv 1995). Comparable results have been obtained for a number of other fish species (see Pitcher 1986; Winfield 1986; Metcalfe et al. 1987; Gotceitas and Godin 1991; Fuiman and Magurran 1994; Bean and Winfield 1995). Although such behaviours may allow fishes to forage even in risky situations, there are inherent implications for the energy intake, nutritional condition, growth and survival of larval and juvenile fishes, with individuals exposed to severe or continuous predation pressure likely to ingest fewer and smaller prey than those able to forage in relative safety (Fuiman and Magurran 1994; Diehl and Eklöv 1995).

Predator–prey relationships

It has long been recognised that planktivorous fishes are a major factor influencing the species and size composition of zooplankton communities (e.g. Hrbáček et al. 1961; Brooks and Dodson 1965; Cushing 1983; Lazzaro 1987). Phenomena frequently attributed to heavy fish predation include suppressed zooplankton biomass, small individual size of plankters, and reduced representation of vulnerable (typically larger) species. Large-bodied zooplankters are more efficient grazers of phytoplankton than are smaller-bodied species and, in the absence of severe predation, dominate the zooplankton (Brooks and Dodson 1965). Zooplanktivorous fishes may, therefore, cause changes in ecosystem functioning by increasing predation on zooplankton, thereby reducing grazing of phytoplankton and causing a reduction in water clarity. This has also been observed for larval and juvenile fishes (Cushing 1983; Bollens 1988; Mills and Forney 1988; Munk and Nielsen 1994; Kurmayer and Wanzenböck 1996; Mehner and Winfield 1997; Mehner and Thiel 1999). Similarly, selection of large zooplankton species and individuals can cause shifts in the species composition of zooplankton communities, as well as reductions in the mean size of individuals of large species and in the assemblage as a whole (Cryer et al. 1986; Romare et al. 1999). In particular, larval and juvenile fishes have the potential to suppress populations of large zooplankton species (Cushing 1983; Bollens 1988; Munk and Nielsen 1994; Mehner and Thiel 1999). Cryer et al. (1986), for example, observed that in summers when 0+ roach were abundant, zooplankton was sparse and dominated by copepods and rotifers, with cladocerans present in only low densities. Moreover, a number of studies (e.g. Cryer et al. 1986; Dettmers and Wahl 1999; Cowan et al. 2000) have suggested a causal link between fish recruitment and zooplankton dynamics, with abundant 0+ fishes sometimes showing poor growth as a result of depression of their prey populations.

Conversely, piscivory can influence the species and size composition of larval and juvenile fish communities. Brabrand (2001), for example, noted that piscivorous 0+ Eurasian perch significantly affected the size distributions of 0+ roach cohorts, because smaller individuals were predated more frequently than larger conspecifics. Furthermore, Beeck et al. (2002) speculated that piscivory by large 0+ Eurasian perch may have contributed to the complete disappearance of smaller individuals and 0+ common bream from a eutrophic lake by the end of the summer. Similarly, Bromley et al. (1997) stated that 0+ gadids can be subjected to substantial levels of piscivory (including cannibalism), potentially of sufficient magnitude to affect recruitment, and Buckel et al. (1999) suggested that predation by 0+ bluefish could have a substantial impact on the recruitment of striped bass. In addition, consumption of zooplanktivorous fishes by piscivorous larvae and juveniles may reduce predation pressure on zooplankton, potentially leading to increased grazing pressure on phytoplankton. Indeed, piscivory by larval and juvenile fishes may prevent the increases in 0+ fish abundance that often occur following depletion of adult fish stocks, which is important to ensure the long-term success of biomanipulation experiments (Beeck et al. 2002).

Apart from the direct affects of fish predation on zooplankton demography, indirect impacts can also occur through shifts in life history (e.g. changes in birth rates, fecundity, size and age at maturity, a switch from parthenogenetic to sexual reproduction, diapause), morphology (e.g. cyclomorphosis) or behaviour (e.g. diel vertical and horizontal migration) (Hanazato et al. 2001; Lass and Spaak 2003), all of which may have implications for ecosystem functioning. In addition, such phenomena may mask the impacts of fish predation. Gliwicz (2001), for example, found that the species-specific density levels of particular zooplankton did not depend upon reproduction rate, since neither increased birth rates nor reproductive effort coincided with an increase in population density; a clear indication that larger numbers of prey were being consumed by fishes at the time of increased reproduction. Furthermore, fish may select ephippial or egg-bearing zooplankters, due to their increased visibility compared with non-ephippial and non-egg-bearing individuals, and thereby influence zooplankton recruitment potential. For example, Gliwicz et al. (2004) revealed strong selection by European smelt for later instars and females of Daphnia spp. and Bosmina sp. with greater clutches, and similar behaviour has been reported in burbot (Ghan and Sprules 1993).

Management implications

For fish populations to persist, habitat must be available in sufficient quality and quantity for the range of activities undertaken during all periods of development. Implicit with this statement is that fish must be able to move between areas of habitat according to ontogenetic or temporal requirements. When this is not the case a bottleneck can occur, and fish population/community size and structure may be restricted. As such, knowledge of the ecology of larval and juvenile fishes has important implications for fisheries management, especially considering their vulnerability to predation, competition, disease and environmental perturbations.

Adult fish stocks are, directly or indirectly, limited by the quality and quantity of habitat and food available to larval and juvenile fishes. Floodplain waterbodies can enhance the recruitment of riverine fish populations through the provision of spawning and nursery habitats, refuge from floods, and higher water temperatures and, importantly, availability of planktonic food resources compared with main river channels. In addition, zooplankton populations often persist longer in floodplain waterbodies than in main river channels, and may be important sources of drifting material, which can colonise main river channels (Bass et al. 1997). Thus, connection of relict channels and man-made waterbodies to rivers has the potential to augment riverine fish populations and communities where natural river-floodplain connectivity has been lost (Grift et al. 2003; Nunn et al. 2007e). Similarly, managed realignment has the potential to enhance fish recruitment in the marine environment by increasing the availability of intertidal habitats, such as mudflats and salt marshes (Dixon et al. 1998; Colclough et al. 2005). It is important, however, that a high diversity of habitats is available to cater for the requirements of all life stages of a range of fish species. Habitats that enhance the diversity, size ranges and abundance of prey should ensure that sufficient food resources are available to larval and juvenile fishes.

It is critical that habitats are complex in structure, to enable segregation of species and minimise the potential for competition, predation and the spread of parasites and diseases. This is particularly important when implementing rehabilitation schemes in engineered ecosystems, where habitats have been modified and invariably simplified. Rehabilitation measures should be designed not just to provide refuge and nursery habitats for fishes, but also to enhance their food base. In general, floodplain waterbodies should be shallow, have complex shorelines with a high variability of flow velocities, and be located close to known spawning areas of key fish species. In addition, such waterbodies should ideally support submerged and emergent macrophytes, and a range in the degree of connectivity with main river channels is desirable (Grift et al. 2003; Bolland et al. 2012). Similarly, managed realignment sites should be designed to encourage the development of mudflats and salt marshes. Indeed, such areas are dynamic and naturally accrete and become colonised by vegetation (French 2006; Garbutt et al. 2006; Mazik et al. 2007). Accretion may increase the amount of habitat available to some fish species (e.g. flatfishes) and also facilitate the development of creeks, which provide areas of deeper water for large-bodied fishes, as well as habitat for small fishes at low water. Establishment of salt marsh further improves habitat complexity, which should increase the importance of managed realignment sites to fishes (Dixon et al. 1998; Colclough et al. 2005). Such habitats frequently support substantial zooplankton populations, thus improving the chances of survival and recruitment of fishes into the adult populations.

The foraging ecology of larval and juvenile fishes also has ramifications for husbandry protocols and stocking densities in aquaculture facilities. There is effectively a trade-off between stocking density and fish growth and survival. Whereas stocking at high densities may produce large numbers of fish, stocking at lower densities may produce larger fish. Thus, it is critical that the correct densities and species of fish are used in aquaculture facilities, and that adequate food is available: if certain fish species affect the growth or condition of other species, they could reduce survival in facilities and once released into the wild, thus reducing stocking success. Priming aquaculture facilities to provide fish larvae with abundant prey, and timing stock-out to coincide with peak zooplankton densities, should maximise survival and growth rates. Rearing fishes in single-species enclosures will eliminate the possibility for interspecific competition during the larval and juvenile periods, although there is still the potential for intraspecific competition to occur. If it is not possible to rear fishes in single-species enclosures, selecting fishes of contrasting behaviour (e.g. benthic and open-water feeders) should minimise any interspecific interactions.

Conclusions and recommendations

Zooplankton serve as a vital food resource for the larvae of most fish species. Even fish species that are piscivorous, herbivorous or detritivorous as juveniles or adults may initially be zooplanktivorous. Notable exceptions include fishes inhabiting the upper, fast-flowing reaches of rivers, where plankton is usually rare. Feeding diversity generally increases during the juvenile period, when there is often an increase in the importance of species-specific dietary traits. There are a large number of species, however, for which there is no information on their foraging ecology during the larval and juvenile periods. Indeed, primary literature was found for only 34 (35%) of the 96 freshwater fish species and 28 (8%) of the 369 marine fish species recorded in the UK, and information is even more sparse for other regions (Froese and Pauly 2011). For example, there appears to be no information for 93% of the 579 freshwater fish species recorded in Europe (Kottelat and Freyhof 2007), and there is also a paucity of knowledge on the foraging ecology of deep-water marine fishes. It is, therefore, recommended that further research is conducted in an attempt to address some of the gaps in the knowledge of the foraging ecology of larval and juvenile fishes.

A range of factors influence the foraging ecology of larval and juvenile fishes, including the ontogeny, morphology, behaviour, availability of potential food items, local habitat complexity and hunger of the fish (Fig. 6). Ontogeny influences the ability of fish to detect, capture, ingest and digest prey. For example, many young larvae have poor vision, which limits their prey-detection capacity. In addition, young larvae are poor swimmers and have small gape size, which limits their prey-capture efficiency and the size of prey that can be ingested. Furthermore, the young larvae of many fish species have poorly developed alimentary tracts and, thus, limited digestive capacity. Once fins are fully developed, associated improvements in swimming performance allow a broadening and diversification of diet spectra, effectively increasing their efficiency and capacity to capture and ingest prey. Similarly, improved vision increases the search volumes of 0+ fishes, effectively increasing prey availability, and development of the alimentary tract and the production of potent digestive enzymes increases digestive capacity. Differences in feeding behaviour (e.g. diurnal vs. nocturnal, filter vs. particulate feeding, vision vs. olfaction) between species, individuals or during ontogeny can also be important.

Fig. 6
figure 6

The complexity of factors (simplified) influencing the prey selection and diet composition of larval and juvenile fishes

Full size image

Prey size, morphology, behaviour and abundance can all influence the prey selection and diet composition of larval and juvenile fishes (Fig. 6). Optimal foraging theory predicts that predators should select prey that maximise the energetic gains available in relation to the energetic costs of capturing, ingesting and digesting the prey. Prey selection is thus determined by the relative profitabilities of particular types and sizes of prey, with maximum profit conferred by prey types and sizes that provide the maximum energy gain per unit handling time. The energetic content of a given prey type increases with prey size. However, certain prey may be consistently selected over others, irrespective of size, suggesting that taxa- as well as size-specific characteristics are important in the prey selection process. For example, cladocerans are common prey of many 0+ freshwater fishes because they are easier to capture than copepods, even though the latter are often numerically abundant and of greater calorific value. Similarly, the larvae of many marine fish species select calanoid over cyclopoid copepods, probably as calanoids swim more consistently and predictably than cyclopoids. It is recommended that further studies and experiments are conducted to establish the relative importance of taxa- and size-specific characteristics in prey selection by larval and juvenile fishes at various stages of development. Such research should consider potential selection for specific components of prey populations, such as egg-bearing or ephippial females, as well as piscivory (including cannibalism). This level of analysis could reveal areas of inter- or intraspecific dietary overlap/competition or, alternatively, resource partitioning. It may also be possible to obtain field evidence of switching, and possibly species-specific threshold densities when fishes change their feeding behaviour.

Temporal and spatial variations in prey availability can have important implications for the prey selection, diet composition, growth, survival, condition and, ultimately, recruitment success of fishes (Fig. 6). Food availability may vary for a number of reasons, including diel, seasonal, annual, microhabitat, mesohabitat, macrohabitat and regional differences in prey abundance. It is recommended that research is conducted to investigate ontogenetic and seasonal changes in diel feeding behaviour. In addition, further research into interannual variations in diet composition and prey availability may allow a causal link between larval and juvenile fish growth and food resources, and possibly threshold densities of key prey taxa, to be identified. Although it is recognised that habitats differ in the composition, abundance and size ranges of potential prey, there is little evidence for an influence of habitat on the species- and development-specific condition of larval and juvenile fishes. Studies of the relative contribution of particular habitats to the diets (e.g. using stable isotope analysis, in combination with traditional diet analyses and estimates of prey availability; Fry 2002) and nutritional condition (e.g. using RNA-DNA ratio analysis; Caldarone et al. 2006) of specific developmental steps of particular fish species may allow key habitats in terms of both food resources and nutritional condition, respectively, to be identified. Such knowledge has fundamental and applied implications and, indeed, is essential if the habitat requirements of all life stages and species of fish are to be met by rehabilitation schemes.

Inter- and intraspecific interactions can also have an important influence on the prey selection and diet composition of larval and juvenile fishes, either through dietary overlap/competition, resource partitioning or predation risk (Fig. 6). It is recommended that experiments are conducted using a range of fish densities and different combinations and ratios of species. This could demonstrate the influence of interspecific interactions on the diet composition and prey selection of different fish species, and identify the densities at which interspecific, and intraspecific, interactions manifest in differences in growth, condition and survival rates. Various combinations of feeding guilds could be studied, for example surface- and benthic-feeding species, to identify the species most likely to compete for resources. Conducting experiments at various densities may also permit species-specific optimal stocking densities to be identified, potentially allowing an increase in production in aquaculture facilities. The effects of density and interspecific interactions, as well as food quantity and quality, on the nutritional condition of fishes could be assessed using RNA-DNA ratio analysis.

Larval and juvenile fishes have the potential to influence the density, species and size composition of prey communities, either directly or indirectly. It is recommended that further studies are initiated to investigate the predator–prey relationships of larval and juvenile fishes, and the cascading impacts on primary production and ecosystem functioning. These could include studies of zooplankton demography to assess the impacts of fish predation or presence on the fecundity, birth and death rates, size and age at maturity, reproductive strategy, morphology and behaviour of key zooplankton groups, as well as piscivory (including cannibalism). The outcomes of such studies will further our understanding of predator–prey relationships and may be of use in the context of biomanipulation and habitat rehabilitation.