to total calorie intake was uncertain. While phytoplankton and protozoans may be consumed in substantial numbers, their volume (and calorie) contributions may in most cases be relatively low, contributing only a small amount to energy intake by larvae (de Figueiredo et al. 2007). A recent, comprehensive analysis of feeding by fish larvae in the Baltic Sea found that protists (ciliates) were a dominant prey for first‐feeding larvae of several species (Zingel et al. 2019). In some cases, larger postlarval stages may continue to include protists in their diets. For example, Chicharo et al. (2012) conducted gut and nutritional analyses on larvae >20 mm in length of Sardina pilchardus, Engraulis encrasicolus and Atherina presbyter in the Ria Formosa Lagoon (Portugal), reporting diverse diets of mostly zooplankton, but noting that each species included phytoplankton in its diet.
The diet reported for the estuary‐dependent, preflexion larvae of the clupeid Brevoortia patronus from the Gulf of Mexico was a mixture of zooplankton and phytoplankton, but with growth the diets of preflexion larvae shifted to feeding exclusively on zooplankton (Govoni et al. 1983, Stoecker and Govoni 1984). However, in the same region, estuary‐dependent larvae of the sciaenids Leiostomus xanthurus and Micropogonias undulatus had fed almost exclusively on zooplankton (Govoni et al. 1983). In another example, Ochoa‐Munoz et al. (2013) reported that feeding by preflexion larvae of a clinid fish Myxodes viridis in a Chilean estuary was predominantly on small phytoplankters (54% of prey in guts); however, by volume of prey the dominant food of these larvae was zooplankton (94%), especially the various developmental stages of copepods. The omnivorous diet of M. viridis had largely shifted to carnivory in flexion‐stage larvae and was completely carnivorous in postflexion larvae, a transition that may be common during ontogeny of many estuarine fishes.
For some fishes, the availability of particular prey to larvae is hypothesised to determine survival potential, e.g. the moronid Morone saxatilis and the lateolabracid Lateolabrax japonicus in which early‐life feeding success depends on adequate concentrations of the copepods Eurytemora carrolleeae (=affinis) and Sinocalanus sinensis, respectively, in association with favoured hydrographic conditions that aggregate and retain larvae and their prey (North & Houde 2003, Islam & Tanaka 2005, Islam et al. 2006, Shoji & Tanaka 2008, Shideler & Houde 2014). An evaluation of bloom dynamics of two key prey species on feeding by larvae of M. saxatilis and Morone americana in the Hudson River demonstrated how locations and bloom dynamics of the strongly selected copepod Eurytemora affinis and abundant cladoceran Bosmina sp. affected feeding variability (Limburg et al. 1997). However, in that case, the hypothesis that zooplankton blooms would lead to better and more successful production and survival of larvae was only partially supported.
In the estuarine portion of the St Lawrence River (Canada), larvae of the osmerid Osmerus mordax encountered better feeding conditions on the copepod Eurytemora affinis and the cladoceran Bosmina longirostris in the Estuarine Turbidity Maximum (ETM) region than down‐estuary of it (Dodson et al. 1989, Sirois & Dodson 2000). Similar results were reported for Morone saxatilis and Morone americana larvae feeding on E. affinis and B. longirostris in the Chesapeake Bay and the St Lawrence system in which the high turbidity and frontal structure of the ETM elevated feeding levels on these prey (North & Houde 2003, Vanalderweireldt et al. 2019a). Aggregation and retention mechanisms in estuaries not only operate on fish larvae but also act on their zooplankton prey. For example, estuarine salt fronts, ETMs and other features may concentrate zooplankton that is favoured prey of fish larvae (Dauvin & Dodson 1990, Sirois & Dodson 2000, North & Houde 2003, Islam & Tanaka 2005, Shoji et al. 2005b, Martino & Houde 2010, Suzuki et al. 2014), thus improving feeding opportunities for fish larvae.
The highly turbid conditions in many estuaries were once thought to inhibit feeding by larval fishes. However, such conditions, as seen in ETM regions, may enhance feeding by larvae. For example, Chesney (2008) found that moderate‐to‐high turbidity enhances feeding ability of the larval engraulid Anchoa mitchilli. This result contrasts with observations on feeding by juvenile and adult fishes, which is generally inhibited at high levels of turbidity. Chesney's (2008) result agrees with other observations and experimental evidence on feeding by estuarine fish larvae (Boehlert & Morgan 1985, Chesney 1989, Hasenbein et al. 2016).
Larvae of some fishes are feeding specialists and may select particular prey types, e.g. appendicularians Oikopleura spp. as reported for estuary‐dependent, pleuronectiform fishes (White 1968, Last 1978a, Llopiz 2013). In the cases of the larval moronid Morone saxatilis and its congener M. americana, the cladoceran Bosmina sp. is an important alternative prey to the copepod Eurytemora affinis (=carrolleeae). The cladoceran can become the dominant prey in some years and under some environmental conditions in North American estuarine tributaries (Limburg et al. 1997, Campfield & Houde 2011, Vanalderweireldt et al. 2019a).
Sufficient availability of suitable planktonic prey is the foundation of Hjort's critical period (Hjort 1914) and related hypotheses. In the sea, concentrations of plankton organisms eaten by fish larvae span at least five orders of magnitude (from <103 to >107 per m3) and concentrations generally are on the higher side in productive estuaries (Houde 2016). Laboratory research in the late twentieth century often concluded that mean concentrations of favoured and suitable prey in the ocean (for example, copepods), which often range from 1 to 100 L−1, were below levels believed at the time to be capable of supporting larval feeding and growth (May 1974, Hunter 1981). In estuaries, prey levels often exceed 100 L−1 and feeding conditions may be better, on average, than in coastal seas and oceans. For example, the mean concentration of zooplankton prey of suitable size to support feeding by fish larvae was 270 L−1 in Biscayne Bay, Florida (Houde & Alpern‐Lovdal 1984). While average prey levels for larvae may be higher in the estuary than in the sea, extreme fluctuations in prey concentrations in response to environmental variability are more likely in estuaries, a possible reason that some estuary‐associated fishes use the sea for spawning and larval production. That said, larval fishes of marine and estuarine origin are capable of varied diet in estuaries, and this flexibility may be key to survival in estuarine nurseries (Strydom et al. 2014b).
Thresholds of prey concentration required for successful feeding by early‐stage fish larvae have been examined for many fish species. Suitable prey at concentrations ≥100 L−1 frequently are reported as required to support growth and survival (Houde 2016). In some fishes, including Baltic Sea species such as the clupeids Sprattus sprattus and Clupea harengus, the required prey concentrations to support larval growth and survival have been analysed and modelled (Peck et al. 2013 and references therein). Larvae of Baltic C. harengus could feed successfully in shallow estuary environments when concentrations of copepod nauplii were only 7.5 L−1, although first‐feeding was delayed up to four days compared to larvae provided prey at 120 L−1 (Kiørboe et al. 1985). Arula et al. (2015) analysed feeding conditions in Baltic C. harengus and found that they were substantially poorer during the first 10 days after larvae initiated feeding relative to conditions encountered by older larvae.
In the clupeid S. sprattus, nutritional condition and survival of larvae were correlated with the abundance of copepod prey (copepodid stages) in the Baltic Bornholm Basin (Voss et al. 2006) where growth rates were low when prey concentrations were <5 L−1 but were near maximum values at 20 L−1 (Hinrichsen et al. 2010). Physiological modelling of first feeding S. sprattus larvae suggested that prey concentrations of 30–40 copepod nauplii L−1 may be required for survival, but these threshold values were strongly dependent upon temperature (Peck & Daewel 2007). In the North American engraulid Anchoa mitchilli, substantial survival of larvae of this estuarine fish was reported when copepod nauplii concentrations exceeded 25 L−1 (Houde 1978). Substantial survival of the moronid Morone saxatilis and the alosine Alosa sapidissima larvae in Chesapeake Bay (USA) may occur if prey levels equal or exceed 50 L−1 (Houde et al. 1996, Leach & Houde 1999).
Prey patchiness at various spatial scales, or processes that elevate encounter rates between larval fishes and prey (e.g. microturbulence), has been