12 Feeding Ecology ofPiscivorous Fishes

FRANCIS JUANES, JEFFREY A . BUCKEL ANDFREDERICK S . SCHARF

12 .1 INTRODUCTIONFish exhibit tremendous diversity in feedinghabits and the morphologies associated withfeeding. A recent book (Gerking 1994) and variousother overviews of fish feeding exist (Wootton1990; Hobson 1991 ; Hart 1993) . However, most ofthese tend to be general reviews of theory or focuson smaller non-piscivorous fishes . Our intent hereis to review the feeding ecology of piscivorous fish,a subject which to our knowledge has never previ-ously been reviewed . We focus on teleost fish, andon species and sizes that consume juvenile andadult prey and do not consider those that feed pri-marily on larvae or fish eggs .

12.1.1 Whypiscivorous fish?Piscivorous fish are broadly distributed phyloge-netically and geographically, occur in most habi-tats and generally occupy the top of most aquatictrophic webs. Piscivorous fish also generallyachieve the largest body size within fish com-munities, are represented by some of the largestspecies (many elasmobranchs, tunas, billfishes)and have potentially large impacts on their com-munities through predation . Finally, many pisci-vorous fishes, because of their ubiquity and largebody size are among the most valuable harvestablespecies in many of the world's fisheries . In somecases, they compete with humans for commer-cially important resource species (V . Christensen1996; Buckel et al . 1999a) . Piscivorous fish as a

group, however, are difficult to categorize anddescribe, which perhaps explains why no reviewsof their ecology exist .

12 .2 ADAPTATIONS FORPISCIVORY

12.2.1 What is a piscivore?We define piscivorous fish as carnivorous fish thatconsume primarily fish prey. Most fish species areopportunistic and flexible in their feeding habits(Dill 1983) and no species consumes only fish prey;however many do ingest fish as the main preyitem. Fish that eat other fish are second in propor-tion to those feeding on benthic invertebratesand are present in a variety of freshwater, estuarineand marine systems . They are equally common intropical and temperate ecosystems . Keast (1985)examined the piscivore feeding guild of small lakesand streams and categorized piscivorous fish intoprimary and secondary piscivores . Primary or'specialized' piscivores shift to piscivory withinthe first few months of life, whereas secondary pis-civores only become fish-eaters later in life . Keastfurther suggests that secondary piscivores switchto fish as a way to maintain energetic efficiencyas they grow. This can only be achieved by eatingprogressively larger prey and, at a certain point,fish are the only prey available . Furthermore, sec-ondary piscivores are not structurally adapted forpiscivory other than acquiring a large mouth asthey age .

268

Chapter 12

12.2.2 What are the adaptationsfor piscivory?

Feeding behaviours andmorphological adaptations

When observed in detail it becomes clear that pis-civore feeding behaviour is complex and flexible indealing with different prey types . Behaviours cangenerally be grouped into the following categories,which are also associated with particular morpho-logical adaptations .

Luring Luring is a sit-and-wait behaviour whereprey are attracted by a 'lure', which consists of astalk topped by a device that resembles a source offood; the lure is a modification of the first dorsalspine (Gerking 1994) . Luring is typical of the an-glerfish (order Lophiiformes) . In addition, amongthe frogfish various morphological features allowthe predator to blend into the rocky environmentsin which they live as a form of camouflage. Angler-fish have a considerable gape and ingest prey sizesthat are large compared to other piscivores .

Stalking Stalking is the unobtrusive pursuit ofprey before the attack occurs . This strategy is com-mon to trumpetfish (Aulostomidae), longnose gar(Lepisosteidae) and needlefish (Belonidae) . Thesespecies have similar morphologies, long slenderbodies with a long snout and sharp teeth .

Chasing Large piscivores are able to chase and'swim-down' prey. This strategy is best representedby the billfish (Xiphiidae and Istiophoridae) andsome tuna (Scombridae) andyellowtail orpompano(Carangidae) . These fish must be able to attainhigh cruising and accelerating speeds . Their bodiescan be described as thunniform or carangiform,where thrust is maximized by a lunate tail with ahigh aspect ratio, a narrow caudal peduncle thatminimizes sideways thrust and a large anteriorbody depth that minimizes recoil of the head end .These same features, along with a relatively rigidand streamlined body, also minimize drag (seeBrix, Chapter4, this volume . Chasers canbe eitherJungers, such as the pike (Esox spp . ), where attacksare startedat close range from an S-shaped position,

strikes occur at high speed and missed prey arerarelypursued, or they can be pursuers such as trout(Salmo trutta ), where attacks start from a C-shapedposition, strikes occur at low speeds from a shortdistanceand missed prey are chased .

Ambush This strategy is used by species thatattack from seclusion, although an element ofchasing may also occur. Examples are morays(Muraenidae), pike (Esocidae) and summer floun-der (Pleuronectidae) . Few morphological similar-ities exist in species that use this behaviour otherthan the ability to use camouflage and high-speedattacks. However, distinct patterns of kinematic,pressure, electromyographic and behavioural pro-files of prey capture are exhibited by ambushersandpursuers .

Other A variety of other rarer feeding habits eachwith their own specialized morphologies, particu-larly in the dentition, include forms of parasitismsuch as blood-sucking in lampreys (Petromyzonti-dae) and catfishes (Trichomycteridae and Cetopsi-dae), scale-eating in many cichlids and characoids,and fin and eye-biting (Gerking 1994 .

A novel feeding behaviour common to plankti-vores has recently been observed for a few speciesof piscivores (Sazima 1998) . Ram suspension (orfilter) feeding is defined as swimming through thewater with the mouth wide open and operclesflared as a way of filtering small prey items out ofthe water column without directing attacks to-wards individual prey. This may be a mechanismused by piscivores feeding on relatively small preyin high concentrations (B . Hanrahan and F. Juanes,personal observation( .

Most piscivores ingest their prey whole . A fewspecies are able to tear off and ingest pieces andgood examples are piranhas (Serrasalmus spp .)and African tiger fish (Hydrocyon vittatus) . Themost detailed analysis of partial prey eating hasbeen performed for bluefish (Pomatomussaltatrix) (Juanes and Conover 1994a ; Scharf et al .1997). Juvenile bluefish switch from taking preywhole to partial prey consumption when the preyto predator size ratio is about 0 .35 independent of

prey type (Scharf et al . 1997( . This ability allowsbluefish to attack much larger prey sizes than canpredators of similar size and is likely to be a func-tion of specialized musculature and dentition .Bluefish are also unique in that they ingest preytail-first, whereas most other piscivores that havebeen examined eat prey head-fast . Tail-first inges-tion may be a result of pursuing prey rather thanambushing it, or a way to reduce prey mobilitythereby increasing prey vulnerability, as has beenobserved in piranhas .

Schoolingin predators

Schooling by prey is generally thought of as anadaptation for evading, confusing and reducing theefficiency of predators (Pitcher and Parrish 1993 .Schooling also has hydrodynamic and foragingfunctions . Little empirical work has been done onwhether schooling by predators enhances preda-tion efficiency in piscivores . Field studies haveshown that, in general, group attacks tend to resultin higher capture success rates (Pitcher and Parrish1993). Eklov (1992) has shown that the foragingefficiency, measured as growth rate, varies be-tween a group-foraging, actively searching pisci-vore (Eurasian perch, Perca fluviatilis) and asolitary-foraging, stalking piscivore (pike, Esoxlucius) . Under similar conditions, grouped perchgrew more than single perch or grouped pike . Incontrast, solitary pike grew better than solitaryperch or grouped pike. Schooling in piscivores mayhave foraging costs and benefits. Prey encounterrates may be maximized but the probability of los-ing a prey item to kleptoparasitism is also greaterwhen predators attack in groups (Juanes andConover 1994a) . There has been anecdotal infor-mation collected that suggests that some piscivo-rous fish hunt cooperatively (Pitcher and Parrish1993), although potential mechanisms have notbeen quantified under controlled conditions .

Active piscivorous pelagic fishes are character-ized by a range of length to maximum depth ratios(the fineness ratio), which ranges from 4 .0 to 6 .5and which maximizes feeding efficiency and mini-mizes drag (Blake 1983) . Bluefish, a primary pisci-vore that switches to fish prey at about 40 mm total

FeedingEcologyofPiscivorous Fishes 269

length, has a fineness ratio between 3 .5 and 5 .0(Juanes et al . 1994) . Interestingly, offshore inverte-brate-feeding juveniles (<40 mm) are already mor-phologically specialized for piscivory, suggestinga trade-off between feeding efficiency and futurediet . A result of this trade-off may be to acceleratethe onset of piscivory.

Life history

Onset of piscivory Most piscivorous fish under-go ontogenetic shifts in diet (Werner and Gilliam1984; Keast 1985 ; Winemiller 1989) . These shiftsgenerally progress from consumption of zooplank-ton to consumption of benthic macrofauna or preyfish, with a concomitant increase in mean preysize as predators grow. There is much variation inthe timing of the shift to piscivory among primaryand secondary piscivores (Mittelbach and Persson1998; Mittelbach, Chapter 11, this volume) . Afew species of scombrids apparently forgo thezooplanktivorous stage and start eating fishes atfirst feeding, whereas others shift early in thelarval period (Tanaka et al . 1996). The shift topiscivory invariably results in an increase in pre-dator growth rate (Buijse and Houthuijzen 1992 ;Juanes and Conover 1994b, Olson 1996). Amongthe scombrids studiedby Tanaka et al . (1996, therewas a direct correlation between the age at theonset of piscivory and early growth, with thosespecies shifting to piscivory at first feeding capableof reaching 100 mm during the first month of life .

Within fish cohorts, the largest individuals areoften able to switch to piscivory while the smallestare delayed and experience reduced growth. Thiseffect leads to the often-observed bimodality insize distribution of juvenile piscivores (Adams andDeAngelis 1987 ; Frankiewicz et al . 1996( . Becausesurvivorship is generally size-selective in fish(Sogard 1997), bimodality can result in increasedmortality of individuals in the smaller size mode(Olson 1996) .

What allows species to shift to piscivory? Thetimingof the onset of piscivory depends onpredatormorphology, predator and prey phenologies, preyabundance and abiotic factors . Clearly predators

270

Chapterl2

have to be larger than their prey in order to be ableto ingest them . Large body size and large mouthgapes are therefore generally considered an impor-tant constraint on prey use, particularly piscivory(Werner 1977; Wainwright and Richard 1995 ;Mittelbach and Persson 1998) . Biomechanical fea-tures such as jaw mechanics and tooth develop-ment have also been implicated (Jenkins et al. 1984 ;Wainwright and Richard 1995) . Among piscivo-rous scombrids, ontogenetic development of thedigestive system (Tanaka et al . 1996) and develop-ment of the visual system (Margulies 1997) havebeen correlated with the onset of piscivory . Wall-eye (Stizostedion vitreum) generally shift to pis-civorybetween 30 and 60 mm in length but are ableto ingest larvae when they are as small as 10 mm ;the earlier onset of piscivory may be related tothe abundance of prey in appropriate size ranges(Mathias and Li 1982) . However, because theswitch to piscivory is most often preceded by aninvertebrate-feeding stage, piscivores must ensurerapid growth during that phase so that when fishprey are available the feeding shift can occur . Ifgrowth is slowed during the invertebrate-feedingstage, the piscivore size advantage over its fish preyis reduced leading to delays in the shift to piscivory(Olson 1996) . Environmental factors such as lowtemperatures early in the growing season canalso reduce growth and delay piscivory (Buijse andHouthuijzen 1992 ; Olson 1996) .

Is acceleration of the onset ofpiscivory adaptive?Because of the dramatic increase in growth ratefollowing the shift to piscivory and because size-selective mortality is so prevalent among juvenilefishes (Sogard 1997), natural selection shouldfavour life-history strategies in piscivores thatminimize the length of the zooplanktivorousphase. One such strategy to attain an accelerationof the onset of piscivory would be to match thetiming and location of spawning with the avail-ability of fish prey. Freshwater primary piscivoreshave evolved the ability to spawn earlier thanother fishes, thereby attaining sufficient size toallow them to consume juvenile fish spawned laterin the same year. This pattern has been observedin largemouth bass (Micropterus salmoides), pike

and walleye (Forney 1971 ; Keast 1985). Amongmarine fish it is much more difficult to examinethe timing of predator and prey spawning . How-ever, bluefish appear to have evolved a life-historystrategy whereby they are spawned at approxi-mately the same temperature as their future preybut at a more southern latitude. This divergenceallows bluefish to attain the size advantage re-quired when they shift to piscivory after beingadvected to higher latitude estuaries where theirprey are abundant.

In a recent review of the freshwater fish litera-ture, Mittelbach and Persson (1998) showed thatthose species born larger and with larger mouthgapes become piscivorous at younger ages andsmaller sizes, although they did not find a relation-ship between spawning temperature, used as an in-dicator of timing, and size at the shift to piscivoryor size at age 1 . However, species that acceleratedthe onset of piscivory by shifting early were largerat age 1 and continued to be larger through laterages .

A potential cost of early onset of piscivory isreduced growth and survival if habitat condi-tions dictate prey abundance . Furthermore, be-cause most piscivores are strongly size-selective(Juanes 1994), if availability of appropriately sizedfish prey is delayed, growth may be reduced(Buckel et al . 1998) .

Resource polymorphisms

Adaptive trophic specialization can in some caseslead to the evolution of resource polymorphisms,where trophically and morphologically special-ized morphs can coexist and ultimately speciationcan occur. Among fishes various examples oftrophic polymorphisms include piscivorousmorphs. For example, in Arctic chary (Salvelinusalpinus), four trophic morphs are often recognized .The four morphs have differing life-history strate-gies, morphologies and behaviours that are gene-tically based (Skulason et al. 1993). Usinglife-history theory, Mangel (1996) has shown thatthe evolution of a large piscivorous morph in trout('ferox' trout) can occur if the growth rates ofthe different asymptotic morph sizes differ and if

there is size-dependent mortality . Other aspects ofalternative life-history evolution caused by com-petitive reproductive behaviour are reviewed byHutchings (Chapter 7, this volume) .

12 .3 COMPONENTS OFPREDATION

12.3.1 Search, encounter, pursuit,capture, handling

For a prey fish to be included in the diet of a pre-dator, it has to be located, pursued, captured,manipulated or handled, and finally digested(Mittelbach, Chapter 11, this volume) . Each ofthese steps must be performed successfully by thepredator in order for the predation process to becomplete. Therefore, prey have several opportun-ities to avoid being eaten during the course of theinteraction . The first of these is to not be detected .Predators have evolved various modes of searchingfor prey . At the same time, prey have evolved anassortment of behaviours to reduce their chancesof being detected by potential predators (Krauseet al., Chapter 13, this volume) . The probabilityof any one prey being encountered will thenbe determined by the morphological and behavi-oural characteristics of both predator andprey .

Most predators that consume motile prey, suchas fish, utilize one of two basic search/encountermodes: ambush or sit-and-wait tactics and cruis-ing tactics as described earlier. The types and sizesof prey eaten by a predator are often associatedwith the foraging tactics employed . For example,Green (1986) showed that invertebrate predatorsthat employed ambush tactics consumed largerprey relative to cruising predators. He furtherdemonstrated that prey encounters for ambushpredators were strongly dependent on prey swim-ming speeds, whereas variation in prey activityhad no discernible effect on encounter rates forcruising predators . Among fish, ambush predatorshave also been shown to take larger prey relative tomore actively searching predators . The inclusionof larger prey items in the diets of ambush preda-tors is thought to be a function of their dependence

Feeding Ecology of Piscivorous Fishes 271

on prey activity producing encounters and the pos-itive relationship that exists between body size,prey movement and detectability.

The estimation of encounter rates between fishpredators and prey remains an elusive problem infish ecology. Specifically, for piscivorous fishes,the mobility of potential prey means that behav-iours and movement patterns of both predator andprey will ultimately affect rates of encounter . Dueto the variable nature of the plethora of environ-mental parameters that can influence the distribu-tion of animals in space and time, encounter ratesin situ are virtually impossible to measure withany certainty. Laboratory tanks used for piscivoresand their prey are generally too small to yield anyrealistic estimates of encounter rate, as all prey areusually within the visual field of a given predator .Therefore, encounter probabilities are often esti-mated using mathematical models that incorpo-rate average swimming velocities of predatorand prey as well as a predator detection radius(Gerritsen and Strickler 1977( . Although theore-tical models can be quite elegant, they remain bur-dened with many assumptions that often cannotbe tested. Because no predation can occur withoutan encounter first, much future research will bedevoted to better understanding the factors thatcontribute to variable rates of encounter .

Once prey are encountered, predators must besufficiently capable of capturing prey for it to beeaten. Piscivorous fishes generally have lowercapture probabilities compared with planktivores .Usually about half of piscivore attacks on averageare successful compared with 70-80% for plank-tivorous fishes . Capture success in piscivores hasalso been shown to be directly related to relativebody sizes of prey and predator, and generally de-clines linearly as the prey size to predator size ratioincreases (Fig . 12 .1) . Similar size-based capturesuccess functions have been observed for severalpiscivore species representing different life stages(Miller et al. 1988; Juanes and Conover 1994a) .Differences in prey type have also been shown toinfluence predator capture success, as prey escapeproficiencies vary among species (Wahl and Stein1988; Scharf et al . 1998) . The strong dependence ofpiscivore capture success on prey size and type

272

N00E0

0

vi

Prey size/predator size

indicates that it plays an important role in deter-mining relative vulnerabilities to predation fordifferent prey.

Similar to capture success, handling time in pis-civorous fishes is affected by both prey size andtype. Because larger prey sizes often need to be ma-nipulated before swallowing, increasing prey sizetypically causes an exponential increase in han-dling time (Fig. 12 .1) . The rate of increase in han-dling time has been shown to be unique for specificprey types (Hoyle and Keast 1987; Scharf et al.1998). Handling time has historically been a criti-cal parameter to estimate, as it has frequently beenused to represent the primary energetic cost offeeding in theoretical foraging models attemptingto predict predator diet (Mittelbach, Chapter 11,this volume) . Although foraging models havedemonstrated some success in predicting diets ofplanktivorous fishes, past models for piscivoreshave often failed (Sih and Christensen 2001) . Morerecent models that incorporate differential captureprobabilities based on prey size and type haveproven more successful in predicting piscivorediets (Rice et al. 1993) . The combination of size-dependent encounter rates, capture probabilities,energy content and handling times can be used toconstruct profitability functions for specific preytypes. For piscivores, these functions are typicallydome-shaped curves that peak at intermediateratios of prey size to predator size (Fig. 12.1( .However, most often encounter rates cannot beestimated with confidence and curves are con-structed using only capture success and handlingtime functions . The specific relation, for each

Chapter 12

Fig . 12 .1 General empiricalrelationships between predationcomponents and relative preysize for piscivores .

prey type, between encounter probability andprey size will determine whether the profitabilitypeak is shifted toward smaller or larger relativeprey sizes. Accurate profitability functions thatincorporate a broad range of prey and predator sizescan be extremely valuable for predicting piscivorediets .

12.3.2 Functional andnumerical responses

Predatory response to variations in prey densitycan be grouped into two categories . The functionalresponse refers to the number of prey eaten perpredatorper unit time as a function of prey density.Changes in number of predators in response tovariations in prey density describes the numericalresponse .

Functional responses

There are several different forms of functional re-sponse. A subset of these have been classified asthe type I, II and III models (Holling 1965) . The typeI response occurs when the number of prey eatenper predator per unit time increases linearly withincreasing prey density (Fig. 12 .2a) . The proportionof prey density consumed stays constant and thisresponse is also called density-independent preda-tion (Fig. 12 .2d) . Most vertebrates have a type II orIII functional response (Hassell 1978) . The type IIresponse is represented by a negatively accelerat-ing curve (Fig. 12.2b) ; therefore, the proportion ofprey density consumed decreases with increasing

0wa6

E

T0.

Ez

(a)

(b)

Type I

Type II

E

sa0d

(d)

(e)

(f)

Type III

Number of prey present

Fig.12.2 /a-c) Shapes of hypothetical functionalresponses : (a) type[, (b) type IIand (c) type III . (d-f)Proportion of prey consumed or percentage mortalitycorresponding to the functional responses : (d) type I,(e) type II and (f) type III .

Feeding Ecology of Piscivorous Fishes

273

Type I

Type II

Type III

prey density and is referred to as negative density-dependent predation . Under this regime the risk ofbeing preyed upon is high at low prey densities butdecreases with increasing prey density )Fig. 12 .2e) .Positive density-dependent predation can resultwhen a predator feeds with a type III functional re-sponse. The shape of this function is sigmoidal(Fig . 12.2c) ; when the proportion of prey densityconsumed is plotted the slope is initially positive,meaning that the risk of being preyed upon is smallat low prey densities but increases up to a certainpoint as prey density increases (Fig . 12 .2f) . Preda-tion components such as attack rate and handlingtime (Section 12 .3.1) can be estimated from thesefunctions (Holling 1965; Hassell 1978) .

In general, piscivorous fishes have a type Hfunctional response. For example, type IIresponseswere found in Arctic chary preying on migratingsockeye salmon (Oncorhynchus nerka) smolts

(Ruggerone and Rogers 1983) and southern floun-der (Paralichthys lethostigma) preying on spot(Leiostomus xanthurus) (Wright et al . 1993) . Laketrout (Salvelinus nemaycush) were capable of highpredation rates at low prey fish densities, suggest-ing a type II response (Eby et al. 1995) . Petersen andDeAngelis (1992) found that type H and III modelsgave similar fits to field data of northern squawfish(Ptychocheilus oregonensis) feeding on juvenilesalmonids .

Although the type III functional response hasnot been observed with regularity, it has beenproposed as a potential mechanism in regulatingpopulation abundance (see Section 12 .6) . Severalfactors can lead to a type III functional response,such as predator learning, prey refuge and the pres -ence of alternative prey (Holling 1965) . The pres-ence and accessibility of a prey refuge has beenhypothesized to be a factor leading to positive den-sity-dependent mortality in some piscivore-preysystems (Bailey 1994 ; Hixon and Carr 1997). Alter-native prey can lead to a type III functional re-sponse through switching behaviour (Murdochand Oaten 1975) . Prey switching occurs when theprey type with the highest relative abundance isincluded disproportionately more in the predator'sdiet than would be expected from random feeding(see Section 12.4.1 for an example) .

Functional responses are critical parts of themultispecies virtual population analysis (MSVPA)method of modelling fish populations (Shepherdand Pope, Chapter 7, Volume 2) . MSVPA differsfrom single-species virtual population analysis(SSVPA) in that interspecific and intraspecific pre-dation are included to estimate levels of naturalmortality, which in SSVPA is assumed constant .Generally, MSVPA has assumed a type II fun-ctional response, which may be reasonable forpredators that do not depend on a small number ofprey species, as in temperate systems, but may notbe adequate for predators that do depend on fewprey species, as in boreal systems . Inclusion of atype III response can have important effects onforecasts based on MSVPA, including producingmore than one solution to MSVPA equations(Magnusson 1995 ; see also Shepherd and Pope,Chapter 7, Volume 2) .

274

Numerical responses

Predator numbers can respond rapidly to increas-ing prey density, as when predators aggregate onprey fish, or can respond slowly, such as increasedreproductive success (Peterman and Gatto 1978) .We know little about numerical responses in pis-civorous fish . This is surprising given its potentialimportance in regulating population levels ; thelack of data stems from the logistical difficultiesof measuring such a response . Whenever a predatorincreases in density in response to increasing preyfish density, the total predation rate response candiffer from the predator's functional response . Asigmoidal response (type III) may more typicallydescribe a piscivore's total response; the totalresponse is a function of both the functional andnumerical response (Bailey 1994( .

12 .4 PREY TYPE ANDSIZE SELECTIVITY

Selective predation can be defined as the consump-tion of prey in different proportions than thoseavailable in the predator's surrounding habitat(Chesson 1978; Mittelbach, Chapter 11, this vol-ume). This is in contrast to random feeding, wherepredators feed indiscriminately on prey in accord-ance with their relative availability . Predatorsthat feed in a random manner with respect to rela-tive prey abundance levels are usually referred toas opportunistic feeders because they adjust theirfeeding habits rapidly to match variation in thelocal prey field . Predators that consume a narrowrange of prey types or sizes regardless of changesin the prey field are thought of as specialists andrepresent the most extreme case of selectivepredation . Most predators fall somewhere along acontinuum between opportunists and specialists .

12.4.1 Observed patternsNumerous instances of prey type and size selec-tion by piscivorous fishes have been reported inboth freshwater and marine communities . Forexample, Wahl and Stein (1988) used laboratory

Chapter 12

experiments and large-scale field manipulationsto determine patterns of prey selection in severalfreshwater predators, including muskellunge(Esox masquinongy), tiger muskellunge, andnorthern pike . Results demonstrated that each ofthe predators consistently selected for shad prey asopposed to Centrarchidprey. The authors conclud-ed that behavioural and morphological differencesbetween prey types resulted in differential vulner-ability to predation and the observed patterns ofselection. Research on the feeding habits ofpiscivorous bluefish in marine waters off thenorthern Atlantic coast of the USA have also re-vealed evidence for selective predation on specificprey types. Buckel et al. (1999b) demonstratedthat, during early summer in shallow estuarinehabitats, juvenile bluefish consumed young-of-the-year striped bass (Morons saxatilis) in greaterproportions than a random sampling of the preyenvironment, and that this positive selection wasdirectly related to striped bass abundance . In otherwords, they showed prey switching behaviour asreferred to in Section 12 .3. During southerly au-tumn migrations in continental shelf waters, juve-nile bluefish have also shown the propensity forselecting specific prey types, particularly bayanchovy (Anchoa mitchilli) (Buckel et al . 1999c( .Selection for specific prey sizes maybe even moreprevalent than prey type selection among piscivo-rous fishes. A recent review of 32 laboratory andfield studies concluded that selection for smallprey from available size distributions was a com-mon phenomenon for both freshwater and marinepiscivores (Juanes 1994) . This study also revealedthat most piscivorous fishes consumed prey small-er than the optimal prey sizes predicted by energymaximization models, independent of bothpredator and prey type as well as predator size(Mittelbach, Chapter 11, this volume) .

12.4.2 Behaviouralmechanismsofselective feeding

Selective feeding can result from both morphologi-cal constraints of predators and behavioural inter-actions between piscivores and their prey . Predatorgape size ultimately limits the sizes of prey in-

gested and can also affect the types of prey con-sumed. However, recent studies indicate that be-havioural capabilities of both predator and preycan regulate the sizes and types of prey eaten beforemorphological constraints become important . In awell-designed series of laboratory experiments, B .Christensen 11996) evaluated the effects of prey an-tipredator behaviours on the sizes of roach (Rutilusrutilus) consumed by piscivorous Eurasian perch .Feeding experiments were conducted at two dif-ferent spatial scales, determined by tank volumes,to control the level of antipredator behaviour ex-pressed by the prey. In smaller arenas that limitedprey escape ability, the author found that perchwere able to consume large roach approaching gapelimitations . In larger arenas, where prey antipreda-tor behaviours were not suppressed, the maximumsize of roach consumed by perch was significantlysmaller. The author concluded that the sizes ofprey eaten were largely dependent upon relativepredator and prey mobility (Mittelbach, Chapter11, this volume) . Juanes (1994) hypothesized thatcommon patterns of selection for small-sized preyby piscivorous fishes were the result of size-dependent vulnerabilities of prey to predatorcapture, rather than predator preferences . Heproposed that many examples of selective feedingin fishes were actually passive selective processesrather than active predator choice, with attackrates being relatively equal among different preysizes but smaller prey being consumed more fre-quently due to ease of capture (see also Hart andHamrin 1990). Prey behaviour can also influencethe rate of attack among different prey typesavailable to a predator. For example, laboratory ex-periments demonstrated higher attack rates bygoby (Gobiusculus flavescens) predators on her-ring (Clupea harengus) larvae compared with cod(Gadus morhua) larvae, with attack proportionsbeing strongly related to differential levels of preyactivity between the species (Utne-Palm 2000) .Predictions of prey selection from most traditionalforaging models assume that optimal choices aremainly the result of behavioural decisions bypredators (Stephens and Krebs 1986) . For piscivo-rous fish predators, the behavioural and foragingcapabilities of both predator and prey can clearly

Feeding Ecology of Piscivorous Fishes 275

alter attack and capture rates of predators and con-tribute significantly to the observed patterns ofselective feeding.

12 .5 PREDATOR-SIZE ANDPREY-SIZE RELATIONSHIPS

Because they are the top predators in many aquaticsystems, knowledge of the sizes of prey included inthe diets of piscivorous fishes is essential in order toidentify their potential impact on prey survival andtheir role in structuring populations at lower troph-ic levels. This is particularly important for theecosystem approach to fisheries management,where knowledge of interactions is critical (Paulyand Christensen, Chapter 10, Volume 2) . Furtherdiscussion of how size-dependent processes struc-ture communities can be found in Persson, Chapter15, this volume. From a behavioural standpoint,relative body sizes of prey and predator can havesignificant effects on predator feeding success .Detection and capture of prey by piscivores areenhanced with increasing size for several reasons,including increased sustained and burst swimmingspeeds and better visual acuity (Webb 1976) . Theescapeproficiencyofpreyalsovarieswithontogenyas reaction distances increase and swimming abili-ties are improved with size (Brix, Chapter 4, thisvolume. Morphological characteristics that influ-ence piscivore-prey interactions, such as predatorgape size and robustness of prey morphologicaldefences (e .g . spines), also scale with ontogeny .Therefore, identifying patterns of prey size use bypiscivorous fishes can provide important clues asto the mechanisms that shape piscivore diets andthe effects of predation by piscivorous fishes oncommunity structure .

12.5.1 General patternsand hypotheses

For piscivorous fishes, the sizes of prey consumedgenerally increase with predator size . In addition,the range of prey sizes eaten typically increases inlarger predators, as maximum prey size often in-creases rapidly while minimum prey size may

276

Predator size

Fig .12 .3 Hypothetical predator size-prey sizerelationship for a piscivore, illustrating typicallyobserved changes in minimum and maximum preysizes consumed with increasing predator size .

change only slightly over a broad range of predatorsizes (Fig. 12.3). For example, Mittelbach andPersson (1998) demonstrated increasing mean andmaximum prey sizes as body size increased for 12species of freshwater piscivores . The authorsfurther noted that for many of the piscivores theyexamined, minimum prey size remained fairlyconstant with increasing predator body size .Scharf et al. (2000 revealed similar predatorsize-prey size patterns for a group of 18 marine pis-civores in continental shelf waters off the Atlanticcoast of the USA . Expanding prey-size ranges thatretain small prey in the diet of larger piscivoresmeans that prey sizes consumed by smaller preda-tors can be a subset of the prey sizes consumedby larger predators, which may result in a competi-tive disadvantage for smaller predators (Wilson1975 . However, when the range of prey sizes eatenis examined as a ratio of predator size rather thanon an absolute scale, ontogenetic changes are moresubtle. For example, Pearre (1986 studied a largegroup of predators that included 43 species oflarvae, juveniles and some small adult fishes inan effort to detect general trends in prey-size distri-

Chapter 12

butions eaten by fish predators of varying bodysizes. Based on ratio-scale analyses, he concludedthat the range of prey sizes eaten did not changesignificantly with increasing body size for mostpredators and that older larger predators shouldnot compete with smaller ones .

Although predator size-prey size relationshipsfor piscivorous fishes are often reported, potentialmechanisms that lead to the commonly observedpattern of increasing maximum prey sizes coupledwith stationary minimum prey sizes are generallylacking . Scharf et al. (2000) noted that the con-tinued inclusion of small prey in the diets of largerpredators contrasts with predictions of optimalforaging models for particulate feeders, which in-dicate that the largest prey available should be con-sumed preferentially to maximize net energeticreturn . The authors hypothesized that the combi-nation of high relative abundance and high captureprobability for small prey relative to large prey maylead to consistently high vulnerability to preda-tion for small prey fishes as already discussed .Because predator handling times also increaserapidly with prey size, they suggested that theretention of small prey in the diet may reflectprofitable foraging decisions by predators becausesearch, capture and handling costs remain low .

12.5.2 What determinesmaximum prey size?

Maximum prey sizes eaten by piscivores can belimited by several factors . Of foremost importanceare morphological limitations imposed by preda-tor gape size and throat width . Most piscivorousfishes swallow prey head first after manipulatingprey so that the largest body depth is positionedlaterally in the mouth (Reimchen 1991) . A strongrelationship between prey body depth and predatorgape width has been detected for several piscivo-rous fishes (e .g . Hambright et al . 1991) . Generalmouth shape and structure can also affect thetypes and sizes of prey that can be ingested . Forexample, Keast and Webb (1966) demonstratedstrong relationships between mouth morphologyand the feeding ecology of several freshwaterfishes . The effects of piscivore gape limitations on

maximum prey sizes eaten can also impact preypopulations. Persson et al. (1996) examined thepotential community-level effects of predatorgapelimitation in afield study of predation in Europeanlakes. Prey fishes in four lakes were exposed to twopredators with different gape sizes and their popu-lations monitored . Results indicated that prey ex-posed to the smaller-gaped predator reached a sizerefuge sooner, were more abundant and had slowergrowth rates due to intraspecific competition com-pared with prey exposed to the larger-gaped preda-tor. The effects of piscivore gape size were alsoobserved at lower trophic levels, indicating thepotential importance of predator gape sizes oncommunity dynamics (Persson, Chapter 15, thisvolume) .

Other factors may restrict maximum prey sizeseaten by fish predators before morphological limi-tations imposed by gape size become important .The behavioural tactics used by predators tosearch for, encounter and attack prey can affectprey sizes consumed . For example, Gaughan andPotter (1997) compared the diets and mouth sizesof five estuarine species of larval fish and con-cluded that mouth width had only a small influ-ence on prey sizes eaten and that disparate feedingpatterns among larvae were likely due to behavi-oural differences . Behavioural capabilities of bothpredator and prey can also contribute considerablyto the sizes of prey eaten. The ability of prey toevade predators is related to body size, with largerprey being more efficient at avoiding predatorstrikes. Therefore, maximum prey sizes eatenby piscivorous fishes can often be considerablysmaller than gape sizes (Juanes and Conover 1995 ;B. Christensen 1996) .

12 .6 POPULATIONREGULATION

Shepherd and Cushing (1990) stated that

there is some basis for the belief that fishpopulations are regulated in some way . Infact, the evidence is twofold: first, that theydo not explode when subjected only to low

Feeding Ecology of Piscivorous Fishes

(natural) levels of mortality : secondly, thatthey do not collapse at all quickly, whensubjected to high levels of mortality .

One of the more well accepted mechanismsput forward to explain population regulation infish is density-dependent mortality in larval andjuvenile stages through interspecific predationor cannibalism (see Section 12 .6 .2; Myers, Chapter6, this volume; Shepherd and Pope, Chapter 8,Volume 2) .

Density-dependent mortality can lead to themaintenance of population stability (Murdoch andOaten 1975 ; Hassell 1978) . Field studies provideevidence for density-dependent mortality inmarine fishes (Myers and Cadigan 1993 ; Forrester1995 ; Hixon and Carr 1997) . Several of these inves-tigations provide evidence that piscivorous fishesare the dominant cause of density-dependent mor-tality. However, identifying the mechanism thatgenerates density-dependent predation mortalityhas been elusive (Bailey 1994). Potential mecha-nisms leading to density-dependent predationmortality as a result of predator behaviour are de-scribed in Section 12 .3 .2 . Density-dependent preyresponses can also influence mortality rates . Forexample, competition for food at high density maylead to reduced growth rates, with prey beingvulnerable to gape-limited piscivores for longerperiods .

12.6.1 Population andcommunityeffects

Piscivorous fishes are known to have a dramaticinfluence on population- and community-leveldynamics (Persson, Chapter 15, this volume) . Theimpacts of piscivores can extend down severaltrophic levels (Zaret and Paine 1973) . In freshwatersystems, the trophic cascade model describeszooplanktivore, zooplankton, and phytoplanktonabundance under high and low levels of piscivoreabundance (see Carpenter et al . 1985 ; Persson,Chapter 15, this volume; Kaiser and Jennings,Chapter 16, Volume 2) . When piscivore abundancelevels are high, density of zooplanktivorous fishis reduced, which leads to increased zooplankton

277

278

Chapter 12

(herbivore) and reduced phytoplankton abundancelevels. Opposite abundance levels to those de-scribed above are observed during periods of lowpiscivore density. Hambright (1994) describes howthe vulnerability of zooplanktivores to piscivorydetermines the magnitude of zooplanktivory andthus its effect on lower trophic levels . Piscivorousfish can influence the life-history strategy of theirprey. Reznick et al . (1990) showed that under sepa-rate piscivore regimes (low vs . high predation),guppies (Poecilia reticulate) evolved different life-history parameters such as age at maturity, repro-ductive effort, brood size and size of offspring(Hutchings, Chapter 7, this volume) .

Cannibalism is interesting because of its poten-tial role in regulating population levels through itscontribution to natural mortality ; however, evi-dence for density-dependent cannibalism in non-captive fishes has not been discovered (Smithand Reay 1991) . Henderson and Corps (1997)found that bass (Dicentrarchus labrax) year-classstrength had a 3-year periodicity, which theybelieved to result from cannibalism within thenursery .

Animal abundance levels are known to exhibitperiodicity as a result of predator-prey or host-par-asite cycles . In a predator-prey cycle, increases inpredator abundance are concurrent with declinesin prey population levels ; subsequent periods oflow prey abundance can lead to predator declinesand prey release . We are aware of only one exampleof a predator-prey cycle for a piscivorous fish .Pacific cod (Gadus macrocephalus) and herring(Clupea harengus pallasi) populations in HecateStrait, British Columbia have patterns of abun-dance that are suggestive of such a cycle (Walterset al . 1986) . The lack of evidence for a piscivore-prey cycle may result from the flexible feedingbehaviour of piscivores, allowing them to remainat stable population levels when one of their preypopulations is at low abundance .

Piscivores can have indirect non-lethal effectson prey fishes. These include all of the behaviouralresponses made by prey when confronted with apiscivore . These behavioural responses can leadto differences in community structure . The domi-nant indirect effect on prey fishes in a northern bog

in the USA after northern pike were introducedwas emigration ; this led to a rapid decline in preyfish abundance and a change in community struc-ture (He and Kitchell 1990) .

12 .7 METHODS OFSTUDYING PREDATION

IN THE FIELD

There are several methods used to determine theeffects of piscivorous fish on prey populations andcommunities . A direct approach is to make com-parisons between treatments where piscivores arepresent or absent . These treatments maybe naturalor artificial . Modelling has been used successfullyto determine the influence of predators on preypopulations (Mittelbach, Chapter 11 and Persson,Chapter 15, this volume). Studies that estimatepredation mortality and compare this to totalprey mortality make up another group of studies .Estimates of consumption rate along with knowl-edge of diet composition, prey sizes and predatorabundance are used to estimate predation mortali-ty. Detailed descriptions of stomach content analy-sis and methods to estimate consumption rates aredescribed by Jobling (Chapter 5, this volume) .

Tethers (usually monofilament fishing line) areused to hold prey fish in place; the fate of the preyfish allows one to assess relative predation inten-sity in both space and time . Danilowicz and Sale(1999) used tethering of French grunt (Haemulonflavolineatum) to determine when, over a dielcycle, predation intensity was highest . They foundthat the risk of being eaten was lowest during diur-nal periods and highest during dusk and nocturnalperiods. Age-0 rainbow trout (Salmo gairdneri)were tethered in British Columbia lakes to deter-mine the spatial and temporal patterns of pis-civory; it was concluded that the risk of predationvaried greatly over both space and time (Post et al .1998) . Clearly, tethering experiments, as withother predator-prey studies, would need to con-form to the animal ethics regulations of the coun-tries and institutions involved .

Filming has been used to both identify preda-tors and gain insight into the intensity of piscivory.

Carr and Hixon (1995) used filming of small patchreefs to aid in identifying pelagic piscivores and formeasuring their visitation rates to experimentalreefs .

12 .8 IMPLICATIONS FORCONSERVATION AND

MANAGEMENT

Fishing alters the age and size structure of popula-tions as older and larger fish are often removedfirst, and eventually the fishery is supported bythe small newly recruited individuals . In manyecosystems, large piscivorous species were theinitial targets of fishing . After the fishing downof these populations, the fisheries shifted tosmaller species at lower trophic levels (Pauly andChristensen, Chapter 10, Volume 2) . Althoughmost marine landings are of small planktivores, itis fishing and the removal of large species at hightrophic levels that affects ecosystem structureand functioning, leading to 'top-down' control(Carpenter et al . 1985; Hixon and Can 1997 ;Kaiser and Jennings, Chapter 16, Volume 2) . Forexample, on Georges Bank, the decline in ground-fish stocks has led to a dramatic increase in foragefish such as herring and mackerel and a replace-ment of gadid and flounder species by small elas-mobranchs (Fogarty and Murawski 1998) . As moreand more stocks become overexploited, fewer toppredators will dominate oceanic food webs andboth direct and indirect community effects will beprevalent. Predator removals can also be a deliber-ate form of fisheries management, whereby de-creases in predation pressure are predicted toresult. However this strategy does not alwaysproceed as predicted (V. Christensen 1996 ; Kaiserand Jennings, Chapter 16 and Cowx, Chapter 17,Volume 2) .

The opposite problem, with similar results,occurs when predators are deliberately or inadver-tently added to ecosystems (Cowx, Chapter 17,Volume 2) . The effects of piscivore introductionhas been documented in temperate, boreal andtropical lakes (Zaret and Paine 1973 ; Mills et al .1994; Vander Zanden et al. 1999). The most dra-

Feeding Ecology ofPiscivorous Fishes 279

matic recent example has been the introduction ofthe Nile perch (Lutes nilotica) to Lake Victoria, theworld's largest tropical lake by surface area . Themain effect of the predator introduction, similarto that observed in most systems where it hasoccurred, has been an accelerated decline of thediverse endemic cichlid fauna of the lake and therecent extinction of at least 200 species (Seehausenet al . 1997) . However, the Nile perch introductionhas also resulted in a fourfold increase in fisheryyield and a doubling of fishery-related employ-ment (Pitcher and Hart 1995 ; Kitchell et al . 1997 .

Piscivore stocking, particularly in freshwatersystems, has also been used as a way to improvewater quality in eutrophic lakes and reservoirs .This process, termed 'biomanipulation', worksthrough trophic cascading and has been quitesuccessful in improving water quality of culturallyeutrophic lakes and reservoirs (Drenner andHambright 1999, although it can also have avariety of unexpected negative impacts .

Marine reserves have been effective in restoringand protecting many reef fishes (Polunin, Chapter14, Volume 2) and they appear to be a promisingtool for marine ecosystem sustainability (NationalResearch Council 1999) . However, reserves, ifnot planned carefully enough, can have negativeimpacts because of potential effects of physicalfactors combined with increased protection ofpiscivorous fishes . For example, in a small reservein Barbados, size and abundance of piscivoreswere greater within the reserve than on adjacentreefs, but recruitment of grunts (Haemulidae)was lower because of predation pressure andoceanographic patterns of larval supply (TupperandJuanes1999) .

12 .9 CONCLUSIONSPiscivores are a diverse group of fish that show dis-tinct behaviours and specialized morphologies .The timing of the onset of piscivory is critical tomany species and may have led to adaptive strate-gies that accelerate the shift to piscivory. Thepredation process can be better understood byexamining the components of predation, and the

280

Chapter 12

predatory response to changes in prey density canbe estimated by generating numerical and func-tional responses. Piscivorous fish exhibit generalpatterns of prey type and size selectivity driven byattack behaviours and prey-specific behaviouralresponses by the prey. Predatory fish can have largeimpacts on prey communities in all habitats wherethey occur and are thus common targets for cons er-vation and management, as they are often the firstspecies to be affected by harvesting ; however, theycan also have both positive and detrimental effectswhen deliberately or inadvertently added toecosystems .

REFERENCES

Adams, S.M. and DeAngelis, D .L . (1987) Indirect effectsof early bass-shad interactions on predator populationstructure and food web dynamics. In : W.C. Kerfoot andA. Sib (eds) Predation: Direct and Indirect Impactson Aquatic Communities . Hanover, NH : UniversityPress of New England, pp . 103-17 .

Bailey, K.M. (1994) Predation on juvenile flatfish and re-cruitment variability . Netherlands Journal of SeaResearch 32,175-89,

Blake, R.W. (1983) Fish Locomotion . Cambridge: Cam-bridge University Press.

Buckel, J.A., Letcher, B .H. and Conover, D .O.(1998) Ef-fects of a delayed onset of piscivory on size of age-0bluefish . Transactions of the American Fisheries Soci-ety 127, 576-87 .

Buckel, J.A., Fogarty, M .J. and Conover, D.O. )1999a)Mutual prey of fish and humans: a comparison ofbiomass consumed by bluefish, Pomatomus saltatrix,with that harvested by fisheries . Fishery Bulletin 97,776-85 .

Buckel, J.A., Conover, DO., Steinberg, N .D . andMcKown, K.A. (1999b) Impact of age-0 bluefish pre-dation on age-0 fishes in the Hudson River estuary :evidence for density-dependent loss of juvenile stripedbass . Canadian Journal of Fisheries and AquaticSciences 56,275-87 .

Buckel, J.A ., Fogarty, M .J . and Conover, D .O. (1999c) For-aging habits of bluefish, Pomatomus saltatrix, on theU.S . east coast continental shelf . Fishery Bulletin 97,758-75 .

Buijse, A.D. and Houthuijzen, R.P. (1992) Piscivory,growth, and size-selective mortality of age 0 pikeperch(Stizostedion lucioperca). Canadian Journal ofFisheries andAquatic Sciences 49, 894-902 .

Carpenter, S.R., Kitchell, J .F. and Hodgson, J.R. (1985)Cascading trophic interactions and lake productivity .BioScience 35,634-9.

Carr, M.H. and Hixon, M .A. (1995) Predation effects onearly post-settlement survivorship of coral-reef fishes .Marine EcologyProgress Series 124,31-42 .

Chesson, J. (1978) Measuring preference in selectivepredation . Ecology 59, 211-15 .

Christensen, B . (1996) Predator foraging capabilities andprey antipredator behaviours : pre- versus postcaptureconstraints on size-dependent predator-prey interac-tions. Oikos 76,368-80 .

Christensen, V. (1996) Managing fisheries involvingpredator and prey species . Reviews in Fish Biology andFisheries 6,417-42 .

Danilowicz, B.S . and Sale, P.F. (1999) Relative intensity ofpredation on the French grunt, Haemulon flavolinea-turn, during diurnal, dusk, and nocturnal periods on acoral reef . Marine Biology 133,337-43,

Dill, L .M. (1983) Adaptive flexibility in the foragingbehavior of fishes . Canadian Journal ofFisheries andAquatic Sciences 40,398-408 .

Drenner, R. W. and Hambright, K .D. (1999) Biomanipula-tion of fish assemblages as a lake restoration tech-nique .Archly fiirHydrobiologie 146, 129-65 .

Eby, L.A ., Rudstam, L.G. and Kitchell, J .F. (1995) Predatorresponses to prey population dynamics : an empiricalanalysis based on lake trout growth rates . CanadianJournal ofFisheries andAquatic Sciences 52,1564-71 .

Eklov, P. (1992) Group foraging versus solitary foragingefficiency in piscivorous predators : the perch, Percafluviatilis, and pike, Esox lucius, patterns . AnimalBehaviour44,313-26 .

Fogarty, M .J . and Murawski, S .A. (1998) Large-scale dis-turbance and the structure of marine systems : fisheryimpacts on Georges Bank. Ecological Applications 8(Suppl .), S6-S22.

Forney, J .L . (1971) Development of dominant year classesin a yellow perch population . Transactions of theAmerican Fisheries Society 100, 739-49 .

Forrester, G .E . (1995) Strong density-dependent survivaland recruitment regulate the abundance of a coral reeffish . Oecologia 103,275-82 .

Frankiewicz, P., Dabrowski, K . and Zalewski, M . (1996)Mechanism of establishing bimodality in a size distri-bution of age-0 pikeperch, Stizostedion lucioperca (L .)in the Sulejow Reservoir, central Poland . AnnalesZoologie Fennici 33,321-7 .

Gaughan, D . J. and Potter, I.C. (1997) Analysis of diet andfeeding strategies within an assemblage of estuarinelarval fish and an objective assessment of dietary nicheoverlap . Fishery Bulletin 95, 722-31 .

Gerking, S .D. (1994) Feeding Ecology of Fish. New York:Academic Press .

Gerritsen, J. and Strickler, J .E . (1977) Encounter prob-abilities and community structure in zooplankton :a mathematical model . Journal of the Fisheries Re-search Boardof Canada 34, 73-82 .

Greene, C.H. (1986) Patterns of prey selection : implica-tions of predatorforaging tactics . American Naturalist128,824-39 .

Hambright, K.D. (1994) Morphological constraints inthe piscivore-planktivore interaction : implicationsfor the trophic cascade hypothesis. Limnology andOceanography 39,897-912 .

Hambright, K.D ., Drenner, R.W., McComas, S.R. andHairston, N.G. (1991) Gape-limited piscivores, plank-tivore size refuges, and the trophic cascade hypothesis .Archiv fur Hydrobiologie 121, 389-404 .

Hart, P.J .B . (1993) Teleost foraging : facts and theories . In :T.J . Pitcher (ed.) Behaviour of Teleost Fishes . London :Chapman &Hall, pp . 253-84 .

Hart, P.J .B . and Hamrin, S .F. (1990) The role of behaviourand morphology in the selection of prey by pike . In :Hughes led.) Behavioural Mechanisms of Food Selec-tion . Heidelberg : Springer-Verlag, pp . 235-54 .

Hassell, M .P. (1978) The Dynamics of Arthropod Preda-tor-Prey Systems. Princeton, NJ: Princeton UniversityPress .

He, X. and Kitchell, J .F. (1990) Direct and indirect effectsof predation on a fish community : a whole lake experi-ment. Transactions of theAmerican Fisheries Society119,825-35 .

Henderson, P. and Corps, M . (1997) The role of tempera-ture and cannibalism in interannual recruitment vari-ation of bass in British waters . Journal of Fish Biology50,280-95 .

Hixon, M. and Carr, M . (1997) Synergistic predation, den-sity dependence, and population regulation in marinefish . Science 277, 946-9 .

Hobson, E .S . (1991) Trophic relationships of fishes spe-cialized to feed on zooplanters above coral reefs . In: P.F.Sale (ed.) The Ecology of Fishes on Coral Reefs . NewYork: Academic Press, pp . 69-95 .

Hulling, C .S . (1965) The functional response of predatorsto prey density and its role in mimicry and populationregulation . Memoirs of the Entomological Society ofCanada 45, 3-60.

Hoyle, J.A. and Keast, A. (1987) The effect of prey mor-phology and size on handling time in a piscivore, thelargemouth bass (Micropterus salmoides). CanadianJournal of Zoology65, 1972-7.

Jenkins, G.P., Milward, N.E. and Hartwick, R .F. (1984)Food of larvae of Spanish mackerels, genus Scombero-morus (Teleostei : Scombridae), in shelf waters of theGreat Barrier Reef . Australian Journal of Marine andFreshwater Research 35,477-82.

Juanes, F. (1994) What determines prey size selectivity in

Feeding Ecology of Piscivorous Fishes 281

piscivorous fishes? In : D .J. Stouder, K.L. Fresh and R .J .Feller (eds) Theory and Application in Studies ofFishFeeding Ecology. Columbia, SC: University of SouthCarolina Press, pp. 79-100 .

Juanes, F. and Conover, D .O. (1994a) Piscivory and preysize selection by young-of-the-year bluefish: predatorpreference or size-dependent capture success? MarineEcologyProgress Series114,59-69 .

Juanes, F. and Conover, D .O. (1994b) Rapid growth, highfeeding rates, and early piscivory in young-of-the-yearbluefish (Pomatomus saltatrix). Canadian Journal ofFisheries and Aquatic Sciences 51, 1752-61 .

Juanes, F. and Conover, D.O. (1995) Size-structured pis-civory: advection and the linkage between predatorand prey recruitment in young-of-the-year bluefish .MarineEcology Progress Series 128,287-304.

Juanes, F., Buckel, J .A. and Conover, D.O. (1994) Acceler-ating the onset of piscivory : intersection of predatorand prey phenologies . Journal of Fish Biology 45(Suppl . A), 41-54 .

Keast, A. (1985) The piscivore guild of fishes in smallfreshwater ecosystems. Environmental Biology ofFishes 12,119-29.

Keast, A . and Webb, D. (1966) Mouth and body form rela-tive to feeding ecology in the fish fauna of a small lake,Lake Opinicon, Ontario . Journal of the Fisheries Re-search Board of Canada 23,1845-74 .

Kitchell, J .F., Schindler, D.S., Ogutu-Ohwayo, R . andReinthal, P.M. (1997) The Nile perch in Lake Victoria :interactions between predation and fisheries. Ecologi-calApplications 7,653-64 .

Magnusson, K.G. (1995) An overview of the multispeciesVPA: theory and applications . Reviews in Fish Biologyand Fisheries 5, 195-212.

Mangel, M. (1996) Life history invariants, age at maturityand the ferox trout . Evolutionary Ecology 10, 249-63 .

Margulies, D. (1997) Development of the visual systemand inferred performance capabilities of larval andearly juvenile scombrids . Marine Behaviour and Phys-iology 30, 75-98 .

Mathias, J.A. and Li, S . (1982) Feeding habits of walleyelarvae and juveniles : comparative laboratory and fieldstudies . Transactions of the American Fisheries Soci-ety 111, 722-35 .

Miller, T.J ., Crowder, L.B ., Rice, J.A. and Marschall, E .A .(1988) Larval size and recruitment mechanisms infishes: toward a conceptual framework . CanadianJournal ofFisheries andAquatic Sciences 45,1657-70 .

Mills, E.L ., Leach, J.H., Carlton, J .T. and Secor, C .L .(1994) Exotic species and the integrity of the GreatLakes . Bioscience 44,666-76 .

Mittelbach, G.G. and Persson, L . (1998) The ontogeny ofpiscivory and its ecological consequences . CanadianJournalofFisheries and Aquatic Sciences 55, 1454-65 .

282

Chapter 12

Murdoch, W. W. and Oaten, A. (1975) Predation and popu-lation stability. Advances in Ecological Research 9,1-131 .

Myers, R.A . and Cadigan, N.C. (1993) Density-dependent juvenile mortality in marine demersal fish.Canadian Journal of Fisheries and Aquatic Sciences50,1576-90 .

National Research Council (1999) Sustaining MarineFisheries . Washington, DC: National Academy Press .

Olson, M.H. (1996) Ontogenetic shifts in largemouthbass: variability and consequences for first-yeargrowth. Ecology77,179-90.

Pearre, S . (1986) Ratio-based trophic niche breadthsof fish, the Sheldon spectrum, and the size-efficiencyhypothesis . Marine Ecology Progress Series 27, 299-314 .

Persson, L., Andersson, J ., Wahistrom, E. and Eklov, P.(1996) Size-specific interactions in lake systems :predator gape limitation and prey growth rate and mor-tality. Ecology 77, 900-1 1 .

Peterman, R .M. and Carte, M. (1978) Estimation offunctional responses of predators on juvenile salmon .Journal ofthe Fisheries Research Board of Canada 35,797-808 .

Petersen, J.H. and DeAngelis, D .L. (1992) Functionalresponse and capture timing in an individual-basedmodel: predation by northern squawfish (Pty-chocheilus oregonensis) on juvenile salmonids in theColumbia River . Canadian Journal of Fisheries andAquatic Sciences 49,2551-65 .

Pitcher, T.J . and Hart, P.J .B . (1995) TheImpacts ofSpeciesChanges in African Lakes. London: Chapman &Hall .

Pitcher, T.J. and Parrish, J .K. (1993) Functions of shoalingbehaviour in teleosts . In : T.J . Pitcher (ed.) The Behav-iour of Teleost Fishes . London: Chapman & Hall, pp .363-439 .

Post, J .R ., Parkinson, E .A. and Johnston, N .T. (1998) Spa-tial and temporal variation in risk to piscivory of age-0rainbow trout: patterns and population level conse-quences . Transactions of the American FisheriesSociety 127,932-42 .

Reimchen, T .E. (1991) Evolutionary attributes of head-first prey manipulation and swallowing in piscivores .Canadian Journal of Zoology 69, 2912-16 .

Reznick, D.A., Bryga, H. and Endler, J .A. (1990) Experi-mentally induced life-history evolution in a naturalpopulation. Nature 346,357-9 .

Rice, J.A., Crowder, L .B . and Rose, K.A. (1993) Interac-tions between size-structured predator and prey popu-lations : experimental test and model comparison .Transactions of the American Fisheries Society 122,481-91 .

Ruggerone, G .T. and Rogers, D .E. (1983) Arctic chart pre-

dation on sockeye salmon smolts at Little TogiakRiver, Alaska . Fishery Bulletin 82, 401-10 .

Sazima, I. (1998) Field evidence for suspension feeding inPseudocaranx dentex, with comments on ram filter-ing in otherjacks (Carangidae) . Environmental BiologyofFishes 53, 225-9 .

Scharf, F.S ., Buckel, J.A ., Juanes, F. and Conover, D.O .(1997) Estimating piscine prey size from partial re-mains : testing for shifts in foraging mode by juvenilebluefish . Environmental Biology of Fishes 49,377-88 .

Scharf, F.S ., Buckel, J.A., Juanes, F. and Conover, DO.(1998) Predation by juvenile piscivorous bluefish(Pomatomus saltatrix) : the influence of prey to preda-tor size ratio and prey type on predator capture successand prey profitability. Canadian Journal of Fisheriesand Aquatic Sciences 55, 1695-703 .

Scharf, F.S ., Juanes, F. and Rountree, R.A. (2000) Predatorsize-prey size relationships of marine fish predators :interspecific variation and the effects of ontogeny andbody size on trophic niche breadth . Marine EcologyProgress Series 208,229-48 .

Seehausen, 0 ., Witte, F., Katunzi, E .F., Smits, J . andBouton, N. (1997) Patterns of remnant cichlid faunain southern Lake Victoria . Conservation Biology 11,890-904 .

Shepherd, J.G. and Cushing, D .H. (1990) Regulation infish populations: myth or mirage? PhilosophicalTransactions of the Royal Society of London Series B330,151-64 .

Sill, A. and Christensen, B . (2001) Optimal diet theory :when does it work and when and why does it fail?Animal Behaviour 61,379-90 .

Skulason, S ., Snorrason, S .S ., Ota, D. and Noakes, D .L.G .(1993) Genetically based differences in foraging behav-iour among sympatric morphs of arctic chary (Pisces :Salmonidae) . AnimalBehaviour 45,1179-92 .

Smith, C. and Reay, P. (1991) Cannibalism in teleost fish .Reviews in Fish Biology andFisheries 1, 41-64 .

Sogard, S.M. (1997) Size-selective mortality in the juve-nile stage of teleost fishes : a review. Bulletin of MarineScience60,1129-57,

Stephens, D.W. and Krebs, JR . (1986) Foraging Theory.Princeton, NJ : Princeton University Press .

Tanaka, M ., Kaji, T., Nakamura, Y . and Takahashi, Y.(1996) Developmental strategies of scombrid larvae :high growth potential related to food habits and preco-cious digestive system development . In : Y. Watanabe,Y. Yamashita and Y . Oozeki (eds) Survival Strategies inEarly Life Stages of Marine Resources . Rotterdam :A.A. Balkema, pp . 125-39 .

Tupper, M. and Juanes, F. (1999) Effects of a marine re-serve on recruitment of grunts (Pisces : Haemulidae)at Barbados, West Indies . Environmental Biology ofFishes 55, 53-63 .

Utne-Palm, A.C. (2000) Prey visibility, activity, size andcatchability's (evasiveness) influence on Gobiusculusflavescens prey choice . Sarsia 85, 157-65 .

Vander Zanden, M.J ., Casselman, J.M. and Rasmussen,J.B. (1999) Stable isotope evidence for the food web con-sequences of species invasions in lakes. Nature 401,464-7 .

Wahl, D.H. and Stein, R .A. (1988) Selective predation bythree esocids : the role of prey behavior and morphol-ogy. Ytansactions of the American Fisheries Society117,142-51,

Wainwright, P.C. and Richard, BA . (1995) Predicting pat-terns of prey use from morphology of fishes . Environ-mental Biology of Fishes 44,97-113 .

Walters, C .J., Stocker, M ., Tyler, A.V. and Westrheim, S.J.(1986) Interaction between Pacific cod (Gadus macro-cephalus) and herring (Clupea harengus pallasi) in theHecate Strait, British Columbia . Canadian Journal ofFisheries andAquatic Sciences 43,830-7 .

Webb, P.W. 11976) The effect of size on the fast-startperformance of rainbow trout, Salmo gairdneri, anda consideration of piscivorous predator-prey interac-tions . Journal of Experimental Biology 65,157-77 .

Feeding Ecology of Piscivorous Fishes 283

Werner, E.E. (1977) Species packing and niche comple-mentarity in three sunfishes . American Naturalist111,553-78 .

Werner, E .E . and Gilliam, J .F . (1984 The ontogeneticniche and species interactions in size-structured popu-lations . Annual Review of Ecology and Systematics15,393-425 .

Wilson, D .S . )1975) The adequacy of body size as a nichedifference . American Naturalist 109, 769-84 .

Winemiller, K.O. (1989) Ontogenetic diet shifts and re-source partitioning among piscivorous fishes in theVenezuelan Llanos . Environmental Biology of Fishes26,177-99 .

Wootton, R .J. (1990) Ecologyof Teleost Fishes . New York :Chapman & Hall .

Wright, R.A., Crowder, L.B. and Martin, T.H. (1993)The effects of predation on the survival and size-distribution of estuarine fishes: an experimentalapproach. Environmental Biology of Fishes 36, 291-300.

Zaret, T.M. and Paine, R .T. (1973) Species introduction ina tropical lake . Science 182,449-55 .

Handbook of Fish Biologyand FisheriesVOLUME 1FISH BIOLOGY

EDITED BYPauIJ .B. Hart

Department of BiologyUniversity of Leicester

AND

John D. ReynoldsSchool of Biological Sciences

University of East Anglia

ft

BlackwellPublishing

© 2002 by Blackwell Science Ltda Blackwell Publishing company

Chapter 8 © British Crown copyright, 1999

350 Main Street, Malden, MA 02148-5018, USA108 Cowley Road, Oxford OX41JF, UK

550 Swanston Street, Carlton South, Melbourne, Victoria 3053, AustraliaKurfUrstendamm 57, 10707 Berlin, Germany

The right of Paul J .B. Hart and John D . Reynolds to be identified as the Authorsof the Editorial Material in this Work has been asserted in accordance with the

UK Copyright, Designs, and Patents Act 1988 .

All rights reserved . No part of this publication maybe reproduced, stored in a retrievalsystem, or transmitted, in any form or by any means, electronic, mechanical,

photocopying, recording or otherwise, except as permitted by the UK Copyright,Designs, and Patents Act 1988, without the prior permission of the publisher .

First published 2002 by Blackwell Science Ltd

Library of Congress Cataloging-in-Publication Data has been applied for .

Volume 1 ISBN 0-632-05412-3 (hbk)Volume 2ISBN 0-632-06482-X (hbk)2-volume set ISBN 0-632-06483-8

A catalogue record for this title is available from the British Library .

Set in 9/11 .5 pt Trump Mediaevalby SNP Best-set Typesetter Ltd, Hong Kong

Printed and bound in the United Kingdom by TJ International Ltd, Padstow, Cornwall .

For further information onBlackwell Publishing, visit our website:h ttp://www.blackwellpublishing .com

P

e

3