Transactions of the American Fisheries Society 123:855-865. 1994 © Copyrigh thy b te American Fisheries Society 1994

Effects of Prey Size, Abundance, and Population Structure on Piscivory by Yellow Perch CYNTHIA A. PASZKOWSKJ AND WILLIAM M. TONN Department of Zoology. University of Alberta Edmonton, Alberta T6G 2E9, Canada

Abstract. —To examine how predator and prey body sizes, prey abundance, and prey population structure affect piscivore-prey interactions, feeding experiments were conducte largn o d e (>190 mm tota lsmald lengthan ) l (150-18TL , yello) 9mm w perch Perca flavescens given small (43-55 mm TL), medium (60-66 mm) largd ,an e (71-8 fathea) 0mm d minnows Pimephales promelass a prey. Handling costs suggested that large yellow perch shoul t discriminatdno e among prey sizes, but small yellow perch should select small fathead minnows over large. Result f feedinso g trials with only one size of prey were consistent with the cost indications. Large yellow perch consumed similar numbers and biomasses of the three size-classes of fathead minnows in trials with only size pref eon o y available. Small yellow perc more hat e small fathead minnows than mediud man more medium than large. Thus, small yellow perch gained less weight with large fathead minnows than with small one theis sa r prey. Whe l threnal e prey sizes were available, both larg smald ean l yellow perch consistently captured more small fathead minnows than large, even when total prey abundance or the size structure of prey populations varied. Overall consumption by large yellow t affectepercno s prey hdb wa y abundance smalt bu , l yellow perch increased consumptio t highena r abundance eatiny sb g more smal mediud lan m prey. Increasin proportioe gth mediuf no m fathead minnows, while decreasin proportioe gth largf o n t affece preyno d t totadi , l consumptio eithey nb r large or small yellow perch. However, small yellow perch responded to this shift by eating more medium and fewer large fathead minnows. Selection by yellow perch for prey fish smaller than expected based on gape limitations and handling times, and the responses of yellow perch to change thein si r prey base under experimental conditions, help explain pattern f predatioso d nan coexistence between this piscivor pres it nature n yi d ean .

Size selection of prey by fish that are large rel- estimated (Tonn and Paszkowski 1986; Persson alive to their food organisms (e.g., zooplankti- et al. 1992). vores) has been studied intensively (e.g., Werner Predation patterns of Perca are known to be Hald an l 1974; . Stei1984al t ne ; Ryer 1988). Pis- affecte y preb d y size (Popova 1967; Tond an n civores are less well studied, especially expert- Paszkowski 1986; Post and Evans 1989; Tonn et mentally. Factors presumed to affect prey size se- al. 1991). This predation can, in turn, have sig- lectio piscivorey nb oftee sar n extrapolated from nificant, long-lasting impact pren so y populations dat n otheo a r foragers rather than documented bV reducing densities, limiting recruitmentd an , from direct observation. Even among piscivorous skewing size distributions toward large, less-vul- fish, relatively large, specialized predators (e.g., nerable body sizes

Downloaded by [University of Alberta] at 09:17 16 December 2011 the attention (e.g., Hoyl Keasd ean t 1988; Wahl « fPP'. selectively alter d habitaan e us t Steid - an ?'as n t clea wel e w l1988)no th activitlho rs i t I . y pattern pref so y fis differenf ho t sizes sumption predictiond san s mad studiey eb largf so e £°™ anf Pa^owski 987; Holopame. al t ne predators apply to smaller, less-specialized pisci- 1991) and thus their availability to piscivores may be indirectly affected. Perca spp. range widely throughout North To further understand factors shaping size-bi- America and Eurasia and are often the most com- ased f1™^ and'*s P°7tial 'T^* On ^ monly encountered specie laken si s withi - re P<>P" na la"<>n*> ** addresse e followinth d g ques- gion (Harvey 1981; Appelberg et al. 1989; Tonn tlons- . 1990)eal t . Although usually viewe generals da o postcapturd - w (1Ho ) e handling- re time d san ists (Keast 1979), yellow perch Perca flavescens or turn rates (biomass of prey ingested per unit Eurasian perch P.fluviatilis can be the most abun- handling time) vary with prey size? dant r onlyo , , piscivor smaln ei lthese lakesAr e ) relationship.(2 Thus e th , s reflecte nume th n -i d importance of predation by Perca spp. in the dy- her and biomass of different sizes of prey con- namics offish communities should not be under- sumed by piscivores? 855 856 PASZKOWSK TOND JAN N

(3) Do predation patterns change when total prey Zoologf o y aquatic facility. Universit Alberta)f yo . abundance is varied but the size distribution Yellow perch were captured in Baptiste and Amisk pref o hels yi d constant? lakes near Athabasca, Alberta, from May through (4) Do predation patterns change when the total September with Windermere perch trap beaca r so h abundanc f preeo s heli y d constan t sizbu te seine. Fathead minnows were captured with min- distribution variede sar ? now trap seina r so e from several small lakes near (5) Do small and large piscivores respond differ- Athabasca. Upon capture, fish were treated with above entlth o yet manipulation preyf so d an , Furan-2 (Aquarium Pharmaceuticals r 12-2)fo 4h if so, how? preveno t t bacteria fungad an l l infections. Yellow perch were held individuall t 15-19°a y a d Can Our study used two sizes of yellow perch and photoperiod of 12 h light: 12 h darkness in aer- three sizes of fathead minnow Pimephales pro- ated tanks, 122 x 51 x 61 cm, with one glass melas. "Small" yellow perch (150-189 mm total side. Before data were recorded, yellow perch were length, TL) represented individuals at the lower acclimated to the observation tank for 1-2 weeks size th e f rango d een that regularly feed fisn o sh d weran e presente e fatheadon d minnow daily. and are especially sensitive to prey size (Tonn and Data collection began whe yellona w perch rou- Paszkowski 1986). "Large" yellow perch (190-229 tinely ate fathead minnows in the presence of an mm) were expected to be able to capture and han- observer withi mi0 n3 f pre no y introduction. l sizeal f fatheae o s dl d minnows. Fathead min- Handling times were determined during day- nows, like other small-bodied, soft-rayed fishes, light hour larg7 r esfo yellow perch (198-21m 0m are particularly vulnerabl piscivoreo et s (Harvey TL) and 15 small yellow perch (156-189 mm TL) 1981; Robinson 1989; Robinson and Tonn 1989; eating fathead minnows that were 43-80 mm TL. Savin Steid oan n 1989). However, yellow perch Experimental yellow perch receive e fatheadon d fathead an d minnows commonly co-occu smaln ri l minnow per day based on a feeding schedule that lakes, where yellow perc oftes h i onle th ny pisci- randomized the order in which the prey size-class- vore (Ton Magnusod nan n 1982; Ton alt ne . 1990). es (see below) were presented. Prior to a trial, prey Becaus e fatheath f eo d minnow's sexual dimor- were weighed and measured for total length. phis semelparoud man s reproductive pattern (Scott beginnine Ath t triala f go fatheaa , d minnow and Crossman 1973; Unger 1983), its populations was transferred into the observation tank, held for are characterized by seasonally distinct size a handis n i -d n netd themi 1an ,n released. Timing tributions (Price et al. 1991). Size-selective or size- began whe yelloe nth w perch capture- im d dan limited piscivory, interacting with size-structured mediately ingested the fathead minnow. If the prey populations, could therefore strongly affect predator repeatedly struck and released the prey both the foraging success of yellow perch and the (which occurre l ingestionn <5al i d f %o - ob s impact of this predation on fathead minnow pop- served), onl finae yth l captur uses measureo dwa t e ulations. handling time. Yellow perch typically captured fathead minnows headfirst l ingestion(87al f %o s Methods observed), sunk toward the bottom of the tank, Handing Times becamd an e immobile while handling preye Ey .

Downloaded by [University of Alberta] at 09:17 16 December 2011 determino T patternf ei f preso y consumption movement decrease operculad dan r movements could be related to a measure of prey profitability, were slo exaggeratedd wan . Handlin deemes gwa d we calculated return rate, the mass of prey in- complete when typical eye movement, respiratory geste unir dpe t time spent handlin preye gth . Han- movement, and swimming resumed (Paszkowski dling time, from capture of prey through inges- . 1989)eal t yelloa f I . w t captur percno d e hdi eth tion, has been used widely in evaluating foraging prey within 30 min, the fathead minnow was re- behavior (e.g., Werner and Hall 1974; Stein et al. moved and the trial was repeated later the same 1984). Prey handling after captur s dictatei e d dayyelloa f I . w perch faile givea o fee t dn o dn largelymorphologiee th y b predatof so pred ran y day fatheaa , d minno same th f ew o pre s sizewa - (Werner 1974), influenced littl y spatiab e l con- sented the next day. straints, and thus can be observed realistically in captivity. Predation Experiments Handling times were measure yellor dfo w perch General protocol. — Predation experiments were eating fathead minnow laboratore th n si y (Mean- conducted in four circular fiberglass tanks, 1.5 m ook Biological Research Statio Departmend nan t in diameter, that were filled with wate depta o rt h PISCIVORY BY YELLOW PERCH 857

of 0.75 m and aerated. Trials were conducted dur- Every 2 d, fathead minnows were captured with ing May-Septembe rooa n ri m illuminate naty db - large hand nets and counted by size-class. The ural daylight and isolated from other activity. Wa- number eate size-clasr npe recordes w swa ne d dan ter temperatures in tanks varied between 14 and fathead minnows were adde restoro dt e popula- 20°C compensato T . seasonar efo l changee th n si tion theio st r original abundance sizd sean distri- condition of the environment and the fish, differ- butions triaa l f o al l d . en Trial e th st lasteA . d d8 ent treatments within an experiment (see below) fish were removed, fathead minnows were count- were assigned in a stratified random design among ed by size-class, and yellow perch were weighed the four tanks and run simultaneously. individually. The total number of fathead min- Experimental fish came from the same lakes as now sfrod eate8 m n ni eac h size-clas- re s wa s those use handlinr dfo g time measurements. Yel- corded and numbers were converted to biomass low perch and fathead minnows were held sepa- by using mean weight r eacfo s h size-class. Fish ratel outdoon yi diametern i r tankm 8 1. s . Every were returne holdino dt g tanks and, afte mina r - 1-2 d, yellow perch were fed fathead minnows and imum of 8 d, were considered available for use in similarly sized northern redbelly dace Phoxinus subsequent trials. Fish were drawn randomly from cos. Fathead minnows received flaked fish food these large stocks (50-150 yellow perch; 500-1,000 daily. fathead minnows) whe groupw nne f predatorso s The day before a trial, six yellow perch, all from and prey were assembled. Thus, we ensured that eithe smale rth largr o l e size-class, were measured group composition changed across trial eacd san h for total length and wet mass, and each was marked group was a unique entity. differentl severiny b y dorsae gon l spine (second Types of'experiments.— thren Wera - e setex f so through seventh). To standardize size distribu- periments for each of the two sizes of yellow perch. tions, each group of six fish comprised three sets Replicates of a given treatment varied between 5 of pairs based on fish size (size ranges of pair sr larg werfo e7 e yellod an w perc d betweehan d an 7 n 190-201, 202-219 d 220-22r largfo an , em 9m 10 for small yellow perch. In experiment 1, where yellow perc 150-162d han , 163-175 176-18d ,an 9 we examine effecte dth f preo s y size, prey popu- r smalfo lm yellom w perch). Large yellow perch lations consiste small0 9 f do , medium r largo , e averaged 120.smald an (±23.5g ) l yelloSD , 6g w fathead minnows experimenn I . - , wher2 tex e ew perch averaged 55.9 g (±13.9 g). Yellow perch amined the effects of prey abundance, fathead were introduced into the experimental tanks and minnow populations were composed of equal pro- held overnight without fathead minnows to stan- portion three th f e so size-classe comprised san d dardize hunger level allod san w acclimation. fish0 experimen, n 18 I 120. 60 r o , , wher3 t e ew We used three size-classes of prey that repre- examined the effects of prey population structure, sented natural groupings foun n locai d l fathead prey abundance was held constant at 90 fathead minnow populations, based on age and sex (Price minnows; 30 of these were small, and the remain- t ale . 1991). Small fathead minnow rangeL (T s , populatioe th f o r medius de nwa largd man e fish 43-55 mm; mean weight ± SD, 1.1 ± 0.3 g) cor- in different proportions 0 mediu0 3 3 ) d (1 : man responde sexuallo dt y immature age-1 - fishMe . mediu large0 mediu0 large4 0 5 ) 2 ) (2 ,d (3 ,m man dium fathead minnows (60-66 mm, 2.5 ± mediu 5 0.5 large 5 4) large 0 (4 d g1 m) r .an d o , Thi an s corresponded primaril o maturt y e femalet bu s proportion series mimicked seasonal changes in Downloaded by [University of Alberta] at 09:17 16 December 2011 could have bee largese nth t juvenile smallesd san t natural populations of fathead minnow, in which sexually mature males. Large fathead minnows large, age-2 individuals die after reproducing in (71-80 mm, 4.3 ± 0.5 g) corresponded to large late spring and age-1 fish grow through the sum- sexually mature male femalesd san alloo T .r wfo mer months (Pric . 1991)al t ee . easy recognition of size-classes, a small nick was cut from the upper lobe of the caudal fin of me- Statistical Analyses dium fathea dtriala minnowsf o ,firse y th da t n O . We used one-way analysis of variance coupled all prey were introduced simultaneousl tane th ko yt with Tukey-Kramer multiple-comparison tests to containing yellow perch. We never observed yel- analyze handling times and return rates, changes low perch attacking newly introduced fathead in weigh yellor fo t w resulte - perchth ex d f so an , minnow indicatioo n therd s san ewa unusuaf no l periment 1. We examined effects of prey size, pop- vulnerabilit predatioo yt t thina s time. During tri- ulation-level treatments e interactionth d an , f o s als, fathead minnows were fed daily with flaked these factor experimentn i stwo-way b 3 d yan s2 fish food. analysi variancef so . Data were logio-transformed 858 PASZKOWSKI AND TONN

Large Yellow Perch Small Yellow Perch

LOOO- 1.000-

— 500 - 500 - c CO

Small Medium Large Small Medium Large

1- 5 X 15

X 1- 0 i 1 . - 5

Small Medium Large Small Medium Large N 6 N ; 5 N -- 4 N 14 N 15 N - II Prev Size Prey Si/e FIGURE 1. —Mean handling times and return rates (biomass of prey consumed per unit handling time) for large (198-21 totam 0m l smald lengthan ) l (156-18TL , yello) 9mm w perch consuming small (43-5 TL)m 5m , medium (60-66 mm), and large (71-80 mm) fathead minnows. Error bars = 95% confidence intervals (calculated from logic-transformed data). Sample sizes (N} represen numbee tth individuaf ro l yellow perch observed; for each yellow perch usee mediae w , dth n value from multiple trials (the mea nindividuan a numbe r fo f trialD o rS l yello± s w perch handling a given size of prey was 4.1 ± 3.5). Means labeled with the same letter did not differ significantly from each other (Tukey-Kramer tests, P > 0.05).

to homogenize variance. In all cases, P < 0.05 fathead minnows offered significantly highe- re r was considered statistically significant. turns than large ones (Figure 1).

Results Experiment Effects1: Prey of Size Handling Times Returnand Rates The total number of and biomass of prey cap- We observed th large7 e yellow perch capture tured and consumed by large yellow perch did not Downloaded by [University of Alberta] at 09:17 16 December 2011 ingesd an fathea9 3 t d minnows (14 small,- 12me differ with prey siz experimenn ei (Figur1 t . e2) dium d 1 3e 1 ,an 5large th smald )an l yellow perch Correspondingly, mean weight change (± SD) of capture and ingest 184 fathead minnows (80 small, individuals within groups of large yellow perch 73 medium 3d 1 an , large) r individuaFo . l yellow did not differ across treatments (small prey of- perch, we calculated median handling times gan;0 mediumd2. ± fered6 large; 1. g ,- ,3 -0. ,2. 9± return rates for each prey size and we use. g) d 6 thes3. e± 8 0. median values for analysis. For both large and suggestes A retury db n rates, small yellow perch small yellow perch, handling time differed signif- proved more sensitive to prey size. Both the total icantly with size-class, increasing with prey size numbe f preo r y capturetotae th l d biomasdan s (Figure 1). consumed differed significantly across treatments In terms of return rate, however, the perfor- in experiment 1 (Figure 2). Most notably, small mances of large yellow perch did not vary signif- yellow perch ate significantly less biomass when icantly with prey size (Figure 1). Return rate dif- only large prey were available than unde othee rth r fered with prey size for small yellow perch: small treatments. Mean weight chang f individualo e s PISCIVOR YELLOYBY W PERCH 859

Large Yellow Perch Small Yellow Perch

I UJ JB 50 - 50 - E 25 25 -

Small Medium Large Small Medium Large

X 3 150- 75 c 22 X [3 100- 50 P , 0.05).

within group f smalso l yellow perch differed sig- As was observed for large yellow perch in this nificantly with prey size, and weight gain (mean experiment, prey size affected significantle th y ± SD) was greater when small fathead minnows number of fathead minnows eaten by small yellow were available than when large ones were offered perch (Figure 3). Consumption differed signifi- (small-prey treatment, 0.7 ± 1.4 g; medium, 0.3 cantly between each prey sizt onl eno termn yi s ± 1.2 g; large, -1. 5± 1.7g) . numbersf o t als termbu ,n o i biomassf so conn I . - e resultth tras r largo t fo ts e yellow perch, prey Experiment Effects2: Prey of Abundance

Downloaded by [University of Alberta] at 09:17 16 December 2011 abundance significantly affected predation pat- Total prey abundanc affect numbee no th td edi r tern smalf so l yellow perc termn hi pref so y num- f fatheao d minnows eate largy nb e yellow perch ber, and in terms of biomass, the effect was almost in experimen . Pre2 t y size, however d affecdi , t significant (P = 0.08). Total biomass ) (meaSD n± numbers eaten. Large yellow perch ingested sig- consumed by small yellow perch was 20.7 ±7.4 nificantly more small fathead minnows than me- g with 60 prey present and increased to 66.3 ± diu r largmo e fathead minnows, even thouge hth 19.4 g with 120 prey and 66.1 ± 24.4 g with 180 size-classes were equally available (Figure Th . e3) prey present. Prey abundance interacted signifi- interaction between sizabundancd ean t no s ewa cantly with prey size. Small yellow perch captured significant. Biomas f preso y eate largy nb e yellow more smal d mediuan l m fathead minnowst bu , perch in this experiment was not significantly af- t morno e larghigheo tw e e rprey th level t a , f o s fected by prey size, abundance, or their interac- prey abundance (Figure 3). Small yellow perch did tion. Large yellow perch did not display significant not, however, display significant differences in difference weighn si t change acros three sth e levels weight change across the three levels of prey abun- f preo y abundance. dance. 860 PASZKOWSKJ AND TONN

Large 30 . Yellow Perch I CO 20 . £ I ,OJ * I 60 120 180 N -5 N -5 N 5

Small 45 . C Yellow Perch

4—• 03 30 .

.£> I 15 J

60 120 180 N 8 N -8 N - 8 Prey Abundance FIGURE 3. — Mean numbers of smail, medium, and large fathead minnows eaten in experiment 2 by groups of six large yellow perch or six small yellow perch in 8 d under three prey abundances (60, 120, or 180 fish, divided equally amon size-classes)e gth . Sample sizes (jV) represen numbee th t groupf ro yellof so w perch tested. Error bars = 95% confidence intervals (calculated from logic-transformed data).

Experiment 3: Effects of Prey numbers differed significantly between each prey Population Structure size, and greater biomasses of small and medium numbee Th r (Figur biomasd an ) e4 f fatheaso d fathead minnows were eaten than large. Unlike minnows eaten by large yellow perch did not differ e patterth r largnfo e yellow perch, predatioy nb in response to changes in the overall structure of small yellow perch displaye da significan t inter- the prey population in experiment 3, but were af- action between prey sizpopulatiod ean n structure Downloaded by [University of Alberta] at 09:17 16 December 2011 fected significantly by prey size. The number of for both number biomasd san pref so y consumed. fathead minnows eaten differed significantly be- As medium fathead minnows became more com- tween each prey size greated ,an r biomas smalf so l within mo nprea y populatio largd nan e fathead and medium fathead minnows than of large prey minnows became rarer, consumption of medium was consumed. The interaction between prey size fathead minnows increased while the consump- and population structure was not significant. Large tio largf no e prey decreased (Figur Thus. e4) , total yellow perch did not display significant differences biomass consume smaly db l yellowl percal r hfo in weight change in response to changes in prey sizes combined remained relatively stable, aver- population structure. aging 58. (±31.6g ) wit mediu0 SD h3 , 6g m fat- As in the case of large yellow perch, prey pop- head minnows available, 41. (±12.8g ) wit3g h ulation structur t affec totae no th d t l edi numbe r , 52.40 3(±37. g 61.d an (±23.6wit ) g , 1g h50 7 biomasd an f fatheaso d minnows eate smaly nb l g) with 55 medium prey. Small yellow perch did yellow perc experimenn hi t prebu y, 3 tsizd eha not display significant differences in weight change significant effects (Figure 4). Consumption by acros seriee sth treatmentsf so . PISCIVOR YELLOY YB W PERCH 861

• Small Prey MediuO m Prcv Large Yellow Peirch • Large Prey 60 a1 40 j* 20 | ( < ( (> • i ' 1 ' ^ * o -1 30:30:30 30:40:20 30:50:10 30555 N - 5 N - 5 N - 5 N - 5 Small Yellow Perch 30 1 3 20 5 i 10 c 1 4i> +1 d) Z o • —— . —————— — . ——————4 — . ——————— . —— • — 30.30:30 30:40:20 30:50:10 30:55:5 8 N- 8 8 N N- N-7 Distribution of Prey Sizes (Small: Medium : Large) FIGURE 4.—Mean number f smallso , medium largd an , e fathead minnows eate experimenn ni groupy b 3 t f so six small yellow perch or six large yellow perch in 8 d. Prey populations consisted of 90 fish and had one of four size distributions (small: medium: large). Sample sizes (N) represent the number of groups of yellow perch tested. Erro confidencr bar95% s= e intervals (calculated from log 10-transformed data).

Discussion aged under a more complex, realistic prey regime choiceand s among prey sizes were possible, return Return rates based on handling times indicated rates were less effective in predicting feeding pat- that large yellow perch feedin n fatheago d min- terns. In these experiments, large as well as small nows should be insensitive to prey size, whereas yellow perch displayed a dichotomy: high con- small yellow perch should discriminate, specifi- sumptio f smalno l fathead minnows termn i , f so cally between smal d largan l e prey. Return rates numbers and biomass, and low consumption of proved to have some utility in predicting feeding large fathead minnows. Although neither sizf eo patterns of yellow perch in experiments 1-3, but yellow perc expectes hwa discriminato dt e inter- also showed appreciable shortcomings. mediate-sized fathead minnows from smalr o l clearese Th t agreement between predictiond san large, the yway n oftei d sn di tha t varied with treat- Downloaded by [University of Alberta] at 09:17 16 December 2011 results occurred in experiment 1, where only one ment and size of predator (Figures 3, 4). Concen- prey availables sizewa . Larger yellow perc- had tratio smaller-than-expecten no d prey, oftee th n juste numbee dth smallf ro , medium r largo , e fat- smallest available consistena s i , t attribut Per-f eo head minnows eate orden ni maintaio rt ncona - ca feeding on fishes (Tonn and Paszkowski 1986; stant intake of biomass and constant weight Ton t alne . 1991 ,f otheo 1992 d r an )predator y change. However, when only large fathead min- fishes feeding on vertebrates (e.g., Mauck and Co- nows were available, small yellow perch were un- ble 1971; Gillen et al. 1981; Hamilton and Powles abl attaio et n this "balance significantld ha d "an y 1983; Hoyle and Keast 1987; Schlosser 1988; Hart less growth. These differences between large and and Hamrin 1990; East and Magnan 1991; Ham- small yellow perch are consistent with the simple bright 1991). This is especially true when the pre- explanation tha largee th tbod e gapd rth yan e sizes dicted size of prey is based soley or partially on e predatorsth f o e greateth , r their handling effi- postcapture handling time. Similar trends have cienc largf yo e prey. been observe r somdfo e predatory invertebrates In experiments 2 and 3, where yellow perch for- (e.g., decapod crustaceans eating molluscs; Juanes 862 PASZKOWSK TOND AN JN

1992). Numerous explanations have been pro- in consumption of all three prey sizes, implying posed nont completele ,bu ear y satisfactory. Hoyle nonselective predation driven primarily by in- and Keasl (1987), for example, related differences creased rates of prey encounter and detection between observed and expected prey size for large- (Magnhagen 1985), wer t observedeno . Onle yth mouth bass Micropterus sahnoides relativee th o t - number of small and medium fathead minnows ly greater availability of small prey because of eaten increased, supportin proposae gth l that yel- population structure and seasonal patterns in na- low perch selectively pursue pre differenf yo t sizes. ture. However, this cannot adequately account for Yet even in the presence of elevated concentra- r resultou experimentn si , whic3 d han suggess2 t tions of smaller prey, which offered higher returns, mora e active rol piscivorer efo shapinn si g their small yellow perc e larghat e fathead minnown si diets (see below). 13 of 16 trials conducted at the two higher prey Escape behavior of prey fish also deserves con- abundances. Foraging experiments have demon- sideration e assumptioTh . n that e stageth f o s strated repeatedly that, although fish may concen- predatory process other than handling (e.g., pur- trat n particularlo e y profitable food types, they suin r subduingo g prey) represent negligibl- in e rarely restrict thei "beste r th die o t "t prey, even vestments in time or energy (Werner 1974) may if it is abundant enough to meet energetic de- be applicable to planktivores but not to piscivores. mands (e.g., Werne Hald ran l 1974). This pattern Fast-start performance (Webb 1978), maneuver- could be due to imperfect discrimination among ability (Moody et al. 1983), and schooling skills prey types by foraging fish, "deliberate" sampling and position (Pitcher and Parrish 1993) improve by foragers, or prey selection based on criteria not with increase sizd thucontribd y ean an sma e dag - measured by the investigator (Stephens and Krebs ut betteo et r escape capabilitie lowed san r capture 1986). probabilitie larger sfo r prey fish. Nonrandom cap- r studOu y offers several insights int nature oth e ture success could explain many patterns observed of piscivory by Perca spp. and its potential impact in our experiments, even if yellow perch attacked fathean o d minnow othed san r prey fishes. First, pre f differeno y t t sizeno t randoma sd di e W . even when prey are continually available, pisciv- quantify attack rates, but selective attack on pre percy yb relativela y s hi or y uncommon eventn i ; of sizes that are likely to be captured and handled all 135 trials, individual yellow perch capturen da * efficiently has been reported for other piscivores average of only one fathead minnow every 35 h. (Eas Magnad tan n 1991) representd ,an spossibla e infrequence Th successfuf yo l piscivory migh- ex t tacti f yelloco w perc wells ha . plai e scarcitnth f fisyo h remain e gutf th o s n i s Large yellow perch were apparently effective yellow perch of even larger size-classes collected enough as piscivores to be insensitive to our ma- in dietary studies (e.g., Keast 1977; Jansen and nipulation f theio s r food bas experimentn ei 2 s MacKay 1992). Small yellow perch, being partic- contrastn I . 3 d ,an feeding pattern smalf so l yellow ularly sensitive siz e sizth e f preth d eo o yean t perch were clearly affected by changes in prey structur f preeo y populations poorlo d y s ya ma , abundanc sizd eean distribution experimenn I . , 3 t piscivores in communities dominated by larger relative shiftth n i s e abundanc f differeneo t sizes prey fishes, even if they are ultimately capable of of prey affected the diet of small yellow perch, not capturing and handling these prey. In such envi- in terms of total number or biomass of prey, but ronments, small yellow perch might concentrate Downloaded by [University of Alberta] at 09:17 16 December 2011 by altering dietary composition. Medium prey were their foraging efforts on benthic invertebrates or substituted for large prey as the medium ones be- seasonally available age-0 fish (Ton . 1992)al t ne . came increasingly mor elarge commoth es a d nan Small yellow perch also appea less-effece b o rt - ones becam etrial6 mor1 f so ewher 0 rare1 n eI . tive piscivore t lowesa r overall prey abundances, large fathead minnows composed 11 r %leso f so which may help explain why the ontogenetic shift the population, they disappeared completely from a piscivorou o t s die s mori t e difficul r percfo t h the die f smalo t l yellow perch. Har Hamrid an t n species in less-productive lakes (Aim 1946; Deed- (1988) reported similar responses by northern pike ler 1951) r resultOu . s from experimen sugges2 t t Esox lucius capturing rudd Scardinius erythroph- that, in addition to a scarcity of benthic and pe- tha/mus from schools with different size struc- lagic invertebrates in such lakes (e.g., Rask 1983; tures. Persson 1987) densitiew lo , s offish precony yma - In experiment 2, small yellow perch foraged tribute to stunting of Perca. more successfully when total prey numbers were Unlike change pren si y abundance, changen si double tripledr do . However, equivalent increases prey population structure that simulated seasonal PISCIVOR YELLOY YB W PERCH 863

shifts in natural populations of fathead minnows lese b s likel supreso yt s reproductio recruitd nan - (shifts caused by the death of large individuals ment totally. Thus, the yellow perch and the fat- after reproduction; Pric t alee .t affec 1991no d t )di head minnow may have morphological and be- total consumptio yelloy nb w perch. Both largd ean havioral characters, undoubtedly evolved small predators were abl traco et progressive kth e independently under different selective regimes, increas f mediuo e m pred decreasan y f largeo e that permit them to live together in appropriate prey. In these manipulations the abundance (ab- habitats variee ligh n th I uniqu .d f do tan e distri- solute and relative) of small fathead minnows re- bution historied san temperatf so e freshwater fish mained constant. However n anothei , r experi- species during and following the Pleistocene, this men piscivorn o t yelloy yb w perch whicn i , e hth typ f "permissivo e e coexistence s probabli " ya population structur pree th yf e(centrao l mudmin- common process shaping fish assemblages. now Umbra limi) shifted from predominantly small (vulnerable) to predominantly large (invul- Acknowledgments nerable) individuals resulte th , s were different- to ; We thank C. Price, K. Price, P. Fok, A. Mc- tal predation, especially by small yellow perch, Pherson, and M. Olson for their help in conduct- declined (Ton Paszkowskd nan i 1986). Predation ing experiments. W. Jansen, C. Robinson, and W. levels by Perca spp., and the impact of that pre- Zyla helped collect fish . LeClai. FielL .K d an r dation on prey populations, should be even more assiste n preparini d e manuscriptgth . Thi- re s sensitive to population structures of large-bodied search was supported by a research grant from the prey, such as many species of European cyprinids Natural Science and Engineering Research Coun- that can achieve absolute refuges from piscivory cil of Canada to W.M.T. by perch through growth (Holopaine . 1991)al t ne . Perca sppe lesar .s specialize d havdan e rela- References tively smaller mouths than many piscivores (Keast 1985). Therefore wits ,a h other prey species (e.g., Aim . 1946G , . Reason occurrence th r sfo f stunteeo d fish populations with special reference to perch. In- central mudminnow, Tonn and Paszkowski 1986; stitute of Freshwater Research Drottningholm Re- yellow perch, Post and Evans 1989), the impact port 25:1-146. of yellow perch piscivory was strongest on small Appeiberg, M., E. Degerman, and P. Nyberg. 1989. fathead minnows corresponding in size to sexually Species composition and relative abundance of the immature fish naturen I . , predatio Percay nb y ma fish fauna in acidified and limed lakes in Sweden. actuall concentratee yb fisn do h even smaller than Pages 143-160 in J. W. S. Longhurst, editor. Acid deposition: sources, effect controlsd an s . Britis- hLi those use thin di s stud d thuhavy yan s sma it e brary, Technical Communication, London. largest effect through limiting recruitment (Tonn Deedler, C. L. 1951. A contribution to the knowledge etal. 1992). of the stunted growth of perch (Perca fluviatilis L.) Other piscivores, such as largemouth bass and in Holland. Hydrobiologia 3:357-378. northern pike, reach large body sizes, have rela- East, P., and P. Magnan. 1991. Some factors regulating tively larger gapes, display their highest return piscivory of brook trout, Salvelinusfontinalis, in the lakes of the Laurentian Shield. Canadian Journal of when handling large prey fishwild an ,l feed pref- Fisheries and Aquatic Sciences 48:1735-1743. erentially on larger fathead minnows (Hambrigh t. CarlineF . Gillen . R Stein A d . . L. A an R ,,, 1981. 1991). Fathead minnows from populations sym- Predatio y pellet-reareb n d tiger muskellungn o e Downloaded by [University of Alberta] at 09:17 16 December 2011 patric with large piscivores display well-devel- minnows and bluegills in experimental systems. oped predator recognition and antipredator re- Transaction Americae th f so n Fisheries Society 110: sponses (Mathi . 1993)al t t surveye sbu , s have 197-209. repeatedly shown that fathead minnow type sar - Hambright . 1991D . K , . Experimental analysi pref so y selection by largemouth bass: role of predator mouth ically absent from lakes where large predatore sar width and prey body depth. Transactions of the present (Harvey 1981; Tonn and Magnuson 1982; American Fisheries Society 120:501-508. Robinso Tond nan n 1989). Strong predation pres- Hamilton, J. G., and P. M. Powles. 1983. Fish pre- sure on large, sexually mature individuals by Esox dation and other distinctive features in the diet of or Micropterus spp. might interfere with the com- Nogies Creek, Ontario, largemouth bass, Microp- plex reproductive behavior of the fathead min- terus salmoides. Canadian Field-Naturalist 97:47- 56. now, particularly obligate parental care (Mc- Hart, P. J. B., and S. F. Hamrin. 1988. Pike as a se- Millan and Smith 1974), and ultimately drive lective predator. Effects of prey size, availability, populations to extinction. Predation by yellow cover and pike jaw dimensions. Oikos 51:220-226. perch, concentrated on smaller individuals, woul . HamrinF d . S Hartd . B.J . an . , P 1990, role f eTh o . 864 PASZKOWSK TOND JAN N

behaviou morphologd ran selectioe th n yi f preno y Paszkowski . HolopainenJ . . TonnI M d . . A.C ,an ,W , . by pike. Pages . Hughes235-25N . R n 4,i editor. 1989 experimentan A . l stud bodf yo y sizfood ean d Behavioural mechanism f fooso d selection. Spring- size relations of crucian carp, Carassius carassius. er-Verlag, Berlin. Environmental Biology of Fishes 24:275-286. Harvey, H. H. 1981. Fish communities of the Bruce Persson, L. 1987. The effect of resource availability Peninsula. Internationale Vereinigung fur theore- and distribution on size class interactions in perch. tisch angewandtd eun e Limnologie Verhandlungen Perca fluviatilis. Oecologia (Berlin) 73:170-177. 21:1222-1230. Persson, L.. S. Diehl, L. Johansson, G. Andersson, and Holopainen . Paszkowski. M. J.A I ,W ., . TonnC d an , . S. F. Hamrin. 1992. Trophic interactions in tem- 1991. Ecological response f cruciaso n carp popu- perate lake ecosystems tes:a foof o t d chain theory. lation predalioo st percy nmanipulateb a n hi d pond. American Naturalist 140:59-84. Internationale Vereinigung fur theoretisch- an d eun . ParrishK Pitcher . J d . .J. T an ,, 1993. Functionf o s gewandte Limnologie Verhandlungen 24:2412- shoaling behaviour in teleosts. Pages 363-440 in T. 2417. J. Pitcher, editor. Behaviour of teleosts fishes. Hoyle, J. A., and A. Keast. 1987. The effects of prey Chapman and Hall, London. morpholog sizd handlinn yan eo g timpiscia n ei - Popova, O. A. 1967. The predator-prey relationship vore largemoute th , h bass (Micropterus salmoides). among fish. Pages 359-376 in S. D. Gerking, editor. Canadian Journa Zoologf o l y 65:1972-1977. The biological basis of freshwater fish production. Hoyle . KeastA . d A.J , .an , 1988. Prey handling time Blackwell, Oxford. UK , in two piscivores, (Esox americanus vermiculatus) Post, J. R., and D. O. Evans. 1989. Experimental ev- (Micropterusd an salmoides). with contrasting mouth idenc size-dependenf eo t predation mortalit- ju n yi morphologies. Canadian Journal of Zoology 66:540- venile yellow perch. Canadian Journal of Zoology 542. 67:521-523. . MacKayC Jansen. W . d A.. W an , 1992. Foraginn gi Price, C. J., W. M. Tonn, and C. A. Paszkowski. 1991. yellow perch, Perca flavescens: biological and phys- Intraspecific patterns of resource use by fathead ical factors affecting diel periodicit feedingn yi , con- minnows in a small boreal lake. Canadian Journal sumption, and movement. Environmental Biology of Zoology 69:2109-2115. of Fishes 34:287-303. Rask, M. 1983. Differences in growth of perch (Perca Juanes, F. 1992. Why do decapod crustaceans prefer fluviatilis L.) in two small forest lakes. Hydrobiolo- small-sized molluscan prey? Marine Ecology Prog- gia 101:139-144. ress Series 87:239-249. Robinson . 1989K . L . . C ,Laborator y surviva f fouo l r Keast. A. 1977. Diet overlap and feeding relationships e presencth pre n i y f northereo n pike. Canadian between the year classes in the yellow perch (Perca Journal of Zoology 67:418^420. flavescens). Environmental Biology of Fishes 2:53- . TonnM Robinson. .W - . d 1989K.in L an e ,. C ,Th . 70. fluence of environmental factors and piscivory in Keast, A. 1979. Patterns of predation in generalist structuring fish assemblage f smalso l Alberta lakes. feeders. Pages 243-255 in H. Clepper, editor. Pred- Canadian Journal of Fisheries and Aquatic Sciences ator-prey system fisherien i s s management. Sport 46:81-89. Fishing Institute, Washington D.C. . Ryer1988H . C , . Pipefish foraging: effects offish size, Keast . 1985 .A piscivor e Th . e feeding fish e guilth -f do prey size and altered habitat complexity. Marine es in small freshwater ecosystems. Environmental Ecology Progress Series 48:37-45. Biology of Fishes 12:119-129. . SteinSavinoA . R . d . F.1989J , an , . Behavioura- in l Magnhagen, C. 1985. Random prey capture or active teractions between fish predator theid san r prey- :ef choice? An experimental study on prey size selec- fect f plano s t density. Animal Behaviour 37:311- tion in three marine fish species. Oikos 45:206-216. 321. Mathis, A., D. P. Olivers, and R. J. F. Smith. 1993. Schlosser, I. J. 1988. Predation risk and habitat selec- Downloaded by [University of Alberta] at 09:17 16 December 2011 Population difference responsen i s fatheaf so d min- tion by two size classes of a stream cyprinid: ex- nows (Pimephales promelas) visuao t chemicad lan l perimental test of a hypothesis. Oikos 52:36-40. stimuli from minnow predators. Etholog y. Crossman J 93:31. E d -an , . ScottB. 1973. W , . Freshwater 40. fishe Canadaf so . Fisheries Research Boar Canf do - Mauck, W. L., and D. W. Coble. 1971. Vulnerability ada Bulletin 184. of some fishe northero t s n pike (Esox lucius) pre- Stein, R. A., C. G. Goodman, and E. A. Marschall. 1984. dation. Journal of the Fisheries Research Board of Using tim energetid ean c measure cosf so esti n i t - Canada 28:957-969. mating prey valu r fisfo eh predators. Ecolog: y65 McMillan . SmithF . J . .R , 1974d V.an , . Agonistid can 702-715. reproductive behavior fatheae th f so d minnow (Pi- Stephens, D. W., and J. R. Krebs. 1986. Foraging the- mephales promelas Rafinesque). Zeitschrift fur ory. Princeton University Press, Princetonw Ne , Tierpsychologie 34:25-58. Jersey. Moody, R. C., J. M. Helland, and R. A. Stein. 1983. . MagnusonJ Tonn. J . M.d W ,an , . 1982. Patternn si Escape tactics use bluegilly db d fatheasan d min- the species compositio richnesd nan s offish assem- now avoio st d predatio tigey nb r muskellunge- En . blage n northeri s n Wisconsin lakes. Ecolog: 63 y vironmental Biolog f Fisheyo s 8:61-65. 1149-1166. PISCIVOR YELLOY YB W PERCH 865

Tonn . MagnusonM.. J . ,W ,J Rask. Toivenen. J ,M d ,an . 1992. Piscivor d recruitmentyan : mechanisms 1990. Intercontinental compariso small-lakf no e fish structuring prey populations in small lakes. Ecology assemblages: the balance between loca regionad lan l 73:951-958. processes. American Naturalist 136:345-375. Unger, L. M. 1983. Nest defense by deceit in the fat- Tonn, W. M., and C. A. Paszkowski. 1986. Size-lim- head minnow, Pimephales promelas. Behavioral ited predation, winterkill organizatioe th d an , f no Ecolog Sociobiologd yan y 13:125-130. Umbra-Perca fish assemblages. Canadian Journal Wahl, D. H., and R. A. Stein. 1988. Selective preda- of Fisherie Aquatid san c Sciences 43:194-202. tion by three esocids: the role of prey behavior and . PaszkowskiA . TonnC . M.d W ,an , . 1987. Habitat morphology. Transactions of the American Fish- use of the central mudminnow (Umbra limi) and eries Society 117:142-151. yellow perch (Perca flavescens) in Umbra-Perca as- Webb. 1978W . P , . Fast-start performanc bodd ean y semblages rolee :th competitionf so , predationd ,an for sevemn i n specie teleosf so t fish. Journa- Ex f lo the abiotic environment. Canadian Journal of Zo- perimental Biology 74:211-226. ology 65:862-870. Werner, E. E. 1974. The fish size, prey size, handling Tonn . Paszkowski A . M. . . HolopainenW ,J C , . I d an , . relation in several sunfishes and some implications. 1989. Response f cruciao s n carp populationo t s Journal of the Fisheries Research Board of Canada differential predation pressure in a manipulated 31:1531-1536. pond. Canadian Journa Zoologf o l y 67:2841-2849. Werner . HallD d . E.. E ,an , 1974. Optimal foraging Tonn, W. M., C. A. Paszkowski, and I. J. Holopainen. size th e d selectioan bluegile pref th n o y yb l sunfish 1991. Selective pisci vor perchy yb : effect predf so - (Lepomis macrochirus). Ecology 55:1216-1232. ator size, prey size, and prey species. Internationale Vereinigun theoretischr gfu e und angewandte Lim- Received Januar , 199y22 3 nologie Verhandlungen 24:2406-2411. Accepted April 4, 1994 Tonn, W. M., C. A. Paszkowski, and I. J. Holopainen. Downloaded by [University of Alberta] at 09:17 16 December 2011