U

Predator-prey relations in fish abattle of intellects

By Johanneke Oosten Supervized by Charlotte Hemelrijk 22-09-2006, Theoretical Biology, RuG

D 931 r

Index

Introduction 2 How shoaling protects fish from, or exposes them to predatIon 3 The many eyes theory Shoal aggregation 4 Shoal-segregation 5

Size-segregation 11 Inspection behaviour 12 Schooling and evasion patterns 15

How predators deal with shoaling prey 17 Predator preferences 18 Hunting co-operatively 20 Ambushing prey 21

How predator-prey interactions are modified by learning 23 Alarm cues and learning by personal experience 24 Social learning 26 Ontogeny 27 Learned or genetic 27 Learning by predators 27 Conclusions 28 Appendix: fish names and families 30 References 32 Introduction Until the 1970's research on predator-prey inter- ics in less predator-dense areas. So it seems protec- actions in fish was rare. But in the last few decades, tion from predation is a major advantage of shoaling interest in the topic has grown explosively. By now (Seghers 1974: guppies; Magurran 1986: minnows). there are thousands of articles on the topic. One as- In this review I will focus on how this protection

pect that has received special attention is learning. by shoaling arises.Iwill show that shoaling does Research makes it increasingly clear that fish are not not always protect, but may expose their members the anti-social, memory-impaired blobs they were to predation instead. In addition Iwill discuss vari- once thought to be, but that they are intelligent crea- ous techniques employed by predators to deal with tures that form social relationships with their shoal shoaling prey. Finally I will show how these predator- mates and perform sophisticated techniques to out- prey interactions are shaped by learning. wit their predators and / or prey (Laland et al 2003; In the course of this review it will become clear Bshary et al 2002). that the theories commonly used to explain the Many fish species live in groups or shoals. Shoals anti-predator benefits of shoaling are difficult to tt al :2004) found that banded killifish provide their members with advantages, such as a prove with empirical data. In their attempts to do crease their shoal size when exposed to choice of breeding partner, as well as disadvantages, so, researchers use experimental set-ups, so tightly Iarrn cues and decrease it when exposed such as food competition (Krause & Ruxton 2002). controlled, that their naturalness becomes doubtful. to food odours. When both were present When shoaling fish detect a predator they aggre- Also, the number of species used in research is lim- intermediate shoals were formed. gate, forming a more compact shoal (e.g. Templeton ited, casting doubt on the universality of our current & Shriner 2005: guppies; Ferrari et al 2005: fathead knowledge. minnows; Brown et al 2004: glowlight tetras; Magur- The text is accompanied by examples from vari- ran & Pitcher 1987: European minnows). Also, several ous studies. These either illustrate the issues dis- studies find that fish living in predator-dense areas cussed or add an extra dimension to them. spend more time aggregated than their conspecif- e

How shoaling protects fish from, or exposesthem to predation

In this chapter I will describe how shoaling fish may respond when they detect a predator. I will discuss how these responses may protect them from or expose them to predation. Then I will show how shoal- ing may aid fish in evading a predator's attack.

7.

fathead minnow

banded killifish

threespine stickleback

northern redbelly dace Image 1: a shoal of mackerel has detected approaching skipjack and has aggregated into a tight ball (from BBC's The Blue Planet)

Shoal aggregation When a fish shoal detects a predator, its members additional benefit of having many eyes is that part often aggregate (e.g. Templeton & Shriner 2005: gup- of the shoal can keep watch while the rest forages, pies; Ferrari et al 2005: fathead minnows) (image 1). rests, etc.. In support of the many-eyes-theory van- Possibly aggregation provides protection from pre- ous studies find that, compared to individuals, shoals 4 dation. I will discuss four theories that explain how detect approaching predators sooner (Godin et al this protection may arise: the-many-eyes theory, the 1988: guppies; Magurran et al 1985: minnows; Tre- selfish herd theory; the dilution effect, the confusion effect. herne & Foster 1980: marine insect, but see Godin & Morgan 1985: banded killifish). Morgan &Godin (1985) found that information trans- Theory 1: the many-eyes-theory Prey will only benefit by their shoal mates' vigi- mission in shoals of banded killifish was independ- According to the many-eyes-theory, compared to an lance if information is transmitted fastacross the ent of shoal size, i.e. independent of the number of individual, a shoal will detect a predator sooner and shoal (Krause & Ruxton 2002). This seems to be the individuals in a shoal, and about twice as high as ap- keep better track of its movements, because a shoal case (Morgan & Godin 1985; Webb 1982). Webb proaching predator speed. with many fish has many eyes watching in many di- (1982) showed that the delay between escape stim- rections. This can save lives: predators' capture suc- ulus, such as a predator, and escape responsewas cess increases when prey are unaware of their pres- smaller for shoals than for individuals. This fast infor- ence (Turesson & Bränmark 2004, Webb 1982: pike; mation transmission was dubbed the Trafalgar effect Krause et al 1998a: rock bass). A fish that is aware of (Treherne & Foster 1981), in reference to theuse of danger can anticipate the moment of a predator's flag signals in Trafalgar's fleet (Krause & Ruxton 2002). strike. The exact tell-tale signs are unclear, but may Possibly the information Is transferred faster if the involve an unnatural stillness in the predator's body shoal mates are doser together. This may beone rea- (Magurran & Pitcher 1987: pike hunting minnows). An son for fish to aggregate on detection of a predator. 1

Foraging fish are less alert Theory 1: the selfish herd theory Foraging reduces vigilance: Godin & Smith Hamilton (1971) suggested that shoal aggrega- (1988) showed that foraging guppies were more tion occurs as a side-effect of fish seeking protection vulnerable to predation than non-foraging conspe- behind each other's backs. He postulated the selfish cifics. Vertically foraging fish are even less respon- herd theory: fish surround themselves by as many sive to their environment than horizontally foraging neighbours as possible so that a predator (that at- ones (Krause & Godin 1996: guppies). Krause & Go- tacks the first prey it encounters) attacks a shoal mate din (1996) suggest that vertically foraging fish may first (image 2). have: Several studies have tried to find empirical sup- - a smaller visual horizon. port for the selfish herd theory by determining if - a greater probability of visual obstacles (such as mortality rate by predation is higher in the shoal's vegetation) periphery than in its centre (table 1). Most studies - an increased difficulty of performing a quick find that the periphery is indeed the most danger- escape. ous area. Only one study found the opposite (Parr- When a predator is detected fish may change ish 1989). She describes how black seabass split up their foraging posture to better observe the preda- shoals of 25 Atlantic silversides by ramming into the Image 2: according to the selfish herd theory fish try tor (Godin 1986: banded killifish; Foam et al 2004: centre. Central fish suddenly found themselves In the to minimize the number of search trajectories lead- convict cichlids; Krause & Godin 1996: guppies). Of- periphery. The author notes, however, that had the ing from the predator to oneself. By aggregating, ten they stop foraging entirely (e.g. Abrahams 1995: shoal been larger, the predator may have attacked many trajectories will lead to shoal mates instead. brook sticklebacks; Godin & Smith 1988: guppies; a more peripheral part, since its main goal probably Godin 1986: banded killifish). Vigilance reduction was to split off a piece of the shoal. Still, predator tac- may also cause starved guppies and guppies enjoy- tics like these make central positions less attractive: ing high food-densities to suffer higher predation the optimal shoal position may in part depend on rates (Godin & Smith 1988: guppies; Godin 1986: the predator's hunting strategy (Parrish 1989). banded killifish). Krause & Ruxton (2002) wandered why fish do Problems proving the selfish herd theory not take advantage of their shoal mates' vigilance, Stankowich (2003) shows that the studies In table while lazing about themselves. They suggest that 1 are difficult to compare, because they differ in their the vigilance of others is not as good as being vigi- definitions of 'periphery' and 'centre He also argues lant yourself. Predators may prefer to attack un- that the selfish herd principle is not required to ex- wary prey (Krause & Godin 1996). Also, transfer of plain their results. He suggests that the increased pre- information across the shoal may be fastest when dation rate in the shoal's periphery could be caused all fish are alert. Inattentive fish may cause small but by a predator's active selection: it may prefer specific fatal delays. phenotypes that commonly reside in the periphery, e.g. smaller, younger or hungrier fish. Table 1: mortality by predation in shoals' centre and periphery (excerpt from Stankowich 2003)

Article Preyspecies Predation rare penphery>

predation rote centre?

Krause & Tegeder 1994 Three-spine sticklebacks Periphery>Centre

Krause 1993 Minnows Periphery>Centre

Barber & Huntingford 1996 Minnows Periphery>Centre

Parrish et al 1989 Flat-iron herring Centre = periphery

Parrish 1989 Atlantic silversides Periphery

Table 2: predators preferences for shoals

Article Predatorspecies Prey species Preference for o2tocking largeitt shoals?

Morgan &Godin 1985 White perch Banded killifish Preferred shoal over solitary fish

Krause & Godin 1995 Blue acara cichlids Trinidadian guppies Preferred larger over smaHer shoal

Botham et al 2005 Pike cichlids. blue acara cichlids Trinidadian guppies Preferred shoal over solitary fish

Botham & Krause 2005a Wolf fish Trinidadian guppies Preferred shoal over solitary fish

Image 3: according to the d lution effect per capita preda- Image 4: a herrng schol is vanquished by yellowtail rockfish and birds (from BBC'sThe Blue planet). tion risk decreases with shoal size

25%

25% 25% Theoi'y2: the dilution effect Many studies find that shoaling fish enjoy a lower met at least for banded killifish. However, there are predation risk, than solitary fish (Krause & Godin 1995: cases in which predators reduce shoals of millions guppies). This benefit is often explained with the di- of fish to nothing (The Blue Planet: herring attacked lution effect. This effect indicates that chances to be by aerial and aquatic predators, image 4). As for the legeder & Krause (1995) observed how, at a criti- attacked decrease with shoal size (Krause & Ruxton second condition: shoals may increase as well as de- cal distance from the larger of two shoals, half the 2002). E.g. if a predator is looking for prey and runs crease detection by predators. Decrease, because a threespine sticklebacks chose the smaller one. into a solitary individual, this fish has a 100% chance randomly swimming predator has to find a localized of being attacked. If the predator runs into a shoal clump in a large water column, increase, because a Northern red-belly dace chose the smaller of two of 100 fish, and still only requires one prey, then the larger group is more conspicuous (Krause & Ruxton shoals shoal if the larger one was too close to a chance of one member being attacked is reduced to 2002). For decreased detection there is little evidence. predator. This preference changed when shoal size 1% (image 3). The existence of the dilution effect is Many predators live in the vicinity of their prey: they difference became large enough (Ashley et al 1993). supported by findings that fish prefer to join the larg- know exactly where to find food at all times and est shoal, i.e. the shoal having the most members, in detection is independent of shoal size (Magurran & Blue acara cichlids were not attracted by shoal the vicinity (Krause 1998b: creek chubs; Tegeder & Pitcher 1987: pike hunting minnows; Seghers 1974: size, but by shoal conspicuousness: they preferred Krause 1995: threespine sticklebacks; Hager & HeIf- various predators hunting guppies). On the other to attack the larger of two prey shoals, but when man 1991: fathead minnows). hand there is much evidence for increased detection it was immersed in cold water, and as a result less There are several conditions that have to be met (Monadjem et al 1996, Remsen 1991: predatory birds; conspicuous than the smaller shoal, their preference for the dilution effect to work (Krause & Ruxton 2002): Magurran 1 990a: freshwater prawns). switched (Krause & Godin 1995). - a predator must not, once it has found the shoal, Predators also seem to be attracted towards prey eat all its members. aggregations (table 2). They may be drawn towards - a shoal of N individuals must not be N times the increased activity and movement of large prey more detectable or attractive than a solitary fish. numbers (Krause & Godin 1995). Predators can sense Godin (1986) showed that the first condition was activity with their lateral line system.

The lateral line system gives fish an idea of hydro- dynamic pressures in their direct vicinity. It has a reach of a few body lengths only (Pitcher 1993). For some predators this sytem is crucial for directing the final strike (Pohlman et al 2004: catfish; Reist 1983: pike). To sabotage this strike an individual prey may freeze': becoming completely motionless and often sinking to the bottom, for up to tens of seconds (Pollock 2001; Smith 1997a). Freezing is commonly observed (e.g. Pollock 2003: fathead minnows; Brown and Smith 1998: rainbow trout, Savino 1989: bluegills; Lehtiniemi 2005: threespine sticklebacks). Image 5: on seeing an enormous amount of similar fish predators become confused and unable to single out one prey (from BBC's The Blue Planet)

Table 3: many predators have a reduced capture success when attacking prey in shoals

Article Prey species Predator species Reduced capture success

with larger shoals?

Nell & Cu lien 1974 Various prey (max. shoal 20) Various predators Yes

Morgan & Godin 1985 Banded killifish (max. shoal 20) White perch No

Landeau &Terborgh 1986 Silvery minnows (max. shoal 15) Largemouth bass Yes

Parrish 1993 Flat-iron herring (>> 100.000) green jack and black Yes skipjack I

Krause & Godin 1995 Trinidadian guppies (max. shoal 16) Blue acara cichllds Yes

Krause et al 1998a Creek chub (max. shoal 13) Rock bass No

Turesson & Bränmark 2004 Roach (max. shoal 16) Perch Yes

Turesson & Bränmark 2004 Roach (max. shoal 16) Pikeperch No

Turesson & BrOnmark 2004 Roach (max. shoal 16) Pike No

Image 6: Krakauer's network (1995) simulates how a vertebrate may proc- ess spatial information. One could interpret input' as the image formed on Input the hunter's retina; transport as the neurons connecting eyes to brain, and I 'output' as the spatial image formed in the brain. Cube 1 receives informa- tion froi, the left part of the visual field, cube 2 information from the left- center partf the visual field, etc Using art algorithm, the network 'tries' to Transport place the information receved through cube 1 into cube '1; through cube 2 into cube 2, . 'tc The chance of success depends on the signal strength. Signal strengtn depends on contrast with the environment. Also, signal Output strength increases if the point/prey is focused on/targeted. Theory 3: the confusion effect Oddity effect Aggregated prey are eaten less, because they Krakauer' (1994) interpreted an odd-looking prey are surrounded by others. Being aggregated may in a shoal as a localized increase of signal intensity. provide them with second benefit: according to the As a result, his simulation predicted that such prey confusion effect predators suffer a reduced capture was easier to target. Several studies found empiri- Krause & Godin (1994) offered banded killifish, a success when attacking shoals instead of solitary fish. cal support for this prediction: they found that 'odd' i:hoice of shoals. The killifish preferred to join a shoal On seeing an enormous amount of similar fish preda- individuals in a shoal suffered higher predation risks composed of heterospecffic, but similar sized indi- tors become confused and unable to single out one (table 4). And others showed that fish preferred to be viduals to either a larger or a conspecific shoal. prey (images). In support of this theory, many studies in shoals composed of similar individuals (Krause & indeed found a reduced capture success for preda- Godin 1994: banded killifish: Wolf 1985: parrotfish tors attacking larger shoals (table 3). Some benefits of and surgeonfish). Notwithstanding the results above, Wolf (1985) described how parrotfish and surgeon- the confusion effect overlap with those from the dilu- 'odd' individuals may still be safer in shoals than fish, after detecting a predator, left their shoal if tion effect: both are supposed to lead to a decreased alone (Landeau &Terborgh 1986: silvery minnows). their conspecffics constituted a minority. per capita predation risk. Both should induce prey to join the largest shoal available. It is not easy to deter- Skittering/dashing mine which effect, dilution or confusion, causes the Fish may increase the confusion effect by skitter- Magurran & Pitcher (1987) described how minnows, greater part of these shared side-effects. ing or dashing (Pitcher 1993). Skittering fish make rap- after detecting pike, inspected it, judged it danger- Krakauer (1995) designed a neural network sim- id, jerky movements of several body lengths and, in ous and then skittered. ulation that could predict the confusion effect. He this way, make It even more difficult for a predator to assumed that predators have a limited ability to lock-on (Templeton & Shriner 2005: threespine stick- process visual information. This could be caused, for lebaclç Dupuch et al 2004: northern redbelly dace; example, by a limited amount of neurons connect- Poltock et al 2003: fathead minnows; Pitcher 1993 pp ing the eyes to the brain (image 6). Increased prey 389-391). Dupuch et al (2004) suggested that skitter- aggregations lead to larger amounts of similar visual ing may also serve to warn shoal mates of a detected information. More and more of this have to pass the predator. same neurons, and visual resolution decreases.

Table 4: increased predation rate of 'odd individuals

Article Predator Prey Increased predation?

Theodorakis 1989 Largemouth bass Fathead minnows Yes Bluntnose minnows Stoneroller minnows

Landeau &Terborgh 1986 Largemouth bass Silvery minnows Yes

Ohguchi 1978 Threespine stickle- Water fleas Yes backs Problems proving the confusion effect Not all studies in table 3 support the existence find evidence for the confusion effect. They often of the confusion effect. Both Neil & Cullen (1974) and use small shoals of prey (table 3). Turesson & Bronmark Turesson & Branmark (2004) studied pike. Neil & Cul- (2004) offered pike shoals composed of up to 16 len's found that pike suffers a reduced capture suc- prey, and found no evidence for the confusion effect. Milinski & Heller (1978) suggested that the atten- cess when attacking larger shoals, Turesson & Brön- Magurran & Pitcher (1987), on the other hand, offered tion-consuming aspect of foraging drives three- mark did not. Turesson & Brönmark attributed this pike shoals composed of up to 50 prey. The latter did spine sticklebacks to switch to a lower-density food inconsistency to an alternative definition of 'capture not look at pike's capture success rate, but they did source when they detect a predator: the stickle- success'. Neil & Cullen (1974) measured capture suc- describe how pike sought to disrupt the schools and backs do not want to be confused by their own cess as number of captures / number of prey encoun- pick off stragglers, a hunting strategy that may well food, and deal with a predator at the same time tered. Since attacking prey in shoals automatically serve to overcome the confusion effect (Magurran & leads to a very high encounter rate, this does not Pitcher 1987). Many researchers probably use such seem like a fair measure. Turesson & Bränmark meas- small shoals, because of laboratory constraints such Milinksi (1990 asked human test-subjects to pin- ured number of prey captured / number of prey at- as tank size. But if the shoal is too small, a specialized, point dots in three densities of two-dimensional tacked. But perhaps the confusion effect makes pike efficient predator such as pike may not be confused swarms. Only t the highest density (8 dots/cm) did hesitate to attack in the first place. Once it manages at all. Shoal sizes in the wild are usually quite a bit the test-subjects oerformance become hampered. to lock on it may get the same success rate. larger than 16, especially after fish have aggregated There is another problem with studies trying to in response to a predator (tableS).

Table 5: shoal sizes of various species observed in the wild

Authors Species Shoal size

Misund 1991 Atlantic herring 1O

Morgan &Godin 1985 Banded Killifish 2-200

Blaxter & Hunter 1982 Clupeoids (including herring, —100 to .106 sardines and anchovies)

Dupuch & Magnan personal observations Northern recibelly dace 50-300

Croft et aI 2003 Trinidad guppies Several-SO U

Size-segregation Pressure exerted by predators may drive the size-segregation in response to predation. One does commonly observed size-segregation in shoals (Wolf not. Many shoals are size-segregated when no pred- 1985; Pitcher et al 1985; Allan & Pitcher 1986; Parrish ator is present. This suggests that there are other 1989; Theodorakis 1989; Svensson et al 2002; Wong reasons for size-segregation as well. Food-competi- et al 2005). Both the oddity effect and the selfish herd tion is often mentioned: small fish do not want to be theory can cause it. The oddity effect predicts that in the same shoal as large fish because the latter are 2 one is safer from predation among similar individu- better competitors (e.g. Krause 1994; Theodorakis als. Thus fish may align themselves with similar-sized 1989). Hemelrijk& Kunz (2004) suggest an interesting individuals to be as inconspicuous as possible. The alternative: size-segregation may arise by self-organi- 0 : o — - - selfish herd theory suggests that the shoals centre is zation. Their model predicts that variations in repul- -I .1 - - the safest place. In a struggle for the best shoal posi- sion area sizes (larger fish have larger repulsion areas 2 .2 tions larger fish may beat the smaller ones, resulting than smaller fish) lead to a shoal structure in which in large fish in the centre and small fish in the periph- larger fish reside in the periphery, and smaller ones ery. Such a shoal organization has been observed in the centre. Such a shoal structure has only been Image /left: differen:es in repulsion areas between (Krause 1994; Theodorakis 1989). If size-segregation observed with water insects (Romey 1997; Sih 1980). large and small fish drive large fish to the periphery indeed originates from anti-predator behaviour, one But if fish are given the additional tendency to avoid and small fish to the centre. Right: if small fish try to would expect it to increase when the shoal feels larger shoal mates more than smaller ones, the struc- avoid large fish, this distribution is reversed (from threatened. Only a few studies have tried to assess ture is reversed: larger fish appear in the centre, and Hemelrijk & Kunz 1994) this (table 6). Three out of four studies show increased smaller ones in the periphery (image 7).

Table 6: increased size-segregation in response to predation

Study Fish Threat Size-segregation?

Pitcher et al (1986b) European minnows Pike Yes

Theodorakis (1989) Bluntnose minnows Largemouth bass Yes

Theodorakis (1989) Fathead minnows Largemouth bass No

Krause (1994) Creek chub Alarm cues Yes '-'I

Inspection behaviour Inspection to gather information Iniage 8.insp€ haviour While the major part of a shoal aggregates in re- Inspection may also serve to gather information sponse to a predator, some shoal members may ap- about a threat. Many fish live in close association with proach it (E.g. Brown & Schwarzbauer 2001: glowlight their predators (Pitcher 1980: roach; Seghers 1974: tetras; Dugatkin & Godin 1992b: guppies; Magurran guppies). Also, anti-predator responses are costly: fish & Pitcher 1987: minnows; Milinski 1987: threespine have to stop foraging and mating (Lima & Dill 1990; Dill sticklebacks) (image 8). This so-called inspection be- 1987). Therefore It may be inefficient to flee as soon haviour is described as a tentative and salutatory ap- as an odd shape is detected (Milinski 1993), and more proach towards and withdrawal from the predator, rewarding to find out what the shape is, if it eats one's either alone or in groups (Dugatkin & Godin 1992a; own species, and if it's hungry (Licht 1989; Magurran Pitcher 1992). I will first discuss why fish inspect. Then & Pitcher 1987). Some aspects of a threat can be deter- I will show how prey modify their inspection behav- mined from greater distances by olfactory cues. E.g. a iour, based on the knowledge they gather. predator's diet can be deduced from chemicals emit- ted by the predator (Brown & Schwarzbauer 2001; Inspection to deter predators Smith & BeIk 2001). Fathead minnows can even deter- Why would fish leave the safety of their shoal to mine a predator's size by their odour alone (Kusch et approach a potential threat? Currently, an important al 2004). But other, important information, such as a reason for inspection is thought to be predator deter- predator's amount of hunger can only be determined rence or mobbing (George 1960). The idea that inspec- visually and one must take a doser look (Smith & Belk tion is a form of predator deterrence is supported by 2001). Possibly, gathered information can be passed There is some evidence that inspectors several studies showing that inspectors are less likely on to the shoal mates (Brown 2003, Brown & Godin have larger chances of getting a mate to be attacked than non-inspectors (Godin & Davis 1999: glowlight tetras; Pitcher et al 1986: minnows). (Godin & Dugatkin 1996: guppies). 1995: guppies; Magurran 1990a: minnows; Brown et Information gathering was originally thought to al 1999: glowlight tetras). be the most important reason for Inspection behav- Some species' inspectors exert physical displays iour (hence the name inspection). To explain its evo- lroAn & Godin (1999) observed how increased hn- in front of their predators: glowlight tetra inspectors lution Milinksi (1987) designed his influential tit-for- fltc king by inspecting glowlight tetras resulted in flick their fins (Brown et al 1999; Motta 1983), and tat model. He assumed that inspectors were facing increased fin-flicking, freezing and aggregating by bluespotted goby inspectors increase their head- a 'prisoner's dilemma'. He described a scenario in shoal mates. bobbing (Smith 1989). Such displays may discourage which two fish inspect together. Either one may stop a predator even further (Motta 1983). Magurran & co-operating (or defect) at any given moment. Pitcher 1993: described how larger inspector Pitcher (1987) showed that minnows in larger shoals - if both fish defect, neither gets information. groups dared approach a predator more closely. inspect in larger groups. Possibly larger inspection - if one inspects and one defect, the defector gets groups are more unnerving to the predator. In con- information without risk. Dominey (1983) describes hov .lcnurl nesting trast to birds, prey fish rarely chase their predators - if both inspect, both get information and share bluegills drive a snapping turtle away troin the nest- away (Dominey 1983) (table 7). the risk. ng grounds Milinski (1987) suggested that fish resolve the dl- lemma, using a turn-based tit-for-tat strategy. ous for the inspectors. The reduced mortality rate of - During the first inspection visit both fish co-operate. inspectors suggests otherwise. - During a subsequent visit, one fish or both fish may defect. Modification of inspection behaviour by predators - If one fish co-operates and the other defects, the Inspectors assess risk from a distance and modify co-operating one may forgive, and co-operate their inspection behaviour based on the knowledge again during the next visit, or it may retaliate and they can gather. If they detect that the predator eats defect as well. their kind they will take longer to initiate inspection - If one fish defects twice in a row, the other fish will (Brown & Godin 1999, Brown & Dreier: glowlight tet- defect during the next visit. ras), and maintain greater distances from the predator For Milinski's model to work, fish must be capable of (Smith & Belk 2001: mosquitofish; Brown & Godin 1999: book-keeping: to remember previous behaviour of shoal glowlight tetras). They will also show greater avoid- Dugatkin & Alfieri (1991) showed that guppies mates and to base their decisions on this knowledge. ance of the area around the predator's head (Brown retaliate against defectors: during a subsequent Various studies confirm this (Dugatkin & Alfieri 1991: & Dreier 2002, Brown & Schwarzbauer 2001: glowlight inspection visit the guppy that had previously guppies; Milinski et al 1990: threespine sticklebacks). tetras; Brown et al 2001a: finescale dace). In this cone- co-operated now defects. Also, the distance from However, evidence against Milinski's model is pil- shaped area, the attack cone, the predator is most the predator at which it defects is similar to the ing up. One of Milinski's major arguments for tit-for-tat likely to attack (Magurran & Seghers 1990a; Dugatkin distance at which its partner had defected during was that sticklebacks, inspecting with a co-operating & Godin 1992b; Pitcher 1992). But even in risky circum- the previous visit. conspecific approached a predator more closely than stances about a third of the inspections is aimed at the sticklebacks inspecting with a defecting conspecific. head (Brown & Dreier 2002: glowlight tetras; Brown et However, Stephenson et al (1997) could reproduce this al 2001a: finescale dace). Brown & Schwa rzbauer (2001) behaviour by assuming that two randomly swimming suggest that predator deterrence may only work if the fish keep a close eye on the predator, while staying predator is approached head-on. Also important in- near each other at the same time. Also, the compari- formation, like posture indicating satiation, may only son of inspection behaviour with a prisoner's dilemma be acquired at the head area. is based on the assumption that inspecting is danger-

Table 7: an overview of sightings of mobbing fish (excerpt from Motta 1983)

Article Pley Predator

Dominey (1983) Colonial nesting bluegills Snapping turtle

EibI-Eibesfeldt (1962) Redbelly yellowtail fusilier Moray eel

Maksimov (1970) Various tropical fishes Various sharks

Fricke (1973) Damselfish Barracuda; triggerfish; octopus image 9: fish fleeing from a predator (from BBC's The Blue Planet)

Table 8: commonly observed evasion patterns in fish schools

Manoeuvre Article Species

Fountain (evading fish turn Magurran & Pitcher (1987) Minnows

around pursuing predator and Nursall (1973) Spottail shiners rejoin the school behind it) Pitcher & Wyche (1983) Sand-eels

Nettestad & Axelsen (1999) Herring

Vacuole': fish surround the Pitcher & Wyche (1983) Sand-eels

predator Nottestad & Axelsen (1999) Herring

Split: school splits up into two or Pitcher & Wyche (1983) Sand-eels

more subschools Nottestad & Axelsen (1999) Herring

Magurran & Pitcher 1987 Minnows

Herd: school flees in front of Pitcher & Wyche (1983) Sand-eels

predator Nøttestad & Axelsen (1999) Herring

Flash expansion: school 'cx- Pitcher & Wyche (1983) Sand-eels plodes' into all directions Magurran & Pitcher 1987 Minnows

Hourglass: school narrows in the Pitcher & Wyche (1983) Sand-eels centre Nottestad & Axelsen (1999) Herring

Image 11: some evasion *Magurran & Pitcher (1987) did not see the minnows perform vacuole. They sug- manoeuvres performed by schools gested that even the largest shoal they experimented with, a shoal of fifty, may ddapted from Magurran & Pitcher 1987) have been too small to surround the predator pike. Schooling and evasion patterns When the predator attacks, a shoal of fish often and flash expansion (Pitcher 1993). How fish manage ly, new techniques using high-resolution multi-beam flees en masse. It turns into an aggregated, polarized to perform these evasion manoeuvres is unclear. sonar systems were developed to study schools in (all members facing the same way) swimming group, They may learn them (Kelley & Magurran 2003b), or the wild (Nøttestad & Axelsen 1999). Hopefully this termed school (as defined by Pitcher 1993. Observed they may emerge through self-organization. Vabø will revive research on the subject. For an overview by e.g. Templeton & Shriner 2005: guppies; Ferrari et and Nøttestad (1997) showed that seemingly com- of the most commonly observed manoeuvres, see al 2005: fathead minnows; Brown et al 2004: glowlight plex schooling patterns, such as herd, vacuole, and table 8 and image 11. tetras; Magurran & Pitcher 1987: European minnows) fountain could be simulated in a cellular automata by (image 9). Pitcher (1993) suggests that schooling may giving prey and predator agents a simple set of in- serve to increase the confusion effect: he describes structions (image 10). how many schooling fish species are thin and silvery, Empirical studies on fish evasion manoeuvres are, so that they become almost invisible during synchro- unfortunately, rare. They are difficult to reproduce in nous turns. Schools may evade the predator by intri- the laboratory, especially when the 'predator' is an cate predator evasion manoeuvres, such as fountain immobile model (Kelley & Magurran 2003b). Recent-

Image 10: a schooling pattern produced by Vabø and Nøttestad (1997). The red clumps are the prey, the black dot is the predator. Predator an prey were given a simple set of rules. Prey:

1 Be attracted to a shoal mate in sight;

2 Swim in the opposite direction from a predator in sight; (predator repul- sion is far stronger than shoal mate attraction) Predator:

1 Follow prey for a limited amount of time, then switch to 2; — ,q' \ 2 Attack single individuals for a limited amount of time, then switch to 1.

15 How predators deal with shoalingprey

Until now, I have mainly discussed the prey's side of predator-prey interactions. In this chapter I will focus on how predators cope with their prey's anti-predator tactics. I will start by debating if and how preda- tors select their prey. Then I will describe two strategies by which predators capture shoaling prey.

Brook trout

Northern pike

argemouth bass

17 Predator preferences In chapter 3 I explained how the confusion effect are more difficult to catch, partly due to larger swim- results in reduced capture success rates for predators ming speeds (Domenici 2001; Scharf et al. 1998; Ellis attacking larger shoals. Still, predators are attracted & Gibson 1997; Juanes & Conover 1994; Folkvord & to larger shoals. How can such an unbeneficial attrac- Hunter 1986). This explanation is supported by stud- tion have evolved? Possibly, predators want to have a ies showing that predators catch more large prey in choice of prey. From large shoals they can pick small, confined laboratory tanks than in open water (Jo- tasty, easy-to-swallow, or easy-to-catch prey. This ex- hansson 2004: pike cichlid; Christensen 1996: perch, planation is only possible if predators actively select but see Nilsson & Brönmark 2000: pike). Secondly, the their prey, a topic that is still under debate. inconsistency could arise by small fish being the first Studies on predator preference have mainly fo- ones that predators encounter, because they often cused on the selection of specific prey sizes and prey reside in the shoal's periphery (Krause 1994; Theo- Krause et al. (1998a) found that the leader of a morphologies, not on selection of specific shoal siz- dorakis 1989). And thirdly predators may more often prey shoal was attacked far more often than fish in es. Still, the findings of these studies may shed some lose large prey to fellow predators due to longer han- other positions. light on how predators select their prey. dling times: kleptoparasitism is common amongst fish (Arnegard & Carlson 2005; Hoyle & Keast 1987; Selection of small fish Nilsson & Bränmark 2000; Major 1978). Several studies show that a predator's gape size de- Researchers often find that predator diets consist termines maximum ingestible prey (Hambright 1991: of smaller prey than what the optimal foraging theo- Selection of fish without armour S largemouth bass; Nilsson et al 2000: pike, Scharf et al ry (OFT) predicts (e.g. Hoyle & Keast 1987: largemouth Predators disproportionately often eat prey with- 2000: various predators). bass; Gillen et al. 1981: tiger muskellunge). The OFT out armour (Hoyle and Keast 1987: largemouth bass; theory suggests that a predator will prefer prey that Lyons 1987: walleyes; Reist 1980: northern pike). Again Crucian carp can increase its body depth, a gives him the largest energy intake (Charnov 1976; this preference can be explained without a preda- change that takes about six weeks, when it is con- Schoener 1971). Might this inconsistency arise from tor's active preference. Armoured fish have longer fronted with chemical cues of predators (Bronmark active selection? Iwill give three alternative expla- handling times (Gillen et al 1981; Wahl & Stein 1988) and Petterson 1994), making it too large to swallow. nations. Firstly, several studies show that larger prey or may not be palatable at all (Bosher et al 2006).

Math is & Chivers (2003) described how threatened brook sticklebacks preferred to join fathead minnow shoals instead of conspecific shoals. They showed that yellow perch preferred the less armoured min- nows as food and concluded that in spite of the oddity effect sticklebacks were safer in shoals of dis- similar individuals. Active selection There is evidence for active prey selection as well. Krause & Godin (1996) showed that blue acara cichlids prefer to attack unwary prey. Arnegard & Carlson (2005) described how nocturnal mormyrid fish first probe their prey electrically. Next, they ei- ther attack or move on to another prey. The authors could not determine what factors induced mormyrid fish to abandon their first prey. They suspected that prey size may be a factor. There is also evidence that predators prefer to attack prey they are familiar with (Warburton & Thomson 2006: silver perch; Nshombo 1994: Plecodus straeleni). In an evolutionary sense it would be strange if predators didn't have a close look at their prey before they attacked. It's not beneficial to keep locking-on to, and invest energy into the pur- suit of prey that is hard to catch or digest. If smaller prey are indeed easier to catch and easier to swallow, Warburton & Thomson (2006) observed how silver predators may learn this in the course of their lives perch preferred to attack a prey species it knew to a and thus passive choice becomes active choice. novel species. Over time this preference changed. Co-operative hunting

By schooling, prey may try to confuse their pred- The cohesiveness of hunting packs varies per spe- ators. But these predators do not sit idly by, letting cies. Electric mormyrid fish remain together for days. themselves be confused. Turesson & Bränmark (2004) Even when two hunting packs meet there is little —U suggested that, in an evolutionary arms-race be- member exchange (Arnegard & Carlson 2005). They tween predator and prey, co-operative hunting may may even have a specialized electrical communica- :1 be a predator's response to schooling. Major (1978) tion system to separate pack mates from unfamiliar 3U describes how groups of co-operatively hunting conspecifics (Arnegard & Carlson 2005) Largemouth jacks aligned into U- or V-shaped attack formations bass, on the other hand, stay together for an average to split up schools of Hawaiian anchovy (image 12). In time of seven minutes (Annet 1998). contrast to attacks by individual jacks, co-operative In many cases the distinction between co-opera- attacks often succeeded. tive hunters and predators that steal each other's food Image 12 Jd( k rmming into Once the school has been disrupted, the prey (kieptoparasitism) becomes irrelevant. Major (1978) de- a school of prey. may seek refuge in the vegetation (Magurran & Pitch- scribes how a ramming attack into a school is initi- er 1987: minnows; Savino 1989: bluegills). Flushing ated by one jack. This leader usually catches the prey. prey out of their hiding places poses an additional But if it misses, this prey is often caught by a fellow Magurran & Pitcher 1987) describe how pike tried challenge. In this, co-operative hunting may prove a pack hunter. Though kleptoparasitism Is disadvanta- to flush out minnows from beneath rocks by direct- valuable tactic as well: Annet (1998) describes how a geous for the robbed predator, it is advantageous for rig jets of water at them. hunting group consisting of both largemouth bass the robber. As such, it may be a reason for predators and bluegill sunfish spread out around some vegeta- to hunt 'co-operativeIy Arnegard & Carlson (2005) tion. One predator dashed into it and flushed out all could find no co-operative hunting techniques with the prey, so that the others could get their fill. It seems packs of mormyrid fish, but kleptoparasitism was that hunting co-operatively with heterospecifics is common. These fish left their pack as soon as they not unheard of. Another fine example of this was caught prey to eat in peace. If they didn't, their prey Mow predator species have increased growth rates given by Bshary et al (2002). He described how red got stolen twelve out of thirteen times (Arnegard & v:Fo-n they hunt near conspecifics (Arnegard & Carl- sea coral groupers and lunartail groupers looked for Carlson 2005). soi .0P: niorrnyrid fish; Foster et al 2001: Australian giant moray eels. Once an eel was found, a grouper Kleptoparasitism may be one of the main con- rfl)j)klov 1992: perch), but many other species would shake its body to wake it up. Grouper and mo- siderations for short handling times: Major (1978) uff' hrrit otham & Krause 2005b: blue a ray eel would then search for prey together. While described that if a predator caught a prey 'wrongly' cichlid; Eklov 1992: pike). the moray eel sneaked through holes (inaccessible oriented, and had to disgorge it repeatedly, its pack to the grouper), the grouper would wait outside to mates would abandon their own chases and try to catch any escaping prey. The authors even observed get part of the prize. how a grouper, after waiting for two minutes for an escaped prey to come out again, fetched the moray eel and led it to the escapees hiding place. /1

Image 13: a predator ambushing its prey Ambushing Co-operative hunting is not the only strategy preda- When pike hunts near conspecffics a size- tors use to outwit their prey. Many predators are ambush- hierarchy is formed in which larger pike ers: they lie motionless at the bottom and launch sur- hold larger territories that are closer to the prise attacks on prey moving overhead (Holliday 2003: prey (Ekläv & Diehl 1994; Eklöv 1992) pike; blue acara cichlid (Krause & Godin 1995: blue acara cichlld; Arnegard & Carlson 2005: mormyrid fish). Webb (1982) describes how pike has evolved a cryptic shape that makes it hard to see. Ambushed prey don't have time to perform anti-predator behaviour, such as aggre- gating, skittering or schooling (Godin & Smith 1988: gup- pies; Milinski & Hetler 1978: threespine sticklebacks). Also, a fish that is aware of danger can anticipate the moment of a predator's strike. The exact tell-tale signs are unclear, but may involve an unnatural stillness in the predator's body (Magurran & Pitcher 1987: pike hunting minnows). Turesson & Brönmark (2004) showed that pike's first sur- prise attack had a far higher success rate than any subse- quent attacks. And Krause & Godin (1996) showed that the ambusher blue acara cichlid prefers to attack unwary prey. The deterring aspect of inspection behaviour may be especially effective with ambushing predators: if such a predator spots the inspector, it may conclude that its cover is blown and search for other prey. How predator-prey interactions aremodified by learning

Every organism is endowed with a set of genes believe that at least part of fish's antipredator behav- that enables it to cope with its environment. But in iour can be fine-tuned by learning (reviewed in Kelley a single lifetime genes cannot adapt to novel envi- & Magurran 2003b). The fitness advantage of good ronmental challenges, such as changes in the habitat learning abilities is dear: experienced fish have high- or new predators. It is therefore likely that any or- er survival rates (Mirza & Chivers 2000: brook trout; ganism that hasn't gone extinct possesses the abil- Mathis & Smith 1993a: fathead minnows). Heavily predated fish populations inspect ity to learn so that it may adjust its behaviour to the Unfortunately, learning by predators is studied predators more often, in larger groups and specific needs of the time and place in which it was far less extensively than learning by prey. As a result, from greater distances, than less-predated born. Many researchers find slight differences In anti- this chapter mainly deals with the latter. I will start by conspecifics (Magurran 1986: minnows; Seg- predator behaviour between fish that live in preda- discussing two well known learning mechanisms in hers 1974: guppies). After an inspection they tor-dense areas, and conspecifics that live in environ- fish: learning by alarm cues and social learning. are less likely to recommence foraging. ments where predators are rare. This has led them to ___

Alarm cues and learning by personal experience Many fish species, when injured by a predator at- Predator recognition Alarm cues are secreted by club-cells, large secretory tack, release chemical compounds called alarm cues. It is as yet unclear if fish need alarm cues to learn cells in the epidermis. These cells lack an opening Alarm cues have been found in many families, includ- to recognize their predators. Some species do (Chlv- at the skin surface and release their contents only ing ostariophysans, salmonids, gobies, pecilids, stick- ers et aI 1995: brook sticklebacks and, Chivers & Smith when the skin is damagaed (Pitcher 1993). lebacks, percids, cottids, cichlids and centrarchids 1994: fathead minnows, learn to recognize northern (Chivers & Smith 1998), but not in swordtails (Pfeiffer pike), but others seem to recognize their predators 1977). The chemical nature of alarm cues is still under by predator odour alone (Vilhunen & Hirvonen 2003: debate (Mirza & Chivers 2001a). Brown et al (2000) arctic char recognizes brown trout; Berejikian et al showed that a nitrogen oxide group may act as the 2003: chinook salmon recognizes northern pike; Mir- age 14 t15f. dgjregdte In resr)nse alarm cues chief molecular trigger in ostariophysans. za & Chivers 2002: brook char). Itis not entirely clear how a fish benefits from sending alarm cues. Chivers et al (1996) found that, Risk assessment I by emitting alarm signals when wounded by pike, At low concentrations alarm cues may not elicit fathead minnows attract other pike that try to steal any behavioural response in fish. This does not mean S •. - . S • .•I the prey. The authors show that the ensuing struggle they are unaware of danger. Brown et al (2004) ob- between the predators increases the minnow's sur- served how glowlight tetras, in the presence of sub- vival probabilities. It is dubious if this phenomenon behavioural threshold concentrations of a chemical is enough reason for alarm cues to evolve so univer- stimulus, responded more strongly to a frightened SIS. sally. conspecific than when no chemicals were present. I Alarm cues are greatly beneficial to the victim's And fathead minnows learned to recognize yellow .1 • shoal mates, however: while killing or digesting a perch by sub-behavioural threshold concentrations meal, predators emit its alarm cues. Their faeces are of alarm cues (Brown et al 2001 b), but not when no Luropean and fathead minnoss c a ie preda- labelled by them as well. The conspecifics of the alarm cues were present. tor recognition (either visual or il) after a predator's meal pick up their own alarm cues, realize At higher concentrations, the alarm cue concen- single exposure to predator odour accompanied that this predator eats their kind, and learn to avoid tration is often correlated to the intensity of a fish's by conspecifc alarm cues (Magurran 1989; Mathis it (Wisenden et al 1994, Mathis and Smith 1993a,b; anti-predator behaviour (Dupuch et al 2004: north- & Smith 1993b) Brown et al 1995: fathead minnows) (image 14). ern redbelly dace; Kusch et al 2004: fathead minnows; Wisenden et al 2004: glowhght tetras), and to its fu- related fish is common (Smith 1999). More interest- Brown & Dreier (2002) showed that predator-naïve ture risk assessment of a predator (Ferrari et al 2005: ingly, fish may respond to alarm cues of non-related tetras only avoided a predator's head area if this fathead minnows). Alarm cue concentration may be species: brook sticklebacks were shown to respond predator was paired with alarm cues, but once the an indicator of predator distance, since its concentra- to alarm cues of fathead minnows and vice versa tetras had learned to recognize the predator they tions are diluted within the space covered (Dupuch (Pollock et al 2003; Wisenden et al 1994). This sensi- always avoided the head area. et al 2004). In confined laboratory tanks, alarm cues tivity to non-related alarm cues is probably learned. dissolve far slower than in the wild (Wisenden et al Brown (2003) suggests four non-mutually exclusive 2004). This partly explains why fish, that are exposed mechanisms by which fish may acquire recognition to alarm cues, respond stronger in the laboratory of heterospecific alarm cues: than in the wild (Wisenden et al 2004: glowlight tet- 1. Fish may learn from heterospecific demonstra- ras; Irving & Magurran 1997: minnows). Alarm cues tors (Mathis et al 1996). may also provide information about the direction of 2.Fish may be exposed to a mixture of conspecific a threat: Krause (1993) set up a current in a laboratory and heterospecific alarm cues. This could hap- tank that transported alarm cues. The redbelly dace pen in mixed species shoals (Smith 1999). responded by swimming 'downstream'. 3.Fish may detect heterospecific alarm cues in the Little is known about how fish react when con- dietary cues of a predator, which they already flicting information about a predator is given. Ferrari know and avoid, or which also emits conspecific & Chivers (2006) found that fathead minnows adopt alarm cues (Mirza & Chivers 2001b: fathead min- a worst-case-scenario strategy if conflicting informa- nows and brook sticklebacks). tion about the seriousness of a threat is offered. Dill (1974) showed that the first few predator encounters How didalarmcue production evolve? of a naïve zebra danio determine the greater part of There are some great benefits for the shoal mates the intensity of its anti-predator behaviour. associated with alarm cue production. So this mech- anism may have evolved as a result of some form of Response to heterospecific alarm cues group selection. Alternatively current research focus- Alarm cues are highly conserved in many species es on family ties within schools. Possibly the sender (e.g. Schutz 1956: various species; Mirza & Chivers itself may not benefit, but its family will. 2001 a). Not surprisingly recognition of alarm cues of Social learning Predator-naive guppies from a low-predation site, Entire shoals were shown to adjust their anti- In addition to recognition of a predator, fish adjusted their time spent in schools and their in- predator behaviour to that of a few experienced may also learn the amount of risk associated with it. soection behaviour to high-predation site demon- conspecifics, or even heterospecifics. (Conspecific They will adjust the intensity of their anti-predator strators (Kelley et al 2003). - Pitcher et al 1986: minnows; Magurran & Higham responses (schooling, time spent hiding, freezing, 1988: minnows; Brown et al 1999: glowlight tetras. etcetera) to that of their demonstrators (Ferrari et al Bowii 20011 showed that fish learned to use escape Heterospecific - Krause 1998b: creek chub demon- 2005: fathead minnows; Vilhunen et al 2005: Arctic ioutes far sooner if they had time to get to know their strators, threespine stickleback observers; Mathis et char; Kelley et al 2003: Trinidadian guppies). To see if environment. In many studies fish are kept in one al 1996: brook stickleback demonstrators, fathead fish can also learn less common anti-predator behav- tank and placed in another for the experiment itself. minnow observers). Kelley et al (2003) showed that iour, several studies examined fishes' ability to learn In tltnse studies, the time that a fish needs to learn to the behavioural differences between fish from high- escape routes. Sugita (1980) showed that guppies use an escape oute is partly invested in overcoming predation and fish from low-predation sites may arise avoid an electric shock by following a demonstrator the problems of an unfamiliar environment. by social learning. into the safe one of two boxes. Brown & Laland (2002) Learning from conspecifics or heterospecifics found that naïve fish, which were accompanied by a has been observed in various taxa and seems to be demonstrator, took the 'right' escape route from a at least as important as learning by personal expe- model predator far more often than a group with- rience. Cultural transmission of gained knowledge out demonstrator. How do fish decide from whom across the shoal is much faster than each member to learn? Vilhunen et al (2005) showed that for social errari et al 2005) observed that predator-naive having to acquire the knowledge by personal experi- learning to occur with arctic char, the demonstrators fathead minnows did not learn to recognize ence (Brown 2003). Brown & Warburton (1999) found had to be a minority, and Swaney et al (2001) found brook char from an experienced demonstrator, that the transfer speed of new knowledge increases that guppies learn from familiar individuals faster that was exposed to alarm cue concentrations, with shoal size in rainbowfish. As such, social learn- than from unfamiliar ones. Brown (2003) argued that too low to elicit a behavioural response. ing may be an important reason for fish to live to- glowlight tetra inspectors may set the example for gether. Brown et al (1997) showed that a population their shoal mates. In this way the entire shoal may ac- ieader & Laland (2000) showed that female gup- of 80.000 fathead minnows acquired predator rec- quire predator recognition, while only a few fish have pies learned a route towards food much faster than ognition of northern pike within 2 to 4 days. Transfer to approach the predator. Nonetheless, inspectors males And fishes that were food-deprived were speed of knowledge seems to be influenced by sex, probably gain more knowledge about a predator fa ter to learn than their less hungry conspecifics. motivation and age (Reader & Laland 2000). than non-inspectors (Brown &Godin 1999).

Brown & Godin (1999) showed that only glowlight tetras that had inspected a tetra-fed predator learned to both visually and chemically recognize it. The shoal acquired only chemical recognition. Ontogeny Genetic or learned Little is known about the ontogeny of fish anti- Many researchers ask themselves which anti- predator behaviour (Kelley & Magurran 2003b). Pos- predator behaviour in fish is embedded within the sibly fry are far more sensitive to predator encoun- genetic material and which behaviour is learned. ters than adult fish. Magurran (1990b) observed To me this distinction between genetic and learned how minnows that were exposed to a pike model in seems problematic. For example: the urge to school childhood performed more inspections 2 years later. when in danger may be genetic, but to repress this In many species, fathers guard their newly hatched urge when the predator is a bird may be learned. offspring for several days: Genes set the boundaries for behaviour that can be - Three-spined sticklebacks: for about 10 days fine-tuned by learning. (Tulley & Huntingford 1987); - Paradise fish: for about five days (Miklosi et al 1997). Mikiosi et al (1997) suggest that this period may Learning by predators be a key-period in the development of anti-preda- Several studies suggest that predators special- Warburton & Thomson (2006) showed that tor behaviour. During this time the father chases the ize on specific prey types. From this type they learn silver perch's capture success of an unfamil- young when they stray, catches them in his mouth life style and anti-predator behaviours (Warburton iar prey increased over time, suggesting that and spits them back into the nest. These experiences & Thomson 2006: perch; Nshombo 1994: Plecodus it developed improved search and handling may provide early anti-predator training (Tulley & Straeleni). Bosher et al (2006) showed that large- skills. Huntingford 1987). Raising by the father can lead to mouth bass became sensitized to spined catfish: Nshombo (1994) described how hunting strongly increased anti-predator behaviour (Tulley & after it had to disgorge prey several times (because techniques of the predator Plecodus strae!eni Huntingford 1987, Benzie 1965: three-spined stickle- of the spine) it began to disgorge catfish from which varied with sex, size and even individuals. He backs; Miklosi et al 1997: paradise fish). the spines were clipped as well. One would expect noted that many predators specialized on To be able to assess the learning abilities of fish, that largemouth bass would learn to avoid catfish al- one or several prey species (this also varied one must take ontogeny into account. Walling et al together. Catfish was the only prey species available. per individual). He attributed this specializa- (2004) caught young and old three-spined stickle- Perhaps, if the predator would have had a choice, it tion on the predators getting to know their backs from a perch sympatric population. Since the would have switched to another prey type. prey's life style and anti-predator behaviours. older ones did not show stronger developed anti- predator responses than the younger ones, the re- searchers conclude that experience is of little impor- Paradise fish that had been raised by their fathers tance for the development of anti-predator respons- showed higher fleeing and backing frequencies to- es. But if anti-predator responses are mainly learned wards a predator than did orphaned fish (Mikiosi et in childhood, these conclusions may be incorrect. al 1997). Conclusions There are more fish species than all other verte- (Parrish 1993, Parrish 1989: black skipjack; Magurran brates combined. Also, the few species that are most & Pitcher 1987: pike; Major 1978: black seabass). Pos- commonly researched (guppies; minnows; stickle- sibly, an increasing ability of predators to cope with backs; pike), and are often supposed to represent schools will drive their prey back towards solitary life fish in general are largely chosen because they live (Eklôv 1992). nearby, or because they are easy research subjects, Inspection implies the counterintuitive behav- not because they are such great examples of 'typical iour of approaching one's predator. Such behaviour fish' (Svensson et at 2000). Therefore caution should seems dangerous, yet several studies showed that in- be exercised in drawing general conclusions on fish spectors were in less danger than their shoal mates! behaviour. Possibly the predator gets confused by the unex- There are, however some research topics, which pected behaviour, but since inspection is quite com- transcend even the fish taxa as a whole, such as living mon, one would expect the predator to get used to in groups. As of yet, researchers are unsure how ad- it. Alternatively, the inspectors may often consist of a vantages relate to disadvantages. They cannot eas- phenotype that is hard to capture, e.g. large, healthy ily be compared (since they influence one another) fish. (Krause & Ruxton 2002): the dilution effect only func- During my study of the literature I regularly en- tions if shoal delectability doesn't increase propor- countered passages or studies that noted the signifi- tionally to shoal size. Odd-looking individuals benefit cant influence of laboratory conditions on fish behav- less from the confusion effect than common-looking iour. Braithwaite & Salvanes (2005) found that North ones. Shoaled fish may confuse predators, but they Sea cod reared in and laboratory tanks do not devel- seem to attract them at the same time. Therefore it op the same learning skills and tendency to explore is hard to assess if the advantages of group life out- as conspecifics reared in more natural conditions. weigh the disadvantages. Since it is so common, one Anti-predator responses were shown to be more in- must assume that it has been advantageous at one tense in the laboratory than in the wild (Wisenden point in time. But predators evolve too. Many stud- et al 2004: glowlight tetras; Irving & Magurran 1997: ies describe how they utilize sophisticated hunting minnows). As mentioned before, increased concen- techniques to split up schools and pick off stragglers trations of alarm cues may explain this. Alternatively, ilhunen (2003) describes how brown trout, ex- fish may be stressed: they cannot flee, and since they posed to pikeperch odour tried to swim away. After are often alone when exposed a predator, they are it found that it could not escape (because it was in a unable to shoal (Smith 1997). laboratory tank) it started to freeze arid dash about. The effects of laboratory conditions may make behaviour of laboratory fish difficult to compare to that of their wild counterparts. An increasing amount of researchers points out the importance of natural conditions for research (Brown 2001; Irving & Magur- ran 1997). The drawback of such studies is that the researcher can control less variables. However, since laboratory conditions seem to add other variables, such as confined spaces and unfamiliar environ- merits, this benefit is reduced. Personally,I found studies performed under (near-)natural conditions such as the excellent studies by Arnegard & Carlson (2005), Parrish (1993) and Major (1978) more informa- tive. Since researchers established that fish are social, clever creatures, research has focused on topics such as fish character and social networks in shoals. Hope- fully these studies will provide new insight into fasci- nating aspects of fish behaviour, such as inspection behaviour and co-operative hunting. --

Appendix: Fish names and families

English name Latin name Dutch name Order Family Arctic charr Salvelinus alpinus Arctische zaimforel Salmoniformes (Zalmachtigen) Salmonidae (Zalmen) Atlantic herring Clupea harengus Atlantische haring Clupeiformes (Haringachtigen) Clupeoidae (haringen) Atlantic silversides Menidia menidia Getijdenaarvis Atheriniformes (Aarvisachtigen) Menidlinae (Koornaarvissen) Australian salmon Arripis trutta Australische zalm Perciformes (baarsachtigen) Carangidae (Horsmakrelen) Banded killifish Fundulus diaphanus Cyprinodontiformes (Tandkarpers) Fundulidae (Killivisjes) Black sea bass Centropristis stnata Zwarte zeebaars Perciformes (baarsachtigen) Serranidae (Zeebaarzen) Black skipjack Euthynnus lineatus Gevlekte tonijn Perciformes (baarsachtigen) Scombridae (Makreelachtigen) Blue acara cichlid Aequidens pulcher Blauwe acara Perciformes (baarsachtigen) Cichlldae (bonte baarzen) Bluegills Lepomis macrochirus blauwkeel zonnebaars Perciformes (baarsachtigen) Centrarchidae (Zonnebaarzen) Bluespotted goby Asterropteryx semipunctata Sterrengrondel Perciformes (baarsachtigen) Gobiidae (zeegrondels) Bluntnose minnow Pimephales notatus Cypriniformes (karperachtigen) Cyprlnidae (Karpers) Brook charr Salvelinus fontinalis Bronforel Salmoniformes (Zalmachtigen) Salmonidae (Zalmen) Brook stickleback Culaea inconstans Beekstekelbaars Gasterosteiformes (Stekelbaarsachtigen) Gasterosteidae (Stekelbaarzen) Brown trout Sairno trutta Zeeforel Salmoniformes (Zalmachtigen) Salmonidae (Zalmen) Catfish Silurus glanis Meerval Cypriniformes (karperachtigen) Siluridae (Echte Meervallen) Chinook salmon Oncorhynchus tshawytscha Quinnat Salmoniformes (Zalmachtigen) Salmonidae (Zalmen) Coalfish Pollachius virens Koolvis Gadiformes (Schelvissen) Gadidae (Kabeljauwachtigen) Cod Gadus morhua Kabeljauw Gadiformes (Schelvissen) Gadidae (Kabeljauwachtigen) Creek chub Semotilus atromaculatus Cypriniformes (karperachtigen) Cyprinidae (Karpers) Crucian carp Carassius carassius Kroeskarper Cypriniformes (karperachtigen) Cyprlnidae (Karpers)

Darter Etheostoma blennioides Grondbaars Perciformes (baarsachtigen) — Percidae (Echte baarzen) European minnow Phoxinus phoxinus Elrits Cypriniformes (karperachtigen) Cyprinidae (Karpers) Fathead minnow Pimephales promelas Amerikaanse dikkop-elrits Cypriniformes (karperachtigen) Cyprinidae (Karpers) Finescale dace Phoxinus neogaeus Cypriniformes (karperachtigen) Cyprinidae (Karpers) Flatiron herring Harengula thrissina Clupeiformes (Haringachtigen) Clupeoidae (haringen) Glowlight tetras Hemigrammus erythrozonus Vuurneon Characiformes (Karperzalmachtigen) Characidae (karperzalmen) Green jack Caranx cabal Groene jack Perciformes (baarsachtigen) Carangidae (Horsmakrelen) Hawaiian anchovy Encrasicholina purpurea Hawaiiaanse anchovis Clupeiformes (Haringachtigen) Clupeoidae (haringen)

Jack Caranx ignobilis Jack Perciformes (baarsachtlgen) Carangidae (Horsmakrelen) Largemouth Bass Micropterus salmoides Forelbaars Perciformes (baarsachtigen) Centrarchidae (Zonnebaarzen) Lunartail grouper Variola louti Wijnrode tandbaars Perciformes (baarsachtigen) Serranidae (Zeebaarzen) English name Latin name Dutch name Order Family Minnow: see European minnow Moray eel Gymnothorax javanicus Reuzemurene Anguilhformes (Pallngachtigen) Muraenldae (murenen) - Mormyrid fish Mormyrops anguilloides Nijisnoek Osteoglossiformes (Beentongvlssen) Mormyridae (Olifantsvissen) Mosquitofish Gambusia affinis Koboldvisje Cyprinodontiformes (Tandkarpers) Poeciliidae (Levendbarende tandkarpers) Northern pike Esox lucius Snoek Esociformes (snoekachtigen) Esocidae (Snoeken) Northern redbelly dace Phoxinus eos Cypriniformes (karperachtigen) Cyprinidae (Karpers) Paradise fish Macropodus opercularis Paradijsvis Perciformes (baarsachtigen) Osphronemidae (Echte goerami's) Perch Perca fluviatilis Rivieraars Perciformes (baarsachtlgen) Percldae (Echte baarzen) Pike: see northern pike Pike cichlld Crenicichla frenata Perciformes (baarsachtigen) Cichlidae (bonte baarzen) Pikeperch Sander lucioperca Snoekbaars Perciformes (baarsachtlgen) Percldae (Echte baarzen) Rainbow trout Oncorhynchus mykiss Regenboogforel Salmoniformes (Zalmachtlgen) Salmonldae (Zalmen) Red sea coral grouper Plectropomus pessuliferus Panter forelbaars Perciformes (baarsachtigen) Zeebaarzen (Serranidae) Redbelly yellowtail fusilier Caesio cuning Perciformes (baarsachtigen) Clchlldae (bonte baarzen) Roach Rutilus rutilus Blankvoorn Cypnnlformes (karperachtigen) Cyprinidae (Karpers) Rock bass Ambloplites rupestris Steenbaars Perciformes (baarsachtigen) Centrarchidae (zonnebaarzen) Sand-eel Ammodytes sp. Zandspieringen Perciformes (baarsachtlgen) Ammodytidae (zandspelrlngen) Scale-eater Plecodus straeleni Perciformes (baarsachtigen) Clchlidae (bonte baarzen) Silver perch Bidyanus bidyanus Zilverbaars Perciformes (baarsachtigen) Terapontidae (Tijgerbaarzen) Silvery minnow Hybognathus nuchalls Cypriniformes (karperachtigen) Cyprinidae (Karpers) Spottail shiners Notropis hudsonius Cypriniformes (karperachtigen) Cyprlnidae (Karpers) Stroneroller minnow Campostoma anomalum Cypriniformes (karperachtigen) Cyprinidae (Karpers) Swordiall Xiphophorus birchmanni Zwaardstaart Cyprinodontiformes (Tandkarpers) Poecillidae (Levendbarende tandkarpers) Three-spined stickleback Gasterosteus aculeatus Driedoornige stekelbaars Gasterostelformes (Stekelbaarsachtigen) Gasterosteldae (Stekelbaarzen) Tiger muskellunge Esox masquinongy Maskulonge Esociformes (snoekachtigen) Esocidae (Snoeken) Trinidadlan guppy Poecilla reticulata Guppy Cyprinodontiformes (Tandkarpers) Poecillidae (Levendbarende tandkarpers) White perch Morone americana Perciformes (baarsachtigen) Moronidae (Moronen)

Wolf fish Hoplias malabaricus Wolfvis Characiformes (Karperzalmachtigen) Erythrinidae (Roofzalmen) - Yellow perch Perca flavescens Gele baars Perciformes (baarsachtlgen) Percidae (Echte baarzen) Zebra danio Danio rerio Zebravis Cypriniformes (karperachtigen) Cyprlnidae (Karpers)

fl References Botham, MS. and Krause, J. (2005a) Shoals receive more at- Brown, G.E., Chivers, D.P. and Smith, R.J.F. (1997) Differential

Abrahams, M.V. (1995) The interaction between antipredator tacks from the Wolf-fish (Hoplias malabaricus Bloch, 1794). learning rates of chemical versus visual cues of a northern behaviour and antipredator morphology: Experiments Ethology, 111,881-890. pike by fathead minnows in a natural habitat. Environmen- with fathead minnows and brook sticklebacks. Canadian Botham, M.S. and Krause, J. (2005b) The effect of competitor tal Biology of Fishes, 49,89-96. Journal of Zoology, 73, 2209-2215. odour on predator choice for grouped prey in blue acara Brown, G.E. and Cowan, J. (2000) Foraging trade-offs and Allan, J.R. and Pitcher, T.J. (1986) Species segregation during cichlids, aequidens puicher (Gill, 1858). Behaviour, 142,441- predator inspection in an Ostariophysan fish: switching predator evasion in cyprinid fish shoals. Freshwater Biology, 453. from chemical to visual cues. Behaviour, 137, 181-1 96. 16,653-659. Braithwaite, V.A. and Salvanes, A.G.V. (2005) Environmental Brown, G.E. and Dreier, V.M. (2002) Chemical predator inspec- Annett, C.A. (1998) Hunting behavior of Florida largemouth variability in the early rearing environment generates tion and attack cone avoidance in a Characin fish: the bass, Micro pterus salmoides floridanus, in a channelized river. behaviourally flexible cod: implications for rehabilitating effects of predator diet. Animal Behaviour, 63, 1175. Environmental biology of fishes, 53, 75-87. wild populations. Proceedings of the Royal SocietyBiologicol Brown, G.E. and Godin, J.G.J. (1999) Who dares, learns; chemi- Arnegard, M.E. and Carison, B.A. (2005) Electric organ dis- Sciences Series B, 1-7. cal inspection behaviour and acquired predator recogni- charge patterns during group hunting by mormyrid fish. Brosnan, S.F., Earley, R.L. and Dugatkln, L.A. (2003) Observa- tion in a characin fish. Animal Behaviour, 57,475-481.

Proceedings of the Royal SocietyofLondon B Biological Sciences, tional Learning and Predator Inspection in Guppies (Po- Brown. G.E., Godin, i.G.J. and Pedersen, J. (1999) Fin-flicking 272, 1305-1314. ecilia reticulata). Ethology, 109,823-833. behaviour: a visual antipredator alarm signal in a characin Ashley. E.J., Kats, L.B. and W. Wolfe, J.W. (1993) Balancing Trade- Brown, C. (2001) Familiarity with the test environment im- fish, Hemigrammus erythrozonus. Animal Behaviour, 58,469- Offs between Risk and Changing Shoal Size in Northern proves escape responses in the crimson spotted rainbow- 475. Red-Belly Dace (Phoxinus eos). Copeia, 2,540-542. fish, Melanotoenia duboulayi. Animal Cognition, 4, 109-113. Brown, G.E., Golub, J.L. and Plata, DL. (2001 a) Attack cone Barber, I. and Huntingford, F.A. (1996) Parasite infection alters Brown, C. and Warburton, K. (1999) Social mechanisms en- avoidance during predator inspection visits by wild fines- schooling behavior: deviant positioning of helminth-in- hance escape responses in shoals of rainbowflsh, Melano- cale dace (Phoxinus neogaues): the effects of predator diet. fected minnows in conspecific groups. Proceedings of the toenia dubouloyi Environmental Biology of Fishes, 56,455-459. Chemical ecology, 27, 1657-1666.

Royal Society of London B Biological Sciences, 263, 1095-1102. Brown, G.E. (2003) Learning about danger: chemical alarm Brown, G.E. and Laland, K.N. (2002) Social learning of a novel Benzie, V. (1965) Some aspects of the anti-predator responses cues and local risk assessment in prey fishes. Fish and fish- avoidance task in the guppy, P. retlculour conformity and of two species of stickleback. University of Oxford. eries, 4, 227-234. social release. Animal Behaviour, 64,41-47. Berejikian, BA., Tezaka, E.P. and LaRaeb, AL. (2003) Innate and Brown, G.E., Adrian, J.C.i., Patton, T. and Chivers, D.P. (2001 b) Brown, G.E., Poirier, J. and Adrian, J.C.J. (2004) Assessment of enhanced predator recognition in hatchery-reared chi- Fathead minnows learn to recognize predator odour local predation risk: the role of subthreshold concentra- nook salmon. Environmental Biology of Fishes, 67,241-251. when exposed to concentrations of artificial alarm phe- tions of chemical alarm cues. Behavioral Ecology, 15,810- Blaxter, J.H.S. and Hunter, iS. (1982) The biology of the dupe- romone below their behavioural threshold. Canadian Jour- 815. oid fishes. Advances in marine biology, 20, 2-223. nal of Zoology, 79, 2239-2245. Brown, G.E. and Schwarzbauer, E.M. (2001) Chemical predator Bosher, B.T., Newton, S.H. and Fine, M.L. (2006) The spines of Brown, G.E., Adrian, iC., Smyth, E., Left, H. and Brennan, S. inspection and attack cone avoidance in a characin fish: channel catfish, Ictalurus punctatus, as an anti-predator (2000) Ostariophysan alarm pheromones: laboratory and the effects of predator diet. Behaviour, 138, 727-739. adaptation: an experimental study. Ethology, 112, 188-195. field tests of the functional significance of nitrogen ox- Brown, G.E. and Smith, RJ.F. (1996) Foraging trade-offs in fat- Botham, MS., Kerfoot, Ci., Louca, V. and Krause, J. (2005) ides. Journal of Chemical ecology, 26, 139—154 head minnows (Pimephales promelas, Osteichthyes, Cyprini- Predator choice in the field; grouping guppies, Poecilia Brown, G.E., Chivers, D.P. and Smith, R.J.F. (1995) Localized dae): acquired predator recognition In the absence of an reticulata, receive more attacks. Behavioral Ecology and So- defication by pike: a response to labelling by cyprinid alarm response. Ethology, 102,776-785.

ciobialogy, 59, 181-184. alarm pheromone? Behavioral Ecology and Sociobiology, 36, Brown, G.E. and Smith, R.J.F. (1998) Acquired predator rec- 105-110. ognition in Juvenile rainbow trout (Oncorhynchus mykiss): conditioning hatchery-reared fish to recognize chemical lour. Canadian Journal of Zoology, 65,803-811. (2005) The role of learning In the development of threat- cues of a predator. Canadian journal of fisheries and aquatic Domenlci, P. (2001)The scaling of locomotor performance in sensitive predator avoidance by fathead minnows. Animal sciences, 55,611-617. predator prey encounters: from fish to killer whales. Com- Behaviour, 70,777-784. Bshary, R., Wickler, W. and Fricke, H. (2002) Fish cognition: a parative Biochemistry and Physiology PartA, 131, 169-182. Foam, P.E., Harvey, M.C., Mirza, R.S. and Brown, G.E. (2004) primate's eye view. Animal Cognition, 5, 1-13. Dominey, WJ. (1983) Mobbing in colonially nesting fishes, Heads up: Juvenile convict cichlids switch to threat-sensi- Charnov, E.L. (1976) Optimal foraging, marginal value theo- especially the blue-gill, Lepomis mocrochirus. Copeio, 4, tive foraging tactics based on chemosensory informa- rem. Theoretical Population Biology, 9, 129-1 36. 1086-1088. tion. Animal Behaviour, 70,601—607. Chivers, D.P., Brown, C. and Smith, RJ.F. (1996) The evolution Dugatkin, L.A. and Alfieri, M. (1991) Guppies and the tit-for-tat Folkvord, A. and Hunter, J.R. (1986) Size specific vulnerability of chemical alarm signals: attracting predators benefits strategy preference based on past interaction. Behavioral of northern anchovy, Engraulis mordax, larvae to preda- alarm signal senders. American naturalist, 148,649-659. Ecology and Sociobiology, 28, 243-246. tion by fishes. Fishery Bulletin, 84,859-869. Chivers, D.P., Brown, G.E. and Smith, RJ.F. (1995) Aquired Dugatkin, L.A. and Godin, J.GJ. (1992a) Prey approaching Foster, E.G., Ritz, D.A., Osbom, i.E. and Swadling, K.M. (2001) recognition of chemical stimuli from pike, Esox lucius, by predators: a cost-benefit perspective. Annals Zoologica Schooling effects the feeding success of Australian salm-

brook sticklebacks, Culea inconstans (Osteichthyes, Gaster- Fennicl, 29, 2 33-252. on (Arripls trutto) when preying on mysid swarms (Paro- oteidae). Ethology, 99,234-242. Dugatkin, L.A. and Godin, J.GJ. (1992b) Predator inspection, mesopodopsls rufa). Journal of ExperimentolMarine Biology and Chivers, D.P., Pullitz, M.H. and Blaustein, A.R. (2000) Chemical shoaling and foraging under predation hazard in the Ecology, 261,93-106. alarm signalling by reticulate sculpins, Coitus perplexus. Tnnidadian guppy, Poecilia reticulata. Environmental biology Fothergill, A. and Atten borough, D. (2001) The Blue Planet.

Envi ronmental biology of fishes, 57,347-352. of fis.hes, 34,265-276. United Kingdom. Chivers, D.P.and Smith, R.J.F. (1 994) Fathead minnows, Dupuch, A., Magnan, P. and Dill, L.M. (2004) Sensitivity of Fricke, H.W. (1973) The Coral Seas. G. P. Putnam's Sons, New Pimephales protnelas, aquired predator recognition when northern redbelly dace, Phoxlnus eos, to chemical alarm York. alarm substance is associated with the sight of unfamiliar cues. Canadian Journal of Zoology, 82,407-415. George, C.J.W. (1960) Behavioural interaction of the pickerel fish. Animal Behaviour, 48,597-605. Eibi-Eibesfeldt, I. (1962) Freiwasserbeobachtungen zur Deu- and the mosquitofish. Harvard University, Vol. PhD. Chivers, D.P. and Smith, R.J.F. (1998) Chemical alrm signalling tung des Schwarmverhaltens verschiedener Fische. Gillen, A.L., Stein, R.A. and Carllne, R.F. (1981) Predation by pel- in aquatic predator-prey systems: a review and prospec- ZeltschnftfurTlerpsychologie, 19, 165-182. let-reared tiger-muskellunge on minnows and bluegills in tus. Ecoscience, 5, 338-352. EklOv, P. (1992) Group foraging versus solitary foraging effi- experimental systems. Transactions of the American Fisheries Christensen, B. (1996) Predator foraging capabilities and prey ciency in piscivorous predators: the perch, PercaJiuviatilis, Society, 110, 197-209. antipredator behaviours: Pre- versus postcapture con- and pike, Esoxlucius, patterns. Animal Behaviour,44, 313- Godin, J.G.J. (1986) Risk of predation and foraging behaviour straints on size-dependent predator-prey interactions. 326. in shoaling banded killifish (Fundulus diaphanus). Canadian Qikos, 76,368-380. Eklöv, P. and Diehi, S. (1994) Piscivore efficiency and ref uging Journal of Zoology, 64, 1675-1678. Croft, D.P., Arrowsmith, BJ., Bielby, J., Skinner, K., White, E., prey: the importance of predator search mode. Oecologia, Godin, J.G.J., Classon, U. and Abrahams, M.V. (1988) Group Couzin, l.D., Magurran, A.E., Ramnarine, I. and Krause, J. 98. 344-353. vigilance and shoal size in a small characin fish. Behaviour, (2003) Mechanisms underlying shoal composition in the Ellis, 1. and Gibson, R.N. (1997) Predation of 0-group flathshes 104,29-40. Trinidadian guppy, Poecilia reticulato. Oikos, 100,429-438. by 0-group cod: Handling times and size-selection. Marine Godin, J.G.J. and Davis, S.A. (1995) Who dares, benefits: preda-

Dill, L.M. (1974) The escape response of zebra danio(Brachy- Ecology Progress Series, 149,83-90. tor approach behaviour in the guppy (Poecilia reticulata) danio rerio) II. The effect of experience. Animal Behaviour, Ferrari, M.C.O. and Chivers, D.P. (2006) Learning threat-sensi- deters predator persuit. Proceedings of the Royal Society of 22,723-730. tive predator avoidance: how do fathead minnows incor- London B Biological Sciences, 259, 193-200. Dill, L.M. (1987) Animal decision making and its ecological porate conflicting information? Animal Behaviour. 71, 19-26. Godin, J.GJ. and Dugatkin, L.A. (1996) Female mating prefer- consequences: the future of aquatic ecology and behav- Ferrari, M.C.O., Trowell, JJ., Brown, G.E. and Chivers, D.P. ence for bold males in the guppy, Poecillareticulata. Pro-

33 ceedings of the NationalAcademy of Sciences of the United States ence or size-dependent capture success? Marine ecology- Krause, J. and Ruxton, G.D. (2002) Living in groups. Oxford se-

of America, 93, 10262-10267. progress series, 114, 59-69. ries in ecology and evolution. Godin, J.G.J. and Morgan, Mi. (1985) Predator avoidance and Kelley, J.L., Evans, J.P., Ramnarine, l.W. and Magurran, A.E. Krause, J., Ruxton, G.D. and Rubenstein, D. (1 998a) Is there foraging behaviour in shoaling banded killifish (Fundulus (2003) Back to school: can antipredator behaviour in gup- always an influence of shoal size on predator hunting suc- diaphnus). Behavioral Ecology and Sociobiology, 16, 105-110. pies be enhanced through social learning? AnimalBehav- cess? Journal of Fish Biology, 52,494-501. Godin, J.G.J. and Smith, S.A. (1988) A fitness cost of foraging in jour, 65,655—662. Krause, J. and Tegeder, R.W. (1994) The mechanism of aggre- the guppy. Nature, 333,69-71. Kelley, J.L. and Magurran, A.E. (2003a) Effects of relaxed pre- gation behaviour in fish shoals: individuals minimize ap- Hager, M.C. and Helfman, G.S. (1991) Safety in numbers: shoal dation pressure on visual predator recognition in the proach time to neighbours. Animal Behaviour, 48,353-359. size choice by minnows under predatory threat. Behavio- guppy. Behavioral Ecology and Sociobiology, 54,225-232. KUlling, D. and Milinski, M. (1992) Size-dependent predation ral Ecology andSociobiology, 29, 271—276. Kelley, J.L. and Magurran, A.E. (2003b) Learned recognition risk and partner quality in predator Inspection of stickle- Hambright, K.D.(1991) Experimental Analysis of Prey Selec- and antipredator responses in fishes. Fish and fisheries, 4, backs. Animal Behaviour, 44,949-955. tion by Largemouth Bass: Role of Predator Mouth Width 216-226. Kusch, R.C., Mirza, R.S. and Chivers, D.P. (2004) Making sense and Prey Body Depth. Transactions oftheAmerican Fisheries Krakauer, D.C. (1995) Groups confuse predators by exploiting of predator scents: investigating the sophistication of Society, 120,500-508. perceptual bottlenecks: a connectionist model of the con- predator assessment abilities of fathead minnows. Behav- Hamilton, W.D. (1971) Geometry for the selfish herd. Journal of fusion effect. Behavioral Ecology and Sociobiology, 36,421-429. ioralecologyandsociobiology, 55, 551-555.

Theoretical Biology, 31.295-311. Krause, J. (1993) The effect of 'Schreckstoff' on the shoaling Laland, K.N., Brown, C. and Krause, J. (2003) Learning in fishes: Hemelrijk, C.K. and Kunz, H. (2005) Density distribution and behaviour of the minnow: a test of Hamilton's selfish herd from three-second memory to culture. Fish and fisheries, 4, size sorting in fish schools: an individual-based model. theory. Animal Behaviour, 45, 1019-1024. 199-202.

Behavioral Ecology, 16, 178-187. Krause, J. (1994) The influence of food competition and pre- Landeau, L. and Terborgh, J. (1986) Oddity and the 'confusion Hoare, Di., Couzin, ID., Godin, J.GJ. and Krause, J. (2004) dation risk on size-assortative shoaling in juvenile chub effect' in predation. Animal Behaviour, 34. Mechanisms underlying shoal composition in the Trinida- (Leuciscus cephalus). Ethology, 96, 105-116. Lawrence, B. and Dill, L.M. (1974) The escape response of the dian guppy. Animal Behaviour, 67, 155-164. Krause, J. and Godin, J.G.J. (1994) Shoal choice in the banded zebra danio (Brachydanio rerlo). II. The effect of experi- Holliday, K. (2003) Time budgets of two coexisting cichlid spe- killifish (Fundulus diophanus, Teleostei, Cyprinodontidae): ence. Animal Behaviour, 22,723-730. cies Crenicichia alto and Aequidens pulcher in the Arima river, effects of predation risk, fish size, species composition Lehtiniemi, M. (2005) Swim or hide: predator cues cause spe-

Trinidad MSc, Universityof Leeds. and size of shoals. Ethology, 98, 128-1 36. cies specific reactions in young fish larvae. Journal of Fish Hoyle, J.A. and Keast, A. (1987) The effect of prey morphology Krause, I. and Godin, J.G.J. (1995) Predator preferences for Biology, 66, 128501299. and size on handling time in a piscivore, the largemouth attacking particular prey group sizes: consequences for Licht, T. (1989) Discrimination between hungry and satiated bass (Micro pterus saimoides). Canadian Journal of Zoology, 65. predator hunting success and prey predation risk. Animal predators: the response of guppies from high and low Irving, P.W. and Magurran, A.E. (1997) Context-dependent Behaviour, 50,465-473. predation sites. Ethology, 82,238-243. fright reactions in captive European minnows: the impor- Krause, J. and Godin, J.G.J. (1996) Influence of prey foraging Lima, S.L. and Dill, L.M. (1990) Behavioral decisions made un- tance of naturalness In laboratory experiments. Animal posture on flight behavior and predation risk: predators der the risk of predatIon: a review and prospectus. Cana-

Behaviour, 53, 1193-1201. take advantage of unwary prey. Behavioral Ecology, 7,264- dian Journal of Zoology, 68,619-640. Johansson, J., Turesson, H. and Persson, A. (2004) Active selec- 271. Lyons, J. (1987) Prey choice among piscivorous juvenile wall-

tion for large guppies, Poeciiia reticulata, by the pike cich- Krause, J., Godin, J.GJ. and Rubenstein, D. (1998b) Group eyes (Stizostedion vitreum). Canadian Journal of Fisheries and lid, Crenicichlo saxotilis. Oikos, 105,595-605. choice as a function of group size difference and assess- Aquatic Sciences,44, 758-764. Juanes, F. and Conover, DO. (1994) Piscivory and prey size ment time in fish: the influence of species vulnerability to Magurran, A.E. (1986) Predator inspection behaviour in min- selection in young-of-the-year bluefish: predator prefer- predation. Ethology, 104,68-74. now shoals: differences between populations and mdi- viduals. Behavioral Ecology and Sociobiology, 19, 267-273. Mathis, A. and Smith, R.J.F. (1993a) Chemical alarm signals conserved within salmonid fishes? Journal of Chemical ecol- Magurran, A.E. (1989) Aquired recognition of predator odour increase the survival time of fathead minnows (Pimephales ogy, 27, 1641-1655. in the European minnow (Phoxinus phoxinus). Ethology, 82, promelas) during encounters with northern pike (Esoxlu- Mirza, R.S. and Chivers, D.P. (2001 b) Learned recognition of

216-233. cius). Behavioral Ecology, 4,260-265. heterospecific alarm signals: the importance of mixed Magurran, A.E. (1990a) The adaptive significance of schooling Mathis, A. and Smith, R.J.F. (1993b) Fathead minnows, predator diet. Ethology, 107, 1007-101 8. as an anti-predator defence in fish. Annals Zoologica Fennici, Pimepholespromelos, learn to recognize northern pike, Exos Misund, O.A. (1991) Dynamics of moving masses; variability in 27,51-66. lucius, as predators on the basis of chemical stimuli from packing density, shape and size among pelagic schools, Magurran, A.E. (1990b) The inheritance and development of minnows In the pike's diet. Animal Behaviour, 46,645-656. In swimming behaviour of schools related to fish capture minnow anti-predator behaviour. Animal Behaviour, 39, Miklósi, A., Berzsenyi, G., Pongracz, Rand Csányi, V. (1995) The and acoustic abundance estimation. 105-132. 834-842. ontogeny of antipredator behavior in paradise fish larvae Monadjem, A., Owen-Smith, N. and Kemp, A.C. (1996) Broad- Magurran, A.E. and Higham, A. (1988) Information transfer (Macropodus operculoris): the recognition of eyespots. Jour- scale distribution patterns of piscivorous birds along a across fish shoals under predator threat. Ethology, 78, 153- nal of Fish Biology, 100,284-294. South African subtropical river during the dry season.

158. Miklósi, A., Csanyi, V. and Gerlal, R. (1997) Antlpredator behav- Tropical Ecology, 37, 135-1 37. Magurran, A.E., Oulton, WJ. and Pitcher, TJ. (1985) Vigilant ior in paradise fish (Mocropodus opercularis) larvae: the role Morgan, M.J. and Godin, J.G.J. (1985) Antlpredator benefits of behaviour and shoal size in minnows. Zeitschrift für Tierpsy- of genetic factors and paternal influence. Behavior Genet- schooling behaviour in a cyprinodontid fish, the banded

chologie, 67, 167-1 78. Ia, 27, 191-200. killifish (Funduius diaphanus). Zeitschrift für Tierpsychologie, Magurran, A.E. and Pitcher, TJ. (1987) Provenance, shoal size Milinksi, M. (1987) Tit-for-tat in sticklebacks and the evolution 70, 236-246. and the sociobiology of predator evasion, behaviour in of cooperation. Nature, 325,433-437. Motta, P.J. (1983) Response of potential prey to coral reef fish minnow shoals. Proceedings of the Royal Society of London B Milinksi, M. (1993) Constraints placed by predators on feeding predators. Animal Behaviour, 31, 1257-1258.

Biological ScIences, 229,439—465. behaviour. In Pitcher, TJ. (ed), The behaviour of teleost fishes. Neil, S.R.S.R. and Cullen, J.M. (1974) Experiments on whether Magurran, A.E. and Seghers, B.H. (1990a) Population differenc- Chapman & Hall, pp.285-302. shooling by their prey affects hunting behaviour of es in predator recognition and attack cone avoidance in Milinksl, M. and Keller, R. (1978) The influence of a predator on cephalopods and fish predators. Journal of zoology, 172, the guppy Poecilia reticulata. Animal Behaviour, 40,443-452. optimal foraging behaviour of sticklebacks (Gasterosteus 549-569. Major, P. (1978) Predator-prey interactions in two schooling aculeatus L.) Behaviour, 142,441-453. Nilsson, P.A. and Brön mark, C. (2000) Prey vulnerability to a fishes, Caranx ignobills and Stolephorus purpureus. Animal Milinskl, M. (1977) Do all members of a swarm suffer the same gape-size limited predator: behavioural and morphologi- Behaviour, 26,760-777. predation? Zeitschrlft für Tlerpsychologie, 45, 373-388. cal impacts on northern pike piscivory. Olkos, 88, 539-546. Maksimov, V.P. (1970) Lotsmani rovanie i travlya akul neko- Milinski, M. (1990) Information overload and food selection. In Nottestad, L. and Axetsen, B.E. (1999) Herring schooling torymi tropicheskimi rybami. Byuileten'Moskovskogo Obsh- Hughes, R.N. ted.), Behavioural mechanisms of food selection. manoeuvres in response to killer whale attack. Canadian chestva lspyroteieIPrirodyOtdel Biologicheskii, 75,30-33. Springer, Berlin Heidelberg New York, Vol.20, pp. 721-736. Journal of Zoology, 77,1540-1546. Mathis, A. and Chivers, D.P. (2003) Overriding the oddity ef- Milinski, M., Pfiuger, D., Kulling, D. and Kettler, R. (1990) Do Nshombo. (1994) Foraging behavior of the scale-eater Pie- fect in mixed-species aggregations: group choice by sticklebacks cooperate repeatedly In reciprocal pairs? codus-straelenl (Cichlidae, Teleostei) in Lake Tanganyika, armored and nonarmored prey. Behavioral Ecology, 14, Beho vioral Ecology and Sociobiology, 27, 17-21. Africa. Environmental biology of fishes, 39,59-72. 334-339. Mirza, R.S. and Chivers, D.P. (2000) Predator-recognition Nursall, J.R. (1973) Some behavioural interactions of spottail Mathis, A., Douglas, P., Chivers, D.P. and Smith, R.J.F. (1996) training enhances survival of brook trout: evidence from shiners (Notropsis hudsonius), yellow perch (Perca flavescens) Cultural transmission of predator recognition in fishes: laboratory and field-enclosure studies. Canadian Journal of and northern pike (Exos lucius). J. Fish. Res. BdCan., 30, 1161- intraspecific and interspecific learning. Animal Beha viour, Zoology, 78,2198-2208. 1178.

51, 185-201. Mirza, R.S. and Chivers, D.P. (2001a) Are chemical alarm signals Ohguchl, 0. (1978) Experiments on the selection against col- 35 our oddity of water fleas by three-spined sticklebacks. recognize chemical alarm cues of introduced brook stick- tion and their distribution. Zeitschrift für Vergleichende Physi- Zeitschrift für Tierpsychologie, 47,254-267. leback, Culaea inconstans. Environmental bioloqyof fishes, 66, ologle, 38,84-135. Parrish, J.K. (1989) Re-examining the Selfish Herd: Are central 313-319. Seghers, B.H. (1974) Schooling behavior in the guppy (Poecilia fish safer? Animal Behaviour, 38, 1048-1053. Radabauch, D.C. (1989) Seasonal colour changes and shifting reticulata): an evolutionary response to predation. Evolu- Parrish, J.K. (1993) Comparison of the hunting behavior of 4 antipredator tactics in darters. Journal of Fish Biology, 34, tion, 28,486-489. piscine predators attacking schooling prey. Ethology, 95, 679-685. Smith, iF. (1999) What good is smelly stuff in the skin? Cross- 223-246. Radakov, D.V. (1958) Adaptive value of the schooling behav- function and cross-taxa effects in fish 'alarm substances'. Pettersson, LB. and Brönmark, C. (1993) Chemical cues from iour of young pollock Pollachius virens (L.) Voprosylkhti- In Johnston, RE., Muller-Schwarze, D. and Sorensen, P.W. piscivores induce a change in morphology in crucian ologil, 11,69-74. (eds.), Advances in chemical signals in vertebrates. Kluwer Aca- carp. Oikos, 70, 396-402. Reader, SM. and Laland, K.N. (2000) Diffusion of foraging in- demic, New York, pp.475-488. Pfeiffer, W. (1977) The distribution of fright reaction and alarm novations In the guppy. Animal Behaviour, 60, 175-180. Smith, M.E. and Belk, MC. (2001) Risk assessment in western substance cells in fishes. Copeia, 653-665. Reist, J.D. (1980) Selective predation upon pelvic phenotypes mosquitofish (Gambusia afflnis): do multiple cues have Pitcher, T.J. (1980) Some ecological consequences of fish of brook sticklebacks, Culea inconstans, by northen pike additive effects? Beha vioral Ecology and Sociobiology, 51,

school volumes. Freshwater Biology, 10, 539-544. Esox lucius. Canadian Journal of Zoology, 58. 101-1 07. Pitcher, Ti. (1992) Who dares, wins: the function and evolu- Reist, J.D. (1983) Behavioral variation in pelvic phenotypes of Smith, R.J.F. (1982) Reaction of Pecina nigrofasciata, Ammocrypta tion of predator inspection behaviour in shoaling fish. brook stickleback, Culaea inconstans, In response to preda- beani, and Etheostoma swami (Percldae, PIsces) to conspe- Netherlands Journal of Zoology, 42,371-391. tion by northern pike, Esox Lucius. Environmentalbiologyof cific and intergeneric skin extracts. Canadian Journal of Pitcher, T.J. (1993) Behaviourof Teleost Fishes. Chapman & Hall, fIshes, 8,255-267. Zoology, 17,2253-2259. London. Remsen, J.V. (1991) Community ecology of neotropical kingflshers. Smith, RJ.F. (1997a) Avoiding and deterring predators. In Pitcher, T.J., Green, D.A. and Magurran, A.E. (1986a) Dicing University of California Press, Berkeley. Godin, J.GJ. (ed.), Behavioral ecology of Teleost fishes. Oxford with death: predator inspection behaviour in minnows Savino, J.F. and Stein, R. (1989) Behavioural interactions be- University Press, London, pp. 163-190. shoals. Journal of Fish Biology, 28,439-448. tween fish predators and their prey effects of plant den- Smith, RJ.F. (1997b) Does one result trump all others? Are- Pitcher, TJ. and J., W.C. (1983) Predator avoidance behaviour sity. Animal Behaviour, 37, 311-321. sponse to Magurran, Irving and Henderson. Proceedings of of sand-eel schools: why schools seldom split? In Noakes, Scharf, F.S., Buckel, J.A., Juanes, F. and Conover, D.C. (1998) the Royal Society of London B BiologicolSciences, 264,445-450. D.L.G., Lindqulst, B.G., Helfman, G.S. and Ward, J.A. (eds.), Predation by juvenile piscivorous bluefish (Pomatomus Smith, RJ.F. and Smith, M.J. (1989) Predator-recognition be-

Predators and prey in fishes. Junk, The Hague. saltatrix): the influence of prey to predator size ratio and haviour In two species of gobies, Asterropteryxsemipuncta- Pitcher, Ti., Magurran, A.E. and Allan, JR. (1986b) Size segre- prey type on predator capture success and prey profit- tus and Gnatholepis anjerenis. Ethology, 81,279-290. gative behaviour in minnow shoals. Journal of Fish Biology, ability Canadian Journal of Fisheries andAquatic Sciences, 55, Stan kowich, T. (2003) Marginal predatIon and the Importance

29, 83-95. 1695-1703. of predator preferences. Animal Behaviour, 66. Pitcher, T.J., Magurran, A.E. and Edwards, J.L. (1985) Schooling Scharf, F.S., Juanes, F. and Rountree, R.A. (2000) Predator size Stephens, D.W., Anderson, J.P. and Benson, K.E. (1997) On the mackerel and herring choose neighbours of similar size. - prey size relationships of marine predators: interspe- spurious occurrence of Tit for Tat in pairs of predator-ap- Marine Biology, 86,319-322. cific variation and effects of ontogeny and body size on proaching fish. Animal Behaviour, 53, 113-131. Pohlmann, K., Atema, J. and Breithaupt, T. (2004) The impor- trophic-niche breadth. Marine Ecology Progress Series, 208, Sugita, Y. (1980) Imitative choice behaviour in guppies. Psycho- tance of the lateral line in nocturnal predation of piscivo- 229-248. logical Research, 22, 7-12. rous catfish. Journal of Experimental Biology, 207,2971-2978. Schoener, T.W. (1971) Large-Billed Insectivorous Birds: a Pre- Svensson, P.A., Barber, I. and Forsgren, E. (2000) Shoaling be- Pollock, MS., Chivers, S.P., Mirza, R.S. and Wisenden, B.D. cipitous Diversity Gradient. Condor, 73, 154-&. haviour of the two-spotted goby. Journal of Fish Biology, 56, (2003) Fathead minnows, Pimephales promelas, learn to Schutz, F. (1956) A comparative study about the fright reac- 1477-1487. Swaney, W., Kendal, J.. Capon, H., Brown, C. and Laland, K.N. Predator Inspection behaviour In three-spined stickle- (2001) Familiarity facilitates social learning of foraging backs (Gasterosteus aculeatus): body size, local predation behaviour in the guppy. AnimalBehaviour, 62, 591 —598. pressure and cooperation. Behavioral Ecology and Sociobiol- Tegeder, R.W. and Krause, J. (1995) Density dependence and ogy, 56, 164-170. numerosity in fright stimulated aggregation behaviour of Warburton, K. and Thomson, C. (2006) Costs of learning: the

shoaling fish. Philosophical transactionsofthe Royal Societyof dynamics of mixed-prey exploitation by silver perch, London, series 8,350, 381-390. Bidyonus bidyanus (Mitchell, 1838). Animal Behaviour, 71,

Templeton, C.N. and Shriner, W.M. (2005) Multiple selection 361 —370. pressures influence Trinidadian guppy (Poeciliareticulata) Webb, P.W. (1982) Avoidance responses of fathead minnows antipredator behavior. Beha vioralEcology,15,673—678. to strikes by four teleost predators. Journal of Comparative Theodorakis, C.W. (1989) SIze segregation and the effects of PhysIology, 147,371-378. oddity on predation risk in minnow schools. Animal Behav- Wisenden, B.D., Chivers. D.P. and Smith, R.J.F. (1994) Risk-sen- iour, 38,496-502. sitive habitat use by brook stickleback (Culaea inconstans) Treherne, i.E. and Foster, W.A. (1981) Group transmission of in areas associated with minnow alarm pheromone. Jour- predator avoidance in a marine insect: the Trafalgar ef- nal of Chemical ecology, 20,2975-2983. fect. Animal Behaviour, 29, 911-917. Wisenden, B.D. and Harter, KR. (2001) Motion, not shape, Tulley, JJ. and Huntingford, F.A. (1987) Paternal care and the facilitates association of predation risk with novel objects development of adaptive variation in anti-predator re- by fathead minnows (Pimephalespromelas).Ethology. 107, sponses in sticklebacks. Animal Behaviour, 35, 1571-1 572. 357-364. Turesson, H. and Brön mark, C. (2004) Foraging Behaviour and Wisenden, B.D.. Vollbrecht, K.A. and Brown, J.L. (2004) Is there capture success in perch, pikeperch and pike and the ef- a fish alarm cue? Affirming evidence from a wild study. fects of prey density. Journal of Fish Biology, 65, 363-375. Animal Behaviour, 67, 59-67. Vabø, R. and Nettestad, L. (1997) An individual based model WoIf N.G. (1985) Odd fish abandon mixed-species groups of fish school reactions: predicting antipredator behav- when threatened. Behavioral Ecology and Sociobiology, 17, iour as observed in nature. Fisheries Oceanography, 6, 155- 47-52. 171. Wong, B.B.M. and Rosenthal, G.G. (2005) Shoal choice in Vilhunen, S. and Hirvonen, H. (2003) Innate antipredator swordtails when preferences conflict. Ethology, 111, 179-

responses of Arctic charr (Salvelinus alpinus) depend on 186. predator species and their dIet. 55, 1-10. Vilhunen, S., Hirvonen, H. and Laakkonen, M.V.-M. (2005) Less is more: social learning of predator recognition requires a low demonstrator to observer ratio in Arctic charr (Salveli- nus alpinus). Behavioral Ecology and Sociobiology, 57,275—282. Wahi, OH. and A., SR. (1988) SelectIve Predation byThree Esocids: The Role of Prey Behavior and Morphology. Trans-

actions of the American Fisheries Society, 117, 142-151. Walling, CA., Dawnay, N., Kazem, A.J.N. and Wright, J. (2004)