Behav Ecol Sociobiol (2006) 61:9–16 DOI 10.1007/s00265-006-0232-y

ORIGINAL ARTICLE

The dynamic nature of antipredator behavior: prey fish integrate threat-sensitive antipredator responses within background levels of predation risk Grant E. Brown & Alix C. Rive & Maud C. O. Ferrari & Douglas P. Chivers

Received: 8 February 2006 / Revised: 5 May 2006 / Accepted: 30 May 2006 / Published online: 28 June 2006 # Springer-Verlag 2006

Abstract Prey animals often have to face a dynamic tradeoff between the costs of antipredator behavior and the benefits of other fitness-related activities such as foraging and reproduction. According to the threat-sensitive predator avoidance hypothesis, prey animals should match the intensity of their antipredator behavior to the degree of immediate threat posed by the predator. Moreover, longerterm temporal variability in predation risk (over days to weeks) can shape the intensity of antipredator behavior. According to the risk allocation hypothesis, changing the background level of risk for several days is often enough to change the response intensity of the prey to a given stimulus. As the background level of risk increases, the response intensity of the prey decreases. In this study, we tested for possible interactions between immediate threatsensitive responses to varying levels of current perceived risk and temporal variability in background risk experienced over the past 3 days. Juvenile convict cichlids were preexposed to either low or high frequencies of predation risk (using conspecific chemical alarm cues) for 3 days and were then tested for a response to one of five concentrations (100, 50, 25, 12.5%, or a distilled water control). According to the threat-sensitive predator avoidance hypothesis, we found greater intensity responses to greater concentrations Communicated by T. Bakker G. E. Brown (*) : A. C. Rive Department of Biology, Concordia University, 7141 Sherbrooke St. West, Montreal, Quebec H4B 1R6, Canada e-mail: [email protected] M. C. O. Ferrari : D. P. Chivers Department of Biology, University of Saskatchewan, 112 Science Place, Saskatoon, Saskatchewan S7N 1E2, Canada

of alarm cues. Moreover, in accordance with the risk allocation hypothesis, we found that cichlids previously exposed to the high background level of risk exhibited a lower overall intensity response to each alarm cue concentration than those exposed to the low background level of risk. It is interesting to note that we found that the background level of risk over the past 3 days influenced the threshold level of response to varying concentrations of alarm cues. Indeed, the minimum stimulus concentration that evoked a behavioral response was lower for fish exposed to high background levels of predation than those exposed to low background levels of predation. These results illustrate a remarkable interplay between immediate (current) risk and background risk in shaping the intensity of antipredator responses. Keywords Antipredator behavior . Threat-sensitive predator avoidance hypothesis . Predation risk . Temporal variability

Introduction Prey animals are often caught between the conflicting demands of detecting and avoiding potential predation threats and meeting energy requirements (Lima and Dill 1990; Sih et al. 2000). The conflict arises because foraging activity leading to maximal food intake may leave prey vulnerable to predation (Godin and Smith 1988; Sih 1992). As such, antipredator response patterns are shaped by the tradeoffs between the benefits associated with the successful detection and avoidance of predation threats and those associated with a suite of fitness-related behavior patterns such as foraging, mating, and/or territorial defense (Lima and Dill 1990; Welton et al. 2003). The ability of prey to assess local predation risk and adjust the intensity of their

10

antipredator response according to the degree of threat is referred to as the threat-sensitive predator avoidance hypothesis (Helfman 1989; Helfman and Winkelman 1997). A great deal of research has addressed the question of threat-sensitive predator avoidance in a wide variety of prey taxa (e.g., Rochette et al. 1997; Persons and Rypstra 2001; Laurila et al. 1997; Bulova 1994), including freshwater fishes (Bishop and Brown 1992; Chivers et al. 2001; Brown et al. 2006). To date, much of this work has focused on the ability of prey to match the intensity of their antipredator responses to immediate predation risk (Vainikka et al. 2005; Ferrari and Chivers 2006). For example, bicolor damselfish (Stegastes partitus) exhibit significantly stronger antipredator responses when presented with a model predator in a vertical “strike” (high risk) vs a horizontal “searching” (lower risk) position (Helfman and Winkelman 1997). Likewise, threespine sticklebacks (Gasterosteus aculeatus) exhibit a reduction in foraging in the presence of an adult conspecific predator that is inversely correlated to the ratio of predator to larvae size (Bishop and Brown 1992). Moreover, the pattern of threat-sensitive response may vary from a graded or directly proportional response to a nongraded or hypersensitive response pattern, depending on factors such as body size, foraging tactic, or prey group size (Helfman and Winkelman 1997; Brown et al. 2006). Regardless of the specific pattern, the intensity of antipredator response is seen to vary according to the degree of perceived risk (Helfman and Winkelman 1997). Recently, several researchers have begun to examine the role background level of predation risk over days to weeks can have on the shaping of individual antipredator response patterns. Given that the level of perceived predation risk under natural conditions can vary over time due to seasonal changes in local predator and/or prey guild membership, prey movements through heterogeneous microhabitats, and/ or movement of potential predators (Sih et al. 2000), the degree of variance in the level of risk should also be expected to influence the tradeoffs between predator avoidance and foraging behavior (Lima and Bednekoff 1999; Sih et al. 2000). According to the risk allocation hypothesis (Lima and Bednekoff 1999), if risky periods are infrequent, prey should be expected to exhibit high intensity antipredator responses during rare periods of risk. Conversely, if risky periods are frequent, prey should be expected to respond less intensely to each predation threat, thereby reducing the overall costs associated with predator avoidance (Lima and Bednekoff 1999; Foam et al. 2005a). As with Helfman’s (1989) threat-sensitivity model, a growing body of literature is providing partial or full support for the risk allocation hypothesis (Hamilton and Heithaus 2001; Sih and McCarthy 2002; Foam et al. 2005a; Laurila et al. 2004; Mirza et al. 2006).

Behav Ecol Sociobiol (2006) 61:9–16

Both the threat-sensitive predator avoidance hypothesis and the risk allocation hypothesis presuppose that prey are able to reliably assess local predation threats. Within aquatic communities, a wide variety of vertebrate and invertebrate prey species rely on damage-released chemical alarm cues to assess risk (Chivers and Smith 1998; Wisenden 2000). Such chemical alarm cues are released after mechanical damage to the prey, as would occur during a predation event, and can elicit dramatic, short-term increases in species-typical antipredator behavior in both conspecifics and some sympatric heterospecifics (Chivers and Smith 1998; Smith 1999). Recently, responses to damage-released chemical alarm cues were studied in the context of both threat-sensitivity (i.e., Dupuch et al. 2004; Zhao and Chivers 2005; Brown et al. 2006) and temporal variation (Foam et al. 2005a; Mirza et al. 2006). The relative concentration of chemical alarm cues detected by prey should provide reliable information regarding local predation threats, as the detected concentration would be directly related to the proximity of a predation threat (Lawrence and Smith 1989; Dupuch et al. 2004). As the concentration of alarm cue detected by individual prey decreases, there would come a point at which overt behavioral responses would no longer be elicited. This concentration is referred to as the “minimum behavioral response threshold” (Brown et al. 2001a; Mirza and Chivers 2003). At concentrations below this point, prey may exhibit covert (Smith 1999) antipredator responses, such as changes in foraging posture (Foam et al. 2005b), increased reliance on visual risk assessment cues (Brown et al. 2004a), or the acquisition of novel predator cues (Brown et al. 2001a; Ferrari et al. 2005). At concentrations above this threshold, prey may exhibit graded responses (i.e., intensity of predator avoidance is proportional to alarm cue concentration (Jachner and Rydz 2002; Dupuch et al. 2004; Zhao and Chivers 2005) or a nongraded or “hypersensitive” (Helfman and Winkelman 1997) response (i.e., Brown et al. 2001b; Mirza and Chivers 2003; Roh et al. 2004). Recently, Brown et al. (2006) showed that juvenile convict cichlids (Archocentrus nigrofasciatus) exhibit flexible threat-sensitive response patterns, shifting from nongraded (hypersensitive; Helfman and Winkelman 1997) to graded responses in response to conspecific alarm cues as group size increases. Foam et al. (2005a) have recently demonstrated that juvenile convict cichlids rely on the frequency of recent predation risk to make antipredator behavioral decisions. When juvenile convict cichlids were exposed to conspecific alarm cues three times per day for 3 days before testing, they exhibited significantly lower intensity antipredator responses than did cichlids exposed to alarm cues only once a day. Similar results were shown for juvenile rainbow trout (Oncorhynchus mykiss) by Mirza et al. (2006). Thus, it

Behav Ecol Sociobiol (2006) 61:9–16

appears that both concentration of alarm cue detected and recent variability in the background level of predation threats both provide reliable information allowing prey to make threat-sensitive adjustments to their predator avoidance decisions. While it is clear that prey fishes are able to adjust their foraging—antipredator tradeoffs based on both immediate perceived risk (threat-sensitivity) and on background levels of risk (temporal variability)—it remains unclear how these factors might combine to influence the predator avoidance decisions made by prey. For example, does a reduced overall intensity of antipredator response that results from different background levels of risk influence the observed minimum behavioral response threshold? If an increase in the background level of predation risk results in an overall reduction in the intensity of predator avoidance behavior, then we might predict that the minimum behavioral response threshold should increase (i.e., individuals should only exhibit overt response to higher perceived risk). Alternatively, an increase in the background level of predation risk may lower the behavioral response threshold of the cichlids to offset the potential costs of reduced antipredator behavior intensity. We tested these hypotheses by exposing juvenile convict cichlids to low (once per day) or high (three times per day) frequency of risk treatments for 3 days and then quantifying their antipredator response to one of four dilutions of conspecific chemical alarm cues (100, 50, 25, and 12.5%) or a distilled water control.

Materials and methods Juvenile convict cichlids were descendants of laboratory cichlids bred with wild cichlids from Costa Rica. All cichlids were held in aerated 110 l aquaria containing gravel substrate and dechlorinated tap water (26°C, pH 7.2). All were exposed to a 12:12 light–dark cycle and were fed daily ad libitum with commercial flake food. Chemical cue preparation Skin extracts (alarm cue) were collected from 30 donor cichlids (mean standard length±SE=4.96±0.12 cm). Donors were humanely killed by cervical dislocation in accordance with Concordia University Animal Research Ethics (protocol AC-2005-BROW). We removed skin fillets from both sides of each fish and immediately placed the fillets in 300 ml of chilled distilled water. Skin fillets were then homogenized and filtered through polyester filter floss to remove any remaining tissue. We collected a total of 224.97 cm2, which was diluted to a final volume of 2,515 ml with the addition of glass-distilled water. Thus, the final concentration was similar to that used by Foam et

11

al. (2005a). We froze the alarm cue in 20-ml aliquots at −20°C until needed. Pretreatment phase Conditioning tanks consisted of a series of white opaque plastic basins (48×38×20 cm) filled with 23 l of dechlorinated tap water. Conditions were the same as described above but no filter boxes were present in the basins. Each basin was equipped with an overflow valve, positioned approximately 5 cm from the upper edge, to facilitate water changes. To create different levels of risk, we exposed groups of 25 cichlids to conspecific alarm cue either once per day (low frequency) or three times per day (high frequency) for 3 days. Though only 20 cichlids from each conditioning basin would be tested, we chose groups of 25 to ensure that we could closely match test fish for size (see below). During this phase, we used the stock concentration, i.e., the same as the 100% concentration treatment during the testing phase (see below). Stimuli were injected next to the airstone in a series of three pulses, each separated by 10 min at 12:00 h for the once a day treatments and at 9:00, 12:00, and 15:00 h for the three times a day treatments. We gave the injections as a series of three 10-ml injections of alarm cues to better simulate natural patterns of risk (Foam et al. 2005a). Ten minutes after each series of injections, we slowly flushed the basins with approximately 25 l of dechlorinated water. Water changes were done on all conditioning tanks, regardless of the treatment to control for any disturbance effects. During this pretreatment phase, cichlids were fed twice daily with commercial flake food. Food was never given in combination with the alarm cues. Approximately 1 h after the final exposure on the third day, we transferred pairs of cichlids, matched for size, to test tanks and allowed them to acclimate overnight before testing on the fourth day. Mean (±SE) standard length at testing was 3.56±0.24 cm. Each conditioning block consisted of one low frequency and one high frequency basin. Each basin yielded ten pairs of cichlids, two pairs per concentration treatment. Thus, a total of six blocks yielded 12 replicates for each frequency × concentration treatment combination. Testing phase Test tanks consisted of a series of 37 l of aquaria, each equipped with a single airstone and an additional length of tubing to allow for the injection of chemical stimuli from a distance of at least 2 m. Test tanks contained a gravel substrate and were filled with dechlorinated tap water but were not filtered. Temperature and lighting were identical to the holding tanks. To facilitate recording area use (see

12

below), we divided each tank into three vertical sections with lines drawn on the exterior of the tanks. In addition, we covered three sides of each test tank with pale orange opaque plastic to prevent transmission of visual information between tanks. Fish were fed equal amounts of commercial flake food the evening before the test day and 1 h before testing. Test fish were fed enough flake food to ensure that some flakes, once saturated with water, would sink to the substrate, providing foraging opportunities during the trials. We conducted a total of 12 replicates per treatment combination, and individual cichlids were tested only once. Trials consisted of a 5-min prestimulus observation period followed by a 5-min poststimulus observation period. Before the prestimulus observation period, we withdrew and discarded 60 ml of tank water through the stimulus injection tube (to remove any residual cues from the tube). We then withdrew and retained an additional 60 ml of water. After the prestimulus observation period, we injected 10 ml of distilled water (control), 10 ml of undiluted stock alarm cue (100%), or 10 ml of alarm cue diluted to 50, 25, or 12.5% of stock concentration with the addition of distilled water. During both the pre- and poststimulus injection observation periods, we recorded the number of foraging attempts (as the per capita rate per minute), vertical area use, and time spent moving. We defined a foraging attempt as the pecking at the gravel substrate with the body inclined at an angle greater than 45° to the substrate (Grant et al. 2002). Area use was recorded every 15 s by assigning each fish a score of 1 (bottom third of the tank) to 3 (top third of the tank). As such, mean area use scores ranged from 2 (both fish near the substrate) to 6 (both fish near the surface). We recorded time moving as the mean time (in seconds) either of the pair of test cichlids spent not being stationary (Brown et al. 2006). Decreased time moving and frequency of foraging attempts and increased time spent near the bottom are indicative of antipredator responses in juvenile cichlids (Wisenden and Sargent 1997; Alemadi and Wisenden 2002; Brown et al. 2004b).

Behav Ecol Sociobiol (2006) 61:9–16

comparisons (post hoc tests based on probabilities from Fisher’s protected least squared differences). We wanted to know whether there was a change in the minimum response threshold based on the significance of differing concentration treatments on the independent variables.

Results MANOVA results revealed a significant overall interaction between frequency of risk during the conditioning phase and the concentration of alarm cue during the testing phase (F4, 110=5.71 and P<0.001). As a result, we examined the effect of alarm cue concentration on the 1× and 3× treatments separately. For cichlids exposed to the low background level of risk (1× treatment), we found a significant effect of alarm cue concentration for change in area use (F4, 55=3.69 and P =0.01), number of foraging attempts (F4, 55=3.57 and P =0.012), and time spent moving (F4, 55=8.53 and P<0.001). Cichlids previously exposed to conspecific alarm cues once a day (infrequent risk) exhibited a significant minimum response threshold at the 25% dilutions (Table 1 and Fig. 1). For all three behavioral measures, we found no significant difference between the distilled water control and the 12.5% dilution. In addition, we found no significant difference in the intensity of response toward alarm cue at dilutions of 25% and above (Table 1 and Fig. 1). For cichlids exposed to the high background level of risk (3× treatment), we found no significant effect of concentration on change in area use (F4, 55=1.78 and P =0.15). However, there was a significant effect of concentration for change in number of foraging attempts (F4, 55=3.37 and P =0.015) and time spent moving (F4, 55=2.75 and P =0.037). According to the moving and foraging data, the cichlids previously exposed to conspecific alarm cues three times a day exhibited a significant minimum response threshold at the 12.5% dilution (Table 1 Table 1 Results of MANOVA and ANOVA comparisons

Statistical analysis To determine whether concentration and frequency had an overall effect on the dependent variables (foraging, area use, and time spent moving), we performed a multivariate analysis of variance (MANOVA). Due to the significant interaction between frequency and concentration, univariate ANOVAs were conducted to see exactly how concentration affected antipredator responses in the 1× and 3× treatments. If the concentration effect was found to be significant, we proceeded by determining the effects of differing concentrations on the individual variables within the two frequency treatments. This was achieved through multiple

MANOVA Frequency Concentration Frequency × concentration ANOVA: 1× Area use Foraging Time moving ANOVA: 3× Area use Foraging Time moving

F

df

P

1.68 18.09 5.71

3, 108 4, 110 4, 110

0.18 <0.001 <0.001

3.70 3.60 8.53

4, 55 4, 55 4, 55

0.01 0.012 <0.001

1.78 3.37 2.75

4, 55 4, 55 4, 55

0.15 0.015 0.037

Behav Ecol Sociobiol (2006) 61:9–16 Fig. 1 Mean (±SE) change in area use (a), time spent moving (b), and per capita foraging rate per minute (c) for juvenile cichlids previously exposed to low (left column) and high (right column) frequency of risk. Means under different bars are significantly different (P<0.05) based on Fisher’s protected least squared differences. N=12 per treatment combination

13

Low frequency of risk (1x per day) 0.5

High frequency of risk (3x per day)

A: Area use

0.25 0 -0.25 -0.5 -0.75 -1 DW 10

12.5

25

50

100

DW

12.5

25

50

100

50

100

DW

12.5

25

50

100

50

100

DW

12.5

25

50

100

B: Time moving (sec)

Mean change in:

5 0 -5 -10 -15 -20 -25 -30 DW 4

12.5

25

C: Per capita foraging rate

2

0

-2

-4

-6 DW

12.5

and Fig. 1). There was no significant difference in the intensity of response to alarm cues at dilutions above 12.5% (Table 1 and Fig. 1). There was a notable difference in the minimum response threshold between the 1× and 3× treatments. Finally, the significant overall interaction between frequency of risk and concentration of alarm cue (see above) suggests that cichlids exposed to the 3× treatment

25

exhibited an overall lower intensity antipredator response when compared to the 1× treatment (Fig. 1). To further examine the effect of frequency of risk on the intensity of antipredator behavior, we compared the behavioral response of cichlids exposed to low vs high frequencies of background risk for the three highest concentration treatments using a post hoc MANOVA. We found a significant effect of frequency of background risk (F3, 64=5.23 and

14

P =0.003) but not effect of concentration (F3, 65=0.18 and P =0.91) nor interaction (F3, 65=0.23 and P =0.88).

Discussion Our results demonstrate that the overall intensity of antipredator response exhibited by cichlids previously exposed to a high frequency of risk (three times a day) treatment was lower than that seen for cichlids previously exposed to a low frequency (once a day) treatment. Overall, cichlids in both the high and low frequency treatments exhibited significant decreases in time spent moving and foraging rate. Also, cichlids in the low frequency treatment showed an overall decrease in area use (movement toward the substrate). The strength of the response was consistently greater for cichlids previously exposed to the infrequent (once a day) treatment compared to the frequent treatment (three times a day). As such, our results are consistent with the theoretical predictions of the risk allocation model (Lima and Bednekoff 1999) and the empirical results of Foam et al. (2005a). These results provide additional support for the hypothesis that the background level of predation risk exerts significant influence on predator avoidance patterns. Our results also provide support for the threat-sensitive predator avoidance hypothesis. We found that cichlids exhibited stronger antipredator responses to increased concentrations of alarm cues. Our most interesting finding is the dynamic interplay between immediate (current) risk and background risk in shaping the intensity of antipredator behavior. Indeed, our results demonstrate a significant effect of background risk on the observed minimum behavioral response threshold required to elicit an overt antipredator response. Cichlids previously exposed to high background levels of risk actually reduced the minimum concentration needed to elicit an overt behavioral response, compared to those previously exposed to low background levels of risk. Under natural conditions, temporal variation in predation risk can result from a variety of nonmutually exclusive environmental and behavioral factors (Sih et al. 2000; Brown and Chivers 2005). The variation in risk may result from predators varying in seasonal abundance, preferred habitats, or predator guild composition (Sih et al. 2000). Likewise, variation in risk might result from movements, microhabitat use, or ontogenetic shifts in habitat preferences of prey themselves (Sih et al. 2000; Brown and Chivers 2005). As such, the presence and intensity of predation risk can fluctuate from year-to-year, season-toseason, or even moment-to-moment. The prediction that prey should reduce the overall intensity of predator avoidance under conditions of high frequency and/or temporally variable risk makes intuitive sense, especially

Behav Ecol Sociobiol (2006) 61:9–16

in the light of short-term tradeoffs between predator avoidance and foraging benefits (Lima and Bednekoff 1999; Sih et al. 2000). However, population differences in the overall intensity and sensitivity toward predation threats are well documented, especially in fishes (Magurran et al. 1995; Kelly and Magurran 2003). For example, threespine sticklebacks from high predation populations show faster escape speeds and more intense antipredator reactions when compared to conspecifics from low predation sites (Giles and Huntingford 1984; Huntingford et al. 1994). Likewise, fathead minnows from high predation populations exhibit a lower minimum behavioral response threshold toward chemical alarm cues compared to those from lower predation sites (Brown et al. 2001a). Presumably, individuals exhibiting an intense antipredator response should gain greater antipredator benefits than those exhibiting a less intense response (Godin 1997). As such, a reduction in antipredator response intensity under conditions of higher predation risk might result in a longterm fitness cost. Our results suggest that individuals may be able to compensate by reducing the amount of alarm cue needed to elicit an overt response. By responding less intensely, but at a lower perceived level of risk (i.e., Lawrence and Smith 1989; Dupuch et al. 2004), prey previously exposed to a high frequency of predation risk may be able to better balance the conflicting demands of predator avoidance and energy gains. One common criticism of the risk allocation model is that any decrease in activity might be due to habituation to predation risk rather than a behavioral decision per se (Hamilton and Heithaus 2001; Sih and McCarthy 2002; Peacor and Hazlett 2003). Our results suggest that this is not the case. If simple habituation were the mechanism accounting for the reduced response intensity in the high predation background treatment, we would have predicted an increase in the minimum behavioral response threshold. Rather, we observed a decrease in the concentration of alarm cue required to elicit an overt behavioral response. As such, our data provide additional support for the risk allocation hypothesis proposed by Lima and Bednekoff (1999). Lima and Steury (2005) have recently argued that one of the most significant pitfalls in the study of predator–prey dynamics is our lack of understanding of how prey perceive or assess local predation risk. This current data set suggests that both questions of “how much” and “how often” risk is encountered are important and that together, they provide information unavailable alone. In addition, recent studies showed that prey fishes are capable of detecting chemosensory cues below the overt behavioral response threshold and can use these “subthreshold” cues to make subtle riskaversive changes in behavior (Brown 2005). Thus, it is becoming increasingly clear that the relative concentration of chemical alarm cues detected by individual prey

Behav Ecol Sociobiol (2006) 61:9–16

provides ample information, allowing for threat-sensitive responses based on immediate tradeoffs. Moreover, our current data suggest that these immediate decisions are influenced by recent experience in the background level of perceived risk. Recently, Brown et al. (2006) showed that juvenile convict cichlids will shift from a hypersensitive to a more graded response to increasing concentrations of conspecific skin extract as group size increases. They demonstrated that singleton cichlids and those in shoals of three exhibited an “all-or-nothing” response, while cichlids in shoals of six exhibited a graded response. Brown et al. (2006) hypothesized that increasing temporal variability in predation risk would also result in a similar shift toward a graded response pattern. However, for both high and low frequency treatments, we found no observable graded response, rather, cichlids exhibited a hypersensitive response pattern (Helfman and Winkelman 1997). This suggests that while temporal variation in perceived predation risk does have an influence on overall response intensity and minimum behavioral response thresholds (current data), the overall pattern of the threat-sensitive decision is influenced by more immediate factors (Vainikka et al. 2005; Ferrari and Chivers 2006). Our results provide evidence that threat-sensitive antipredator responses are influenced by the background level of risk that prey experience over the past several days. The fascinating next step in understanding the nature of prey antipredator responses is to combine variation in immediate risk (threat-sensitive responses) with background levels of risk occurring over longer term ecological (lifetime) and evolutionary scales. This would allow us to gain a full appreciation of the role of temporal variation in mediating the intensity of antipredator behavior. Acknowledgements We wish to thank Antoine Leduc, Meaghan Vavrek, and James Grant for helpful comments on earlier versions of this manuscript. Financial support was provided by Concordia University, the University of Saskatchewan, and Natural Sciences and Engineering Research Council of Canada research grants to G. E. B. and D. P. C. All work reported herein was conducted in accordance with Concordia University AREC protocol # BROW-2005.

References Alemadi SD, Wisenden BD (2002) Antipredator response to injuryreleased chemical alarm cues by convict cichlid young before and after independence from parental protection. Behaviour 139:603– 611 Bishop TD, Brown JA (1992) Threat sensitive foraging by larval threespine sticklebacks (Gasterosteus aculeatus). Behav Ecol Sociobiol 31:133–138 Brown GE (2005) Local predation risk assessment based on low concentration chemical alarm cues in prey fishes: evidence for threat-sensitivity. In: Mason RT, LeMaster MP, Müller-Schwartze

15 D (eds) Chemical signals in vertebrates, vol 10. Springer, Berlin Heidelberg New York, pp 313–320 Brown GE, Chivers DP (2005) Learning as an adaptive response to predation. In: Barbosa P, Castellanos I (eds) Ecology of predator–prey interactions. Oxford University Press, Oxford, pp 34–54 Brown GE, Adrian JC Jr, Patton T, Chivers DP (2001a) Fathead minnows learn to recognize predator odour when exposed to concentrations of artificial alarm pheromone below their behavioural-response threshold. Can J Zool 79:2239–2245 Brown GE, Adrian JC Jr, Shih ML (2001b) Behavioural responses of fathead minnows (Pimephales promelas) to hypoxanthine-3-Noxide at varying concentrations. J Fish Biol 58:1465–1470 Brown GE, Poirier J-F, Adrian JC Jr (2004a) Assessment of local predation risk: the role of subthreshold concentrations of chemical alarm cues. Behav Ecol 15:810–815 Brown GE, Foam PE, Cowell HE, Guevara-Fiore P, Chivers DP (2004b) Production of chemical alarm cues in juvenile convict cichlids: the effect of diet, condition, and ontogeny. Ann Zool Fenn 41:487–499 Brown GE, Bongiorno T, DiCapua DM, Ivan LI, Roh E (2006) Effects of group size on the threat-sensitive response to varying concentrations of chemical alarm cues by juvenile convict cichlids. Can J Zool 84:1–8 Bulova SJ (1994) Ecological correlates of population and individual variation in antipredator behavior of two species of desert lizards. Copeia 1994:980–992 Chivers DP, Smith RJF (1998) Chemical alarm signaling in aquatic predator–prey systems: a review and prospectus. Écoscience 5:338–352 Chivers DP, Mirza RS, Bryer PJ, Kiesecker JM (2001) Threatsensitive predator avoidance by slimy sculpins: understanding the importance of visual versus chemical information. Can J Zool 79:867–873 Dupuch A, Magnan P, Dill LM (2004) Sensitivity of northern redbelly dace, Phoxinus eos, to chemical alarm cues. Can J Zool 82:407–415 Ferrari MCO, Chivers DP (2006) Learning threat-sensitive predator avoidance: how do fathead minnows incorporate conflicting information? Anim Behav 71:19–26 Ferrari MCO, Trowell JJ, Brown GE, Chivers DP (2005) The role of learning in the development of threat-sensitive predator avoidance by fathead minnows. Anim Behav 70:777–784 Foam PE, Harvey MC, Mirza RS, Brown GE (2005b) Heads up: juvenile convict cichlids switch to threat-sensitive foraging tactics based on chemosensory information. Anim Behav 70:601–607 Foam PE, Mirza RS, Chivers DP, Brown GE (2005a) Juvenile convict cichlids (Archocentrus nigrofasciatus) allocate foraging and antipredator behaviour in response to temporal variation in predation risk. Behaviour 142:139–144 Giles N, Huntingford FA (1984) Predation risk and inter-population variation in anti-predator behaviour in the three-spined stickleback, Gasterosteus aculeatus L. Anim Behav 32:264–274 Godin J-GJ (1997) Evading predators. In: Godin J-GJ (ed) Behavioural ecology of teleost fishes. Oxford University Press, Oxford, pp 191–236 Godin J-GJ, Smith SA (1988) A fitness cost of foraging in the guppy. Nature 333:69–71 Grant JWA, Girard IL, Breau C, Weir LK (2002) Influence of food abundance on competitive aggression in juvenile convict cichlids. Anim Behav 63:323–330 Hamilton IM, Heithaus M (2001) The effects of temporal variation in predation risk on anti-predator behaviour: an empirical test using marine snails. Proc R Soc Lond B 268:2585–2588 Helfman GS (1989) Threat-sensitive predator avoidance in damselfish–trumpetfish interactions. Behav Ecol Sociobiol 24:47–58

16 Helfman GS, Winkelman DL (1997) Threat sensitivity in bicolor damselfish: effects of sociality and body size. Ethology 103: 369–383 Huntingford FA, Wright PJ, Tierney JF (1994) Adaptive variation in antipredator behaviour in threespine stickleback. In: Bell MA, Foster SA (eds) The evolutionary biology of the threespine stickleback. Oxford University Press, Oxford, pp 227–296 Jachner A, Rydz MA (2002) Behavioural response of roach (Cyprinidae) to different doses of chemical alarm cues (Schreckstoff). Arch Hydrobiol 155:369–381 Kelly JL, Magurran AE (2003) Learned predator recognition and antipredator responses in fishes. Fish Fish 4:216–226 Laurila A, Kujasalo J, Ranta E (1997) Different antipredator behaviour in two anuran tadpoles: effects of predator diet. Behav Ecol Sociobiol 40:329–336 Laurila A, Järvi-Laturi M, Pakkasmaa S, Merilä J (2004) Temporal variation in predation risk: stage-dependency, graded responses and fitness costs in tadpole antipredator defences. Oikos 107:90–99 Lawrence BJ, Smith RJF (1989) Behavioral response of solitary fathead minnows, Pimephales promelas, to alarm substance. J Chem Ecol 15:209–219 Lima SL, Bednekoff PA (1999) Temporal variation in danger drives anti-predator behavior: the predator risk allocation hypothesis. Am Nat 153:649–659 Lima SL, Dill LM (1990) Behavioral decisions made under the risk of predation: a review and prospectus. Can J Zool 68:619–640 Lima SL, Steury TD (2005) Perception of predation risk: the foundation of nonlethal predator–prey interactions. In: Barbosa P, Castellanos I (eds) Ecology of predator–prey interactions. Oxford University Press, Oxford, pp 166–188 Magurran AE, Seghers BH, Shaw PW, Carvalho GR (1995) The behavioural diversity and evolution of guppy Poecilia reticulata populations in Trinidad. Adv Study Behav 24:155–202 Mirza RS, Chivers DP (2003) Response of juvenile rainbow trout to varying concentrations of chemical alarm cue: response thresholds and survival during encounters with predators. Can J Zool 81:88–95 Mirza RS, Mathis A, Chivers DP (2006) Does temporal variation in predation risk influence the intensity of anti-predator responses? A test of the risk allocation hypothesis. Ethology 112:44–51 Peacor KW, Hazlett BA (2003) Frequency of encounter with risk and the tradeoff between pursuit and anti-predator behaviours in

Behav Ecol Sociobiol (2006) 61:9–16 crayfish: a test of the risk allocation hypothesis. Ethology 109:97–106 Persons MH, Rypstra AL (2001) Wolf spiders show graded antipredator behavior in the presence of chemical cues from different sized predators. J Chem Ecol 27:2493–2504 Rochette R, Dill LM, Himmelman JH (1997) A field test of threat sensitivity in a marine gastropod. Anim Behav 54:1053–1062 Roh E, Mirza RS, Brown GE (2004) Quality or quantity? The role of donor condition in the production of chemical alarm cues in juvenile convict cichlids. Behaviour 141:1235–1248 Sih A (1992) Prey uncertainty and the balancing of antipredator and feeding needs. Am Nat 139:1052–1069 Sih A, McCarthy TM (2002) Prey responses to pulses of risk and safety: testing the risk allocation hypothesis. Anim Behav 63:437–443 Sih A, Ziemba R, Harding KC (2000) New insights on how temporal variation in predation risk shapes prey behaviour. Trends Ecol Evol 15:3–4 Smith RJF (1999) What good is smelly stuff in the skin? Cross function and cross taxa effects in fish “alarm substances”. In: Johnston RE, Müller-Schwarze D, Sorensen PW (eds) Advances in chemical signals in vertebrates. Kluwer Academic, New York, pp 475–488 Vainikka A, Jokelainen T, Kortet R, Ylönen H (2005) Predation risk allocation or direct vigilance response in the predator interaction between perch (Perca fluviatilis L.) and pike (Esox lucius L.)? Ecol Freshw Fish 14:225–232 Welton NJ, McNamara JM, Houston AI (2003) Assessing predation risk: optimal behaviour and rules of thumb. Theor Popul Biol 64:417–430 Wisenden BD (2000) Olfactory assessment of predation risk in the aquatic environment. Philos Trans R Soc Lond B 355:1205–1208 Wisenden BD, Sargent RC (1997) Antipredator behaviour and suppressed aggression by convict cichlids in response to injuryreleased chemical cues of conspecifics but not those of an allopatric heterospecific. Ethology 103:283–291 Zhao X, Chivers DP (2005) Response of juvenile goldfish (Carassius auratus) to chemical alarm cues: relationship between response intensity, response duration and the level of predation risk. In: Mason RT, LeMaster MP, Müller-Schwartze D (eds) Chemical signals in vertebrates, vol 10. Springer, Berlin Heidelberg New York, pp 334–341