Behav Ecol Sociobiol (2006) 60: 522–527 DOI 10.1007/s00265-006-0195-z

ORIGINA L ARTI CLE

Maud C. O. Ferrari . Terence Capitania-Kwok . Douglas P. Chivers

The role of learning in the acquisition of threat-sensitive responses to predator odours Received: 21 September 2005 / Revised: 10 February 2006 / Accepted: 7 March 2006 / Published online: 12 April 2006 # Springer-Verlag 2006

Abstract The supposition that prey animals respond to a predator with an intensity that matches the risk posed by the predator is known as the threat-sensitive predator avoidance hypothesis. Many studies have provided support for this hypothesis; yet, few studies have attempted to determine how such abilities are acquired by prey species. In this study, we investigated whether fathead minnows (Pimephales promelas) could learn to recognize an unknown predator (northern pike, Esox lucius) in such a way that they could match the intensity of their antipredator response with the threat posed by the predator. We exposed pike-naïve minnows to conspecific alarm cues paired with either a high or low concentration of pike odor. The following day, both groups were tested for a response to either high or low concentration of pike odor alone. We found that minnows conditioned with alarm cues paired with a given concentration of pike odor subsequently responded with a higher intensity to higher concentrations of pike odor, and with a lower intensity to lower concentrations of pike odor. These results demonstrate that during a single conditioning trial, minnows learn the identity of the predator in a threat-sensitive manner. Minnows use predator odor concentrations that they experience in subsequent interactions to adjust the intensity of their antipredator behavior.

Introduction Predation is an important selective force acting on the physiology, behavior, morphology, and life history of prey species. Prey animals often face a trade-off between the Communicated by K. Lindström M. C. O. Ferrari (*) . T. Capitania-Kwok . D. P. Chivers Department of Biology, University of Saskatchewan, 112 Science Place, Saskatoon, SK S7N 5E2, Canada e-mail: [email protected] Tel.: +1-306-9661430 Fax: +1-306-9664461

benefits and costs of antipredator responses (Lima and Dill 1990). If they fail to respond to predation threat, they may be captured. However, prey must also allocate time and energy for essential activities such as foraging, reproducing, and/or defending territories. Consequently, the ability to accurately assess the level of risk associated with a predation threat should be beneficial for prey. Helfman (1989) developed his threat-sensitive predator avoidance hypothesis, stating that prey should respond to a predation threat with an intensity that matches the risk posed by the predator. This hypothesis has been tested and validated in diverse aquatic organisms, including freshwater isopods (Holomuzki and Short 1990), mayflies (McIntosh et al. 1999), crustaceans (Wahle 1992), amphibians (Kats et al. 1994; Anholt et al. 1996; Puttlitz et al. 1999; Mathis and Vincent 2000; Amo et al. 2004), and fishes (Williams and Brown 1991; Hartman and Abrahams 2000; Chivers et al. 2001; Golub and Brown 2003; Marcus and Brown 2003). Some studies have focused on the ability of aquatic organisms to assess the level of risk they are exposed to through the relative concentrations of chemicals present in their environment. Kusch et al. (2004) showed that wildcaught pike-experienced fathead minnows (Pimephales promelas) used the concentration of pike (Esox lucius) odor they were exposed to as a risk assessment tool. Indeed, minnows exposed to low concentrations of pike odor responded with a lower intensity than minnows exposed to higher concentrations of pike odor. Similarly, redbelly dace (Phoxinus eos) (Dupuch et al. 2004), goldfish (Carassius auratus) (Zhao and Chivers 2005), and fathead minnows (Ferrari et al. 2005) exhibit stronger antipredator behaviors to increased concentrations of conspecific alarm cues. Chemical alarm cues are found in many aquatic organisms including protozoans, flatworms, annelids, arthropods, mollusks, fishes, and amphibians (Wisenden 2003). In fishes, they have been documented in ostariophysans, salmonids, gobies, poeciliids, sticklebacks, percids, cottids, cichlids, and centrarchids (review Chivers and Smith 1998). These chemicals are located in the epidermis, and, thus, can be released in the water column when the skin is damaged after capture by a predator. An

523

interesting characteristic of those alarm cues is that they elicit innate antipredator responses when detected by conspecifics. Also, they have been demonstrated to facilitate the learned recognition of predator identities (review Chivers and Smith 1998). Many fish species do not show innate recognition of their predators (Mathis et al. 1993; Chivers and Smith 1994a; however, see Vilhunen and Hirvonen 2003; Berejikian et al. 2003). Thus, learning is a necessary step for them to identify their potential predators. One way of learning is through conditioning with conspecific alarm cues paired with visual and/or chemical cues of the predator (review Chivers and Smith 1998; Smith 1999; Kelley and Magurran 2003). For example, European minnows (Phoxinus phoxinus) and fathead minnows acquire the chemical recognition of a novel predator after a single exposure to the predator odor paired with conspecific alarm cues (Magurran 1989; Mathis and Smith 1993; Chivers and Smith 1994b). Despite the large number of studies showing threatsensitive antipredator responses to predation cues, only two studies have investigated how such threat sensitivity develops. Ferrari et al. (2005) conditioned trout-naïve fathead minnows with different concentrations of conspecific skin extract (containing alarm cues) paired with trout odor and recorded the intensity of their antipredator response. Minnows exposed to higher concentrations of skin extract responded with higher intensities of response. When subsequently exposed to trout odor alone, minnows responded with an intensity that matched the intensity they initially displayed during their conditioning trials. This study showed that minnows learn to identify their predators and highlighted the importance of alarm cue concentrations in the acquisition of information regarding a level of threat posed by a predator. A subsequent study by Ferrari and Chivers (2006) showed how minnows incorporate conflicting information when conditioned multiple times with different concentrations of conspecific alarm cues paired with predator odor. Both of these studies investigated the effect of changing the alarm cue concentrations that were paired with a constant predator odor concentration during conditioning trials. The present study attempts to extend our knowledge of the acquisition of threat-sensitive predator learning in prey animals. In this study, we focus on understanding the effect of manipulating predator odor concentrations on the acquisition of threat-sensitive predator avoidance in fathead minnows. We conditioned minnows with conspecific skin extract paired with either high (H) or low (L) concentration of pike odor. Then, we subsequently tested them with either high or low concentration of pike odor alone and documented the intensity of their response. We predicted that minnows initially conditioned with a high concentration of pike odor should respond to high pikeodor concentration with an equal intensity as those conditioned with a low concentration and tested with a

low concentration. Moreover, minnows conditioned with a particular concentration of pike odor should subsequently respond with a higher intensity response to higher concentrations of pike odor and with a lower intensity response to lower concentrations of pike odor.

Materials and methods Test fish Fathead minnows were captured from Feedlot Pond, located in the University of Saskatchewan campus, using minnow traps in September 2004. This population of minnows originated from the South Saskatchewan River when the pond was filled in 1959 to provide water for agricultural purposes. Intensive trapping and gill netting over the past 10 years revealed the presence of only one other fish species in the pond, brook stickleback (Culaea inconstans), and the absence of any predatory fish. Moreover, previous experiments (e.g., Chivers and Smith 1995; Brown et al. 1997) have demonstrated that minnows from this pond and nearby ponds do not show innate recognition to pike cues. The minnows were housed in an 18,000-l flow-through pool filled with dechlorinated tap water at 12°C and fed ad libitum once a day with commercial flakes. The photoperiod was adjusted to 14:10 h light:dark cycle. Northern pike were captured in September 2004 from Eagle Creek, Saskatchewan, Canada, using seine nets. They were housed in a 6,000-l flowthrough pool filled with dechlorinated tap water at 12°C. Pike were fed ad libitum with live fathead minnows. We ensured that some minnows were always present in the pool to avoid pike cannibalism. Stimulus collection Minnow skin extract We collected skin extract from six fathead minnows (fork length: mean±SD=4.76±0.43 cm). Minnows were killed with a blow on the head (in accordance with the Canadian Council on Animal Care) and skin fillets were removed from either side of the body and placed in 100 ml of chilled distilled water. Skin fillets were then homogenized and filtered through glass wool to remove any remaining tissues. We collected a total of 19.14 cm2 of skin in a total of 383 ml of distilled water, which constituted our standardized stock solution containing ∼1 cm2 of skin per 20 ml of distilled water. This solution was diluted to make the experimental solution containing ∼1 cm2 of skin per 40 l. We chose this concentration as it has already been shown to elicit an overt antipredator response in fathead minnows (Ferrari et al. 2005). Skin extracts were frozen in 20 ml aliquots at −20°C until required.

524

Pike odor Prey animals often exhibit antipredator responses to chemical cues of predators fed with conspecifics of the prey but not those fed with another diet (Chivers and Mirza 2001). Thus, two arbitrarily chosen pike (20.0 and 22.0 cm standard length) were kept independently in two 145-l tanks for 7 days before stimulus collection and fed with two convict cichlids (Archocentrus nigrofasciatus) (standard length: mean±SD=3.93±0.64 cm) once at day 2 and once at day 5 of the 7-day period. Both pike were then rinsed and transferred to a single 72-l tank containing 60 l of dechlorinated tap water and equipped with an air stone but no filter. The pike remained in the stimulus collection tank for 24 h, at which time they were transferred back to their initial holding pool. Water containing pike odor was frozen at −20°C in 60 ml aliquots until required.

to inject test stimuli into the tanks. The tanks were also equipped with a shelter that consisted of a 10×20-cm ceramic tile mounted on three 3.5 cm long cylindrical glass legs. Before conditioning, minnows were acclimated for a 24-h period in their conditioning tanks. We fed the fish after their transfer into the tank and an hour before testing (to reduce the potential trade-off between foraging and predator avoidance—Brown and Smith 1996). We tested 30 minnows for each of the four treatments (fork length: mean±SD=5.75±0.07 cm). Before the stimuli injection, we withdrew and discarded 60 ml of water from the injection tubes (to remove any stagnant water) and then withdrew and retained an additional 60 ml. We then injected 5 ml of conspecific skin extract immediately followed by 60 ml of either high or low concentration of pike odor into the tank. We used the retained tank water to slowly flush the stimuli into the tank. The injection of high or low concentration of pike odor was randomized among the tanks.

Experimental protocol The experiment consisted of a two by two design. Minnows were conditioned with conspecific alarm cues paired with either a high or low concentration of pike odor. During subsequent recognition trials, minnows from each group were tested for a response to either a high or low concentration of pike odor. Thus, we had four treatments: minnows conditioned with skin extract paired with low concentration of pike odor and tested with low concentration of pike odor (LL), minnows conditioned with skin extract paired with low concentration of pike odor and tested with high concentration of pike odor (LH), minnows conditioned with skin extract paired with high concentration of pike odor and tested with low concentration of pike odor (HL), and finally minnows conditioned with skin extract paired with high concentration of pike odor and tested with high concentration of pike odor (HH). To obtain the high and low concentrations of pike odor, we used 60ml syringes to inject either 20 ml (low) or 60 ml (high) of the prepared pike odor. To control for volumes injected into the tanks, the syringe containing 20 ml of pike odor was filled with an additional 40 ml of tank water. Consequently, our high concentration stimulus injected into the tanks was three times as concentrated as our low concentration of pike odor. Fathead minnows were held in a different room than the experimental room. To allow the fish to acclimate to the new room temperature (water temperature at 14–15°C, same photoperiod as mentioned above), groups of about 20 minnows were transferred into a 145-l holding tank located in the experimental room, 1 week before being tested. Conditioning trials Individual minnows were placed in 37-l tanks filled with dechlorinated tap water. The tanks contained a gravel substrate and were equipped with an air stone, near to which we attached a 2-m long piece of plastic tubing used

Recognition trials An hour after the end of the conditioning trials, the fish were moved to similar 37-l tanks filled with clean dechlorinated tap water. After 24 h, recognition trials were performed. Observations consisted of an 8-min preand an 8-min poststimulus injection period. After the prestimulus period, we injected 60 ml of either high or low concentration of pike odor into the tank using the same procedure as used in the conditioning trials. Once the stimuli were fully injected, we began the poststimulus observation period. The most common antipredator response displayed by minnows exposed to predator cues is shelter use, if a shelter is present, or a decrease in activity, if a shelter is not available (Chivers and Smith 1998). When hiding under shelter, fish might still be ‘active’, displaying foraging behavior for example, making the activity measure less powerful when a shelter is present. In this study, we measured time spent under shelter as the primary antipredator response variable, but also recorded time spent moving as a secondary variable. Time under shelter and time spent moving were recorded during the 8-min pre- and 8-min postinjection periods. The experimenter was blind to the treatments when recording fish behaviors. In a recent experiment, Ferrari et al. (2005) demonstrated that the intensity of response of naïve minnows to novel predator odors combined with alarm cues is determined primarily by alarm cue concentration. They also demonstrated that the intensity of antipredator response in the conditioning trials matched the intensity of response during the recognition trials when fish were exposed to the same concentration of predator odor. In this experiment, we hypothesized that minnows do not innately recognize highconcentration vs low-concentration predator odors (Chivers and Smith 1994a). Consequently, we predicted that fish conditioned with alarm cues to recognize high concentrations of pike odor during the conditioning trials and

525

Statistical analysis For all behavioral measures, we calculated the change from the prestimulus baseline. An increase in shelter use or a reduction in time spent moving would indicate an increase in antipredator behavior (review Chivers and Smith 1998). The data for change in time spent moving and shelter use were parametric (normally distributed with equal variances among treatments), thus we performed a 2×2 ANOVA for each of the variables, followed by post hoc Bonferonni tests for our three preplanned comparisons.

Results The results of the ANOVA showed a significant effect of conditioning and testing concentrations on both shelter use (for conditioning, F(1,113)=4.86, P=0.029; for testing, F(1,113)=19.94, P<0.001) and time spent moving (for conditioning, F(1,113)=18.55, P<0.001; for testing, F(1,113)=13.59, P<0.001). However, no interaction between conditioning and recognition concentrations was found for either shelter use (F(1,113)=0.053, P=0.819) or time spent moving (F(1,113)=1.25, P=0.265). The results of the post hoc comparisons showed that minnows conditioned with skin extract paired with a low concentration of pike odor and tested with a low concentration of pike odor (LL) did not respond with a different intensity than minnows conditioned with skin extract paired with a high concentration of pike odor and tested with a high concentration of pike odor (HH) for change in shelter use (Diff=−63.9, P=0.676; Fig. 1a) or time spent moving (Diff=−16.5, P=1.000; Fig. 1b). Minnows conditioned with skin extract paired with a low concentration of pike odor and tested with a low concentration of pike odor (LL) responded with a lower intensity than minnows conditioned with skin extract paired with a low concentration of pike odor and tested with a high concentration of pike odor (LH) for change in shelter use (Diff=−119.8, P=0.020). However, we did not find a significant difference for time spent moving (Diff=68.3, P=0.432).

a Change in shelter use

300 250 200 150 100 50 0 LL

LH

HL

HH

LL

LH

HL

HH

b 50 Change in time spent moving

exposed to a high concentration of pike odor during the recognition trials should respond with an equal intensity response as fish conditioned with a low concentration of pike odor and exposed to low concentration of pike odor during recognition trials. We tested this by comparing the intensity of response of the minnows in the LL vs HH treatments. We then compared the LL vs LH treatments to determine the effect of an increase in predator cue concentration between conditioning and recognition trials (predator cues represent a higher threat in recognition trials than conditioning trials). Finally, we compared the HH vs HL treatments to determine the effect of a decrease in predator cue concentration between conditioning and recognition trials (predator cues represent a lower threat during recognition trials than conditioning trials).

0 -50 -100 -150 -200 -250 -300 -350

Fig. 1 Mean (±SE) change in a shelter use and b time spent moving for minnows conditioned with skin extract paired with low concentration of pike odor and tested with low concentration of pike odor (LL), minnows conditioned with skin extract paired with low concentration of pike odor and tested with high concentration of pike odor (LH), minnows conditioned with skin extract paired with high concentration of pike odor and tested with low concentration of pike odor (HL), and finally minnows conditioned with skin extract paired with high concentration of pike odor and tested with high concentration of pike odor (HH)

Minnows conditioned with skin extract paired with a high concentration of pike odor and tested with a low concentration of pike odor (HL) respond with a lower intensity than minnows conditioned with skin extract paired with a high concentration of pike odor and tested with a high concentration of pike odor (HH) for change in shelter use (Diff=−132.8, P=0.007) and time spent moving (Diff=127.8, P=0.006). It is important to note that we used parametric tests because our data were normally distributed and the variances among treatments were equal. From this, we can also conclude that the change in observed antipredator intensities are due to individual fish changing their response intensity and not a difference in the proportion of fish responding. Indeed, this would have led to different variances among treatments.

Discussion Our results confirm earlier studies in that minnows learn to recognize predator odor based on a single conditioning event (Chivers and Smith 1998). More interesting is the fact that minnows acquire recognition in a threat-sensitive manner. When conditioned with skin extract and a given

526

concentration of pike odor, minnows increased the intensity of their antipredator response when subsequently exposed to higher concentrations of pike odor and decreased their intensity of response when subsequently exposed to lower concentrations of pike odor. It is interesting to note that no statistical difference was found between LL and HH treatments. Our experiment suggests that the initial intensity of response to a novel predation threat is determined primarily by the concentration of alarm cues and not the concentration of predator odor. We acknowledge the fact that only one concentration of skin extract was used in our experiment, limiting our power of conclusion with this data set only. However, Ferrari et al. (2005) demonstrated that the intensity of response of naïve minnows to novel predator odors paired with alarm cues is determined primarily by alarm cues concentration, and the intensity of antipredator response in the conditioning trials matched the intensity of response during the recognition trials when fish were exposed to the same concentration of predator odor. Taken with the results of this experiment, we can conclude that the concentration of unknown predator odor does not affect the intensity of response of naïve fish to novel predator odors because no difference was found between the LL and HH treatments. In addition, no interaction was found between the concentration of predator odor used during conditioning and testing. If predator odor concentration was driving the intensity of response, we would have seen an interaction between conditioning and recognition factors. Put together, this strongly supports the fact that naïve minnows do not have a priori knowledge that higher concentrations of novel predator odors represent a bigger threat and that they respond to change in relative concentration rather than the actual concentration of predator odors. Our results extend those of Ferrari et al. (2005). In their study, minnows learned the identity of the predator as well as the concentration of conspecific alarm cues used during conditioning. In this study, we showed that minnows also learned the concentration of predator odor they are exposed to during conditioning. Learning the identity of potential predators is essential for prey species with no innate recognition of their predators. Such learning mechanisms should be as efficient as possible, i.e., allowing the prey to get as much information as possible at once. During learning through conditioning by skin extract paired with predator odor, the intensity of the fright response is linked to the concentrations of each of the stimuli. During social learning, Ferrari et al. (2005) showed that there was a positive correlation between the intensity of response displayed by the ‘tutors’ and the intensity of response displayed by the ‘observers’ during subsequent exposures to the predator odor. Once again, fish learn more than the identity of the predator; they also learn the intensity associated with their fright response. Helfman’s (1989) threat-sensitive predator avoidance hypothesis pointed out the importance of displaying an adaptive intensity of antipredator response. Knowing what intensity of response to display when exposed to predator cues is almost as important as being able to identify the

predator. For obvious reasons, effective predator avoidance is under strong selection. However, prey are also under selection to avoid overresponding to predation threat. Time spent hiding, freezing, and dashing is time spent away from fitness-related activities such as foraging and reproducing. Thus, prey should respond to predation threats in a threatsensitive manner to optimize the trade-off between antipredator behavior and other fitness-related activities. In fact, Ferrari and Chivers (2006) pointed out the importance of continuous learning and the priority given to recent information on the ability of prey fishes like minnows to stay up-to-date regarding a predation threat and avoid perceiving a neutral nondangerous stimulus as a threat. Continuous learning seems to be the key to display threatsensitive behaviors. Helfman (1989) showed that threespot damselfish (Stegastes planifrons) exposed to a predatory model increased the intensity of their antipredator response when the model was larger, closer, or in a strike pose. Similarly, cueing on chemical concentrations to respond to a predation threat can be seen as an adaptive way to assess the level of risk a prey is exposed to. Indeed, stronger concentrations of predator odors can be indicative of the close proximity of the predator or a larger number of predators (Ferrari et al. in press). Moreover, prey living in complex or murky habitats should rely heavily on chemical cues because visual cues are limited in these types of environments. However, prey living in moving water might deal with chemical cues in a slightly different way. Depending on the type of current, odors are more or less diluted as they travel downstream. Moreover, light counter currents might contain low concentrations of the odor of a predator located in close proximity downstream (Dahl et al. 1998). For these reasons, it might be possible that prey living in different habitats rely differently on chemical cue concentrations to which they are exposed to assess the risk of predation. When detecting predator odors, prey species can use the concentration to assess the risk they are exposed to, but predator odors sometimes contain information like size of the predators (Kusch et al. 2004), and proximity and density of predators (Ferrari et al. in press). Much information about predators is acquired by prey through exposure to predator odor. It would be interesting to know what else prey fish can learn from a ‘simple’ odor. The development of threat-sensitive predator avoidance deserves more work in a diversity of predator/prey systems so that we get a complete understanding of how prey deal with the dynamic trade-offs between the costs and benefits of antipredator behaviors. Acknowledgements This work is dedicated to the memory of Dr. Joseph A. Brown who passed away unexpectedly on September 4, 2005. Dr. Brown was a pioneer in fish behavioral ecology. His contribution to science will be sorely missed. The Natural Sciences and Engineering Research Council of Canada and the University of Saskatchewan provided financial support to D. P. Chivers. All work reported herein was in accordance with the University of Saskatchewan Committee of Animal Care and Supply protocol #19920077.

527

References Amo L, Lopez P, Martin J (2004) Wall lizards combine chemical and visual cues of ambush snake predators to avoid overestimating risk inside refuges. Anim Behav 67:647–653 Anholt BR, Skelly DK, Werner EE (1996) Factors modifying antipredator behavior in larval toads. Herpetologica 52:301–313 Berejikian BA, Tezaka EP, LaRaeb AL (2003) Innate and enhanced predator recognition in hatchery-reared chinook salmon. Environ Biol Fishes 67:241–251 Brown GE, Smith RJF (1996) Foraging trade-offs in fathead minnows (Pimephales promelas, Osteichthyes, Cyprinidae): acquired predator recognition in the absence of an alarm response. Ethology 102:776–785 Brown GE, Chivers DP, Smith RJF (1997) Differential learning rates of chemical versus visual cues from a northern pike by fathead minnows in a natural habitat. Environ Biol Fishes 49:89–96 Chivers DP, Mirza RS (2001) Predator diet cues and the assessment of predation risk by aquatic vertebrates: a review and prospectus. In: Marchlewska-Koj DA, Lepri JJ, MüllerSchwarze D (eds) Chemical signals in vertebrates, vol. 9. Plenum, New York, pp 277–284 Chivers DP, Smith RJF (1994a) The role of experience and chemical alarm signalling in predator recognition by fathead minnows, Pimephales promelas. J Fish Biol 44:273–285 Chivers DP, Smith RJF (1994b) Fathead minnows, Pimephales promelas, acquire predator recognition when alarm substance is associated with the sight of unfamiliar fish. Anim Behav 48:597–605 Chivers DP, Smith RJF (1995) Free-living fathead minnows rapidly learn to recognize pike as predators. J Fish Biol 46:949–954 Chivers DP, Smith RJF (1998) Chemical alarm signalling in aquatic predator–prey systems: a review and prospectus. Ecoscience 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 Dahl J, Nilsson PA, Petterson LB (1998) Against the flow: chemical detection of downstream predators in running waters. Proc R Soc Lond B Biol Sci 265:1339–1344 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) The role of learning in the development of threat-sensitive predator avoidance: how do fathead minnows incorporate conflicting information? Anim Behav 71:19–26 Ferrari MCO, Messier F, Chivers DP (in press) The nose knows: minnows determine predator proximity and density through detection of predator odours. Anim Behav 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 Golub JL, Brown GE (2003) Are all signals the same? Ontogenetic change in the response to conspecific and heterospecific chemical alarm signals by juvenile green sunfish (Lepomis cyanellus). Behav Ecol Sociobiol 54:113–118 Hartman EJ, Abrahams MV (2000) Sensory compensation and the detection of predators: the interaction between chemical and visual information. Proc R Soc Lond B Biol Sci 267:571–575 Helfman GS (1989) Threat-sensitive predator avoidance in damselfish–trumpetfish interactions. Behav Ecol Sociobiol 24:47–58

Holomuzki JR, Short TM (1990) Ontogenetic shifts in habitat use and activity in a stream-dwelling isopod. Holarct Ecol 13: 300–307 Kats LB, Breeding JA, Hanson KM, Smith P (1994) Ontogenetic change in California newts (Taricha torosa) in response to chemical cues from conspecific predators. J North Am Benthol Soc 13:321–325 Kelley JL, Magurran AE (2003) Learned predator recognition and antipredator responses in fishes. Fish Fish 4:216–226 Kusch RC, Mirza RS, Chivers DP (2004) Making sense of the predator scents: investigating the sophistication of predator assessment abilities of fathead minnows. Behav Ecol Sociobiol 55:551–555 Lima SL, Dill LM (1990) Behavioral decisions made under the risk of predation: a review and prospectus. Can J Zool 68:619–640 Magurran AE (1989) Acquired recognition of predator odour in the European minnows (Phoxinus phoxinus). Ethology 82:216–233 Marcus J, Brown GE (2003) Response of pumpkinseed sunfish to conspecific chemical alarm cues: an interaction between ontogeny and stimulus concentration. Can J Zool 81: 1671–1677 Mathis A, Smith RJF (1993) Chemical alarm signals increase the survival time in fathead minnows (Pimephales promelas) during encounters with northern pike (Esox lucius). Behav Ecol 4:260–265 Mathis A, Vincent F (2000) Differential use of visual and chemical cues in predator recognition and threat-sensitive predator-avoidance responses by larval newts (Notophthalmus viridescens). Can J Zool 78:1646–1652 Mathis A, Chivers DP, Smith RJF (1993) Population differences in responses of fathead minnows (Pimephales promelas) to visual and chemical stimuli from predators. Ethology 93:31–40 McIntosh AR, Peckarsky BL, Taylor BW (1999) Rapid size-specific changes in the drift of Baetis bicaudatus (Ephemeroptera) caused by alteration in fish odour concentration. Oecologia 118:256–264 Puttlitz MH, Chivers DP, Kiesecker JM, Blaustein AR (1999) Threat-sensitive predator avoidance by larval Pacific treefrogs (Amphibia, Hylidae). Ethology 105:449–456 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, New York, pp 475–488 Vilhunen S, Hirvonen H (2003) Innate antipredator responses of Arctic charr (Salvelinus alpinus) depend on predator species and their diet. Behav Ecol Sociobiol 55:1–10 Wahle RA (1992) Body-size dependent anti-predator mechanisms of the American lobster. Oikos 65:52–60 Williams PJ, Brown JA (1991) Developmental changes in foraging —predator avoidance trade-offs in larval lumpfish, Cyclopterus lumpus. Mar Ecol Prog Ser 76:53–60 Wisenden BD (2003) Chemically-mediated strategies to counter predation. In: Collin SP, Marshall NJ (eds) Sensory processing in the aquatic environment. Springer, Berlin Heidelberg New York, pp 236–251 Zhao X, Chivers DP (2005) Response of juvenile goldfish (Carassius auratus) to chemical cues: relationship between response intensity, response duration and the level of predation risk. In: Mason RT, LeMaster M, Müller-Scharze D (eds) Chemical signals in vertebrates, vol 10. Plenum, New York, pp 334–341