Behav Ecol Sociobiol DOI 10.1007/s00265-009-0870-y

ORIGINAL PAPER

The ghost of predation future: threat-sensitive and temporal assessment of risk by embryonic woodfrogs Maud C. O. Ferrari & Douglas P. Chivers

Received: 11 May 2009 / Revised: 31 August 2009 / Accepted: 17 September 2009 # Springer-Verlag 2009

Abstract Amphibians are able to learn to recognize their future predators during their embryonic development (the ghost of predation future). Here, we investigate whether amphibian embryos can also acquire additional information about their future predators, such as the level of threat associated with them and the time of day at which they would be the most dangerous. We exposed woodfrog embryos (Rana sylvatica) to different concentrations of injured tadpole cues paired with the odor of a tiger salamander (Ambystoma tigrinum) between 1500 and 1700 hours for five consecutive days and raised them for 9 days after hatching. First, we showed that embryos exposed to predator odor paired with increasing concentrations of injured cues during their embryonic development subsequently display stronger antipredator responses to the salamander as tadpoles, thereby demonstrating threatsensitive learning by embryonic amphibians. Second, we showed that the learned responses of tadpoles were stronger when the tadpoles were exposed to salamander odor between 1500 and 1700 hours, the time at which the embryos were exposed to the salamander, than during earlier (1100–1300 hours) or later (1900–2100 hours) periods. Our results highlight the amazing sophistication of learned predator recognition by prey and emphasize the importance of temporal considerations in experiments examining risk assessment by prey. Communicated by J. Christensen-Dalsgaard M. C. O. Ferrari (*) Department of Environmental Science and Policy, University of California, Davis, CA 95616, USA e-mail: [email protected] D. P. Chivers Department of Biology, University of Saskatchewan, Saskatoon, SK S7N 5E2, Canada

Keywords Predator recognition . Learning . Embryo . Woodfrog Rana sylvatica

Introduction Predation is a strong selection pressure acting on the behavior, morphology, and life-history of prey species (Lima and Dill 1990; Brown 2003). A prerequisite for prey to adaptively respond to predators is for them to be able to recognize predators from nonpredators. While many species require learning to recognize predators as threatening, a number of prey species, including birds (Goth 2001), amphibians (Kiesecker and Blaustein 1997), and fishes (Berejikian et al. 2003), have been shown to display antipredator responses towards certain predatory species prior to any experience with them. Species possessing such “innate” predator recognition may be at a selective advantage, especially during their first encounter with a predator. The prey does not need the first encounter to learn that the predator represents a threat. However, not all species possess this ability, indicating either (1) a cost to genetically fixing the recognition of a specific predator and/ or (2) the need for a long and stable evolutionary history between the prey and the predator to allow genetic fixation of the recognition (Ferrari et al. 2007a). Prey living in environments where the predator composition is rapidly changing (e.g., as would occur in ephemeral ponds) will be exposed to a variety of species, many of which are potential predators. In such situations, selection for innate recognition may be limited, and hence, prey living in these types of environments usually requires learning to recognize predators. Until recently, learned predator recognition was thought to allow the recognition of one predatory species at a time, but recent work has indicated that prey may have wide recognition templates, allowing them to

Behav Ecol Sociobiol

learn one particular species and generalize their recognition to closely related species of that predator (Griffin et al. 2001; Ferrari et al. 2007a, 2008a). Hence, these results dramatically changed our understanding of prey's perception of potential predators. Our understanding of learned predator recognition was recently pushed a step further, with the first documentation that prey animals may acquire predator recognition as embryos. In that study, ringed salamanders, Ambystoma annulatum, and woodfrogs, Rana sylvatica, learn to recognize their future predators while still in the egg (the ghost of predation future; Mathis et al. 2008). While much of the research on embryonic learning has tested whether pre-exposure to a stimulus (taste or smell) subsequently results in a preference of the individual for this stimulus as adult (Hepper and Waldman 1992; Sneddon et al. 1998), Mathis et al. (2008) were the first to show learned predator recognition at such an early developmental stage. Essentially, amphibians that lack innate predator recognition can hatch and still avoid predators upon their first “physical” encounter with them. While predator recognition is a prerequisite for prey to respond to predators, recognition by itself is not sufficient to allow prey to exhibit threat-sensitive tradeoffs between costly predator avoidance and other fitness-related activities such as foraging or mating. The ability of prey to display antipredator response intensities matching the risk posed by the predators is known as threat-sensitive predator avoidance (Helfman 1989; Anholt et al. 1996; Chivers et al. 2001; Van Buskirk and Arioli 2002; Relyea 2003; Ferrari et al. 2006). Many prey species have been shown to make sophisticated adjustments in the intensity of their antipredator response according to the predator's size (Helfman 1989; Anholt et al. 1996; Kusch et al. 2004), the predators' proximity and density (Ferrari et al. 2006), the predator's posture (Helfman 1989), the predator's diet (Chivers and Mirza 2001b; Kiesecker et al. 2002; Chivers et al. 2007), the overall risk associated with the predator (Peacor and Werner 1997; Van Buskirk et al. 2002; Ferrari et al. 2005), and the time at which the predator is more active and foraging (Sih and McCarthy 2002; Sullivan et al. 2005; Ferrari et al. 2008b; Ferrari and Chivers 2009) While woodfrogs have been shown to learn to recognize their predators as embryos, we asked whether they could also learn the risk posed by the predator during their embryonic development (i.e., the ghost of predation future). While many modes of learning are available for prey to learn to recognize predators, a widespread and well-studied mode of learning for aquatic species is through the pairing of the novel predator cues (odor or sight) with injured conspecific cues (Mathis and Smith 1993; Chivers and Smith 1998). An amazing characteristic of this learning mode is that one pairing event is enough for prey to acquire

the recognition of the predator. Recent studies have shown that prey could learn to recognize the novel predator as a threat and could also adjust their responses to the predator according to the concentration of injured conspecific cues experienced during the learning phase. Specifically, prey conditioned with predator odors paired with a high concentration of injured conspecific cues will display stronger antipredator behaviors when subsequently exposed to the predator compared to prey conditioned with lower concentrations of injured cues. This threat-sensitive learning phenomenon has been demonstrated in fishes (Ferrari et al. 2005; Zhao et al. 2006), larval amphibians (Ferrari and Chivers 2009), and larval mosquitoes (Kesavaraju et al. 2007; Ferrari et al. 2008c). To test whether embryonic woodfrogs could show such threat-sensitive learning as embryos, we conditioned groups of embryos with the odor of a tiger salamander (Ambystoma tigrinum) paired with different concentrations of injured tadpole cues and subsequently tested the larvae for their responses to the predator. We identified variation in the responses of tadpoles to predators at different times of day and hence, further explored whether the intensity of their responses varied at different times of day.

Methods Water and predators Ten days prior to starting the experiments, a 1,900-l tub was filled with well water and left outdoors. The tub was seeded with zooplankton and phytoplankton from a local pond using a fine mesh dip net. This ensured that the holding and testing water contained a full array of algae and plankton but that no salamander cues were present in the water. A tiger salamander (snout-vent length 14.1 cm) was obtained from a commercial supplier and housed in the Biology Department Animal Care Unit at the University of Saskatchewan. The salamander was transported to the field site and maintained on a diet of earthworms. Test subjects and stimulus preparation All eggs and tadpoles used in this experiment were collected from a single pond in Central Alberta in May 2008. Our previous work has established that woodfrog tadpoles from this pond do not show recognition of predators without prior experience (Ferrari et al. 2007c; Ferrari and Chivers 2008). Egg laying began on May 2, 2008 and was completed on May 11, 2008. Four-egg clutches laid early in the season were transferred into a pool containing pond water and aquatic plants and were left to develop until hatching. This was done to ensure that

Behav Ecol Sociobiol

tadpoles were available for the stimulus preparation. The tadpole stimulus was prepared daily immediately prior to being used, by crushing 50 tadpoles (approximately 8– 12 mm in length) in 100 ml of well water using a mortar and pestle. This stimulus was run through filter floss and divided into aliquots for injection into the holding buckets (see below). The salamander odor used for the egg treatment phase was obtained by putting the salamander into 500 ml of conditioned well water for 24 h. The water was changed every day.

immediately upon hatching. The treatments stopped after day 5. Embryos started hatching the following morning, and all the embryos were hatched by the evening. Tadpoles were provided with rabbit food, and the water was partially changed every second day. Tadpoles were raised for 9 days to Gosner stage 25 (Gosner 1960) and subsequently tested for a response to salamander odor. Experiment 1: Threat-sensitive learning of predator recognition by embryonic woodfrogs

Five egg clutches laid the previous night were collected, and each clutch was divided into five groups (or subclutches) of approximately 50–60 eggs. Each subclutch was then transferred into a 3.5-l bucket filled with 3 l of conditioned well water. The subclutches consisted of a single mass of eggs with the egg jelly intact. Temperature checks done twice daily revealed only a 0.1°C difference between the 25 buckets. Examinations of eggs revealed that the five clutches were at Gosner developmental stage 10–11. At this stage, the neural tube is not yet formed.

The goal of this experiment was to determine whether the conditioning treatment received as embryos would affect the level of responses that tadpoles had learned to display in response to the predator odor. Tadpoles from each of the five treatments were tested for their responses to salamander odor. We predicted that tadpoles that did not experience the odor of the predator paired with injured conspecific cues (treatments WW and W + SO) as embryos would fail to recognize the salamander as a threat (Mathis et al. 2008). Furthermore, we predicted that the concentration of injured cues experienced as embryos would subsequently affect the intensity of the antipredator response that tadpoles display to the predator. Specifically, we predict that the higher the concentration of injured cues, the stronger the learned response to the predator.

Experimental procedure

Behavioral assay

The layout of the experiment followed a randomized block design with replication, in which each clutch represents a block. One of the five subclutches from each clutch was randomly assigned to one of five treatments and treated with the following stimuli: (1) WW: 32 ml of well water, (2) W + SO: 12 ml of well water paired with 20 ml of salamander odor, (3) 1TP + SO: 2 ml of crushed tadpole solution (equivalent to one crushed tadpole), 10 ml of well water, and 20 ml of salamander odor, (4) 3TP + SO: 6 ml of crushed tadpole solution (equivalent to three crushed tadpoles), 6 ml of well water, and 20 ml of salamander odor, and (5) 6TP + SO: 12 ml of crushed tadpole solution (equivalent to six crushed tadpoles) paired with 20 ml of salamander odor. Additional water injections were made to ensure that each treatment group received the same volume of 32 ml. Eggs were treated daily at 1500 hours for 5 days (days 1 to 5); the stimuli were slowly injected on the side of the buckets to minimize disturbance to the eggs. At 1700 hours, a 100% water change was performed. The experimenter was wearing latex gloves to avoid the transfer of any odor to the embryos. Eggs were treated until the embryo within the eggs appeared fully formed (approximately Gosner stage 22/24) but had not hatched. We verified that the embryos had not hatched as they were curled up inside their egg shell. Tadpoles straighten out

Tadpoles were tested using a well established protocol (Chivers and Mirza 2001a; Ferrari et al. 2007b, c; Mathis et al. 2008, Ferrari et al. 2008b). Individual tadpoles were transferred into plastic cups containing 0.5 l of conditioned well water and left to acclimate for 45 min. The trials consisted of a 4-min prestimulus followed by a 4-min poststimulus injection period during which the behavior (activity) of the tadpoles was recorded. The two periods were separated by a 30-sec injection period, during which the content of a 5-ml syringe was emptied slowly on the side of the cup to minimize disturbance. All tadpoles were exposed to 5 ml of salamander odor. The salamander odor used was obtained by putting the salamander into 2 l of conditioned well water for 24 h. While a single salamander was used in the present experiment, previous work suggests that the defensive response elicited in tadpoles by salamander predators is not unique to the individual predator being examined (Ferrari and Chivers 2008, 2009). Tadpoles have been shown to decrease activity in response to predation cues. Hence, a line was drawn in the middle of the cup, and the number of line crosses was counted during the two observation periods. We considered that a tadpole crossed a line when its entire body was on the other side of the line. The trials (N = 30 per treatment) were

Experimental setup

Mean proportion change (+/- SE) in line crosses

Behav Ecol Sociobiol c

0.2 c

0 -0.2 b

-0.4 a

a

-0.6 6TP+SO

3TP+SO

1TP+SO

W+SO

WW

Fig. 1 Mean proportion change (±SE) in line crosses from the prestimulus baseline for tadpoles exposed to salamander odor. Tadpoles were exposed as embryos to well water (WW) or to different concentration of injured tadpoles cues (1TP, 3TP, or 6TP) or well water (W) paired with salamander odor (SO). Different letters indicate statistical differences at a 0.05 level

performed outdoors, between 1430 and 1830 hours. The order of the treatments was randomized throughout the day. The observer was blind to the treatment. Statistical analysis We calculated the change in proportion of line crosses from the prestimulus baseline. The data was normally distributed, and the variances were homogeneous among treatments. Hence, we analyzed the data using a randomized block design analysis of variance (ANOVA) with replication, in which we assessed the effect of treatment (fixed factor) and clutch (block). Significant effects of treatments were further analyzed using post hoc Tukey's tests. Experiment 2: Temporal learning of predation risk by embryonic woodfrogs When looking at the data of experiment 1, we noticed that the intensity of the antipredator responses were variable throughout the day, with the response intensities tending to be lower at the end of the testing period. Hence, we decided to test whether there was a temporal aspect of the tadpole's responses to predator cues. If so, we predicted that the strongest response to the Table 1 Table of the nonparametric ANOVA performed on the three testing periods

P values with significant difference were set to bold

df

Conditioning Cue Clutch Conditioning×cue Conditioning×clutch Cue×clutch Conditioning×cue× clutch

1,4 2,8 4,8 2,8 4,8 8,8 8,93

salamander cues would match the time of day at which the embryos experienced the odor of the salamander paired with injured conspecific cues (i.e., 1500–1700 hours). Tadpoles were tested for their response to salamander odor at three testing periods: 1100–1300 hours (2–4 h prior to embryonic conditioning time), 1500–1700 hours (embryonic conditioning time) and 1900–2100 hours (2–4 h after embryonic conditioning time). Due to the short testing time (2-hr window for each period), we had to select a low number of experimental groups to be tested to avoid having the experiment to go on for numerous days. Thus, we tested tadpoles from two experimental groups only: tadpoles from the 6TP + SO group represented our positive control (tadpoles having acquired recognition of the salamander as a threat) and tadpoles from the W + SO group represented our negative control (tadpoles not showing recognition of the salamander as a threat). At each time period, tadpoles from each treatment were tested for their response to well water (negative control for disturbance), injured conspecific cues (positive control for threatening stimuli; Ferrari et al. 2007b) and salamander odor. The behavioral assay used was identical to the one used in experiment 1. The salamander odor used in this experiment was made prior to the experiment and frozen to eliminate any concentration effects and to ensure that the “quality” of the salamander odor was constant throughout the day. A salamander was placed into 2 l of conditioned well water for 24 h. The conditioned water was then bagged in 20-ml aliquots and frozen at −20°C until required. This procedure was repeated twice. The stimulus bags were thawed 10 min prior to use (ensuring the stimulus was at ambient temperature before use). Injured conspecific cues were prepared fresh a few minutes before being used by crushing one tadpole in 5 ml of well water, using a mortar and pestle. Tadpoles were exposed to 5 ml of well water, injured conspecific cues or salamander odor. Data were collected over 4 days (N = 19–24 per treatment combination). Temperature checks over the 4 days indi-

1100–1300hours

1500–1700hours

1900–2100hours

F

Pvalue

F

Pvalue

F

Pvalue

0.1 98.5 1.6 1.4 1.1 0.5 0.9

0.93 <0.001 0.55 0.30 0.42 0.82 0.51

9.6 16 0.1 13.9 2.8 2.8 0.9

0.03 0.002 0.98 0.002 0.10 0.08 0.54

0.5 156.8 0.9 1.2 1.1 0.2 0.6

0.52 <0.001 0.85 0.35 0.43 0.99 0.15

Behav Ecol Sociobiol

Results Experiment 1 The ANOVA revealed a significant effect of treatment (F4, 16 = 19.2, P < 0.001), no effect of clutch (F4, 16 = 1.4, P > 0.2), and no interaction between the two factors (F16, 124 = 0.8, P > 0.6, Fig. 1). Tadpoles from the WW and W + SO treatments did not respond differently to the salamander odor (P > 0.7), but tadpoles from the 1TP + SO, 3TP + SO, and 6TP + SO treatments all responded with a stronger intensity to salamander odor than tadpoles from the WW treatment (all P < 0.05). Tadpoles from the 3TP + SO and 6TP + SO treatment responded stronger to the cue than tadpoles from the 1TP + SO treatment (P < 0.001). Tadpoles from the 6TP + SO treatment did not respond differently to the salamander cue than tadpoles from the 3TP + SO treatment (P > 0.9). Experiment 2 The nonparametric ANOVA revealed that both in the early (1100–1300 hours) and late (1900–2100 hours) testing periods, “cue” was the only significant factor (see Table 1, Fig. 2). During these periods, tadpoles responded to the injured conspecific cues with a greater intensity of response than to water or salamander odor, and tadpoles did not respond differently to water and salamander odor. Tests conducted between 1500 and 1700 hours, the time of embryonic conditioning, revealed that both “cue” and “conditioning” groups had a significant effect on the responses of tadpoles (see Table 1, Fig. 2). The responses of tadpoles from both the W + SO group and the 6TP + SO group were affected by “cue” (P < 0.001, P = 0.005,

Mean proportion change (+/- SE) in line crosses

We calculated the change in proportion of line crosses from the prestimulus baseline. The data was normally distributed, but the variances were not homogeneous among treatments. Hence, we analyzed the data using a nonparametric ANOVA approach (Scheirer–Ray–Hare extension of the Kruskal–Wallis test; Sokal and Rohlf 2003) using the rank value of the data. The effect of conditioning (W + SO and 6TP + SO—fixed factor), cue (water, injured cues, and salamander odor—fixed factor), and clutch (random factor) were assessed at each of the three testing periods.

0 -0.2 -0.4 -0.6 -0.8 Water

Salamander

Injured cues

Tadpoles tested between 1500 and 1700 hr Mean proportion change (+/- SE) in line crosses

Statistical analysis

Tadpoles tested between 1100 and 1300 hr

0 -0.2 -0.4 -0.6 -0.8 Water

Salamander

Injured cues

Tadpoles tested between 1900 and 2100 hr Mean proportion change (+/- SE) in line crosses

cated that the water temperatures during the testing periods ranged between 11–15°C for the 1100–1300-hour period, 15–20°C for the 1500–1700-hour period, and 14– 18°C for the 1900–2100-hour period.

0 -0.2 -0.4 -0.6 -0.8 Water

Salamander

Injured cues

Fig. 2 Mean proportion change (±SE) in line crosses from the prestimulus baseline for tadpoles exposed to water, salamander odor, or injured conspecific cues between 1100 and 1300 hours (top panel), between 1500 and 1700 hours (middle panel), or between 1900 and 2100 hours (bottom panel). Tadpoles were exposed as embryos to well water paired with salamander odor (open bars) or to the cues of six injured tadpoles paired with salamander odor (solid bars) between 1500 and 1700 hours

respectively) but not clutch (both P > 0.4, interaction term P > 0.05).

Discussion Mathis et al. (2008) provided the first evidence that prey animals can learn to recognize unknown predators during their embryonic development. In that experiment, the prey learned as embryos, hatched, and subsequently adjusted

Behav Ecol Sociobiol

their behavioral responses to predators as larvae. The present data indicates that this ability is much more sophisticated than the simple learning of an unknown predator. In our first experiment, we showed that embryonic woodfrogs can use the concentration of injured cues experienced to assess the level of risk that the predator poses and subsequently respond to the predator in a threatsensitive manner as larvae. This threat-sensitive learning ability has been demonstrated in some aquatic species; however, this is the first demonstration that such a seemingly complex phenomenon can be acquired at such an early age during the embryonic stage. The present results have important consequences for the dichotomy existing between learned and innate predator recognition in many amphibian species. This topic is particularly fascinating, as embryos of frogs and salamanders collected from certain populations occurring with predators seem to “innately” respond to the predator cues, whereas embryos collected from populations not occurring with predators do not (Kats et al. 1988; Kiesecker and Blaustein 1997). However, the present data and recent findings by Mathis et al. (2008) suggest that learned predator recognition occurs during embryonic development, questioning the existence of a true innate response to predator odors by larval amphibians. The results of our second experiment underscore the sophistication of learning, strongly suggesting that the time of day at which the embryos experienced the predator will subsequently affect the intensity of response that the larvae will display in response to the predator at different times of day. We acknowledge that further experiments are needed to fully validate this hypothesis. Specifically, groups of embryos need to be conditioned at different times of day, and the emergent larvae need to be subsequently tested at different times of day to provide full support for a plastic temporal adjustment of antipredator responses. Even though not demonstrating the plasticity of this temporal adjustment of learning, the present data at least demonstrate temporal adjustment of antipredator responses to conditioning occurring at midday. Our understanding of the importance of temporal variability of predation risk in mediating predator/prey interactions is slowly growing (Chivers et al. 2008). Lima and Bednekoff (1999) were the first to attempt to predict the effect of a prey's history of predation risk (in terms of intensity and frequency of risk) on its decision making through the risk allocation hypothesis. While empirical tests of the model have provided mixed support (Hamilton and Heithaus 2001; Sih and McCarthy 2002; Van Buskirk et al. 2002; Laurila et al. 2004; Ferrari et al. 2009), the work has emphasized the need for behavioral ecologists to incorporate temporal/predictable aspects of predation threat into studies of decision-making. Accordingly, Ferrari et al.

(2008b) and Ferrari and Chivers (2009) have demonstrated that woodfrog larvae have the ability to learn to adjust the intensity of their antipredator responses according to the time of day they experience the predators. Hence, all these results strengthen the need for researchers to take into account the temporal aspects of their experiments. A number of studies have reported amazingly sophisticated abilities of some prey species to deal with predators. However, these species share a common characteristic: all are “successful species.” For example, fathead minnows, woodfrogs, larval mosquitoes are widespread throughout their range, regardless of the predator community encountered. There is no doubt that their ability to deal with predators is linked to their success. However, very little is known about the ability of less successful species to deal with predators. Since we now know the extent to which some prey can adjust their responses to predators, comparative studies between successful species and species at risk (or reversely invasive species) would be helpful to highlight the specific characteristics that make a prey susceptible (or reversely resistant) to variable predation pressures. An area particularly ripe for investigation may be the study of behavioral phenotypes associated with the potential success of species to novel environmental conditions (e.g., Sih et al. (2004). Studies on amphibians may be particularly fruitful as one of the reasons put forward for the global decline of amphibian populations is their inability to cope with introduced competitors and predators (Blaustein and Kiesecker 2002; Blaustein and Bancroft 2007). Acknowledgments We thank Jean and Glen Chivers for their help and support and for letting us invade their wetlands for the duration of our field season. A special thank you goes to Jean, Christopher Chivers, and Trouble for their help and assistance in the field. Research funding was provided to F. Messier and D. Chivers through the Natural Sciences and Engineering Research Council of Canada. All work reported herein was in accordance with the Guidelines to the Care and Use of Experimental Animals published by the Canadian Council on Animal Care and was conducted under the University of Saskatchewan Committee of Animal Care and Supply protocol no. 20060014.

References Anholt BR, Skelly DK, Werner EE (1996) Factors modifying antipredator behavior in larval toads. Herpetologica 52:301– 313 Berejikian BA, Tezak EP, LaRae AL (2003) Innate and enhanced predator recognition in hatchery-reared chinook salmon. Environ Biol Fisches 67:241–251 Blaustein AR, Bancroft BA (2007) Amphibian population declines: evolutionary considerations. Bioscience 57:437–444 Blaustein AR, Kiesecker JM (2002) Complexity in conservation: lessons from the global decline of amphibian populations. Ecol Lett 5:597–608 Brown GE (2003) Learning about danger: chemical alarm cues and local risk assessment in prey fishes. Fish Fish 4:227–234

Behav Ecol Sociobiol Chivers DP, Mirza RS (2001a) Importance of predator diet cues in responses of larval wood frogs to fish and invertebrate predators. J Chem Ecol 27:45–51 Chivers DP, Mirza RS (2001b) Predator diet cues and the assessment of predation risk by aquatic vertebrates: a review and prospectus. In: Marchlewska DA, Lepri JJ, Müller-Schwarze D (eds) Chemical signals in vertebrates. Plenum Press, New York, pp 277–284 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 Chivers DP, Zhao X, Ferrari MCO (2007) Linking morphological and behavioural defences: prey fish detect the morphology of conspecifics in the odour signature of their predators. Ethology 113:733–739 Chivers DP, Zhao X, Brown GE, Marchant TA, Ferrari MCO (2008) Predator-induced changes in morphology of a prey fish: the effects of food level and temporal frequency of predation risk. Evol Ecol 22:561–574 Ferrari MCO, Chivers DP (2008) Cultural learning of predator recognition in mixed-species assemblages of frogs: the effect of tutor-to-observer ratio. Anim Behav 75:1921–1925 Ferrari MCO, Chivers DP (2009) Temporal variability, threatsensitivity and conflicting information about the nature of risk: understanding the dynamics of tadpole antipredator behaviour. Anim Behav 78:11–16 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 Ferrari MCO, Messier F, Chivers DP (2006) The nose knows: minnows determine predator proximity and density through detection of predator odours. Anim Behav 72:927–932 Ferrari MCO, Gonzalo A, Messier F, Chivers DP (2007a) Generalization of learned predator recognition: an experimental test and framework for future studies. Proc R Soc B 274:1853–1859 Ferrari MCO, Messier F, Chivers DP (2007b) Degradation of chemical alarm cues under natural conditions: risk assessment by larval woodfrogs. Chemoecology 17:263–266 Ferrari MCO, Messier F, Chivers DP (2007c) First documentation of cultural transmission of predator recognition by larval amphibians. Ethology 113:621–627 Ferrari MCO, Messier F, Chivers DP (2008a) Can prey exhibit threatsensitive generalization of predator recognition? Extending the predator recognition continuum hypothesis. Proc R Soc B 275:1811–1816 Ferrari MCO, Messier F, Chivers DP (2008b) Larval amphibians learn to match antipredator response intensity to temporal patterns of risk. Behav Ecol 19:980–983 Ferrari MCO, Messier F, Chivers DP (2008c) Threat-sensitive learning of predators by larval mosquitoes Culex restuans. Behav Ecol Sociobiol 62:1079–1083 Ferrari MCO, Sih A, Chivers DP (2009) The paradox of risk allocation: a review and prospectus. Anim Behav 78:579–585 Gosner KL (1960) A simplified table for staging anuran embryos and larvae with notes on identification. Herpetologica 16:183–190 Goth A (2001) Innate predator-recognition in Australian brush-turkey (Alectura lathami, Megapodiidae) hatchlings. Behaviour 138:117–136 Griffin AS, Evans CS, Blumstein DT (2001) Learning specificity in acquired predator recognition. Anim Behav 62:577–589

Hamilton IM, Heithaus MR (2001) The effects of temporal variation in predation risk on anti-predator behaviour: an empirical test using marine snails. Proc R Soc B 268:2585–2588 Helfman GS (1989) Threat-sensitive predator avoidance in damselfish– trumpetfish interactions. Behav Ecol Sociobiol 24:47–58 Hepper PG, Waldman B (1992) Embryonic olfactory learning in frogs. Q J Exp Psychol 44B:179–197 Kats LB, Petranka JW, Sih A (1988) Antipredator defenses and the persistence of amphibian larvae with fishes. Ecology 69:1865– 1870 Kesavaraju B, Damal K, Juliano SA (2007) Threat-sensitive behavioral responses to concentrations of water-borne cues from predation. Ethology 113:199–206 Kiesecker JM, Blaustein AR (1997) Population differences in responses of red-legged frogs (Rana aurora) to introduced bullfrogs. Ecology 78:1752–1760 Kiesecker JM, Chivers DP, Anderson M, Blaustein AR (2002) Effect of predator diet on life history shifts of red-legged frogs, Rana aurora. J Chem Ecol 28:1007–1015 Kusch RC, Mirza RS, Chivers DP (2004) Making sense of predator scents: investigating the sophistication of predator assessment abilities of fathead minnows. Behav Ecol Sociobiol 55:551–555 Laurila A, Jarvi-Laturi M, Pakkasmaa S, Merila J (2004) Temporal variation in predation risk: stage-dependency, graded responses and fitness costs in tadpole antipredator defences. Oikos 107:90–99 Lima SL, Bednekoff PA (1999) Temporal variation in danger drives antipredator behavior: the predation 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 Mathis A, Smith RJF (1993) Fathead minnows, Pimephales promelas, learn to recognize Northern pike, Esox lucius, as predators on the basis of chemical stimuli from minnows in the pike's diet. Anim Behav 46:645–656 Mathis A, Ferrari MCO, Windel N, Messier F, Chivers DP (2008) Learning by embryos and the ghost of predation future. Proc R Soc B 275:2603–2607 Peacor SD, Werner EE (1997) Trait-mediated indirect interactions in a simple aquatic food web. Ecology 78:1146–1156 Relyea RA (2003) How prey respond to combined predators: a review and an empirical test. Ecology 84:1827–1839 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, Bell A, Johnson JC (2004) Behavioral syndrome: an ecological and evolutionary overview. Trends Ecol Evol 19:372–378 Sneddon H, Hadden R, Hepper PG (1998) Chemosensory learning in the chicken embryo. Physiol Behav 64:133–139 Sokal RR, Rohlf FJ (2003) Biometry: the principles and practice of statistics in biological research, 3rd edn. Freeman and Co, New York Sullivan AM, Madison DM, Maerz JC (2005) Nocturnal shift in the antipredator response to predator-diet cues in laboratory and field trials. In: Mason RT, LeMaster MP, Müller-Schwartze D (eds) Chemical signals in vertebrates. Springer Verlag, New York, pp 349–356 Van Buskirk J, Arioli M (2002) Dosage response of an induced defense: how sensitive are tadpoles to predation risk? Ecology 83:1580–1585 Van Buskirk J, Muller C, Portmann A, Surbeck M (2002) A test of the risk allocation hypothesis: tadpole responses to temporal change in predation risk. Behav Ecol 13:526–530 Zhao X, Ferrari MCO, Chivers DP (2006) Threat-sensitive learning of predator odours by a prey fish. Behaviour 143:1103–1121