Olfactory cues of habitats facilitate learning about landscapes of fear

Olfactory cues of habitats facilitate learning about landscapes of fear Abstract Across landscapes, prey are exposed to different levels of predation risk within different habitats. However, little is known about how prey learn about risk in different habitat types. Here, we examined if wood frog tadpoles, Lithobates sylvatica, use olfactory cues from 2 distinct, plant-dominated habitats (cattail and pond weed) to learn about the overall risk within a habitat and the risk posed by a specific predator species within different habitats. In our first experiment, tadpoles experienced both a high-risk and a low-risk habitat before being tested for habitat-specific neophobic responses, a cognitive trait expressed in high-risk but not low-risk environments. In the second experiment, we taught tadpoles to recognize a predator in one habitat while the other one was never associated with a predator. Tadpoles were then tested for their responses to the predator and a control in both habitats. Our results showed that high-risk cattail tadpoles developed habitat-specific neophobia. However, high-risk pond weed tadpoles developed a generalized neophobia, responding to the novel cues irrespective of the habitat where they were tested. We also found that the habitat in which prey learned the identity of a specific predator did not affect their responses to that predator when tested in different habitats. Our results provide support for the use of olfactory habitat cues by prey to learn about predation risk across landscapes, suggesting unrecognized nuances to how prey use such cues to learn about predation risk. INTRODUCTION Predation is a major ecological process that shapes communities. Beyond consuming prey, predators can affect prey by changing their behavior, morphology, and life-history (Werner and Peacor 2003; Preisser et al. 2005). For instance, in the presence of predators, prey can invest in morphological changes that increase escape speed (e.g., tail depth [Dayton et al. 2005]) or become harder to consume (e.g., body size, amour thickness and spines (Tollrian 1993; Laforsch et al. 2004)). Some prey species may alter the timing of transitions between developmental stages to avoid predators, while others change the timing of their sexual maturity and reproduction (Abrams and Rowe 1996; Laurila et al. 1998). The trade-off between avoiding predators and investing in other activities such as foraging, growth, and reproductive output that enhance fitness means that, in many cases, prey must adjust their time and energetic investment in antipredator responses to match the level of threat they experience (Brown et al. 1999; Lima and Bednekoff 1999; Beauchamp and Ruxton 2011). However, as predation risk varies with both time and space, prey should adjust their antipredator responses to match the level of risk experienced in a given ecological context, such as specific habitats (Lima and Bednekoff 1999; Laundré et al. 2010). Across landscapes, predation risk varies due to the distribution and foraging efficiency of predators in different habitats. Furthermore, the structural complexity of the habitat can constrain the ability of prey to detect predators, the availability/proximity of shelter from predators or the optimal escape tactics of prey (Denno et al. 2005; Rilov et al. 2007). In response to this variation in risk, prey can either become evasive, moving to avoid predators, or modify their antipredator responses (e.g., vigilance) to reduce the chances of capture (Werner et al. 1983; Huntingford and Wright 1989; Bland and Temple 1990; Laundré et al. 2010; Martin et al. 2010; Schmitz and Trussell 2016). For prey to optimize their behavioral response to risk across landscapes, they must have knowledge of the risk levels associated with each different habitat. While prey can assess the baseline level of risk by assessing characteristics such as field of view or availability of shelter, ultimately, prey have imperfect knowledge of risk and thus can benefit from learning about the spatial variation in risk posed by predators (Dall 2010; Laundré et al. 2010). This is especially true given that predator hunting performance is not consistent across habitats (Preisser et al. 2007; Orrock et al. 2013); for example, ambush predators are generally more effective in complex habitats while pursuit predators perform better in open habitats (Preisser et al. 2007). Hence, prey should learn about the overall risk as well as the danger associated with specific predators in a specific habitat. Prey are remarkably efficient at learning about predation threats, learning associated visual and olfactory cues and the intensity of risk after a single encounter (Engstrom-Ost and Lehtiniemi 2004; Ferrari et al. 2010). Prey are also able to learn to avoid locations where they have previously encountered known predators (Huntingford and Wright 1989; Mathis and Unger 2012). However, there is still little information regarding how prey learn about predation risk across different contexts and what cues they might use to do so. When prey learn about predators, they do so in an information-rich environment where they are surrounded by cues, which may provide information on the time of day or local environment (Skow and Jakob 2005). By incorporating environmental information into their assessment of predation risk, prey should be able to learn spatial and temporal patterns of risk in their environment and regarding their predators. Few studies have looked at how the context of encounters with predators or other risk-associated stimuli can influence how prey learn their predators, and those that have explored the role context plays in learning about risk provide conflicting results. Previous studies have shown that prey can adjust their antipredator response to match temporal variation in risk posed by a predator (Ferrari et al. 2009; Bosiger et al. 2012). In contrast, wild prey still respond to predator cues under laboratory settings, suggesting that prey may not associate habitat contexts with the identity of learned predators (Griffin 2004). In this study, we examined if the olfactory cues from different habitats influence how prey learn about specific predators and whether prey use these olfactory cues to learn about the overall risk of habitats in general. Prey exposed to high levels of predation risk develop what is considered a “high-risk” phenotype (Ferrari et al. 2015) that involves behavioral, cognitive and physiological changes such as increased lateralization, increased retention of responses to learned predator cues, faster metabolic recovery following stress, altered morphology and the development of neophobic responses (the “fear” of novel stimuli) (Brown and Braithwaite 2005; Brown et al. 2013; Ferrari et al. 2015; Mitchell et al. 2016; Crane and Ferrari 2017). Furthermore, this high-risk phenotype can be rapidly induced over a matter of days (Brown et al. 2013; Brown et al. 2015; Ferrari et al. 2015; Mitchell et al. 2016), suggesting its expression is context dependent. While these traits enhance prey survival in high-risk environments, they are costly (e.g., reduced competitive ability and spatial learning) and reduce fitness in low-risk habitats (Brown and Braithwaite 2005; Chivers et al. 2017). As these traits are responsive to temporal variation in risk, we might expect that they should also be induced as prey move from habitats of low risk to high risk, particularly the more plastic cognitive traits such as neophobia. Using wood frog tadpoles, Lithobates sylvatica, we explored how risk experienced in different habitats shapes the antipredator response of prey. Wood frogs have a broad distribution across North America, breed in a range of habitats (Baldwin et al. 2006), and are exposed to a range of predators with different foraging strategies including sit and wait, sit and pursue and active pursuit predators that likely differ in their predation success in various habitats (Chivers and Mirza 2001; Relyea 2001). Hence, their response to each predator should be habitat-specific and their expression of antipredator phenotypes dependent on the overall risk encountered within each habitat (e.g., hunting efficiency and density of predators). Tadpoles have also been shown to develop neophobia following periods of high predation risk (Mitchell et al. 2016). To test if tadpoles use olfactory cues from habitats to develop habitat-specific antipredator responses, we conditioned tadpoles over a period of days to odor from a high-risk habitat and a low-risk habitat (experiment 1), or we taught tadpoles to recognize a predator in one habitat while allowing them to experience the other habitat as low-risk (experiment 2). In the presence of one of the 2 habitat odors, we then tested tadpole responses to a novel odor or a control (experiment 1), or to the learned predator odor or a novel odor control (experiment 2). We predicted that prey experiencing high-risk habitat odor will exhibit habitat-specific neophobia to a novel odor, whereas prey will not elicit a response to the same novel odor when tested in a low-risk habitat odor. We also predicted that tadpoles will associate predators with specific habitat odors and thus show a more pronounced antipredator response to the predator odor when tested in the same habitat odor where the predator cue was first encountered. We did not expect that the context in which prey learn about a predator would completely inhibit predator recognition in a different habitat, as the cost of failed recognition of a known predator is high. METHODS Collection and maintenance We collected 4 wood frog egg clutches (~200–400 eggs per clutch) from small ephemeral ponds around Saskatoon, Canada and transported then to the RJF Smith Centre for Aquatic Ecology at the University of Saskatchewan (Saskatoon, SK). Each clutch was placed in individual outdoor 67-L pools filled with 60 L of aged dechlorinated tap water and maintained under natural light and temperature conditions. The eggs were left until they hatched, at which point the tadpoles were split between six 67-L pools to reduce tadpole density, and algal discs (Wardley) were added to provide food. Tadpoles were given 3 weeks to develop before the experiment began (Gosner stage 25). Cue production To induce high-risk environments, we used repeated exposures of injured tadpole alarm cues, prepared directly before their use in conditioning. Tadpoles were euthanized by a concussive blow to the head using a pestle, which killed the tadpoles instantly, and were then ground up using a pestle and mortar. Twenty milliliters of dechlorinated water were then added to the crushed tadpoles and the solution was filtered to produce our alarm cue stock. For novel odors and conditioned predator odors, we used lake sturgeon, Acipenser fulvescens, and rainbow trout, Oncorhynchus mykiss, which were housed in the RJF Smith Centre for Aquatic Ecology in groups of 4–8. Wood frog tadpoles do not innately recognize rainbow trout (an introduced predator) as a predator (Chivers et al. 2015), and neither rainbow trout or lake sturgeon are found at our collection sites, precluding the potential for embryonic learning. To produce the trout and sturgeon odor, water was collected directly from the housing tanks prior to each conditioning or testing period. Sturgeon cues were diluted by a factor of 12 to standardize cue concentrations to relative body size and number of fish. To produce the habitat cues, we collected dead cattails, Typha latifolia, and a mixture of floating pond weed (including duckweed, Lemna sp., coontail, Ceratophyllum demersum, and water milfoil, Myriophyllum sp.) from Pike Lake, Saskatchewan. The 2 plant groups were selected to represent the odor profiles of 2 common habitats found around the local area but were not present at the site where the egg clutches were collected. This controlled for the fact that tadpoles can learn about risk during embryonic development (Mathis et al. 2008). The pond weed and cattails were then added to two 190-L pools filled with 150 L of aged dechlorinated water. These pools provided the stock habitat water for the conditioning tanks and testing arenas. Approximately half of the water was used each day and the pools were refilled at the end of the testing days. Experiment 1: Test of habitat-specific neophobia The aim of this experiment was to test if tadpoles use olfactory cues of habitats to learn about risk and develop habitat-specific neophobia. Tadpoles from all clutches were randomly added to 5 L glass conditioning tanks containing 3 L of either cattail odor or pond weed odor (24 tanks, 12 of each habitat odor and 20 tadpoles per tank). Each conditioning tank had a small amount of food and a small basket made of black plastic mesh that excluded tadpoles and was filled with plant material matching the stock habitat water. This ensured that habitat odors were constantly present during the conditioning period. The following day (day 1), tadpoles from 12 tanks (6 pond weed water and 6 cattail water) received 20 mL of alarm cues (“high-risk”) at 3 random times during the day (minimum of 1.5 h between exposures). The other 12 tanks received 20 mL dechlorinated water (“low-risk”). Odors were added slowly using a 60-mL syringe. At the end of the day, tadpoles were moved into a clean tank containing matching, fresh habitat odor water. The following day (day 2), tadpoles received the same conditioning cues as the day before; however, at the end of the day tadpoles were moved to tanks containing the opposite habitat odor. On days 3 and 4, tadpoles were conditioned with the alternate risk level cue to the one they received on days 1 and 2. This cycle of 2 days of high-risk and 2 days low-risk was repeated for a second cycle, meaning tadpoles were conditioned for a total of 8 days, with 4 days in a high-risk habitat and 4 days in a low-risk habitat. This exposure regime represented a fully balanced experimental design, with 12 tanks of tadpoles conditioned to recognize the cattail habitat as risky and the 12 tanks of tadpoles conditioned to recognize the pond weed habitat as risky. This design also controlled for any potential order effects associated with the onset of risk. The day after the conditioning phase ended, tadpoles were tested for their response to a novel odor or a control. Tadpoles from all conditioning regimes were transferred into 0.5 L circular testing arenas containing either cattail water or pond weed water. The tadpoles were left to acclimate for at least 1 h before testing. The testing period consisted of a 4-min behavioral observation, where the number of times tadpoles crossed the medial line of the arena was counted. Following the initial observation period, 5 mL of trout odor (novel odor) or dechlorinated water (control) was slowly injected down the side of the arena using a syringe, and tadpole activity was observed for another 4 min. The number of times tadpoles cross the median line provides one measure of locomotive activity levels. Changes (reductions) in activity from pre- to postcue exposure is a well-established assay for measuring the tadpole antipredator response, and hence tadpoles that reduced activity when exposed to a novel odor were considered to be neophobic (Gonzalo et al. 2010; Chivers et al. 2016; Mitchell et al. 2016). Tadpoles were additionally tested in dechlorinated water for their responses to pond weed odor, cattail odor, the novel odor or the water control. This allowed us to determine if tadpoles had learned to associate predation risk with the habitat odors and if the neophobic response was elicited in a novel habitat (dechlorinated water). A total of 221 tadpoles were tested (n = 25–30 per treatment). Experiment 2: Test of habitat specific predator recognition The aim of this experiment was to test whether the habitat in which prey learn to recognize a predator influences how tadpoles respond to the predator during future encounters. The experiment followed a similar design to that of experiment 1. Tadpoles from all clutches were randomly added to 5-L glass tanks containing either pond weed water or cattail water and a basket of the corresponding plant material, as described above. The following day (day 1) tadpoles in 12 tanks (6 cattail and 6 pond weed) were conditioned to recognize trout odor as a predator by adding 20 mL of tadpole alarm cues (containing 3 tadpoles) paired with 10 mL of trout odor. In the remaining 12 tanks, we added 20 mL of dechlorinated water as a control. At the end of the day tadpoles were moved to new tanks containing food and the opposite habitat odor. Alternating the habitats daily allowed tadpoles to gain experience with both the predator-associated habitat and the non-predator habitat. On day 2, tadpoles received the opposite conditioning treatment to the one they received the day before, i.e., tadpoles conditioned to recognize trout on day 1 received the water control, and tadpoles conditioned with water on day 1 were conditioned with the trout odor and alarm cue. Again, at the end of the day, tadpoles were moved to a new tank containing the alternative habitat odors. This process was repeated for the next 4 days (6 days total conditioning period) so that tadpoles experienced 3 days in each habitat and were conditioned to recognize trout in one habitat 3 times. The testing phase started the following day and followed the same procedure as experiment 1, except using different odors as test cues. Tadpoles were placed in arenas containing water from either habitat and then tested with the addition of either trout odor (learned predator) or sturgeon odor (novel odor control). Sturgeon odor acted as a control for both the mechanical disturbance caused by introducing the stimuli and the addition of a chemical odor. A total of 221 tadpoles were tested (n = 24–36 per treatment). Ethical statement Experimental methods followed ASAB and UCACS guidelines for the ethical treatment of animals (UCACS protocol 2015031). Throughout the early developmental period, water was changed daily (50%), and excess food was removed to ensure water parameters were optimal. During the experiment tadpoles were captured using small hand nets and moved between tanks using water filled containers. To produce the alarm cues tadpoles were euthanized according to amphibian guidelines using standard physical methods (Chivers et al. 2016; Crane et al. 2017). Chemical anesthesia could not be used as a method for euthanasia as it may interfere with the chemical alarm cues and the behavior of the tadpoles. There was no tadpole mortality during the collection, hatching and experimental periods. However, there was some minimal mortality (<3%) due to natural causes during early tadpole development prior to the onset of the experiment. At the end of the experiment, all tadpoles were returned to the ponds where they were collected. Fish odors were collected from species already housed in the RJF Smith Centre. Lake sturgeon and rainbow trout were held in 2500-L and 950-L tanks respectively and supplied with dechlorinated water on a flow-through system. Fish were fed daily to satiation, with excess food removed. No trout or sturgeon was handled during the experiment, and there was no mortality. Following the end of the experiment, all fish remained in the aquatics facility to be used in future experiments. Statistical analysis For both experiments, we calculated the proportional change in line crosses from the prestimulus baseline ([post ˗ pre]/pre) and used these values as our response variable for the analyses. All data met the assumptions for homogeneity of variance and normality. Data analysis was conducted using the statistical software package IBM SPSS statistics (IBM). For all experiments, we used nested analysis of variance (ANOVA) designs with Type I sum of squares where the conditioning tank was included as a random factor (nested under treatment) to account for the fact that tadpoles were conditioned in batches and therefore not independent. For experiment 1, we first tested differences in prestimulus activity among treatment groups. We used a 3-way ANOVA to test the effects of high-risk conditioning habitat (either cattails versus pond weed conditioned as high-risk), test habitat (cattails vs. pond weed as habitat odor present during testing), and test odor (novel odor vs. dechlorinated water added to the testing arena after pre-exposure observations) on tadpole line crosses. To interpret significant interactions among factors, we split the data by conditioning habitat and performed post-hoc 2-way nested ANOVAs for high-risk-associated cattail and pond weed habitats, followed by independent sample t-tests to identify differences between testing odors in each testing habitat. We used a Bonferroni correction on the post-hoc t-tests to account for multiple tests being run on the same data (adjusted α = 0.0125). To determine whether tadpoles responded to habitat cues, we used a 2-way nested ANOVA to test the effects of conditioning habitat (cattails vs. pond weed) and test odor (cattail vs. pond weed vs. novel odor vs. water) on the proportional change in line crosses. For experiment 2, we used 3-way nested ANOVAs to test the effects of predator-learning habitat (pond weed vs. cattail), testing habitat (pond weed vs. cattail), and test odors (trout vs. sturgeon) on both the prestimulus baseline activity and the proportional change in lines crossed data (2 separate analyses). RESULTS Experiment 1: Test for habitat specific neophobia Tadpoles did not differ in the number of line crosses during prestimulus observations (high-risk conditioning habitat, all terms P > 0.05). For the change in lines crossed due to the testing odor, the 3-way ANOVA revealed a significant interaction between conditioning habitat, test habitat, and test odor (P = 0.049; Table 1, a; Figure 1). For tadpoles exposed to high predation risk in the presence of pond weed odor, there was a significant effect of test odor (P < 0.005; Table 1, b; Figure 1a) but no effect of test habitat, test habitat × test odor interaction (both P > 0.05) or conditioning tank (P = 0.06). In other words, tadpoles that had experienced high-risk pond weed habitat significantly reduced the number of line crosses when exposed to the novel odor relative to the water control, demonstrating a neophobic response. However, this response was not habitat-specific during testing, as tadpoles tested in both habitat types responded similarly (Figure 1a). Table 1 Results of habitat-specific neophobia experiment, with ANOVAs testing for (a) the effect of high-risk conditioning habitat, test habitat, and test odor on the proportional change in tadpole activity, (b) the effect of test habitat and test odor on the proportional change in tadpole activity for tadpoles conditioned to high-risk in the pond weed habitat, and (c) the effect of test habitat and test odor on the proportional change in tadpole activity for tadpoles conditioned to high-risk in the cattail habitat Source SS df F P (a) Full model High-risk conditioning habitat 0.000 1 0.001 0.972 Error 5.624 21.320 Test habitat 0.093 1 0.469 0.494 Error 38.493 194.103 Test odor 3.193 1 15.642 <0.001 Error 43.277 211.998 High-risk conditioning habitat × Test habitat 0.002 1 0.011 0.915 Error 38.550 194.357 High-risk conditioning habitat × Test odor 0.000 1 0.001 0.969 Error 43.267 211.880 Test habitat × Test odor 0.871 1 4.367 0.038 Error 40.352 202.238 High-risk conditioning habitat × Test habitat × Test odor 0.781 1 3.914 0.049 Error 40.338 202.178 Conditioning tank 5.774 22 1.326 0.158 Error 37.800 191 (b) Pond weed high-risk habitat Test habitat 0.078 1 0.430 0.513 Error 17.878 98.893 Test odor 1.584 1 8.271 0.005 Error 20.445 106.735 Test habitat × Test odor 0.002 1 0.010 0.919 Error 19.440 105.391 Conditioning tank 3.589 11 1.813 0.062 Error 17.457 97 (c) Cattail high-risk habitat Test habitat 0.023 1 0.106 0.745 Error 20.589 95.194 Test odor 1.604 1 7.482 0.007 Error 22.504 104.991 Test habitat × Test odor 1.650 1 7.638 0.007 Error 20.992 97.155 Conditioning tank 2.185 11 0.918 0.527 Error 20.343 94 Source SS df F P (a) Full model High-risk conditioning habitat 0.000 1 0.001 0.972 Error 5.624 21.320 Test habitat 0.093 1 0.469 0.494 Error 38.493 194.103 Test odor 3.193 1 15.642 <0.001 Error 43.277 211.998 High-risk conditioning habitat × Test habitat 0.002 1 0.011 0.915 Error 38.550 194.357 High-risk conditioning habitat × Test odor 0.000 1 0.001 0.969 Error 43.267 211.880 Test habitat × Test odor 0.871 1 4.367 0.038 Error 40.352 202.238 High-risk conditioning habitat × Test habitat × Test odor 0.781 1 3.914 0.049 Error 40.338 202.178 Conditioning tank 5.774 22 1.326 0.158 Error 37.800 191 (b) Pond weed high-risk habitat Test habitat 0.078 1 0.430 0.513 Error 17.878 98.893 Test odor 1.584 1 8.271 0.005 Error 20.445 106.735 Test habitat × Test odor 0.002 1 0.010 0.919 Error 19.440 105.391 Conditioning tank 3.589 11 1.813 0.062 Error 17.457 97 (c) Cattail high-risk habitat Test habitat 0.023 1 0.106 0.745 Error 20.589 95.194 Test odor 1.604 1 7.482 0.007 Error 22.504 104.991 Test habitat × Test odor 1.650 1 7.638 0.007 Error 20.992 97.155 Conditioning tank 2.185 11 0.918 0.527 Error 20.343 94 View Large Table 1 Results of habitat-specific neophobia experiment, with ANOVAs testing for (a) the effect of high-risk conditioning habitat, test habitat, and test odor on the proportional change in tadpole activity, (b) the effect of test habitat and test odor on the proportional change in tadpole activity for tadpoles conditioned to high-risk in the pond weed habitat, and (c) the effect of test habitat and test odor on the proportional change in tadpole activity for tadpoles conditioned to high-risk in the cattail habitat Source SS df F P (a) Full model High-risk conditioning habitat 0.000 1 0.001 0.972 Error 5.624 21.320 Test habitat 0.093 1 0.469 0.494 Error 38.493 194.103 Test odor 3.193 1 15.642 <0.001 Error 43.277 211.998 High-risk conditioning habitat × Test habitat 0.002 1 0.011 0.915 Error 38.550 194.357 High-risk conditioning habitat × Test odor 0.000 1 0.001 0.969 Error 43.267 211.880 Test habitat × Test odor 0.871 1 4.367 0.038 Error 40.352 202.238 High-risk conditioning habitat × Test habitat × Test odor 0.781 1 3.914 0.049 Error 40.338 202.178 Conditioning tank 5.774 22 1.326 0.158 Error 37.800 191 (b) Pond weed high-risk habitat Test habitat 0.078 1 0.430 0.513 Error 17.878 98.893 Test odor 1.584 1 8.271 0.005 Error 20.445 106.735 Test habitat × Test odor 0.002 1 0.010 0.919 Error 19.440 105.391 Conditioning tank 3.589 11 1.813 0.062 Error 17.457 97 (c) Cattail high-risk habitat Test habitat 0.023 1 0.106 0.745 Error 20.589 95.194 Test odor 1.604 1 7.482 0.007 Error 22.504 104.991 Test habitat × Test odor 1.650 1 7.638 0.007 Error 20.992 97.155 Conditioning tank 2.185 11 0.918 0.527 Error 20.343 94 Source SS df F P (a) Full model High-risk conditioning habitat 0.000 1 0.001 0.972 Error 5.624 21.320 Test habitat 0.093 1 0.469 0.494 Error 38.493 194.103 Test odor 3.193 1 15.642 <0.001 Error 43.277 211.998 High-risk conditioning habitat × Test habitat 0.002 1 0.011 0.915 Error 38.550 194.357 High-risk conditioning habitat × Test odor 0.000 1 0.001 0.969 Error 43.267 211.880 Test habitat × Test odor 0.871 1 4.367 0.038 Error 40.352 202.238 High-risk conditioning habitat × Test habitat × Test odor 0.781 1 3.914 0.049 Error 40.338 202.178 Conditioning tank 5.774 22 1.326 0.158 Error 37.800 191 (b) Pond weed high-risk habitat Test habitat 0.078 1 0.430 0.513 Error 17.878 98.893 Test odor 1.584 1 8.271 0.005 Error 20.445 106.735 Test habitat × Test odor 0.002 1 0.010 0.919 Error 19.440 105.391 Conditioning tank 3.589 11 1.813 0.062 Error 17.457 97 (c) Cattail high-risk habitat Test habitat 0.023 1 0.106 0.745 Error 20.589 95.194 Test odor 1.604 1 7.482 0.007 Error 22.504 104.991 Test habitat × Test odor 1.650 1 7.638 0.007 Error 20.992 97.155 Conditioning tank 2.185 11 0.918 0.527 Error 20.343 94 View Large Figure 1 View largeDownload slide Test for habitat-specific neophobia (experiment 1). Mean (± 1 SE) proportional change in line crosses for tadpoles exposed to either a novel odor (gray) or control water (white) when tested in water from either pond weed habitat or cattail habitat. Tadpoles were initially conditioned to learn the odor of pond weed odor (top panel) or cattail habitat (bottom panel) as a high-risk habitat and the alternative habitat as low-risk. Figure 1 View largeDownload slide Test for habitat-specific neophobia (experiment 1). Mean (± 1 SE) proportional change in line crosses for tadpoles exposed to either a novel odor (gray) or control water (white) when tested in water from either pond weed habitat or cattail habitat. Tadpoles were initially conditioned to learn the odor of pond weed odor (top panel) or cattail habitat (bottom panel) as a high-risk habitat and the alternative habitat as low-risk. When tadpoles experienced high risk in the cattail habitat odor, there was a significant interaction between test habitat and test odor (P = 0.007; Table 1, c; Figure 1b). Tadpoles that had experienced high-risk in the cattail habitat and were tested in cattail odor water reduced the number of line crosses when exposed to the novel odor versus water (t53 = 3.63, P = 0.001; adjusted α = 0.0125). However, when tested in pond weed habitat odor, this same group of tadpoles did not respond to the novel odor relative to the water control (t52 = −0.042, P = 0.97; adjusted α = 0.0125). There was no difference in the response of high-risk cattail habitat tadpoles to the water control when tested in cattail versus pond weed (t51 = −1.96, P = 0.055: adjusted α = 0.0125). Tadpoles showed a stronger but nonsignificant response to the novel odor when tested in the cattail habitat odor than in the pond weed odor (t54 = 2.03, P = 0.048; adjusted α = 0.0125). Together, these results demonstrate that tadpoles developed habitat-specific neophobia after experiencing high risk in the cattail habitat. When tadpoles were tested in control water for their response to habitat odors and the novel odor, we found that there was no effect of conditioning habitat (F1, 21.07 = 0.29, P = 0.60), test odor (F3, 220.64 = 0.91, p = 0.44), or interaction between the two (F3, 2221.09 = 0.77, P = 0.51) (Figure 2). Additionally, there was no effect of conditioning tank (F22, 203 = 1.41, P = 0.11). Figure 2 View largeDownload slide Learned responses to habitat odors (experiment 1). Mean (± 1 SE) proportional change in line crosses for tadpoles exposed to either pond weed water, cattail water, novel odor or control water when tested in control water. Tadpoles were initially conditioned to learn the odor of pond weed odor (top panel) or cattail habitat (bottom panel) as a high-risk habitat and the alternative habitat as low-risk. Figure 2 View largeDownload slide Learned responses to habitat odors (experiment 1). Mean (± 1 SE) proportional change in line crosses for tadpoles exposed to either pond weed water, cattail water, novel odor or control water when tested in control water. Tadpoles were initially conditioned to learn the odor of pond weed odor (top panel) or cattail habitat (bottom panel) as a high-risk habitat and the alternative habitat as low-risk. Experiment 2: Test for habitat specific predator learning There was no difference in tadpole line crosses during the prestimulus observations (high-risk conditioning habitat, all terms P > 0.05). The change in line crosses made by tadpoles was significantly affected by the test odor (F1, 207.3 = 20.39, P < 0.001; Figure 3). Across all groups, tadpoles reduced line crosses when exposed to the predator odor (~30–40% reduction) versus the novel odor (<15% reduction in activity). There was no effect of conditioning habitat, testing habitat or any interactions among factors (all P > 0.05). There was an effect of conditioning tank (F22, 191 = 1.73, P = 0.015). The significant effect of odor indicates that the tadpoles recognized the trout as a predator but the absence of an interaction among the 3 factors indicates that the tadpoles did not learn to perceive the predator as more dangerous in a specific habitat. Figure 3 View largeDownload slide Test for habitat specific predator learning (experiment 2). Mean (± 1 SE) proportional change in line crosses for tadpoles exposed to either the predator odor (gray) or the novel odor (white) when tested in water from either pond weed habitat or cattail habitat. Tadpoles were initially conditioned to learn the predator odor as a threat in either pond weed odor (top panel) or cattail habitat (bottom panel) and experienced only control water (low risk) in the alternative habitat. Figure 3 View largeDownload slide Test for habitat specific predator learning (experiment 2). Mean (± 1 SE) proportional change in line crosses for tadpoles exposed to either the predator odor (gray) or the novel odor (white) when tested in water from either pond weed habitat or cattail habitat. Tadpoles were initially conditioned to learn the predator odor as a threat in either pond weed odor (top panel) or cattail habitat (bottom panel) and experienced only control water (low risk) in the alternative habitat. DISCUSSION Many studies have investigated how prey alter their behavior under predation risk as it varies across landscapes, yet little is known about how prey learn habitat-specific risk or which cues prey use to learn about predators across contexts. Our results demonstrate that prey use olfactory cues from habitats to learn about changes in predation risk between habitats. In experiment 1, tadpoles that experienced high risk in the cattail habitat water and were subsequently tested in the cattail water reduced activity in response to a novel odor (neophobic response), but these tadpoles did not respond to the novel odor when tested in the pond weed water. Tadpoles also showed no behavioral response to the novel odor when tested in the control water, irrespective of the habitat in which tadpoles experienced risk. Thus, the onset of neophobia was unique to the habitat where tadpoles experienced high risk. Neophobic responses have been shown to increase survival (Ferrari et al. 2014) and are one of several antipredator traits that is expressed when prey experience periods of high predation risk (Ferrari et al. 2015). The rapid onset of neophobia in tadpoles entering high-risk habitats should enhance survival while they adjust to their new surroundings. However, such traits are maladaptive in low-risk environments (Chivers et al. 2014), hence the lack of neophobic responses in the low-risk pond weed habitat or control water. The results presented here demonstrate that tadpoles can use olfactory cues from habitats to transition between these high-risk and low-risk antipredator phenotypes and optimize their performance in response to variable risk at landscape scales. While there was clear evidence for habitat-specific neophobia, antipredator responses were not consistent across the background habitat treatments. Tadpoles that had experienced high risk in the pond weed habitat displayed a neophobic response to the novel odor when tested in both pond weed and cattail water, perceiving the cattail habitat as high risk despite the fact they did not experience high risk in that habitat. However, there were no neophobic responses when tadpoles were tested in control water, indicating that the neophobic responses were limited to the habitats experienced during conditioning. These apparently conflicting results may be in part explained by how prey develop antipredator responses to different cues. It is possible that tadpoles innately recognized the cattail habitat as risky. Consequently, tadpoles would display neophobic responses without having to experience risk in the cattail habitat. Studies have shown that animals can innately recognize cues from specific habitats, e.g., both sea lamprey, Petromyzon marinus, and anemonefish, Amphiprion percula, show innate recognition of olfactory cues from preferred settlement habitats (Vrieze et al. 2010; Dixson et al. 2014). For tadpoles to possess innate recognition of cattail habitat odors as risky, exposure to consistently high predation risk in cattail habitats over multiple generations is likely required. Given the diversity and distribution of predators that consume tadpoles across different habitats, we suggest that such an innate response is unlikely to have developed for cattail habitats and not pond weed habitats. Alternatively, tadpoles may have generalized recognition of risk from the pond weed cues to the cattail cues but in that direction only. Generalized recognition occurs when a stimulus has enough similarity to a conditioned odor that it elicits a similar response to that of the conditioned stimulus (Ghirlanda and Enquist 2003; Shettleworth 2010). Unidirectional generalization may have occurred if a chemical was a common component of the pond weed odor but was present only as a minor component of the cattail odor. Tadpoles would then generalize high-risk from the pond weed to the cattail habitat but generalization would not go in the opposite direction. Finally, the conflicting results may have been due to variation in chemical stability of habitat cues during the experiment. As the cattail plants were dead, the suite of chemical compounds released may have change over time as the plants decayed. Previous studies have shown that variation in the stimuli being learned can slow or inhibit the learning process (Craig 1994a; Craig 1994b). For example, stingless bees, Trigona fuscipennis, can learn to avoid webs spun by orb spinner, Argiope argentata, but the learning process is delayed due to A. argentata using different silk patterns to decorate its web each day. Additionally, learning about safe stimuli requires a greater number of conditioning events than learning about risky stimuli (Ferrari and Chivers 2011), and is highly context sensitive (Lubow 1989). A combination of the variation in the odor complex being learned and the information being encoded (safety of the habitat) may have prevented tadpoles from learning the difference in risk levels between the 2 habitats and resulted in learning risk across both habitats. Variation in cattail cues would not have prevented tadpoles from learning the cattail habitat as risky, as only one conditioning event is sufficient to cause associative learning of an odor as risky. In experiment 2, we found that habitat had no effect on how prey learned about predator odors. Irrespective of the habitat in which prey were conditioned to recognize the predator, prey responded to predator cues with similar reductions in activity when tested in either habitat odor. The fact that wild caught animals still respond to predator cues in captive settings also suggests that context does not influence learned predator recognition (Griffin 2004). Ecologically, such responses are appropriate, as prey should generally adopt a cautionary approach and overestimate risk (Bouskila and Blumstein 1992). Hence, when tadpoles encounter a known predator in a safe environment, tadpoles should respond to the predator as in previous encounters in other habitats. Despite this, it should be noted that the lack of variation in response to predator cues across habitats does not mean prey do not incorporate information from habitats when learning about predator specific risk, simply that the optimal response in our experiment was to respond with an equal intensity across both habitats. Future research should assess conflicting information about a predator across habitats, as has been done for prey learning temporal patterns of risk (Ferrari et al. 2009; Bosiger et al. 2012). Despite the potential costs, learning about risk provides an adaptive strategy for prey to modify their antipredator responses to their current environment. While studies have shown how prey respond to changes in risk through time, few have looked at how prey learn about risk across different contexts. As demonstrated in this study olfactory cues from habitats provide information that allows prey to learn about risk in different habitats. Yet our results also show that there are nuances to how prey use olfactory cues from different habitats which require further investigation. It might be that prey may need additional information from other sensory systems to resolve such issues. To understand how prey respond to variable risk between habitats, additional work exploring how prey use ancillary cues from their environment to inform antipredator responses is required. Such studies will also provide insights into how human altered environments might impact predator–prey interactions. The potential value of such studies is highlighted by recent findings from coral reefs, where changes in the chemical milieu of habitats following a disturbance event appears to impair the ability of fish to detect chemical alarm cues (Ferrari et al. 2017; McCormick et al. 2017). FUNDING This work was supported by Natural Sciences and Engineering Research Council of Canada grants to M.C.O.F. and D.P.C. Data accessibility: Analyses reported in this article can be reproduced using the data provided by Mitchell et al. (2018). We thank Laurel Sacco and Phil Anderson for their assistance hunting for wood frog egg clutches. REFERENCES Abrams PA , Rowe L . 1996 . The effects of predation on the age and size of maturity of prey . Evolution . 50 : 1052 – 1061 . Google Scholar CrossRef Search ADS PubMed Baldwin RF , Calhoun AJ , deMaynadier PG . 2006 . Conservation planning for amphibian species with complex habitat requirements: a case study using movements and habitat selection of the wood frog Rana sylvatica . J Herpetol . 40 : 442 – 453 . Google Scholar CrossRef Search ADS Beauchamp G , Ruxton GD . 2011 . A reassessment of the predation risk allocation hypothesis: a comment on Lima and Bednekoff . Am Nat . 177 : 143 – 146 . Google Scholar CrossRef Search ADS PubMed Bland JD , Temple SA . 1990 . Effects of predation-risk on habitat use by Himalayan Snowcocks . Oecologia . 82 : 187 – 191 . Google Scholar CrossRef Search ADS PubMed Bosiger YJ , Lonnstedt OM , McCormick MI , Ferrari MC . 2012 . Learning temporal patterns of risk in a predator-diverse environment . PLoS One . 7 : e34535 . Google Scholar CrossRef Search ADS PubMed Bouskila A , Blumstein DT . 1992 . Rules of thumb for predation hazard assessment: predictions from a dynamic model . Am Nat . 139 : 161 – 176 . Google Scholar CrossRef Search ADS Brown C , Braithwaite VA . 2005 . Effects of predation pressure on the cognitive ability of the poeciliid Brachyraphis episcopi . Behav Ecol . 16 : 482 – 487 . Google Scholar CrossRef Search ADS Brown GE , Elvidge CK , Ramnarine I , Ferrari MCO , Chivers DP . 2015 . Background risk and recent experience influences retention of neophobic responses to predators . Behav Ecol Sociobiol . 69 : 737 – 745 . Google Scholar CrossRef Search ADS Brown GE , Ferrari MC , Elvidge CK , Ramnarine I , Chivers DP . 2013 . Phenotypically plastic neophobia: a response to variable predation risk . Proc Biol Sci . 280 : 20122712 . Google Scholar CrossRef Search ADS PubMed Brown JS , Laundre JW , Gurung M . 1999 . The Ecology of Fear: optimal foraging, game theory, and trophic interactions . J Mammal . 80 : 385 – 399 . Google Scholar CrossRef Search ADS Chivers DP , Mathiron A , Sloychuk JR , Ferrari MC . 2015 . Responses of tadpoles to hybrid predator odours: strong maternal signatures and the potential risk/response mismatch . Proc Biol Sci . 282 : 20150365 . Google Scholar CrossRef Search ADS PubMed Chivers DP , McCormick MI , Mitchell MD , Ramasamy RA , Ferrari MCO . 2014 . Background level of risk determines how prey categorize predators and non-predators . Proc R Soc B Biol Sci . 281 : 20140355 . Google Scholar CrossRef Search ADS Chivers DP , McCormick MI , Warren DT , Allan BJM , Ramasamy RA , Arvizu BK , Glue M , Ferrari MCO . 2017 . Competitive superiority versus predation savvy: the two sides of behavioural lateralization . Anim Behav . 130 : 9 – 15 . Google Scholar CrossRef Search ADS Chivers DP , Mirza RS . 2001 . Importance of predator diet cues in responses of larval wood frogs to fish and invertebrate predators . J Chem Ecol . 27 : 45 – 51 . Google Scholar CrossRef Search ADS PubMed Chivers DP , Mitchell MD , Lucon-Xiccato T , Brown GE , Ferrari MCO . 2016 . Background risk influences learning but not generalization of predators . Anim Behav . 121 : 185 – 189 . Google Scholar CrossRef Search ADS Craig CL . 1994a . Predator foraging behavior in response to perception and learning by its prey: interactions between orb-spinning spiders and stingless bees . Behav Ecol Sociobiol . 35 : 45 – 52 . Google Scholar CrossRef Search ADS Craig CL . 1994b . Limits to learning: effects of predator pattern and colour on perception and avoidance-learning by prey . Anim Behav . 47 : 1087 – 1099 . Google Scholar CrossRef Search ADS Crane AL , Demuth BS , Ferrari MCO . 2017 . Experience with predators shapes learning rules in larval amphibians . Behav Ecol . 28 : 312 – 318 . Google Scholar CrossRef Search ADS Crane AL , Ferrari MCO . 2017 . Patterns of predator neophobia: a meta-analytic review . Proc R Soc B Biol Sci . 284 : 20170583 . Google Scholar CrossRef Search ADS Dall SR . 2010 . Managing risk: the perils of uncertainty . In: Westneat DF , Fox CW , editors. Evolutionary behavioral ecology . New York (NY) : Oxford University Press, Inc . p. 194 – 206 . Dayton GH , Saenz D , Baum KA , Langerhans RB , DeWitt TJ . 2005 . Body shape, burst speed and escape behavior of larval anurans . Oikos . 111 : 582 – 591 . Google Scholar CrossRef Search ADS Denno RF , Finke DL , Langellotto GA . 2005 . Direct and indirect effects of vegetation structure and habitat complexity on predator-prey and predator-predator interactions . In: Barbosa P , Castellanos I , editors. Ecology of predator-prey interactions . Oxford : Oxford University Press . p. 211 – 239 . Dixson DL , Jones GP , Munday PL , Planes S , Pratchett MS , Thorrold SR . 2014 . Experimental evaluation of imprinting and the role innate preference plays in habitat selection in a coral reef fish . Oecologia . 174 : 99 – 107 . Google Scholar CrossRef Search ADS PubMed Engstrom-Ost J , Lehtiniemi M . 2004 . Threat-sensitive predator avoidance by pike larvae . J Fish Biol . 65 : 251 – 261 . Google Scholar CrossRef Search ADS Ferrari MCO , Chivers DP . 2011 . Learning about non-predators and safe places: the forgotten elements of risk assessment . Anim Cogn . 14 : 309 – 316 . Google Scholar CrossRef Search ADS PubMed Ferrari MCO , Chivers DP , Wisenden BD . 2010 . Chemical ecology of predator-prey interactions in aquatic ecosystems: a review and prospectus . Can . 88 : 698 – 724 . Ferrari MCO , Manek AK , Chivers DP . 2009 . Temporal learning of predation risk by embryonic amphibians . Biol Lett . 6 : 308 – 310 . Google Scholar CrossRef Search ADS PubMed Ferrari MCO , McCormick MI , Allan BJM , Chivers DP . 2017 . Not equal in the face of habitat change: closely related fishes differ in their ability to use predation-related information in degraded coral . Proc R Soc B Biol Sci . 284 : 20162758 . Google Scholar CrossRef Search ADS Ferrari MC , McCormick MI , Allan BJ , Choi R , Ramasamy RA , Johansen JL , Mitchell MD , Chivers DP . 2015 . Living in a risky world: the onset and ontogeny of an integrated antipredator phenotype in a coral reef fish . Sci Rep . 5 : 15537 . Google Scholar CrossRef Search ADS PubMed Ferrari MCO , McCormick MI , Meekan MG , Chivers DP . 2014 . Background level of risk and the survival of predator-naive prey: can neophobia compensate for predator naivety in juvenile coral reef fishes ? Proc R Soc B Biol Sci . 282 : 20142197 . Google Scholar CrossRef Search ADS Ghirlanda S , Enquist M . 2003 . A century of generalization . Anim Behav . 66 : 15 – 36 . Google Scholar CrossRef Search ADS Gonzalo A , López P , Martín J . 2010 . Risk level of chemical cues determines retention of recognition of new predators in Iberian green frog tadpoles . Behav Ecol Sociobiol . 64 : 1117 – 1123 . Google Scholar CrossRef Search ADS Griffin AS . 2004 . Social learning about predators: a review and prospectus . Anim. Learn Behav . 32 : 131 – 140 . Google Scholar CrossRef Search ADS Huntingford FA , Wright PJ . 1989 . How sticklebacks learn to avoid dangerous feeding patches . Behav Processes . 19 : 181 – 189 . Google Scholar CrossRef Search ADS PubMed Laforsch C , Ngwa W , Grill W , Tollrian R . 2004 . An acoustic microscopy technique reveals hidden morphological defenses in Daphnia . Proc Natl Acad Sci USA . 101 : 15911 – 15914 . Google Scholar CrossRef Search ADS PubMed Laundré JW , Hernández L , Ripple WJ . 2010 . The landscape of fear: ecological implications of being afraid . Open Ecol J . 3 : 1 – 7 . Google Scholar CrossRef Search ADS Laurila A , Kujasalo J , Ranta E . 1998 . Predator-induced changes in life history in two anuran tadpoles: effects of predator diet . Oikos . 83 : 307 . Google Scholar CrossRef Search ADS Lima SL , Bednekoff PA . 1999 . Temporal variation in danger drives antipredator behavior- The Predation Risk Allocation Hypothesis . Am Nat . 153 : 649 – 659 . Google Scholar CrossRef Search ADS PubMed Lubow RE . 1989 . Latent inhibition and conditioned attention theory . Cambridge : Cambridge University Press . Google Scholar CrossRef Search ADS Martin CW , Fodrie FJ , Heck KL Jr , Mattila J . 2010 . Differential habitat use and antipredator response of juvenile roach (Rutilus rutilus) to olfactory and visual cues from multiple predators . Oecologia . 162 : 893 – 902 . Google Scholar CrossRef Search ADS PubMed Mathis A , Ferrari MCO , Windel N , Messier F , Chivers DP . 2008 . Learning by embryos and the ghost of predation future . Proc R Soc B Biol Sci . 275 : 2603 – 2607 . Google Scholar CrossRef Search ADS Mathis A , Unger S . 2012 . Learning to avoid dangerous habitat types by aquatic salamanders, Eurycea tynerensis: learning to avoid dangerous habitats by salamanders . Ethology . 118 : 57 – 62 . Google Scholar CrossRef Search ADS McCormick MI , Chivers DP , Allan BJ , Ferrari MC . 2017 . Habitat degradation disrupts neophobia in juvenile coral reef fish . Glob Chang Biol . 23 : 719 – 727 . Google Scholar CrossRef Search ADS PubMed Mitchell MD , Chivers DP , Brown GE , Ferrari MCO . 2016 . Living on the edge: how does environmental risk affect the behavioural and cognitive ecology of prey ? Anim Behav . 115 : 185 – 192 . Google Scholar CrossRef Search ADS Mitchell MD , Crane AL , Bairos-Novak KR , Ferrari MCO , Chivers DP . 2018 . Data from: olfactory cues of habitats facilitate learning about landscapes of fear . Dryad Digital Repository . http://dx.doi.org/10.5061/dryad.218p0v2 Orrock JL , Preisser EL , Grabowski JH , Trussell GC . 2013 . The cost of safety: refuges increase the impact of predation risk in aquatic systems . Ecology . 94 : 573 – 579 . Google Scholar CrossRef Search ADS PubMed Preisser EL , Bolnick DI , Benard MF . 2005 . Scared to death? The effects of intimidation and consumption in predator-prey interactions . Ecology . 86 : 501 – 509 . Google Scholar CrossRef Search ADS Preisser EL , Orrock JL , Schmitz OJ . 2007 . Predator hunting mode and habitat domain alter nonconsumptive effects in predator-prey interactions . Ecology . 88 : 2744 – 2751 . Google Scholar CrossRef Search ADS PubMed Relyea RA . 2001 . Morphological and behavioral plasticity of larval anurans in response to different predators . Ecology . 82 : 523 – 540 . Google Scholar CrossRef Search ADS Rilov G , Figueira WF , Lyman SJ , Crowder L . 2007 . Complex habitats may not always benefit prey: linking visual field with reef fish behavior and distribution . Mar Ecol Prog Ser . 329 : 225 – 238 . Google Scholar CrossRef Search ADS Schmitz OJ , Trussell GC . 2016 . Multiple stressors, state-dependence and predation risk — foraging trade-offs: toward a modern concept of trait-mediated indirect effects in communities and ecosystems . Curr Opin Behav Sci . 12 : 6 – 11 . Google Scholar CrossRef Search ADS Shettleworth SJ . 2010 . Cognition, evolution and behavior . 2nd ed . New York : Oxford University Press . Skow CD , Jakob EM . 2005 . Jumping spiders attend to context during learned avoidance of aposematic prey . Behav Ecol . 17 : 34 – 40 . Google Scholar CrossRef Search ADS Tollrian R . 1993 . Neckteeth formation in Daphnia pulex as an example of continuous phenotypic plasticity: morphological effects of Chaoborus kairomone concentration and their quantification . J Plankton Res . 15 : 1309 – 1318 . Google Scholar CrossRef Search ADS Vrieze LA , Bjerselius R , Sorensen PW . 2010 . Importance of the olfactory sense to migratory sea lampreys Petromyzon marinus seeking riverine spawning habitat . J Fish Biol . 76 : 949 – 964 . Google Scholar CrossRef Search ADS Werner EE , Gilliam JF , Hall DJ , Mittelbach GG . 1983 . An experimental test of the effects of predation risk on habitat use in fish . Ecology . 64 : 1540 – 1548 . Google Scholar CrossRef Search ADS Werner EE , Peacor SD . 2003 . A review of trait-mediated indirect interactions in ecological communities . Ecology . 84 : 1083 – 1100 . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Behavioral Ecology Oxford University Press

Olfactory cues of habitats facilitate learning about landscapes of fear

Loading next page...
 
/lp/ou_press/olfactory-cues-of-habitats-facilitate-learning-about-landscapes-of-u6yo9znQU0
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1045-2249
eISSN
1465-7279
D.O.I.
10.1093/beheco/ary024
Publisher site
See Article on Publisher Site

Abstract

Abstract Across landscapes, prey are exposed to different levels of predation risk within different habitats. However, little is known about how prey learn about risk in different habitat types. Here, we examined if wood frog tadpoles, Lithobates sylvatica, use olfactory cues from 2 distinct, plant-dominated habitats (cattail and pond weed) to learn about the overall risk within a habitat and the risk posed by a specific predator species within different habitats. In our first experiment, tadpoles experienced both a high-risk and a low-risk habitat before being tested for habitat-specific neophobic responses, a cognitive trait expressed in high-risk but not low-risk environments. In the second experiment, we taught tadpoles to recognize a predator in one habitat while the other one was never associated with a predator. Tadpoles were then tested for their responses to the predator and a control in both habitats. Our results showed that high-risk cattail tadpoles developed habitat-specific neophobia. However, high-risk pond weed tadpoles developed a generalized neophobia, responding to the novel cues irrespective of the habitat where they were tested. We also found that the habitat in which prey learned the identity of a specific predator did not affect their responses to that predator when tested in different habitats. Our results provide support for the use of olfactory habitat cues by prey to learn about predation risk across landscapes, suggesting unrecognized nuances to how prey use such cues to learn about predation risk. INTRODUCTION Predation is a major ecological process that shapes communities. Beyond consuming prey, predators can affect prey by changing their behavior, morphology, and life-history (Werner and Peacor 2003; Preisser et al. 2005). For instance, in the presence of predators, prey can invest in morphological changes that increase escape speed (e.g., tail depth [Dayton et al. 2005]) or become harder to consume (e.g., body size, amour thickness and spines (Tollrian 1993; Laforsch et al. 2004)). Some prey species may alter the timing of transitions between developmental stages to avoid predators, while others change the timing of their sexual maturity and reproduction (Abrams and Rowe 1996; Laurila et al. 1998). The trade-off between avoiding predators and investing in other activities such as foraging, growth, and reproductive output that enhance fitness means that, in many cases, prey must adjust their time and energetic investment in antipredator responses to match the level of threat they experience (Brown et al. 1999; Lima and Bednekoff 1999; Beauchamp and Ruxton 2011). However, as predation risk varies with both time and space, prey should adjust their antipredator responses to match the level of risk experienced in a given ecological context, such as specific habitats (Lima and Bednekoff 1999; Laundré et al. 2010). Across landscapes, predation risk varies due to the distribution and foraging efficiency of predators in different habitats. Furthermore, the structural complexity of the habitat can constrain the ability of prey to detect predators, the availability/proximity of shelter from predators or the optimal escape tactics of prey (Denno et al. 2005; Rilov et al. 2007). In response to this variation in risk, prey can either become evasive, moving to avoid predators, or modify their antipredator responses (e.g., vigilance) to reduce the chances of capture (Werner et al. 1983; Huntingford and Wright 1989; Bland and Temple 1990; Laundré et al. 2010; Martin et al. 2010; Schmitz and Trussell 2016). For prey to optimize their behavioral response to risk across landscapes, they must have knowledge of the risk levels associated with each different habitat. While prey can assess the baseline level of risk by assessing characteristics such as field of view or availability of shelter, ultimately, prey have imperfect knowledge of risk and thus can benefit from learning about the spatial variation in risk posed by predators (Dall 2010; Laundré et al. 2010). This is especially true given that predator hunting performance is not consistent across habitats (Preisser et al. 2007; Orrock et al. 2013); for example, ambush predators are generally more effective in complex habitats while pursuit predators perform better in open habitats (Preisser et al. 2007). Hence, prey should learn about the overall risk as well as the danger associated with specific predators in a specific habitat. Prey are remarkably efficient at learning about predation threats, learning associated visual and olfactory cues and the intensity of risk after a single encounter (Engstrom-Ost and Lehtiniemi 2004; Ferrari et al. 2010). Prey are also able to learn to avoid locations where they have previously encountered known predators (Huntingford and Wright 1989; Mathis and Unger 2012). However, there is still little information regarding how prey learn about predation risk across different contexts and what cues they might use to do so. When prey learn about predators, they do so in an information-rich environment where they are surrounded by cues, which may provide information on the time of day or local environment (Skow and Jakob 2005). By incorporating environmental information into their assessment of predation risk, prey should be able to learn spatial and temporal patterns of risk in their environment and regarding their predators. Few studies have looked at how the context of encounters with predators or other risk-associated stimuli can influence how prey learn their predators, and those that have explored the role context plays in learning about risk provide conflicting results. Previous studies have shown that prey can adjust their antipredator response to match temporal variation in risk posed by a predator (Ferrari et al. 2009; Bosiger et al. 2012). In contrast, wild prey still respond to predator cues under laboratory settings, suggesting that prey may not associate habitat contexts with the identity of learned predators (Griffin 2004). In this study, we examined if the olfactory cues from different habitats influence how prey learn about specific predators and whether prey use these olfactory cues to learn about the overall risk of habitats in general. Prey exposed to high levels of predation risk develop what is considered a “high-risk” phenotype (Ferrari et al. 2015) that involves behavioral, cognitive and physiological changes such as increased lateralization, increased retention of responses to learned predator cues, faster metabolic recovery following stress, altered morphology and the development of neophobic responses (the “fear” of novel stimuli) (Brown and Braithwaite 2005; Brown et al. 2013; Ferrari et al. 2015; Mitchell et al. 2016; Crane and Ferrari 2017). Furthermore, this high-risk phenotype can be rapidly induced over a matter of days (Brown et al. 2013; Brown et al. 2015; Ferrari et al. 2015; Mitchell et al. 2016), suggesting its expression is context dependent. While these traits enhance prey survival in high-risk environments, they are costly (e.g., reduced competitive ability and spatial learning) and reduce fitness in low-risk habitats (Brown and Braithwaite 2005; Chivers et al. 2017). As these traits are responsive to temporal variation in risk, we might expect that they should also be induced as prey move from habitats of low risk to high risk, particularly the more plastic cognitive traits such as neophobia. Using wood frog tadpoles, Lithobates sylvatica, we explored how risk experienced in different habitats shapes the antipredator response of prey. Wood frogs have a broad distribution across North America, breed in a range of habitats (Baldwin et al. 2006), and are exposed to a range of predators with different foraging strategies including sit and wait, sit and pursue and active pursuit predators that likely differ in their predation success in various habitats (Chivers and Mirza 2001; Relyea 2001). Hence, their response to each predator should be habitat-specific and their expression of antipredator phenotypes dependent on the overall risk encountered within each habitat (e.g., hunting efficiency and density of predators). Tadpoles have also been shown to develop neophobia following periods of high predation risk (Mitchell et al. 2016). To test if tadpoles use olfactory cues from habitats to develop habitat-specific antipredator responses, we conditioned tadpoles over a period of days to odor from a high-risk habitat and a low-risk habitat (experiment 1), or we taught tadpoles to recognize a predator in one habitat while allowing them to experience the other habitat as low-risk (experiment 2). In the presence of one of the 2 habitat odors, we then tested tadpole responses to a novel odor or a control (experiment 1), or to the learned predator odor or a novel odor control (experiment 2). We predicted that prey experiencing high-risk habitat odor will exhibit habitat-specific neophobia to a novel odor, whereas prey will not elicit a response to the same novel odor when tested in a low-risk habitat odor. We also predicted that tadpoles will associate predators with specific habitat odors and thus show a more pronounced antipredator response to the predator odor when tested in the same habitat odor where the predator cue was first encountered. We did not expect that the context in which prey learn about a predator would completely inhibit predator recognition in a different habitat, as the cost of failed recognition of a known predator is high. METHODS Collection and maintenance We collected 4 wood frog egg clutches (~200–400 eggs per clutch) from small ephemeral ponds around Saskatoon, Canada and transported then to the RJF Smith Centre for Aquatic Ecology at the University of Saskatchewan (Saskatoon, SK). Each clutch was placed in individual outdoor 67-L pools filled with 60 L of aged dechlorinated tap water and maintained under natural light and temperature conditions. The eggs were left until they hatched, at which point the tadpoles were split between six 67-L pools to reduce tadpole density, and algal discs (Wardley) were added to provide food. Tadpoles were given 3 weeks to develop before the experiment began (Gosner stage 25). Cue production To induce high-risk environments, we used repeated exposures of injured tadpole alarm cues, prepared directly before their use in conditioning. Tadpoles were euthanized by a concussive blow to the head using a pestle, which killed the tadpoles instantly, and were then ground up using a pestle and mortar. Twenty milliliters of dechlorinated water were then added to the crushed tadpoles and the solution was filtered to produce our alarm cue stock. For novel odors and conditioned predator odors, we used lake sturgeon, Acipenser fulvescens, and rainbow trout, Oncorhynchus mykiss, which were housed in the RJF Smith Centre for Aquatic Ecology in groups of 4–8. Wood frog tadpoles do not innately recognize rainbow trout (an introduced predator) as a predator (Chivers et al. 2015), and neither rainbow trout or lake sturgeon are found at our collection sites, precluding the potential for embryonic learning. To produce the trout and sturgeon odor, water was collected directly from the housing tanks prior to each conditioning or testing period. Sturgeon cues were diluted by a factor of 12 to standardize cue concentrations to relative body size and number of fish. To produce the habitat cues, we collected dead cattails, Typha latifolia, and a mixture of floating pond weed (including duckweed, Lemna sp., coontail, Ceratophyllum demersum, and water milfoil, Myriophyllum sp.) from Pike Lake, Saskatchewan. The 2 plant groups were selected to represent the odor profiles of 2 common habitats found around the local area but were not present at the site where the egg clutches were collected. This controlled for the fact that tadpoles can learn about risk during embryonic development (Mathis et al. 2008). The pond weed and cattails were then added to two 190-L pools filled with 150 L of aged dechlorinated water. These pools provided the stock habitat water for the conditioning tanks and testing arenas. Approximately half of the water was used each day and the pools were refilled at the end of the testing days. Experiment 1: Test of habitat-specific neophobia The aim of this experiment was to test if tadpoles use olfactory cues of habitats to learn about risk and develop habitat-specific neophobia. Tadpoles from all clutches were randomly added to 5 L glass conditioning tanks containing 3 L of either cattail odor or pond weed odor (24 tanks, 12 of each habitat odor and 20 tadpoles per tank). Each conditioning tank had a small amount of food and a small basket made of black plastic mesh that excluded tadpoles and was filled with plant material matching the stock habitat water. This ensured that habitat odors were constantly present during the conditioning period. The following day (day 1), tadpoles from 12 tanks (6 pond weed water and 6 cattail water) received 20 mL of alarm cues (“high-risk”) at 3 random times during the day (minimum of 1.5 h between exposures). The other 12 tanks received 20 mL dechlorinated water (“low-risk”). Odors were added slowly using a 60-mL syringe. At the end of the day, tadpoles were moved into a clean tank containing matching, fresh habitat odor water. The following day (day 2), tadpoles received the same conditioning cues as the day before; however, at the end of the day tadpoles were moved to tanks containing the opposite habitat odor. On days 3 and 4, tadpoles were conditioned with the alternate risk level cue to the one they received on days 1 and 2. This cycle of 2 days of high-risk and 2 days low-risk was repeated for a second cycle, meaning tadpoles were conditioned for a total of 8 days, with 4 days in a high-risk habitat and 4 days in a low-risk habitat. This exposure regime represented a fully balanced experimental design, with 12 tanks of tadpoles conditioned to recognize the cattail habitat as risky and the 12 tanks of tadpoles conditioned to recognize the pond weed habitat as risky. This design also controlled for any potential order effects associated with the onset of risk. The day after the conditioning phase ended, tadpoles were tested for their response to a novel odor or a control. Tadpoles from all conditioning regimes were transferred into 0.5 L circular testing arenas containing either cattail water or pond weed water. The tadpoles were left to acclimate for at least 1 h before testing. The testing period consisted of a 4-min behavioral observation, where the number of times tadpoles crossed the medial line of the arena was counted. Following the initial observation period, 5 mL of trout odor (novel odor) or dechlorinated water (control) was slowly injected down the side of the arena using a syringe, and tadpole activity was observed for another 4 min. The number of times tadpoles cross the median line provides one measure of locomotive activity levels. Changes (reductions) in activity from pre- to postcue exposure is a well-established assay for measuring the tadpole antipredator response, and hence tadpoles that reduced activity when exposed to a novel odor were considered to be neophobic (Gonzalo et al. 2010; Chivers et al. 2016; Mitchell et al. 2016). Tadpoles were additionally tested in dechlorinated water for their responses to pond weed odor, cattail odor, the novel odor or the water control. This allowed us to determine if tadpoles had learned to associate predation risk with the habitat odors and if the neophobic response was elicited in a novel habitat (dechlorinated water). A total of 221 tadpoles were tested (n = 25–30 per treatment). Experiment 2: Test of habitat specific predator recognition The aim of this experiment was to test whether the habitat in which prey learn to recognize a predator influences how tadpoles respond to the predator during future encounters. The experiment followed a similar design to that of experiment 1. Tadpoles from all clutches were randomly added to 5-L glass tanks containing either pond weed water or cattail water and a basket of the corresponding plant material, as described above. The following day (day 1) tadpoles in 12 tanks (6 cattail and 6 pond weed) were conditioned to recognize trout odor as a predator by adding 20 mL of tadpole alarm cues (containing 3 tadpoles) paired with 10 mL of trout odor. In the remaining 12 tanks, we added 20 mL of dechlorinated water as a control. At the end of the day tadpoles were moved to new tanks containing food and the opposite habitat odor. Alternating the habitats daily allowed tadpoles to gain experience with both the predator-associated habitat and the non-predator habitat. On day 2, tadpoles received the opposite conditioning treatment to the one they received the day before, i.e., tadpoles conditioned to recognize trout on day 1 received the water control, and tadpoles conditioned with water on day 1 were conditioned with the trout odor and alarm cue. Again, at the end of the day, tadpoles were moved to a new tank containing the alternative habitat odors. This process was repeated for the next 4 days (6 days total conditioning period) so that tadpoles experienced 3 days in each habitat and were conditioned to recognize trout in one habitat 3 times. The testing phase started the following day and followed the same procedure as experiment 1, except using different odors as test cues. Tadpoles were placed in arenas containing water from either habitat and then tested with the addition of either trout odor (learned predator) or sturgeon odor (novel odor control). Sturgeon odor acted as a control for both the mechanical disturbance caused by introducing the stimuli and the addition of a chemical odor. A total of 221 tadpoles were tested (n = 24–36 per treatment). Ethical statement Experimental methods followed ASAB and UCACS guidelines for the ethical treatment of animals (UCACS protocol 2015031). Throughout the early developmental period, water was changed daily (50%), and excess food was removed to ensure water parameters were optimal. During the experiment tadpoles were captured using small hand nets and moved between tanks using water filled containers. To produce the alarm cues tadpoles were euthanized according to amphibian guidelines using standard physical methods (Chivers et al. 2016; Crane et al. 2017). Chemical anesthesia could not be used as a method for euthanasia as it may interfere with the chemical alarm cues and the behavior of the tadpoles. There was no tadpole mortality during the collection, hatching and experimental periods. However, there was some minimal mortality (<3%) due to natural causes during early tadpole development prior to the onset of the experiment. At the end of the experiment, all tadpoles were returned to the ponds where they were collected. Fish odors were collected from species already housed in the RJF Smith Centre. Lake sturgeon and rainbow trout were held in 2500-L and 950-L tanks respectively and supplied with dechlorinated water on a flow-through system. Fish were fed daily to satiation, with excess food removed. No trout or sturgeon was handled during the experiment, and there was no mortality. Following the end of the experiment, all fish remained in the aquatics facility to be used in future experiments. Statistical analysis For both experiments, we calculated the proportional change in line crosses from the prestimulus baseline ([post ˗ pre]/pre) and used these values as our response variable for the analyses. All data met the assumptions for homogeneity of variance and normality. Data analysis was conducted using the statistical software package IBM SPSS statistics (IBM). For all experiments, we used nested analysis of variance (ANOVA) designs with Type I sum of squares where the conditioning tank was included as a random factor (nested under treatment) to account for the fact that tadpoles were conditioned in batches and therefore not independent. For experiment 1, we first tested differences in prestimulus activity among treatment groups. We used a 3-way ANOVA to test the effects of high-risk conditioning habitat (either cattails versus pond weed conditioned as high-risk), test habitat (cattails vs. pond weed as habitat odor present during testing), and test odor (novel odor vs. dechlorinated water added to the testing arena after pre-exposure observations) on tadpole line crosses. To interpret significant interactions among factors, we split the data by conditioning habitat and performed post-hoc 2-way nested ANOVAs for high-risk-associated cattail and pond weed habitats, followed by independent sample t-tests to identify differences between testing odors in each testing habitat. We used a Bonferroni correction on the post-hoc t-tests to account for multiple tests being run on the same data (adjusted α = 0.0125). To determine whether tadpoles responded to habitat cues, we used a 2-way nested ANOVA to test the effects of conditioning habitat (cattails vs. pond weed) and test odor (cattail vs. pond weed vs. novel odor vs. water) on the proportional change in line crosses. For experiment 2, we used 3-way nested ANOVAs to test the effects of predator-learning habitat (pond weed vs. cattail), testing habitat (pond weed vs. cattail), and test odors (trout vs. sturgeon) on both the prestimulus baseline activity and the proportional change in lines crossed data (2 separate analyses). RESULTS Experiment 1: Test for habitat specific neophobia Tadpoles did not differ in the number of line crosses during prestimulus observations (high-risk conditioning habitat, all terms P > 0.05). For the change in lines crossed due to the testing odor, the 3-way ANOVA revealed a significant interaction between conditioning habitat, test habitat, and test odor (P = 0.049; Table 1, a; Figure 1). For tadpoles exposed to high predation risk in the presence of pond weed odor, there was a significant effect of test odor (P < 0.005; Table 1, b; Figure 1a) but no effect of test habitat, test habitat × test odor interaction (both P > 0.05) or conditioning tank (P = 0.06). In other words, tadpoles that had experienced high-risk pond weed habitat significantly reduced the number of line crosses when exposed to the novel odor relative to the water control, demonstrating a neophobic response. However, this response was not habitat-specific during testing, as tadpoles tested in both habitat types responded similarly (Figure 1a). Table 1 Results of habitat-specific neophobia experiment, with ANOVAs testing for (a) the effect of high-risk conditioning habitat, test habitat, and test odor on the proportional change in tadpole activity, (b) the effect of test habitat and test odor on the proportional change in tadpole activity for tadpoles conditioned to high-risk in the pond weed habitat, and (c) the effect of test habitat and test odor on the proportional change in tadpole activity for tadpoles conditioned to high-risk in the cattail habitat Source SS df F P (a) Full model High-risk conditioning habitat 0.000 1 0.001 0.972 Error 5.624 21.320 Test habitat 0.093 1 0.469 0.494 Error 38.493 194.103 Test odor 3.193 1 15.642 <0.001 Error 43.277 211.998 High-risk conditioning habitat × Test habitat 0.002 1 0.011 0.915 Error 38.550 194.357 High-risk conditioning habitat × Test odor 0.000 1 0.001 0.969 Error 43.267 211.880 Test habitat × Test odor 0.871 1 4.367 0.038 Error 40.352 202.238 High-risk conditioning habitat × Test habitat × Test odor 0.781 1 3.914 0.049 Error 40.338 202.178 Conditioning tank 5.774 22 1.326 0.158 Error 37.800 191 (b) Pond weed high-risk habitat Test habitat 0.078 1 0.430 0.513 Error 17.878 98.893 Test odor 1.584 1 8.271 0.005 Error 20.445 106.735 Test habitat × Test odor 0.002 1 0.010 0.919 Error 19.440 105.391 Conditioning tank 3.589 11 1.813 0.062 Error 17.457 97 (c) Cattail high-risk habitat Test habitat 0.023 1 0.106 0.745 Error 20.589 95.194 Test odor 1.604 1 7.482 0.007 Error 22.504 104.991 Test habitat × Test odor 1.650 1 7.638 0.007 Error 20.992 97.155 Conditioning tank 2.185 11 0.918 0.527 Error 20.343 94 Source SS df F P (a) Full model High-risk conditioning habitat 0.000 1 0.001 0.972 Error 5.624 21.320 Test habitat 0.093 1 0.469 0.494 Error 38.493 194.103 Test odor 3.193 1 15.642 <0.001 Error 43.277 211.998 High-risk conditioning habitat × Test habitat 0.002 1 0.011 0.915 Error 38.550 194.357 High-risk conditioning habitat × Test odor 0.000 1 0.001 0.969 Error 43.267 211.880 Test habitat × Test odor 0.871 1 4.367 0.038 Error 40.352 202.238 High-risk conditioning habitat × Test habitat × Test odor 0.781 1 3.914 0.049 Error 40.338 202.178 Conditioning tank 5.774 22 1.326 0.158 Error 37.800 191 (b) Pond weed high-risk habitat Test habitat 0.078 1 0.430 0.513 Error 17.878 98.893 Test odor 1.584 1 8.271 0.005 Error 20.445 106.735 Test habitat × Test odor 0.002 1 0.010 0.919 Error 19.440 105.391 Conditioning tank 3.589 11 1.813 0.062 Error 17.457 97 (c) Cattail high-risk habitat Test habitat 0.023 1 0.106 0.745 Error 20.589 95.194 Test odor 1.604 1 7.482 0.007 Error 22.504 104.991 Test habitat × Test odor 1.650 1 7.638 0.007 Error 20.992 97.155 Conditioning tank 2.185 11 0.918 0.527 Error 20.343 94 View Large Table 1 Results of habitat-specific neophobia experiment, with ANOVAs testing for (a) the effect of high-risk conditioning habitat, test habitat, and test odor on the proportional change in tadpole activity, (b) the effect of test habitat and test odor on the proportional change in tadpole activity for tadpoles conditioned to high-risk in the pond weed habitat, and (c) the effect of test habitat and test odor on the proportional change in tadpole activity for tadpoles conditioned to high-risk in the cattail habitat Source SS df F P (a) Full model High-risk conditioning habitat 0.000 1 0.001 0.972 Error 5.624 21.320 Test habitat 0.093 1 0.469 0.494 Error 38.493 194.103 Test odor 3.193 1 15.642 <0.001 Error 43.277 211.998 High-risk conditioning habitat × Test habitat 0.002 1 0.011 0.915 Error 38.550 194.357 High-risk conditioning habitat × Test odor 0.000 1 0.001 0.969 Error 43.267 211.880 Test habitat × Test odor 0.871 1 4.367 0.038 Error 40.352 202.238 High-risk conditioning habitat × Test habitat × Test odor 0.781 1 3.914 0.049 Error 40.338 202.178 Conditioning tank 5.774 22 1.326 0.158 Error 37.800 191 (b) Pond weed high-risk habitat Test habitat 0.078 1 0.430 0.513 Error 17.878 98.893 Test odor 1.584 1 8.271 0.005 Error 20.445 106.735 Test habitat × Test odor 0.002 1 0.010 0.919 Error 19.440 105.391 Conditioning tank 3.589 11 1.813 0.062 Error 17.457 97 (c) Cattail high-risk habitat Test habitat 0.023 1 0.106 0.745 Error 20.589 95.194 Test odor 1.604 1 7.482 0.007 Error 22.504 104.991 Test habitat × Test odor 1.650 1 7.638 0.007 Error 20.992 97.155 Conditioning tank 2.185 11 0.918 0.527 Error 20.343 94 Source SS df F P (a) Full model High-risk conditioning habitat 0.000 1 0.001 0.972 Error 5.624 21.320 Test habitat 0.093 1 0.469 0.494 Error 38.493 194.103 Test odor 3.193 1 15.642 <0.001 Error 43.277 211.998 High-risk conditioning habitat × Test habitat 0.002 1 0.011 0.915 Error 38.550 194.357 High-risk conditioning habitat × Test odor 0.000 1 0.001 0.969 Error 43.267 211.880 Test habitat × Test odor 0.871 1 4.367 0.038 Error 40.352 202.238 High-risk conditioning habitat × Test habitat × Test odor 0.781 1 3.914 0.049 Error 40.338 202.178 Conditioning tank 5.774 22 1.326 0.158 Error 37.800 191 (b) Pond weed high-risk habitat Test habitat 0.078 1 0.430 0.513 Error 17.878 98.893 Test odor 1.584 1 8.271 0.005 Error 20.445 106.735 Test habitat × Test odor 0.002 1 0.010 0.919 Error 19.440 105.391 Conditioning tank 3.589 11 1.813 0.062 Error 17.457 97 (c) Cattail high-risk habitat Test habitat 0.023 1 0.106 0.745 Error 20.589 95.194 Test odor 1.604 1 7.482 0.007 Error 22.504 104.991 Test habitat × Test odor 1.650 1 7.638 0.007 Error 20.992 97.155 Conditioning tank 2.185 11 0.918 0.527 Error 20.343 94 View Large Figure 1 View largeDownload slide Test for habitat-specific neophobia (experiment 1). Mean (± 1 SE) proportional change in line crosses for tadpoles exposed to either a novel odor (gray) or control water (white) when tested in water from either pond weed habitat or cattail habitat. Tadpoles were initially conditioned to learn the odor of pond weed odor (top panel) or cattail habitat (bottom panel) as a high-risk habitat and the alternative habitat as low-risk. Figure 1 View largeDownload slide Test for habitat-specific neophobia (experiment 1). Mean (± 1 SE) proportional change in line crosses for tadpoles exposed to either a novel odor (gray) or control water (white) when tested in water from either pond weed habitat or cattail habitat. Tadpoles were initially conditioned to learn the odor of pond weed odor (top panel) or cattail habitat (bottom panel) as a high-risk habitat and the alternative habitat as low-risk. When tadpoles experienced high risk in the cattail habitat odor, there was a significant interaction between test habitat and test odor (P = 0.007; Table 1, c; Figure 1b). Tadpoles that had experienced high-risk in the cattail habitat and were tested in cattail odor water reduced the number of line crosses when exposed to the novel odor versus water (t53 = 3.63, P = 0.001; adjusted α = 0.0125). However, when tested in pond weed habitat odor, this same group of tadpoles did not respond to the novel odor relative to the water control (t52 = −0.042, P = 0.97; adjusted α = 0.0125). There was no difference in the response of high-risk cattail habitat tadpoles to the water control when tested in cattail versus pond weed (t51 = −1.96, P = 0.055: adjusted α = 0.0125). Tadpoles showed a stronger but nonsignificant response to the novel odor when tested in the cattail habitat odor than in the pond weed odor (t54 = 2.03, P = 0.048; adjusted α = 0.0125). Together, these results demonstrate that tadpoles developed habitat-specific neophobia after experiencing high risk in the cattail habitat. When tadpoles were tested in control water for their response to habitat odors and the novel odor, we found that there was no effect of conditioning habitat (F1, 21.07 = 0.29, P = 0.60), test odor (F3, 220.64 = 0.91, p = 0.44), or interaction between the two (F3, 2221.09 = 0.77, P = 0.51) (Figure 2). Additionally, there was no effect of conditioning tank (F22, 203 = 1.41, P = 0.11). Figure 2 View largeDownload slide Learned responses to habitat odors (experiment 1). Mean (± 1 SE) proportional change in line crosses for tadpoles exposed to either pond weed water, cattail water, novel odor or control water when tested in control water. Tadpoles were initially conditioned to learn the odor of pond weed odor (top panel) or cattail habitat (bottom panel) as a high-risk habitat and the alternative habitat as low-risk. Figure 2 View largeDownload slide Learned responses to habitat odors (experiment 1). Mean (± 1 SE) proportional change in line crosses for tadpoles exposed to either pond weed water, cattail water, novel odor or control water when tested in control water. Tadpoles were initially conditioned to learn the odor of pond weed odor (top panel) or cattail habitat (bottom panel) as a high-risk habitat and the alternative habitat as low-risk. Experiment 2: Test for habitat specific predator learning There was no difference in tadpole line crosses during the prestimulus observations (high-risk conditioning habitat, all terms P > 0.05). The change in line crosses made by tadpoles was significantly affected by the test odor (F1, 207.3 = 20.39, P < 0.001; Figure 3). Across all groups, tadpoles reduced line crosses when exposed to the predator odor (~30–40% reduction) versus the novel odor (<15% reduction in activity). There was no effect of conditioning habitat, testing habitat or any interactions among factors (all P > 0.05). There was an effect of conditioning tank (F22, 191 = 1.73, P = 0.015). The significant effect of odor indicates that the tadpoles recognized the trout as a predator but the absence of an interaction among the 3 factors indicates that the tadpoles did not learn to perceive the predator as more dangerous in a specific habitat. Figure 3 View largeDownload slide Test for habitat specific predator learning (experiment 2). Mean (± 1 SE) proportional change in line crosses for tadpoles exposed to either the predator odor (gray) or the novel odor (white) when tested in water from either pond weed habitat or cattail habitat. Tadpoles were initially conditioned to learn the predator odor as a threat in either pond weed odor (top panel) or cattail habitat (bottom panel) and experienced only control water (low risk) in the alternative habitat. Figure 3 View largeDownload slide Test for habitat specific predator learning (experiment 2). Mean (± 1 SE) proportional change in line crosses for tadpoles exposed to either the predator odor (gray) or the novel odor (white) when tested in water from either pond weed habitat or cattail habitat. Tadpoles were initially conditioned to learn the predator odor as a threat in either pond weed odor (top panel) or cattail habitat (bottom panel) and experienced only control water (low risk) in the alternative habitat. DISCUSSION Many studies have investigated how prey alter their behavior under predation risk as it varies across landscapes, yet little is known about how prey learn habitat-specific risk or which cues prey use to learn about predators across contexts. Our results demonstrate that prey use olfactory cues from habitats to learn about changes in predation risk between habitats. In experiment 1, tadpoles that experienced high risk in the cattail habitat water and were subsequently tested in the cattail water reduced activity in response to a novel odor (neophobic response), but these tadpoles did not respond to the novel odor when tested in the pond weed water. Tadpoles also showed no behavioral response to the novel odor when tested in the control water, irrespective of the habitat in which tadpoles experienced risk. Thus, the onset of neophobia was unique to the habitat where tadpoles experienced high risk. Neophobic responses have been shown to increase survival (Ferrari et al. 2014) and are one of several antipredator traits that is expressed when prey experience periods of high predation risk (Ferrari et al. 2015). The rapid onset of neophobia in tadpoles entering high-risk habitats should enhance survival while they adjust to their new surroundings. However, such traits are maladaptive in low-risk environments (Chivers et al. 2014), hence the lack of neophobic responses in the low-risk pond weed habitat or control water. The results presented here demonstrate that tadpoles can use olfactory cues from habitats to transition between these high-risk and low-risk antipredator phenotypes and optimize their performance in response to variable risk at landscape scales. While there was clear evidence for habitat-specific neophobia, antipredator responses were not consistent across the background habitat treatments. Tadpoles that had experienced high risk in the pond weed habitat displayed a neophobic response to the novel odor when tested in both pond weed and cattail water, perceiving the cattail habitat as high risk despite the fact they did not experience high risk in that habitat. However, there were no neophobic responses when tadpoles were tested in control water, indicating that the neophobic responses were limited to the habitats experienced during conditioning. These apparently conflicting results may be in part explained by how prey develop antipredator responses to different cues. It is possible that tadpoles innately recognized the cattail habitat as risky. Consequently, tadpoles would display neophobic responses without having to experience risk in the cattail habitat. Studies have shown that animals can innately recognize cues from specific habitats, e.g., both sea lamprey, Petromyzon marinus, and anemonefish, Amphiprion percula, show innate recognition of olfactory cues from preferred settlement habitats (Vrieze et al. 2010; Dixson et al. 2014). For tadpoles to possess innate recognition of cattail habitat odors as risky, exposure to consistently high predation risk in cattail habitats over multiple generations is likely required. Given the diversity and distribution of predators that consume tadpoles across different habitats, we suggest that such an innate response is unlikely to have developed for cattail habitats and not pond weed habitats. Alternatively, tadpoles may have generalized recognition of risk from the pond weed cues to the cattail cues but in that direction only. Generalized recognition occurs when a stimulus has enough similarity to a conditioned odor that it elicits a similar response to that of the conditioned stimulus (Ghirlanda and Enquist 2003; Shettleworth 2010). Unidirectional generalization may have occurred if a chemical was a common component of the pond weed odor but was present only as a minor component of the cattail odor. Tadpoles would then generalize high-risk from the pond weed to the cattail habitat but generalization would not go in the opposite direction. Finally, the conflicting results may have been due to variation in chemical stability of habitat cues during the experiment. As the cattail plants were dead, the suite of chemical compounds released may have change over time as the plants decayed. Previous studies have shown that variation in the stimuli being learned can slow or inhibit the learning process (Craig 1994a; Craig 1994b). For example, stingless bees, Trigona fuscipennis, can learn to avoid webs spun by orb spinner, Argiope argentata, but the learning process is delayed due to A. argentata using different silk patterns to decorate its web each day. Additionally, learning about safe stimuli requires a greater number of conditioning events than learning about risky stimuli (Ferrari and Chivers 2011), and is highly context sensitive (Lubow 1989). A combination of the variation in the odor complex being learned and the information being encoded (safety of the habitat) may have prevented tadpoles from learning the difference in risk levels between the 2 habitats and resulted in learning risk across both habitats. Variation in cattail cues would not have prevented tadpoles from learning the cattail habitat as risky, as only one conditioning event is sufficient to cause associative learning of an odor as risky. In experiment 2, we found that habitat had no effect on how prey learned about predator odors. Irrespective of the habitat in which prey were conditioned to recognize the predator, prey responded to predator cues with similar reductions in activity when tested in either habitat odor. The fact that wild caught animals still respond to predator cues in captive settings also suggests that context does not influence learned predator recognition (Griffin 2004). Ecologically, such responses are appropriate, as prey should generally adopt a cautionary approach and overestimate risk (Bouskila and Blumstein 1992). Hence, when tadpoles encounter a known predator in a safe environment, tadpoles should respond to the predator as in previous encounters in other habitats. Despite this, it should be noted that the lack of variation in response to predator cues across habitats does not mean prey do not incorporate information from habitats when learning about predator specific risk, simply that the optimal response in our experiment was to respond with an equal intensity across both habitats. Future research should assess conflicting information about a predator across habitats, as has been done for prey learning temporal patterns of risk (Ferrari et al. 2009; Bosiger et al. 2012). Despite the potential costs, learning about risk provides an adaptive strategy for prey to modify their antipredator responses to their current environment. While studies have shown how prey respond to changes in risk through time, few have looked at how prey learn about risk across different contexts. As demonstrated in this study olfactory cues from habitats provide information that allows prey to learn about risk in different habitats. Yet our results also show that there are nuances to how prey use olfactory cues from different habitats which require further investigation. It might be that prey may need additional information from other sensory systems to resolve such issues. To understand how prey respond to variable risk between habitats, additional work exploring how prey use ancillary cues from their environment to inform antipredator responses is required. Such studies will also provide insights into how human altered environments might impact predator–prey interactions. The potential value of such studies is highlighted by recent findings from coral reefs, where changes in the chemical milieu of habitats following a disturbance event appears to impair the ability of fish to detect chemical alarm cues (Ferrari et al. 2017; McCormick et al. 2017). FUNDING This work was supported by Natural Sciences and Engineering Research Council of Canada grants to M.C.O.F. and D.P.C. Data accessibility: Analyses reported in this article can be reproduced using the data provided by Mitchell et al. (2018). We thank Laurel Sacco and Phil Anderson for their assistance hunting for wood frog egg clutches. REFERENCES Abrams PA , Rowe L . 1996 . The effects of predation on the age and size of maturity of prey . Evolution . 50 : 1052 – 1061 . Google Scholar CrossRef Search ADS PubMed Baldwin RF , Calhoun AJ , deMaynadier PG . 2006 . Conservation planning for amphibian species with complex habitat requirements: a case study using movements and habitat selection of the wood frog Rana sylvatica . J Herpetol . 40 : 442 – 453 . Google Scholar CrossRef Search ADS Beauchamp G , Ruxton GD . 2011 . A reassessment of the predation risk allocation hypothesis: a comment on Lima and Bednekoff . Am Nat . 177 : 143 – 146 . Google Scholar CrossRef Search ADS PubMed Bland JD , Temple SA . 1990 . Effects of predation-risk on habitat use by Himalayan Snowcocks . Oecologia . 82 : 187 – 191 . Google Scholar CrossRef Search ADS PubMed Bosiger YJ , Lonnstedt OM , McCormick MI , Ferrari MC . 2012 . Learning temporal patterns of risk in a predator-diverse environment . PLoS One . 7 : e34535 . Google Scholar CrossRef Search ADS PubMed Bouskila A , Blumstein DT . 1992 . Rules of thumb for predation hazard assessment: predictions from a dynamic model . Am Nat . 139 : 161 – 176 . Google Scholar CrossRef Search ADS Brown C , Braithwaite VA . 2005 . Effects of predation pressure on the cognitive ability of the poeciliid Brachyraphis episcopi . Behav Ecol . 16 : 482 – 487 . Google Scholar CrossRef Search ADS Brown GE , Elvidge CK , Ramnarine I , Ferrari MCO , Chivers DP . 2015 . Background risk and recent experience influences retention of neophobic responses to predators . Behav Ecol Sociobiol . 69 : 737 – 745 . Google Scholar CrossRef Search ADS Brown GE , Ferrari MC , Elvidge CK , Ramnarine I , Chivers DP . 2013 . Phenotypically plastic neophobia: a response to variable predation risk . Proc Biol Sci . 280 : 20122712 . Google Scholar CrossRef Search ADS PubMed Brown JS , Laundre JW , Gurung M . 1999 . The Ecology of Fear: optimal foraging, game theory, and trophic interactions . J Mammal . 80 : 385 – 399 . Google Scholar CrossRef Search ADS Chivers DP , Mathiron A , Sloychuk JR , Ferrari MC . 2015 . Responses of tadpoles to hybrid predator odours: strong maternal signatures and the potential risk/response mismatch . Proc Biol Sci . 282 : 20150365 . Google Scholar CrossRef Search ADS PubMed Chivers DP , McCormick MI , Mitchell MD , Ramasamy RA , Ferrari MCO . 2014 . Background level of risk determines how prey categorize predators and non-predators . Proc R Soc B Biol Sci . 281 : 20140355 . Google Scholar CrossRef Search ADS Chivers DP , McCormick MI , Warren DT , Allan BJM , Ramasamy RA , Arvizu BK , Glue M , Ferrari MCO . 2017 . Competitive superiority versus predation savvy: the two sides of behavioural lateralization . Anim Behav . 130 : 9 – 15 . Google Scholar CrossRef Search ADS Chivers DP , Mirza RS . 2001 . Importance of predator diet cues in responses of larval wood frogs to fish and invertebrate predators . J Chem Ecol . 27 : 45 – 51 . Google Scholar CrossRef Search ADS PubMed Chivers DP , Mitchell MD , Lucon-Xiccato T , Brown GE , Ferrari MCO . 2016 . Background risk influences learning but not generalization of predators . Anim Behav . 121 : 185 – 189 . Google Scholar CrossRef Search ADS Craig CL . 1994a . Predator foraging behavior in response to perception and learning by its prey: interactions between orb-spinning spiders and stingless bees . Behav Ecol Sociobiol . 35 : 45 – 52 . Google Scholar CrossRef Search ADS Craig CL . 1994b . Limits to learning: effects of predator pattern and colour on perception and avoidance-learning by prey . Anim Behav . 47 : 1087 – 1099 . Google Scholar CrossRef Search ADS Crane AL , Demuth BS , Ferrari MCO . 2017 . Experience with predators shapes learning rules in larval amphibians . Behav Ecol . 28 : 312 – 318 . Google Scholar CrossRef Search ADS Crane AL , Ferrari MCO . 2017 . Patterns of predator neophobia: a meta-analytic review . Proc R Soc B Biol Sci . 284 : 20170583 . Google Scholar CrossRef Search ADS Dall SR . 2010 . Managing risk: the perils of uncertainty . In: Westneat DF , Fox CW , editors. Evolutionary behavioral ecology . New York (NY) : Oxford University Press, Inc . p. 194 – 206 . Dayton GH , Saenz D , Baum KA , Langerhans RB , DeWitt TJ . 2005 . Body shape, burst speed and escape behavior of larval anurans . Oikos . 111 : 582 – 591 . Google Scholar CrossRef Search ADS Denno RF , Finke DL , Langellotto GA . 2005 . Direct and indirect effects of vegetation structure and habitat complexity on predator-prey and predator-predator interactions . In: Barbosa P , Castellanos I , editors. Ecology of predator-prey interactions . Oxford : Oxford University Press . p. 211 – 239 . Dixson DL , Jones GP , Munday PL , Planes S , Pratchett MS , Thorrold SR . 2014 . Experimental evaluation of imprinting and the role innate preference plays in habitat selection in a coral reef fish . Oecologia . 174 : 99 – 107 . Google Scholar CrossRef Search ADS PubMed Engstrom-Ost J , Lehtiniemi M . 2004 . Threat-sensitive predator avoidance by pike larvae . J Fish Biol . 65 : 251 – 261 . Google Scholar CrossRef Search ADS Ferrari MCO , Chivers DP . 2011 . Learning about non-predators and safe places: the forgotten elements of risk assessment . Anim Cogn . 14 : 309 – 316 . Google Scholar CrossRef Search ADS PubMed Ferrari MCO , Chivers DP , Wisenden BD . 2010 . Chemical ecology of predator-prey interactions in aquatic ecosystems: a review and prospectus . Can . 88 : 698 – 724 . Ferrari MCO , Manek AK , Chivers DP . 2009 . Temporal learning of predation risk by embryonic amphibians . Biol Lett . 6 : 308 – 310 . Google Scholar CrossRef Search ADS PubMed Ferrari MCO , McCormick MI , Allan BJM , Chivers DP . 2017 . Not equal in the face of habitat change: closely related fishes differ in their ability to use predation-related information in degraded coral . Proc R Soc B Biol Sci . 284 : 20162758 . Google Scholar CrossRef Search ADS Ferrari MC , McCormick MI , Allan BJ , Choi R , Ramasamy RA , Johansen JL , Mitchell MD , Chivers DP . 2015 . Living in a risky world: the onset and ontogeny of an integrated antipredator phenotype in a coral reef fish . Sci Rep . 5 : 15537 . Google Scholar CrossRef Search ADS PubMed Ferrari MCO , McCormick MI , Meekan MG , Chivers DP . 2014 . Background level of risk and the survival of predator-naive prey: can neophobia compensate for predator naivety in juvenile coral reef fishes ? Proc R Soc B Biol Sci . 282 : 20142197 . Google Scholar CrossRef Search ADS Ghirlanda S , Enquist M . 2003 . A century of generalization . Anim Behav . 66 : 15 – 36 . Google Scholar CrossRef Search ADS Gonzalo A , López P , Martín J . 2010 . Risk level of chemical cues determines retention of recognition of new predators in Iberian green frog tadpoles . Behav Ecol Sociobiol . 64 : 1117 – 1123 . Google Scholar CrossRef Search ADS Griffin AS . 2004 . Social learning about predators: a review and prospectus . Anim. Learn Behav . 32 : 131 – 140 . Google Scholar CrossRef Search ADS Huntingford FA , Wright PJ . 1989 . How sticklebacks learn to avoid dangerous feeding patches . Behav Processes . 19 : 181 – 189 . Google Scholar CrossRef Search ADS PubMed Laforsch C , Ngwa W , Grill W , Tollrian R . 2004 . An acoustic microscopy technique reveals hidden morphological defenses in Daphnia . Proc Natl Acad Sci USA . 101 : 15911 – 15914 . Google Scholar CrossRef Search ADS PubMed Laundré JW , Hernández L , Ripple WJ . 2010 . The landscape of fear: ecological implications of being afraid . Open Ecol J . 3 : 1 – 7 . Google Scholar CrossRef Search ADS Laurila A , Kujasalo J , Ranta E . 1998 . Predator-induced changes in life history in two anuran tadpoles: effects of predator diet . Oikos . 83 : 307 . Google Scholar CrossRef Search ADS Lima SL , Bednekoff PA . 1999 . Temporal variation in danger drives antipredator behavior- The Predation Risk Allocation Hypothesis . Am Nat . 153 : 649 – 659 . Google Scholar CrossRef Search ADS PubMed Lubow RE . 1989 . Latent inhibition and conditioned attention theory . Cambridge : Cambridge University Press . Google Scholar CrossRef Search ADS Martin CW , Fodrie FJ , Heck KL Jr , Mattila J . 2010 . Differential habitat use and antipredator response of juvenile roach (Rutilus rutilus) to olfactory and visual cues from multiple predators . Oecologia . 162 : 893 – 902 . Google Scholar CrossRef Search ADS PubMed Mathis A , Ferrari MCO , Windel N , Messier F , Chivers DP . 2008 . Learning by embryos and the ghost of predation future . Proc R Soc B Biol Sci . 275 : 2603 – 2607 . Google Scholar CrossRef Search ADS Mathis A , Unger S . 2012 . Learning to avoid dangerous habitat types by aquatic salamanders, Eurycea tynerensis: learning to avoid dangerous habitats by salamanders . Ethology . 118 : 57 – 62 . Google Scholar CrossRef Search ADS McCormick MI , Chivers DP , Allan BJ , Ferrari MC . 2017 . Habitat degradation disrupts neophobia in juvenile coral reef fish . Glob Chang Biol . 23 : 719 – 727 . Google Scholar CrossRef Search ADS PubMed Mitchell MD , Chivers DP , Brown GE , Ferrari MCO . 2016 . Living on the edge: how does environmental risk affect the behavioural and cognitive ecology of prey ? Anim Behav . 115 : 185 – 192 . Google Scholar CrossRef Search ADS Mitchell MD , Crane AL , Bairos-Novak KR , Ferrari MCO , Chivers DP . 2018 . Data from: olfactory cues of habitats facilitate learning about landscapes of fear . Dryad Digital Repository . http://dx.doi.org/10.5061/dryad.218p0v2 Orrock JL , Preisser EL , Grabowski JH , Trussell GC . 2013 . The cost of safety: refuges increase the impact of predation risk in aquatic systems . Ecology . 94 : 573 – 579 . Google Scholar CrossRef Search ADS PubMed Preisser EL , Bolnick DI , Benard MF . 2005 . Scared to death? The effects of intimidation and consumption in predator-prey interactions . Ecology . 86 : 501 – 509 . Google Scholar CrossRef Search ADS Preisser EL , Orrock JL , Schmitz OJ . 2007 . Predator hunting mode and habitat domain alter nonconsumptive effects in predator-prey interactions . Ecology . 88 : 2744 – 2751 . Google Scholar CrossRef Search ADS PubMed Relyea RA . 2001 . Morphological and behavioral plasticity of larval anurans in response to different predators . Ecology . 82 : 523 – 540 . Google Scholar CrossRef Search ADS Rilov G , Figueira WF , Lyman SJ , Crowder L . 2007 . Complex habitats may not always benefit prey: linking visual field with reef fish behavior and distribution . Mar Ecol Prog Ser . 329 : 225 – 238 . Google Scholar CrossRef Search ADS Schmitz OJ , Trussell GC . 2016 . Multiple stressors, state-dependence and predation risk — foraging trade-offs: toward a modern concept of trait-mediated indirect effects in communities and ecosystems . Curr Opin Behav Sci . 12 : 6 – 11 . Google Scholar CrossRef Search ADS Shettleworth SJ . 2010 . Cognition, evolution and behavior . 2nd ed . New York : Oxford University Press . Skow CD , Jakob EM . 2005 . Jumping spiders attend to context during learned avoidance of aposematic prey . Behav Ecol . 17 : 34 – 40 . Google Scholar CrossRef Search ADS Tollrian R . 1993 . Neckteeth formation in Daphnia pulex as an example of continuous phenotypic plasticity: morphological effects of Chaoborus kairomone concentration and their quantification . J Plankton Res . 15 : 1309 – 1318 . Google Scholar CrossRef Search ADS Vrieze LA , Bjerselius R , Sorensen PW . 2010 . Importance of the olfactory sense to migratory sea lampreys Petromyzon marinus seeking riverine spawning habitat . J Fish Biol . 76 : 949 – 964 . Google Scholar CrossRef Search ADS Werner EE , Gilliam JF , Hall DJ , Mittelbach GG . 1983 . An experimental test of the effects of predation risk on habitat use in fish . Ecology . 64 : 1540 – 1548 . Google Scholar CrossRef Search ADS Werner EE , Peacor SD . 2003 . A review of trait-mediated indirect interactions in ecological communities . Ecology . 84 : 1083 – 1100 . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Behavioral EcologyOxford University Press

Published: Apr 5, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off