Behavioral Ecology Advance Access originally published online on December 12, 2007
Behavioral Ecology 2008 19(2):263-271; doi:10.1093/beheco/arm133
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Survival benefits and divergence of predator-induced behavior between pumpkinseed sunfish ecomorphs
Department of Integrative Biology, University of Guelph, Guelph, Ontario N1G 2W1, Canada
Address correspondence to B.W. Robinson. E-mail: berenrob{at}uoguelph.ca. J.C. Koblitz is now at the Animal Physiology, Department of Biology, University of Tübingen, Auf der Morgenstelle 28, 72076 Tuebingen, Germany.
Received 13 April 2007; revised 16 October 2007; accepted 25 October 2007.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Resource use is widely thought to influence adaptive phenotypic divergence, whereas other ecological factors, such as predation, are frequently overlooked, particularly in studies of polyphenism in fishes. Juvenile pumpkinseed sunfish (Lepomis gibbosus) reared with predatory walleye (Sander vitreus) increase body depth and dorsal spine length, indicating that developmental responses to predation can shape phenotype. Body form responses to the same predator cues though have also evolutionarily diverged between sunfish ecomorphs that coexist in single lake populations by inhabiting either littoral or pelagic habitats, suggesting that predation risk varies between habitats. Here, we test if prior exposure to predator cues influences the development of behavior in juvenile pumpkinseed sunfish, if behavioral responses to the same predator cues vary between ecomorphs, and if induced phenotypic variation affects survival under predation. Behavior depended strongly on prior exposure to predator cues, but this effect varied between sunfish ecomorphs, indicating that ecomorphs have different responses to the same predator cues. Predator-induced phenotypes had higher survival than control phenotypes under simulated littoral but not pelagic conditions. Predator-induced phenotypic responses are candidate-inducible defenses, and divergent responses between ecomorphs suggest that they can evolve in response to selection imposed by differences in habitat-specific predation risk.
Key words: evolution, induced defense, Lepomis gibbosus, phenotypic plasticity, polymorphism, predation.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Predation has important ecological (Werner 1994
) and evolutionary consequences in fishes (Langerhans et al. 2004; Vamosi and Schluter 2004; Vamosi 2005), mainly because traits related to accomplishing and especially avoiding predation are likely to be under strong selection because of their large effect on fitness (Abrams 2000). Evolutionary responses include genetic differentiation of developmentally canalized "hardwired" defensive traits (Levins 1968) or the evolution of adaptive phenotypic plasticity or flexibility in response to predation cues (Stearns 1989), the classic dichotomy of "nature" (or traits thought to be of evolutionary importance) versus "nurture" (of developmental importance). A third hypothesis combines these by asking if populations can genetically diverge with respect to their phenotypic or developmentally plastic responses, such as larger (increased plasticity) or smaller responses (increased canalization) and the direction of phenotypic responses (West-Eberhard 1989; Schlichting and Pigliucci 1998; Parsons and Robinson 2006). Tests of this "nature of nurture" hypothesis are rare but useful because they provide insights into how environmental influences on development can evolve and contribute to or constrain diversification (Papaj 1994; West-Eberhard 2003; Pigliucci 2005).
Studies of the genetics and evolution of environmentally mediated development or phenotypic plasticity attempt to consolidate the aspects of developmental genetics, ecology and evolution and are contributing to a nascent field recently termed "eco-devo" (Gilbert 2001). This synthetic approach can be applied to any environmentally responsive trait that influences performance and ultimately fitness including flexible behavior such as learning. Adaptive learning has referred both to changes in behavior in an individual that result from prior experience (reflecting development or nurture) and to changes in learning in a population occurring over generations that result from evolutionary responses to selection (evolution of the nature of nurture) (Papaj 1994). The extent to which learning can evolve and contribute to the genetic divergence of populations remains an exciting question that can be addressed in the context of behavioral responses by prey to predators.
Predator-induced responses that affect survival are referred to as inducible defenses and are best known in plants and aquatic invertebrates (Tollrian and Harvell 1999). Morphological and chemical responses are more common in plants, whereas a suite of behavioral, life history, and morphological responses occur in aquatic invertebrates (Straile and Halbich 2000). Vertebrates are better known for behavioral rather than morphological responses to predation cues, such as changes in activity, refuge, and habitat use (Werner 1994; Eklöv and Svanbäck 2006). The chemical cues underlying behavioral responses to predation are well studied in fish (Pfeiffer 1977; Mathis and Smith 1993; Brown et al. 1995, 2001; Brown and Brennan 2000; Chivers et al. 2001; Smith and Belk 2001; Brown and Zachar 2002; Mirza and Chivers 2002; Leduc et al. 2003; Marcus and Brown 2003). Behavioral responses to the same cues can also diverge between populations with and without local predators (Magurran 1990). Predator-induced morphological responses in fish though appear more rare. The best example involves the crucian carp (Carassius carassius), which reduces mortality in the presence of piscivorous pike (Esox lucius) by increasing body depth, thus superseding the gape limit of pike (Brönmark et al. 1999). This response appears adaptive because carp develop a slimmer body form that may reduce the energetic costs of swimming in the absence of pike. Juvenile perch (Perca fluviatilis) and roach (Rutilus rutilus) can also change body form in response to pike predation (Eklöv and Jonsson 2007). Behavioral responses to predation risk can indirectly influence morphological variation, however, when predator-induced shifts in habitat lead to shifts in diet with subsequent affects on growth and morphological development of prey (Eklöv and Svanbäck 2006). Thus, induced morphological responses by prey fish to predation cues seem rare and likely covary with induced behavioral responses.
Evolutionary models of adaptive inducible defenses usually require 1) a developmental system that can produce alternate phenotypes, 2) heritable variation in plasticity, 3) a system to match phenotype to environment using reliable cues, and 4) that divergent selection favors alternate phenotypes under different and stochastically available environments (although stochasticity is not a necessary prerequisite for the evolution of plastic responses; Parsons and Robinson 2006). Criteria (1) and (2) have been verified in various fish (Robinson and Parsons 2002), and criterion (3) is met in a variety of fish that detect and behaviorally respond to chemical cues (see above). Criterion (4) is more challenging because it can be difficult to assess how predation risk favors different phenotypic optima (Brönmark and Miner 1992; Magurran et al. 1993; Svanbäck and Eklöv 2004; Vamosi and Schluter 2004). Nonetheless, this criterion predicts that induced character states influence performance and ultimately fitness in alternate environments.
We are involved in a long-term study of the proximal and ultimate causes of variation in polyphenic pumpkinseed sunfish (Lepomis gibbosus) in the postglacial lakes of North America (Jastrebski and Robinson 2004; Parsons and Robinson 2006). In numerous populations, divergent ecomorphs coexist by inhabiting either inshore shallow water (hereafter "littoral") or offshore deeper water ("pelagic") habitats.
In a previous study (Januszkiewicz and Robinson 2007), we demonstrated that juvenile pumpkinseed sunfish change their body form in response to predation cues released by walleye (Sander vitreus; Nelson et al. 2003) fed conspecific sunfish. Walleye cues induced an approximately 14% increase in body depth in both ecomorphs compared with control sunfish. However, the response of dorsal spine length to predation cues was 10% greater in pelagic compared with littoral ecomorphs (in the absence of predator cues, spine lengths were shorter in pelagic compared with littoral ecomorphs). If these ecomorphs have otherwise shared a common environment, then this suggests that induced spine responses have genetically diverged between ecomorphs, perhaps, in response to variation in predation risk between lake habitats. For example, small sunfish experience greater predation risk in the open pelagic compared with shallow and vegetated littoral habitats (Werner 1994; Chipps et al. 2004).
Predation risk is reduced whenever the chance of a prey being detected, identified, approached, subjugated, or consumed are reduced (Endler 1986). Increased prey body depth provides a size refuge against gape-limited predators in fish for purely physical reasons (Brönmark and Miner 1992; Nilsson et al. 1995; Einfalt and Wahl 1997; Brönmark et al. 1999) but also because larger prey cause predators to select less costly alternate prey that require reduced handling time (Wahl and Stein 1988; Brönmark et al. 1999) or cause predators to redirect their attacks away from the center of prey mass thereby reducing the effectiveness of attack (Webb 1986). Spines reduce mortality in small invertebrates under fish predation (Tollrian and Harvell 1999; Straile and Halbich 2000) although this may not be true in macroinvertebrates (Mikolajewski et al. 2006). Spines can also reduce mortality in fish prey by influencing how predators select, attack, and manipulate prey after capture (Reimchen 1991; Einfalt and Wahl 1997; Vamosi and Schluter 2004). Larger dorsal spines also increase apparent size when extended and so should increase the handling costs of predators through increased risk of injury during prey manipulation and swallowing (Hoyle and Keast 1987; Reimchen 1991; Johansson and Samuelsson 1994). Behaviors that reduce predation risk include prey shifts to lower risk habitats, decreased activity, increased vigilance, and formation of schools (Werner 1994; Chipps et al. 2004; Eklöv and Svanbäck 2006).
Both resource availability and predator risk can vary between lake habitats and so contribute together to diversifying selection that results in the adaptive divergence of populations (Schluter 1996; Vamosi and Schluter 2002; Rundle et al. 2003; Chipps et al. 2004; Vamosi 2005). Lake habitats vary with respect to a variety of biotic (e.g., predators and prey resources) and abiotic conditions (e.g., light, temperature, oxygen concentrations; (Smith and Todd 1984) that influence the trade-off between energy gain and predation risk (Dill 1983; Endler 1986; Lima 1998). In littoral habitats, fish can use structural complexity to feed without detection by employing cryptic coloration and by reducing activity (Savino and Stein 1982, 1989a, 1989b). Small sunfishes (<51–83 mm body length) refuge from fish predators in littoral habitats (Werner and Hall 1988), and this can result in severe juvenile competitive bottlenecks that influences population dynamics (Mittelbach et al. 1988; Osenberg et al. 1992). Conversely, evasion is improved in the structurally simple open water habitats by schooling and increasing the distance between prey and potential predators (Savino and Stein 1989b; Einfalt and Wahl 1997; Chipps et al. 2004).
We build on our previous study of predator-induced morphological variation in juvenile sunfish by first asking if behavior expressed by pumpkinseed sunfish is influenced by prior exposure to predator cues. In other words, do juvenile sunfish learn from or acclimate to prior experience? We then compare the forms of induced behavioral responses between 2 sunfish ecomorphs to test if these responses have diverged. Finally, using artificial pools that simulate littoral and pelagic conditions, we test if phenotypic variation induced by predator cues is related to survival of sunfish ecomorphs in their respective native habitats. If induced characters are advantageous, then we predict that under walleye predation, the survival of predator-induced phenotypes will be greater than that of control phenotypes reared without predation cues.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Source of predator-induced and -naive sunfish phenotypes
Fish in this study were used to test for both predator-induced morphological and behavioral responses in sunfish ecomorphs. The analysis of morphology is presented in (Januszkiewicz and Robinson 2007
; reviewed above). The sunfish were young of year 0+ and 1+ juveniles collected from Salmon Trout Lake in the late summer and early fall of 2003. This population was chosen because it coexists with a significant predator, the walleye S. vitreus (Lyons 1987; Johnson et al. 1988) and because these sunfish have a deeper body form than other local populations (Jastrebski and Robinson 2004). Lab-reared sunfish from artificial crosses do not survive well, and so we collected juveniles using minnow traps (2.5-cm entrance diameter) and seines (0.75-cm mesh size) from shallow water inshore littoral (L) and deeper offshore pelagic (P) sites. Replicated groups composed of 16 fish of the same ecomorph and age class (0+ or 1+) were reared for 6 months (September 2003–April 2004) in the Hagan Aqualab at the University of Guelph. Groups were reared in 100-l plastic tubs with (+ or experienced) or without (– or control) caged walleye fed a diet of euthanized juvenile sunfish for a total of 4 rearing treatments: littoral fish with and without walleye predators (L+ vs. L–) and pelagic fish with and without walleye (P+ vs. P–). Similar sized pumpkinseed sunfish were used in place of the walleye as a control in the predator-absent (–) treatments. Each rearing treatment was replicated 7 times (further details in Januszkiewicz and Robinson [2007]).
Behavioral study
Fish used in our behavioral study were taken temporarily from the rearing study after approximately 4 months of growth. Behavior was recorded in real time for a pair of focal sunfish removed from a group and placed in a 90-l glass aquarium (72 L x 28 W x 44 D cm), filled to a depth of 30 cm with 20 °C water (hereafter the "test aquarium"). A second glass aquarium (36 L x 20 W x 41 D cm) was placed at one end of the test aquarium to house a walleye. A single walleye (18 cm total length [TL]), acclimated for 11 days, was used for all behavioral trials to reduce predator variability. Visual contact between the sunfish and walleye predator was manipulated by removing a sliding screen between the 2 aquaria. A 1-cm square grid drawn on the long walls of the sunfish test aquarium permitted spatial measurements.
Each pair of focal sunfish were usually the largest (5–7.5 cm TL) fish from their rearing group and were within 1 cm TL of each other in order to reduce dominance interactions. Focal pairs were acclimated for at least 5 h in the test aquarium before each trial, during which they were visually isolated from the predator by the screen. Each trial had 2 phases during which behavior was recorded at 30-s intervals by J.C.K. from behind a blind: prior to visual exposure to the walleye for 5 min (10 observations) and followed by visual exposure to the predator for 10 min (20 observations).
Five behavioral variables (2 spatial and 3 nonspatial) were recorded every 30 s from a randomly selected focal sunfish. The 2 spatial behavioral variables were recorded on a continuous numeric scale corresponding to the grid on the aquarium walls (±0.5 cm), whereas the 3 nonspatial variables were tallied in mutually exclusive binomial categories. The spatial variables were proximity of the focal sunfish to the walleye tank (0 cm at the wall adjacent to the walleye and 72 cm at maximum distance) and depth of focal sunfish below water surface (30 cm at the tank bottom). The nonspatial variables were aggregation of focal sunfish (aggregated was defined as sunfish being within 2 body lengths of each other and moving in a coordinated fashion vs. being dispersed), activity of the sunfish closest to the predator (active refers to a horizontal or vertical change in position by >1 cm over the previous 5 s vs. no such change), and orientation of the closest sunfish toward the predator tank (either facing away vs. toward or lateral to the walleye).
The 10 observations before and 20 observations after exposure to the predator walleye were summarized in each trial for each behavioral variable as follows. For the 2 spatial variables, mean proximity to the walleye tank and the mean depth in the water column before and after exposure were each computed. For the 3 nonspatial behaviors, we computed proportions of observations in each category before and after predator exposure.
We compared the behavior of sunfish from 14 rearing tubs (3 replicate tubs each of the L+ and P+ treatments and 4 replicate tubs each of L– and P–). Smaller and generally younger fish were not used because of increased risk of mortality. A total of 23 pairs of sunfish were tested including duplicate pairs of sunfish drawn from 9 of the rearing groups. We further summarized the behavioral data for duplicate pairs from the same tub by calculating the mean of observations among the pairs of sunfish drawn from the same tub to keep the rearing group as the independent unit (n = 14). We used a repeated-measures analysis of variance model to test the effects of sunfish ecomorph, prior predator exposure (predator naive vs. experienced) and observation phase (before vs. after predator exposure). Ecomorph and prior exposure (and their interaction) were treated as among-subjects factors, whereas observation phase (and its 2- and 3-way interactions) were nested within subjects (Zar 1999).
Predation study
We performed "choice" trials to test if survival varied between predator-induced and control phenotypes for a given ecomorph after rearing for 6 months with or without predation cues. The effect of phenotype on survival was tested by releasing equal numbers of predator-induced and control phenotypes (predator +/–) of a particular ecomorph into a pool for approximately 72 h with 2 walleye predators. Pelagic forms were stocked into pools that simulated pelagic conditions, and littoral forms were stocked in simulated littoral conditions. We did not create mixed groups of pelagic and littoral ecomorphs as we wanted to test for the effect of predator-induced phenotype not sunfish ecomorph. We used 6 in-ground circular black polyvinylchloride-lined pools (1.2 m deep x 3.9 m diameter) at the University of Guelph Turfgrass Institute (Hanson et al. 2002). Cleaned pools were filled to a depth of 1 m (12 000 l volume) with filtered water from an adjacent pond. Three replicate pelagic treatment pools each included a simulated shoal made from 6 standard concrete blocks (40 x 20 x 20 cm) containing 2 square holes (15 x 15 cm) stacked haphazardly at the center. Pelagic sunfish aggregate around rocky shoals that punctuate open water lake habitats where juveniles appear to hide from predators (Robinson et al. 1993; Gillespie and Fox 2003). We simulated submerged macrophyte stems in 3 littoral treatment pools by hanging 4- x 86-cm strips of weighted burlap fabric from ropes stretched across each pool to the bottom at an evenly spaced density of 100 strips per m2.
Seven replicates of mixed groups of littoral sunfish reared with and without predation cues (L+ and L–) were tested in the simulated littoral pools, and 7 replicates mixed groups of pelagic ecomorphs (P+ and P–) were tested in the simulated pelagic pools. In the littoral pools, mixed group size was 14 littoral sunfish (7L+ and 7L–) for 6 replicate trials and 10 sunfish (5L+ and 5L–) for 1 trial. In the pelagic pools, mixed group size was 14 pelagic sunfish (7P+ and 7P–) for 3 replicates and 10 sunfish (5P+ and 5P–) for the remaining 4 replicates because fewer pelagic sunfish were available from our rearing study. Predator-induced and control phenotypes were identified within each mixed group of sunfish by clipping 0.5 cm of the ninth or tenth dorsal spine (clipping was assigned to different treatments in different replicates). Sunfish (mean fork length [FL] of 5.3 cm; ±1.2 mm standard error [SE]) were fed and acclimated in the pools for 48 h prior to the release of pairs of hatchery-raised walleye (Wade Leonard Walleye Culture, Hartington Ontario, Canada). Unfed walleye (21–26 cm FL) were acclimated for 48 h in mesh enclosures within each pool and then released for 72 h. Pools were then drained and surviving sunfish were collected, identified, and counted. Surviving sunfish and walleye were not reused in any subsequent trials. Littoral and pelagic replicate trials were performed concurrently over a 2-week period in early May 2004.
We compared survival between predator-induced and control phenotypes in each mixed group using a paired design that treated the pool replicate as the independent unit. The proportions surviving of induced and control phenotypes in each replicate trial were compared using a paired t-test of the 7 replicate trials of littoral pools and separately for 7 replicated trials in pelagic pools to minimize possible effects on survival of group size, composition, and variation among predators. A 1-tailed alternative hypothesis was justified because we expect predator-induced phenotypes to have higher survival if predator-induced responses confer survival advantages. The mean proportion of survivors between littoral and pelagic ecomorphs and simulated environments were also compared using a 2-tailed t-test after combining induced and control phenotypes. Proportions were transformed to their arcsine for all analyses (where 0/n and n/n were, respectively, replaced by 1/4n and 1 – 1/4n; Zar 1999).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Variation in sunfish behavior
Prior experience with predation cues strongly influenced behavioral responses in these pumpkinseed sunfish, but these responses could also differ between sunfish ecomorphs and pre- versus postsighting of predator observation phases. For example, distance to the walleye end of the aquaria generally increased on sighting the walleye predator as expected (Figure 1A; observation phase F1,10 = 10.7, P = 0.008). But prior experience also influenced the sunfish proximity to the walleye predator in different ways between sunfish ecomorphs because rank predator distances depended on prior experience and on observation phase (pre- vs. postsighting). The proximity behavior of littoral ecomorphs was much more flexible than in pelagic ecomorphs. After visual exposure, littoral ecomorphs without prior experience L–post fish maintained an average distance from the predator end of the tank of 43.4 cm (±1 SE = 12.5 cm), whereas L+post fish held a much closer average distance of 16.1 cm (±11.6 cm) from the walleye. In contrast, pelagic ecomorphs (P–post and P+post) kept similar stations on sighting the walleye (31.5 ± 4.9 cm and 29.0 ± 5.0 cm, respectively), indicating that prior experience had little effect on their proximity behavior. Interestingly, behavioral differences also existed between ecomorphs prior to sighting the walleye, although in the reverse order, so that L+pre > P+pre and P–pre > L–pre, whereas P+post > L+post and L–post > P–post (Figure 1A; 3-way interaction of ecomorph, rearing treatment, and observation phase: F1,10 = 6.69, P = 0.027). This demonstrates that the flexibility of proximity behavior in sunfish can have 2 distinct components: a response to prior experience with predation cues and a response contingent on current predation risk conditions and that their interaction varies between pumpkinseed sunfish ecomorphs.
|
Variation in movement activity also reflected a complex interaction between a learned effect and a response to current situation that was most noticeable in the pelagic ecomorph (Figure 1B; 3-way interaction of ecomorph, rearing treatment, and observation phase: F1,10 = 8.73, P = 0.014). Here, the activity levels of inexperienced fish (P–pre, P–post, L–pre, and L–post) were all similarly low regardless of whether they were assessed before or after sighting the predator (range in mean proportion of active intervals among groups: 0.28–0.49), whereas the activity levels of predation-experienced fish were high and generally similar L+pre, L+post, and P+post (range in mean proportion of active intervals: 0.56–0.65), except for P+pre fish who before sighting the predator were remarkably inactive (mean proportion of active intervals: 0.05 ± 0.05).
There was also evidence that prior exposure to predator cues during rearing influenced orientation direction (Figure 1E; F1,10 = 6.91, P = 0.051). P+ and L+ individuals both showed similar high frequencies of facing the walleye end of the tank after sighting the walleye (respective proportions: 0.90 ± 0.03 and 0.94 ± 0.05) compared with inexperienced P– and L– individuals (respective proportions: 0.74 ± 0.03 and 0.81 ± 0.06). Finally, there was weak statistical evidence of variation in the vertical position behavior among these groups as well (Figure 1C; 3-way interaction of ecomorph, rearing treatment, and observation phase: F1,10 = 3.69, P = 0.08). P– fish were generally closer to the surface (off the bottom) compared with L– fish regardless of observation phase (respective mean surface distances: 24.3 ± 0.72 cm and 26.6 ± 0.43 cm). P+ and L+ fish were both generally closer to the surface (off the bottom) regardless of observation phase (mean = 22.6 ± 0.89 cm), except for P+pre fish, which remained at the bottom prior to sighting of the walleye (mean = 25.3 ± 0.37 cm, consistent with their lower activity level above). There was no evidence that prior exposure to predator cues, ecomorph, or observation phase had any consistent effects on the aggregation behavior of juvenile sunfish under these conditions.
Variation in survival under predation
No mortality of focal sunfish or walleye was observed during the 48-h acclimation period in the outdoor pools prior to release of walleye predators. Water temperature fluctuated between 11 and 15 °C over the 2-week study. The mean proportion (±1 SE) of fish surviving 72 h of walleye predation across all treatments and replicates was 48.1% (±6.2%). There was no significant difference in the mean proportion of fish surviving in littoral compared with pelagic experiments when the predator-experienced and -naive phenotypes where pooled in each ecomorph (littoral treatment = 47.1 ± 8.6%; pelagic treatment = 49.0 ± 9.4%; t26 = 0.23, P = 0.8).
In the simulated littoral pools stocked with littoral ecomorphs, predator-induced phenotypes (L+) had on average 58.4% survival (±12.8% SE), whereas only 35.9% (±10.6%) of the control phenotypes (L–) survived 72 h of walleye predation (t6 = 2.12, P = 0.039; Figure 2). In the simulated pelagic pools, induced phenotypic variation was not related to survival. Predator-induced phenotypes (P+) had on average 51.0% survival (±11.3%) similar to the 46.9% (±15.8%) survival of P– phenotypes (t6 = 0.36, P = 0.37).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Prior experience with predator cues changed spatial, activity, and orientation behavior in these juvenile pumpkinseed sunfish in addition to the variation in morphology identified in a previous study (Januszkiewicz and Robinson 2007
). This is consistent with predator-induced variation in behavior in other pumpkinseed sunfish studies (Marcus and Brown 2003). Of greater interest is our finding that the nature of the learned responses to the same predator cues varied between ecomorphs. This suggests that learned responses to predation cues have evolutionarily diverged. Littoral ecomorphs had strong responses to prior exposure because L+ were the least intimidated by predator proximity, whereas L– fish kept the greatest mean distance of all groups. The proximity behavior of pelagic ecomorphs was unaffected by prior predator experience and so exhibited a more canalized or congenital-like response. Predator-induced phenotypes had greater probability of survival under simulated littoral but not pelagic conditions, suggesting that phenotypic responses to predation cues during rearing can provide fitness advantages in at least one lake habitat. Such survival advantages may be due to changes in behavior or morphology. Januszkiewicz and Robinson (2007) found that predator-induced pumpkinseed phenotypes had deeper bodies and longer dorsal spines than fish reared without cues. This potentially provides a 70-mm long sunfish with a body depth normally associated with larger 80-mm sunfish, a size expected to reduce predation by smallmouth bass (Mittelbach 1984). Increased body depth and dorsal spine length may enhance survival by restricting gape-limited predators to feed on smaller fish (Brönmark and Miner 1992; Nilsson et al. 1995; Einfalt and Wahl 1997; Brönmark et al. 1999) or enhance escape after attack (Webb 1986). As far as we know, this is the first evidence that predator-induced phenotypic changes can influence survival under predation in sunfishes.
The potential advantages of predator-induced morphology are potentially confounded with variation in induced behavior because both develop in response to predator cues. They are difficult to decouple here because we focused on survival rather than on the specific mechanics of predation. Littoral ecomorphs with prior experience (L+) tolerated a much shorter predator distance and expressed the highest levels of activity, behaviors that were associated with reduced rates of mortality under littoral conditions. However, induced behavioral responses may have increased survival of L+ sunfish independent of morphology because induced morphological variation was not consistently related to survival. For example, survival was not enhanced in P+ sunfish as it was in L+ sunfish despite both ecomorphs receiving the advantages of enhanced body depths (Januszkiewicz and Robinson 2007). In addition, spine length is similar between L+ and L– types and so could not account for survival differences between these 2 forms. We conclude then that induced variation in body form did not account for variation in survival among littoral forms.
The spatial behavior of L+ and L– types were dramatically different, suggesting that L+ types are either more acclimated to the presence of predators or perhaps are more successful at predator inspection. Predator inspection may increase survival through alarm signaling (Smith and Smith 1989), predator deterrence (Godin and Davis 1995), predation risk assessment, and predator recognition (Brown and Godin 1999). Pelagic ecomorphs on the other hand held intermediate predator–prey distances that were unaffected by prior predator experience. In other words, littoral ecomorphs change spatial behavior in response to prior experience with predation cues or express learning or acclimation, whereas pelagic ecomorphs do not. A similar pattern emerged with vertical position behavior in sight of a predator with littoral ecomorphs showing a greater change in depth position than pelagic ecomorphs with prior experience. Littoral ecomorphs with prior experience also increased their activity levels slightly more than pelagic ecomorphs in sight of the walleye.
The divergence in learned behavioral responses between sunfish ecomorphs suggests that they may have evolved under different risks of predation between lake habitats. Variation in predation risk is possible because piscivorous walleye and bass hunt more commonly in littoral than pelagic habitats (Lyons 1987; Johnson et al. 1988; Werner 1994; Chipps et al. 2004) and because predator attack and prey response tactics may be specific to the physical structure of each habitat. For example, predator inspection may be beneficial in the littoral habitat if its complex structure provides more refuges in case of attack than are available in the open pelagic habitat. In the pelagic habitat, increasing the distance from and focus on a potential predator may be a better prey response. This response to variation in predation risk is also interesting because there do not appear to be any evolutionary responses in life history traits between ecomorphs in the pumpkinseed polyphenism (Gillespie and Fox 2003).
We can begin to address more generally the evolution of antipredator behavior in sunfishes by contrasting our study with another involving sympatric bluegill sunfish ecomorphs (Lepomis macrochirus) that are ecologically analogous to the pumpkinseed ecomorphs used here (Chipps et al. 2004). It is important to note that their study did not set out to test the degree to which antipredator behaviors were learned (because ecomorphs were collected from the wild) or if the nature of learned responses varied between bluegill ecomorphs. Nonetheless, their study demonstrated that behavioral variation among littoral and pelagic ecomorphs of bluegill sunfish influenced survival under predation by largemouth bass (Micropterus salmoides). Behavioral variation in evasiveness between bluegill ecomorphs was positively related to survival in simulated pelagic but not littoral pools (contrary to our predation results). In their pelagic pools, pelagic bluegill ecomorphs schooled more densely and maintained greater distances from bass predators than did littoral ecomorphs. In our behavioral study, P+ pumpkinseed sunfish also maintained a greater distance from the predator than L+ sunfish although we found no significant variation in aggregation behavior (2 individuals and a small aquarium size may have limited aggregation behavior here). Bluegill antipredator tactics also varied between habitats because surfacing behavior (a tendency to rise closer to the water surface and even leap from the water during pursuit by predators) was observed in both bluegill ecomorphs under pelagic but not littoral conditions. In bluegill sunfish, surfacing behavior appears to respond to current conditions like structural simplicity or habitat-specific predator attack tactics. Both of our pumpkinseed ecomorphs showed some tendency to rise from the bottom with prior experience with predators although there appear to be innate differences between the ecomorphs in the absence of prior predator cues. Finally, under littoral conditions, Chipps et al. (2004) found that antipredator behavior did not vary between ecomorphs even though littoral bluegill survived longer than pelagic bluegill. This suggests that under littoral conditions, morphology may play a greater role in bluegill sunfish survival than in pumpkinseed sunfish survival. A possible explanation for these differences is that the bluegill study focused on the consequences of variation between ecomorphs collected from the field, whereas we focused on variation within ecomorphs induced by a single factor (predation cues) in the laboratory. Regardless, both studies suggest that predation risk likely imposes diversifying selection between lake habitats that could favor the evolution of divergent phenotypes.
Our conclusion that pumpkinseed sunfish ecomorphs in Salmon Trout Lake have genetically diverged with respect to how prior exposure to predation cues shapes behavior rests on there being no other explanation for these phenotypic differences. We did not rear our fish from birth in a common laboratory environment (because this is not yet possible), and so we cannot rule out that variation in learning between ecomorphs resulted from different early experiences in their native lake habitats prior to collection. This does not affect conclusions about behavior and survival between predator-exposed and -naive treatments within each ecomorph, however, because they did share a completely common history. Certain aspects of our study reduce but do not eliminate the problem between ecomorphs. Juvenile sunfish proceed through a series of ontogenetic niche shifts mediated by body size that generally causes similar size individuals to share common habitats and diets (Werner 1994). Below 30 mm TL, sunfish forage on planktonic prey in the water column, switching to benthic invertebrate prey at a body size of 35 mm that not only improves energy gain but also reduces predation risk because of better opportunities to hide (reviewed in Januszkiewicz and Robinson [2007]). Small pumpkinseed sunfish are cautious in the field and difficult to observe and catch in both habitats (Robinson BW, personal observation). This suggests that they share high risk of predation. Our fish were on average 44 mm in length, suggesting that most would therefore refuge near the safer lake bottom regardless of habitat. Finally, all treatments included a range of body sizes that broadly overlapped, further reducing the chance that early experience influenced ecomorph responses here. Nevertheless, this does not rule out early environmental affects and so our conclusions of genetic divergence must be treated as preliminary.
Our study is also not the first to conclude that learned responses by small fish to the same predator cues genetically diverge between populations. Magurran (1990) found that the learned antipredator behavior of populations of minnows from lakes in Great Britain with and without fish predators had also diverged. A novel feature of our study is that divergence is taking place between sympatric ecomorphs. These studies demonstrate that learned responses to predator cues can evolve to become greater (more flexible) or lesser (more congenital) under selection. A single universally responsive mechanism, such as learning, used to generate optimal antipredator responses under all circumstances is not supported by these results.
Variation in behavior and survival were obviously assessed under artificial conditions here. Natural habitats are larger and more complex than we could possibly achieve in our tests, and survival in littoral habitats may be related to the scale of structural complexity (Savino and Stein 1982, 1989b). Densities of sunfish prey in our pools (8.3–11.7 sunfish per 10 m2) were comparable to natural densities in littoral habitats (1–18 fish per 10 m2; Fox and Keast 1991; Osenberg et al. 1992), but walleye density may have been higher than normal (2 per 10 m2). It is unlikely that survival was enhanced by the relatively small size (
25 cm FL) of walleye. The ratio of prey sunfish (5 cm FL) to predator body size was 0.2, near the bottom limit of the optimal prey to predator size ratios of 0.25–0.3 for other temperate fishes (Hoyle and Keast 1987).
Inappropriate conditions may explain why induced phenotype did not affect survival in our simulated pelagic pools. Alternatively, induced phenotypes did not provide functional antipredator advantages for pelagic ecomorphs under pelagic conditions or, perhaps, we elicited inadequate or incorrect responses compared with responses induced by multifarious or stronger cues in the field. The more inflexible behavioral responses to prior predation cues of pelagic ecomorphs are consistent with this explanation (Figure 1). However, various unrealistic conditions in pelagic pools may have also affected survival. For example, the holes in the cement blocks were large enough for the walleye to enter and so did not provide secure refugia for sunfish (e.g.,Nemeth 1998; Steele 1999). The clear water, the homogeneity in background color of the black pond liner, and its contrast with the white cinder blocks may also intensify the silhouette of pumpkinseeds (Seapy and Young 1986). Finally, the small volume of our pools may limit the effectiveness of antipredator behavior. For example, we observed prolonged chases culminating in successful attacks on sunfish, which suggests that walleye hunted sunfish by tiring them in the absence of a secure refuge. In the field, startled pelagic sunfish ecomorphs dive toward the bottom where visual predators may lose sight of them either in the gloom or against the substrate. Our observations on pumpkinseed sunfish and studies with bluegill sunfish (Chipps et al. 2004) suggest that pelagic ecomorphs reduce predation by maintaining greater distances than may have been possible in our pools.
Developmental responses to predation cues appear to influence survival under predation in littoral habitats in addition to the previously known antipredator effects of body size and habitat shifts in sunfishes (Mittelbach and Chesson 1987; Werner 1994; Chipps et al. 2004). This suggests that induced phenotypes may serve an antipredator function. Our most novel result though is that coexisting sunfish ecomorphs appear to have genetically diverged with respect to how prior experience influences the development of behavior as well as body form. This suggests that selection can act on genetic variation in developmental responses that produce adaptive and maladaptive phenotypes (West-Eberhard 2003; Parsons and Robinson 2006). Studies of divergence in flexible behaviors are ideal for testing this idea further because the evolvability of complex flexible traits are still poorly documented and understood in animals. Selection from predation influences diversity in many fishes and seems to contribute to polyphenism in sunfish. Perhaps, many other "resource polymorphisms" in fishes have also evolved under diversifying selection imposed by habitat-specific predation risk in addition to resource use.
|
FUNDING |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
National Sciences and Engineering Research Council of Canada Discovery grants program (NSERC 216890-03) and a PREAward from the province of Ontario (PREA 0015-0751) to B.W.R; the Landesstiftung Baden-Württemberg, Germany, to J.C.K.
|
ACKNOWLEDGEMENTS |
|---|
T. Nudds, R. McLaughlin, D. L. G. Noakes, D. Weese, K. Peiman, M. Moles, R. Svanbäck, J. Field, and an anonymous reviewer all provided valuable comments on the manuscript. S. Csinos, M. Palka, M. Washburn, R. Wall, A. Kwasniewska, and T. Habib all provided assistance in the laboratory and field, and K. Peiman prepared the figures. All animal-use protocols met the guidelines for animal care and research in Canada and were approved by the Animal Care Committee of the University of Guelph.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Abrams PA. The evolution of predator-prey interactions: theory and evidence. Annu Rev Ecol Syst (2000) 2000:79–105.
Brönmark C, Miner JG. Predator-induced phenotypic change in body morphology in crucian carp. Science (1992) 258:1348–1350.
Brönmark C, Pettersson LB, Nilsson PA. Predator-induced defense in crucian carp. In: The ecology and evolution of inducible defenses—Tollrian R, Harvell CD, eds. (1999) Princeton (NJ): Princeton University Press. 203–217.
Brown GE, Brennan S. Chemical alarm signals in juvenile green sunfish (Lepomis cyanellus, Centrarchidae). Copeia (2000) 4:1079–1082.
Brown GE, Chivers DP, Smith RJF. Fathead minnows avoid conspecific and heterospecific alarm pheromones in the faeces of northern pike. J Fish Biol (1995) 47:387–393.[Web of Science]
Brown GE, Godin J-GJ. Who dares learns: chemical inspection behaviour and acquired predator recognition in a characin fish. Anim Behav (1999) 137:181–196.
Brown GE, Golub JL, Plata DL. Attack cone avoidance during predator inspection visits by wild finescale dace (Phoxinus neogaeus): the effects of predator diet. J Chem Ecol (2001) 27:1657–1666.[CrossRef][Web of Science][Medline]
Brown GE, Zachar MM. Chemical predator inspection in a characin fish (Hemingrammus erythrozonus, Characidae, Ostariophysi): the effects of mixed predator diets. Ethology (2002) 108:451–461.[CrossRef][Web of Science]
Chipps SR, Dunbar JA, Wahl DH. Phenotypic variation and vulnerability to predation in juvenile bluegill sunfish (Lepomis macrochirus). Oecologia (2004) 138:32–38.[CrossRef][Web of Science][Medline]
Chivers DP, Mirza RS, Bryer PJ, Kiesecker JM. Threat-sensitive predator avoidance by slimy sculpins: understanding the importance of visual versus chemical information. Can J Zool (2001) 79:867–873.[CrossRef]
Dill LM. Adaptive flexibility in the foraging behavior of fishes. Can J Fish Aquat Sci (1983) 40:398–408.
Einfalt LM, Wahl DH. Prey selection by juvenile walleye as influenced by prey morphology and behavior. Can J Fish Aquat Sci (1997) 54:2618–2626.[CrossRef]
Eklöv P, Jonsson P. Pike predators induce morphological changes in young perch and roach. J Fish Biol (2007) 70:155–164.
Eklöv P, Svanbäck R. Predation risk influences adaptive morphological variation in fish populations. Am Nat (2006) 167:440–452.[CrossRef][Web of Science][Medline]
Endler JA. Defense against predators. In: Predator-prey relationships—Feder GLM, ed. (1986) Chicago (IL): Chicago Univ. Press. 109–134.
Fox MG, Keast A. Effects of overwinter mortality on reproductive life history characteristics of pumpkinseed (Lepomis gibbosus) populations. Can J Fish Aquat Sci (1991) 48:1792–1799.[CrossRef]
Gilbert SF. Ecological developmental biology: developmental biology meets the real world. Dev Biol (2001) 233:1–12.[CrossRef][Web of Science][Medline]
Gillespie GJ, Fox MG. Morphological and life history differentiation in populations of pumpkinseed (Lepomis gibbosus). J Fish Biol (2003) 62:1099–1115.[CrossRef][Web of Science]
Godin J-GJ, Davis SA. Who dares, benefits: predator approach behaviour in the guppy (Poecilia reticulata) deters predator pursuit. Proc R Soc Lond Ser B Biol Sci (1995) 259:193–200.[CrossRef]
Hanson ML, Sibley PK, Ellis DA, Mabury SA, Muir DC, Solomon KR. Evaluation of monochloracetic acid (MCA) degradation and toxicity to Lemna gibba, Myriophyllum spicatum, and Myriophyllum sibiricum in aquatic microcosms. Aquat Toxicol (2002) 61:251–273.[CrossRef][Web of Science][Medline]
Hoyle JA, Keast A. The effect of prey morphology and size on handling time in a piscivore, the largemouth bass (Micropterus salmoides). Can J Zool (1987) 65:1972–1977.[CrossRef]
Januszkiewicz AJ, Robinson BW. Divergent walleye (Sander vitreus) mediated inducible defenses in the centrarchid pumpkinseed sunfish (Lepomis gibbosus). Biol J Linn Soc (2007) 90:25–36.[CrossRef][Web of Science]
Jastrebski CJ, Robinson BW. Natural selection and the evolution of replicated trophic polymorphisms in pumpkinseed sunfish (Lepomis gibbosus). Evol. Ecol. Res (2004) 6:1–22.
Johansson F, Samuelsson L. Fish-induced variation in abdominal spine length of Leucorrhinia dubia (Odonata) larvae? Oecologia (1994) 100:74–79.[CrossRef][Web of Science]
Johnson BL, Smith DL, Carline RF. Habitat preferences, survival, growth, foods, and harvests of walleyes and walleye X sauger hybrids. N Am J Fish Manag (1988) 8:292–304.[CrossRef]
Langerhans RB, Layman CA, Shokrollahi AM, DeWitt TJ. Predator-driven phenotypic diversification in Gambusia affinis. Evolution (2004) 58:2305–2318.[Web of Science][Medline]
Leduc AOHC, Noseworthy MK, Adrian JC, Brown GE. Detection of conspecific and heterospecific alarm signals by juvenile pumpkinseed under weak acidic conditions. J Fish Biol (2003) 63:1331–1336.[CrossRef][Web of Science]
Levins R. Evolution in changing environments. (1968) Princeton (NJ): Princeton University Press.
Lima SL. Stress and decision-making under the risk of predation: recent developments from behavioral, reproductive and ecological perspectives. Adv Stud Behav (1998) 27:215–290.[CrossRef]
Lyons J. Prey choice among piscivorous juvenile walleyes (Stizostedion vitreum). Can J Fish Aquat Sci (1987) 44:758–764.[CrossRef]
Magurran A, Seghers BH, Carbalho GR, Shaw PW. Evolution of adaptive variation in antipredator behavior. Mar Behav Physiol (1993) 23:29–44.[CrossRef]
Magurran AE. The inheritance and development of minnow anti-predator behavior. Anim Behav (1990) 39:828–834.[CrossRef][Web of Science]
Marcus JP, Brown GE. Response of pumpkinseed sunfish to conspecific chemical alarm cues: an interaction between ontogeny and stimulus concentration. Can J Zool (2003) 81:1671–1677.[CrossRef]
Mathis A, Smith RJF. Chemical alarm signals increase the survival time of fathead minnows (Pimephales promelas) during encounters with northern pike (Esox lucius). Behav Ecol (1993) 4:260–265.
Mikolajewski DJ, Johansson F, Wohlfahrt B, Stoks R. Invertebrate predation selects for the loss of a morphological antipredator trait. Evolution (2006) 60:1306–1310.[Web of Science][Medline]
Mirza RS, Chivers DP. Brook charr (Salvelinus fontinalis) can differentiate chemical alarm cues produced by different age/classes of conspecifics. J Chem Ecol (2002) 28:555–564.[CrossRef][Web of Science][Medline]
Mittelbach GG. Predation and resource partitioning in two sunfishes (Centrarchidae). Ecology (1984) 65:499–513.[CrossRef][Web of Science]
Mittelbach GG, Chesson PL. Predation risk: indirect effects on fish populations. In: Predation, direct and indirect impacts on aquatic communities—Kerfoot WC, Sih A, eds. (1987) Hanover (NH): University Press of New England. 315–332.
Mittelbach GG, Osenberg CW, Leibold MA. Trophic relations and ontogenetic niche shifts in aquatic ecosystems. In: Size-structured populations: ecology and evolution—Ebenman B, Persson L, eds. (1988) Berlin (Germany): Springer-Verlag. 219–235.
Nelson JS, Crossman EJ, Expinosa-Perez H, Findley LT, Gilbert CR, Lea RN, Williams JD. The ‘names of fishes’ list, including recommended changes in fish names: chinook salmon for chinook salmon, and Sander to replace Stizostedion for the sauger and walleye. Fisheries (2003) 28:38–39.
Nemeth RS. The effect of natural variation in substrate architecture on the survival of juvenile bicolor damselfish. Environ Biol Fishes (1998) 53:129–141.[CrossRef]
Nilsson PA, Bronmark C, Pettersson LB. Benefits of a predator-induced morphology in crucian carp. Oecologia (1995) 104:291–296.[CrossRef][Web of Science]
Osenberg CW, Mittelbach GG, Wainwright PC. Two-stage life histories in fish: the interaction between juvenile competition and adult diet, performance, and morphology. Ecology (1992) 73:255–267.[CrossRef][Web of Science]
Papaj DR. Optimizing learning and its effect on evolutionary change in behavior. In: Behavioral mechanisms in evolutionary ecology—Real LA, ed. (1994) Chicago (IL): University of Chicago Press. 133–153.
Parsons KJ, Robinson BW. Replicated evolution of integrated plastic responses during early adaptive evolution. Evolution (2006) 60:801–813.[Web of Science][Medline]
Pfeiffer W. The distribution of fright reaction and alarm substances in fishes. Copeia (1977) 4:653–665.
Pigliucci M. Evolution of phenotypic plasticity: where are we going now? Trend Ecol Evol (2005) 20:481–486.[CrossRef][Medline]
Reimchen TE. Trout foraging failures and the evolution of body size in stickleback. Copeia (1991) 1991:1098–1104.[CrossRef]
Robinson BW, Parsons KJ. Changing times, spaces, and faces: tests and implications of adaptive morphological plasticity in the fishes of northern postglacial lakes. Can J Fish Aquat Sci (2002) 59:1819–1833.[CrossRef]
Robinson BW, Wilson DS, Margosian AS, Lotito PT. Ecological and morphological differentiation of pumpkinseed sunfish in lakes without bluegill sunfish. Evol. Ecol (1993) 7:451–464.
Rundle HD, Vamosi SM, Schluter D. Experimental test of predation's effect on divergent selection during character displacement in sticklebacks. Proc Natl Acad Sci USA (2003) 100:14943–14948.
Savino JF, Stein RA. Predator-prey interactions between largemouth bass and bluegills as influenced by simulated submersed vegetation. Trans Am Fish Soc (1982) 111:255–266.[CrossRef]
Savino JF, Stein RA. Behavior of fish predators and their prey: habitat choice between open water and dense vegetation. Environ Biol Fishes (1989a) 24:287–293.[CrossRef]
Savino JF, Stein RA. Behavioural interactions between fish predators and their prey: effects of plant density. Anim Behav (1989b) 37:311–321.[CrossRef][Web of Science]
Schlichting C, Pigliucci M. Phenotypic evolution: A reaction norm perspective (1998) Sunderland (MA): Sinauer Associates, Inc.
Schluter D. Ecological speciation in postglacial fishes. Philos T Roy Soc B (1996) 351:807–814.[CrossRef]
Seapy RR, Young RE. Concealment in epipelagic pterotracheid heteropods (Gastropoda) and cranchiid squids (Cephalopoda). J Zool (1986) 210:137–147.[Web of Science]
Smith GR, Todd TN. Evolution of fish species flocks of fishes in north temperate lakes. In: The evolution of fish species flocks—Echelle AA, Kornfield I, eds. (1984) Orono (ME): University of Maine Press. 45–68.
Smith ME, Belk MC. Risk assessment in western mosquitofish (Gambusia affinis): do multiple cues have additive effects? Behav Ecol Sociobiol (2001) 51:101–107.[CrossRef][Web of Science]
Smith RJF, Smith MJ. Predator-recognition behaviour in two species of gobiid fishes, Asterropteryx semipunctatus and Gnatholepis anjerensis. Ethology (1989) 83:19–30.[Web of Science]
Stearns SC. The evolutionary significance of phenotypic plasticity. BioScience (1989) 39:436–445.[CrossRef][Web of Science]
Steele MA. Effects of shelter and predators on reef fishes. J Exp Mar Biol Ecol (1999) 233:65–79.[CrossRef][Web of Science]
Straile D, Halbich A. Life history and multiple antipredator defenses of an invertebrate pelagic predator, Bythotrephes longimanus. Ecology (2000) 81:150–163.[Web of Science]
Svanbäck R, Eklöv P. Morphology in perch affects habitat-specific feeding efficiency. Funct Ecol (2004) 18:503–510.[CrossRef]
Tollrian R, Harvell CD. The evolution of inducible defenses: current ideas. In: The ecology and evolution of inducible defenses—Tollrian R, Harvell CD, eds. (1999) Princeton (NJ): Princeton University Press.
Vamosi SM. On the role of enemies in divergence and diversification of prey: a review and synthesis. Can J Zool (2005) 83:894–910.[CrossRef]
Vamosi SM, Schluter D. Impacts of trout predation on fitness of sympatric sticklebacks and their hybrids. P Roy Soc Lond B Biol (2002) 269:923–930.
Vamosi SM, Schluter D. Character shifts in the defensive armor of sympatric sticklebacks. Evolution (2004) 58:376–385.[CrossRef][Web of Science][Medline]
Wahl DH, Stein RA. Selective predation by three esocides: the role of prey behavior and morphology. Trans Am Fish Soc (1988) 117:142–151.[CrossRef]
Webb PW. Effect of body form and response threshold on the vulnerability of four species of teleost prey attacked by largemouth bass (Micropterus salmoides). Can J Fish Aquat Sci (1986) 43:763–771.
Werner EE. Individual behavior and higher-order species interactions. In: Behavioral mechanisms in evolutionary ecology—Real LA, ed. (1994) Chicago (IL): University of Chicago Press. 297–324.
Werner EE, Hall DJ. Ontogenetic habitat shifts in bluegill: the foraging rate-predation risk trade-off. Ecology (1988) 69:1352–1366.[CrossRef][Web of Science]
West-Eberhard MJ. Phenotypic plasticity and the origins of diversity. Annu Rev Ecol Syst (1989) 20:249–278.[CrossRef][Web of Science]
West-Eberhard MJ. Developmental plasticity and evolution. (2003) New York: Oxford University Press.
Zar JH. Biostatistical analysis. (1999) Upper Saddle River (NJ): Prentice Hall.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||