| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Behavioral Ecology Vol. 10 No. 3: 251-262
© 1999 International Society for Behavioral Ecology
The development of behavioral defenses: a mechanistic analysis of vulnerability in red-eyed tree frog hatchlings
Department of Zoology, University of Texas, Austin, TX 78712, USA
K. M. Warkentin is now at the School of Biological Sciences, University of Kentucky, Lexington KY 40506-0225, USA. E-mail: rana{at}ceeb.uky.edu
Received 20 November 1997; accepted 12 October 1998.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
I examined the development and effectiveness of behavioral defenses of red-eyed tree frog hatchlings, Agalychnis callidryas, against predatory shrimp Macrobrachium americanum. Arboreal eggs of A. callidryas hatch early if attacked by egg predators and later if undisturbed, producing tadpoles that enter the water at different developmental stages. Older hatchlings survive better than young hatchlings with aquatic predators, including shrimp. Hatchlings respond to shrimp by both increasing activity and avoiding the bottom microhabitat. Older hatchlings are more active and, in the presence of shrimp, avoid the bottom more than young hatchlings. These ontogenetic changes in behavior improve survival. Specifically, the likelihood of fleeing from shrimp increases with hatchling age, and fleeing is an effective defense. In contrast to results from experiments with odonates, immobility does not reduce risk of shrimp attack, thus there is no trade-off between fleeing and motionless crypsis. Shrimp spend most of their time on the bottom, where attacks are more often successful. Avoidance of the bottom microhabitat by tadpoles therefore improves survival. Evasive maneuvers also function in defense, but evasiveness does not change with age. Morphology may limit microhabitat use by younger hatchlings. Failure to flee may reflect unresponsiveness to disturbance, which would reduce unnecessarily early hatching and limit exposure of young A. callidryas to aquatic predators.
Key words: activity, Agalychnis, defense, development, Macrobrachium, microhabitat, predation, shrimp, tadpoles.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Growth and development profoundly influence predator-prey interactions (Polis, 1984
;
Polis et. al, 1989;
Werner, 1988;
Werner and Gilliam, 1984). A
widespread pattern is that young animals suffer high predation, and
survivorship improves ontogenetically
(Polis et. al, 1989;
Werner and Gilliam, 1984).
This common pattern does not, however, have a single underlying cause.
For instance, prey achieve size refuges from gape-limited predators through
growth
(Brönmark and
Miner, 1992; Osenberg and
Mittelbach, 1989). Prey develop behavioral, morphological, and
chemical defenses against predators (Ash,
1989;
Boppré,
1990; Brodie et al.,
1978; Dyer, 1995;
Formanowicz and Brodie, 1982;
Fuiman and Magurran, 1994;
Palmer, 1990). Experience or
learning also contributes to defense
(Chivers et al., 1995;
Craig, 1994;
Punzo, 1997;
Wisendon et al., 1997).
Different defense mechanisms can have different ecological and evolutionary
effects. For instance, defenses may protect against multiple predators or
protect against one predator at the expense of increased vulnerability to
another predator. Selection for defense may be reinforced or opposed by
selection for other functions, such as feeding and growth
(Anholt and Werner, 1995;
Skelly, 1992;
Werner and Hall, 1988).
Understanding the mechanistic basis of differences in vulnerability therefore
aids in interpreting the ecological function and the evolution of
defenses.
Amphibian larvae are widely used as model systems for the study of
predator-prey interactions. Predation on premetamorphic amphibians typically
declines ontogenetically (Calef,
1973; Crump, 1984;
Semlitsch and Gibbons, 1988),
but the mechanisms by which development reduces predation vary. Growth alone
improves survival with gape-limited predators
(Caldwell et al., 1980;
Semlitsch and Gibbons, 1988).
However, vulnerability also decreases ontogenetically without gapelimitation
of predators (Cronin and Travis,
1986; Formanowicz,
1986; Semlitsch,
1990), and without substantial change in prey size
(Warkentin, 1995). Amphibian
larvae employ a variety of behavioral antipredator defenses, including reduced
activity (Anholt and Werner,
1995; Skelly,
1994), increased activity
(Hews, 1988), spatial
avoidance of predators, and refuge use
(Horat and Semlitsch, 1994;
Kats et al., 1988;
Kiesecker et al., 1996;
Petranka et al., 1987). Use of
behavioral antipredator defenses may change with development
(Bridges and Gutzke, 1997).
Most studies of the ontogeny of behavioral defenses in amphibian larvae
have examined swimming performance. Sprint speed typically improves with
growth and development and aids in escape from predators
(Hoff, 1987;
Huey, 1980;
Parichy and Kaplan, 1995;
Wassersug and Sperry, 1977;
Watkins, 1996). However,
evasiveness can decline with age (Brown and
Taylor, 1995), presumably increasing risk
(Feder, 1983;
Watkins, 1996). The ontogeny
of refuge use has been examined in two species, and in neither could it
explain decreasing predation with age
(Bridges and Gutzke, 1997;
Sih and Kats, 1994).
The study of ontogenetic changes in defense is complicated by the effects
of experience. In nature, animals gain experience as they develop. Even under
controlled conditions, effects of experience and development are not easily
separated. Exposure to predators or predator cues can affect development, such
that the morphology of naive prey makes them more vulnerable to predation
(Brönmark and
Miner, 1992; Harvell,
1990; McCollum and Van
Buskirk, 1996; Smith and Van
Buskirk, 1995). However, prey exposed to predator cues without
mortality may habituate, showing weaker behavioral responses to predators than
do naive prey (Bridges and Gutzke,
1997; Jackson and Semlitsch,
1993; Magurran and Girling,
1986). A solution to this problem is to study prey that switch
habitats at variable stages of development. This allows a natural separation
of development and experience, at least with respect to predators restricted
to the second habitat.
In this study I examined the development of behavioral antipredator
defenses in one such prey, the red-eyed tree frog Agalychnis
callidryas, specifically addressing the role of prey movement and
microhabitat use. Red-eyed tree frogs have a variable period of embryonic
development and switch habitats at hatching. A. callidryas inhabit
lowland, wet forest from the Yucatan through Panama, breeding in temporary
ponds and swamps, and sometimes in persistent water such as pools in stream
beds. Eggs are attached to vegetation overhanging water, where embryos are
threatened by arboreal and aerial predators. At hatching, tadpoles fall into
the water where they are threatened by aquatic predators. In Costa Rica, most
undisturbed embryos hatch at 7 or 8 days, but when attacked by egg predators
tadpoles hatch as early as 5 days
((Warkentin, 1995).
Late-hatched A. callidryas tadpoles are more developed and survive
better than early-hatched tadpoles in the presence of both a poeciliid fish,
Brachyraphis rhabdophora, and their most common predator, the
freshwater shrimp Macrobrachium americanum
(Warkentin, 1995). The
variation in hatching age in red-eyed tree frogs and the isolation of embryos
from aquatic predators allows examination of the effects of development on
defense without confounding effects of experience.
This study had two goals. The first goal was to determine which, if any,
behaviors serve as antishrimp defenses. The second was to assess the
mechanistic basis of the ontogenetic decline in vulnerability, or how tadpole
defenses develop. Three experiments were performed to reach these goals. (1) I
conducted a predation experiment to determine if the coverfree context
necessary for behavioral observations alters the ontogenetic pattern of
predation risk. This was necessary because the experiment that demonstrated an
ontogenetic decrease in vulnerability of A. callidryas to shrimp was
performed in a structurally complex environment, in which few behavioral
interactions were visible (Warkentin,
1995). (2) I observed the behavior of hatchlings of different ages
in the presence of shrimp and in a predator-free context to determine if, and
how, tadpoles modify their behavior when exposed to shrimp, and if any such
response varies developmentally. (3) I videotaped newly hatched tadpoles
interacting with shrimp to determine how tadpole behavior affects shrimp
hunting success and how tadpole behavioral defenses change with age.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
This research was conducted from 1992 to 1995 under permits from the Costa Rica National Park Service and Ministry of Natural Resources. All experiments were conducted at Sirena Biological Station, in the lowland, wet forest of Corcovado National Park, Costa Rica. All shrimp and young egg clutches were collected from A. callidryas breeding ponds and housed in an open-air laboratory. Egg clutches were hung from the sides of plastic tubs containing a small amount of stream water and misted with water daily to prevent desiccation. Shrimp and surviving tadpoles were returned to their collection sites after use.
All experiments compare newly hatched A. callidryas (hatchlings)
of 4 different ages: 5, 6, 7, and 8 days. Age is measured from oviposition.
A. callidryas in Corcovado become capable of hatching at 5 days old,
and most hatch by age 8 days (Warkentin,
1995). I mechanically induced hatching to obtain hatchlings of
each age. Induced versus spontaneous hatching does not appear to alter
characteristics of A. callidryas tadpoles, and much natural variation
in hatching age is induced by egg predators
(Warkentin, 1995). To induce
hatching in groups of eggs, I manually jiggled the egg mass and vigorously
sprayed it with water. To obtain a single hatchling, I removed an egg from its
clutch with forceps and, if necessary, rolled or gently pressed and prodded
the egg until it hatched. Most 7- and 8-day embryos and some younger ones
hatched without additional stimulation once removed from the clutch.
Predation experiments
I conducted predation experiments in bowls with and without leaf litter to
assess the effect of age on predation in a structurally simple context and to
assess the effect of habitat structure on predation. Shrimp were collected
27-29 h before use, housed individually in 250 ml of stream water in plastic
cups, and left unfed to standardize hunger levels. This time is adequate for
complete gut clearance. I transferred each shrimp to a bowl with 3.25 l of
water 2-3 h before experiments began. Bowls were filled with leaf litter to
two-thirds water depth for the complex habitat structure treatment and
contained a single small leaf as a perch and cover item for the shrimp in the
simple habitat structure treatment. Shrimp were 28.6±0.3 mm total
length (mean±SE throughout). Only shrimp with both pincers were used.
Recently molted shrimp with soft exoskeletons were not used, and replicates
with shrimp that molted during the experiment were discarded because shrimp do
not hunt while or immediately after molting.
Eggs were induced to hatch in the early evening and equalaged hatchlings
from two to five clutches were pooled. Sets of 12 tadpoles were drawn randomly
from this pool and placed in the bowls with shrimp within an hour of hatching.
Experiments ran for 24 h. I conducted batches of up to five replicates per age
simultaneously, depending on availability of eggs and shrimp. There were 10
replicates without cover (using 5-11 clutches/age) and 5 replicates with cover
(using 2-5 clutches/age) for each age. The data on replicates with cover is
reported in Warkentin (1995),
along with data on predation by smaller shrimp. Effects of age, habitat
structure, and their interaction were tested by analysis of variance, with
shrimp size as a covariate. Assumptions of normality and homoscedasticity were
confirmed with Kolmogrov-Smirnov and Bartlett's tests. All analyses were
performed in Systat 5.2.1 (Systat Inc., Evanston, Illinois).
Although this experiment used unequal numbers of clutches per age, an age
effect on survival is unlikely to be a spurious result of sibship effects; if
anything, uncontrolled sibship effects might obscure an age effect. The effect
of habitat structure on predation depends on both the perceptual abilities and
hunting techniques of shrimp and the morphology and behavior of tadpoles;
thus, uncontrolled differences in morphology or behavior among sibships of
tadpoles might obscure an effect of habitat on predation. Hatchling morphology
varies little among sibships, although it changes substantially with age
(Warkentin, 1998). I tested
for an effect of sibship on behavior in the following experiment.
Behavioral observations
I compared the behavior of hatchlings in the presence of pincerless shrimp
with that in predator-free controls across different hatchling ages to assess
the development of tadpole behavioral responses to shrimp. Pincerless shrimp
provide physical and chemical cues and behaviorally interact with tadpoles but
are relatively unsuccessful hunters.
Shrimp readily autotomize pincers; lost pincers regenerate and are complete
by the next molt (McVean,
1982). When possible, I collected shrimp that had lost pincers
naturally; otherwise I induced pincer autotomy shortly after collecting the
shrimp. Shrimp were housed individually in 2 l of water and fed Tetramin fish
food ad libitum. Each shrimp was acclimated to laboratory conditions for at
least 26 h and was used 4 times, once with each age of tadpoles, except in 3
cases where shrimp that molted were replaced with new pincerless shrimp.
Shrimp were transferred to experimental containers shortly before tadpoles
were introduced and were returned to holding containers and fed between
experiments. Shrimp were 32±0.6 mm in length.
I divided A. callidryas egg clutches into 4 groups of 10 or more eggs shortly after collection (n = 10 clutches). I planned to use one part of each clutch at each age, from 5-8 days. However, some eggs assigned to hatch at 8 days hatched earlier, leaving insufficient 8-day embryos. Other clutches were used to complete the experiment. Tadpoles were observed in bowls filled with 4 l of stream water. Hatching was induced in the early evening, and two sets of five siblings were drawn from each clutch. I placed one set into a bowl with a shrimp and the other into a predator-free bowl within 5 min after hatching. Behavioral observations were made under dim, indirect light to minimize disturbance of the shrimp. I scanned each bowl and recorded positions of all tadpoles 10 times, at 10-min intervals starting 5-10 min after hatching. Tadpole positions were defined as (1) midwater, not touching any surface, (2) touching the bottom, (3) touching the water surface, or (4) touching the container side. I also observed focal tadpoles for 5-min periods, counting periods of movement (moves) and using a stopwatch to record time spent moving. Focal animal observations were repeated twice per bowl. In each case I chose a focal tadpole that was initially motionless and in the position shared by most other tadpoles in the bowl. In two cases the clawless shrimp caught tadpoles during the experiment; I replaced the missing tadpoles immediately. Data were analyzed using multivariate analysis of variance. Movement data were log-transformed to improve normality and homoscedasticity. Because some 8-day hatchlings were from unrelated clutches, I used only data from 5- to 7-day hatchlings to test if behavior varied among sibships.
Videotaped predator-prey interactions
Videotape recording
I recorded and analyzed videotapes of interactions between shrimp and newly
hatched tadpoles of different ages to assess developmental changes in tadpole
behavior and effects of tadpole behavior on predation by shrimp.
Animal collection and embryo housing were as above. Shrimp were housed individually in round plexiglas tanks 19.1 cm in internal diameter by 15.7 cm deep, filled to 6 cm depth with stream water. Shrimp were acclimated to laboratory conditions for at least 5 days before experiments. During this time they were fed A. callidryas hatchlings each afternoon. Experiments commenced once shrimp were feeding well under conditions allowing recording. I measured each shrimp (pincer length, antennae length, and body length, from tail tip to the tip of the rostral horn) either before or after its recording series. Exoskeletal dimensions do not change between molts, and shrimp that molted during the recording process were not used.
Videotapes were recorded at 30 frames/s using a Sony V801 Hi8 videocamera
mounted above the tank. A mirror was placed at a 45° angle beside the tank
so that a lateral view, providing depth information, was recorded
simultaneously with the view from above. A 1-mm grid was placed below the tank
and lines marked at 2-cm intervals on the side of the tank for scale. A time
trace on individual frames was recorded on the tape. All recordings were of a
single shrimp and a single tadpole in the tank. Although most A.
callidryas hatch at night (Warkentin,
1995), recordings were made under natural light in the afternoon
because shrimp would not hunt at night under artificial light adequate for
recording.
Individual eggs were removed from clutches, induced to hatch when necessary, and the tadpoles allowed to fall into the water. Each tadpole was taped for 0-15 min, depending on survivorship. I removed the tadpole and replaced it with another if a shrimp failed to consume it within 15 min. When a shrimp caught a tadpole, I waited 5 min before presenting the next tadpole to ensure that the shrimp was ready to hunt again; shrimp consume tadpoles within a minute. Younger (5 or 6 days) and older (7 or 8 days) tadpoles were presented alternately. If a shrimp neither attacked nor searched for a tadpole within 15 min, it was offered no more tadpoles that day. To control hunger levels, if a shrimp consumed four tadpoles no more were offered that day, and if a shrimp failed to eat at least two tadpoles it was fed embryos to make up the difference after recording was over for the day. Shrimp were offered a maximum of five tadpoles per day.
I recorded three tadpoles of each of the four hatching ages exposed to each of eight shrimp. In three cases, one of the first three tadpoles fell on the shrimp upon entering the water and was eaten immediately. In these cases I recorded a fourth tadpole (n = 99 tadpoles: 3x4x8 + 3). Experiments used only one tadpole of each age per sibship (24-25 sibships/age).
Videotape analysis
All interactions between shrimp and tadpoles were examined. Behavioral data
were also recorded every 60 s to provide a baseline measure independent of
interactions.
Tadpole behavior was categorized as motionless, coasting (moving without beating tail, after swimming), spontaneous swimming (not in response to shrimp, typically slow), or fleeing (sprinting in response to nearby shrimp movement). Shrimp behavior was divided into hunting and other. Hunting shrimp walked or swam with antennae extended or antennal movement and attacked tadpoles they encountered. Most other behavior was resting and grooming.
Three types of interactions were analyzed. (1) Close encounters: shrimp and tadpole came within 4 cm horizontally and 2 cm vertically. In the absence of a behavioral indicator of prey detection independent of attack, I assumed that probability of detection increases with proximity and chose a spatial definition of "close" based on the distribution of a preliminary subset of attack data and on shrimp body size. (2) Attacks: shrimp oriented toward tadpole and approached with pincers extended. I divided attacks into initial and secondary attacks. In initial attacks a shrimp attacked a newly located tadpole. In secondary attacks a shrimp had previously located and attacked a tadpole, which fled, and the shrimp followed and attacked again. (3) Fleeing: tadpole sprinted in response to nearby shrimp movements.
Shrimp position was specified as mouth position, and tadpole position as the center of the body. Positions were measured as height above the bottom and distance from the wall. For interactions, the time, positions of both animals, distance between animals, immediately preceding tadpole behavior, and interaction outcome were recorded. In attacks without a clear initiation point, data were taken for both the earliest and latest possible attack initiation point. For baseline data, position and behavior of both animals were recorded. Distances and positions were measured directly from a monitor against the recorded grid using calipers.
Due to the curvature of the tank, there were two small areas that were not visible in the mirror. Thus for some interactions not all data were obtained. Interactions in these parts of the tank should not differ systematically from interactions in any other part of the tank. Prior recordings of shrimp and tadpoles in a rectangular tank, in which all interactions were visible in the mirror, indicated an artifactually high mortality rate in the corners (K. M. Warkentin, unpublished data).
Analysis of data from videotapes
I tested the hypotheses that tadpole behavior changed with age and that
tadpole behavior affected the probability and the success of shrimp attacks. I
analyzed 504 attacks, 618 shrimpinduced sprints (fleeing), 375 close
encounters, and 795 sets of baseline position data. These included 63
fatalities. To test for developmental differences in tadpole behavior, I
averaged behavioral measurements within individual tadpoles, pooled tadpoles
across shrimp, and tested for age effects across tadpoles with Kruskal-Wallis
one-way ANOVA. I pooled tadpoles within shrimp to test for effects of tadpole
behavior on the hunting success of shrimp or survival of tadpoles. For each
shrimp, I calculated averages of its successes and its failures, then compared
successes with failures across shrimp with a Friedman two-way ANOVA. Using
sequential Bonferroni corrections for multiple tests
(Holm, 1979), all results
remain significant (p <.05), except in one case which is noted. I
report original significance levels in the results.
I used only initial attacks to assess factors affecting prey detection because in secondary attacks tadpoles were detected earlier and followed. As a measure of the distance from which shrimp attacked tadpoles in attacks without a clear initiation point, I averaged distances at earliest and latest possible attack initiation (n = 94, 14 without vertical distance). I combined these averages with measured distances at attack initiation from clear cases (n = 285, 16 without vertical distance). Using only clear cases gives similar results.
Because shrimp are largely benthic, tadpoles on the bottom may be more frequently encountered and attacked. I determined if the proportion of close encounters and attacks that occurred on the bottom differed from a null hypothesis based on tadpole microhabitat use, using baseline position data (recorded every 60 s) to estimate the tadpole distribution to which each hunting shrimp was exposed.
To assess effect of tadpole movement on shrimp attack, I compared the
frequency of initial attacks on motionless and moving tadpoles by nearby
shrimp and the distance from which shrimp attacked motionless and moving
tadpoles. For attack frequency I examined 375 close encounters, using only
cases in which the sequence of attack and sprint, if any, was unambiguous
(
1/30 s apart). "Attacked, moving" tadpoles moved before
attacked; this includes tadpoles that fled before attacked plus coasting and
spontaneously swimming tadpoles. "Attacked, motionless" tadpoles
moved only after attacked, if at all. For attack distance, I examined 324
initial attacks. As above, I used average distance in attacks without a clear
initiation point.
To test if use of the bottom microhabitat changes ontogenetically, I compared the proportion of baseline position data at 0 cm height across tadpole ages. To test if development affects the likelihood that tadpoles will flee from shrimp, I compared the proportion of times that tadpoles of each age fled in close encounters and in attacks. I compared the number of times tadpoles fled from attacks before capture across tadpole ages for tadpoles caught because they failed to flee. I also calculated the average distance at which each tadpole fled from attacking shrimp (headstart distance), then compared across ages.
I pooled all attacks to estimate the relative importance of position, fleeing, headstart distance, and evasive maneuvers in tadpole survival using partial regression coefficients. For each case, tadpole survival was coded as 1 (survived) or 0 (died). Position (up or on the bottom), evasive maneuvers (performed or not), and fled (or not) were likewise categorical.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Predation experiment
Older tadpoles survived better with shrimp than younger tadpoles (Figure 1; F1,55 = 12.432, p =.001). There was no significant effect of habitat structure, either alone or as an interaction with age (Figure 1; structure: F1,55 = 2.217, p =. 142; age*structure: F1,55 = 2.274, p =.137).
|
Behavioral observations
Behavior did not vary significantly among sibships (MANOVA for main effect
of sibship and all interactions including sibship, all p >.27).
Tadpoles rested on the container bottom, adhered to the water surface using
their external mouthparts, adhered to the container side using mouthparts or
cement glands, floated mid-water, or swam. Both tadpole development and the
presence of shrimp influenced tadpole activity (Figure
2A,
2B and
Table 1). In the absence of
shrimp, tadpoles moved infrequently and spent little time moving. The youngest
hatchlings either lay on the bottom (Figure
2C) or adhered to a surface
(Figure 2D). Older hatchlings
spent a little time floating mid-water or slowly swimming
(Figure 2E). In the presence of
shrimp, tadpoles moved more frequently and spent more time moving (Figure
2A,
2B and
Table 1). Most movement was
sprinting in response to contact with or approach by the shrimp or disturbance
by another active tadpole.
|
|
There was a significant shrimp-by-age interaction effect on tadpole microhabitat use (Table 1). Tadpoles spent less time on the bottom in the presence of shrimp, and this response was stronger in older, more developed tadpoles (Figure 2C). Reduced use of the bottom is necessarily coupled with increased use of other microhabitats. In the presence of shrimp, older tadpoles increased use of the mid-water microhabitat more than did younger tadpoles (Figure 2E; Kruskal-Wallis test of age effect on change in behavior, p =.001). Occasionally, 5-day hatchlings at the water surface fell to the bottom; this was never observed in older hatchlings.
Videotaped predator-prey interactions
Prey detection
Shrimp seldom responded immediately when a tadpole fell into the water, but
began antennating a few moments later, and hunting shortly thereafter. Thus
tadpoles were detected shortly after they entered the water, although shrimp
usually took longer to locate tadpoles precisely. I could not reliably
determine when a shrimp had located a tadpole except through its attack
behavior: orientation and approach with pincers extended. Often attacking
shrimp also folded their antennae back along their body, away from the
tadpole.
Most initial attacks occurred when the animals were within 4 cm horizontally (68±5%) and within 2 cm vertically (68±4%, n = 8 shrimp; Figure 3). I considered tadpoles within this space close to the shrimp; 47±6% of initial attacks were on close tadpoles. Shrimp were 33.9±0.99 mm in total length, with pincers 18.3±0.9 mm and antennae 43.8±2.7 mm (n = 8).
|
Microhabitat use and effect on predation
Tadpoles mostly hung from the water surface or rested on the bottom; they
floated and swam mid-water and hung from the sides of the tank less often
(Figure 4A). Development
affected microhabitat use; younger tadpoles spent more time on the bottom of
the tank than did older tadpoles (Figure
4A, and Table 2).
Shrimp spent most of their time on the bottom
(Figure 4B).
|
|
Shrimp did not come close to or attack tadpoles on the bottom (
1 cm
up) more often than expected from the proportion of time tadpoles spent there
(for both comparisons, Friedman test statistic = 2, p =.157). Attacks
on tadpoles on the bottom were, however, more often fatal than other attacks
(Figure 5 and
Table 3). To control for any
correlation between tadpole position and fleeing, I considered only cases when
tadpoles initially fled. Regardless, shrimp were more successful attacking
tadpoles on the bottom (Figure
5, Table 3).
Restricting the analysis to cases when tadpoles fled nonstop under pursuit,
the trend is nonsignificant (Figure
5 and Table 3),
with only six shrimp catching tadpoles.
|
|
Movement: effect on shrimp attack
Nearby shrimp attacked motionless tadpoles more often than moving tadpoles
(Figure 6A). Using a sequential
Bonferroni correction for three tests of effects of tadpole behavior on attack
probability, this effect is marginally nonsignificant (Friedman test statistic
= 4.5, p =.034 versus. 
=.017); however, tadpole movement
clearly does not increase the risk of attack by nearby shrimp. There was no
significant difference in the distance from which shrimp attacked motionless
and moving tadpoles (Figure 6B;
Friedman test statistic = 2, p =.157).
|
Movement: ontogeny
Older tadpoles more often fled when close to or attacked by shrimp
(Figure 7A and
Table 2). Older tadpoles fled
from more attacks before failing to flee
(Figure 7B and
Table 2). Older tadpoles also
began fleeing at greater distances from the shrimp, giving themselves a
greater headstart than younger tadpoles
(Figure 7C and
Table 2).
|
Movement: effect on mortality
Most mortality occurred when tadpoles failed to flee
(Figure 8 and
Table 4). Every attacked
tadpole that did not flee was caught and eaten. This includes both tadpoles
that never fled and those that stopped while being pursued and did not flee
again. This 100% success rate of attacks on tadpoles that did not flee
contrasts sharply with the 5.6±0.9% success rate of attacks on tadpoles
that initially fled, and the 2.5±0.7% success rate of attacks on
tadpoles that fled continuously while pursued
(Table 3). Shrimp often did not
chase fleeing tadpoles (Figure
8). Even considering only chases, the capture rate was only
13.0±4.3% for initially fleeing tadpoles and 4.3±1.1% for
tadpoles that continued to flee. There were two cases in which shrimp briefly
caught a fleeing tadpole which then immediately escaped their grasp. No
motionless or coasting tadpole grasped by a shrimp escaped.
|
|
Considering only attacks from which the tadpole fled, the headstart distance of tadpoles in fatal attacks was less than their headstart distance in nonfatal attacks (Figure 7C and Table 3).
Tadpoles that fled and were followed by shrimp made evasive maneuvers,
defined as sudden changes of direction
(Feder, 1983;
Watkins, 1996) 21.2±4%
of the time. The chance of making such maneuvers did not change significantly
with age (Table 2). Chasing
shrimp were more likely to cease following tadpoles that made evasive
maneuvers, compared with tadpoles that did not
(Figure 8 and
Table 3).
Which behavior most affects survival?
Fleeing explained much of the variance in survival and was clearly more
important than any other factor. In a model including fleeing and position,
r2 =.755 and standardized regression coefficients were.857
and.050, respectively (n = 488 cases). For a second model including
position, headstart distance, and evasion and using only cases where the
tadpole fled (n = 392), r2 =.031 and coefficients
were.095,.128, and.049, respectively.
Figure 8 summarizes the sequence of tadpole and shrimp behavior following an attack, showing the ways in which tadpoles escaped or were caught. By far the most common cause of death was failure to flee. In contrast, tadpoles that fled usually survived. On average 50% of tadpoles that fled were not chased (56.3±8% for initial attacks). Infrequent evasive maneuvers strongly affected a tadpole's chance of escape: shrimp pursued tadpoles through only 11% of evasive maneuvers versus 67% of nonevasive fleeing. Nonetheless, relatively few tadpoles were caught while fleeing. In many cases tadpoles stopped or slowed to coasting while still pursued. A substantial number were caught because they were attacked again and failed to flee. Others fled again and many were not chased; stopping and fleeing again was slightly but not significantly more effective in escaping shrimp than just continuing to flee (45.3±11% versus 33±12% escaped, Friedman test statistic = 2, p =.157).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
The overwhelming factor affecting survival of red-eyed tree frog tadpoles in the presence of shrimp is whether or not the tadpoles flee when attacked. By far the majority of tadpoles that flee escape predation. In contrast, I found no evidence that tadpoles that fail to flee ever escape. Tadpole activity does not increase the risk of shrimp attack; if anything, the risk is reduced. Thus tadpoles exposed to shrimp do not face a trade-off between remaining motionless to avoid attack and swimming away to avoid capture. Likewise, there is no tradeoff between activity-dependent feeding and an inactivity-dependent defense. This is in sharp contrast to most studies of tadpoles with odonate predators, where complete immobility substantially improves survival by reducing risk of attack, fleeing appears less important, and a feeding versus defense trade-offs exists (Anholt and Werner, 1995
;
Chovanec, 1992;
Skelly, 1992,
1994,
1995;
Skelly and Werner, 1990;
Werner, 1991;
Werner and Anholt, 1996). Even
with odonates, however, increased activity sometimes reduces predator success
(Hews, 1988).
Because habitat complexity did not alter tadpole survivorship, leaf litter
should not be considered a refuge from shrimp, and it is reasonable to examine
behavioral defenses in a structurally simple environment. Also, because cover
does not reduce mortality of the youngest, least active tadpoles, the failure
of immobility as a defense in the videotaped interactions cannot be attributed
to lack of cover. Shrimp forage actively and will eat plant material and
scavenge as well as attack live prey
(Garces and Heinen, 1993;
New, 1976). Food searching and
grasping are elicited by chemical stimuli in a congner
(Harpaz and Steiner, 1990),
and in M. americanum prey detection and recognition are clearly not
dependent on movement. Given the low frequency with which pursuing shrimp
catch fleeing tadpoles, the average energetic gain may be low, and the
energetic costs of attack and pursuit may select against such behavior. In
contrast, the 100% success of attacks on tadpoles that fail to flee makes
attacking motionless tadpoles advantageous.
Younger tadpoles flee from shrimp attacks significantly less often than do older tadpoles; 8-day hatchlings fail to flee in only 4% of attacks, whereas 5-day hatchlings fail to flee in almost half (Figure 7A). This difference seems sufficiently large to explain the roughly twofold higher mortality of 5-versus 8-day tadpoles in predation experiments. Indeed, the lower mortality of older tadpoles is largely due to fewer "failed to flee" fatalities (Table 4).
Other factors explain relatively little of the variation in survival, but still differ significantly between fatal and non-fatal attacks. Tadpoles on the bottom are less likely to survive shrimp attacks than are tadpoles in other locations (Figure 5). This could contribute to improved survival for older hatchlings because the proportion of time spent on the bottom declines with development. Although shrimp spend most of their time on the bottom, I did not detect an effect of position on likelihood of attack; shrimp on the bottom swam up to attack tadpoles above them. My experimental design probably underestimates the importance of position because tadpole-shrimp interactions were videotaped in water only 6 cm deep, and many A. callidryas breeding ponds are much deeper. Microhabitat use of shrimp and A. callidryas hatchlings in the wild parallels that in my tanks; shrimp use the bottom and other substrates, and hatchlings mostly lie on the substrate or hang from the water surface (K. M. Warkentin, personal observation). In deeper water hatchlings at the surface are much farther from shrimp on the bottom than they were in my experiment, which may reduce attack probability.
Although A. callidryas tadpoles infrequently perform evasive
maneuvers (sudden changes of direction), such maneuvers dramatically increase
the likelihood of escape from pursuing shrimp
(Figure 8,
Table 3). Evasive maneuvers are
also important in the escape of tadpoles from garter snakes and turtles
(Feder, 1983;
Watkins, 1996). These
predators, like shrimp and unlike odonates, are active foragers that chase
tadpoles. Evasiveness of red-eyed tree frog tadpoles did not change
developmentally, and so evasiveness does not contribute to the lower mortality
of older hatchlings.
Hatchlings respond behaviorally to shrimp in two ways that reduce mortality. First, tadpoles of all ages are more active in the presence of shrimp (Figure 2A, 2B). Increased activity is largely due to tadpoles fleeing from shrimp, which clearly reduces the success of attacks and may also make attack less likely. Second, tadpoles of all ages reduce use of the bottom microhabitat in the presence of shrimp, and this response is stronger in older tadpoles (Figure 2C). I detected a relatively weak effect of tadpole position on attack success, but in deeper water tadpoles may also avoid attacks by avoiding the bottom, and other substrates.
The ontogenetic change in microhabitat use may be influenced by three
aspects of development—specifically lung and tail growth and
differentiation of external mouthparts
(Warkentin, 1998,
in press). Lung size in A.
callidryas embryos increases substantially from 5 to 8 days, allowing
older hatchlings to achieve neutral and even positive buoyancy, whereas 5-day
hatchlings consistently sink (Warkentin,
1998). To avoid the bottom and other substrates, negatively
buoyant young tadpoles must either swim or adhere to the water surface. The
smaller tails of 5-day hatchlings may hamper swimming performance, and their
less developed mouthparts sometimes fail to adhere to the water surface
(Warkentin, 1998). Thus
bottom-avoidance is more energetically costly for younger tadpoles, and
morphology appears to limit behavior.
The cause of the ontogenetic change in propensity to flee from shrimp is less obvious. Because there is no trade-off between avoiding and escaping attacks, natural selection by shrimp predation is directional and favors fleeing. Why do the youngest hatchlings flee so infrequently? I consider three hypotheses. (1) Not fleeing is adaptive; other, stronger, selection pressures in the water oppose selection by shrimp predation and favor immobility. (2) Not fleeing reflects developmental limitations. (3) Not fleeing is a pleiotropic effect of selection on embryos.
Although tadpole immobility increases predation by shrimp, it might be
advantageous in a larger context. Specifically, other predators may
selectively prey on active tadpoles. Odonates and other sit-and-wait predators
are likely candidates for such agents of selection. Shrimp, however, are by
far the most abundant predator of tadpoles in the study area, with dense
populations in most A. callidryas breeding sites
(Warkentin, 1995). As well, in
preliminary experiments shrimp ate more tadpoles per capita than did other
invertebrate predators. Poeciliid fish co-occur with and prey on A.
callidryas tadpoles in some breeding sites
(Warkentin, 1995). However,
most poeciliids are omnivores and, like shrimp, will readily consume immobile
food items (Meffe and Snelson,
1989). More importantly, if inactivity truly confers a net
benefit, why are older hatchlings so active? Because of the delayed hatchling
of undisturbed clutches, young A. callidryas are less likely than
older animals to be in the water. Selection on tadpoles in the water should
thus be more effective on later stages.
A variant of the first, adaptive, hypothesis is that selection pressures on tadpoles change with age such that immobility confers a net advantage only for the youngest hatchlings. This could occur if fleeing were a less effective defense for younger tadpoles. The frequency with which fleeing tadpoles were captured declined slightly with age; however, even fleeing 5-day hatchlings were captured only 6.6% of the time. Compared with the 100% mortality of nonfleeing tadpoles, an opposing selection pressure would need to be extremely strong for failure to flee to be an adaptation to aquatic conditions. There are no obvious candidates for such a strong selection pressure.
A second hypothesis is that failure to flee results from incomplete
development of sensory and motor systems, not direct selection. The youngest
tadpoles may simply not be sufficiently developed to consistently detect or
respond to approaching shrimp. Nonetheless, these tadpoles did flee from
shrimp attacks half the time, suggesting that they do detect shrimp and can
respond, and even the youngest hatchlings have well-developed neuromasts and
eyes (Warkentin, 1998).
A third hypothesis is that selection pressures on embryos may oppose those
on tadpoles, making young hatchlings less likely to flee than they otherwise
would be. An undisturbed arboreal egg clutch is a safe place, much safer than
most ponds for small, young tadpoles. Undisturbed clutches do not hatch at 5
days, and few undisturbed embryos hatch at 6 days; most wait until 7 or 8 days
(Warkentin, 1995). If
remaining in a safe egg is favored through 7 days, younger arboreal embryos
should not hatch in response to minor disturbances such as wind, rain, or
spiders walking on them; indeed, they do not
(Warkentin, 1995).
Rapid early hatching is a behavioral response to disturbance by egg
predators. Agalychnis callidryas embryos are attached to vegetation
by tough, elastic jelly. In removing eggs from the clutch, predators create a
substantial disturbance that induces rapid hatching
(Warkentin, 1995). A.
callidryas hatch by vigorous movements that rupture the egg membrane;
motionless embryos do not hatch. The hatching movements appear similar to the
escape response of posthatching larvae, as in fish
(Eaton and Nissanov, 1985).
This suggests that escape hatching and fleeing from aquatic predators may
share an underlying sensory and motor mechanism, a common neuromuscular
pathway. If so, selection against unnecessarily early hatching might reduce
sensitivity or responsiveness of this pathway in younger embryos, with
correlated effects on the response of young hatchlings to aquatic predators.
After locating a motionless tadpole, shrimp often approach slowly and do not
touch the tadpole until the fatal strike. In such cases the vibratory cues
transmitted to the tadpole must be substantially weaker than the disturbance
to embryos during snake attack. Selection by aquatic predators on young
hatchlings for increased sensitivity to disturbance would fall only on the
fraction of tadpoles actually in the water at that stage, those just attacked
in the egg. Older animals are more likely to be in the water, thus selection
by aquatic predators would be more effective at later ages, and we expect
older hatchlings to be better adapted to aquatic life.
The importance of fleeing as a defense against predatory shrimp contrasts
strikingly with the defensive role of inactivity with odonate naiads
(Anholt and Werner, 1995;
Skelly, 1994). Clearly,
behavioral defenses are not universally effective; behavior that improves
survival with one predator can seriously reduce survival with another. Thus
the efficacy of putative defenses must be considered in the context of an
entire suite of predators, and particular behaviors should not be considered
defenses without predator-specific demonstration of their effectiveness. When
different predators impose opposing selection pressures on prey defenses, prey
may evolve specific responses to different predator types. Such response
specificity is known in primates (Cheney
and Seyfarth, 1990) and in birds
(Evans et al., 1993). Some
amphibian larvae also distinguish among predators, responding behaviorally
only to those predators against which the behavior functions in defense
(Lefcort, 1996), and showing
stronger responses to more dangerous predators
(Kiesecker et al., 1996). Only
by recognizing the often restricted effectiveness of antipredator defenses
will we be able to determine to what extent multiple predators have shaped
prey phenotypes and how conflicting selection pressures generated by different
predators affect both response specificity of prey and their ecological
interactions with a suite of predators.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
I thank the National Park Service of Costa Rica for permission to work at Sirena and the staff of Sirena Station, especially P. Valverde and G. Fonseca. H. Gould, J. Strom, K. Keaton, A. Ginzel, and F. Macdonald helped collect data from the videotapes. I thank M. Ryan, G. Rosenthal, R. Dudley, D. Cannatella, R. Buskirk, C. Gabor, L. Higgins, and R. Brown for comments on the manuscript and R. Dudley and G. Rosenthal for access to equipment. I was supported during this work by the Natural Sciences and Engineering Research Council of Canada and the University of Texas. The research was funded by the University of Texas Department of Zoology and Institute of Latin American Studies, Sigma Xi, the Percy Sladen Memorial Fund of the Linnean Society, the Gaige fund of the American Society of Ichthyologists and Herpetologists, and a Dissertation Improvement Grant from the National Science Foundation (NSF). Laboratory facilities at Sirena were made possible by an NSF grant to L. Gilbert.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Anholt BR, Werner EE, 1995. Interaction between food availability and predation mortality mediated by adaptive behavior.Ecology 76:2230 -2234.
Ash VB, 1989. Resistance to shell breaking in two intertidal snails. Zool J Linn Soc 96:167-184.
Boppré M, 1990. Lepidoptera and pyrrolizidine alkaloids: exemplification of complexity in chemical ecology. J Chem Ecol 16:165-185.
Bridges CM, Gutzke WHN, 1997. Effects of environmental history, sibship, and age on predator-avoidance responses of tadpoles.Can J Zool 75:87-93.
Brodie ED Jr, Formanowicz DR Jr, Brodie ED III, 1978. The development of noxiousness of Bufo americanus to aquatic insect predators. Herpetologica 34:302-306.
Brönmark C, Miner JG,1992
. Predator-induced phenotypical change in body morphology in
crucian carp. Science
258:1348-1350.
Brown RM, Taylor DH, 1995. Compensatory escape mode trade-offs between swimming performance and maneuvering behavior through larval ontogeny of the wood frog, Rana sylvatica.Copeia 1995:1-7.
Caldwell JP, Thorp JH, Jervey TO, 1980. Predator-prey relationships among larval dragonflies, salamanders and frogs.Oecologia 46:285-289.
Calef GW, 1973. Natural mortality of tadpoles in a population of Rana aurora. Ecology 54:741-758.
Cheney DL, Seyfarth RM, 1990. How monkeys see the world. Chicago: University of Chicago Press.
Chivers DP, Brown GE, Smith RJF, 1995. Acquired recognition of chemical stimuli from pike, Esox lucius, by brook sticklebacks, Culaea inconstans (Osteichthyes, Gasterosteidae).Ethology 99:234-242.
Chovanec A, 1992. The influence of tadpole swimming behaviour on predation by dragonfly nymphs. Amphib Reptil 13:341-349.
Craig CL, 1994. Limits to learning: effects of predator pattern and color on perception and avoidance-learning by prey.Anim Behav 47:1087-1099.
Cronin JT, Travis J, 1986. Size-limited predation on larval Rana areolata (Anura: Ranidae) by two species of backswimmer (Insecta: Hemiptera: Notonectidae). Herpetologica 42:171-174.
Crump ML, 1984. Ontogenetic changes in vulnerability to predation in tadpoles of Hyla pseudopuma.Herpetologica 40:265-271.
Dyer LA, 1995. Tasty generalists and nasty specialists? Antipredator mechanisms in tropical lepidopteran larvae.Ecology 76:1483-1496.[ISI]
Eaton RC, Nissanov J, 1985. A review of Mauthner-initiated escape behavior and its possible role in hatching in the immature zebrafish, Brachydanio rerio. Environ Biol Fishes 12:265-279.
Evans CS, Evans L, Marler P, 1993. On the meaning of alarm calls: functional reference in an avian vocal system. Anim Behav 46:23-38.
Feder ME, 1983. The relation of air breathing and locomotion to predation on tadpoles, Rana berlandieri, by turtles.Physiol Zool 56:522-531.
Formanowicz DR Jr, 1986. Anuran tadpole/aquatic insect predator-prey interactions: tadpole size and predator capture success.Herpetologica 42:367-373.
Formanowicz DR Jr, Brodie ED Jr, 1982. Relative palatabilities of members of a larval amphibian community.Copeia 1982:91-97.
Fuiman LA, Magurran AE, 1994. Development of predator defences in fishes. Rev Fish Biol Fish 4:145-183.
Garces CL, Heinen JM, 1993. Evaluation of beef liver, fruits and vegetables as dietary supplements for indoor nursery culture of freshwater prawns Macrobrachium rosenbergi. J World Aquacult Soc 24:58-65.
Harpaz S, Steiner JE, 1990. Analysis of betaine-induced feeding behavior in the prawn Macrobrachium rosenbergi (de Man, 1879) (Decapoda, Caridea). Crustaceana 58:175-185.
Harvell CD, 1990. The ecology and evolution of inducible defenses. Q Rev Biol 65:323-339.[Medline]
Hews DK, 1988. Alarm response in larval western toads, Bufo boreas: release of larval chemicals by a natural predator and its effect on predator capture efficiency. Anim Behav 36:125-133.
Hoff K, 1987. Morphological determinants of fast-start performance in anuran tadpoles (PhD dissertation). Halifax: Dalhousie University.
Holm S, 1979. A simple sequentially rejective multiple test procedure. Scand J Stat 6:65-70.
Horat P, Semlitsch RD, 1994. Effects of predation risk and hunger on the behaviour of two species of tadpoles. Behav Ecol Sociobiol 34:393-401.
Huey RB, 1980. Sprint velocity of tadpoles (Bufo boreas) through metamorphosis. Copeia 1980:537-540.
Jackson ME, Semlitsch RD, 1993. Paedomorphosis in the salamander Ambystoma talpoideum: effects of a fish predator.Ecology 74:342-350.
Kats LB, Petranka JW, Sih A, 1988. Antipredator defenses and the persistence of amphibian larvae with fishes.Ecology 69:1865-1870.
Kiesecker JM, Chivers DP, Blaustein AR, 1996. The use of chemical cues in predator recognition by western toad tadpoles. Anim Behav 52:1237-1245.
Lefcort H, 1996. Adaptive, chemically mediated fright response in tadpoles of the southern leopard frog, Rana utricularia.Copeia 1996:455-459.
Magurran AE, Girling SL, 1986. Predator model recognition and response habituation in shoaling minnows. Anim Behav 34:510-518.
McCollum SA, Van Buskirk J, 1996. Costs and benefits of a predator-induced polyphenism in the gray treefrog Hyla chrysoscelis. Evolution 50:583-593.
McVean A, 1982. Autotomy. In: The biology of crustacea (Sandeman DC, Atwood HL, eds) New York: Academic Press;107 -132.
Meffe GK, Snelson FF Jr, 1989. An ecological overview of poeciliid fishes. In: Ecology and evolution of livebearing fishes (Poeciliidae) (Meffe GK, Snelson FF Jr, eds). Englewood Cliffs, New Jersey: Prentice-Hall; 13-31.
New MB, 1976. A review of dietary studies with shrimp and prawns. Aquaculture 9:101-144.
Osenberg CW, Mittelbach GG, 1989. Effects of body size on the predator-prey interaction between pumpkinseed sunfish and gastropods.Ecol Monogr 59:405-432.
Palmer RA, 1990. Predator size, prey size, and the scaling of vulnerability: hatchling gastropods vs. barnacles.Ecology 71:759-775.
Parichy DM, Kaplan RH, 1995. Maternal investment and developmental plasticity: functional consequences for locomotor performance of hatchling frog larvae. Funct Ecol 9:606-617.
Petranka JW, Kats LB, Sih A, 1987. Predator-prey interactions among fish and larval amphibians: use of chemical cues to detect predatory fish. Anim Behav 35:420-425.
Polis GA, 1984. Age structure component of niche width and intraspecific resource partitioning: can age groups function as ecological species? Am Nat 123:541-564.
Polis GA, Myers CA, Holt RD, 1989. The ecology and evolution of intraguild predation: potential competitors that eat each other.Annu Rev Ecol Syst 20:297-330.[ISI]
Punzo F, 1997. Leg autotomy and avoidance behavior in response to a predator in the wolf spider, Schizocosa avida (Araneae, Lycosidae). J Arachnol 25:202-205.
Semlitsch RD, 1990. Effects of body size, sibship and tail injury on the susceptibility of tadpoles to dragonfly predation.Can J Zool 68:1027-1030.
Semlitsch RD, Gibbons JW, 1988. Fish predation in size-structured populations of treefrog tadpoles. Oecologia 75:321-326.
Sih A, Kats LB, 1994. Age, experience, and the response of streamside salamander hatchlings to chemical cues from predatory sunfish. Ethology 96:253-259.
Skelly DK, 1992. Field evidence for a cost of behavioral antipredator response in a larval amphibian. Ecology 73:704-708.
Skelly DK, 1994. Activity level and the susceptibility of anuran larvae to predation. Anim Behav 47:465-468.
Skelly DK, 1995. A behavioral trade-off and its consequences for the distribution of Psuedacris treefrog larvae.Ecology 76:150-164.
Skelly DK, Werner EE, 1990. Behavioral and life-historical responses of larval American toads to an odonate predator.Ecology 71:2313-2322.[ISI]
Smith DC, Van Buskirk J, 1995. Phenotypic design, plasticity, and ecological performance in two tadpole species. Am Nat 145:211-233.
Warkentin KM, 1995. Adaptive plasticity in hatching
age: a response to predation risk trade-offs. Proc Natl Acad Sci
USA
92:3507-3510.
Warkentin KM, 1998. Phenotypic plasticity at hatching in the red-eyed treefrog, Agalychnis callidryas: life history, behavior and development (PhD dissertation). Austin: University of Texas.
Warkentin KM, in press. Effects of hatching age on development and hatchling morphology in the red-eyed treefrog, Agalychnis callidryas.Biol J Linn Soc .
Wassersug RJ, Sperry DG, 1977. The relationship of locomotion to differential predation on Pseudacris triseriata (Anura: Hylidae). Ecology 58:830-839.
Watkins TB, 1996. Predator-mediated selection on burst swimming performance in tadpoles of the Pacific tree frog, Psuedacris regilla. Physiol Zool 69:154-167.
Werner EE, 1988. Size, scaling, and the evolution of complex life cycles. In: Size-structured populations (Ebenman B, Persson L, eds). Berlin: Springer-Verlag;60 -81.
Werner EE, 1991. Nonlethal effects of a predator on competitive interactions between two anuran larvae. Ecology 72:1709-1720.
Werner EE, Anholt BR, 1996. Predator-induced behavioral indirect effects: consequences to competitive interactions in anuran larvae. Ecology 77:157-169.
Werner EE, Gilliam JF, 1984. The ontogenetic niche and species interactions in size structured populations. Annu Rev Ecol Syst 15:393-425.[ISI]
Werner EE, Hall DJ, 1988. Ontogenetic habitat shifts in bluegill: the foraging rate-predation risk trade-off.Ecology 69:1352-1366.
Wisendon BD, Chivers DP, Smith RJF, 1997. Learned recognition of predation risk by Enallagma damselfly larvae (Odonata: Zygoptera) on the basis of chemical cues. J Chem Ecol 23:137-151.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. L. Prudic, J. C. Oliver, and F. A. H. Sperling From the Cover: The signal environment is more important than diet or chemical specialization in the evolution of warning coloration PNAS, December 4, 200 |