Behavioral Ecology Vol. 12 No. 1: 51-58
© 2001 International Society for Behavioral Ecology
A comparative study of activity levels in larval anurans and response to the presence of different predators
Department of Biological Sciences, Dartmouth College, Hanover, NH 03755, USA
Address correspondence to J. M. L. Richardson, Department of Zoology, University of Toronto, 25 Harbord St., Toronto, Ontario M5S 3G5, Canada. E-mail: jmlr{at}zool.utoronto.ca .
Received 18 November 1999; revised 18 May 2000; accepted 26 May 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Activity level is a key behavioral trait in many animals which mediates a trade-off between finding food and avoiding predation. Optimal activity level will therefore depend on environment, and plasticity in response may increase fitness (if an organism encounters multiple environments in a lifetime). One group in which activity level, and its relationship to foraging and predation risk, has been well studied is larval anurans. Anurans inhabit a range of distinct freshwater aquatic community types that are created by differences in pond permanency and top predator. Species segregate across these pond types and therefore tadpoles from different species encounter different selection regimes. I hypothesized that species from different pond types would therefore differ in activity behavior, and in plasticity of this behavior. I tested this in a phylogenetic framework to consider the evolution of plasticity in anurans diversifying into different pond types. Time spent active was quantified for larvae of each of 13 anuran species (from three taxonomic families) in four conditions: when no predator was present, and in the non-lethal presence of a dragonfly, newt, or fish predator. Species nested within pond type by taxonomic family differed significantly in time spent active. A significant interaction between predator treatment and taxonomic family was also observed. A phylogenetic analysis of change in behavior revealed strong positive correlations in evolution of these behaviors and suggests constraints on the ability of larval anurans to independently modify activity levels in the presence versus absence of predators.
Key words: Anura, behavior, activity, anti-predator behavior, comparative study, tadpoles, phenotypic plasticity.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
All organisms face a trade-off between finding food and becoming food for a predator (Abrams, 1991
;
Lima, 1998;
Lima and Dill, 1990;
Werner and Anholt, 1993). In
many animals, this trade-off is mediated through behavioral changes in
activity levels (Anholt and Werner,
1995; Skelly,
1994). The trade-off is intuitive: individuals increase the
chances of finding food as they move about more actively, but in doing so they
also increase the chances of being detected by a predator. Theoretical models
that incorporate these trade-offs predict optimal activity levels that vary
with food availability and predator presence
(Abrams, 1991;
Werner and Anholt, 1993).
These models therefore predict that species encountering different selective
regimes will show species-specific differences in activity level. For example,
those species found in habitats where the selective force of predation risk is
strong relative to the selective force imposed by food limitation, are
predicted to be relatively less active than those species found in
environments where food limitation is relatively more important as a selective
agent.
Further, since both food levels and predation risk can vary temporally
within an organism's lifespan, and because behavior is labile, we might expect
that an optimal organism would display behavioral plasticity for the trait.
Indeed, empirical data show this can often be observed (reviewed in
Lima, 1998). For example,
Lawler (1989) showed that
amount of movement was significantly lower in tadpoles when a predator was
present. We can quantify plasticity in traits by viewing traits expressed in
each environment as separate traits and measuring the across environment
correlation of the traits; this is the approach taken in quantitative genetic
models of the evolution of plasticity
(Falconer, 1952;
Scheiner, 1993;
Van Tienderen and Koelewijn,
1994; Via and Lande,
1985). Both quantitative genetic and optimality models predict
that, for labile traits, plasticity will always be favored (reviewed in
Scheiner, 1993). However,
predictions of optimal plasticity assume no cost to plasticity and sufficient
genetic variance (Scheiner,
1993). Since these assumptions are unlikely to be true, we might
expect that organisms evolving under different selective regimes may display
differing levels of phenotypic plasticity.
The influence of selective regime and shared ancestry on the evolution of
traits (including plasticity) can be assessed using a comparative study
(Huey, 1987;
Pagel and Harvey, 1988).
Comparative studies require a system in which data can be collected from
several species, of known phylogenetic relationship, experiencing different
selection regimes. Larval anurans provide an ideal study system for
consideration of the evolution of plasticity in related species evolving under
different selective regimes. Breeding anurans tend to have nonoverlapping
distributions along a gradient of pond permanency. At one end of this gradient
are ephemeral ponds that dry annually, while at the other end of the spectrum
are permanent ponds and lakes. Correlated with these changes in hydroperiod
are changes in predators (McPeek,
1990a,
b;
Skelly, 1997;
Wellborn et al., 1996).
Variation in hydroperiod and predators creates different selective forces for
organisms in ponds from different positions along the gradient. Activity level
affects both foraging and predator avoidance traits, and is a behavioral trait
known to show plasticity in response to environmental conditions. Therefore,
activity level is expected to reflect the differences in habitat type across
the gradient of pond permanency (Collins
and Wilbur, 1979; Skelly,
1997; Wellborn et al.,
1996). For example, in permanent ponds with fish, fish predation
is a dominant selective force (e.g., McPeek,
1990a,
b), and the permanency of the
pond allows a strategy of slow growth rates to be successful. Thus, tadpoles
using these fish ponds should have relatively lower activity rates than
species using ponds with less predation pressure. Species using ponds with
more variable predation pressure, such as ponds with aquatic insect larvae as
top predators whose threat varies as the cohort matures, should be more likely
to display plasticity in activity levels than those species using ponds where
predation risk is relatively constant.
Larval anurans are also an ideal study system for addressing questions
about the evolution of plasticity because we have good information on
phylogenetic relationships among species in the three most speciose families
in North America (Hylidae, Ranidae, Bufonidae;
Figure 1). Species in each of
these three families have diversified into several different pond types (see
Methods for characterization of pond types). Thus, the group provides
replicate divergence into similar habitat types, allowing tests of phenotypic
change corresponding to habitat type. Furthermore, previous studies showing
differences in activity levels among anuran species, as well as differences
among species in behavioral response to predators
(Anholt et al., 2000;
Lawler, 1989;
Skelly, 1995;
Woodward, 1983) suggest
activity level is a trait which has been influenced by selection. Decreased
activity in tadpoles reduces predation risk
(Anholt and Werner, 1995;
Skelly, 1994).
|
In the present study, I quantify time spent active in larvae of 13 anuran
species when no predator is present and in the presence of different
predators. I test for effects of lineage and habitat type on tadpole
behavioral responses. Because activity rates influence both predation risk and
foraging rates (Anholt and Werner,
1995), I expect tadpoles from vernal ponds to be most active and
those from fish ponds to be least active. I quantified time active in the
laboratory using four treatment levels: (1) no predator present; (2) presence
of a dragonfly larva (Anax junius); (3) presence of a newt
(Notophthalmus viridescens); and (4) and presence of a sunfish
(Lepomis spp.). The three species represent the most common predators
encountered by tadpoles in the field across the pond types: Anax and
Notophthalmus in dragonfly ponds, and Lepomis in fish ponds
(Werner and McPeek, 1994).
Quantifying activity of each anuran species in the presence of each predator
allowed me to test whether species differentially responded to predators found
in the same pond type versus those found in different pond types, and also
provided data on whether species differed in plasticity for this trait.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
From 1996 to 1998, I collected tadpoles of 13 anuran species from eastern North America. Late stage tadpoles (Gosner stages 34-38, Gosner, 1960
) were caught by
dipnet, returned to the lab, and housed in 38 liter plastic tubs (36 x
50 x 21 cm) until placed in experimental tanks. Tadpoles were kept in
50% filtered pond water and 50% aged tap or well water; the same water was
used in all experiments. I fed animals `tadpole food' (pellets of processed
grain and plant products supplemented with vitamins; Carolina Biological
Supply Company, Burlington, NC, USA). Water was changed as needed (e.g.,
whenever water became fouled), typically every 4-7 days. All species thrived
under this regime. All tadpoles used in this study were at Gosner stage 36-38
(i.e., in hindlimb development; Gosner,
1960). I selected this stage because previous work has shown that
species differences in activity level are greatest at this developmental stage
(Lawler, 1989). To maximize the number of species from which I could collect data, I conducted experiments in New Hampshire, USA, during 1996 and 1998 (six species), and in Florida, USA during 1997 (seven species) (species listed in Figure 2). Tadpoles from New Hampshire, USA were housed at Dartmouth College in a room maintained at 20°C on a 14:10 light:dark cycle. Species from Florida were housed in the Florida State University Mission Road Greenhouse facility. The greenhouse received natural daylight and was maintained within a temperature range of approximately 23-27°C with heating elements and cooling fans; on rare occasion the temperature rose to between 28-30°C, but water temperature in greenhouse tanks was never greater than that measured in local, permanent ponds on the same day.
|
Tadpoles inhabit ponds that vary along a gradient of pond permanency; this
gradient has been well described and, although it is continuous in nature, two
clear transitions are seen along the gradient
(Semlitsch et al., 1996;
Skelly, 1997;
Wellborn et al., 1996).
Wellborn et al. (1996) termed
these the permanence transition and the predator transition and these two
transitions lead to three categories of pond which they label as
"Temporary habitats," "Permanent, fishless habitats,"
and "Habitats with fish" (see
Wellborn et al., 1996:
Figure 2). The difference in
prey species' assemblage between permanent fishless and fish ponds has been
well established for many different taxonomic groups (e.g., cladocerans;
Brooks and Dodson, 1965;
odonates; McPeek, 1990a,
b; anurans; Collins and
Wilbur, 1974; Morin, 1983;
Skelly, 1995;
Werner and McPeek, 1994).
Evidence suggests that it is the change in top predators which leads to
differential species' assemblages among these two types of permanent ponds
(reviewed in Wellborn et al.,
1996) and thus I will refer to habitats with fish as "fish
ponds" and permanent habitats without fish as "dragonfly
ponds" to emphasize the role of large aquatic insects, such as dragonfly
larvae, as the top predators in permanent fishless ponds (McPeek,
1990a,
b;
Skelly, 1996;
Werner and McPeek, 1994). Note
that pond permanency is a gradient and so that, although relatively permanent,
dragonfly ponds may dry on occasion (i.e., during drought conditions). At the
extreme of the gradient, on the other side of the "permanence
transition," are extremely temporary habitats; these dry predictably
every year and I will subsequently call these "vernal ponds."
I used these broad categories to classify the tadpole species I collected
according to pond type. Although I did not quantify predators in the ponds
from which I collected, I sampled the ponds from which I collected tadpoles
extensively for the 9 months I was in Florida and for two breeding seasons in
New Hampshire. I had strict and repeatable criteria for the classification of
pond types. Vernal ponds were only those habitats that were dry both in the
winter preceding the breeding season and in the winter after the breeding
season; these were shallow bodies of water (less than two feet deep) that
would clearly not overwinter. In other words, a pond was only considered
vernal if it dried predictably every year. In Florida, these were mostly low
lying areas that filled with water during the rainy season. In New Hampshire,
these included rock pools that filled in spring when rivers overflowed the
bank and deep tire tracks that filled with rain or snow melt. I classify
H. squirella as a vernal pond species because I found it only in such
ponds; in my sampling of over 30 ponds in Florida, I never collected H.
squirella in dragonfly ponds. In New Hampshire, the only species I found
in such vernal ponds was B. americanus (which is also found in fish
ponds). Dragonfly ponds were those ponds that held water both before and after
the breeding season, had had water continuously for several years prior, and
had no predatory fish (personal observation; M. McPeek, personal communication
for New Hampshire ponds; J. Travis, personal communication for Florida ponds).
Fish ponds are the most permanent ponds and all contained predatory fish
(personal observation; M. McPeek, personal communication; J. Travis, personal
communication). After 2 years of drought, subsequent to this study, the fish
ponds in northern Florida from which I collected all still had water, while
the dragonfly ponds from which I collected all became dry. Species that I
classified as dragonfly pond species were found in many different dragonfly
ponds, but were never found in fish ponds. Likewise, fish pond species were
found in several different fish ponds, but never in dragonfly ponds. Some
species were consistently found in two of these pond types. For example,
B. americanus was found both in vernal ponds and in fish ponds
throughout my sampling area in New Hampshire. Pseudacris crucifer
were found in both dragonfly and fish ponds that I sampled in both New
Hampshire and Florida. Species such as these that were found in two different
pond types were labeled "multiple pond" species. Thus, I have four
levels of habitat type: vernal, dragonfly, fish, and multiple pond species.
For this study, individuals from each species were collected from several
different sites. In 4 years of sampling > 30 sites for the species found in
this study, I never encountered species in a pond type inconsistent with my
categorization of them, using the criterion for pond types outlined above.
Further, my classification of species is consistent with classifications used
in previously published work (e.g.,
Skelly, 1996 [P.
crucifer]; Werner and McPeek,
1994 [R. clamitans, R. catesbeiana]).
Experiments were run in the same room in which tadpoles were housed, using 40-l aquaria. Each aquarium was divided into two compartments using a fiberglass screen stiffened at each side with doweling and fastened with silicon to the insides of each aquarium approximately 10 cm from one end. Predators, when present, were placed in the smaller section and tadpoles were placed in the larger section of the tank; the screen divider allowed flow of water and water-borne chemical cues from the predator to the tadpole side, but prevented predators from preying on tadpoles. Aquaria had a 3 cm layer of washed sand as substrate and Elodea-style artificial plants as places of refuge for the tadpoles. Tanks (along with sand and artificial plants) were randomly assigned to predator treatments at the start of experiments and then dedicated to that predator treatment (to guard against the transfer of predator chemical cues). For filming, tanks were set up on four metal shelving units with each shelf holding two aquaria. Eight tanks were set up simultaneously to allow filming of two complete replicates at a time.
For each trial, aquaria were randomly assigned to shelves. Aquaria were filled with approximately 22 l of water. I divided the tadpole side of the aquaria into four quadrates and placed two stems of artificial plants haphazardly in two diagonally opposed quadrates. Plants and sand were well rinsed with water and shelves and tank positions were re-assigned to treatments every second trial. I placed tadpoles and predators in the aquaria 18 to 24 h prior to recording tadpole behavior to allow tadpoles to acclimate to aquaria. I also added a small piece of pressed algae disc (Wardley® Spirulina discs) as food for tadpoles. Predators were given no food for the duration of the trial.
Three tadpoles were placed in each aquarium; this mimics natural densities
for most tadpoles (Morin,
1983). However, for some species I was unable to collect enough
tadpoles at the correct stage to do this; thus, some species were tested using
only two individuals per tank (A. gryllus, R. catesbeiana, and some
replicates of H. gratiosa) or only one individual per tank (R.
sphenocephala). I checked for an effect of number of individuals per tank
on activity behavior using H. gratiosa (because trials were done in
this species with both two and three individuals per tank) and found no effect
(F1,8 = 0.18, p >.68). Nonetheless, tadpoles
respond to the presence of conspecifics and thus R. sphenocephala
were tested under somewhat different conditions than all other species; this
should be borne in mind when interpreting results.
A blind shielded all tanks from disturbance during the acclimatization period. Behavior was filmed using a video camera and video cassette recorder. Two tanks were filmed simultaneously. Each tank was recorded for a 15 min period and then the camera was moved to the next position (shelves were assigned to treatment at random, and then filmed in order). In 1996 at Dartmouth, the light cycle of tadpoles was switched so that for each species half the trials were done in the morning and half were done in the evening. Analysis of these results revealed no effect of time of day and so subsequently (1997 and 1998) the light cycle was kept constant. All trials were started approximately 2 h after the lights switched on and took approximately 75 min to complete.
Fish predators were collected from local lakes by seining, housed in 40-l aquaria in the lab, and fed a combination of invertebrate prey, earthworms and TetraBits® (Tetra, Blacksburg, VA, USA). The pumpkinseed sunfish, Lepomis gibbosus, was used in New Hampshire experiments and the spotted sunfish, Lepomis punctatus in the Florida experiments. Notophthalmus (Notophthalmus viridescens viridescens in New Hampshire and N. v. louisianensis in Florida) were also collected from local populations by dipnet, housed in 20-l aquaria and fed invertebrate prey collected from local ponds. Anax junius larvae in the final or penultimate stage were collected by dipnet, kept individually in small plastic tubs (9 x 9 x 9 cm) and fed invertebrate prey or small fish. Anax were collected throughout the season to replace individuals as they matured. No individual predator was used more than twice with the same anuran species; in the majority of cases predators were used only once with each species.
All experimental protocols were approved by the Animal Care and Use Committees of Dartmouth College (protocol 98-03-10) and Florida State University (protocol 9701).
Statistical analyses
Time spent active was quantified using a simple event recording program I
wrote to keep track of the length of time active by one, two, or three
tadpoles. A tadpole was considered active whenever it was moving through the
water column in any direction. Tadpoles grazing, but not actually moving any
distance were not considered active, because identifying whether or not a
tadpole remaining in one spot was actually grazing was difficult. Previous
research in both Bufo and Hyla species has found that
swimming is positively correlated with growth rate
(Skelly, 1996;
Skelly and Werner, 1990).
Mean time active (s) per tank (out of a possible 900 s) was used as the response variable in analyses after transformation with the natural logarithm to make the data normally distributed. I looked for differences in mean time active using a three-way ANOVA with taxonomic family, habitat, and predator treatments as fixed effects and species nested within family by habitat as a random effect. Thus, the appropriate error term for habitat and family effects was species (family by habitat).
Evolutionary contrasts
Species' trait values are influenced by shared common ancestry and thus
species can not be considered independent data points
(Felsenstein, 1985). Many
comparative methods have now been developed to statistically control for this
non-independence (reviewed in Felsenstein,
1988; Gittleman and Luh,
1992; Harvey and Pagel,
1991; Huey, 1987).
Felsenstein's independent contrast method has been repeatedly found the most
unbiased and robust of these methods
(Díaz-Uriarte and Garland,
1996,
1998;
Gittleman and Luh, 1992;
Martins and Garland, 1991).
Felsenstein's independent contrasts transform species' trait means into n-1
independent contrast values (Felsenstein,
1985). Further, these contrast values, when standardized by branch
length, actually estimate evolutionary rates of the traits in question
(Martins, 1994;
McPeek, 1995). I therefore
used these contrast values to examine correlations in the evolutionary change
of activity level in the presence of different predators (i.e., I considered
time active under each treatment level a trait, giving four trait values per
species).
Felsenstein's independent contrast method depends on knowledge of the
branch lengths in the phylogeny and requires that all branch lengths be
measured on the same scale (Martins,
1996). The phylogenetic tree I use for anurans
(Figure 1) was put together
from three different studies (one for each family clade), each of which used a
different molecular technique (Graybeal,
1997; Hedges,
1986; Hillis and Davis,
1986). Thus, the branch lengths for each clade are clearly not
based on the same scale (i.e., we can not assume that proteins and DNA evolve
at the same rate). Therefore, I could not use branch lengths from these
published phylogenies in a combined comparative analysis. Although it is
possible to assign branch lengths arbitrarily
(Garland et al., 1992;
Grafen, 1989), this makes
assumptions about character evolution that may or may not be true. Instead I
chose to test correlations in evolutionary change using a randomization
procedure that would be robust to branch length differences, allowing me to
determine which correlations are present independent of relative branch
lengths (Losos, 1994;
Martins, 1996). In this way,
my tests make no assumptions about relative branch lengths. I did this by
calculating contrasts on trees with branch lengths assigned random values
between zero and one. I calculated contrast values using this tree with
randomly assigned branch lengths for two traits and then calculated the
correlation coefficient between the contrasts of the two traits. This
procedure was repeated 2000 times for each pair of trait values and a
distribution of contrast correlations was generated. A confidence interval for
each correlation was determined by ranking the resultant correlation
coefficients and finding the two values that encompassed 95% of the values
(Manly, 1991).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Family and habitat effects
I collected data for this study from two potentially different conditions: in Florida and in New Hampshire. Comparative studies depend on including as many species as possible (Pagel and Harvey, 1988
) and thus the benefits of collecting data on more
species in Florida outweighed the inherent problems of combining data from two
different locales. To ensure that data were not systematically biased by the
locale in which they were collected, I tested for a locale effect using
species nested within locale as the error term (since species differed
significantly). This ANOVA found no significant effect of locale on time
active (F1,11 = 4.12, p >.067). The power of
this test to detect an effect of the size seen (mean time active differed by
138 s between groups) is relatively low (
0.50), but a thorough
inspection of the data revealed no systematic effects of locale that might
affect results described here. No interaction was present (using species
nested within the interaction term as the denominator) between locale and
predator (F3,44 = 0.30, p >.82), locale and
family (F2,7 = 0.36, p >.71), or locale and
habitat (F2,6 = 0.05, p >.95). Also, the main
effects of family, habitat and predator when tested on each locale separately
were very similar to presented results of combined data (interaction between
family and habitat could not be tested in these analyses because of too many
missing cells in the data). Therefore, in all results below I combine species
from both locales. Species nested within habitat by family treatments differed strongly in time spent active (Table 1, Figure 3). Time spent active also varied significantly by family, and by predator, and a significant interaction between family and predator was present (Table 1). The effect of habitat was not statistically significant (Table 1). Because the full design has several missing cells (not all families have species in all habitats; Figure 2), I also tested a fully crossed subset of the data to confirm results from the unbalanced dataset. Results from an ANOVA of this subset of data (ranids and hylids in dragonfly, fish, and multiple ponds) were very similar to those from the whole dataset.
|
|
The presence of strong species differences within a family by habitat combination makes interpretation of the dataset more challenging. Although the habitat effect was not significant in the above analysis, this tells us only that on average species' responses tend to be similar across habitats. This could mean that every species in a family has a similar trait value, no matter which habitat they use. However, we would also get this result if species in one habitat type had low and high trait values, while species in another habitat type had medium trait values. The strong species within habitat by family effect made the latter result seem highly likely for these data. Therefore, to interpret the data better, I ran analyses on differences between species.
First, I looked for species' differences in time spent active when no predator is present. I used activity when no predator is present as "baseline" activity because species are expected to maximize activity levels under this condition, and it provides a controlled environment that allows species' comparisons. A one-factor ANOVA confirmed strong species differences in time spent active (F12,68 = 11.93, p <.0001). Tukey tests found significant differences mostly between the least active species, A. gryllus, and the most active species, R. sylvatica, and other species (Table 2). Looking at both the Tukey test results (Table 2) and Figure 3 suggests that A. gryllus, P. crucifer, (multiple pond species) and H. squirella (vernal pond species) can be grouped as the least active, while R. clamitans and R. sylvatica (dragonfly pond species) can be grouped as the most active species.
|
To consider how species differ in behavioral response to predators, I
analyzed predator effects within each species (each species is a separate
experiment in my study). These analyses found significant changes in time
active with changing predators for the vernal pond species H. squirella
(F3,13 = 4.06, p <.040) and the dragonfly pond
species R. sylvatica (F3,28 = 4.35, p <.012),
and a nearly significant effect for the dragonfly pond species H.
versicolor (F3,8 = 3.67, p <.063). No other
species revealed significant predator effects
(Figure 4), but across species
the presence of predators significantly reduced time active
(Table 1; Tukey's test revealed
the no predator and Notophthalmus treatments differed from the
Lepomis treatment). Note that failure to detect more within species
differences in activity in the presence of predators may be due to low power.
Using the error MS for each species to determine the power of this design to
detect a predator effect of 100 s (= 11% of the time observed), I found that
power ranged from approximately 10% for R. catesbeiana to nearly 100%
for H. versicolor (Zar,
1984). Power to reject the null hypothesis under this alternative
hypothesis was also high (> 98%) for B. terrestris, R. clamitans,
and P. crucifer. Power ranged from 0.25 to 0.45 for B.
americanus, H. cinerea, H. femoralis, H. gratiosa, and R.
sphenocephala.
|
Evolutionary contrasts
Contrast values for time spent active in the different predator treatments
revealed strongly significant positive correlations
(Table 3). This was true for
all possible pairings of the four sets of contrast values. Time active with
Lepomis and time active in each of the other treatments showed the
strongest relationships. Time active with Notophthalmus and time
active when no predator was present were relatively weakly correlated, with a
much wider confidence interval than the other contrast correlations
(Table 3).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This study confirms the expected presence of strong species differences in activity levels, and that tadpoles modify activity levels in the presence of a predator. However, phylogenetic analyses revealed tight positive correlations in the evolution of activity levels under different predator conditions. Thus, species that evolve to be more active when alone also evolve to be more active in the presence of predators. This cross-environment correlation may represent a constraint in the degree of plasticity that can evolve in tadpoles. Evolution of higher activity levels when no predator is present (beneficial because foraging rates increase) will lead to concurrent evolution of higher activity levels in the presence of predators (detrimental because predation risk goes up). Likewise, the degree to which tadpoles can reduce activity in the presence of predators would be limited because some minimum level of activity when no predator is present is necessary. Previous research has found that baseline activity level (activity level when no predator is present) is the best predictor of tadpole predation risk (Lawler, 1989
). That evolution
of activity rates were found to be correlated in this study supports this
hypothesis.
This potential constraint in evolution of plasticity has not prevented
species from acquiring some plasticity in behavioral responses to predators.
Tadpoles, in general, decreased the amount of time spent active in the
presence of predators, especially Lepomis. This was true even for
those species found only in fishless ponds. This may reflect the presence of
ancestral traits that selection has not acted against. Or, tadpoles may
respond to a general predator cue that is common to fish, newt and insect
predators. Decreased activity in the presence of predators implies that those
species which frequently encounter predators will suffer reduced growth rates
via reduced foraging rates, which are a consequence of decreased activity
levels (Anholt and Werner,
1995; Werner and Anholt,
1993).
Variance among species in activity levels was large relative to variance between habitats. This implies that even within a lineage, species can evolve to have different trait values within the same selective environment. However, a significant family by predator effect implies some phylogenetic inertia in trait values. It suggests that change in activity level when predators are present differs among lineages. This may reflect the fact that some hylid species greatly increased activity levels in the presence of some predators, while ranid and bufonid species did not. Change in activity with predators was significant for the vernal pond species H. squirella, as expected given the more variable nature of predator presence in vernal ponds. However, with no predators present, activity levels in H. squirella were significantly lower than species from other pond types. This was a surprising result because vernal pond species must develop quickly and activity levels are generally correlated with growth rates. Other data confirm that H. squirella do not have an exceptionally high growth rate, but rather develop and metamorphose at a small size (Richardson, unpublished data).
As expected, ranids in dragonfly ponds (R. clamitans and R. sylvatica) were more active than most species from fish or multiple pond types. However, hylids in dragonfly ponds were not significantly more active than either fish pond species, or bufonids and ranids in multiple ponds. Similarly, hylids in multiple ponds (P. crucifer and A. gryllus) were less active than other species, as expected, but bufonids and the ranid using multiple ponds were not less active than fish or dragonfly pond species. It appears that ranids in dragonfly ponds and hylids in multiple ponds occupy the predicted adaptive peak, while other species in those pond types have perhaps found alternative adaptive peaks.
The relative activity levels of species in this study were qualitatively
similar to previously published interspecific observations (reviewed in
Skelly, 1997). This finding
suggests that activity level is a repeatable species-specific trait. For
example, a variety of studies have now revealed P. crucifer to have
extremely low activity levels whether a predator is present or not
(Lawler, 1989;
Skelly, 1995; this study). In
general, tadpoles of Bufo were among the most active in this study,
as has been observed in European anurans
(Laurila et al., 1997).
Bufo in this study also responded weakly to different predator
treatments. Some bufonids form cohesive swarms as an anti-predator response
(Watt et al., 1997) and thus
Bufo appears to use an anti-predator mechanism other than reducing
activity level (but see Anholt et al.,
1996).
Response to predators varied from predictions in some species. Not all
species reduced activity in the presence of predators. In the present study,
H. squirella and R. catesbeiana both increased activity
levels in the presence of a newt predator. This is counter to the dogma of
prey reducing activity in the presence of predators, but increased activity in
the presence of a predator has also been observed in some invertebrate prey;
this response could reflect refuge-seeking activity
(Lima, 1998;
Lima and Dill, 1990). Prey
that can escape predators through swimming (as seems likely for tadpoles with
Notophthalmus and Anax predators) will be induced to swim
whenever disturbed by a predator. Thus, if predators are active, prey
disturbance will increase, and prey activity levels will increase. This does
not require physical contact between predator and prey; the movement of water
created by moving predators can induce tadpoles to swim away from the
disturbance (personal observation).
One difficulty inherent in this study is the need to use a "common
garden" experiment. Species can only be compared when data are collected
under controlled conditions. Although I tried to ensure that conditions
matched the natural situation, there is no way to confirm if behavior in the
lab is identical to behavior in the field. Any measures of behavior in the
field are necessarily confounded by difference in the habitat type in which
species are found. Furthermore, activity levels cannot be measured for many
species in the field; the only examples are of species found in shallow, clear
rock pools (Smith and Van Buskirk,
1995). Species found in deeper and murkier waters cannot be
observed this way. In addition, these sorts of observations are confounded by
observer presence, something that can be avoided in the lab by using a video
camera. For the purposes of this study, it is not vital that measures of
activity rates are quantitatively precise because we are interested in
activity rates of species relative to one another. Thus, we must assume only
that any lab effect on behavior is similar for all species considered.
This study also did not control for early developmental environment of
individuals. Phenotypic plasticity can occur in traits that are fixed during
development, as well as in labile traits, such as behavior, that change as
quickly as the environment (Scheiner,
1993). Phenotypic plasticity in fixed morphological traits has
been documented in larvae of several anuran species
(McCollum and Van Buskirk,
1996; Smith and Van Buskirk,
1995; Van Buskirk et al.,
1997; Van Buskirk and Relyea,
1998). Some aspects of behavior may also be fixed through
differences in the environment experienced during development. Such responses
can lead to inter-population differences within a species. Because I could not
control for this variability in developmental environment, I instead
randomized for this effect. For each species I collected individuals from two
or more disjunct (> 10 miles apart) populations. This reduced the chance
that data collected might reflect an aberrant population mean instead of
characterizing the species' mean. Note that ignoring this within species
variation has likely increased "noise" in my data, leading to
estimates of within species variance that will be higher than the true
variance. The presence of species differences that are evident despite this
uncontrolled variation within species reveals that differences between species
surpasses intra-specific population differences.
In summary, this study finds that tadpoles in general respond to the
presence of predators (including those they would not typically encounter) by
reducing activity levels. The study further reveals that despite its lability,
activity behavior displays a consistent pattern among species and thus is a
suitable metric for characterizing species' differences. Finally, phylogenetic
analysis in this study reveals that, at least for tadpoles of the species
considered, behavioral response to the environment may be evolutionarily
constrained. This suggests that behavioral traits may not always be more
easily and quickly modified during evolution than morphological or
physiological traits, as has been typically assumed
(Gittleman et al., 1996;
Wcislo, 1989).
|
ACKNOWLEDGEMENTS |
|---|
Thanks to the many folks who provided help in the field: Charlie Baer, Steve Fradkin, Nick Friedenberg, Howard Horne, Lisa Horth, Mark McPeek, and Tom Waltzcek. And thanks to all who read and commented on this manuscript: Brad Anholt, Doug Bolger, Carol Folt, Richard Holmes, Jeff Leips, Mark McPeek, Joe Travis, David Westneat and two anonymous reviewers. This research was supported by an Animal Behavior Society Research award to J.M.L.R., an NSF dissertation improvement grant (IBN-97008777) to Mark A. McPeek and J.M.L.R. and by NSF grants (DEB-9419318, IBN-97007787) to Mark A. McPeek.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abrams PA, 1991. Strengths of indirect effects generated by optimal foraging. Oikos 62: 167-176.
Anholt BR, Skelly DK, Werner EE, 1996. Factors modifying antipredator behavior in larval toads. Herpetologica 52: 301-313.
Anholt BR, Skelly DK, Werner EE, 2000. Changes in larval activity of four species of ranid frogs in response to food availability and predator density. Ecology, in press.
Anholt BR, Werner EE, 1995. Interaction between food availability and predation mortality mediated by adaptive behavior. Ecology 76: 2230-2334.
Brooks JL, Dodson SI, 1965. Predation, body size, and
composition of plankton. Science 150:
28-35.
Collins JP, Wilbur HM, 1979. Breeding habits and habitats of the amphibians of the Edwin S. George Reserve, Michigan, with notes on the local distributions of fishes. Occ Pap Mus Zool Univ Mich 686: 1-34.
Díaz-Uriarte R, Garland T Jr, 1996. Testing hypotheses of correlated evolution using phylogenetically independent contrasts: sensitivity to deviations from Brownian motion. Syst Biol 45: 27-47.
Díaz-Uriarte R, Garland T Jr, 1998. Effects of branch length errors on the performance of phylogenetically independent contrasts. Syst Biol 47: 654-672.[ISI][Medline]
Falconer DS, 1952. The problem of environment and selection. Am Nat 86: 293-298.[ISI]
Felsenstein J, 1985. Phylogenies and the comparative method. Am Nat 125: 1-15.
Felsenstein J, 1988. Phylogenies and quantitative characters. Annu Rev Ecol Syst 19: 445-471.[ISI]
Garland T Jr, Harvey PH, Ives AR, 1992. Procedures for the analysis of comparative data using phylogenetically independent contrasts. Syst Biol 41: 18-32.
Gittleman JL, Luh, H-K, 1992. On comparing comparative methods. Annu Rev Ecol Syst 23: 383-404.[ISI]
Gittleman JL, Anderson CG, Kot M, Luh H-K, 1996. Phylogenetic lability and rates of evolution: a comparison of behavioral, morphological and life history traits. In: Phylogenies and comparative method in animal behavior. (Martins EP, ed). Oxford: Oxford University Press; 166-205.
Gosner N, 1960. A simplified table for staging anuran embryos and larvae with notes on identification. Herpetologica 16: 183-190.
Grafen A, 1989. The phylogenetic regression. Phil Trans R Soc Lond B 326: 119-157.[ISI][Medline]
Graybeal A, 1997. Phylogenetic relationships of bufonid frogs and tests of alternate macroevolutionary hypotheses characterizing their radiation. Zool J Linn Soc 119: 297-338.
Harvey PH, Pagel MD, 1991. The comparative method in evolutionary biology. Oxford: Oxford University Press.
Hedges SB, 1986. An electrophoretic analysis of holarctic hylid frog evolution. Syst Zool 35: 1-21.
Hillis DM, Davis SK, 1986. Evolution of ribosomal DNA: fifty million years of recorded history in the frog genus Rana. Evolution 40: 1275-1288.
Huey RB, 1987. Phylogeny, history, and the comparative method. In: New directions in ecological physiology. (Feder ME, Bennett AF, Burggren WW, Huey RB, eds). Cambridge: Cambridge University Press; 76-98.
Laurila A, Kujasalo J, Ranta E, 1997. Different antipredator behaviour in two anuran tadpoles: effects of predator diet. Behav Ecol Sociobiol 40: 329-336.
Lawler SP, 1989. Behavioural responses to predators and predation risk in four species of larval anurans. Anim Behav 38: 1039-1047.
Lima SL, 1998. Stress and decision-making under the risk of predation: recent developments from behavioral, reproductive, and ecological perspectives. Adv Study of Behav 27: 215-290.
Lima SL, Dill LM, 1990. Behavioral decision made under the risk of predation: a review and prospectus. Can J Zool 68: 619-640.[ISI]
Losos JB, 1994. An approach to the analysis of comparative data when a phylogeny is unavailable or incomplete. Syst Biol 43: 117-123.
Manly BFJ, 1991. Randomization and Monte Carlo Methods in Biology. New York: Chapman & Hall.
Martins EP, 1994. Estimating the rate of phenotypic evolution from comparative data. Am Nat 144: 193-209.
Martins EP, 1996. Conducting phylogenetic comparative studies when the phylogeny is not known. Evolution 50: 12-22.
Martins EP, Garland TG, 1991. Phylogenetic analyses of the correlated evolution of continuous characters: a simulation study. Evolution 45: 534-557.[ISI]
McCollum SA, Van Buskirk J, 1996. Costs and benefits of a predator-induced polyphenism in the gray treefrog Hyla chrysoscelis. Evolution 50: 583-593.
McPeek MA, 1990a. Determination of species composition in the Enallagma damselfly assemblages of permanent lakes. Ecology 71: 83-98.
McPeek MA, 1990b. Behavioral differences between Enallagma species (Odonata) influencing differential vulnerability to predators. Ecology 71: 1714-1726.
McPeek MA, 1995. Testing hypotheses about evolutionary change on single branches of a phylogeny using evolutionary contrasts. Am Nat 145: 686-703.
Morin PJ, 1983. Predation, competition, and the composition of larval anuran guilds. Ecol Monogr 53: 119-138.
Pagel MD, Harvey PH, 1988. Recent developments in the analysis of comparative data. Q Rev Biol 63: 413-440.[Medline]
Scheiner SM, 1993. Genetics and evolution of phenotypic plasticity. Annu Rev Ecol Syst 24: 35-68.[ISI]
Semlitsch RD, Scott DE, Pechman JHK, Gibbons JW, 1996. Structure and Dynamics of an Amphibian Community. In: Long-term studies of vertebrate communities. (Cody ML, Smallwood JA, eds). San Diego: Academic Press; 217-248.
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 Pseudacris treefrog larvae. Ecology 76: 150-164.
Skelly DK, 1996. Pond drying, predators, and the distribution of Pseudacris tadpoles. Copeia 1996: 599-605.
Skelly DK, 1997. Tadpole communities. Am Sci 85: 36-45.
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.
Van Buskirk J, McCollum SA, Werner EE, 1997. Natural selection for environmentally induced phenotypes in tadpoles. Evolution 51: 1983-1992.
Van Buskirk J, Relyea RA, 1998. Selection for phenotypic plasticity in Rana sylvatica tadpoles. Biol J Linn Soc 65: 301-328.
Van Tienderen PH, Koelewijn HP, 1994. Selection on reaction norms, genetic correlations and constraint. Genet Res Camb 64: 115-125.[ISI][Medline]
Via S, Lande R, 1985. Genotype-environment interaction and the evolution of phenotypic plasticity. Evolution 39: 505-522.[ISI]
Watt PJ, Nottingham SF, Young S, 1997. Toad tadpole aggregation behaviour: evidence for a predator avoidance function. Anim Behav 54: 865-872.[ISI][Medline]
Wcislo WT, 1989. Behavioral environments and evolutionary change. Annu Rev Ecol Syst 20: 137-69.[ISI]
Wellborn GA, Skelly DK, Werner EE, 1996. Mechanism creating community structure across a freshwater habitat gradient. Annu Rev Ecol Syst 27: 337-363.[ISI]
Werner EE, 1991. Nonlethal effects of a predator on competitive interactions between two anuran larvae. Ecology 72: 1709-1720.
Werner EE, Anholt BR, 1993. Ecological consequences of the tradeoff between growth and mortality rates mediated by foraging activity. Am Nat 142: 242-272.[ISI]
Werner EE, Anholt BR, 1996. Predator-induced behavioral indirect effects: consequences to competitive interactions in anuran larvae. Ecology 77: 157-169.
Werner EE, McPeek MA, 1994. Direct and indirect effects of predators on two anuran species along an environmental gradient. Ecology 75: 1368-1382.
Woodward BD, 1983. Predator-prey interactions and breeding-pond use of temporary-pond species in a desert anuran community. Ecology 64: 1549-1555.
Zar JH, 1984. Biostatistical Analysis, 2nd edition. Englewood Cliffs, New Jersey: Prentice-Hall.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  G. J. Tattersall, P. C. Eterovick, and D. V. de Andrade Tribute to R. G. Boutilier: Skin colour and body temperature changes in basking Bokermannohyla alvarengai (Bokermann 1956) J. Exp. Biol., April 1, 2006; 209(7): 1185 - 1196. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||