Behavioral Ecology Vol. 11 No. 2: 169-177
© 2000 International Society for Behavioral Ecology
Male green frogs lower the pitch of acoustic signals in defense of territories: a possible dishonest signal of size?
a Division of Biological Sciences, University of Missouri, Columbia, MO 65211, USA b Department of Biological Sciences, Butler University, Indianapolis, IN 46208, USA
Address correspondence to M. A. Bee, 105 Tucker Hall, University of Missouri-Columbia, Columbia, MO 65211-7400, USA. E-mail: c669112{at}showme.missouri.edu .
Received 10 January 1999; revised 18 July 1999; accepted 7 August 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In responses to broadcasts of conspecific advertisement calls, male green frogs (Rana clamitans) lower the dominant frequency of their calls. Because dominant frequency is negatively correlated with male body size in green frogs, frequency alteration provides a means of potentially exaggerating size during territorial contests. In field playback experiments, we broadcast synthetic stimuli representing small, medium, or large intruders to territorial residents. We tested the hypotheses that males use frequency alteration to provide honest signals of their size or their size-independent fighting ability, or to dishonestly signal size. Dominant frequency did not better predict male size in response calls than in unsolicited calls. The magnitude of frequency alteration was not related to body size, general condition, or an indirect measure of fighting ability. Thus, males did not use frequency alteration to provide honest information about body size or size-independent fighting ability. However, males significantly increased their apparent size by producing lower frequency calls. Small males produced relatively lower frequency calls in response to the large-male stimulus (compared to the small-male and medium-male stimuli), but large males did not. Further, the magnitudes of frequency alteration were significantly greater in responses to the large-male stimulus, primarily because small males responded with a greater decrease in frequency to the large-male stimulus than to the small-male and medium-male stimuli. These results support several predictions of the dishonest signal hypothesis and suggest that dishonesty may be a conditional strategy used by small males.
Key words: bluffing, dishonesty, frequency alteration, green frogs, Rana clamitans, territoriality.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Differences in body size and weaponry are important sources of asymmetry in fighting ability or resource holding potential (RHP) during animal contests (Parker, 1974
; reviewed in
Andersson, 1994;
Archer, 1988;
Reichert, 1998). Animals often
evaluate these asymmetries by assessing communication signals that convey
information about an opponent's fighting ability. In anuran amphibians,
difference in body size influences the outcome of male-male contests; large
males usually defeat smaller males (e.g.,
Emlen, 1976;
Given, 1988;
Howard, 1978;
Wagner, 1989;
Wells, 1978). The fundamental
frequency of anuran acoustic signals is partially determined by the shape and
mass of the laryngeal apparatus, which is in turn related to body size
(Duellman and Trueb, 1986;
Martin, 1972). Spectral call
properties and body size are negatively correlated in many species
(Gerhardt, 1994). Not
surprisingly, many male anurans rely on spectral properties of acoustic
signals to assess an opponent's size during aggressive encounters
(Arak, 1983;
Bee et al., 1999;
Davies and Halliday, 1978;
Given, 1987;
Ramer et al., 1983;
Robertson, 1986;
Wagner, 1989).
There is, however, increasing empirical evidence that male frogs alter
spectral properties of their acoustic signals during interactions with other
males (Bee and Perrill, 1996;
Bee et al., 1999;
Given, 1999;
Howard and Young, 1998;
Lopez et al., 1988; Wagner,
1989,
1992). Male green frogs,
Rana clamitans (Anura, Ranidae), significantly decrease the dominant
frequency of their calls during aggressive territorial encounters
(Bee and Perrill, 1996;
Bee et al., 1999). Furthermore,
males differentially alter their behavior according to the perceived size of
simulated intruders based on differences in the frequency spectra of acoustic
signals, suggesting that males obtain information about an intruder's body
size from the frequency of its calls (Bee
et al., 1999; Ramer et al.,
1983). Our goal in this study was to test three hypotheses offered
by Wagner (1992) to explain
the role of frequency alteration in anuran communication.
The "signal of size" hypothesis proposes that males lower the
frequency of their calls to provide honest signals of their size to opponents
(Wagner, 1992). Lower
frequency response calls may better predict male size than unsolicited calls
in either of two ways. First, the lower frequency of calls produced during
aggressive encounters may be more highly correlated with body size than is the
frequency of unsolicited calls. Hence, the predictability of size based on
dominant frequency should increase in altered response calls during agonistic
interactions. Second, males may encode information about their body size in
the magnitude of frequency alteration. Thus, a relationship between size and
frequency change is predicted.
The "signal of size-independent fighting ability" hypothesis
states that males use the magnitude of frequency alteration to indicate their
true fighting ability, determined by experience, motivation, or physiological
condition, and not by size (Wagner,
1992). Assuming that males of greater fighting ability are more
likely to escalate encounters to physical fighting, this hypothesis predicts
that frequency alteration should be related to fighting propensity or the
probability that a male will attack an opponent, independent of male body
size. Furthermore, to the extent that fighting ability is a function of a
male's general body condition, this hypothesis predicts that condition and
frequency alteration should be related.
Because there is a strong negative relationship between call frequency and
body size in green frogs (R2 =.70;
Bee et al., 1999), the
"dishonest signal of size" hypothesis states that frequency
alteration potentially functions as a means of producing exaggerated signals
of size. A prediction of this hypothesis is that a male's predicted size
should be larger based on an assessment of altered frequency response calls
than unsolicited calls. While a nonsignificant result would refute this
hypothesis, a significant result does not exclude other possibilities because
frequency decreases for any reason will result in an increase in apparent size
if frequency and size are negatively related, as they are in green frogs.
There are, however, additional predictions that can be derived from the
dishonest signal hypothesis. One model of deceitful communication predicts
that males with the lowest fighting ability should use the greatest
"exaggeration factor" in attempts to inflate their apparent RHP
(Bond, 1989). Thus, if
frequency alteration is a means of communicating false information about body
size, we predict that small male green frogs should lower their frequency more
often or by a larger magnitude than large males. According to Wagner
(1992), if lower frequency
responses function as bluffs of size, males should decrease the frequency of
their calls most when confronted by larger or similar-sized opponents because
these pose the greatest threats. Moreover, the predictability of body size
from dominant frequency is expected to decrease if frequency alteration
functions to exaggerate size (Wagner,
1992). Together, the predictions of Bond
(1989) and Wagner
(1992) suggest that frequency
alteration is context-dependent: exaggerated signals should be used
disproportionately by smaller males (Bond,
1989) and when males are confronted by large opponents
(Wagner, 1992).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Study organism
The green frog mating system is resource defense polygyny. Males defend territories in the breeding pond that females use as oviposition sites, and some males fail to obtain mates, while others mate several times during a single breeding season (Wells, 1977
,
1978). The vocal repertoire of
male green frogs has been described in detail elsewhere
(Bee and Perrill, 1996,
Wells, 1978). The most common
call is the type I advertisement call, which is a short, broad-band signal
produced several times per minute as a single-note or multiple-note call. The
dominant frequency, which we define as the harmonic of greatest relative
amplitude, corresponds to the second harmonic. The mean dominant frequency and
mean note duration of type I advertisement calls in our study population are
approximately 400 Hz and 160 ms, respectively
(Bee and Perrill, 1996). Males
also occasionally produce a distinctly different type II high intensity
advertisement call (Wells,
1978). We excluded type II calls from our analyses because they
rarely occurred during playback trials, and their function in communication
remains ambiguous (Bee and Perrill,
1996). Males use type III encounter calls in close range territory
defense against intruding males (Bee and
Perrill, 1996; Bee et al.,
1999; Wells,
1978). Type III encounter calls are spectrally and temporally
similar to type I advertisement calls but can be distinguished from type I
advertisement calls by their relatively lower frequency, longer duration, and
lower sound pressure levels (SPL) (see sonograms in
Bee and Perrill, 1996).
Playback experiments
During June and July of 1995 and 1996, we conducted two field playback
experiments in four ponds located in Eagle Creek Park in Indianapolis,
Indiana, USA. Each experiment consisted of three consecutive 4-min test
periods. Test period 1 was always a control period during which we recorded a
male's unsolicited vocalizations. These calls were nearly always type I
advertisement calls, and they were never type III encounter calls. During test
periods 2 and 3 we broadcast two synthetic type I advertisement calls that
differed only in fundamental frequency. We modeled the synthetic stimuli after
natural type I advertisement calls recorded in our study population. Each
stimulus was a single note (160 ms duration) that consisted of the first nine
harmonics. The relative amplitudes of the harmonics were the same for each
stimulus (see Bee et al., 1999;
Figure 1D). Additional details
regarding the synthesis of stimulus calls are provided in Bee et al.
(1999). In experiment 1
(N = 30), the two test stimuli had dominant frequencies of 450 Hz and
350 Hz; in experiment 2 (N = 30), the dominant frequencies of the
stimuli were 400 Hz and 350 Hz. The sequence of stimulus presentation in test
periods 2 and 3 was balanced between subjects to minimize any effects of
stimulus presentation order on responses. No such effects were found.
|
The 450, 400, and 350 Hz stimuli correspond to the natural calls of males with approximate SVLs of 66.6 (small), 76.2 (medium), and 85.7 mm (large), respectively (population mean ± SD = 75.2 ± 7.7 mm, range = 56-97 mm, N = 83). Individual males were not tested in more than one experiment per night, and only two males were tested in both experiments. After each playback test, we determined the subject's snout-to-vent length (SVL) (to the nearest 1 mm) and mass (to the nearest 1 g). One small male in experiment 2 escaped before we could measure its SVL. The mean (± SD, here and throughout) SVLs of males tested in experiment 1 (74.0 ± 8.0 mm) and experiment 2 (77.9 ± 8.0 mm) were not significantly different (t57 = 1.87, p =.07). Playback tests were conducted under ambient light levels between 2000 and 0330 h, Eastern Standard Time.
We used a Marantz PMD430 recorder to broadcast stimuli through a Sony
APM-090 amplified speaker, which floated on a styrofoam platform at a distance
of 1 m from the subject. The speaker was always positioned before the
beginning of the control period. We broadcast stimuli at a rate of 1/15 s at
84 dB (SPL re 20 µPa, "fast" RMS) at 1 m, which corresponds to
the SPL of natural calls at that distance
(Bee and Perrill, 1996). We
calibrated playback SPLs with a Brüel and
Kjær Precision Integrating Sound Level Meter Type 2230. During the three
consecutive 4-min test periods, we continually recorded a male's vocalizations
onto a second Marantz PMD430 cassette recorder using an Audio-technica AT815
condenser microphone.
We determined the value of the dominant frequency for each type I
advertisement and type III encounter call recorded from each male during each
test period. We limited our analyses to the first note of multiple-note calls
because the frequency of additional notes progressively decreases during these
calls. In an earlier study (Bee and
Perrill, 1996), males did not respond differently to natural
single-note and multiple-note type I advertisement calls. Hence, we assume
that differences in note number are not important during territorial contests.
Dominant frequencies were determined from spectrograms generated with
SoundEdit 2.0 (for calls recorded in 1995; ±10 Hz) or a Kay DSP
Sona-Graph Model 5500 (for calls recorded in 1996; ±5 Hz). Spectrograms
were generated by averaging the frequency spectrum over a thin time section
(approximately 50 ms) during the first half of the call. We used the Kay
Sona-Graph to reanalyze a subset of the calls originally analyzed with the
SoundEdit software to confirm that the two methods yielded similar
results.
During playback tests, some males called and jumped toward or around the
speaker and approached to within 5 cm of the speaker face, while other males
remained stationary and called from their original calling positions. We
counted the number of moves a male made during the playback of each stimulus
as an indicator of its propensity to fight for all 30 males in experiment 1
and for 17 of 30 males in experiment 2. We have observed similar aggressive
movements that were associated with territorial disputes and wrestling bouts
between male green frogs (see also Wells,
1978). We assume that these movements reflect a male's propensity
to fight and that fighting propensity is related to true fighting ability.
Statistical analysis
For each male, we determined the mean dominant frequency of all calls
(excluding type II calls) produced within each test period. The magnitude of
frequency alteration to a stimulus was determined by subtracting the mean
dominant frequency of a male's control calls from the mean dominant frequency
of its responses to that stimulus. Hence, most values are negative, and a more
negative value indicates a greater decrease. The results of preliminary
analyses using the magnitude of frequency alteration expressed as a percentage
of a male's mean dominant frequency were qualitatively similar to the results
reported below. We estimated male condition by dividing the residuals of a
regression of the cube root of mass on SVL by SVL to obtain a condition index
(after Baker, 1992; see also
Howard et al., 1997;
Howard and Young, 1998). The
distributions of the number of moves made during each test period of each
experiment departed from normality and were subsequently square-root
transformed [Y' = (Y + 0.5)
;
Zar, 1984]. After this
transformation, the number of moves made during the 450 Hz test period of
experiment 1 still departed significantly from normality (one-sample
Kolmogorov-Smirnov test: d = 0.25, p <.05), although much
less severely than without the transformation (all other p >.05).
Because standard parametric procedures are robust to deviations from normality
at reasonable sample sizes (Lindman,
1974), we analyzed the data using these techniques (
=.05
for all analyses). For multiple comparisons involving the same variable (e.g.,
SVL), we applied a sequential Bonferroni correction to control the type I
error rate (Rice, 1989).
Additional details of statistical analyses are provided in the Results.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Signal of size hypothesis
We tested the hypothesis that the predictability of SVL increased during responses to our stimuli by comparing the variance associated with the residuals from regressions of SVL on mean dominant frequency for responses to each stimulus to that based on control calls (Zar, 1984
). If frequency
alteration functions to increase the predictability of size, then the variance
of the residuals from the regression of SVL on dominant frequency should be
lower for response calls than for unsolicited control calls. We used a
two-tailed test, however, because the dishonest signal hypothesis predicts an
increase in residual variance for responses
(Wagner, 1992).
In experiment 1 there were significant negative relationships between SVL
and dominant frequency for calls produced during the control
(R2 =.78, p <.01), 450 Hz stimulus
(R2 =.77, p <.01), and 350 Hz stimulus
(R2 =.81, p <.01) periods
(Figure 1). As a predictor of
body size, dominant frequency explained 78%, 77%, and 81% of the variation in
SVL during the control, 450 Hz, and 350 Hz periods, respectively. The variance
of the residuals from responses was not significantly different from that of
control calls (450 Hz versus control: 
2 = 29.79, df = 28,
p >.50; 350 Hz versus control: 2 = 25.01, df = 28,
p >.25). Similarly, in experiment 2 there were significant
negative relationships between SVL and dominant frequency during the control
(R2 =.57, p <.01), 400 Hz stimulus
(R2 =.45, p <.01), and 350 Hz stimulus
(R2 =.42, p <.01) periods, and dominant
frequency explained 57%, 45%, and 42% of the variation in SVL in these three
periods, respectively (Figure
1). As in experiment 1, the variance of the residuals of responses
did not differ significantly from control calls (400 Hz versus control:
2 = 34.43, df = 27, p >.75; 350 Hz versus control:
2 = 36.48, df = 27, p >.75).
Because males could provide accurate information about their size in the
actual magnitude of frequency alteration, we examined the regression of SVL on
the magnitude of dominant frequency change in responses to each stimulus in
both experiments (Figure 2A). There were no significant relationships in responses to the 450 Hz
(R2 <.01, p =.74) and 350 Hz
(R2 =.03, p =.36) stimuli in experiment 1 or the
400 Hz (R2 =.03, p =.38) and 350 Hz
(R2 =.15, p =.04, critical =.025) stimuli
in experiment 2.
|
Signal of size-independent fighting ability hypothesis
We used multiple regression to examine the size-independent relationship
between fighting ability and frequency alteration. The number of moves a male
made during the playback of each stimulus was regressed on both SVL and the
magnitude of frequency alteration to that stimulus. There were no significant
relationships in experiment 1 (450 Hz: R2 =.05, p
=.53; 350 Hz: R2 =.06, p =.44), but this was not
the case for experiment 2 (400 Hz: R2 =.55, p
<.01; 350 Hz: R2 =.45, p <.01). The partial
correlations (Figure 2B)
relating the magnitude of frequency alteration and the number of moves after
controlling for body size, which we take to indicate the degree to which
frequency alteration may signal size-independent fighting ability, were
nonsignificant in experiment 1 (450 Hz: t27 =.45,
p =.66; 350 Hz: t27 = -.25, p =.81) and
in experiment 2 (400 Hz: t14 =.15, p =.88; 350
Hz: t14 =.30, p =.77). The significance of the
multiple regression models for experiment 2 was due almost entirely to
significant positive relationships between SVL and the number of moves in
responses to both the 400 Hz stimulus (R2 =.55, p
<.01) and the 350 Hz stimulus (R2 =.44, p
<.01). We did not find any relationship between the magnitude of frequency
alteration and male body condition in responses to either stimulus in
experiment 1 (450 Hz: R2 <.01, p =.84; 350 Hz:
R2 <.01, p =.65) and in experiment 2 (400 Hz:
R2 =.02, p =.52; 350 Hz: R2
<.01, p =.70; Figure
2C).
Dishonest signal of size hypothesis
We compared each male's SVLs predicted from the dominant frequency of his
calls during the three test periods using repeated-measures ANOVA. This is
analogous to a test of differences in dominant frequency
(Bee et al., 1999), but the
conversion to SVL presents the analysis in the terms of the size-related
information that is potentially communicated by dominant frequency and results
in a more straightforward test of whether frequency alteration increases
apparent size. The regression of the dominant frequency of unsolicited calls
(y, in Hz) on SVL (x, in mm) for all males recorded in our
study population between 1994 and 1996 (y = -5.24x + 799.11,
R2 =.70, N = 83; see
Bee et al., 1999:
Figure 2) was used to predict
SVL. Because univariate F tests often violate additional assumptions
of repeated-measures analyses with more than two levels of the within-subjects
factor, we corrected the obtained p values using the method of
Greenhouse and Geisser (1959).
We used planned contrasts to compare predicted values of actual and apparent
SVLs (Rosenthal and Rosnow,
1985).
In experiment 1 there were significant differences in the predicted SVLs (F2,58 = 35.3, p <.01; Figure 3A). The mean apparent SVL based on responses to the 450 Hz stimulus (77.3 ± 9.0 mm) was significantly larger than the predicted actual SVL based on control calls (73.2 ± 8.6 mm; F1,29 = 29.14, p <.01). The mean apparent SVL determined from responses to the 350 Hz stimulus (78.6 ± 7.7 mm) was significantly larger than both the predicted actual SVL (F1,29 = 56.07, p <.01), and the apparent SVL based on responses to the 450 Hz stimulus (F1,29 = 6.36, p =.02). A similar pattern of significant differences between predicted SVLs was observed in experiment 2 (F2,58 = 33.24, p <.01; Figure 3B). The mean apparent SVLs based on responses to the 400 Hz stimulus (84.0 ± 7.5 mm) and the 350 Hz stimulus (85.2 ± 6.8 mm) were both significantly larger than the predicted actual SVL (80.2 ± 7.5 mm) (400 Hz: F1,29 = 25.41, p <.01; 350 Hz: F1,29 = 60.69, p <.01). In addition, the mean apparent SVL based on responses to the 350 Hz stimulus was significantly larger than that based on responses to the 400 Hz stimulus (F1,29 = 6.33, p =.02).
|
We further tested the hypothesis that frequency alteration functions as an
exaggerated signal of size by examining whether frequency alteration depended
on male body size (Bond, 1989)
or on the perceived size of a simulated intruder
(Wagner, 1992). Because there
was no significant relationship between the magnitude of frequency alteration
and SVL (Figure 2A), we asked
whether males in different size classes responded differently to the two
stimuli in each experiment. The 30 subjects in each experiment were separated
into two size classes (small and large) at the median SVL for males in that
experiment. We used binomial tests to test the hypothesis that the proportion
of males that produced lower frequency calls in response to the 350 Hz
stimulus, compared to the 450 Hz and 400 Hz stimuli, exceeded the chance
expectation of 50%. We also incorporated size class as a between-subjects
factor in a 2 (size class) x 2 (stimulus) ANOVA, with stimulus as a
within-subjects factor. The magnitude of frequency alteration was the
dependent variable. We used planned contrasts to compare the responses within
each size class. This analysis was performed to answer three questions. First,
do males of different size classes respond with different magnitudes of
frequency alteration to conspecific intruders (size class effect)
(Bond, 1989)? Second, do males
respond differently to conspecific intruders of different size classes
(stimulus effect) (Wagner,
1992)? Third, does the magnitude of frequency alteration in
response to a particular sized intruder depend on the size of the resident
male (size class x stimulus interaction)?
In experiment 1, 19 of 30 males produced relatively lower frequency calls in response to the 350 Hz stimulus (p =.07). Of these, 7 of 15 large males (SVL > 75.5 mm; p =.70) and 12 of 15 small males (SVL < 75.5 mm; p =.02) produced relatively lower frequency calls in response to the 350 Hz stimulus. The ANOVA revealed no significant effect of size class on the magnitude of frequency alteration (F1,28 =.05, p =.83). There was a significant stimulus effect (F1,28 = 7.34, p =.01), which indicates that males responded with larger decreases in frequency to playbacks of the 350 Hz (large) stimulus (Figure 4A). This analysis also revealed a significant size class x stimulus interaction (F1,28 = 5.52, p =.03; Figure 5A). Small males decreased the frequency of their calls significantly more in response to the 350 Hz stimulus (-30.6 ± 24.1 Hz) than in response to the 450 Hz stimulus (-18.3 ± 22.9 Hz; F1,28 = 12.80, p <.01). Large males, however, did not vary the magnitude of frequency change in their responses based on the perceived size of the stimulus (F1,28 =.07, p =.80). In responses to the 450 Hz and 350 Hz stimuli, large males decreased their dominant frequencies by means of -25.7 ± 21.8 Hz and -26.6 ± 17.8 Hz, respectively.
|
|
In experiment 2, 20 of 29 males produced relatively lower frequency calls in response to the 350 Hz stimulus (p =.02). Of these, 8 of 14 large males (SVL > 78.8 mm; p =.40) and 12 of 15 small males (SVL < 78.8 mm; p =.02) produced relatively lower frequency calls in response to the 350 Hz stimulus. One male was excluded from the binomial test because the mean dominant frequencies of its responses to the 400 Hz and the 350 Hz stimuli were the same. The ANOVA revealed no significant effect of size class on responses (F1,28 = 1.41, p =.25). As in experiment 1, there was a significant stimulus effect (F1,28 = 7.04, p =.01; Figure 4B). There was also a significant size class x stimulus interaction (F1,28 = 4.22, p =.049; Figure 5B). Small males decreased the dominant frequency of their response calls significantly more to the 350 Hz stimulus (-32.4 ± 18.3 Hz) than to the 400 Hz stimulus (-20.5 ± 23.0 Hz; F1,28 = 11.08, p <.01). Large males, however, did not differentially decrease their dominant frequency in responses to the 400 Hz stimulus (-17.9 ± 19.3 Hz) and the 350 Hz stimulus (-19.4 ± 16.1 Hz; F1,28 =.18, p =.68).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The results reported here suggest that male green frogs can discriminate a frequency difference of about 12-14%, which corresponds to the 50 Hz difference in dominant frequency between the 400 Hz and 350 Hz stimuli (Bee et al., 1999
). Males of
the closely related bullfrog R. catesbeiana appear to be capable of
discriminating differences in fundamental frequency as small as 5-10% in field
playback experiments (Bee MA and Gerhardt HC, unpublished data), and Wagner
(1992) demonstrated that male
cricket frogs respond differently to two stimuli differing by 5.7% in
fundamental frequency. Whether green frogs can discriminate such small
differences in frequency is not known. In responses to the 350 Hz stimulus,
13.8% of males (8 of 58) lowered the frequency of their calls more than 12%.
Unless a 12-14% difference in frequency approaches the minimum difference
required for frequency resolution by green frogs, however, we expect that a
higher proportion of males are altering their dominant frequency by a
magnitude large enough to potentially produce meaningful differences between
calls (Nelson and Marler,
1990).
Signal of size hypothesis
In experiment 1, the dominant frequency of control calls explained 78% of
the variation in SVL, and the dominant frequency of responses to the 450 Hz
and 350 Hz stimuli explained 77% and 81% of the variation in SVL,
respectively. These results suggest that SVL is predicted about equally well,
and certainly not much better, by solicited response calls compared to
unsolicited calls. Although not a significant reduction, the proportion of
variation in SVL explained by dominant frequency in experiment 2 decreased
from 57% in control calls to 45% and 42% in responses to the 400 Hz and 350 Hz
stimuli, respectively, indicating that the predictability of size based on the
dominant frequency of a male's calls may decrease during aggressive
interactions. The dominant frequency of unsolicited calls appears to be an
equally good or better predictor of SVL than that of a male's aggressive calls
(Figure 1). In addition, the
magnitude of frequency alteration was not related to SVL and therefore does
not function as a reliable indicator of body size
(Figure 2A). These two
experiments suggest that frequency alteration does not provide opponents with
more accurate information about a male's size during aggressive encounters.
Based on these results, and those from our previous study
(Bee and Perrill, 1996), we
reject the signal of size hypothesis as an explanation for frequency
alteration in communication between male green frogs.
Signal of size-independent fighting ability hypothesis
There was no relationship between the magnitude of frequency alteration and
the number of moves made after controlling for the effects of body size
(Figure 2B), and frequency
alteration and male body condition were unrelated
(Figure 2C). Thus, our results
do not support the hypothesis that frequency alteration functions as an honest
signal of a male's size-independent fighting ability. Two limitations of our
and Wagner's (1992) test of
this hypothesis, however, are the untested assumptions that (1) movement
toward a speaker accurately represents fighting propensity, and (2) fighting
propensity accurately reflects fighting ability. If these assumptions hold, we
reject the signal of size-independent fighting ability hypothesis as an
explanation for frequency alteration in green frogs. This conclusion is in
contrast to the those of Wagner's
(1992) study of frequency
alteration in cricket frogs.
Dishonest signal of size hypothesis
Maynard Smith and Parker
(1976: 169) claim that the
essential feature of a bluffed display is that "it should increase
apparent size (or whatever feature is being used to settle conflicts without
escalation) without altering RHP in an escalated contest." As a
behavioral response during simulated territorial contests, frequency
alteration by male green frogs possesses the essential feature of a
exaggerated signal of size: the production of lower frequency calls leads to
an increase in apparent size (Figure
3). Although this result is an obvious consequence of frequency
reduction when size and frequency are strongly negatively related, the
implications are nonetheless interesting. Previous studies have demonstrated
that male green frogs modify aspects of their calling behavior according to
the size of simulated intruders, as conveyed by the frequency of natural and
synthetic acoustic signals (Bee et al.,
1999; Ramer et al.,
1983). Given that larger male green frogs win territorial contests
with smaller males (Wells,
1978), and that males attend to size-related information in the
frequency spectrum of acoustic signals, frequency decreases resulting in an
increase in apparent size have the potential to deceive receivers during a
conflict. This result is particularly interesting because fundamental
frequency is commonly regarded as a good example of an honest signal of size
due to constraints imposed by the morphology of sound production
(Cheney and Seyfarth, 1991;
Krebs and Dawkins, 1984;
Wiley, 1983). It is worth
noting that, although the increases in apparent SVL resulting from frequency
alteration are typically small, some males increased their apparent SVL by
more than a centimeter. Increases of this magnitude are equivalent to the
difference between the apparent sizes of our simulated large and mediumsized
intruders, which was large enough to effectively elicit differences in
behavior (Bee et al.,
1999).
Our results provide mixed support for the prediction that, if lower
frequency calls represent exaggerated signals, frequency alteration should be
context dependent (Bond, 1989;
Wagner, 1992). Males responded
with a greater decrease in frequency to a perceived large intruder than to the
calls of a small or medium-sized male
(Figure 4), which is consistent
with Wagner's (1992)
prediction that males should exaggerate their size more in response to large
opponents. Moreover, only males in the small size class produced relatively
lower frequency calls in response to the 350 Hz stimulus significantly more
often than expected by chance, and the greatest magnitude of frequency
decrease consistently occurred in the calls of small males responding to the
large stimulus (Figure 5).
These observations are consistent with Bond's
(1989) prediction that animals
with the lowest probability of winning an escalated encounter should be more
likely to exaggerate their RHP. It is not clear, however, why large males
should lower frequency at all, much less why they did so more than small males
in response to the 450 Hz (small) stimulus in experiment 1. Our results also
fail to demonstrate the predicted decrease in the predictability of size from
dominant frequency (Wagner,
1992), although a nonsignificant trend in this direction was
evident in experiment 2. Compared to unsolicited calls, the frequency of
aggressive calls is not an unreliable indicator of body size. We note,
however, that even if the predictability of size does not change, receivers
could be deceived by lower frequency calls if they do not also modify the
decoding rules they use to predict size from frequency during agonistic
encounters. Taken together, our results suggest the possibility that frequency
alteration may be a conditional strategy differentially used by smaller males,
especially when they interact with large opponents.
We are currently unable to reject bluffing as a possible function of
frequency alteration in communication between male green frogs. We urge
caution in accepting this hypothesis, however, because dishonest signals are
generally expected to occur rarely in animal communication
(Dawkins and Krebs, 1978;
Grafen, 1990;
Wiley, 1983;
Zahavi, 1977; but see
Cheney and Seyfarth, 1991;
Dawkins and Guilford, 1991;
Johnstone, 1998;
Wiley, 1994). Not
surprisingly, there are few examples of dishonest signals of RHP that are both
intraspecific and intrasexual (but see
Adams and Caldwell, 1990;
Caldwell, 1986;
Steger and Caldwell, 1983).
Despite the expected rarity of dishonest signaling and the limited empirical
evidence of its occurrence, several models have been developed to show that
the use of dishonest signals can be stable in a population, especially if they
occur at low frequencies (e.g., Adams and
Mesterton-Gibbons, 1995; Bond,
1989; Gardner and Morris,
1989; Johnstone and Grafen,
1993). Interestingly, two of these models predict the use of
dishonest signals by members of the population that are the most vulnerable in
escalated contests (Adams and
Mesterton-Gibbons, 1995; Bond,
1989). Such may be the case for frequency alteration by small male
green frogs. Before accepting the dishonest signal hypothesis, however,
several alternative hypotheses deserve consideration.
Alternatives to dishonesty
There are at least four additional alternative explanations for frequency
alteration that have not been considered here. First, large males may reduce
their frequency by smaller magnitudes because large body size imposes physical
constraints on frequency alteration. Although this hypothesis can explain the
limited frequency alteration by large males, it fails to explain the
consistent differences by which small males lowered their frequency in
response to stimuli representing males of different sizes
(Figure 5).
Second, in their aggressive responses, male green frogs also produce
significantly longer calls at lower SPLs
(Bee and Perrill, 1996).
Frequency alteration could be a by-product of changes in note duration or
amplitude that function in male-male communication
(Martin, 1972). Note duration
and amplitude are not correlated with SVL
(Bee and Perrill, 1996), and
males did not differentially alter their behavior when a pair of synthetic
stimuli differed only in note duration (Bee MA and Perrill SA, unpublished
data). Moreover, reduced amplitude likely represents a cost to producing lower
frequency calls in terms of limiting the distance of sound propagation and may
explain why males do not always produce the lowest frequency calls they are
able to produce.
Third, frequency alteration may be related to a male's propensity to
abandon calling or retreat when facing an opponent of much greater fighting
ability, which may be determined by an internal state analogous to fear.
Although this hypothesis explains the tendency for smaller males to decrease
their frequency more to the larger stimulus, it is at odds with certain
motivation-structural rules of signal design derived from studies of birds and
mammals, which suggest that fear is associated with the production of higher,
not lower, frequency sounds (Morton,
1977). None of these alternatives appears to be a suitable
explanation for frequency alteration in green frogs.
Payne and Pagel (1996)
offer an interesting alternative explanation. According to their optimality
model of signal escalation, frequency alteration represents a costly increase
in the magnitude of the signal during agonistic encounters in which the signal
functions to facilitate assessment of an opponent's quality. A key prediction
of their model is that low-quality individuals should pay this cost early
during an encounter by immediately escalating to the greatest magnitude signal
they can produce, especially when confronted by superior opponents.
High-quality individuals can forgo some of these costs by escalating to a
greater magnitude signal only when their initial signals prove unsuccessful in
settling the conflict. When small males were confronted by our simulated large
intruders, they decreased their frequency to a greater magnitude than when
confronted by small or medium-sized males
(Figure 5). These results
accord well with predictions of Payne and Pagel's
(1996) model, in which the
smaller magnitude of frequency alteration in responses to low-quality
opponents represents a "false modesty" on the part of small males.
The difficulty in applying this model to green frogs becomes evident upon
considering the responses of large males. The smallest magnitude of frequency
alteration is predicted to occur when large males respond to small intruders.
Moreover, large males are expected to lower their frequency in responses to
other large males because these represent high-quality opponents. Our results,
however, do not support these predictions
(Figure 5).
The failure of our results to unequivocally support the predictions of
either the dishonest signal of size hypothesis or the model of Payne and Pagel
(1996) warrants further
investigation into the role of frequency alteration in communication during
agonistic encounters between male frogs. We have not investigated the ability
of frequency alteration to deter encroachment or escalation by intruders on a
male's territory, and, therefore, we do not know whether frequency alteration
might deceive opponents (but see Wagner,
1992). The interesting possibility remains that frequency
alteration represents an instance of dishonesty in signaling, either by
conveying false information about size or withholding accurate information
about general quality.
|
ACKNOWLEDGEMENTS |
|---|
We thank B. Buchanan, C. Gerhardt, M. Given, C. Murphy, J. Schwartz, W. Wagner Jr., A. Welch, and two anonymous reviewers for their helpful comments on previous versions of the manuscript. This manuscript benefited from discussions with R. Howard, W. Wagner Jr., and especially C. Gerhardt. J. Schwartz kindly provided software and technical expertise for synthesizing stimulus calls and the Indianapolis Parks Department provided generous access to the ponds in Eagle Creek Park. This research was funded by an National Science Foundation Graduate Research Fellowship to M.B., a Butler University Research Grant to S.P., and a grant from the Butler Summer Institute to P.O.
|
FOOTNOTES |
|---|
P. C. Owen is now at the Department of Ecology and Evolutionary Biology, University of Connecticut, 75 N. Eagleville Road, U-43, Storrs, CT 06296-3043, USA.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Adams ES, Caldwell RL, 1990. Deceptive communication in asymmetric fights of the stomatopod crustacean Gonodactylus bredini. Anim Behav 39: 706-716.[Web of Science]
Adams ES, Mesterton-Gibbons M, 1995. The cost of threat displays and the stability of deceptive communication. J Theor Biol 175: 405-421.
Andersson M, 1994. Sexual selection. Princeton, New Jersey: Princeton University Press.
Arak A, 1983. Sexual selection by male-male competition in natterjack toad choruses. Nature 306: 261-262.
Archer J, 1988. The behavioural biology of aggression. Cambridge: Cambridge University Press.
Baker JMR, 1992. Body condition and tail height in great crested newts, Triturus cristatus. Anim Behav 43: 157-159.
Bee MA, Perrill SA, 1996. Responses to conspecific advertisement calls in the green frog (Rana clamitans) and their role in male-male communication. Behaviour 133: 283-301.[Web of Science]
Bee MA, Perrill SA, Owen PC, 1999. Size assessment in simulated territorial encounters between male green frogs (Rana clamitans). Behav Ecol Sociobiol 45: 177-184.[Web of Science]
Bond A, 1989. Toward a resolution of the paradox of aggressive displays: I. Optimal deceit in the communication of fighting ability. Ethology 81: 29-46.[Web of Science]
Caldwell RL, 1986. The deceptive use of reputation by stomatopods. In: Deception: perspectives in human and nonhuman deceit (Mitchell RW, Thompson NS, eds). Albany: State University of New York Press; 129-245.
Cheney DL, Seyfarth RM, 1991. Truth and deception in animal communication. In: Cognitive ethology: the minds of other animals (Ristau CA, ed). Hillsdale, New Jersey: Lawrence Erlbaum Associates; 127-151.
Davies NB, Halliday TR, 1978. Deep croaks and fighting assessment in toads Bufo bufo. Nature 274: 683-685.
Dawkins MS, Guilford T, 1991. The corruption of honest signalling. Anim Behav 41: 865-873.
Dawkins R, Krebs JR, 1978. Animal signals: information or manipulation? In: Behavioural ecology: an evolutionary approach, 1st ed (Krebs JR, Davies NB, eds). Oxford: Blackwell; 282-309.
Duellman WE, Trueb L, 1986. Biology of amphibians. Baltimore: Johns Hopkins University Press.
Emlen ST, 1976. Lek organization and mating strategies in the bullfrog. Behav Ecol Sociobiol 1: 283-313.[Web of Science]
Gardner R, Morris MR, 1989. The evolution of bluffing in animal contests: an ESS approach. J Theor Biol 137: 235-243.
Gerhardt HC, 1994. The evolution of vocalization in frogs and toads. Annu Rev Ecol Syst 25: 293-324.[Web of Science]
Given MF, 1987. Vocalizations and acoustic interactions of the carpenter frog, Rana virgatipes. Herpetologica 43: 467-481.[Web of Science]
Given MF, 1988. Territoriality and aggressive interactions of male carpenter frogs, Rana virgatipes. Copeia 1988: 411-421.
Given MF, 1999. Frequency alteration of the advertisement call in the carpenter frog, Rana virgatipes. Herpetologica 55: 304-317.[Web of Science]
Grafen A, 1990. Biological signals as handicaps. J Theor Biol 144: 517-546.[Web of Science][Medline]
Greenhouse SW, Geisser S, 1959. On methods in the analysis of profile data. Psychometrica 24: 95-112.[Web of Science]
Howard RD, 1978. The evolution of mating strategies in bullfrogs, Rana catesbeiana. Evolution 32: 850-871.[Web of Science]
Howard RD, Moorman RS, Whiteman HH, 1997. Differential effects of mate competition and mate choice on eastern tiger salamanders Anim Behav 53: 1345-1356.[Web of Science][Medline]
Howard RD, Young JR, 1998. Individual variation in male vocal traits and female mating preferences in Bufo americanus. Anim Behav 55: 1165-1179.[Web of Science][Medline]
Johnstone RA, 1998. Game theory and communication. In: Game theory and animal behavior (Dugatkin LA, Reeve HK, eds). New York: Oxford University Press; 94-117.
Johnstone RA, Grafen A, 1993. Dishonesty and the handicap principle. Anim Behav 46: 759-764.
Krebs JR, Dawkins R, 1984. Animal signals: mind-reading and manipulation. In: Behavioural ecology: an evolutionary approach, 2nd ed (Krebs JR, Davies NB, eds). Oxford: Blackwell Scientific; 380-402.
Lindman HR, 1974. Analysis of variance in complex experimental designs. San Francisco: W. H. Freeman.
Lopez PT, Narins PM, Lewis ER, Moore SW, 1988. Acoustically induced call modification in the white-lipped frog, Leptodactylus albilabris. Anim Behav 36: 1295-1308.
Martin WF, 1972. Evolution of vocalization in the genus Bufo. In: Evolution in the genus Bufo (Blair WF, ed). Austin: University of Texas Press; 279-309.
Maynard Smith J, Parker GA, 1976. The logic of asymmetric contests. Anim Behav 24: 159-175.
Morton ES, 1977. On the occurrence and significance of motivation-structural rules in some bird and mammal sounds. Am Nat 111: 855-869.
Nelson DA, Marler P, 1990. The perception of birdsong and an ecological concept of signal space. In: Comparative perception, vol. 2 (Stebbins WC, Berkley MA, eds). New York: Wiley; 443-478.
Parker GA, 1974. Assessment strategy and the evolution of fighting behaviour. J Theor Biol 47: 223-243.[Web of Science][Medline]
Payne RJH, Pagel M, 1996. When is false modesty a
false economy? An optimality model of escalating signals. Proc R Soc
Lond B 263:
1545-1550
Ramer JD, Jenssen TA, Hurst CJ, 1983. Size-related variation in the advertisement call of Rana clamitans (Anura: Ranidae), and its effect on conspecific males. Copeia 1983: 141-155.
Reichert SE, 1998. Game theory and animal contests. In: Game theory and animal behavior (Dugatkin LA, Reeve HK, eds). New York: Oxford University Press; 64-93.
Rice WR, 1989. Analyzing tables of statistical tests. Evolution 43: 223-225.[Web of Science]
Robertson JGM, 1986. Male territoriality, fighting and assessment of fighting ability in the Australian frog Uperoleia rugosa. Anim Behav 34: 763-772.
Rosenthal R, Rosnow RL, 1985. Contrast analysis: focused comparisons in the analysis of variance. Cambridge: Cambridge University Press.
Steger R, Caldwell RL, 1983. Intraspecific deceit by
bluffing: a defense strategy of newly molted stomatopods (Arthropoda:
Crustacea). Science 221:
558-560.
Wagner WE Jr, 1989. Fighting, assessment, and frequency alteration in Blanchard's cricket frog. Behav Ecol Sociobiol 25: 429-436.[Web of Science]
Wagner WE Jr, 1992. Deceptive or honest signalling of fighting ability? A test of alternative hypotheses for the function of changes in call dominant frequency by male cricket frogs. Anim Behav 44: 449-462.
Wells KD, 1977. Territoriality and male mating success in the green frog (Rana clamitans). Ecology 58: 1053-1066.
Wells KD, 1978. Territoriality in the green frog (Rana clamitans): vocalizations and agonistic behaviour. Anim Behav 26: 1051-1063.
Wiley RH, 1983. The evolution of communication: information and manipulation. In: Animal behaviour, vol.2. Communication (Halliday TR, Slater PJB, eds). New York: W. H. Freeman; 156-189.
Wiley RH, 1994. Errors, exaggeration, and deception in animal communication. In: Behavioural mechanisms in evolutionary ecology (Real LA, ed). Chicago: University of Chicago Press; 157-189.
Zahavi A, 1977. Reliability in communication systems and the evolution of altruism. In: Evolutionary ecology (Stonehouse B, Perrins C, eds). Baltimore: University Park Press; 253-259.
Zar JH, 1984. Biostatistical analysis. Englewood Cliffs, New Jersey: Prentice-Hall.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. A. Bee Territorial male bullfrogs (Rana catesbeiana) do not assess fighting ability based on size-related variation in acoustic signals Behav. Ecol., January 1, 2002; 13(1): 109 - 124. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||