Behavioral Ecology Vol. 13 No. 4: 472-480
© 2002 International Society for Behavioral Ecology
Mate sampling by female barking treefrogs (Hyla gratiosa)
a Department of Biology, James Madison University, Harrisonburg, VA 22807, USA b Division of Biological Sciences, University of Missouri at Columbia, Columbia, MO 65211, USA
Address correspondence to C.G. Murphy. E-mail: murphycg{at}jmu.edu .
Received 2 March 2001; revised 6 September 2001; accepted 13 September 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGENERAL DISCUSSIONREFERENCES |
|---|
Despite intense interest in mate choice, relatively little is known about how individuals sample prospective mates. Indeed, a key issue is whether females sample males or simply mate with the first male encountered. We investigated mate sampling by female barking treefrogs (Hyla gratiosa). Females choosing mates in natural choruses did not move between males but instead mated with the first male they approached closely. Most females mated with the male closest to them at the start of their mate-choice process, and females were more likely to mate with the closest male when the distance to other males was large. These observations are consistent with the hypothesis that females do not sample potential mates but instead mate with the first male they distinguish from the rest of the chorus. To test this initial detection hypothesis, we conducted a playback experiment in which we offered females a choice between two calls, one of which was detectable above the background chorus sound at the female's release point, and one of which became detectable only as females moved toward the initially detectable call. Females did not prefer the initially detectable call, thus ruling out the initial detection hypothesis and implicating sampling of potential mates by females. Based on the behavior of females in natural choruses, we hypothesize that females approach the chorus, move to locations where they are able to detect the calls of several males simultaneously, and choose a mate from among these males at some distance from the males. Such simultaneous sampling may be common in lekking and chorusing species, which have been the subjects of many studies of sexual selection.
Key words: Hyla gratiosa, mate choice, mate sampling, simultaneous sampling, treefrogs.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGENERAL DISCUSSIONREFERENCES |
|---|
Most theoretical and empirical work on mate choice has focused on determining which traits influence mate choice and on understanding the evolution of mate choice (reviewed by Andersson, 1994
). There is
growing interest, however, in mate-sampling behavior: how individuals gather
information about potential mates and make decisions based on that information
(reviewed by Gibson and Langen,
1996; Jennions and Petrie,
1997). Knowledge about mate-sampling behavior is critical to our
understanding of mate choice for several reasons. First, the outcomes of
coevolutionary models of male traits and female preferences are influenced by
the sampling tactic assumed for females
(Pomiankowski, 1987;
Seger, 1985). Second, it may
be difficult to determine whether mate choice in nature is based on particular
male traits unless we know the extent and duration of sampling
(Gibson, 1996;
Morris, 1989). Third, the
proper design of mate-choice experiments requires a knowledge of sampling
behavior, including the number of mates normally assessed by females and the
spatial relationships of sampled males.
The importance of studies of mate-sampling also extends beyond the field of
sexual selection. The extent of sampling is influenced by cognitive factors
such as expectations of the distribution of male quality (Dombrovsky and
Perrin, 1995; Real, 1990),
perceptual abilities, and memory limitations. Sampling is also affected by
economic factors (Real, 1990)
such as energy expenditure (Milinski and
Bakker, 1992), time available for sampling
(Alatalo et al., 1988;
Real, 1990;
Sullivan, 1994), perceived
risk of predation (e.g., Berglund,
1993; Hedrick and Dill,
1993), and loss of mating opportunities
(Real, 1990). Therefore,
studies that determine how females balance the costs and benefits of mate
sampling will provide us with an understanding of the mechanisms animals use
to solve complex tasks.
The simplest sampling tactic is for females to mate with the first male
encountered (i.e., to sample only a single male;
Janetos, 1980). Alternatively,
females may use any one of a variety of sampling tactics. For example, females
may adopt a threshold-criterion tactic, mating with the first male that
exceeds some threshold (Janetos,
1980; Real, 1990;
Wittenberger, 1983). Females
could also use a relative comparison tactic, in which they sample a number of
males before choosing the male of highest quality from among those sampled
(Downhower and Lank, 1994).
Females adopting this type of tactic might sample males sequentially and
compare only the last n males (e.g., the last two males;
Brown, 1981;
Choudhury and Black, 1993;
Wittenberger, 1983), or they
might assess n males sequentially (or assess males for a fixed amount
of time) and then return to mate with the highest quality male from among
those previously sampled (i.e., a best-of-n tactic;
Janetos, 1980;
Real, 1990). Recent
theoretical work has refined the predictions that can be used to distinguish
between sampling tactics (Wiegmann et al.,
1996), and models have been developed that incorporate uncertainty
in male signals and the ability of females to update information about the
distribution of male traits in a population
(Dombrovsky and Perrin, 1994;
Luttbeg, 1996;
Mazalov et al., 1996).
A number of studies have sought to determine which mate-sampling tactics
females use, but the results have often been ambiguous
(Gibson and Langen, 1996;
Jennions and Petrie, 1997;
Reid and Stamps, 1997;
Wiegmann et al., 1996). This
difficulty arises in part because observations of mate-sampling behavior alone
may be insufficient to distinguish among possible tactics. Because different
models may make the same predictions about the behavior of females under
natural conditions (see Wiegmann et al.,
1996), experiments are often required to distinguish among
alternatives.
In this article, we first present observations of mate-sampling of female barking treefrogs (Hyla gratiosa) in natural choruses. These observations suggest that females use a very simple tactic to choose mates. We then present the results of playback experiments that test and refute this hypothesis. We suggest a more complex sampling tactic that is consistent with the behavior of females in natural choruses.
Study site and species
Male H. gratiosa typically call from ponds that are 0.5-2 m deep
and 20-50 m diam. Chorusing occurs on nearly every night from late April until
mid-August, with mating females present on 75-95% of these nights. Shortly
after sunset, males begin arriving at breeding ponds from the surrounding
forest, with most males arriving within a 40-min period
(Murphy, 1999). Choruses last
3-4 h, with individual males calling for an average of 2.5 h; males return to
the forest once they finish calling for the night. Males float on the surface
of the water and produce advertisement calls consisting of a short (170 ms)
burst of sound with two or three spectral peaks, the fundamental and one or
two of the upper harmonics (Gerhardt,
1981; Oldham and Gerhardt,
1975). In the population we studied, the average fundamental
frequency is 470 Hz (range: 410-550), and males produce an average of 55 calls
per min (range: 42-66 calls/min) (Murphy
and Gerhardt, 1996).
Females arrive at the pond later than males, but there is considerable
overlap in the distribution of arrival times of the two sexes
(Murphy, 1999). Females choose
mates without interference from males, which rarely exhibit satellite
behavior, do not search actively for females, and do not displace other males
from amplexus (Murphy, 1992).
Males do not defend resources of value to females, and females do not have the
opportunity to copy the mate choice of others. Once oviposition is completed,
both members of mated pairs return to the forest. Neither sex provides
parental care.
Mate sampling in natural choruses
Methods
We conducted observations of the mate-sampling behavior of female H.
gratiosa in natural choruses during two consecutive seasons at two
different ponds in the Apalachicola National Forest, near Tallahassee,
Florida, USA (30°22' N, 84°21' W). Females approaching the
pond were captured in a fence located 5-20 m from the edge of the pond
(Murphy, 1993). To help us
follow females, we attached a reflective dot (5 mm diam) to the heads of focal
females with tissue cement. Females were placed inside the fence within 2-5
min of capture (median = 3 min); handling involved in capturing and marking
females disturbed them only momentarily. We observed females from distances of
3-20 m with binoculars and a headlamp dimmed with a rheostat to the point
where the female was just visible. The light used to illuminate females did
not appear to disturb them or calling males; this conclusion is further
substantiated by data on the distances traveled by females not under direct
observation (Murphy, unpublished data). We followed females until they entered
amplexus. During this time, one observer monitored the fence continuously,
immediately placing inside the fence any H. gratiosa intercepted by
the fence. Once females entered amplexus, we mapped the positions of all
calling males; searches were completed within a median of 13.5 min after
females entered amplexus (range: 9-21 min). Females were captured,
freeze-branded to permit identification if encountered later in the season,
and released immediately with their mate.
We mapped paths taken by females and the positions of calling males using a polar grid that covered the entire pond and the shore between the fence and the pond. Markers were placed every 0.5 m along the spokes of the grid, whose angular separation was such that, at 25 m from the center of the grid, they were separated by 1 m.
To determine whether the distance to other males affected whether a female
mated with the male closest to her, we calculated the extra distance
individual females would have had to travel to reach more distant mates by
subtracting the distance to each male from the distance between the female's
location and the closest male. We used logistic regression to determine
whether this additional distance influenced the probability that females mated
with the male closest to them. Hosmer-Lemeshow goodness-of-fit tests
(Hosmer and Lemeshow, 1989)
indicated that a logistic regression adequately fit the data in all tests (all
p 
.25). Sample sizes varied among analyses because not all pieces
of information were collected for all females.
Results
The pattern of mate-sampling behaviors was consistent for all 33 females
across both years (7 in the first season; 26 in the second). After their
release inside the fence, females interspersed short movements between
extended periods of sitting, then made a final movement directly to a male and
initiated amplexus with that male (Figures
1 and
2). No female moved to another
male after approaching a male closely. The average time between a female's
release inside the fence and her initiation of amplexus was 14.8 min (range =
3.5-32.6 min, n = 28), and females spent an average of 83.0% (range =
38.5-98.6%, n = 24) of this time sitting on the shore of the pond.
Females traveled a median of 20.5 m from their release point to their mate
(range: 8.7-40.3 m, n = 25).
|
|
Females contacted their mates after a rapid, final approach (Figures 1 and 2). This final movement covered a median of 5.6 m (interquartile range [IQR] = 2.6-8.0; range = 0.4-30.9, N = 21), or 32.0% (IQR = 12.1-53.3%; range = 2.0-84.4%) of the total distance traveled and required a median of only 0.7 min (IQR = 0.3-1.1; range = 0.1-5.7), or 4.5% (IQR = 2.0-11.5%; range = 0.6-48.2%) of the total time between their release inside the fence and mating. Females moved much more rapidly during their final movement (median = 8.1 m/min; IQR = 5.4-11.1; range = 1.5-31.4) than during the rest of their approach to their mate (median = 1.0 m/min; IQR = 0.7-1.5; range = 0.2-2.3; p =.0001, Wilcoxon matched-pairs signed-rank test). One female was killed by an unidentified turtle as she made her final approach to a male, representing 3.0% of all females observed.
Of the 21 females for which we recorded the positions of all calling males at the pond, 15 (71%) mated with the male that was closest to their release point inside the fence, five females (24%) mated with the second closest male, and one female (5%) mated with the fourth closest male. Sixteen females (76%) mated with the male closest to them at the start of their final movement, four (19%) mated with the second closest male, and one female (5%) mated with the third closest male. For 13 (62%) of the 21 females, the female's eventual mate was the closest male throughout her approach to that male; for the remaining eight (38%) females, the female's mate was the closest male for a median of only 23% of her approach to that male (range: 10.4-97.3%).
We conducted two logistic regression analyses to determine whether the
probability that females mated with the closest male was influenced by the
additional distance that females had to travel to reach more distant males,
measured from the point where females initiated their final movement to their
mate. In the first analysis, we used the additional distance to the second
closest male, and in the second analysis, we used the average additional
distance to the second and third closest males. We chose these two groups
because, at the point of initiation of the final move, most females mated with
the second closest male, and none mated with a male more distant than the
third closest male. To control family-wide error rates, we adjusted 
for each analysis using the sequential Bonferroni technique
(Rice, 1989).
The probability that a female mated with the closest male was positively
related to the extra distance that she had to travel to reach other males.
Females were significantly more likely to mate with the male closest to them
at the beginning of their final movement when the additional distance they had
to travel to reach the second closest male was greater
(Figure 3a; logistic regression
analysis; b =.393, 
2 = 5.21, df = 1, p =.046,
sb =.05, R2 =.26) and when the average
additional distance to the second and third closest males was greater
(Figure 3b; logistic regression
analysis: b =.466, 2 = 6.05, df = 1, p =.014,
sb =.025, R2 =.38).
|
Discussion of field observations
Female H. gratiosa did not move from male to male but instead
approached closely only one male, their mate. This pattern of movement is
similar to that of female painted reed frogs (Hyperolius marmoratus),
which also spend relatively little time choosing mates and often mate with the
male closest to them at their release point
(Grafe, 1997). Mate choice by
female painted reed frogs is also influenced by distance to potential mates
(Grafe, 1997). In contrast,
females of several other species of anuran amphibians have been observed
approaching more than one male closely before making a choice (e.g., Bufo
calamita, Arak, 1988;
Uperolia rugosa, Robertson,
1986; Physalaemus pustulosus,
Ryan, 1985).
One hypothesis for the lack of movement among males by female H.
gratiosa is that females do not sample males but instead orient toward
the first male whose calls they can distinguish from the rest of the chorus
during their approach to the chorus (analogous to Parker's
[1982,
1983] passive attraction
hypothesis and Forrest and Raspert's
[1994; Forrest TG, personal
communication] random response rule). Under this hypothesis, females make
their mating decisions after perceiving only a single male and thus do not
sample more than one male, even though the calls of other males are likely to
become detectable as females move toward the male they first detect
(Forrest and Raspert, 1994;
Parker, 1982,
1983; see
Grafe, 1997, for further
discussion).
This initial detection hypothesis can explain why most females mated with the male closest to them at their release point inside the fence and why females were more likely to mate with the closest male when the distance to other males was greater. Because the intensity of a sound decreases with increasing distance between a receiver and the sound source, the calls of males that are closer to a female will, on average, be more intense and hence be more likely to exceed the intensity of chorus sounds than will the calls of more distant males. When males differ greatly in their distances from an approaching female, there will be large differences among males in their detectability by females, and females that approach the first male whose calls they distinguish from the sound of the chorus will be more likely to mate with closer males than with more distant males. However, when males are at similar distances from an approaching female, the conspicuousness of their calls will be determined predominately by differences among the males in call characteristics such as call repetition rate and the intensity at which they produce calls. In this situation, the calls of the male closest to an approaching female may not be the most intense and therefore may not be the first calls the female distinguishes from the rest of the chorus. Thus, females will be less likely to mate with the closest male when distances between neighboring males are small than when males are more widely separated.
In the next section, we present the results of three playback experiments that tested hypotheses suggested by the field data. The first experiment tested whether choice by females is influenced by the distance to potential males. The second determined the intensity that a call must be relative to the background chorus for it to be detected by females. This information was then used to design the third playback experiment, which explicitly tested the initial detection hypothesis.
Influence of distance on mate choice
Methods
To test the hypothesis that the distance to potential mates affects mate
choice, we tested each female twice with a choice between two speakers
broadcasting synthesized calls differing in their attractiveness to females.
In the first trial, the speakers were equidistant from the female's release
point, and in the second trial, we doubled the distance to the speaker that
she had chosen in the first trial (Figure
4a). We compared the proportion of females that switched choices
between trials in this experiment with the proportion doing so in another
experiment in which speakers were equidistant from the release point during
both trials. With the exception of having different treatments (speakers moved
or not moved between trials) and being conducted in different years, the two
experiments were conducted identically.
|
We used calls that differed in their attractiveness, rather than identical calls, because we wanted to determine whether differences in distance to potential mates could override preferences for call characteristics, a hypothesis that cannot be tested if identical calls are broadcast at different distances. Results supportive of an overriding influence of distance would necessarily demonstrate that distance affects choice between identical calls.
The two calls differed only in fundamental frequency (500 Hz and 550 Hz).
We used these fundamental frequencies because previous experiments had
revealed a preference for 500 Hz over 550 Hz. Calls were synthesized with
custom software (Schwartz J, unpublished data; sampling rate of 20 kHz, 8
bits/sample). Calls lasted 175 ms and had a pulsatile beginning followed by a
harmonic series consisting of the fundamental plus the third and fourth
harmonic, both of which were attenuated 6 dB relative to the fundamental.
Previous results (Gerhardt,
1981) in another population indicated that these calls are as
attractive to females as are natural calls. Calls were stored as digital files
on the hard drive of a laptop computer and played back from the computer
through a Sony XM-3021 stereo amplifier and broadcast from Realistic Minimus 7
speakers. Calls from each speaker were broadcast at a rate of 60 per min,
which is slightly above the population average (see above), and calls
broadcast from each speaker alternated temporally. The root-mean square (RMS)
sound pressure level (SPL, re 20 µPa, fast response, C weighting) of the
two choices was equalized at 70 dB at the release point with a Realistic sound
level meter. To control for side biases, we switched stimuli between speakers
for each female.
We obtained females for playback experiments by collecting pairs in
amplexus; females collected from amplexus are as discriminating as are females
that have just begun their matechoice process
(Murphy and Gerhardt, 1996).
We transported the mated pair to an outdoor playback arena located 0.45 km
from the pond, where we separated the female from her mate before testing. The
playback arena was located at the intersection of two unpaved roads, lacked
any walls, and had a sandy substrate without vegetation.
At the beginning of a trial, we placed the female in an acoustically transparent, hardware-cloth cage. We released the female from the restraining cage after 10 s (i.e., after she heard five repetitions of each call) by manually removing its lid. We observed females using a headlamp dimmed with a rheostat to an intensity at which the female was just visible. We scored a female as having chosen a speaker when she either passed within 10 cm of the speaker or circled it; most females actually contacted the speaker. In natural choruses, a female exhibiting such behavior near a calling male would almost always enter amplexus with that male. We scored a "no response" if the female did not leave the release cage after 3 min or if she hopped outside the playback arena. After testing, the female was freeze-branded and returned along with her mate to the pond so that the pair could oviposit.
Results and discussion
The choices of females with respect to fundamental frequency differed
depending on the relative distances to the two choices. In the first trial, in
which the two speakers were equidistant, 12 of the 15 females chose the 500-Hz
call, and 3 chose the 550-Hz call, a significant preference for the 500-Hz
call (p =.035, binomial test). In the second trial, in which we
doubled the distance to the speaker chosen by the female in first trial, 10
(67%) females moved to the closer speaker (i.e., switched their choice of
stimuli between test); the others moved to the more distant speaker (i.e.,
chose the same stimulus in both trials). The proportion of females switching
choices in this experiment (67%) was significantly greater (p =.003,
Fisher's Exact test) than the proportion (2 of 16 = 13%) doing so in a
separate experiment in which speakers were equidistant from the release point
during both trials.
These results indicate that differences among males in their distances from approaching females can override preferences for particular call characteristics, thereby corroborating the observations of female H. gratiosa in natural choruses. Of course, we cannot determine from our results whether females assessed the actual distance to the speakers or other, correlated cues such as call intensity at the female's location. Additional experiments are needed to determine the specific cue that influences female choice with respect to distance.
Relative call intensity required for detection
Methods
To test the hypothesis that females approach the first male they
distinguish from the sound of the chorus, we first had to determine the
intensity that a call needed to be relative to that of the chorus for the call
to be detected by females. We gave females a choice between one speaker
broadcasting chorus sounds and another broadcasting the same chorus sounds
plus a synthesized male call (Figure
4b). The chorus sound was a looped segment (10 s) recorded 22 m
from a natural chorus of 20-30 males and included calls of cricket frogs
(Acris gryllus) and several unidentified insect species. The
recording was digitized (22.050 kHz, 16 bits) and placed into both the right
and left channels of a SoundEdit16 file; the chorus sound was perfectly
correlated in the two channels (i.e., the segments began and ended at the same
time and hence sounded exactly the same at any point in time). The synthesized
male call was placed into one of the channels at a rate of 60 calls/min, and
its RMS SPL was set in SoundEdit16 such that it was, depending on the playback
series, +6, +3, 0, or -6 dB relative to the chorus sounds broadcast from both
channels. The RMS SPL of the chorus was set at 64 dB at the release point with
a Realistic sound level meter, with that of the male call being 70, 67, 64, or
58 dB, depending on the playback series. For these playbacks, the synthesized
call had a duration of 175 ms, with a pulsatile beginning followed by a
harmonic series consisting of the fundamental (450 Hz) plus the second through
sixth harmonics, which were attenuated relative to the fundamental by -25,
-20, -11, -27, and -35 dB (these values are similar to the population
averages). We chose a fundamental frequency of 450 Hz because this value is
close to the population average (Murphy
and Gerhardt, 1996). We used a six-harmonic call, rather than a
three-harmonic call, because male calls consist of six or more harmonics and
because two-choice playbacks conducted after the distance experiment described
above indicated that females preferred the six-harmonic call (21 females) over
the three-harmonic call (9 females, p =.043). To control for side
biases, we alternated between females which speaker broadcast the male call.
Playback equipment and testing procedures were the same as those described
above, except that we amplified calls with a Kenwood KAC-606 stereo amplifier
and held females in the restraining cage for 30 s after playback commenced. We
used longer wait times in this experiment because females had to discriminate
calls from chorus sounds and may therefore have needed more time for initial
assessment of calls than did females in the distance experiment.
Results and discussion
Females were able to detect a call whose RMS SPL was at least 3 dB greater
than the chorus background (Table
1). These results are similar to those from two-choice playback
experiments with other species of hylid treefrogs. Gerhardt and Klump
(1988) found that female
Hyla cinerea could detect a male's call when the call's RMS SPL was
equal to that of the background chorus sounds but not when the call's RMS SPL
was 6 dB less than the chorus sound. Wollerman
(1999) found that female
H. ebracatta could detect a male call when its SPL was 3 dB greater
than the background chorus sounds but not when the call was 1.5 dB greater
than the chorus sound.
|
Test of the initial detection hypothesis
Methods
Having determined the relative intensity necessary for calls to be detected
above the sound of the chorus, we were able to design an experiment to test
directly the hypothesis that females approach the first male they distinguish
from the rest of the chorus. We conducted three series of playbacks in which
we gave females a choice between two synthetic calls, each played from a
separate speaker, that differed only in whether, at the female's release
point, they were detectable above the sound of a chorus, which was broadcast
from a third speaker (Figure
5). In series 1, the RMS SPL of calls broadcast from a speaker
located 8 m from the release point was 3 dB greater than the chorus sound at
the release point; females were thus able to detect these calls at the release
point. In contrast, calls broadcast from a speaker located 4 m from the
release point had the same RMS SPL as the chorus sound at the release point
and were thus not detectable. If females approach the first male they
distinguish from the rest of the chorus, then females in this series should
move to the 8-m speaker more often than to the 4-m speaker. Calls from the 4-m
speaker would become detectable once females moved about 2 m from the release
point (Figure 5) and would
therefore become detectable to a female moving toward the 8-m speaker.
However, if females choose the first male they distinguish from the chorus,
then they should continue on to the 8-m speaker, even after they have detected
calls from the 4-m speaker.
|
In series 2 and 3, the SPL of calls from both speakers were at least 3 dB greater than the chorus sound at the release point, and thus calls from both speakers were detectable at the release point. In series 2, the intensities of calls from the two speakers were identical at the release point (3 dB above the chorus sounds), whereas in series 3, calls from the 4-m speaker were, at the release point, 9 dB greater than the chorus sound and 6 dB greater than the calls from the 8-m speaker. If females approach the first male they distinguish from the sound of the chorus, then females in the series 2 and 3 should move to the 4-m and 8-m speakers equally often because, in both series, calls from both speakers were detectable at the release point. The fact that calls from the two speakers differ in intensity at the release point in the third series should not influence the decisions of females under this hypothesis; if females do move to the speaker whose calls are more intense at the release point, then the initial detectability of a signal cannot have been a choice criterion.
All three playback series were required to fully test the initial detection hypothesis. In the first series, calls from the 8-m speaker had a higher intensity at their source and at the release point than did calls from the 4-m speaker. Hence, a preference for calls emitted from the 8-m speaker in this series would be consistent with both the initial detection hypothesis and choice based on intensity. The second and third series allowed us to distinguish these alternatives. If females base their choice on intensity of calls at the release point, then they should exhibit no preference in series 2 but should prefer calls played back from the 4-m speaker in series 3. If females base their choice on source intensity, they should prefer calls from the 8-m speaker in series 2 but exhibit no preference in series 3.
The single speaker we used to broadcast chorus sounds did not accurately mimic a natural chorus, in which each male is a separate sound source and males are distributed at various distances and angular separations from an approaching female. However, our design was adequate for testing the hypothesis in question. The arrangement of the speakers mimicked a situation in which two males called from the side of a pond nearest an approaching female, with the rest of the chorus on the opposite side of the pond. In such a situation, which does occur in natural choruses, females choosing the first male they distinguish from the sound of the chorus should have no difficulty choosing between the two speakers in a manner consistent with this tactic.
We used the same looped segment of chorus sounds and the same synthetic call as in the experiment that determined the relative intensity necessary for a call to be distinguished from the chorus sounds. Calls were output from a laptop computer and amplified with a Kenwood KAC-606 stereo amplifier, whereas chorus sounds were played back from a Sony WM-DC6 cassette recorder and amplified with a Sony XM-3021 stereo amplifier. All sounds were broadcast from Realistic Minimus 7 speakers. We followed the same procedures for testing females as describe above. We had no difficulty distinguishing between females that chose the 4-m speaker and those that passed close to this speaker on their way to the 8-m speaker. To control for side biases, we alternated the sides of the arena in which we placed the 4-m and 8-m speakers.
Results and discussion
Females did not prefer the call that was initially distinguishable from the
sound of the chorus over the call that was not
(Table 2). When both calls were
distinguishable at the release point, females exhibited a preference if the
SPL of one of the calls was 6 dB more intense than that of the other at the
release point, but they did not exhibit a preference if both calls were of
equal intensity at the release point (Table
2).
|
These results refute the initial detection hypothesis. The lack of preference for calls emitted by the 8-m speaker in series 1 indicates that female H. gratiosa do not choose the first male whose calls they can distinguish from the rest of the chorus. The lack of a preference in series 2 is consistent with this hypothesis, but the strong preference for calls broadcast by the 4-m speaker in series 3 is inconsistent with the lack of preference predicted by this hypothesis.
Preferences based on distance and source intensity cannot explain the results of the three-speaker experiments. Females did not prefer calls broadcast by the 4-m speaker in all three series, as would be expected if choice were based solely on distance to calling males. Females did not prefer the 8-m speaker in series 1 and 2, as would be expected if females based their choice solely on the intensity of calls at their source. Furthermore, females in series 3 exhibited a very strong preference for calls broadcast from the 4-m speaker over calls from the 8-m speaker, even though the intensities of calls broadcast from the two speakers were the same at the source.
Preference for higher relative intensity among detectable calls may explain the results of the three-speaker experiments. This hypothesis is supported by the lack of preference in series 2, in which calls from the two speakers were of equal intensity at the release point, and by the strong preference in series 3 for the 4-m speaker, whose calls were more intense at the release point than were the calls from the 8-m speaker. At first glance, the results of series 1 do not appear to support this hypothesis because females did not exhibit a preference for the calls played from the 8-m speaker; these calls were more intense at the release point than were calls played from the 4-m speaker. However, in this series, females were unable to detect calls from the 4-m speaker at the release point. Initially, females may have headed for the 8-m speaker, which broadcast calls that females could detect at the release point. As they moved in the direction of the 8-m speaker, some females may have switched to the 4-m speaker when calls from this speaker became distinguishable from the sound of the chorus. Females may have switched to the 4-m speaker at this point because calls from that speaker may have been more intense at the female's location than calls from the 8-m speaker (Figure 5) or because the 4-m speaker was closer to females than the 8-m speaker at that location. Our records of the paths of females during playbacks were not precise enough to allow us to test either of these hypotheses.
|
GENERAL DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONGENERAL DISCUSSIONREFERENCES |
|---|
The results of the three-speaker playback experiments refute the initial detection hypothesis, thereby ruling out a lack of sampling as an explanation for the behavior of female H. gratiosa in natural choruses and implicating, by default, comparisons of potential mates. Based on the behavior of females in natural choruses and the outcome of playback experiments in this and previous studies, we hypothesize that females use a sampling tactic that involves simultaneous assessment of males from a distance. We suggest that females approach the chorus, move to one or more locations where they can assess several males simultaneously, choose a male, and approach that male directly.
This hypothesis is consistent with the behavior of females in natural choruses. Females sat for much of the time during their approach to their mates (Figure 1). We interpret the period of sitting that occurred immediately after release of females inside the fence (Figures 1 and 2) as a response to handling, but subsequent periods of sitting could be periods of assessment of males. During periods earlier in their approach, females may have assessed the chorus as a whole, choosing a general portion of the chorus or a subgroup of males to approach more closely. Later periods of sitting, especially those that occurred immediately before the final movement to a mate (Figures 1 and 2), may represent a period during which the female chooses a mate from among the males she can detect at her location. It is likely that, at this point, females assess relatively few males because the females we observed did not mate with a male more distant than the third closest to them at the beginning of their final movement.
Simultaneous sampling from a distance can also explain why females are more likely to mate with the closest male when the distance to other males was greater. If one male is much closer than others at the point where a female makes her decision, then the cost of traveling to more distant males (e.g., predation risk) may more than offset any gain in male quality (as indicated by call characteristics). However, if several males are nearby, then the benefit of selecting a higher quality male may more than compensate for a slight increase in travel costs.
Similar arguments can be made if females base their choice of mates on relative call intensity at the point of sampling, as suggested by the three-speaker playbacks. When some males are substantially closer than others to an approaching female, there will be large differences in the intensities of their calls at the female's location, and females will be likely to mate with the closest male. However, when males are at similar distances from an approaching female, then differences among males in call intensity at the female's location will be determined predominately by differences among the males in the intensity at which they produce calls (i.e., source intensity). The calls of the closest male will therefore not always be the most intense at the female's decision point, and females will sometimes mate with more distant males.
We did not record the calls of males while females were choosing mates in
natural choruses, so we are unable to determine which call characteristics
might be important to females choosing mates under these circumstances.
However, females exhibit clear preferences for a number of call
characteristics in two-choice playbacks
(Gerhardt, 1981; Murphy and
Gerhardt, 1996,
2000;
Oldham and Gerhardt, 1975),
and these preferences are relative, with the attractiveness of a call
depending on the alternative with which it competes
(Gerhardt, 1981;
Murphy and Gerhardt, 2000).
Relative preferences are consistent with the hypothesis that females sample
males simultaneously and choose the best individual from a group of males.
Our working hypothesis is that females sample males simultaneously from a
distance and incorporate into their decisions both the characteristics of male
calls and the distance to potential mates (or correlates of distance such as
call intensity). Additional experiments, which are in progress, are needed to
test this hypothesis explicitly; the three-speaker experiments presented above
were designed to test the initial detection hypothesis and not to demonstrate
simultaneous sampling. We predict that differences in call rate are especially
likely to mediate female choice. Not only do females show directional
preferences for higher call rates in quiet conditions
(Murphy and Gerhardt, 2000),
but, all things being equal, calls with a high repetition rate should be more
detectable in noisy conditions than calls with a low repetition rate (e.g.,
Grafe, 1997).
Although most discussions and models of mate-sampling tactics have focused
on sequential sampling tactics (see Forrest
and Raspert, 1994 for a notable exception), simultaneous sampling
may be common in lekking or chorusing species, as females of these species
have the opportunity to compare simultaneously the signals of several males
(e.g., Gibson, 1996;
Rintamäki et al., 1995;
Trail and Adams, 1989). In
fact, when attempting to determine the sampling tactics used by females of
such species, it may be critical to include simultaneous sampling in the list
of alternatives. Movement among potential mates is usually taken as evidence
of mate sampling (Gibson and Langen,
1996), and lack of such movement as evidence of lack of sampling
(e.g., Arak, 1988). However,
the absence of movement among males rules out only sequential sampling of
males; it is entirely consistent with simultaneous sampling. Thus, failure to
consider simultaneous sampling may lead to the erroneous conclusion that
females do not sample males and thus do not exhibit mate choice. For example,
Arak (1988) concluded that
female natterjack toads (Bufo calamita) that mated with the first
male they approached closely (21 of the 41 females observed) did not sample
mates. However, it is possible that these females sampled several males
simultaneously from a distance. Given that lekking and chorusing species have
been the focus of much of the theoretical and empirical research in sexual
selection (Andersson, 1994;
Kirkpatrick and Ryan, 1991),
further theoretical and empirical work on simultaneous sampling is likely to
enhance our general understanding of sexual selection.
Further theoretical treatment of simultaneous sampling is also needed
because the perceptual and economic problems faced by females differ for
sequential and simultaneous sampling. Females sampling sequentially need to
process the signals of only one male at a time, whereas females sampling
simultaneously must extract information about individual males from a noisy
background (e.g., Gerhardt and Klump,
1988; Römer et al.,
1989; Schwartz and Gerhardt,
1989). For sequential sampling, the number of males sampled is
expected to decrease as the cost of sampling each male increases
(Real, 1990). However, for
simultaneous sampling, the extent of sampling is unlikely to be limited by the
cost of sampling each additional male because this cost will be low (as
females do not move between males). Instead, the extent of sampling is more
likely to be limited by the perceptual abilities of females to distinguish
males from the background noise. For sequential sampling, the major economic
decision facing females is to balance the improvement in mate quality with the
cost of additional sampling (Real,
1990), whereas females sampling simultaneously must balance
improvement in mate quality with the cost of reaching males.
|
ACKNOWLEDGEMENTS |
|---|
We thank the U.S. Forest Service for permission to conduct this study; Cris Copley, Scott Hamilton, and Dana Otto-Allender for assistance in the field; Joe Travis and the Department of Biology at Florida State University for logistical support; Josh Schwartz for use of his program to synthesize calls; and David Westneat and three anonymous reviewers for helpful comments on the manuscript. This study was supported by a Summer Faculty Assistance Grant from the College of Science and Mathematics, James Madison University, to C.G.M. and by a National Science Foundation grant (IBN-9122219) to H.C.G.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONGENERAL DISCUSSIONREFERENCES |
|---|
Alatalo RV, Carlson A, Lundberg A, 1988. The search cost in mate choice of the pied flycatcher. Anim Behav 36: 289-291.[Web of Science]
Andersson M, 1994. Sexual selection. Princeton, New Jersey: Princeton University Press.
Arak A, 1988. Female mate choice in the natterjack toad: active choice or passive attraction? Behav Ecol Sociobiol 22: 317-327.[Web of Science]
Berglund A, 1993. Risky sex: male pipefishes mate at random in the presence of a predator. Anim Behav 46: 169-175.[Web of Science]
Brown L, 1981. Patterns of female choice in mottled sculpins (Cottidae, Teleostei). Anim Behav 29: 375-382.
Choudhury S, Black JM, 1993. Mate-selection behaviour and sampling strategies in geese. Anim Behav 46: 747-757.[Web of Science]
Dombrovsky Y, Perrin N, 1994. On adaptive search and optimal stopping in sequential mate choice. Am Nat 144: 355-361.[Web of Science]
Downhower JF, Lank DB, 1994. Effect of previous experience on mate choice by female mottled sculpins. Anim Behav 47: 369-372.
Forrest TG, Raspert R, 1994. Models of female choice
in acoustic communication. Behav Ecol 5:
293-303.
Gerhardt HC, 1981. Mating call recognition in the barking treefrog, Hyla gratiosa: Responses to synthetic calls and comparisons with the green treefrog, Hyla cinerea. J Comp Physiol A Sens Neural Behav Physiol 144: 17-26.
Gerhardt HC, Klump GM, 1988. Masking of acoustic signals by the chorus background noise in the green treefrog: a limitation on mate choice. Anim Behav 36: 1247-1249.
Gibson RM, 1996. Female choice in sage grouse: the roles of attraction and active comparison. Behav Ecol Sociobiol 39: 55-59.
Gibson RM, Langen TA, 1996. How do animals choose their mates? Trends Ecol Evol 11: 648-470.
Grafe TU, 1997. Costs and benefits of mate choice in the lek-breeding frog, Hyperolius marmoratus. Anim Behav 53: 1103-1117.
Hedrick AV, Dill LM, 1993. Mate choice by female crickets is influenced by predation risk. Anim Behav 46: 193-196.[Web of Science]
Hosmer DW Jr, Lemeshow S, 1989. Applied logistic regression. New York: John Wiley.
Janetos AC, 1980. Strategies of female mate choice: a theoretical analysis. Behav Ecol Sociobiol 7: 107-112.
Jennions MD, Petrie M, 1997. Variation in mate choice and matting preferences: a review of causes and consequences. Biol Rev 72: 283-327.[Medline]
Kirkpatrick M, Ryan MJ, 1991. The evolution of mating preferences and the paradox of the lek. Nature 350: 33-38.
Luttbeg B, 1996. A comparative Bayes tactic for mate
assessment and choice. Behav Ecol 7:
451-460.
Mazalov V, Perrin N, Dombrovsky Y, 1996. Adaptive search and information updating in sequential mate choice. Am Nat 148: 123-137.[Web of Science]
Milinski M, Bakker TCM, 1992. Costs influence
sequential mate choice in sticklebacks, Gasterosteus aculeatus.
Proc R Soc Lond B 250:
229-233.
Morris MR, 1989. Female choice of large males in the treefrog Hyla chrysoscelis: the importance of identifying the scale of choice. Behav Ecol Sociobiol 25: 275-281.[Web of Science]
Murphy CG, 1992. The mating system of the barking treefrog (Hyla gratiosa) (PhD dissertation). Ithaca, New York: Cornell University.
Murphy CG, 1993. A modified drift fence for capturing treefrogs. Herpetol Rev 24: 143-145.
Murphy CG, 1999. Nightly timing of chorusing by male barking treefrogs (Hyla gratiosa): the influence of female arrival and energy. Copeia 1999: 333-347
Murphy CG, Gerhardt HC, 1996. Evaluating the design of mate-choice experiments: the effect of amplexus on mate choice by female barking treefrogs, Hyla gratiosa. Anim Behav 51: 881-890.
Murphy CG, Gerhardt HC, 2000. Mating preference functions of individual female barking treefrogs, Hyla gratiosa for two properties of male advertisement calls. Evolution 54: 660-669.[Web of Science][Medline]
Oldham RS, Gerhardt HC, 1975. Behavioral isolating mechanisms of the treefrogs Hyla cinerea and H. gratiosa. Copeia 1975: 223-231.
Parker GA, 1982. Phenotype-limited evolutionary stable strategies. In: Current problems in sociobiology. (King's College Sociobiology Group, eds). Cambridge: Cambridge University Press, 173-201.
Parker GA, 1983. Mate quality and mating decisions. In: Mate choice (Batson P, ed). Cambridge: Cambridge University Press; 141-166.
Pomiankowski AN, 1987. The costs of choice in sexual selection. J Theor Biol 128: 195-218.[Web of Science][Medline]
Real L, 1990. Search theory and mate choice. I. Models of single-sex discrimination. Am Nat 136: 376-404.[Web of Science]
Reid ML, Stamps JA, 1997. Female mate choice tactics in a resource-based mating system: field tests of alternative models. Am Nat 150: 98-121.[Web of Science][Medline]
Rice WR, 1989. Analyzing tables of statistical tests. Evolution 43: 223-225.[Web of Science]
Rintamäki PT, Alatalo RV, Höglund J, Lundberg A, 1995. Mate sampling behaviour of black grouse females (Tetrao tetrix). Behav Ecol Sociobiol 37: 209-215.[Web of Science]
Robertson JGM, 1986. Female choice, male strategies and the role of vocalizations in the Australian frog Uperoleia rugosa. Anim Behav 34: 773-784.[Web of Science]
Römer H, Bailey WJ, Dadour I, 1989. Insect hearing in the field. III. Masking by noise. J Comp Physiol A Sens Neural Behav Physiol 164: 609-620.
Ryan MJ, 1985. The Túngara frog: a study in sexual selection and communication. Chicago: University of Chicago Press.
Schwartz JJ, Gerhardt HC, 1989. Spatially mediated release from auditory masking in an anuran amphibian. J Comp Physiol A Sens Neural Behav Physiol 166: 37-41.
Seger J, 1985. Unifying genetic models for the evolution of female choice. Evolution 39: 1185-1193.[Web of Science]
Sullivan MS, 1994. Mate choice as an information gathering process under time constraint: implications for behaviour and signal design. Anim Behav 47: 141-151.
Trail PW, Adams ES, 1989. Active mate choice at cock-of-the-rock leks: tactics of sampling and comparison. Behav Ecol Sociobiol 25: 283-292.[Web of Science]
Wiegmann DD, Real LA, Capone TA, Ellner S, 1996. Some distinguishing features of search behavior and mate choice. Am Nat 147: 188-204.[Web of Science]
Wittenberger JF, 1983. Tactics of mate choice. In: Mate choice (Batson P, ed). Cambridge: Cambridge University Press; 435-447.
Wollerman L, 1999. Acoustic interference limits call detection in a Neotropical frog Hyla ebraccata. Anim Behav 57: 529-536.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  C. G. Murphy The cause of correlations between nightly numbers of male and female barking treefrogs (Hyla gratiosa) attending choruses Behav. Ecol., March 1, 2003; 14(2): 274 - 281. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||