Frog Auditory Behavior
by David D. Olmsted (Copyright - 2000, 2006. Free to use for personal and
Last Revised October 28, 2006
Purpose of the Frog's Auditory System
The purpose of the frog's auditory system is to recognize and properly respond
to the various calls of its own species. These calls (some of which are shown below)
are distinct with regard to their frequency and temporal characteristics and include
calls for mating (advertisement), territory, release, warning, rain, and distress
(Bogert - 1960, Capranica - 1965, Gerhardt - 1968). Unlike the auditory system
in reptiles and mammals which are able to characterize general sound patterns the
frog auditory system is specialized only for recognizing their own calls. Frogs
cannot learn to associate arbitrary sound patterns to various situations (Yerkes
The best summary of the strategies used by frog auditory systems is
given in a review by Albert Feng (1990) below:
"The particular cues utilized
by frogs for call discrimination have been revealed by behavioral studies
employing female phonotaxis (seeking out the sound source), or the evoked
calling response of males, as an indicator of behavioral selectivity and hence
cue utilization. The results of such studies demonstrate that cues utilized
by a given species for call discrimination are highly dependent upon the characteristics
for the mating calls employed by other frog species calling at the same time and
place. In many geographic regions, frog calls are parceled into different frequency
bands; the power spectrum for the call alone suffices to provide cues for recognition."
"For example, the bullfrog mating call has two prominent peaks; a low frequency
band at 250-300 Hz and a higher frequency band at 1000-1500 Hz (Capranica
- 1965). Playback experiments with synthetic and natural sounds show that
the simultaneous presentation of both the low and high frequency bands is both
sufficient and necessary to elicit other male bullfrogs to vocalize (Capranica -
1965). Other frequency combinations, or presentation of a single spectral
band of any frequency range and intensity, fail to evoke male calling behavior.
Similarly, in the green treefrog (Hyla cinerea), two choice experiments also
show that the presence of energy in two frequency bands is most attractive
to females (Gerhardt - 1974). Unlike bullfrogs, however, green treefrogs will
respond to a single frequency band if the sound is sufficiently intense."
"When the dominant frequencies of the calls of a mixed assemblage of species
are similar, mating call discrimination is primarily mediated by temporal
differences. For example, in various hylid frogs the pulse repetition rate
is the most essential cue for discrimination (Blair - 1964; Loftus-Hills and
Little-John - 1971; Gerhardt - 1978) Other studies have demonstrated the importance
of the calling rate (Schneider - 1982), pulse number (Fouquette - 1975), pulse
duration (Narins and Capranica - 1978) and envelope rise time (Gerhardt and
Doherty - 1988)."
"Finally, in many species both the spectral and temporal
features of the mating call are important for call discrimination. For example,
calling by male European grass frogs (R. temporaria) can be elicited only by
stimuli with the appropriate spectral component and pulse repetition rate (Walkowiak
and Brzoska - 1982). In the Puerto Rican "Co Qui" treefrog, Eleutherodactylus
coqui, the duration of the first note "Co", is critical in eliciting male
territorial behavior, while the spectral content of the second note, "Qui",
is crucial in eliciting positive phonotaxic responses from females."
most species of frogs rely heavily upon a single call parameter for recognition,
they do not do so exclusively. Recent behavioral evidence (Gerhardt and Doherty
- 1988) points out that frogs can utilize many of the cues available to them, but
these cues are weighted, and become significant under certain conditions. For examples,
at "cool" temperatures the pulse repetition rate of the H. versicolor mating
call overlaps that of H. chrysoscelis, a sympatric species. Thus, at cooler temperatures
there is a greater possibility that a female of H. versicolor will mistake the call
of a male H. chrysolscellis for that of a male H. versicolor. Under these
conditions other temporal parameters such as pulse duration and the rise time of
the pulse envelope also contribute to call discrimination, thereby reducing the
possibility of a species mismatch. These findings indicate that call discrimination
by frogs is often a complex process, a fact that must be taken into account
when studying the neural basis of this behavior."
The Calls of the Northern Leopard Frog Rana pipiens are shown below
in figure 1. On the left are sonograms giving the frequency and time distributions
for three calls of Rana pipiens while the right side shows their
spectrograms giving the power distributions. The mating call is given by males to
attract females while the release call is given when a frog is unexpectedly
grabbed from behind (as when a nonreceptive female is clasped by an amorous
male or a male is clasped by a male or either is grabbed by a researcher).
The mating trill and chuckle calls were obtained from the record "Voices of the
Night" by P.P. Kellogg and A.A. Allen (Library of natural sounds, Laboratory
of Ornithology, Cornell University, Ithaca, New York. The release call was
recorded in the laboratory.
Figure 2 compares the mating call from three species of frogs.
The Calls of the Northern Leopard Frog Rana pipiens. (Mudry, Karen M., Constantine-Paton, Martha, and Capranica, Robert, R. - 1977)
Single vocalizations from Rana pipiens, Rana blairi, and Rana catesbeiana. (Feng, Albert S., Hall, Jim C., and Gooler, David M. - 1990)
The mating call of the Green Treefrog Hyla cinerea is shown below in figure 3. The top sonogram shows
the frequency and time distribution of four vocalizations. The bottom spectrographs
show the power distribution in the first and third vocalizations. They were recorded
by H.C. Gerhardt.
Mating Call of the Green Treefrog Hyla cinerea.
The Calls of the Grass Frog Rana temporaria are shown below in figure 4
The darker the area the more energy
at that frequency. Sections "a" and "b" of figure 1 show variations in the
mating call given by male frogs to attract female frogs. Section "c"
gives the male release call while section "d" gives the female release call.
Release calls are is given when a frog is unexpectedly grabbed from behind
such as when a nonreceptive female is clasped by an amorous male or a male
is clasped by a male or either is grabbed by a researcher. Section "e"
shows the territorial call which is given when an intruding male approaches
another male. If the intruder does not withdraw the two frogs fight until
one withdraws. These calls may also be uttered between two mating calls even
though no other male is threatening. Lastly, the warning call is shown in section
"f' which is sometimes given as a startled frog dives to safety in its pond.
The Calls of the Grass Frog Rana temporaria. (Brzoska, J., Walkowiak, W. and Schneider, H. - 1977)
Behavioral Experiments n the Frog
For being such small creatures female frogs do remarkably well in tracking
down calling males yet their approach does not occur in a perfectly straight line
as shown in figures 5 and 6. In this series of experiments (Rheinlaender, et. al.-
1979) frogs approach a speaker on the floor producing
only three frequencies (0.9, 2.7, and 3.0 kHz) of their mating call. The results are
described in as follows:
"Shortly after the beginning of sound presentation
the females in the release cage elevated their heads. Most of them made scanning
movements with their heads, turned directly towards the sound source and climbed
onto the edge of the releasing cage. In only one trial did a female fail to turn
directly from her release orientation toward the speaker. In this exceptional
trial she circled 270 degrees in the wrong direction until eventually facing
" In those trials in which females were released when facing
180 degrees away from the loudspeaker, they first turned either to the left
or to the right side until they oriented directly toward it. In those trials in
which females were released when facing the speaker, no orienting turns were observed.
So turning and not hopping until facing the sound source seems to be a general,
fundamental prelude to subsequent phonotaxis. This orienting behavior indicates
that these animals are able to resolve reliably the ambiguity between frontal
versus posterior locations of the source."
"After climbing onto the edge of
the cage, a female typically again made scanning movements from side to side
with her head; this scanning behavior generally lasted for 1 to 3 minutes.
Upon leaping from the edge of the cage onto the floor, she then began her phonotactic
approach to the loudspeaker in a zig-zag manner, as earlier observed by Feng,
et. al. (1976), until she reached the speaker location. Thus the actual jump
direction frequently deviated slightly from the orientation of the head and/or
body (long axis). This zig-zag mode of approach became more and more pronounced
as the female advanced toward the sound source."
"The average length of a female's
jump was 28 cm and she took, on the average, 14 jumps to reach the speaker
location. Generally the females paused between successive jumps; in about
one-fourth of these pauses they made head scanning movements, particularly
during the early stages of their approach. However, head scanning movements
were entirely absent during 4 of 82 runs. Thus, head scanning is not an absolute
requirement for accurate sound localization."
"Normally a female listened to several
calls before each jump, but her jumps were not synchronized with the stimuli.
That is, there was neither a systematic time delay between sound presentation and
her jump nor a fixed number of call repetitions to elicit a jump. If a female deliberately
hopped twice during the silent interval between repetitive stimuli, she obviously
would not have received acoustic cues regarding her new position between the
two jumps. Such double jumps occurred in only 2% of all leaps, indicating
that females usually rely on acoustic cues for reorientation following each
"The total elapsed time from the moment the females left the screened cage
until they reached the wall of acoustic wedges in which the speaker was located
varied between 8 and 360 seconds, with an average of 51 seconds. Jumps tended
to occur more rapidly as a female approached the sound source. In general the first
half of the approach lasted twice as long as the second half. In a few cases some
females crawled short distances of 5-20 cm during the trial. When exhibiting
this mode of phonotaxis, their head orientation and locomotion axes were,
in general, very accurate."
Approach Paths of 41 Frogs (Hyla cinerea) to Their Mating Call. (Rheinlaender, et al - 1979)
The Best and Worst Approach Paths of Hyla cinerea. The range of any head turning is shown by the small lines. The arrow direction indicates the head direction at the jump while the tail indicates the body direction (Rheinlaender, et al - 1979)
If the frog is happens to scan its head back and forth, the head angle relative to
the sound source at jump tends to be more accurate with a mean of 8.4 degrees (figure 7 A). The resulting jumps towards the sound also tend to be more accurate
with a mean of 11.8 degrees even though there is a wide variation
(figure 7 B).
As might be expected the jump angle without scanning is less accurate
averaging 17.6 degrees and it shows an even wider variation having a standard deviation
of 15.6. Significantly, Hyla cinerea will approach the speaker just as well
even if even if the sound consists of only one frequency at 0.9 kHz (compare this with the real mating
Orientations of Head Angles at Jump, and Jump Angles for Hyla cinerea. (Reinlaender, et al - 1979)
The purpose of the frog's head scan would seem to be for determining the side having the
greatest sound intensity since that would indicate the general direction of the sound.
In experiments by Albert Feng (1976) frogs with one ear plugged by a layer
of grease will turn in the opposite direction as shown in figure 8. Albert
Feng (1980) also measured the difference in sound intensity between the sides
of a the head of Rana pipiens to be 4 dB at 1900 Hz and 1 to 2 dB at 170 Hz. The
sound level (RMS measurement) of mating calls for most frogs and toads is between
100 and 85 dB at 50 cm (Gerhardt - 1975) with a 6 dB decline each time the distance
Blocking Sound Into One Ear Produces Repeated Turning. (a) eardrums untouched, (b) Right eardrum coated with grease, (c) grease removed from right eardrum, (d) left eardrum coated with grease. (Feng, et al - 1976)
Relative mating call intensity levels needed for recognition by the frog Hyla cinerea
were determined by Ehret and Gerhardt. Using a 900 Hz tone , the mating
call must be an average of 33 dB more intense than a 55 dB noise background, 40
dB more intense than a 65 dB noise background, and 45 dB more intense than a 75
dB background (Ehret and Gerhardt - 1980).
This sound intensity differential
is further modified by the phase differential of the sound waves. A 1000 Hz sound
wave has a wavelength of 3.3 cm meaning that the maximum phase pressure differential
occurs every half wavelength which is 1.6 cm - about the width of a typical frogs
head. Yet these phase effects are periodic as the wave passes around the frog
such that the phase differential will at one time increase the intensity differential
yet 1 millisecond later (half wavelength) it will decrease the differential.
of sound by the arrival time differential between ears would not seem to be
a major factor due to the small time interval involved. This differential
is only 60 microseconds for frog ears spaced 2 cm apart at 0 degrees C (sound
travels at 332 m/sec at 0 degrees C). This is much too small to be analyzed by the
brain which operates in the range of milliseconds. Yet the proper arrangement of
the synapses on a target neuron can allow this time differential to amplify the
intensity differential (discussed on the page reviewing the Superior Olivary
The sound intensity differences between the ears when the head
is not at right angles to the sound source must be minuscule. This suggests
that an additional sound localizing mechanism must be at work yet nothing
else seems to have been suggested in the literature. The simplest and most
likely mechanism would be to have the frog follow the sound intensity gradient
using cascaded ever increasing auditory thresholds. As shown in figure 9 the
ear's output neurons have a nearly continuous range of intensity thresholds,
much more than is needed just to maintain a large dynamic range from 0 to 100 dB.
Thresholds and Best Excitatory Frequencies (BEF) from 177 Eighth Nerve Axons in Hyla cinerea for Both Pure Tones (A) and Broadband Noise (B). (Ehret and Capranica - 1980)
With cascaded thresholds additional neurons would fire as the frog approaches the
sound source with each triggering more jumps in the same direction. If no more auditory
neurons fired then the frog could stop and use the head scanning method since that
would indicate that the frog was no longer approaching the sound source. If this mechanism is truly at work one
would expect the frog to make more jumps with less scanning as it approached
the sound source since after each jump more auditory neurons would be activated
due to the higher sound gradient near the speaker. As described above in the last
paragraph of the quote from Rheinlaender, et. al. (1979) this is indeed what happens.
Bogert, C.M. (1960) The Influence of Sound on the Behavior of Amphibians
and Reptiles. In Animal Sounds and Communication, pp137-320. Eds: W.E. Lanyon
and W.N. Tavolga. Amer. Inst. Biol. Sci.: Washington D.C.
Brzoska, J., Walkowiak,
W. and Schneider, H. (1977) Acoustic Communication in the Grass Frog (Rana temporaria
L.): Call, Auditory Thresholds and Behavioral Responses. J Comp. Physiol.
Capranica, R.R. (1965) The Evoked Vocal Response of the Bullfrog:
A Study of Communication in Anurans. Research Monographs, 33, MIT Press: Cambridge
Capranica, R.R., and Moffat, J.M. (1983) Neurobehavioral Correlates of Sound
Communication in Anurans. In Advances in Vertebrate Neuroethology edited by
J-P Ewert, R.R. Capranica, D.J. Ingle, Plenum Press - New York
and Capranica, R.R. (1980) Masking Patterns and Filter Characteristics of
Auditory Nerve Fibers in the Green Treefrog (Hyla cinerea). J. Comp. Physiol.
Ehret, G. and Gerhardt, H.C. (1980) Auditory Masking and Effects
of Noise on Responses of the Green Treefrog (Hyla cinerea) to Synthetic Mating
Calls. J. Comp. Physiol. 141:13-18
Feng, A.S., Gerhardt, H.C., and Capranica R.S.
(1976) Sound Localization Behavior of the Green Treefrog (Hyla cinerea) and
the Barking Treefrog (H. gratiosa).
J. Comp. Physiol. 107:241-252
Directional Characteristics of the Acoustic Receiver of the Leopard frog (Rana pipiens): a Study of Eighth Nerve Auditory Responses. J. Acoust. Soc. Am. 68:1107-1114
Feng, Albert S., Hall, Jim C., and Gooler, David M. (1990) Neural Basis of
Sound Pattern Recognition in Anurans. Progress in Neurobiology, 34:313-329
Gerhardt, H.C.(1975) Sound Pressure Levels and Radiation Patterns of the Vocalizations
of Some North American Frogs and Toads. J. Comp. Physiol. 102:1-12
Karen M., Constantine-Paton, Martha, and Capranica, Robert, R. (1977) Auditory
Sensitivity of the Diencephalon of the Leopard Frog, Rana pipiens. J. Comp.
Rheinlaender, J., Gerhardt, H.C., Yager, D.D., and Capranica,
R.R. (1979) Accuracy of Phonotaxis by the Green Treefrog (Hyla cinerea). J.
Comp. Physiol. 133:247-255