Content uploaded by Henry Heffner
Author content
All content in this area was uploaded by Henry Heffner on Oct 18, 2019
Content may be subject to copyright.
Audiograms of ¢ve species of rodents: implications for the evolution of
hearing and the perception of pitch
R.S. He¡ner, G. Koay, H.E. He¡ner *
Department of Psychology, University of Toledo, Toledo, OH 43606 USA
Received 30 January 2001; accepted 10 April 2001
Abstract
Behavioral audiograms were determined for five species of rodents : groundhog (Marmota monax), chipmunk (Tamias striatus),
Darwin's leaf-eared mouse (Phyllotis darwinii), golden hamster (Mesocricetus auratus), and Egyptian spiny mouse (Acomys
cahirinus). The high-frequency hearing of these animals was found to vary inversely with interaural distance, a typical mammalian
pattern. With regard to low-frequency hearing, the animals fell into two groups: those with extended low-frequency hearing
(chipmunks, groundhogs, and hamsters hear below 100 Hz) and those with restricted low-frequency hearing (spiny and leaf-eared
mice do not hear appreciably below 1 kHz). An analysis of mammalian hearing reveals that the distribution of low-frequency hearing
limits is bimodal with the two distributions separated by a gap from 125 to 500 Hz. The correspondence of this dichotomy with
studies of temporal coding raises the possibility that mammals that do not hear below 500 Hz do not use temporal encoding for the
perception of pitch. ß 2001 Published by Elsevier Science B.V.
Key words: Audiogram ; Evolution ; Frequency coding; Groundhog; Chipmunk; Leaf-eared mouse; Hamster ; Spiny mouse
1. Introduction
Over the last few years, we have had the opportunity
to test the hearing of ¢ve species of rodents. This brings
the total number of species in the order Rodentia with
reasonably complete behavioral audiograms to 20. Each
of the ¢ve species tested was selected to broaden the
sample of mammals whose hearing is known. The east-
ern chipmunk (Tamias striatus) and the groundhog
(Marmota monax), both sciurids, dig their own nest
burrows and were tested to increase the number of
burrowing rodents in the sample. In addition, the
groundhog is a large rodent (7 kg) and its inclusion
expands the size range of the rodents in the sample.
Darwin's leaf-eared mouse (Phyllotis darwinii), a murid,
increases the sample of very small rodents ( 650 g) to
four. The golden hamster (Mesocricetus auratus), a
murid rodent, is the only common laboratory animal
used in auditory research whose behavioral audiogram
has not previously been reported. Finally, the spiny
mouse (Acomys cahirinus), also a murid, is a precocious
rodent and was tested to provide information to those
interested in an animal that is well developed at birth.
In this report we describe the audiograms of these
¢ve species and compare them to those of other mam-
mals to gain insight into the selective pressures involved
in the evolution of mammalian hearing. Of particular
interest is the ¢nding that mammalian audiograms fall
into two distinct groups based on whether or not they
show good low-frequency hearing. Analysis of this di-
chotomous distribution raises the possibility that ani-
mals that do not hear below 500 Hz do not use tempo-
ral coding for the perception of pitch.
2. Methods
All of the animals were tested using a conditioned
suppression/avoidance procedure in which a thirsty an-
imal was trained to lick a water spout in order to re-
0378-5955 / 01 / $ ^ see front matter ß 2001 Published by Elsevier Science B.V.
PII: S0378-5955(01)00298-2
* Corresponding author. Tel.: (419) 530-2257 ;
Fax: (419) 825-1659 ; E-mail : hhe¡ne@pop3.utoledo.edu
Abbreviations: pps, pulses per second ; r, correlation coe¤cient;
SPL, sound pressure level re 20 WN/m2
HEARES 3706 12-7-01
Hearing Research 157 (2001) 138^152
www.elsevier.com/locate/heares
ceive a steady trickle of water (He¡ner and He¡ner,
1995). Pure tones were then presented at random inter-
vals and followed at their o¡set by a mild electric shock
delivered through the spout. The animal learned to
avoid the shock by breaking contact with the spout
when a tone was presented, a response indicating that
it had heard the tone. Absolute thresholds were then
determined for tones throughout each species' hearing
range.
2.1. Subjects
Eastern chipmunks (T. striatus). Three animals of
undetermined sex (designated A, B, and C), weighing
85^111 g, were wild-trapped in Lucas County, OH,
USA. They were housed in glass tanks (50.8U
25.4U30.5 cm) with corncob bedding and provided
small wooden nest boxes (20U9U7.6 cm). The nest
boxes were equipped with sliding doors and were used
to transfer the animals from their home cage to the test
cage. They were given access to rodent blocks, sun-
£ower seeds, and mixed nuts with occasional supple-
ments of fruits and vegetables.
Groundhogs, also known as marmots or woodchucks
(M. monax). Four young males (designated A, B, C,
and D), weighing 2.7^5.7 kg, were wild-trapped in Lu-
cas and Fulton Counties, OH, USA. They were housed
in large glass tanks (91U32U43 cm) with corncob bed-
ding and provided free access to rodent blocks, and
monkey chow, with occasional supplements of fruits
and vegetables. The groundhogs went into hibernation
during the fall and winter months during which time
they stopped eating and drinking and became torpid.
Thus, all testing was conducted during the spring and
summer months.
Hamsters (M. auratus). Eight male Syrian golden
hamsters (designated A through H), weighing 117^
140 g, were obtained from Charles River Laboratory
and housed in standard solid bottom cages
(33U21.6U19 cm) with corncob bedding. They were
given free access to rodent blocks and occasional pieces
of apple.
Darwin's leaf-eared mice (P. darwinii). Two females
(designated A and B), weighing 35^49 g, were pur-
chased from a local animal supplier. They were housed
in the same type of glass tanks and nest boxes used for
chipmunks and given free access to rodent blocks, with
occasional supplements of seeds and vegetables.
Spiny mice (A. cahirinus). Four animals, two males
(designated A and C) and two females (designated B
and D), weighing 50^69 g, were obtained from a local
animal supplier. They were housed and fed in the same
manner as the leaf-eared mice.
The animals received water in the test sessions and
were weighed before each session to monitor their dep-
rivational state. The care and use of the animals in this
study were approved by the University of Toledo Insti-
tutional Animal Care and Use Committee.
2.2. Behavioral apparatus
Testing was conducted in a carpeted, double-walled
chamber (IAC model 1204; Industrial Acoustics Co.,
Bronx, NY, USA; 2.55U2.75U2.05 m). To reduce
sound re£ections, the walls and ceiling were lined with
acoustic foam. The equipment for behavioral and stim-
ulus control was located outside the chamber and the
animals were observed over closed-circuit television.
The animals were tested in wire cages mounted ap-
proximately 1 m above the chamber £oor to minimize
sound-re£ecting surfaces in the vicinity of the animal.
The groundhogs were tested in a cage (74U38.5U24
cm) constructed of 1-inch (2.54-cm) welded wire fencing
and mounted on four narrow wooden legs. The other
rodents were tested in a cage (35U21U24 cm) con-
structed of half-inch (1.27-cm) wire mesh on a support-
ing frame of 1/8-inch (3.2-mm) brazing rods and
mounted on a camera tripod. When testing hamsters
and chipmunks, the width of the cage was restricted
by a narrow (7-cm), shoulder-high wire mesh fence
that ensured they were directly facing a loudspeaker
when licking the water spout. The leaf-eared mice
and spiny mice were positioned in front of the loud-
speaker by requiring them to climb onto a narrow
wire mesh platform (16 cmU6U6 cm) to reach the
water spout.
A water spout, consisting of 15-gauge brass tubing
with a small brass oval `lick plate' at the top, came up
through the £oor in the front of a cage (Fig. 1). The
spout was adjusted to a level that permitted an animal
to drink comfortably with its head facing towards a
loudspeaker. This position placed the spout low enough
Fig. 1. Drawing of the cage used to test the hamsters and chip-
munks.
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152 139
that it did not interfere with the sound reaching the
animal's ears. For the mice, chipmunks, and ground-
hogs, water was delivered to the spout from a constant-
pressure water reservoir (Marriotte bottle) through an
electrically operated water valve with the £ow rate con-
trolled by varying the rate of operating pulses sent to
the water valve (e.g., 2^3 per s). For the hamsters, water
was delivered using a 25-ml syringe pump with an ad-
justable drive. The £ow rate was adjusted so that an
animal could obtain adequate water in a single test
session lasting 30^60 min. Requiring the animals to
keep their mouths on the water spout served to ¢x their
heads in the sound ¢eld, allowing precise measurement
of the intensity of the sound at their ears.
A contact circuit, connected between the water spout
and cage £oor, turned on the water whenever an animal
touched the spout. Mild shock, which was provided by
a shock generator connected between the spout and the
cage £oor, could be avoided or escaped by breaking
contact with the spout. A 15-W light, mounted 0.5 m
below the cage, was turned on whenever the shock was
on and the animals learned to return to the spout fol-
lowing a shock as soon as the `shock light' was turned
o¡.
2.3. Acoustical apparatus
Sine waves from 16 to 80 000 Hz were generated by a
tone generator (Hewlett Packard 209A or Krohn-Hite
2400) and the frequency veri¢ed with a frequency coun-
ter (Fluke 1900A). The signal was shaped by a rise/fall
gate (Coulbourn S84-04, cosine gating) with a 10-ms
rise/fall time for frequencies of 1 kHz and higher. For
frequencies from 63 to 500 Hz, rise/fall times were used
that allowed at least 10 cycles during signal onset and
o¡set. For 16 and 32 Hz, rise/fall times of 270 and
160 ms, respectively, were used with the signal gated
on at zero crossing (i.e., when the phase of the sine
wave was at zero voltage).
For frequencies of 125 Hz and higher, four pulses of
400-ms duration each were presented, with a 100-ms
interval between pulses. To accommodate the longer
rise/fall times, the intertrial interval was lengthened to
160 ms for 32 and 63 Hz. For 16 Hz, the four pulses
were 500 ms in duration with a 270-ms inter-pulse in-
terval.
The intensity of the pure tones was adjusted in 5-dB
steps using an attenuator (Hewlett Packard 350D), the
linearity of which was calibrated throughout the range
used in testing. The electrical signal was band-pass-
¢ltered (Krohn-Hite 3550 ; þ 1/3 octave), ampli¢ed
(Crown D75), and sent to a loudspeaker. The electrical
signal to the loudspeaker was monitored with an oscil-
loscope. One of the following loudspeakers was used
each session depending on the frequency being tested :
15-inch (38-cm) or 12-inch (30.5-cm) woofer, 6-inch
(15.2-cm) or 3-inch (7.6-cm) midrange, ribbon tweeter
or piezoelectric tweeter. The loudspeaker was placed at
ear level 1 m in front of the animal for frequencies of
63 Hz and above. At 16 and 32 Hz, the 15-inch loud-
speaker, which was mounted in a 0.45-m3enclosure,
was turned to face into a corner of the acoustic cham-
ber. The test cage was then placed in a standing wave,
which was located using the measuring microphone.
This arrangement was necessary to achieve a su¤ciently
intense, but undistorted sound at the location of the
animal. It may be noted that sound-measuring micro-
phones are not directionally sensitive at such low fre-
quencies and, indeed, varying the orientation of the
microphone did not change the sound pressure level
(SPL) reading. As a precaution against substrate-borne
vibrations, 8-cm thick foam pads were placed under the
woofer and the legs of the test cage when testing 16 and
32 Hz.
The SPL (re 20 WN/m2) was measured daily using a
Bru
«el and Kjaer 1/4-inch (0.64-cm) microphone (model
4135) or 1-inch (2.54-cm) microphone (model 4231),
preampli¢er (model 2618), microphone ampli¢er (model
2608), and ¢lter (Krohn-Hite 3202 set to pass one oc-
tave above and below the test frequency). The measur-
ing system was calibrated with a pistonphone (Bru
«el
and Kjaer model 4230). Sound measurements were tak-
en by placing the microphone in the position occupied
by an animal's head when it was drinking and pointing
it directly toward the loudspeaker (0³ incidence). Care
was taken to produce an homogeneous sound ¢eld
( þ 1 dB) in the area occupied by the animal's head
and ears. The acoustic signal was also analyzed for
overtones by sending the un¢ltered output of the
sound level meter to a spectrum analyzer (Zonic
3525). This analysis indicated that overtones, which
were present when low-frequency tones were produced
at high intensity, were at least 20 dB below the animal's
threshold.
2.4. Behavioral procedure
A thirsty animal was trained to make steady contact
with the water spout in order to obtain a slow but
steady trickle of water. A train of four tone pulses
was presented at random intervals and followed at its
o¡set by a mild electric shock (300 ms maximum dura-
tion) delivered between the spout and the platform. The
animal soon learned to avoid the shock by breaking
contact with the spout whenever it heard a tone. The
shock was adjusted for each animal to the lowest level
that reliably produced an avoidance response. The
mildness of the shock was con¢rmed by the willingness
with which an animal returned to the spout after the
shock had been delivered.
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152140
Test sessions were divided into trials lasting 2^3 s
(depending on the frequency being tested) and sepa-
rated by 1.5-s intertrial intervals. Approximately 22%
of the trials contained a pulsing tone (warning signal)
while the remaining trials contained only silence (safe
signal). The contact circuit detected whether an animal
was in contact with the spout during the ¢nal 150 ms of
every trial. If an animal broke contact for more than
half of the 150-ms response period, an avoidance re-
sponse was recorded. This response was classi¢ed as a
hit if the trial contained a tone (warning signal) or as a
false alarm if the trial consisted of silence (safe signal).
Typically, the same tone (i.e., same frequency and in-
tensity) was presented for 6^8 successive warning trials
and approximately 30 associated safe trials following
which the hit and false alarm rates were calculated.
The hit rate was then corrected for false alarms to pro-
duce a performance measure for that stimulus using the
formula:
Performance Hit rate3False alarm rateUHit rate
This measure proportionately reduces the hit rate by
the false alarm rate observed for a particular stimulus
and can range from 0 (no hits) to 1 (100% hit rate with
no false alarms, i.e., perfect performance).
Auditory thresholds were determined by reducing the
intensity of the tone in successive blocks of 6^8 warning
trials until the animal no longer responded to the warn-
ing signal above the 0.01 level of chance, that is, the hit
rate was no longer signi¢cantly higher than the false
alarm rate (binomial distribution). After a preliminary
threshold had been obtained, ¢nal threshold determina-
tion was conducted by presenting tones varying in 5-dB
increments from 10 dB above to 10 dB below the esti-
mated threshold. A typical test session for a trained
animal would consist of approximately 35^75 tone
(warning) trials and approximately four times as
many silent (safe) trials. Threshold was de¢ned as the
intensity at which the performance measure equaled
0.50, which was usually obtained by interpolation.
For a particular frequency, initial testing was consid-
ered complete when the thresholds obtained in at least
three di¡erent sessions were within 3 dB of each other.
Once an entire audiogram had been completed for an
individual animal, each frequency was retested at least
once to ensure threshold reliability.
3. Results
The audiograms of the ¢ve species presented
here in graphical form are available in tabular form
on the Internet (http://www.utoledo.edu/psychology/
animalhearing/).
3.1. Eastern chipmunk
Beginning at 16 Hz, the audiograms of the three east-
ern chipmunks show a gradual increase in sensitivity as
frequency is increased to about 250 Hz with little
change in sensitivity between 250 Hz and 45 kHz
(Fig. 2). Sensitivity declines rapidly for frequencies
above 45 kHz with 56 kHz being the highest frequency
to which an animal responded. At a level of 60 dB SPL,
the chipmunks have a broad hearing range extending
from 39 Hz to 52 kHz (10.4 octaves) with an average
best sensitivity of 16.7 dB at 1 kHz. Compared with
other mammals, chipmunks have good low-frequency
hearing, but relatively poor overall sensitivity as only
500 Hz and 1 kHz are audible at levels below 20 dB
SPL.
Because it is unusual (although not unknown) for
small animals to hear low frequencies (He¡ner and
He¡ner, 1998), the chipmunk was tested down to
16 Hz. In doing this, it was necessary to rule out the
possibility that the chipmunks might have been re-
sponding to extraneous cues. Because it can be di¤cult
to generate tones below 63 Hz at high intensities with-
out harmonics, one possibility is that the animals were
responding to high-frequency harmonics in the signal.
However, a spectrum analysis indicated that all har-
monics were at least 20 dB below the animals' thresh-
olds making it unlikely that the animals were respond-
ing to them. A second possibility is that the animals
were responding to non-auditory cues, such as cage
vibrations, air movement on the animals' facial hair
or vibrissae (which could occur if the animals were in
the `near-¢eld' portion of the sound ¢eld), or that they
could see the movement of the loudspeaker diaphragm
at low frequencies. However, it is unlikely that the
animals were responding to air currents or that they
could see the speaker vibrating because at 16 and
32 Hz the speaker was facing away from the animals.
Although it is not possible to completely rule out
cage vibrations, it should be noted that laboratory
rats tested concurrently with the same apparatus never
responded to frequencies below 250 Hz even at in-
tensities of 100 dB SPL (He¡ner et al., 1994). Thus,
we are con¢dent that the low-frequency thresholds
are valid.
3.2. Groundhog
Beginning at 32 Hz, the audiograms of the four
groundhogs show a gradual increase in sensitivity as
frequency is increased with their best hearing occurring
at 4 and 8 kHz (Fig. 2). At higher frequencies, sensitiv-
ity decreases slightly at 16 kHz, followed by a small
increase in sensitivity at 22.4 kHz. Above 22.4 kHz,
sensitivity declines rapidly with 32 kHz being the high-
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152 141
est frequency to which they responded. At a level of
60 dB SPL, the groundhogs have a broad hearing range
extending from 40 Hz to 27.5 kHz (9.4 octaves) with an
average best sensitivity of 21.5 dB at 4 kHz. Like chip-
munks, groundhogs have good low-frequency hearing
and relatively poor sensitivity as they do not hear ap-
preciably below 20 dB SPL. However, their 60-dB high-
frequency limit is about one octave lower than that of
the chipmunks, making them noticeably less sensitive at
high frequencies.
Fig. 2. Audiograms of the ¢ve species of rodents. Individual animals are designated by letters. Note that the three species in the left column
have more extensive low-frequency hearing than the two species in the right column.
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152142
3.3. Hamster
The complete audiograms of four hamsters (hamsters
A, B, C, and D) are shown in Fig. 2 along with the
partial audiograms of four additional animals (hamsters
E^H). Beginning at 32 Hz, the audiograms show a
gradual increase in sensitivity as frequency increases
up to a comparatively well-de¢ned point of best hearing
at 10 kHz. Sensitivity declines noticeably at 16 kHz
followed by improvement from 20 to 32 kHz. Above
32 kHz, sensitivity declines rapidly to 50 kHz, the high-
est frequency to which they responded. At a level of 60
dB SPL, the hamsters show a broad hearing range ex-
tending from 96 Hz to 46.5 kHz (8.9 octaves) with an
average best sensitivity of 1 dB at 10 kHz. Although the
hamsters' low-frequency hearing is not quite as good as
that of chipmunks and groundhogs, they have much
better sensitivity in their range of best hearing with
frequencies from 4 to 12.5 kHz audible at a level below
20 dB SPL.
3.4. Darwin's leaf-eared mouse
Beginning at 1 kHz, the audiograms of the two leaf-
eared mice show a comparatively sharp increase in sen-
sitivity as frequency is increased with a well-de¢ned
point of best hearing around 11 kHz (Fig. 2). Sensitivity
declines at 16 kHz followed by a plateau and an im-
provement in sensitivity at 45 kHz. Above 64 kHz,
sensitivity declines rapidly to 80 kHz, the highest
frequency to which one animal responded. At a level
of 60 dB SPL, their hearing range extends from 1.55
kHz to 73.5 kHz (5.5 octaves) with an average best
sensitivity of 33.5 dB at 11 kHz. Compared with the
previous three rodents, Darwin's leaf-eared mice have
better high-frequency sensitivity and much poorer low-
frequency sensitivity. In addition, they have superior
best sensitivity although their ability to hear below
20 dB SPL is limited to a narrow range around 8^11
kHz.
3.5. Spiny mouse
Beginning at 1 kHz, the audiograms of the four spiny
mice, like those of Darwin's leaf-eared mice, show a
sharp increase in sensitivity as frequency is increased
to 8 kHz, their frequency of best sensitivity (Fig. 2).
Sensitivity then declines gradually to 32 kHz with a
small improvement at 45 kHz. Above 45 kHz, sensitiv-
ity declines rapidly to 80 kHz, the highest frequency to
which the animals responded. At a level of 60 dB SPL,
their hearing range extends from 2.3 kHz to 71 kHz
(4.9 octaves) with an average best sensitivity of 14 dB
at 8 kHz. Spiny mice are able to hear below 20 dB SPL
at two frequencies, 8 and 16 kHz.
4. Discussion
The audiograms of these rodents and those of other
mammals are discussed with respect to ¢ve issues:
(1) the variation in high-frequency hearing and its rela-
tion to sound localization, (2) the occurrence of second-
ary peaks of sensitivity that are apparently due to the
pinnae, (3) a dichotomy in the distribution of mamma-
lian low-frequency hearing that suggests species di¡er-
ences in the mechanisms used in the perception of pitch,
and (4) the variation in low-frequency hearing and its
relation to high-frequency hearing.
4.1. Variation in high-frequency hearing
Rodents show more variation in high-frequency hear-
ing than any other order of mammals. Using the high-
est frequency audible at 60 dB SPL as a standard for
comparison, rodent high-frequency hearing limits ex-
tend from 5.9 kHz for the blind mole rat (the poorest
high-frequency limit of any mammal) to 92 kHz for the
wild house mouse, a range of 3.96 octaves (He¡ner and
He¡ner, 1998). Indeed, only echolocating bats and ce-
tacea are known to hear higher frequencies than ro-
dents and including them in the comparison extends
the high-frequency hearing limits of mammals to
150 kHz, only 0.71 octave higher (Bitter et al., 2001).
Good high-frequency hearing is common in rodents, as
it is in mammals as a whole, and the 20 rodent species
for which behavioral data are available have a median
upper limit of 52 kHz.
The variation in mammalian high-frequency hearing
has been attributed to the selective pressures involved in
sound localization (e.g., He¡ner and He¡ner, 1998 ;
Masterton et al., 1969). Brie£y, mammals with small
heads and pinnae need to hear higher frequencies
than larger mammals in order to use interaural intensity
di¡erences and pinna cues to localize sounds. That is,
the interaural intensity-di¡erence cue is e¡ective only if
an animal can hear frequencies that are high enough to
be shadowed by its head thereby resulting in a di¡er-
ence in the intensity of a sound reaching the two ears.
Similarly, pinna cues are available only if an animal
hears sounds that are high enough to be modi¢ed by
the pinna as a function of the angle of the sound source
relative to the head. Just how high an animal needs to
hear in order to use these two cues depends on the size
of its head and pinnae ; the smaller they are, the higher
the animal must hear to use these two cues.
As illustrated in Fig. 3, the ¢ve rodents tested here
conform to the relationship between head size and high-
frequency hearing. This ¢gure shows the relationship
between `functional' head size and high-frequency hear-
ing, with functional head size de¢ned as time required
for sound to travel around the head from one ear to the
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152 143
other, a measure that is directly related to head size and
indirectly related to pinna size. As can be seen, mam-
mals with functionally small heads hear higher frequen-
cies than those with large heads and, in general, larger
pinnae (correlation coe¤cient, r=30.786, P60.0001).
Although most rodents conform to the relationship
between functional head size and high-frequency hear-
ing, subterranean rodents are a notable exception (Fig.
3). Interestingly, the failure of the naked mole rat,
pocket gopher, and blind mole rat to hear as high as
predicted by their head size supports the idea that high-
frequency hearing evolved in mammals for sound local-
ization. This is because subterranean mammals, which
live exclusively in burrows where directional responses
to sound are limited, have also lost the ability to local-
ize sound (e.g., He¡ner and He¡ner, 1990, 1992, 1993,
1998). The observation that mammals that do not local-
ize sound lose the ability to hear high frequencies sup-
ports the theory that high-frequency hearing evolved in
mammals for sound localization.
4.2. Secondary peaks of high-frequency sensitivity
Examination of the audiograms of the ¢ve rodents
tested here reveals the existence of secondary peaks of
sensitivity at frequencies well above the animals' fre-
quencies of best hearing (e.g., at 50 kHz for the leaf-
eared mouse and at 45 kHz for the spiny mouse in Fig.
2). Such secondary peaks have been noted in other spe-
cies and their occurrence has been attributed by some
to the specialization of the audiogram for ultrasonic
communication (e.g., Brown, 1970 ; Floody, 1979).
Recent evidence, however, indicates that these peaks
result from the directionality of the pinnae, which en-
ables animals to localize sound in the vertical plane and
to reduce front^back confusions (e.g., Butler, 1975,
1999; He¡ner et al., 1996 ; Musicant and Butler,
1984; Ro¥er and Butler, 1968). Speci¢cally, these sec-
ondary peaks of sensitivity have been shown in bats to
vary with the elevation of the sound source (Koay et
al., 1998; Wotton et al., 1995). Furthermore, the view
that such peaks are due to the external ear and are not
necessarily associated with communication is supported
by the existence of a secondary peak of high-frequency
sensitivity in the human audiogram at 13 kHz that is
attributed to the acoustic characteristics of the auditory
canal (Shaw, 1974). Thus, the existence of high-fre-
quency peaks does not provide convincing evidence
that the hearing of rodents was modi¢ed by selective
pressure for intraspeci¢c communication. Instead, the
presence of ultrasonic vocalizations in rodents may rep-
resent the co-optation of high-frequency hearing, orig-
Fig. 3. High-frequency hearing limit (highest frequency audible at 60 dB SPL) varies with functional interaural distance (the number of Ws re-
quired for a sound to travel from one auditory meatus to the other). Mammals with small interaural distances hear higher frequencies than
larger mammals. Open circles indicate the ¢ve species in this study and are labeled. Animals that use echolocation (bats and cetacea) are indi-
cated by the letter e. Additional species are labeled for reference. Note that the three subterranean species, blind mole rat, naked mole rat, and
pocket gopher, were not included in the correlation (for references to individual species, see Koay et al., 1998).
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152144
inally evolved for sound localization, for use in commu-
nication. Indeed, the frequency of a species' vocaliza-
tions seems to be determined by its audiogram, not the
other way around, as naked mole rats, which lack high-
frequency hearing because they do not need to localize
sound, have developed an extensive repertoire of low-
frequency communication calls (He¡ner and He¡ner,
1993; Pepper et al., 1991).
4.3. Distribution of low-frequency hearing limits
The variation in mammalian low-frequency hearing is
even greater than that for high-frequency hearing.
Among rodents, the 60-dB low-frequency hearing limit
extends from 28 Hz (black-tailed prairie dog) to 2.3 kHz
(spiny mouse and wild house mouse), a range of
6.36 octaves. For mammals as a whole, low-frequency
hearing limits extend from 17 Hz (Indian elephant) to
10.3 kHz (little brown bat), a range of 9.24 octaves that
is almost twice the 4.67-octave range in high-frequency
hearing (He¡ner and He¡ner, 1998).
In attempting to explain the variation in low-fre-
quency hearing, it has been noted that high- and low-
frequency hearing are correlated such that animals with
good high-frequency hearing tend to have poor low-
frequency hearing, and vice versa (He¡ner and Master-
ton, 1980; Koay et al., 1998). However, before address-
ing this relationship, it should be noted that mammals
appear to fall into two groups based on whether they
have good or poor low-frequency hearing, a dichotomy
that is especially well illustrated by the ¢ve rodents
tested here (Fig. 4).
4.3.1. Dichotomy in the distribution of low-frequency
hearing limits
For some time we have noticed a gap in the distri-
bution of low-frequency hearing limits in the region
from 125 to 500 Hz and conservatively assumed it
was due to sampling error that would disappear as
the audiograms of additional species became available.
However, as more audiograms were added to the sam-
ple, the gap persisted, and the distribution took on a
distinctly bimodal appearance. Fig. 5A shows the dis-
tribution of 60-dB low-frequency limits for mammals
(only audiograms conducted in air are shown because
of the di¤culty in equating air and water thresholds).
Of the 58 species for which low-frequency limits are
available, 38 have extended low-frequency hearing
with 60-dB hearing limits below 125 Hz (Fig. 6). An-
other 19 species have restricted low-frequency hearing
with 60-dB limits above 500 Hz. Only one species falls
within the two-octave gap from 125 to 500 Hz the
pocket gopher (Geomys bursarius), an animal whose
Fig. 4. Audiograms of the ¢ve rodents in this study. Note the di-
chotomy in low-frequency hearing. C, chipmunk; G, groundhog ;
H, hamster; L, Darwin's leaf-eared mouse; S, spiny mouse.
Fig. 5. (A) Bimodal distribution of mammalian low-frequency hear-
ing limits (lowest frequency audible at 60 dB SPL). Note that only
audiograms conducted in air are represented. Histogram bin widths
are 2/3 octave. (B) Distribution of mammalian high-frequency hear-
ing limits (highest frequency audible at 60 dB SPL). The distribu-
tion appears to be unimodal with slight skew towards the low-fre-
quencies due primarily to the subterranean animals which have
relinquished high-frequency hearing (cf. Fig. 3). Histogram bin
widths are 1/3 octave.
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152 145
hearing is adapted to a subterranean environment
(He¡ner and He¡ner, 1990). In contrast to the bimodal
distribution of low-frequency limits, no similar dichot-
omy is apparent in the distribution of high-frequency
hearing limits (Fig. 5B). Thus, it appears that mammals
can be divided into two groups based on their low-fre-
quency hearing limits.
It should be noted that, for purposes of comparison,
we have de¢ned the low-frequency hearing limit as the
lowest frequency audible at 60-dB SPL, although ani-
mals can actually hear lower frequencies at higher in-
tensities. For example, using an 80-dB de¢nition lowers
the low-frequency limit for the laboratory rat from 540
to 290 Hz (He¡ner et al., 1994) and for the chipmunk
from 39 Hz to 14 Hz. The 60-dB de¢nition was chosen
because few animals have been tested beyond this limit
and the lowest frequency audible at any intensity is
usually not known. However, the 60-dB limit is not
an unreasonable de¢nition of useful hearing because
the sounds an animal must hear to survive (i.e., sounds
produced by other animals) tend not to be very loud.
Nevertheless, the exact location of the gap between the
two distributions depends on the de¢nition of hearing
limits. Interestingly, using a higher intensity to de¢ne
hearing limits would increase the size of the gap. This is
because the two groups di¡er in the steepness of the
low-frequency slopes of their audiograms with the re-
stricted low-frequency animals having steeper slopes
than the extended low-frequency animals (a di¡erence
that can be seen in Fig. 4). This di¡erence is reliable as
demonstrated by a statistical comparison of the slopes
of the audiograms, as de¢ned by the di¡erence (in oc-
taves) between the lowest frequency audible at 30 dB
and at lowest audible at 60 dB (n= 58, P= 0.0001,
Mann^Whitney U). Thus, using a higher intensity to
de¢ne the hearing limits would increase the size of the
gap between the two groups while shifting it to slightly
lower frequencies.
4.3.2. Potential explanations of the dichotomy in
low-frequency hearing
One possible explanation for the dichotomy in low-
frequency hearing is that the two groups di¡er along
some dimension such as body size, phyletic lineage, or
lifestyle. However, there is no obvious feature that dis-
tinguishes the two groups (Fig. 6). Although the ani-
mals in the group with restricted low-frequency hearing
tend to be small (with the exception of the opossum),
the group with extended low-frequency hearing also
contains small animals (e.g., gerbil, mole rats, least wea-
sel). In addition, although the species belonging to a
particular Order tend to fall into one group or the
other, rodents are found in both, thus making a phylo-
genetic division di¤cult. Nor is the dichotomy based on
trophic level as predators and prey can be found in
both groups. Similarly, the two groups do not divide
along other lines such as type of habitat or activity
cycle (e.g., nocturnal vs. diurnal). Thus, at this time
we are unable to account for the dichotomy in low-
Fig. 6. Distribution of mammalian low-frequency hearing limits
(lowest frequency audible at 60 dB SPL) with individual species in-
dicated. Note that rodents, shown on the left, fall into both the ex-
tended and restricted low-frequency hearing groups (for references
to individual audiograms, see Koay et al., 1998).
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152146
frequency hearing in terms of morphology, phylogeny,
or ecology.
Another possibility is that the two groups di¡er in
the mechanisms they use to perceive the pitch of low-
frequency sounds. Brie£y, there are two di¡erent neural
mechanisms that may underlie pitch (e.g., Moore, 1993 ;
Wever, 1949). In one, frequency is encoded by temporal
mechanisms that are based on phase-locking. Here
nerve ¢ring tends to occur at a particular phase of the
stimulating waveform, and the intervals between succes-
sive neural impulses are thus a multiple of the tone
period (1/frequency). However, temporal coding is lim-
ited to low frequencies because phase locking declines
as frequency increases (e.g., Rose et al., 1967). In the
second, higher frequencies are encoded by a spatial or
place mechanism in which tones of di¡erent frequencies
excite hair cells and ¢bers at di¡erent locations along
the basilar membrane. However, the actual frequency
ranges over which either the temporal or the place
mechanisms are dominant for the perception of pitch
have not been agreed upon in theory nor determined in
practice. Some observations suggest that the upper limit
of the temporal mechanism for the perception of pitch
is 4^5 kHz (e.g., Moore, 1993). However, as described
below, there is also reason to believe that the upper
limit of the temporal mechanism for pitch perception
may be much lower.
4.3.2.1. Upper limit in the use of temporal information
for pitch perception. The predominant view, summa-
rized by Moore (1993, 1997), is that the upper limit
of temporal coding for the perception of pitch is about
5 kHz. Evidence for this limit includes the following
observations: (1) the upper limit of neural phase lock-
ing in the squirrel monkey auditory nerve is 4^5 kHz
(Rose et al., 1967); (2) human frequency di¡erence
limens for tone bursts increase above 5 kHz, an obser-
vation consistent with the belief that place coding of
frequency is less precise than temporal coding (e.g.,
MÖller, 2000); (3) humans have no clear sense of mel-
ody in tones above 5 kHz; and (4) the residue pitch or
missing fundamental resulting from combination tones
is only observed when the combination tones are below
about 5 kHz.
Central to this view is the assumption that the upper
limit of phase locking in the human auditory nerve is
the same as in the squirrel monkey. Although initially a
reasonable assumption, it is now apparent that the
upper limit of phase locking varies between species.
For example, phase locking in the guinea pig begins
to decline at about 600 Hz and is no longer detectable
above 3.5 kHz, which is almost one octave lower than
in the squirrel monkey (Palmer and Russell, 1986).
Moreover, it is di¤cult to determine the highest fre-
quency for which neural phase locking is actually
used: Is it the highest frequency at which phase locking
can be detected or the frequency at which it begins to
decline?
The lack of a ¢rm value for the upper limit of phase
locking in humans weakens the psychophysical evidence
regarding the upper limit of temporal coding. The ob-
servation that human frequency di¡erence limens in-
crease above 5 kHz and that humans have no clear
sense of melody above 5 kHz may be due to a lack of
selective pressure to maintain these abilities above 5 kHz
rather than to any inherent limitation of the place
mechanism. Similarly, it is possible to explain the resi-
due pitch in terms of a temporal or a place mechanism
^ it is the correspondence of the upper limit of this
phenomenon with the upper limit of phase locking in
the squirrel monkey auditory nerve that makes a tem-
poral explanation attractive (Moore, 1997). In short, we
do not know if the human auditory nerve phase locks
to 5 kHz and therefore cannot be con¢dent that the
temporal coding of frequency extends to 5 kHz.
On the other hand, there are at least two observa-
tions suggesting a more restricted upper limit for the
role of temporal mechanisms in pitch. The ¢rst has to
do with the perception of the pitch of click trains con-
structed so that they can potentially be matched in
pitch either on the basis of pulse rate, which is a tem-
poral basis, or on the basis of fundamental frequency,
which is interpreted as a spatial or place basis. For click
trains composed of all positive or all negative clicks,
pulse rate and repetition rate are identical. However,
pulse rate and repetition rate di¡er when patterns of
alternating positive and negative clicks are used. For
example, a train of alternating positive and negative
clicks has a pulse rate twice its repetition rate. When
asked to match the pitch of two click trains of less than
100 pulses per second (pps), one with uniform-polarity
clicks and the other with a pattern of positive and neg-
ative clicks, subjects match them by setting the trains to
equal pulse rates, a temporal basis, regardless of di¡er-
ences in fundamental frequency (Flanagan and Gutt-
man, 1960). However, at pulse rates of 200 pps or
more, subjects adjust click trains by equating their fun-
damental frequencies, regardless of di¡ering pulse rates,
which was interpreted as evidence for place coding of
frequency. This result suggests that the upper limit of
the temporal mechanism for pitch perception may be
between 100 and 200 Hz. Consistent with this result is
the physiological ¢nding that there are two forms of
click train encoding in auditory cortex, one for click
trains below 1003200 pps, which is independent of
click polarity, the other for higher click train rates,
which is dependent on click polarity (Steinschneider et
al., 1998)
A second line of evidence comes from psychophysical
studies of cochlear implant patients. These studies in-
dicate that pitch changes with the frequency of stimuli
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152 147
only up to about 300 Hz (e.g., Shannon, 1983). Taken
together, the observations of Flanagan and Guttman
and those of Shannon suggest an upper limit of the
temporal mechanism for pitch perception of around
200^300 Hz. Because this upper limit corresponds to
the gap in mammalian low-frequency limits, it suggests
the possibility that the animals in the group with re-
stricted low-frequency hearing may not use temporal
coding for pitch perception.
4.3.2.2. Some mammals do not use temporal coding
for pitch perception. Some mammals do not use tem-
poral coding for pitch perception because they do not
hear low enough for phase locking to occur. For exam-
ple, the low-frequency hearing limit of the little brown
bat (10.3 kHz) is too high for phase locking, and there-
fore temporal encoding, to occur (Fig. 6). The same
conclusion may hold for other mammals that do not
hear low frequencies, such as the big brown bat and
marmosa opossum, which do not hear below 4 kHz.
Thus, the question arises as to which other mammals
might not use temporal encoding for pitch perception.
The correspondence of the gap in mammalian low-
frequency hearing limits with the evidence for a 200^
300-Hz upper limit for the temporal mechanism for
pitch perception suggests the possibility that the two
phenomena may be related. Speci¢cally, whereas mam-
mals with extended low-frequency hearing are probably
using both the place and temporal mechanism for pitch
perception, those with restricted low-frequency hearing
may be using only the place mechanism.
It should be noted that neural phase locking may still
occur in animals in the restricted low-frequency hearing
group because it is used in sound localization to extract
the binaural phase-di¡erence cue. Thus, it is possible
that the upper limit of phase locking in some animals
is determined by its use for sound localization rather
than by its use for pitch perception. Sound localization
studies have indicated that the upper limit of the use of
binaural phase di¡erence in mammals varies with the
size of an animal's head with small animals using it at
higher frequencies than large animals (Brown, 1994) ^
the current upper limits extend from 500 Hz for cattle
to 6.3 kHz for the Jamaican fruit bat, although some
very small mammals, such as the big brown bat, are
unable to use the binaural phase cue at all (He¡ner et
al., 1999, 2000).
Because some animals do not use phase locking for
sound localization and others may not use it for pitch
perception, the following combinations may occur (can-
didate species are listed in parentheses) : (1) Animals
that use phase locking in both pitch perception and
sound localization (humans, cats, and chinchillas).
(2) Animals that use phase locking in sound localization
but not in pitch perception (laboratory rat, Egyptian
and Jamaican fruit bats). (3) Animals that use phase
locking in pitch perception but not for sound localiza-
tion, a situation that may be found in subterranean
mammals that have good low-frequency hearing, but
have lost the ability to localize brief sounds (gophers
and mole rats). (4) Animals that do not use phase lock-
ing in either pitch perception or sound localization (big
brown bat).
If the restricted and extended low-frequency animals
di¡er in terms of the mechanisms used to encode fre-
quency, then they might also be expected to show other
di¡erences in auditory processing, particularly in the
cochlea. Indeed, frequency maps of the cochlea, which
show the projection of frequency along the cochlear
partition, indicate that mammals fall into di¡erent
groups as de¢ned by the normalized slope of the rela-
tionship (i.e., by the coe¤cient of the exponential term
when cochlear distance is expressed as a proportion or
percentage of total partition length ; Greenwood, 1996).
In one such group, which includes the Norway rat and
the marsupial, Monodelphis domestica, octaves subtend
a larger percentage of the cochlear length than they do
in a second group, which includes human, cat, guinea
pig, chinchilla, monkey, and gerbil (Greenwood, 1996 ;
Mu
«ller, 1996). As can be seen in Fig. 6, the animals in
the ¢rst group have restricted low-frequency hearing
while those in the second group all have extended
low-frequency hearing. Thus, whether or not an animal
uses temporal coding for pitch may be re£ected in its
cochlear frequency map. However, whether an animal
uses temporal and/or place coding for pitch perception
must ultimately be determined behaviorally.
4.4. Relation between high- and low-frequency hearing
In attempting to explain the variation in low-fre-
quency hearing, previous studies identi¢ed two poten-
tial explanatory factors : functional head size and high-
frequency hearing limit (He¡ner and He¡ner, 1985 ;
He¡ner and Masterton, 1980). These studies found
that although both factors are signi¢cantly correlated
with low-frequency hearing limit, high-frequency hear-
ing is directly correlated with low-frequency hearing
while functional head size is indirectly correlated with
low-frequency hearing through its correlation with
high-frequency hearing. Thus, animals with good
high-frequency hearing generally have poor low-fre-
quency hearing. With the addition of the ¢ve rodents
in this report, as well as the division of mammals into
two groups based on low-frequency hearing, it is useful
to reexamine this relationship.
The relationship between high- and low-frequency
hearing in mammals was examined separately for ani-
mals with extended and restricted low-frequency hear-
ing, de¢ned by whether an animal's 60-dB low-fre-
quency limit was below 250 Hz. Functional head size
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152148
was de¢ned as the shortest distance around the head
from one auditory meatus to the other. Aquatic audio-
grams were excluded because of the di¤culty in equat-
ing underwater thresholds with those obtained in air.
Subterranean mammals were also excluded from the
statistical analysis (but are depicted in the graphical
presentation) because they have vestigial high-frequency
hearing.
Among mammals with restricted low-frequency hear-
ing, high- and low-frequency hearing are reliably corre-
lated (n= 18, r= 0.691, P= 0.0015). The slope of the
regression line is relatively steep with a tradeo¡ of
1.71 octaves of low-frequency hearing for each octave
of high-frequency hearing gained or lost (Fig. 7). Partial
correlation analysis indicates that this relationship re-
mains signi¢cant when functional head size is held
constant (P= 0.022). On the other hand, although
head size is reliably correlated with low-frequency hear-
ing (r=30.562, P= 0.015), this correlation falls to
chance when high-frequency hearing is held constant
(P= 0.246).
For mammals with extended low-frequency hearing,
high- and low-frequency hearing are less closely corre-
lated, although the relationship is still strong (n= 33,
r= 0.567, P= 0.0006). The slope of the regression line
for this group is also less steep showing a tradeo¡ of
only 0.72 octaves of low-frequency hearing for each
octave of high-frequency hearing gained or lost (Fig.
7). Partial correlation analysis indicates that this rela-
tionship remains signi¢cant when functional head size is
held constant (P= 0.0007). On the other hand, function-
al head size is not reliably correlated with low-fre-
quency hearing even without partialling out high-fre-
quency hearing (r=30.293, P= 0.0873).
These results indicate that high-frequency hearing is
related to low-frequency hearing for both groups of
animals. Functional head size, on the other hand, is
indirectly related to low-frequency hearing through its
correlation with high-frequency hearing (but only for
the group with restricted low-frequency hearing).
Thus, in attempting to explain the variation in low-fre-
quency hearing, only high-frequency hearing need be
considered as head size appears to play no direct role.
High-frequency hearing accounts for 46% and 30% of
the variance in low-frequency hearing for the restricted
and extended low-frequency hearing groups, respec-
tively, indicating that it is a major factor in£uencing
low-frequency hearing. In seeking reasons why high-
and low-frequency hearing are related, it is necessary
to consider both anatomical and functional factors.
4.4.1. Anatomical considerations in the relation between
high- and low-frequency hearing
One possible explanation of the relationship between
Fig. 7. Relation between the highest and lowest frequencies audible at 60 dB SPL for 51 species of mammals. Correlations are shown sepa-
rately for animals with restricted and extended low-frequency hearing (upper and lower regression lines, respectively). Subterranean rodents
(blind mole rat, naked mole rat, and pocket gopher) are depicted, but not included in the calculations. Open circles indicate the ¢ve species in
this study. Echolocating bats indicated by the letter e.
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152 149
high- and low-frequency hearing is that although the
mammalian ear can be adapted to hearing very high
or very low frequencies, no single ear can e¤ciently
transduce and encode both. Such an incompatibility
could arise in the middle ear if the mechanical con¢g-
urations that are e¤cient at transmitting low frequen-
cies are not e¡ective at high frequencies (e.g., Fleischer,
1978; Nummela, 1999; Rosowski, 1992). Alternatively,
there may be morphological constraints in the mamma-
lian basilar membrane such that species cannot hear
both high and low frequencies, at least not without
loss of overall sensitivity (Hemila et al., 1995; Numme-
la, 1999; West, 1985).
Attractive as these hypotheses may be, the idea that
mammals cannot hear well at both high and low fre-
quencies is contradicted by the existence of species that
do. Animals that hear in the top quartile for both high-
and low-frequency hearing include the least weasel (50
Hz to 60 kHz), domestic cat, (55 Hz to 79 kHz), and
bushbaby (92 Hz to 65 kHz). Thus, the implication that
hearing range should be relatively constant across spe-
cies is not supported. Moreover, contrary to expecta-
tions (Hemila et al., 1995 ; Nummela, 1999), broad
hearing ranges are not achieved at the expense of sen-
sitivity as hearing range and best sensitivity are not
signi¢cantly correlated (n= 55, r=30.219, P= 0.1351).
Thus, the evidence so far indicates that the variation in
high- and low-frequency hearing is not due to anatom-
ical or physiological constraints in the mammalian ear,
but is instead determined by what animals need to hear
in order to survive, i.e., by selective pressure.
This is not to say that the anatomical characteristics
of the ear have no e¡ect on an animal's hearing. On the
contrary, all of the animals with restricted low-fre-
quency and good high-frequency hearing that have
been examined (Virginia opossum, house mouse, Nor-
way rat, horseshoe bat, and Egyptian fruit bat) have
`microtype' middle ears with low compliance and a rel-
atively small incus making them best suited to transmit
high frequencies (Fleischer, 1978 ; Rosowski, 1992).
Similarly, all of the animals with extended low-fre-
quency hearing that have been examined are known
to have middle ears described as either freely mobile
and compliant, making them well-suited to transmit
low frequencies (guinea pig, chinchilla, kangaroo rat,
human, macaques, gerbil, weasels, and chimpanzee) or
as intermediate between the two types of ears (horse,
cat, bushbaby, and tree shrew) (Fleischer, 1978 ; Ro-
sowski, 1992). However, it is important to note that
the structure of the ear is ultimately determined by
what an animal needs to hear in order to survive. The
idea that the hearing ability of an animal is determined
by the size of its ear which, in turn, is determined by the
size of its head is contradicted by the existence of small
mammals with good low-frequency hearing (e.g., gerbil
and least weasel). Furthermore, there is no obvious
physical factor that prevents large animals from having
ears suited for high-frequency hearing, as demonstrated
by the domestic cat. It seems that animals that require
good high- or low-frequency hearing evolve the type of
ear capable of transducing the sounds they need in or-
der to survive. To state that an animal hears low fre-
quencies because it has an ear suitable for transmitting
low frequencies addresses the question of how, but not
why, it hears as it does.
4.4.2. Functional considerations in low-frequency hearing
The existence of animals with restricted low-fre-
quency hearing suggests that they either have no need
to hear low frequencies or that good low-frequency
hearing might actually be detrimental. With regard to
the ¢rst possibility, the fact that most vertebrates hear
low frequencies demonstrates that it has advantages.
Indeed, not only would low-frequency hearing be ex-
pected to aid the ability of an animal to detect many,
if not most, sounds made by other animals, but low
frequencies travel farther and could alert an animal to
more distant danger. It is possible that very early mam-
mals retained reptilian low-frequency hearing while
adding the ability to hear high frequencies (Manley,
2000), which is the condition found so far in most ex-
tant mammals. Because low-frequency sensitivity is
nearly universal among vertebrates, there is every rea-
son to believe that it has strong adaptive value.
On the other hand, it is conceivable that good
low-frequency hearing might be detrimental in situa-
tions where low-frequency sounds might mask or
otherwise interfere with the analysis of high-frequency
sounds. Indeed, it is well known that a sound has
a greater masking e¡ect on higher than lower frequen-
cies (Wegel and Lane, 1924). Thus, animals in the re-
stricted low-frequency group may be giving up 1.71
octaves of low-frequency hearing for each octave in-
crease in high frequency to prevent low frequencies
from masking the high frequencies they use to localize
sound. Indeed, it has been noted that animals often
localize high-pass noise more accurately than broad-
band noise (He¡ner et al., 1995). Animals in the
extended low-frequency group may additionally be
a¡ected by the high levels of ambient low-frequency
noise that exist in the environment (Martin, 1984) and
by a lack of biologically relevant sounds at very low
frequencies. As a result, for each octave of high-fre-
quency hearing which they relinquish because it is no
longer required for sound localization, their low-fre-
quency hearing is extended only 0.72 octaves into the
low frequencies. Thus, the advantages of being able to
detect low frequencies must be weighed against the
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152150
disadvantage that they may mask important higher
frequency sounds.
Acknowledgements
We thank and J. Conesa, C. Contos, K. Flohe, and
S. Mooney for their help in these experiments. In addi-
tion, we thank D. Greenwood for bibliographic refer-
ences and useful comments on a previous draft. Sup-
ported by NIH Grants R01 DC02960, R21 DC03258,
and BNSF Grant 95-00188.
References
Bitter, K.S., He¡ner, R.S., Koay, G., He¡ner, H.E., 2001. Behavioral
audiogram of the short-tailed fruit bat, Carollia perspicillata. ARO
Abstr. 24, 63^64.
Brown, A.M., 1970. Bimodal cochlear response curves in rodents.
Nature 228, 576^577.
Brown, C.H., 1994. Sound localization. In: Fay, R.R., Popper, A.N.
(Eds.), Comparative Hearing: Mammals. Springer, New York, pp.
57^96.
Butler, R.A., 1975. The in£uence of the external and middle ear on
auditory discriminations. In: Keidel, W.D., Ne¡, W.D. (Eds.),
Handbook of Physiology, Auditory System, Vol. V/2. Springer,
New York, pp. 247^260.
Butler, R.A., 1999. The unfolding of an auditory illusion. Perspect.
Biol. Med. 42, 157^173.
Flanagan, J.L., Guttman, N., 1960. On the pitch of periodic pulses.
J. Acoust. Soc. Am. 32, 1308^1319.
Fleischer, G., 1978. Evolutionary principles of the mammalian middle
ear. Adv. Anat. Embryol. Cell Biol. 55, 1^70.
Floody, O.R., 1979. Behavioral and physiological analyzes of ultra-
sound production by female hamsters (Mesocricetus auratus). Am.
Zool. 19, 443^455.
Greenwood, D.D., 1996. Comparing octaves, frequency ranges, and
cochlear-map curvature across species. Hear. Res. 94, 157^162.
He¡ner, H.E., He¡ner, R.S., 1995. Conditioned avoidance. In:
Klump, G.M., Dooling, R.J., Fay, R.R., Stebbins, W.C. (Eds.),
Methods in Comparative Psychoacoustics. Birkhau
«ser, Basel, pp.
73^87.
He¡ner, H.E., He¡ner, R.S., 1998. Hearing. In: Greenberg, G., Har-
away, M.M. (Eds.), Comparative Psychology, A Handbook. Gar-
land, New York, pp. 290^303.
He¡ner, H.E., He¡ner, R.S., Contos, C., Ott, T., 1994. Audiogram of
the hooded Norway rat. Hear. Res. 73, 244^247.
He¡ner, H., Masterton, B., 1980. Hearing in glires: Domestic rabbit,
cotton rat, feral house mouse, and kangaroo rat. J. Acoust. Soc.
Am. 68, 1584^1599.
He¡ner, R.S., He¡ner, H.E., 1985. Hearing in mammals: The least
weasel. J. Mammal. 66, 745^755.
He¡ner, R.S., He¡ner, H.E., 1990. Vestigial hearing in a fossorial
mammal, the pocket gopher (Geomys bursarius). Hear. Res. 46,
239^252.
He¡ner, R.S., He¡ner, H.E., 1992. Hearing and sound localization in
blind mole rats, Spalax ehrenbergi. Hear. Res. 62, 206^216.
He¡ner, R.S., He¡ner, H.E., 1993. Degenerate hearing and sound
localization in naked mole rats (Heterocephalus glaber), with an
overview of central auditory structures. J. Comp. Neurol. 331,
418^433.
He¡ner, R.S., He¡ner, H.E., Koay, G., 1995. Sound localization in
chinchillas. II: Front/back and vertical localization. Hear. Res. 88,
190^198.
He¡ner, R.S., Koay, G., He¡ner, H.E., 1996. Sound localization in
chinchillas. III: E¡ect of pinna removal on sound localization.
Hear. Res. 99, 13^21.
He¡ner, R.S., Koay, G., He¡ner, H.E., 1999. Sound localization in an
old-world fruit bat Rousettus aegyptiacus: Acuity, use of binaural
cues, and relationship to vision. J. Comp. Psychol. 113, 297^
306.
He¡ner, R.S., Koay, G., He¡ner, H.E., 2000. Sound localization in a
new-world frugivorous bat, Artibeus jamaicensis: Acuity, use of
binaural cues, and relationship to vision. J. Acoust. Soc. Am.
109, 412^421.
Hemila, S., Nummela, S., Reuter, T., 1995. What middle ear param-
eters tell about impedance matching and high frequency hearing.
Hear. Res. 85, 31^44.
Koay, G., He¡ner, R.S., He¡ner, H.E., 1998. Hearing in a megachir-
opteran fruit bat Rousettus aegyptiacus. J. Comp. Psychol. 112,
371^382.
Manley, G.A., 2000. Cochlear mechanisms from a phylogenetic view-
point. Proc. Natl. Acad. Sci. USA 97, 11736^11743.
Martin, G.R., 1984. An environmental limit on absolute auditory
sensitivity in non-aquatic vertebrates. Behav. Process. 9, 205^
221.
Masterton, B., He¡ner, H., Ravizza, R., 1969. The evolution of hu-
man hearing. J. Acoust. Soc. Am. 45, 966^985.
MÖller, A.R., 2000. Hearing. Academic Press, New York.
Moore, B.C.J., 1993. Frequency analysis and pitch perception. In:
Yost, W.A., Popper, A.N., Fay, R.R. (Eds.), Human Psychophy-
sics. Springer, New York, pp. 56^115.
Moore, B.C.J., 1997. An Introduction to the Psychology of Hearing,
4th edn. Academic Press, London.
Mu
«ller, M., 1996. The cochlear place-frequency map of the adult and
developing mongolian gerbil. Hear. Res. 94, 148^156.
Musicant, A.D., Butler, R.A., 1984. The in£uence of pinnae-based
spectral cues on sound localization. J. Acoust. Soc. Am. 75,
1195^1200.
Nummela, S., 1999. Scaling and modelling of the mammalian middle
ear. Doctoral Dissertation presented to the Faculty of Science of
the University of Helsinki, Helsinki.
Palmer, A.R., Russell, I.J., 1986. Phase-locking in the cochlear nerve
of the guinea-pig and its relation to the receptor potential of inner
hair-cells. Hear. Res. 24, 1^15.
Pepper, J.W., Braude, S.H., Lacey, E.A., Sherman, P.W., 1991. Vocal-
izations of the naked mole-rat. In: Sherman, P.W., Jarvis, J.U.M.,
Alexander, R.D. (Eds.), The Biology of the Naked Mole-Rat,
Princeton University Press, Princeton, NJ, pp. 243^276.
Ro¥er, S.K., Butler, R.A., 1968. Factors that in£uence the localiza-
tion of sound in the vertical plane. J. Acoust. Soc. Am. 43, 1255^
1259.
Rose, J.E., Brugge, J.F., Anderson, D.J., Hind, J.E., 1967. Phase-
locked response to low-frequency tones in single auditory nerve
¢bers of the squirrel monkey. J. Neurophysiol. 30, 769^793.
Rosowski, J.J., 1992. Hearing in transitional mammals : Predictions
from the middle-ear anatomy and hearing capabilities of extant
mammals. In: Popper, A.N., Fay, R.R., Webster, D.B. (Eds.), The
Evolutionary Biology of Hearing. Springer, New York, pp. 615^
631.
Shannon, R.V., 1983. Multichannel electrical stimulation of the audi-
tory nerve in man. I. Basic psychophysics. Hear. Res. 11, 157^
189.
Shaw, E.A.C., 1974. The external ear. In: Keidel, W.D., Ne¡, W.D.
(Eds.), Handbook of Sensory Physiology, Vol. V(1). Springer,
New York, pp. 450^490.
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152 151
Steinschneider, M., Reser, D.H., Fishman, Y.I., Schroeder, C.E.,
Arezzo, J.C., 1998. Click train encoding in primary auditory cor-
tex of the awake monkey: Evidence for two mechanisms sub-
serving pitch perception. J. Acoust. Soc. Am. 104, 2935^2955.
Wegel, R.L., Lane, C.E., 1924. The auditory masking of one pure
tone by another and its probable relation to the dynamics of the
inner ear. Phys. Rev. 23, 266^285.
West, C.D., 1985. The relationship of the spiral turns of the cochlea
and the length of the basilar membrane to the range of audible
frequencies in ground dwelling mammals. J. Acoust. Soc. Am. 77,
1091^1101.
Wever, E.G., 1949. Theory of Hearing. Wiley, New York.
Wotton, J.M., Haresign, T., Simmons, J.A., 1995. Spatially dependent
acoustic cues generated by the external ear of the big brown bat,
Eptesicus fuscus. J. Acoust. Soc. Am. 98, 1423^1455.
HEARES 3706 12-7-01
R.S. He¡ner et al. / Hearing Research 157 (2001) 138^152152
