AMER. ZOOC, IS: 1215-1226 (1973).
Mechanisms of Sound Production by Echolocating Bafs
RODERICK A. SUTHERS AND JAMES M. FATTU
Department of Anatomy and Physiology, Indiana University,
Bloomington, Indiana 47401
SYNOPSIS. The echolocaitive pulses emitted at high repetition rates 'by bats pose a number of questions regarding the mechanism oE their production. We have investigated
some physiologic parameters of vocalization in the North American Vespertilionid bat
Eptesicus fuscus, including pulse insertion within the respiratory cycle and subglottic pressure changes accompanying these intense sounds. Sutoglottic pressures are
considerably higher than those characteristic of human speech and, therefore, require a substantial structure to act as a glottal stop. We suggest that this is the
role of the vocal fold and that the thin, paired vocal and ventricular membranes are
the ultrasonic generators. The cricothyroid muscle, which is thought to influence
sound frequency by altering the tension on these membranes, contracts just prior
to each ultrasonic vocalization and relaxes during phonation. Cricothyroid muscle
relaxation may gradually decrease the tension on the membranes and create the
downward frequency sweep characteristic of most pulses.
Although the ability of bats to orient
acoustically by echolocation has been
known for many years, little is yet known
about the laryngeal mechanisms by which
the ultrasonic orientation sounds or pulses
are produced. The physiology of phonation in echolocating bats becomes particularly interesting when one considers the
nature of the echolocative sounds.
Common North American species of the
family Vespertilionidae, for example, produce pulses having a duration from less
than one to several msec. The initial frequency is between 60 and 80 kHz but
sweeps downward about an octave within
the brief duration of the pulse, the intensity of which often exceeds 110 db in reference to 0.0002 dynes/cm2 measured 10
to 15 cm in front of the bat's mouth. During straight flight in an uncluttered environment, orientation pulses are normally emitted at a rate of several per second.
This repetition rate increases dramatically,
however, when landing, negotiating obstacles in the flight path, or pursuing insect prey. Maximum pulse repetition rates
of about 200/sec are characteristic of these
We thank Dr. H. David Potter for kindly photographing histologic sections and Dr. James Sidie
for the opportunity to examine his data relative
to this paper.
situations. This is indeed an impressive
vocal performance and poses many interesting questions regarding
laryngeal
mechanisms and their control.
We wish here to provide a preliminary
report of some of our recent research on
the mechanism of vocalization by the insectivorous big brown bat Eptesicus fuscus
(Vespertilionidae).
The anatomy of the microchiropteran
larynx has been described by Elias (1908),
Fischer and Vomel (1961), and Fischer and
Gerken (1961), yet our present physiological investigation has been hampered by the
lack of a more detailed knowledge of the
functional anatomy. Salient morphological
•features of the microchiropteran larynx
include unusually large cricoid and thyroid cartilages. The cricothyroid muscle,
which bridges the joint between these cartilages, is also greatly hypertrophied and
envelopes the whole larynx, hiding all the
other intrinsic muscles from view.
Fischer and Vomel (1961) and Fischer
and Gerken (1961) studied the laryngeal
anatomy of Myotis myotis and hypothesized that the caudal lip of the ventricle
acted as a whistle with audible components
being filtered out by the vocal tract. They
contended that phonation was turned on
and off by adduction and abduction, re-
1215
1216
RODERICK A.
SUTHERS AND JAMES i\f.
FATTU
50 MS
FIG. 1. Oscillographic display of echolocative
pulses emitted by Eplesicus fuscus superimposed
on output of therrnislor in front ot animals mouth.
Expiration is accompanied by an upward slope of.
Ihe thermistor trace. Pulses are often emitted near
the reversal point between inspiration and expiration.
spectively, of the vocal folds under control
of the vocalis muscle, rather than by rotation of the arytenoid cartilages, which
they considered to be relatively immobile
in the bat.
Griffin and Novick (Griffin, 1958; Novick and Griffin, 1961) suggested an alternate, and we believe more likely, mechanism of sound production. They observed
that in Myolis lucifugus and Eplesicus
fuscus the vocal and ventricular folds are
modified to form two pairs of very thin
membranes which cover the anterior and
posterior edges of the ventricle to form a
miniature "drum head" on each side of
the ventricular lumen. Each membrane is
about 6 to 8 micra thick, 0.5 mm wide, and
2.0 mm long. These membranes are
stretched by the contraction of the cricothyroid muscles, and Griffin has estimated
that their resonant frequency might reasonably be expected to lie in the range of
frequency modulated (FM) orientation
pulses which he hypothesizes are generated by vibration of these membranes.
The vocal fold which vibrates in the human larynx is a relatively substantial structure containing a ligament and muscle fibers. It functions not only as a sound generator, but also as a source of airway resistance and forms the glottal stop. The
vocal folds thus play a vital role in controlling subglottic pressure during phonation. Subglottic pressure is one of the factors determining the intensity of the
sound. The fragile laryngeal membranes
in the bat described by Griffin could hardly sustain the relatively high subglottic
pressures one might expect to accompany
their very intense echolocative sounds.
We have, therefore, sought to determine
the subglottic pressures accompanying
phonation by Eptesicus and the structure
in the airway acting as the resistive ele-
FIG. 2. Fluctuations in subglottic pressure associated with the production of three FM pulses
at a low repetition rate of 10/sec. The first pulse
is shown on an expanded time base in center and
on a sonagraph at bottom. (A similar format is
used in Figs. 3-5; note time bars.) An increase
in subglottic pressure causes an upward deflection
of pressure trace. In this and the next three figures
the acoustic signal lags 0.5 msec behind the subglottic pressure since the microphone was 15 cm
in front of the bat.
FIG. 3. Fluctuations in subglottic pressure during
phonation of five FM pulses at a repetition rate of
about 50 pulses/sec. Pressure drop before first
pulse was caused by electrical stimulus to brain
which activated part of motor apparatus for vocalization but did not elicit actual sound production. Note ihe rapid rise of subglottic pressure
after each pulse in preparation for emission of
next pulse. Pressure is rising during last half of
pulses. Low frequency sinusoidal pressure oscillation during third pulse is associated with terminal
fricative portion of sonagram.
1217
BAT SOUND PRODUCTION
CN
OS
1 I I I
I
I I
CO
CN
CO
CN
CN
I •,
I
1218
RODERICK A. SUTHERS AND JAMES M. FATTU
ment which forms the glottal stop and creates these pressures. We have also investigated the way in which orientation pulses
are inserted into the respiratory cycle and
have obtained some preliminary electrophysiological data on the cricothyroid
muscles which, we believe, are relevant to
understanding their function.
ration depending on the position of the
stimulating electrodes. Phonation is usually accompanied by appropriate ear and lip
movements. Ultrasonic pulses similar to
those used in echolocation can be elicited
at various repetition rates by this technique.
METHODS
PLACEMENT OF PULSES IN THE RESPIRATORY
CYCLE
We measured the respiratory dynamics
of awake, restrained bats. Prior to the experiment, surgical procedures were performed under ether anesthesia. At this
time a T-cannula made of 18 gauge stainless steel tubing was inserted into the trachea three to five rings below the larynx.
An incision was also made in the scalp,
the muscles were reflected, and a portion
of the cranium was removed to expose the
inferior and superior colliculi. After the
bat had recovered from the effects of the
anesthetic, it was restrained by a head
holding device 15 cm in front of a quarterinch, calibrated Bruel and Kjaer condenser
microphone. The side arm of the tracheal
cannula was attached to a Pitran (model
PT-H2 MO4) pressure sensitive transistor, having a flat frequency response from
DC to 400 Hz and operated in a temperature-compensated circuit. The amplified
outputs of this pressure transducer and the
microphone were displayed on an oscilloscope and recorded on magnetic tape with
an instrumentation tape recorder (Precision Instrument Co., model PI6104). The
frequency response of the pressure recording system was relatively flat from DC to
10 kHz, and that of the audio channel was
flat from 200 Hz to 100 kHz.
The restrained, awake bat was induced
to vocalize by delivering brief trains of
electrical pulses to appropriate regions of
the midbrain, as described by Suga and
Schlegel (1972). When a pair of fine stimulating electrodes is inserted into a particular part of the anterior portion of the inferior colliculus, repetitive stimulation
with trains of pulses causes the animal to
vocalize. These vocalizations may be audible or ultrasonic and of lone or short du-
Respiration in an echolocating bat must
be responsive not only to the organism's
metabolic demands, which are extremely
high during flight, but also to the need
for continual vocalization. The rate at
which orientation pulses are emitted is determined by the animal's sensory needs
which vary independently from its metabolic needs. It is, therefore, pertinent to
know the extent to which pulse emission
is dependent on the respiratory cycle. The
respiratory cycle of restrained Eptesicus
was monitored by a small thermistor bead
(Fenwal GB34J3) positioned a millimeter
or two in front of the mouth. The output
of the thermistor was telemetered, displayed, and recorded as described by Suthers et al. (1972). An upward displacement
of the thermistor trace in Figure 1 is caused
by warming during expiration. The downward deflections indicate cooling due either to a momentary cessation of expiratory flow or inspiration. Although one
must use caution in interpreting records
such as this, Figure 1 shows that our restrained bats tended to emit sounds at the
reversal point between the end of expiration and beginning of inspiration. At higher pulse repetition rates the inspiratory
phase may be interrupted by a brief expiration coinciding with the production of
an additional orientation pulse.
In some experiments bats were placed
in a body plethysmograph making it possible to measure respiratory tidal volume.
The tidal volume was typically about 0.3
cc during silent respiratory cycles and those
containing only one FM pulse. When several pulses were emitted during a respiratory cycle, as occurs during high pulse
BAT SOUND PRODUCTION
1219
BOi-
01-
10 1 8
FIG. 4. Subglottic pressure during pulses of relatively long duration and limited FM. Subglottic
pressure drops gradually during pulse but remains
positive.
repetition rates, the tidal volume was increased immediately prior to and following the burst of pulses. Similarly, audible
pulses with 5 to 10 times the duration of
FM pulses were associated with increased
tidal volumes of up to 0.6 cc or more.
Nearly all vocalizations appeared to be
produced during laryngeal air flow in the
expiratory direction. When several pulses
were produced during a single respiratory
cycle, some occurred during the basic inspiratory phase, but plethysmograph and
subglottic pressure measurements indicate
that such pulses were accompanied by
small expiratory maneuvers or minibreaths which momentarily interrupted the
respiratory cycle. In only one instance
have we observed phonation during what
clearly appears to be inspiratory flow. This
sound was one of a series of brief wide
spectrum "clicks" quite unlike normal
orientation pulses.
AIRWAY PRESSURES DURING VOCALIZATION
Air pressure changes during phonation
were measured in the nostrils, mouth, posterior pharynx, and trachea. Only the.tracheal pressure rises prior to sound production. Posterior pharyngeal pressure" rises
concurrently with the onset of phonation,
reaches about one-half the magnitude of
the tracheal pressure, and declines to ambient pressure following the vocalization.
Mouth pressure increases only during phonation and seldom exceeds 5 cm H 2 O. Na-
1220
RODERICK A.
SUTHERS AND JAMES M.
FATTU
•
(
i
RODERICK A. SUTHERS AND JAMES M.
FATTU
1221
FIG. 5. Subglottic pressures accompanying vocalizations containing prominent audible components
with wide frequency spectrum. Low frequency
pressure oscillation corresponds to maximum excursions of complex acoustic waveform.
FIG. 6. Section through larynx of Eptesicus fiis-
cus. Stained with haemotoxylin and eosin. A, vocal
membrane; B, vocal fold; C, ventricular membrane;
D, cricothyroid muscle; E, thyroarytenoid muscle;
F, ventricle of Morgagni; G, thyroid cartilage; H,
cricoid cartilage; I, arytenoid cartilage. Bar equals
500 micra.
sal pressure fluctuations are slight and
consist primarily of a small negative pressure developed shortly after phonation.
We shall hereafter refer to the tracheal
pressure as the subglottic pressure since
we believe it reflects the variable airway
resistance produced by the glottal stop.
Figure 2 shows subglottic pressure during
the production of three brief ultrasonic
FM orientation pulses emitted at a relatively low repetition rate of about 10
pulses/sec. A sound spectrograph of the
first pulse is presented in the lower part
of the figure to show its frequency structure. Subglottic pressure starts increasing
shortly after the onset of expiration and
rises during the 30 to 50 msec before the
pulse to a peak value between 25 and 40
cm H 2 O. It drops abruptly at the onset of
phonation and then more gradually during the terminal portion of the pulse. Subglottic pressure drops by about 20 cm H2O
during a typical FM pulse whose frequency
sweeps through 0.8 octave or more within
5 msec at a sound pressure level of 105 db
measured 15 cm from the mouth.
At higher pulse repetition rates the subglottic pressure fluctuations must be adjusted to accommodate several orientation
pulses in one respiratory cycle. Figure 3
shows subglottic pressure during a group
of five pulses at a repetition rate of about
50/sec. We have noted that tidal volume
may be doubled in anticipation of these
bursts of pulses. We have studied pulses
emitted at repetition rates up to 130/sec.
The transient drop in subglottic pressure
prior to this group is due to the fact that
brain stimulation often elicits many of the
motor patterns of vocalization prior to actual vocalization. Peak subglottic pressures
attain higher values during multiple pulse
respiratory cycles—often up to 60 or 70 cm
H2O—than during single pulse cycles. Sub-
glottic pressure commences to increase
with the onset of expiration, and large
transient pressure drops of 20 to 45 cm
H2O occur during the first 2 to 3 msec of
each pulse. Furthermore, the subglottic
pressure returns within 3 to 5 msec to
nearly its value at the onset of the pulse if
another pulse is to be produced; if not,
the pressure declines to normal respiratory
levels.
Stimulation often elicited longer duration pulses with only a little FM and possessing several harmonically related components, the lowest of which were often
audible. An example of these is shown in
FIG. 7. Vocal fold of Eptesicus fuscus with thin
vocal membrane extending into laryngeal ventricle.
Note the stratified squamous epithelium covering
the muscular hump of the vocal fold. Haemotoxylin and eosin stain. A, vocal membrane; B,
vocal fold. Bar equals 50 micra.
1222
RODERICK A. SI.'THKRS AND JAMES M. FATTU
SLN.
CTMV
50 MS
I'IG. 8. Nerve and muscle recordings during vocalizations induced by brain stimulation. SLN,
compound action potential of motor twig of su-
perior laryngeal nerve; CTiYr, electromyograph of
ventrolateial portion of cricothyroid muscle; V,
vocalizations.
Figure 4, which also illustrates that the
subglottic pressure, after a brief dip at the
onset of phonation, declines gradually until the end of the pulse.
Finally, there were vocalizations which
—though they sometimes began with discrete frequency components—contained
substantial portions with a wide-band,
poorly defined frequency spectrum, large
portions of which were audible (Fig. 5).
We have called these "protest cries" as
they sound similar to audible, rasping
squeaks commonly emitted voluntarily by
a hand-held, struggling bat. An interesting
aspect of these sounds is the appearance
of a subglottic pressure oscillation with a
period of approximately 4 kHz superimposed on a slowly declining subglottic pressure between 30 and 60 cm H2O at the
onset of phonation. Cry No. 2 is particularly interesting since it begins with a discrete FM component which abruptly
changes after 4 msec into a protest cry of
wider frequency spectrum. The change in
the nature of the sound is accompanied by
a transient increase in subglottic pressure
which then begins to oscillate with a period corresponding to that of the maximum
excursions of the acoustic waveform. It is
tempting to speculate that a new resistive
element is introduced into the airway during protest cries and that the 4 kHz oscillation in subglottic pressure reflects the
fundamental frequency of this acoustic
generator. Fundamental frequencies above
10 kHz which might be associated with ul-
trasonic pulses would not be reproduced
by the FM system we used to record subglottic pressure.
IVTORPHOLOGICAL EVIDENCE FOR SEPARATE
RESISTIVE AND VIBRATORY STRUCTURES
Analysis of subglottic pressure and sound
pressure levels in many pulses shows that
sound intensity is a function of subglottic
pressure. The high subglottic pressures we
have measured during phonation are probably an important adaptation for producing intense orientation pulses. The fragile
ventricular membranes described by Griffin (1958) can hardly be the resistive elements responsible for these subglottic pressures. Those membranes associated with
the vocal fold can be seen in our histological cross sections of the larynx (Fig. 6) to be
thinner (6 to 10 microns) and wider (200
microns in the midportion of the ventricle)
than those on the ventricular fold. Figure
6 also shows the hypertrophied cricothyroid muscle and portions of the thyroid
and cricoid cartilages. The thyroarytenoid
muscle is present inside the thyroid laminae.
A view, under higher magnification
(Fig. 7), of the vocal fold and vocal membrane demonstrates one possibility for
anatomical separation of the functions of
increasing airway resistance for pressure
build-up and of the vibration generating
the sound. The vocal fold immediately below the ventricle forms a muscular hump
covered by a stratified squamous epitheli-
BAT SOUND PRODUCTION
1223
50 MS
FIG. 9. Development of tension in cricothyroid
muscle immediately prior to ultrasonic pulse emission. Ventrolateral portion of cricothyroid appears
to be relaxing during phonation. Both bars equal
50 msec.
um as it is in human and other larynges.
A similar squamous epithelium is also
found at the cranial margin of the ventricle of most species, but in Eptesicus this is
replaced by the pseudostratified ciliated
columnar epithelium which, thus, lines all
the laryngeal cavity except the vocal folds.
One would expect squamous, rather than
ciliated, epithelium to form the laryngeal
lining in a region subjected to the mechanical stresses associated with the glottis. The muscular hump at the base of the
vocal membrane may thus be an important resistive element during phonation.
The pharyngeal constrictors above the
larynx may also be important elements
but our data do not yet permit us to evaluate their roles.
in which section of the recurrent nerves
usually distorts normal speech. Denervation of the cricothyroid muscles by cutting the superior laryngeal nerves, however, caused the frequencies of the emitted
pulses to be lowered considerably so that
echolocative sounds became clearly audible
and relatively constant in frequency. Pulse
duration was relatively unaffected, however. Novick and Griffin (1961) also recorded electromyograms from the region of
the cricothyroid muscles during phonation.
This electrical activity was correlated with
pulse emission, but it was not known if
the electrical activity was only from the
cricothyroid or also from adjacent muscles. These action potentials often began
about 20 msec before the onset of the pulse
and continued during the pulse, sometimes ending after the pulse. Action potentials in the motor branch of the superior
laryngeal nerve innervating the cricothyroid were also highly correlated with pulse
emission. Griffin and Novick observed
that one of the actions of the cricothyroid
upon electrical stimulation was to exert
tension on the two pairs of ventricular
membranes. They hypothesized that the
frequency of the orientation pulse was
controlled by the tension this muscle ex-
ACTIVATION OF THE CRICOTHYROID MUSCLE
DURING PHONATION
Griffin and Novick (Griffin, 1958; Novick and Griffin, 1961) demonstrated that
bilateral neurotomy of the recurrent laryngeal nerves, which innervate all the intrinsic laryngeal muscles except the cricothyroid, had little effect on the ultrasonic
pulses or audible vocalizations. This indicates the mechanism of phonation in bats is
different from that classically seen in man,
1224
RODERICK A.
SLTIIIKS AND [AMUS M.
FATTU
50 MS
FIG. 10. Temporal relationship between electromyograph (upper trace) and vocalizations (lower
trace) in diMcrenl portions of cricothyroid muscle.
See text.
erted on these membranes. This hypothesis is consistent with our finding that the
frequency of orientation pulses is not proportional to the subglottic pressure, i.e.,
some other factor such as variable tension
on the ventricular membranes must be involved in the control of frequency.
We are now utilizing the technique of
eliciting vocalizations by electrical stimuli
delivered to the brain to make a detailed
study of the patterns of excitation and contraction of various laryngcal muscles during phonation. We would like to describe
some of our preliminary results which have
a bearing on the possible role of the cricothyroid muscles in controlling pulse frequency. The larynx can be surgically exposed through a ventral incision in an
anesthetized bat, and recording electrodes
can be placed on various twigs of the
nerves innervating it and on the cricothyroid muscle.
Figure 8 shows the action potentials in
the large lateral ventral portion of the
cricothyroid together with the electrical ac-
tivity in the motor twig of the superior
laryngeal nerve supplying these fibers and
the associated vocalizations. Note that the
electrical activity begins about 50 msec
prior to vocalization and ends before the
vocalization starts. Thus, this portion of
of the cricothyroid apparently is not contracting during vocalization.
Often the whole larynx moves cranially
during phonation. One Eplesicus, however,
emitted ultrasonic pulses with no detectable "elevation" of the larynx. Jt was thus
possible to use an RCA 5734 force transducer triode to measure the actual contraction of the lateral ventral portion of the
cricothyroid muscle. Figure 9 shows that
tension increases steadily during the period
of electrical activity before phonation,
reaches its peak just prior to sound production and declines during emission of
the pulse. Although the vector in which
these tension measurements were made is
not quite parallel to the long axis of the
muscle fibers and although we have not
yet sufficient data to quantitatively com-
BAT SOUND PRODUCTION
1225
electrodes are again in the large lateral
ventral portion just discussed and electrical activity precedes each pulse. In B, the
recording electrodes were located ventromedially on a small portion at the posterior of the larynx. Here the electromyogram
appears 15 to 30 msec before each pulse,
continues during the pulse and for about
10 msec or more after the pulse is finished.
In C, the recording electrodes were kept
near the posterior of the larynx, but moved
a little more laterally. Electrical activity
here seems to be associated with bursts of
pulses in that it starts before a group of
several pulses is emitted and continues until after the last pulse in the group is finished. The functional significance of these
differences is not yet clear.
SPEED OF CONTRACTION OF THE BAT
CRICOTHYROID
FIG. 11. Mechanical response o£ cricothyroid
muscle to repetitive stimulation of motor branch
of superior laryngeal nerve. Upper trace illustrates
typical mechanical response. Lower trace gives
diagrammatic representation of biphasic compound muscle action potential following each
stimulus. The stimulus rates from top to bottom
are 30, 60, 130, and 230/sec. The horizontal bar
equals 20 msec, the vertical bar represents a tension of about 0.5 gram. (From unpublished data
of James Sidie.)
pare tension and sound frequency, the decline of tension during pulse emission supports Griffin's (1958) suggestion that by
varying the tension on the vocal folds the
cricothyroid controls the frequency of the
orientation pulse.
Although this may be an important
function of the cricothyroid muscle, it may
also play other roles in vocalization. This
muscle consists of several bundles and recordings of electrical activity suggest these
contract at different times relative to phonation. Figure 10 shows electromyograms
of three different parts of the cricothyroid
during phonation. In A, the recording
The cricothyroid muscle of bats seems
well adapted for the rapid contractions required of it if it is to follow pulse repetition rates up to 200/sec. The sarcoplasmic
reticulum is extremely well developed
(Revel, 1962). Sidie, working in our laboratory, has found that the mean contraction time of the ventrolateral portion of
the cricothyroid muscle in Eptesicus under
barbiturate anesthesia is 6.5 msec in response to electrical stimulation of the motor nerve. This is less than one-fourth the
contraction time of dog (Martensson and
Skoglund, 1964; Hast, 1966) or rabbit
(Hall-Craggs, 1968) cricothyroid and compares favorably with the fastest contracting vertebrate muscles, such as the toadfish
swimbladder (Skoglund, 1961). The total
contraction/relaxation time of the bat
cricothyroid is 12 to 16 msec, which is still
considerably longer than the 4 to 5 msec
interpulse interval at maximum pulse repetition rates. Using repetitive stimulation
of the motor nerve, Sidie has shown that
although a partial tetanus begins to develop at stimulation rates of 55 to 65/sec,
there is a partial relaxation between each
stimulus until a fusion frequency of 220
to 240/sec is reached (Fig. 11). Although
1226
RODERICK A. SUTHERS AND JAMES M. FATTU
one must be cautious in extrapolating
from contractile performance during electrical stimulation of the motor nerve to
that accompanying voluntary phonation,
Sidie's data suggest that partial muscle relaxation might still be responsible for frequency modulation of pulses having interpulse intervals as short as 5 msec.
The experiments described here shed
some light on the respiratory dynamics and
motor control of phonation by echolocating bats, but many important problems
remain unsolved. The use of brain stimulation to elicit vocalizations in restrained
animals provides a powerful tool through
which we hope to find answers to some of
these problems.
REFERENCES
Elias, H. 1908. Zur Anatomie des Kehlkopfes tier
Mikrochiropteran. Morphol. Jahrb. 37:70-119.
Fischer, H., and H. Gerken. 1961. Le larynx de la
chauve-souris (Myotis myolis) et le larynx humain. Ann. Oto-Laryngol. 78:577-585.
Fischer, H., and H. J. Vomel. 1961. Der Ultra-
schallapparat des Larynx von Myotis myotis.
Gegenbaurs Jahrb. Morphol. Mikr. Anat. Abt 1.
102:200-226.
Griflin, D. R. 1958. Listening in the dark. Yale
Univ. Press, New Haven.
Hall-Craggs, E. C. B. 1968. The contraction times
and enzyme activity of two rabbit laryngeal
muscles. J. Anat. 102:241-255.
Hast, M. H. 1966. Mechanical properties of the
cricothyroid muscle. Laryngoscope 76:537-548.
Martensson, A., and C. R. Skoglund. 1964. Contraction properties of intrinsic laryngeal muscles.
Acta Physiol. Scand. 60:318-336.
Novick, A., and D. R. Griffin. 1961. Laryngeal
mechanisms in bats for the production of orientation sounds. J. Exp. Zool. 148:125-145.
Revel, J. P. 1962. The sarcoplasmic reticulum of
the bat cricothyroid muscle. J. Cell. Biol. 12:
571-588.
Skoglund, C. R. 1961. Functional analysis of swimbladder muscles engaged in sound production
of the toadfish. J. Biophys. Biochem. Cytol. 10
(Suppl.) : 187-200.
Suga, N., and P. Schlegel. 1972. Neural attenuation of responses to emitted sounds in echolocating bats. Science 177:82-84.
Suthers, R. A., S. P. Thomas, and B. J. Suthers.
1972. Respiration, wing-beat and ultrasonic pulse
emission in an echolocating bat. J. Exp. Biol.
56:37-48.