Echolocation signals of wild Atlantic spotted dolphin
Whitlow W. L. Aua)
Marine Mammal Research Program, Hawaii Institute of Marine Biology, University of Hawaii,
P.O. Box 1109, Kailua, Hawaii 96734
Denise L. Herzingb)
Department of Biological Sciences, Florida Atlantic University, 777 Glades Road, Boca Raton,
?Received 27 April 2002; accepted for publication 9 September 2002?
An array of four hydrophones arranged in a symmetrical star configuration was used to measure the
echolocation signals of theAtlantic spotted dolphin ?Stenella frontalis? in the Bahamas. The spacing
between the center hydrophone and the other hydrophones was 45.7 cm. A video camera was
attached to the array and a video tape recorder was time synchronized with the computer used to
digitize the acoustic signals. The echolocation signals had bi-modal frequency spectra with a
low-frequency peak between 40 and 50 kHz and a high-frequency peak between 110 and 130 kHz.
The low-frequency peak was dominant when the signal the source level was low and the
high-frequency peak dominated when the source level was high. Peak-to-peak source levels as high
as 210 dB re 1 ?Pa were measured. The source level varied in amplitude approximately as a
function of the one-way transmission loss for signals traveling from the animals to the array. The
characteristics of the signals were similar to those of captive Tursiops truncatus, Delphinapterus
leucas and Pseudorca crassidens measured in open waters under controlled conditions. © 2003
Acoustical Society of America. ?DOI: 10.1121/1.1518980?
PACS numbers: 43.80.Ev, 43.80.Ka, 43.80.Jz ?FD?
The Bahama Islands are an archipelago in the tropical
West Atlantic east of Florida that are surrounded by deep
water while the Bahamas banks are relatively shallow
(?15 m). The banks are thick, submerged platforms of cal-
careous rock providing diverse habitats, including fringe and
patch reefs, atolls, grassy flats, and ledges. Since 1985, a
resident community of approximately 200 spotted and 200
bottlenose dolphins have been identified, sexed, and ob-
served in a variety of behavioral contexts in this area ?see
Fig. 1?. Basic life history and age class categories for the
Atlantic spotted dolphin, S. frontalis, have been described
?Herzing, 1997?. Underwater behavior and correlated sound
?narrow-band frequency ?20 kHz) have also been described
?whistles?, ?2? excitement/distress ?whistle-squawks?, ?3?
pursuit/herding ?buzzes?, ?4? aggression ?burst-pulses?, ?5?
group synchrony ?synch pulses?, ?6? interspecific interactions
?barks, screams, squawks?, ?7? nonvocal sounds ?tail-slaps?,
and ?8? foraging/feeding ?echolocation clicks? ?Herzing,
1996, 2000?. Previous studies on the behavior and sound
production of S. frontalis in captivity also exist ?Wood, 1953;
Caldwell and Caldwell, 1966, 1971; Caldwell et al., 1973?.
The echolocation signals used by S. frontalis in the Ba-
hama banks, measured on a broadband basis, will be consid-
ered in this paper. The echolocation characteristics of del-
phinid species have been studied primarily in captivity ?Au,
1993?. Measurements from stationary dolphins in captivity
have shown that echolocation clicks are emitted in a direc-
tional beam and signals measured off-axis are distorted with
respect to the signals measured along the major axis of the
beam. Therefore, it is very difficult to obtain accurate mea-
surements of free-ranging, fast moving dolphins in the wild.
Another complicating factor is associated with the broadband
nature of echolocation signals and the possibility that the
center frequency of echolocation clicks tends to vary with
the intensity of the emitted clicks. Au et al. ?1985? found that
higher intensity clicks emitted by a beluga whale ?Delphi-
napterus leucas? in Hawaii had higher frequencies than the
lower intensity clicks used by the same animal in San Diego
Bay. Au et al. ?1995? also found a nearly linear relationship
between the center frequency and source level of a false
killer whale ?Pseudora crassidens?, i.e., the higher the source
level, the higher the center frequency of the emitted clicks.
Therefore, any measurements of the spectra of echolocation
clicks should be accompanied by an estimate of the source
A. Measurement system
A four-hydrophone array with the hudrophones arranged
as a symmetrical star was used to measure the echolocation
signals of S. frontalis. Such a sensor geometry was used
successfully by Aubauer ?1995? in tracking echolocating bats
in the field. The array structure resembled the letter ‘‘Y,’’
with each arm being 45.7 cm in length and separated by an
angle of 120 degrees as shown in Fig. 2. The arms of the
a?Electronic mail: email@example.com
b?Electronic mail: firstname.lastname@example.org
598J. Acoust. Soc. Am. 113 (1), January 20030001-4966/2003/113(1)/598/7/$19.00 © 2003 Acoustical Society of America
array were constructed out of 1.27-cm o.d. PVC pipe with a
spherical hydrophone connected to the end of each pipe and
the cable running through the center of the pipe. Another
hydrophone was connected at the geometric center of the
‘‘Y.’’ The PVC pipes fit into holes drilled into a 2.54-cm-
thick delrin plate. The range of a sound source can be deter-
mined by measuring the time of arrival differences between
the signal at the center and the three other hydrophones. If
the arrival time difference between the center and the other
hydrophones is denoted as ?0i, where i?1, 2, and 3, then the
range, R, of the source can be expressed as ?see Appendix?
where c is the speed of sound, and a is the distance between
the center and the other hydrophones.
An underwater housing connected to the back of the
hydrophone mounting plate contained an amplifier and line-
driver for each of the hydrophones. A CCD video camera in
an underwater housing was mounted next to the center hy-
drophone. A multi-conductor cable, 77 m in length, consist-
ing of five coaxial lines and two d.c. power lines, connected
the array to an adjustable amplifier-filter box containing a
Echolocation signals were digitized with two Gage-
1210, 12 bit dual simultaneous sampling data acquisition
boards that were connected to a ‘‘lunch box’’ computer via
two EISA slots. The data acquisition system operated at a
sample rate of 500 kHz with a pretrigger capability. When
the computer signaled the Gage-1210 to collect data, four
channels of acoustic signals were simultaneously and con-
tinuously digitized with the results going into separate circu-
lar memories on each Gage-1210 board. When an echoloca-
tion signal was detected by the center hydrophone, it
triggered the data acquisition board. Two hundred pretrig-
gered points and two hundred posttrigger points were col-
lected for each channel and downloaded into the computer. A
total of 80 clicks could be downloaded for each episode be-
fore the data had to be stored on the hard drive. A specially
constructed ISA board was also used to measure the time
interval between the clicks being acquired and to cause a
light-emitting diode to flash, indicating that clicks were be-
ing captured. The interclick interval data was also down-
loaded and stored on the hard drive. The time of capture ?to
the closest 18-ms interval of the computer timing system? of
each click was also saved and stored on the hard drive. The
clock on a portable VCR was synchronized to the computer’s
clock so that the video images could be synchronized with
the acoustic data.
B. Acoustic measurements
Measurement of echolocation signals was conducted
from the Wild Dolphin Project’s 60-ft power catamaran,
FIG. 1. Map of the study site in the
banks area of the Bahama Island.
FIG. 2. Schematic diagram of the four-hydrophone symmetrical star array
along with a video camera on a pole.
599J. Acoust. Soc. Am., Vol. 113, No. 1, January 2003 W. W. L. Au and D. L. Herzing: Atlantic spotted dolphin echolocation
Stenella. A map of the field site showing the Bahama bank
where the study was conducted is shown in Fig. 1. All the
electronics including the lunch-box computer and video tape
recorder were housed on the deck of the catamaran. The field
work was performed in the summers of 1996 and 1997.
The four-hydrophone array was deployed by two meth-
ods. In the first method, the array was connected to a 1-m
long, 2.54-cm o.d. aluminum pipe with a handle grip at one
end of the pipe. Dolphins were first located by patrolling the
sandbank, an area of approximately 644 km2. When the ani-
mals were located, the speed of the catamaran was reduced,
and the dolphins were encouraged to bow ride for a short
period while swimmers prepared to enter the water. The en-
gine of the catamaran was then placed into idle and swim-
mers entered the water equipped with swim fins and snor-
kels. The array was then handed over to one of the swimmers
who then swam towards the dolphins. At the same time, an
operator controlled the data acquisition sequence by arming
the computer to start the data acquisition process and by
starting the video tape recorder.
The second method involved positioning the catamaran
at night along the edge of the drop-off from the sandbank to
deeper waters. A spotlight was directed into the water next to
the catamaran, attracting various small fishes and squid to
the surface. Spotted dolphins were also attracted by either
the spotlight or the congregation of micronekton. The array
was attached to a 3-m-long aluminum pole and placed in the
water along side the boat close to where the spotlight inter-
sected the water’s surface. The center of the array was low-
ered to a depth of 1 to 1.5 m below the surface. Spotted
dolphins often milled about 20–30 m from the boat foraging
on prey, and also made ‘‘runs’’ into the lit area towards the
array, continuously echolocating as they did so.
A total of 43 files of echolocation clicks were collected
on three field trips. The quantity of data collected was lim-
ited by periods of equipment malfunction requiring repair.
There were also occasions when a dolphin approached the
array only a few degrees from the plane of the array so that
one of the hydrophones would not detect the click because of
the direction of the animal’s beam. The number of clicks
collected per file varied considerably from as low as 3 to a
high of 80, the maximum number of clicks that the system
could handle. Only echolocation events in which the ampli-
tude of the signals received by the center hydrophone was
either the highest or within 3 dB of the highest were accepted
for analysis. This criterion was chosen to ensure that a dol-
phin beam was directed at the array. The beam patterns mea-
sured for three different odontocete species ?Au, 1993; Au
et al., 1995? indicate that when the major axis of the beam is
directed to within ?5 degrees of the center hydrophone, the
signal received by the center hydrophone will either have the
highest or will be within 3 dB of the highest amplitude. A
total of 1277 clicks met the appropriate criterion.
Three typical types of clicks are shown in Fig. 3 with the
signal waveforms on the right and the frequency spectra on
the left. These clicks are very brief, generally less than 70 ?s
in duration, with broad frequency spectra. Clicks with bimo-
dal spectra are obvious in the spectra plots. Some of the
bimodal spectra have relatively high peak frequencies
(?80 kHz) whereas some have low peak frequencies
(?40 kHz). The majority (?80%) of the clicks examined
had bimodal spectra. The click waveforms resemble those
used by other odontocetes such as Tursiops truncatus, Del-
phinapterus leucas ?Au, 1993?, Pseudora crassidens ?Au
et al., 1995?, and Lagoringcus albirostris ?Rasmussen et al.,
The peak-to-peak source level as a function of range
between an echolocating dolphin and the array is shown in
Fig. 4. As the dolphin’s range to the array decreased, the
source level also decreased. The solid curve in Fig. 4 is a
regression curve represented by the equation
and has an r2value of 0.52, where SL is the source level in
dB re 1 ?Pa and R is the range in meters. The decrease in
FIG. 3. Examples of the some representative waveforms and frequency
spectra emitted by Stenella frontalis in the Bahamas banks.
FIG. 4. Scatter plot of source level as a function of the range between an
echolocating dolphin and the hydrophone array. The solid curve represents
the one-way spherical spreading curve-fitted to the data in a least-square
fashion with r2?0.56.
600J. Acoust. Soc. Am., Vol. 113, No. 1, January 2003 W. W. L. Au and D. L. Herzing: Atlantic spotted dolphin echolocation
SL corresponded to the decrease in the one-way spherical
spreading loss. Therefore, the amplitude of the echoes return-
ing to the dolphins increased in magnitude as the range de-
creased, suggesting that the dolphins prefer to receive echoes
that have increasing signal-to-noise. The results also suggest
that the dolphins were echolocating on the hydrophone array
and not on some other objects since the source level de-
creased as the range to the array decreased. Also, the inter-
click intervals were always greater than the two-way travel
time from the animals to the array and back, which is con-
sistent with the notion that the dolphins were echolocating on
the array. The fitted curve was constrained to vary as
20logR, however, the ‘‘best-fit’’ logarithm’s curve would be
very similar to the 20logR fit and the difference in r2would
be in the third decimal place.
The distributions of peak and center frequencies of the
echolocation signals are shown in Fig. 5. Peak frequency is
defined as the frequency at which the frequency spectrum of
a signal has its maximum amplitude. Center frequency is
defined as that frequency which divides the energy in a fre-
quency spectrum into two equal parts; it is defined math-
where S(f ) is the Fourier transform of the echolocation sig-
nal and f is the instantaneous frequency. Seventy-six percent
of the signals had center frequencies greater than 50 kHz and
23% had center frequencies greater than 80 kHz. The center
frequency extends to much higher frequencies than the peak
frequency and this property is indicative of signals with bi-
modal spectra. Eighty percent of the echolocation clicks had
bimodal spectra in which the amplitudes of the secondary
peaks were within 50% of the amplitude of the primary peak.
Therefore, center frequency is a more representative measure
of signals with bimodal spectra ?Au et al., 1995?.
The variation of the center frequency as a function of the
source level is shown in Fig. 6. A third order polynomial
regression with an r2value of 0.14 is also shown in the
figure. The regression line suggests that the center frequency
in independent of the source level for levels lower than 200
dB re 1 ?Pa. However, as the source level increases beyond
205 dB, the center frequency also increases. If only signals
with source levels equal to or greater than 205 dB are con-
sidered, the dependence of center frequency on source level
becomes stronger with an r2value of 0.25. The dependence
of the center frequency on source level is weaker than for
Pseudora crassidens ?Au et al., 1995? where a linear regres-
sion line has an r2value of 0.44.
Histograms of the 3-dB bandwidth and the rms band-
width are shown in Fig. 7. The 3-dB bandwidth is the width
of the frequency band between the two points that are 3-dB
lower than the maximum amplitude of a spectrum. The 3-dB
points are also referred to as the half-power points. The rms
bandwidth is a measure of the frequency width about the
FIG. 5. Histograms of peak and center frequencies of echolocation signals.
FIG. 6. Scatter plot of center frequency versus source level. A third-order
polynomial is fitted for all data and has an r2value of 0.14 while a linear
regression curve is shown for source levels greater than and equal to 205 dB
FIG. 7. Histograms of 3-dB and rms bandwidth of echolocation signals.
601 J. Acoust. Soc. Am., Vol. 113, No. 1, January 2003W. W. L. Au and D. L. Herzing: Atlantic spotted dolphin echolocation
center frequency. It is defined as ?Rihaczek, 1969?
where fois the center frequency given in Eq. ?3?. The 3-dB
bandwidth for bimodal spectra can often provide a misrepre-
sentation of the width of the signal since the bandwidth
might cover only the frequency range about the peak fre-
quency. The rms bandwidth is probably a better measure of
the width of signals with bimodal spectra. The histograms
clearly show higher values for the rms bandwidth than the
A scatter plot of rms bandwidth as a function of center
frequency along with a linear regression line is shown in Fig.
8. The linear regression has a relatively high r2value of
0.44, indicating a strong relationship between rms bandwith
and center frequency. As the frequency increased, the band-
width also increased almost linearly. The range or temporal
resolution of a signal is related to the inverse of the band-
width so that the wider the bandwidth, the smaller the tem-
poral resolution capability of the signal ?Burdic, 1969?.
IV. DISCUSSION AND CONCLUSIONS
Atlantic spotted dolphin ?Stenella frontalis? project
broadband, short duration echolocation signals similar to
those of other odontocetes. Most of the signals have a bimo-
dal frequency distribution, which also contributes to the
broadband nature of the signals. The broad bandwidth of the
echolocation signal provides a good range resolution capa-
bility ?Au, 1993? that should enable Stenella frontalis to be
able to perform fine target discrimination in a shallow water
environment where bottom reverberation can be trouble-
some. Spotted dolphin also project relatively high-amplitude
signals with maximum source level measured about 223 dB
re 1 ?Pa, although most of the source levels were between
200 and 210 dB re 1 ?Pa.
The peak-to-peak source levels measured for Stenella
frontalis are comparable to those measured for Tursiops trun-
catus in open-water captive echolocation experiments ?Au,
1980, 1993?, for comparable target ranges. For target ranges
between 6 and 20 m, Tursiops source levels varied from
about 204 to 216 dB re 1 ?Pa, which are similar to that of
Stenella frontalis. However, there is a large difference be-
tween the target strengths of targets used in the echolocation
experiments for Tursiops and the target strength of the array
assembly used to measure signals in the field. Although the
target strength of the array was not measured, the theoretical
target strength of an aluminum pipe, connected to a flat
plexiglass container mounted on a flat delrin plate along with
a camera holder, should be approximately 15–20 dB greater
than the small cylinders and spheres used in the Tursiops
experiments ?Au, 1980, 1993?. This comparison suggests the
importance of range on the source levels utilized by dol-
phins. Despite the higher target strength of the array, the
spotted dolphins emitted similar levels of echolocation sig-
nals as Tursiops echolocating on much weaker targets at
The variation of source level as a function of the one-
way transmission loss is similar to that of the white beaked
dolphin, Lagenorhynchus albirostris ?Rasmussen et al.,
2002? and killer whales, Orcinus orca ?Au et al., 2001?. This
type of variation in source levels is also similar to variations
found with captive dolphins. If the data shown in Fig. 7.14 of
Au ?1993? are rearranged into a plot of source level versus
range, the variation with range will also be a function of the
one-way transmission loss.
Our results suggest that several basic signal parameters
are interrelated in a complex relationship. Source level is
dependent on target range, center frequency is dependent on
source level ?at least for source levels greater than 205 dB?,
and rms bandwidth is dependent on center frequency. How-
ever, it seems that the most basic parameter in this interrela-
tionship is target range. Therefore, it is important to be able
to ascertain the range of echolocating dolphins when mea-
suring echolocation signals, even for on-axis signals.
The results of this study also clearly demonstrate the
utility of using a multi-hydrophone array to measure echolo-
cation signals of dolphins in the wild. The symmetrical star
array used in this study is relatively compact and easy to
handle, and can provide information on whether a specific
received echolocation signal originated in the vicinity of the
major axis of the animal’s transmission beam. Time of ar-
rival differences between hydrophones were easily ascer-
tained because of the rapid onset of the echolocation signals.
Our results suggest that it is very difficult to obtain reliable
data on echolocation signals without the use of some kind of
Grateful thanks are given to the crew and staff of the
Wild Dolphin Project. Special thanks to Lisa Harrod, Alice
and Ed Crawford, Randy and Michelle Wells, Will Engleby,
and Tim Barrett. The senior author thanks Dr. Roland
Aubauer for his suggestion of using a symmetrical star ge-
ometry and for various discussions associated with his re-
search on detection flying bats. The assistance of Michiel
Schotten in testing and calibrating the array is also appreci-
ated. This work was conducted under Bahamas Department
of Fisheries Research Permit No. MAF/FIS 12 and funded
by the Office of Naval Research, Dr. Robert Gisiner, pro-
gram manager. This is HIMB Contribution No. 1136.
FIG. 8. Scatter plot of rms bandwidth versus center frequency. The linear
regression line has an r2of 0.44.
602 J. Acoust. Soc. Am., Vol. 113, No. 1, January 2003 W. W. L. Au and D. L. Herzing: Atlantic spotted dolphin echolocation
Let us consider the four hydrophone array with on the
x-y plane and a sound source located at coordinates
(Sx,Sy,Sz) as shown in Fig. 9. Let Ribe the range from the
source to the ith hydrophone, where i?1,2,3. Then the range
to each hydrophone can be expressed by the equation
where c is the speed of sound in water, ?0iis the time of
arrival difference between the center hydrophone and the ith
hydrophone, and l?)/2?cos30°. Using the technique of
Watkins and Schevill ?1972?, the top equation of the system
of equations in ?A1? is subtracted from the other equations to
give ?after some rearranging?
This is a system of three equations with three unknowns.
There are a variety of methods to solve for the unknowns Sx,
Sy, and t0. The equations of ?A2? can be expressed in a
matrix format as
Using Cramer’s rule ?Kreyszig, 1983? we can solve for t0by
solving the determinant equation
where ? is the characteristic determinant defined by
Solving the determinant in Eq. ?A4? for t0using the relation-
ship of R?t0c, we obtain the equation for the range from the
source to the center hydrophone in the array as
For a given set of delay times, solutions for Szwill have a ?
ambiguity, indicating that the source can be either above or
below the X-Y plane of Fig. 9.
The accuracy of using the symmetrical star array to de-
termine the range of a sound source was measured by pro-
jecting a simulated dolphin echolocation signal at different
ranges from the array and using Eq. ?A6? to estimate the
range. The results of the measurements shown in Fig. 10
suggest that this technique can give very accurate results out
to about 17.5 m. Ten pings were measured at each range. If
the purpose of estimating R is to obtain the transmission loss
due to spherical spreading, then the difference in the esti-
mated and actual ranges will result in only a 1.2-dB error for
an actual range of 25 m.
Au, W. W. L. ?1980?. ‘‘Echolocation signals of the Atlantic bottlenose dol-
phin ?Tursiops truncatus? in open waters,’’ in Animal Sonar Systems, ed-
ited by R. G. Busnel and J. F. Fish, ?Plenum, New York?, pp. 251–282.
Au, W. W. L. ?1993?. The Sonar of Dolphins ?Springer-Verlag, New York?.
Au, W. W. L., Ford, J., andAllman, K. ?2001?. ‘‘Echolocation signal of killer
whales ?Orcinus orca? in Johnstone Strait, Canada,’’in Proc. 14th Biennial
Conf. on the Biology of Marine Mammals, Vancouver, BC, Canada, pp.
Au, W. W. L., Carder, D. A., Penner, R. H., and Scronce, B. L. ?1985?.
‘‘Demonstration of adaptation in Beluga whale echolocation signals,’’ J.
Acoust. Soc. Am. 77, 726–730.
Au, W. W. L., Pawloski, J., Nachtigall, P. E., Blonz, M., and Gisner, R.
?1995?. ‘‘Echolocation signal and transmission beam pattern of a false
killer whale ?Pseudorca crassidens?,’’ J. Acoust. Soc. Am. 98, 51–59.
Aubauer, R. ?1995?. ‘‘Korrelationsverfahren zur Flugbahnverfolgung echoo-
rtender Fledemause,’’ Fortschr.-Ber. VDI Reihe 17 Nr. 132 Dusseldorf:
FIG. 9. Geometry for determining the location of a sound source from time
of arrival difference information with a symmetrical star array.
FIG. 10. Calibration measurement for the symmetrical star array used in this
study. The mean and standard deviation of the estimated range from ten
measures are plotted against the actual range.
603J. Acoust. Soc. Am., Vol. 113, No. 1, January 2003 W. W. L. Au and D. L. Herzing: Atlantic spotted dolphin echolocation
Burdic, W. S. ?1969?. Radar Signal Analysis ?Prentice–Hall, Englewood
Caldwell, D. K., and Caldwell, M. C. ?1966?. ‘‘Observations on the distri-
bution, coloration, behavior and audible sound production of the spotted
dolphin, Stenella plagiodon,’’ Contrib. Sci. 104, 1–28.
Caldwell, D. K., and Caldwell, M. C. ?1971?. ‘‘Underwater pulsed sounds
produced by captive spotted dolphins, Stenella plagiodon,’’ Cetol. 1, 1–7.
Caldwell, M. C., Caldwell, D. K., and Miller, J. F. ?1973?. ‘‘Statistical evi-
dence for individual signature whistles in the spotted dolphin, Stenella
plagiodon,’’ Cetol. 16, 1–21.
Herzing, D. L. ?1996?. ‘‘Underwater behavioral observations and associated
vocalizations of free-ranging Atlantic spotted dolphins, Stenella frontalis,
and bottlenose dolphin, Tursiops truncatus,’’ Aq. Mamm. 22, 61–79.
Herzing, D. L. ?1997?. ‘‘The natural history of free-ranging Atlantic spotted
dolphins, ?Stenella frontalis?: Age classes, color phases and female repro-
duction,’’ Marine Mammal Sci. 13, 40–59.
Herzing, D. L. ?2000?. ‘‘Acoustics and Social Behavior of wild dolphins:
implications for a sound society,’’ in Hearing by Whales and Dolphins,
edited by W. W. L. Au, A. N. Popper, and R. R. Fay ?Springer, New York?.
Kreyszig, E. ?1983?. Advanced Engineering Mathematics, 5th ed. ?Wiley,
Rasmussen, M. H., Miller, L.A., andAu, W. W. L. ?2002?. ‘‘Source levels of
clicks from free-ranging white-beaked dolphins ?Lagenorhynchus albiros-
tris, Gray, 1846? recorded in Icelandic waters,’’ J. Acoust. Soc. Am. 111,
Rihaczek, A. W. ?1969?. Principles of High-Resolution Radar ?McGraw-
Hill, New York?.
Watkins, W. A., and Schevill, W. E. ?1972?. ‘‘Sound source location by
arrival-times on a non-rigid three-dimensional hydrophone array,’’ Deep-
Sea Res. Oceanogr. Abstr. 19, 691–705.
Wood, Jr., F. G. ?1953?. ‘‘Underwater sound production and concurrent be-
havior of captive porpoise, Tursiops truncatus and Stenella plagiodon,’’
Bull. Mar. Sci. Gulf Carib. 3, 120–133.
604 J. Acoust. Soc. Am., Vol. 113, No. 1, January 2003 W. W. L. Au and D. L. Herzing: Atlantic spotted dolphin echolocation