22
w.w.L. Au
The signature whistle hypothesis has been challenged by McCowan and
Reiss (1995). They analyzed the whistles of 10 adult dolphins in captivity
from three different social groups. The normalized spectrogram data were
subjected to a K-cluster analysis and a discriminant analysis. They determined whistle categories within social groups and across social groups. Their
results were not consistent with the signature whistle hypothesis that suggested most of the whistles produced by dolphins are signature whistles,
since their subjects not only produced many different whistle types but they
also shared several whistle types within as well as across social groups.
4.4.3 Social Sounds: Burst Pulses
Burst pulse sounds are another major category of sound emissions by odontocetes. All odontocetes seem to produce burst pulse sounds. Dolphins and
small whales produce burst pulse sounds that are characterized by a high
repetition rate (greater than about 300 pulses per second) or low interpulse
intervals (less than about 3 ms). To the human ear, burst pulse sounds
resemble squawks, blats, squeals, cracks, snaps, bleats, barks, groans, and
moans, and have been described as such in the literature.
Researchers have suggested that burst pulse sounds are an important part
of dolphin sound emissions (see Herman and Tavolga 1980), yet these
sounds have not been studied as much as whistles. The heavy emphasis on
studying whistles has led many to suppose that whistles are the primary
mode of communication in dolphins and small whales. Burst pulse sounds
can have frequency components that extend beyond 100kHz (Au et al.
1999). An example of such a burst pulse sequence is depicted in Fig. 1.10.
Therefore, special wideband recording instruments and wideband hydrophones must be used to capture the entire spectrum of these pulses.
The burst pulse sound of Fig. 1.10 from a wild Hawaiian spinner dolphin
(S longirostiris) has no frequency components in the human audio range
and only frequency components in the ultrasonic range.
There is a fine region of demarcation between echolocation clicks and
burst pulse sounds, and one type of signal may merge into the other during
a sound emission bout. It would not be surprising if we eventually learn that
burst pulse sounds and echolocation clicks are produced by the same mechanism. The distinction between the two types of pulse sounds can be found
in the interclick intervals and intensity. When a dolphin echolocates, the
interclick intervals are usually longer than the time for the signal to travel
from the animal and back plus a lag time (Au, Chapter 9). Lag time is
defined as the difference in time between the two-way travel time and the
interclick interval. Evans and Powell (1967) found an almost constant mean
lag time of 15.4ms for target ranges of 1.4 to O.4m. Only when the animal
approached within 0.4 to 0.03 m of the target did the mean lag time decrease
to a minimum of 2.5 ms. However, it is doubtful that at these short ranges
a dolphin can meaningfully echolocate since the animal's transmitter is on
w.w.L. Au
The signature whistle hypothesis has been challenged by McCowan and
Reiss (1995). They analyzed the whistles of 10 adult dolphins in captivity
from three different social groups. The normalized spectrogram data were
subjected to a K-cluster analysis and a discriminant analysis. They determined whistle categories within social groups and across social groups. Their
results were not consistent with the signature whistle hypothesis that suggested most of the whistles produced by dolphins are signature whistles,
since their subjects not only produced many different whistle types but they
also shared several whistle types within as well as across social groups.
4.4.3 Social Sounds: Burst Pulses
Burst pulse sounds are another major category of sound emissions by odontocetes. All odontocetes seem to produce burst pulse sounds. Dolphins and
small whales produce burst pulse sounds that are characterized by a high
repetition rate (greater than about 300 pulses per second) or low interpulse
intervals (less than about 3 ms). To the human ear, burst pulse sounds
resemble squawks, blats, squeals, cracks, snaps, bleats, barks, groans, and
moans, and have been described as such in the literature.
Researchers have suggested that burst pulse sounds are an important part
of dolphin sound emissions (see Herman and Tavolga 1980), yet these
sounds have not been studied as much as whistles. The heavy emphasis on
studying whistles has led many to suppose that whistles are the primary
mode of communication in dolphins and small whales. Burst pulse sounds
can have frequency components that extend beyond 100kHz (Au et al.
1999). An example of such a burst pulse sequence is depicted in Fig. 1.10.
Therefore, special wideband recording instruments and wideband hydrophones must be used to capture the entire spectrum of these pulses.
The burst pulse sound of Fig. 1.10 from a wild Hawaiian spinner dolphin
(S longirostiris) has no frequency components in the human audio range
and only frequency components in the ultrasonic range.
There is a fine region of demarcation between echolocation clicks and
burst pulse sounds, and one type of signal may merge into the other during
a sound emission bout. It would not be surprising if we eventually learn that
burst pulse sounds and echolocation clicks are produced by the same mechanism. The distinction between the two types of pulse sounds can be found
in the interclick intervals and intensity. When a dolphin echolocates, the
interclick intervals are usually longer than the time for the signal to travel
from the animal and back plus a lag time (Au, Chapter 9). Lag time is
defined as the difference in time between the two-way travel time and the
interclick interval. Evans and Powell (1967) found an almost constant mean
lag time of 15.4ms for target ranges of 1.4 to O.4m. Only when the animal
approached within 0.4 to 0.03 m of the target did the mean lag time decrease
to a minimum of 2.5 ms. However, it is doubtful that at these short ranges
a dolphin can meaningfully echolocate since the animal's transmitter is on