18
WWL. Au
minor variations to another contour, and in determining to what specific
category a particular contour belongs. Classification of sound repertoire has
also suffered from nomenclature difficulties involved with naming of a
sound without sufficient spectral and structural information to allow for
consistency between researchers (Herzing 1988). However, if very broad
general categories are used, then whistles can be lumped into six or seven
different general broad contours (Bazua-Duran 1997) as shown in Figure
1.9. Figures 1.9b-f are in separate categories that are based on the slope of
the whistle and the number of inflection points, where an inflection point
is defined as a point at which the slope of the contour reverses direction
(e.g., an upsweep changing to a downsweep [negative infection] or a downsweep changing to an upsweep [positive inflection]). The contour of
Figure 1.9g belongs to a class of short-duration emissions (Caldwell and
Caldwell 1965).
A more quantitative method of categorizing whistle signals involves a
determination of various parameters to describe a signal. Some of the parameters include: (a) starting frequency, (b) ending frequency, (c) minimum
frequency, (d) maximum frequency, (e) number of inflection points, (f)
frequency of inflection points, (g) duration, (h) break in contour, and (i)
the presence of harmonics. These parameters are then fed to different
statistical software programs for analysis. Steiner (1981) studied the speciesspecific differences in pure tonal whistles of five western North Atlantic
Dolphin species using a multivariate discriminant analysis and parameters
a, b, c, d, e, and h. The discriminant analysis easily discriminated the
whistles of the long-finned pilot whale (Globicephala melaena) and
the Atlantic bottlenose dolphin (T. truncatus) from whistles of the Atlantic
white-sided dolphin (Lagenorhynchus acutus), the Atlantic spotted dolphin
(Stene/la frontalis), and the spinner dolphin (S. longirostris). The analysis
did not discriminate between whistles of L. acutus, S. longirostris, and
S. frontalis. Wang et al. (1995) compared the whistles of seven odontocete
species recorded in the wild using the parameters a-i except for f,
and included two other parameters that were not clearly defined: beginning
sweep and end sweep. They also used a similar discriminant analysis
program as Steiner (1981) in order to make their results somewhat
comparable. An interesting result from their study is the strong correlation
between the maximum frequency of the whistles with body length.
The longer the body length the lower the maximum frequency the various
animals emitted. They also found that species-specific differences existed
in the whistles and the degree of differences correlated with taxonomic
relations, habitats, and body length. Species with close taxonomical relations
had similar whistles. Pelagic species had whistles of higher frequency
and greater frequency modulation than coastal species and riverine species
had much lower frequency range and lower modulation than oceanic
species.
WWL. Au
minor variations to another contour, and in determining to what specific
category a particular contour belongs. Classification of sound repertoire has
also suffered from nomenclature difficulties involved with naming of a
sound without sufficient spectral and structural information to allow for
consistency between researchers (Herzing 1988). However, if very broad
general categories are used, then whistles can be lumped into six or seven
different general broad contours (Bazua-Duran 1997) as shown in Figure
1.9. Figures 1.9b-f are in separate categories that are based on the slope of
the whistle and the number of inflection points, where an inflection point
is defined as a point at which the slope of the contour reverses direction
(e.g., an upsweep changing to a downsweep [negative infection] or a downsweep changing to an upsweep [positive inflection]). The contour of
Figure 1.9g belongs to a class of short-duration emissions (Caldwell and
Caldwell 1965).
A more quantitative method of categorizing whistle signals involves a
determination of various parameters to describe a signal. Some of the parameters include: (a) starting frequency, (b) ending frequency, (c) minimum
frequency, (d) maximum frequency, (e) number of inflection points, (f)
frequency of inflection points, (g) duration, (h) break in contour, and (i)
the presence of harmonics. These parameters are then fed to different
statistical software programs for analysis. Steiner (1981) studied the speciesspecific differences in pure tonal whistles of five western North Atlantic
Dolphin species using a multivariate discriminant analysis and parameters
a, b, c, d, e, and h. The discriminant analysis easily discriminated the
whistles of the long-finned pilot whale (Globicephala melaena) and
the Atlantic bottlenose dolphin (T. truncatus) from whistles of the Atlantic
white-sided dolphin (Lagenorhynchus acutus), the Atlantic spotted dolphin
(Stene/la frontalis), and the spinner dolphin (S. longirostris). The analysis
did not discriminate between whistles of L. acutus, S. longirostris, and
S. frontalis. Wang et al. (1995) compared the whistles of seven odontocete
species recorded in the wild using the parameters a-i except for f,
and included two other parameters that were not clearly defined: beginning
sweep and end sweep. They also used a similar discriminant analysis
program as Steiner (1981) in order to make their results somewhat
comparable. An interesting result from their study is the strong correlation
between the maximum frequency of the whistles with body length.
The longer the body length the lower the maximum frequency the various
animals emitted. They also found that species-specific differences existed
in the whistles and the degree of differences correlated with taxonomic
relations, habitats, and body length. Species with close taxonomical relations
had similar whistles. Pelagic species had whistles of higher frequency
and greater frequency modulation than coastal species and riverine species
had much lower frequency range and lower modulation than oceanic
species.