The signature hypothesis has been reassessed recently (Caldwell et al.
1990; McCowan and Reiss 1995a, 2001). McCowan and Reiss (1995a) argue
that repertoires include many whistles shared among the members of a
social group or even across social groups. These whistles are not likely to
be individual signatures; rather, hypotheses suggest they may be referential
(Tyack 1997) or affiliative (Smolker 1993; Tyack 1997). McCowan and Reiss
(1995a) also provide data indicating that the proportion of whistles unique
to an individual can be quite low. They suggest that the preponderance of
signature whistles documented by some researchers is due to recording
context. When restrained and isolated from others, both adults and infants
produce many signature whistles (McCowan and Reiss 1995a; Janik and
Slater 1998). When freely interacting with others, dolphins produce many
shared whistles (McCowan and Reiss 1995a, 1995b). Thus, defining a signature as the whistle most commonly produced is problematic.
Much of the evidence for vocal learning in Tursiops comes from observational studies of captive animals (Tyack 1986; Reiss and McCowan 1993;
McCowan and Reiss 1995a, 1995b). Observational studies of communication are being incorporated into long-term studies of two wild populations,
one in Sarasota, Florida (Wells et al. 1987; Sayigh et al. 1990, 1995; Wells
1991), and the other in Shark Bay, Australia (Connor et al. 1992a; Smolker
1993; Connor and Smolker 1995, 1996). Experimental work on vocal learning is restricted primarily to eliciting specific whistles by operant conditioning (Richards et al. 1984; Reiss and McCowan 1993). Experimental
work is hampered by two factors: small sample sizes due to difficulties
keeping animals in captivity, and neither social and acoustic isolation nor
deafening are possible due to ethical considerations or because they induce
abnormal social and acoustic behavior.
Vocal development has been studied for both captive and free-living
dolphins. Infant bottlenose dolphins produce whistles when they are only a
few days old (Caldwell and Caldwell 1979), suggesting that auditory experience is not required for whistle production. Like adults, calves give some
whistles that are unique to each individual (73% of whistle types produced
during the first year), but do not retain one particular whistle as their
individual signature (McCowan and Reiss 1995b). A high proportion of
whistles are shared within social groups (23%), and many whistles are
shared by calves in the three groups studied (17%). Certain whistle types
are produced only by very young infants, and the same type is given by all
infants. The proportion of this infant whistle gradually decreases and the
proportion of other whistles increases as the calves age, resulting in a
turnover in the repertoire. Calves continue to acquire new whistle types
even as they discontinue use of earlier types. This process continues past
the first year, as repertoires of 12-month-old calves differ from adults. At
any one developmental stage, calves have larger repertoires than adults and
by 12 months have given as many as 22–55 different types of whistles
(McCowan and Reiss 1995b). This suggests that individual repertoires can
160
J.W. Boughman and C.F. Moss
1990; McCowan and Reiss 1995a, 2001). McCowan and Reiss (1995a) argue
that repertoires include many whistles shared among the members of a
social group or even across social groups. These whistles are not likely to
be individual signatures; rather, hypotheses suggest they may be referential
(Tyack 1997) or affiliative (Smolker 1993; Tyack 1997). McCowan and Reiss
(1995a) also provide data indicating that the proportion of whistles unique
to an individual can be quite low. They suggest that the preponderance of
signature whistles documented by some researchers is due to recording
context. When restrained and isolated from others, both adults and infants
produce many signature whistles (McCowan and Reiss 1995a; Janik and
Slater 1998). When freely interacting with others, dolphins produce many
shared whistles (McCowan and Reiss 1995a, 1995b). Thus, defining a signature as the whistle most commonly produced is problematic.
Much of the evidence for vocal learning in Tursiops comes from observational studies of captive animals (Tyack 1986; Reiss and McCowan 1993;
McCowan and Reiss 1995a, 1995b). Observational studies of communication are being incorporated into long-term studies of two wild populations,
one in Sarasota, Florida (Wells et al. 1987; Sayigh et al. 1990, 1995; Wells
1991), and the other in Shark Bay, Australia (Connor et al. 1992a; Smolker
1993; Connor and Smolker 1995, 1996). Experimental work on vocal learning is restricted primarily to eliciting specific whistles by operant conditioning (Richards et al. 1984; Reiss and McCowan 1993). Experimental
work is hampered by two factors: small sample sizes due to difficulties
keeping animals in captivity, and neither social and acoustic isolation nor
deafening are possible due to ethical considerations or because they induce
abnormal social and acoustic behavior.
Vocal development has been studied for both captive and free-living
dolphins. Infant bottlenose dolphins produce whistles when they are only a
few days old (Caldwell and Caldwell 1979), suggesting that auditory experience is not required for whistle production. Like adults, calves give some
whistles that are unique to each individual (73% of whistle types produced
during the first year), but do not retain one particular whistle as their
individual signature (McCowan and Reiss 1995b). A high proportion of
whistles are shared within social groups (23%), and many whistles are
shared by calves in the three groups studied (17%). Certain whistle types
are produced only by very young infants, and the same type is given by all
infants. The proportion of this infant whistle gradually decreases and the
proportion of other whistles increases as the calves age, resulting in a
turnover in the repertoire. Calves continue to acquire new whistle types
even as they discontinue use of earlier types. This process continues past
the first year, as repertoires of 12-month-old calves differ from adults. At
any one developmental stage, calves have larger repertoires than adults and
by 12 months have given as many as 22–55 different types of whistles
(McCowan and Reiss 1995b). This suggests that individual repertoires can
160
J.W. Boughman and C.F. Moss