matrilines. Matrilineal groups are genetically differentiated (Olivier et al.
1981). Females remain in their natal group, whereas males disperse. Female
offspring assume the dominance rank of their mothers; members of a matriline assist one another in agonistic interactions to actively defend their
dominance rank. Social groups generally avoid one another rather than
actively defend territories. Aggressive encounters do occur, and females
usually focus their aggression on other females when defending a clumped
food source. Females sometimes act aggressively toward immigrant males,
and males almost always do so (Cheney 1987).
Japanese (M. fuscata) and rhesus (M. mulatta) macaques give calls,
termed “coos,” in various social contexts and when encountering food
(Hauser 1996). Coos can be used to recognize individuals (Hansen 1976;
Masataka 1985). Studies of call structure provide conflicting results with
respect to vocal learning. Green (1975) argues that social transmission has
produced differences in frequency modulation of coos in three troops of
Japanese macaques (M. fuscata). Unfortunately, he provides no analysis of
acoustic features and illustrates single examples from each troop, which
makes quantitative assessment of troop variation impossible. The vocal
differences he describes are related to the provisioning context and are the
only acoustic differences among troops noted. Individuals give some coos
with the “locale-specific” features and some without. Other vocalizations
do not differ among troops.
Subsequent work to investigate vocal learning has used cross-fostering
among two species, the Japanese macaque, M. fuscata, and rhesus macaque,
M. mulatta (Masataka and Fujita 1989; Owren et al. 1992, 1993). These
studies ask whether offspring raised by heterospecifics produce coos like
those of their foster mothers. Masataka and Fujita (1989) cross-fostered a
single Japanese and two rhesus macaques. They describe differences in
peak frequency of the fundamental frequency, but not in duration, of the
two species’ calls, and they claim that cross-fostered infants produce coos
with frequency characteristics similar to their adoptive mothers’. They also
present playback data suggesting that others respond to the cross-fostered
monkeys as though they are conspecifics. These data are interpreted as
providing strong evidence of vocal learning.
However, Owren et al. (1992, 1993) challenge these results. Although
calls of infants raised by conspecifics are individually distinctive and differ
between species, there are no statistically significant differences between
calls given by adults of the two species for peak frequency or duration, nor
for five other frequency variables related to fundamental frequency (Owren
et al. 1992). Each species shows substantial variability and overlaps the
frequency range of the other. Rather than being species-specific, adult coos
differ among individuals, although classification accuracy in a discriminant
analysis is low. Coos are not particularly good vocalizations to use in a crossfostering study because there are no consistent differences between species
in adult calls (Owren et al. 1992). Thus, they cannot serve as distinctive
models for developing species-specific characteristics. In addition, similar4. Comparative Vocal Learning
165
1981). Females remain in their natal group, whereas males disperse. Female
offspring assume the dominance rank of their mothers; members of a matriline assist one another in agonistic interactions to actively defend their
dominance rank. Social groups generally avoid one another rather than
actively defend territories. Aggressive encounters do occur, and females
usually focus their aggression on other females when defending a clumped
food source. Females sometimes act aggressively toward immigrant males,
and males almost always do so (Cheney 1987).
Japanese (M. fuscata) and rhesus (M. mulatta) macaques give calls,
termed “coos,” in various social contexts and when encountering food
(Hauser 1996). Coos can be used to recognize individuals (Hansen 1976;
Masataka 1985). Studies of call structure provide conflicting results with
respect to vocal learning. Green (1975) argues that social transmission has
produced differences in frequency modulation of coos in three troops of
Japanese macaques (M. fuscata). Unfortunately, he provides no analysis of
acoustic features and illustrates single examples from each troop, which
makes quantitative assessment of troop variation impossible. The vocal
differences he describes are related to the provisioning context and are the
only acoustic differences among troops noted. Individuals give some coos
with the “locale-specific” features and some without. Other vocalizations
do not differ among troops.
Subsequent work to investigate vocal learning has used cross-fostering
among two species, the Japanese macaque, M. fuscata, and rhesus macaque,
M. mulatta (Masataka and Fujita 1989; Owren et al. 1992, 1993). These
studies ask whether offspring raised by heterospecifics produce coos like
those of their foster mothers. Masataka and Fujita (1989) cross-fostered a
single Japanese and two rhesus macaques. They describe differences in
peak frequency of the fundamental frequency, but not in duration, of the
two species’ calls, and they claim that cross-fostered infants produce coos
with frequency characteristics similar to their adoptive mothers’. They also
present playback data suggesting that others respond to the cross-fostered
monkeys as though they are conspecifics. These data are interpreted as
providing strong evidence of vocal learning.
However, Owren et al. (1992, 1993) challenge these results. Although
calls of infants raised by conspecifics are individually distinctive and differ
between species, there are no statistically significant differences between
calls given by adults of the two species for peak frequency or duration, nor
for five other frequency variables related to fundamental frequency (Owren
et al. 1992). Each species shows substantial variability and overlaps the
frequency range of the other. Rather than being species-specific, adult coos
differ among individuals, although classification accuracy in a discriminant
analysis is low. Coos are not particularly good vocalizations to use in a crossfostering study because there are no consistent differences between species
in adult calls (Owren et al. 1992). Thus, they cannot serve as distinctive
models for developing species-specific characteristics. In addition, similar4. Comparative Vocal Learning
165