played,adults preferentially turned the left ear,whereas infants failed to
show bias;the turnstone’s call is familiar to rhesus,but is a signal that is
clearly not from a conspecific.Together,these results provide additional
support for the idea that macaques have right ear bias for perceiving
conspecific signals,implying left hemisphere dominance for processing
conspecific calls.
To determine which acoustic features of a signal influence the prefer-
ential head-turning response and thus the suggested hemispheric bias
underlying perception in rhesus (Hauser and Andersson 1994),a second
experiment was carried out (Hauser,Agnetta and Perez,in press).
Digital signal-editing tools (Beeman 1996) were used to modify the
structure of naturally produced calls.The idea,in a nutshell,is this.Call
types within the repertoire are characterized by a suite of parametric fea-
tures,including both temporal and spectral ones.We hypothesized that
when particular features of a signal are manipulated beyond the range
of natural variation,such signals will no longer be perceived as conspe-
cific calls;call types within the repertoire will differ in terms of their char-
acteristic defining features and consequently,no single manipulation is
likely to be meaningful across all call types,except at extremes.Given
the observation that rhesus respond to playbacks of one avian species’
alarm call by preferentially turning their left ear to listen (Hauser and
Andersson 1994),we predicted that playbacks of calls shifted outside the
species-typical range would also elicit left ear bias;such manipulations
may lead to no response bias if the acoustic signal caused significant acti-
vation in both hemispheres as a result different causal factors.Calls that
have been manipulated,but remain within the species-typical range,
would continue to elicit right ear bias,that is,continue to be classified as
conspecific calls.
The focus of this experiment (Hauser,Agnetta,and Perez,in press)
was the salience of temporal parameters in call classification.All three
call types presented are characterized by pulses of energy separated by
silence.For each call type,we started with a naturally recorded call con-
sisting of three pulses,together with pulse and interpulse intervals that
fell close to the population mean.We then shrunk as well as stretched
interpulse intervals to create four additional stimuli.Calls with reduced
interpulse intervals were reduced to the minimum observed in the pop-
ulation or were completely eliminated.Calls with stretched interpulse
intervals were increased to the maximum in the population or twice the
maximum.
Figure 6.5 shows representative spectrograms of the three call types
used in this experiment: grunt,shrill bark, and copulation scream
(Hauser 1993b;Hauser and Marler 1993a;Bercovitch,Hauser,and Jones
1995).Three factors guided our decision to use these particular call types.91 Primate Vocalizations in Emotion and Thought