Journal of Comparative Psychology
2000, Vol. 114, No. 1, 36-46
Copyright 2000 by the American Psychological Association, Inc.
0735-7036/00/$5.00 DOI: 10.1037//0735-7036.114.1.36
Proximate Factors Mediating "Contact" Calls in Adult Female Baboons
(Papio cynocephalus ursinus) and Their Infants
Drew Rendall, Dorothy L. Cheney, and Robert M. Seyfarth
University of Pennsylvania
"Contact" calls are widespread in social mammals and birds, but the proximate factors that
motivate call production and mediate their contact function remain poorly specified. Field
study of chacma baboons (Papio cynocephalus ursinus) revealed that contact barks in adult
females were motivated by separation both from the group at large and from their dependent
infants. A variety of social and ecological factors affect the probability of separation from
either one or both. Results of simultaneous observations and a playback experiment indicate
that the contact function of calling between mothers and infants was mediated by occasional
maternal retrieval rather than coordinated call exchange. Mothers recognized the contact barks
of their own infants and often were strongly motivated to locate them. However, mothers did
not produce contact barks in reply unless they themselves were at risk of becoming separated
from the group.
Vocalizations labeled contact calls are widespread among
primates and other social mammals and birds. In primates,
they include loud calls given periodically by widely sepa-
rated individuals, as well as comparatively quiet calls given
at high rates while groups move or forage in dense vegeta-
tion in which the risk of becoming separated is high (e.g.,
Boinski, 1991, 1993; Byrne, 1981; Dittus, 1988; Gautier &
Gautier, 1977; Itani, 1963; Marler & Hobbett, 1975; Palom-
bit, 1992; Robinson, 1982). Given the individualistic nature
of social relationships in many primate groups, researchers
often have suggested that, in addition to allowing individu-
als to maintain contact with the group at large, contact calls
serve the more specific function of allowing individuals to
maintain contact with particular social companions (e.g.,
Cheney, Seyfarth, & Palombit, 1996; Dittus, 1988; Rendall,
Rodman, & Emond, 1996). This hypothesis has received
indirect empirical support from several studies demonstrat-
Drew Rendall, Dorothy L. Cheney, and Robert M. Seyfarth,
Departments of Psychology and Biology, University of Pennsylvania.
This research was supported by the Natural Sciences and
Engineering Research Council of Canada, the National Science
Foundation, National Institutes of Health Grant HD-29483, the
National Geographic Society, and both the Research Foundation
and the Institute for Research in Cognitive Science of the Univer-
sity of Pennsylvania.
We are grateful to the Office of the President and the Department
of Wildlife and National Parks of the Republic of Botswana for
granting us permission to conduct research in the Moremi Game
Reserve. We are also grateful to Bill Hamilton for access to the field
site and background data, to Mokopi Mokopi for assisting with data
collection, and to Klaus Zuberbiihler for his comments on this
article. Karen Rendall was instrumental in all aspects of the field
research and deserves special thanks.
Correspondence concerning this article should be addressed to
Drew Rendall, who is now at the Department of Psychology and
Neuroscience, University of Lethbridge, 4401 University Drive,
Lethbridge, Alberta, Canada T1K 3M4. Electronic mall may be
sent to firstname.lastname@example.org.
ing individual differences in the acoustic structure of contact
calls (reviewed in Snowdon, 1986) and, in a few cases,
explicit vocal recognition (Hansen, 1976; Rendall et al.,
1996). Hence, the structure of contact calls is often compat-
ible with monitoring the location of specific group members.
However, few studies have tested the extent to which calling
is in fact motivated by separation from particular individu-
als, as opposed to the group at large (cf. Cheney et al., 1996;
Mitani & Nishida, 1993).
At the same time, although some studies have reported
chorused, or antiphonal, calling implying active vocal
exchanges (Biben, 1993; Smith, Newman, & Symmes,
1982; Snowdon & Hodun, 1981; Sugiura & Masataka, 1995;
Winter, Ploog, & Latta, 1966), few studies have tested
whether such "exchanges" result from a synchronization of
activity and underlying motivational state among group
members or whether they occur because individuals are
selectively answering the calls of specific social compan-
ions. The distinctions between these various alternatives
need not affect ultimate explanations of the contact function
of calling. However, they may reflect important differences
in the proximate mechanisms governing call production.
In this article, we report on a study of the contact barks of
wild chacma baboons. Contact barks are loud, harmonically
rich calls (see Figure 1) that are given at low rates, are
transmitted over considerable distances (>200 m), and often
are accompanied by overt behaviors consistent with a
contact function, such as visual scanning and climbing to an
elevated position from which to survey the area (Byrne,
1981; Cheney et al., 1996; Hall & DeVore, 1965; Ransom,
1981). At times, widely separated individuals can be heard
calling at about the same time, creating the impression that
they are exchanging calls.
In the only systematic research to date on the contact
barks of chacma baboons, Cheney et al. (1996) found that
adult females called primarily when alone or when moving
in the last one third of the travel progression, suggesting that
separation from the group at large (or at least the risk of
PROXIMATE FACTORS MEDIATING BABOON "CONTACT" CALLS
barks to address more systematically the proximate factors
mediating contact calls in female baboons and their infants.
We begin with behavioral observations examining the
factors that affect call production in adult females. We then
report on the production of contact barks by infants and on
the responses of mothers from natural observations and a
Part 1: Proximate Factors Affecting Call Production in
Figure 1. Narrow-band (21-Hz) spectrogram of baboon contact
barks for an adult female and an infant.
becoming separated) was the principal determinant of call-
ing. However, females occasionally gave contact barks
when they were in the midst of the group, suggesting that
calling also may have been motivated by separation from
particular social companions. Given the general importance
of kinship in the regulation of spatial and social relationships
among female baboons, Cheney et al. examined whether
separation from adult female kin might underlie calling by
testing whether females answered the calls of their adult
female relatives. Although the authors could not conclude
that separation from adult female kin never motivated
females to call, they found little evidence that females
selectively replied to the calls of their female relatives.
Cheney et al.'s (1996) findings suggest that female
baboons did not answer the calls of social companions but
instead called primarily only when they themselves were at
risk of becoming separated from the group. However, there
are other individuals from whom separation might be
important and with whom a mechanism of selective call
exchange might be especially beneficial, namely, infants.
Female baboons invest heavily in their infants' develop-
ment, and selection should favor mechanisms that allow
mothers to efficiently maintain contact with their infants to
provide support when needed. Such mechanisms might be
particularly important when the risks of predation and
infanticide are high, as they are for baboons in the Okavango
Delta (Busse, 1980; Busse & Hamilton, 1981; Palombit et
al., in press; Palombit, Seyfarth, & Cheney, 1997; Tarara,
In fact, female baboons have been found to produce
contact barks at times when their infants are away from them
(Cbeney et al., 1996; Ransom, 1981), and they often behave
at such times as though they are trying to locate their infants.
Likewise, infants sometimes call and appear to be anxious
when they become separated from their mothers. These
observations suggest that the production of contact barks in
adult female baboons may be motivated by separation not
only from the group at large but also from their infants. They
also suggest that mothers and infants may exchange calls to
In this article, we extend the study of baboon contact
Study site and subjects. We conducted research on free-ranging
baboons (Papio cynocephalus ursinus) in the Okavango Delta of
northern Botswana, a vast wetland created by seasonal flooding of
the Okavango River. The habitat is a mixture of grassy floodplains
and wooded "islands" that rise a few meters above the floodplains
and in years of heavy flooding are completely surrounded by water.
Because of low flood levels during this study, the islands were
surrounded by open grasslands.
Baboons at this site have been studied since 1977 (Bulger &
Hamilton, 1988; Busse & Hamilton, 1981; Cheney, Seyfarth, &
Silk, 1995; Rendall, Seyfarth, Cheney, & Owren, 1999). Hence,
they are fully habituated to human observers and are readily
identified on an individual basis. The matrilineal relatedness of all
natal baboons is known. Subjects were the 22 adult females of one
group that numbered approximately 75 individuals at the time of
this study (February 1996-March 1997). Adult females could be
ranked in a stable, linear dominance hierarchy on the basis of the
outcome of approach-retreat interactions (Silk, Seyfarth, & Che-
Observational protocol. The goal of behavioral observation of
the adult females was to identify factors responsible for call
production, including in particular the importance of separation
from their infants versus separation from the group at large.
Observational data were gathered during 1-hr "focal animal"
samples (Altmann, 1974) on the adult females sampled in random
order. Observations were conducted throughout the day but were
concentrated between 6 a.m. and 2 p.m. because our routine was to
locate and then follow the baboons for 5-7 hr beginning early in the
day as they left their sleeping site. A total of 734 hr of data were
gathered, split roughly evenly among the 22 females. Because
several females gave birth during the course of the study, our
sample included data from 27 different immature offspring (infants
or young juveniles) ranging in age from birth to 671 days. At the
beginning of the study, 15 females had young infants between 0 and
6 months old, 5 females had older infants or young juveniles
between 7 and 18 months old, and 2 females had no immature
offspring. Three infants died or disappeared during the study. One
is suspected to have been the victim of leopard predation, another
died from wounds inflicted by an infanticidal male, and the third
disappeared with no prior signs of ill health.
Observations included continuous data on the focal female's
general activity state (e.g., resting, foraging, grooming), social
behavior, and production of contact barks. In addition, point
samples were taken at 15-min intervals during the 1-hr samples and
included a global positioning satellite (GPS) reading (accurate to
within 100 m), a measure of habitat visibility, and estimates of the
focal female's proximity both to the group and to her infant.
We used the GPS readings to establish the female's average rate
of travel and location within the home range during each 15-min
interval. We predicted that both factors would affect spacing and
calling behavior. Travel rates for each interval were calculated
38 RENDALL, CHENEY, AND SEYFARTH
using trigonometric relations to determine the straight line distance
between GPS readings from successive point samples. The result-
ing value was then multiplied by 4 to express the distance traveled
during that 15-min interval as an hourly rate (m/hr). To establish
the female's location within the home range, we used the complete
sample of GPS readings to create a scatter plot map of the group's
ranging activities over the course of the study. Using this map and a
scaled aerial photo of the region, we delineated a core region
encompassing two thirds of all GPS readings and many of the
group's preferred sleeping, resting, and feeding sites. Locations
inside of this core region were labeled central, whereas locations
outside of this core region were labeled peripheral.
At each point sample, we scored the type of habitat where the
focal female was located. Experience indicated that certain gross
distinctions in habitat corresponded to different levels of visibility,
which we predicted would also affect spacing and calling behavior.
We delineated four habitat types reflecting a continuum from high
to low visibility: (a) floodplain--the open grasslands; (b) island
edge--the transitional area surrounding an island, separating it
from the floodplain; (c) island wood--forested areas on the raised
islands; and (d) island scrub.--areas of dense bush in the middle of
We scored a female's proximity to the group in one of two
categories: either "with the group" if she was within 50 m of at
least one other adult or "alone" if she was not. Females that were
not within 50 m of another adult still may have been within 50 m of
one or more juveniles or subadults and thus were not always truly
alone. However, our observations suggested that separation from
other adults by more than 50 m precluded visual contact in most
habitats and posed a real risk of losing track of the direction of
We also scored a female's proximity to her infant in two
categories, but we used a different criterion of separation. Mother
and infant were scored as "in proximity" if they were within 5 m of
one another. Otherwise, they were scored as "separated." Of
course, mothers that were more than 5 m from their infants were
not always out of sight of their infants. However, several factors
made it difficult to evaluate separation between mother and infant
in a fashion more comparable to that between a female and the rest
of the group. For example, it was relatively easy to identify a
baboon as an adult at distances up to (and beyond) 50 m, but it was
considerably more difficult to locate and determine the individual
identity of small infants at similar distances because of their size. In
fact, in low-visibility habitats, it was sometimes difficult (for us,
and potentially also for mothers) to see and identify small infants at
a distance of more than a few meters. Thus, we evaluated
mother-infant separation only with respect to 5-m proximity.
The only departure from application of this measure of mother-
infant separation concerned episodes of calling. Each time a female
produced a contact bark, we immediately scored her proximity both
to the group and to her infant. The former was scored as outlined
above. In scoring the latter, however, we made a special effort to
establish whether the infant was truly out of sight by searching the
surrounding area. It is important to note that this disparity in our
evaluations of mother-infant separation between baseline condi-
tions and episodes of calling was conservative with respect to the
research hypothesis. That is, if separation from one's infant is an
important determinant of calling, then a female should call more
when she is separated from her infant than when the two are in
proximity to each other. Our measure of separation in baseline
conditions using the 5-m criterion, however, greatly overestimated
the proportion of time that mothers and infants were truly out of
sight. As a result, our calculation of the rate at which females called
in this condition was necessarily low. This served to reduce the
difference between the rates of calling when mothers and infants
were in proximity versus separated and made it harder for us to
detect an effect of infant separation on call production.
We conducted the statistics by using the Number Cruncher
Statistical System (Hintze, 1989) and SYSTAT (Wilkinson, 1992)
software packages. All tests were two-tailed except where noted.
Because of a strongly right-skewed distribution, average travel
rates were square-root transformed before statistical testing.
A total of 328 calls were recorded from 17 different
females during focal sampling. Some females contributed
many calls, whereas others contributed very few calls.
Although contact barks are given singly, females often call
repeatedly over short intervals such that their calls are
clumped in time (Cheney et al., 1996). To improve the
independence of calling episodes, we lumped calls produced
by the focal female within each 15-min interval into a single
calling bout, yielding 86 different bouts of calling. Hence,
our primary measure of calling behavior in the analyses was
whether or not females gave at least one contact bark in a
given 15-rain interval. In cases in which females gave
multiple calls in a given interval, we used the number of
calls produced as a supplemental measure of calling. Of the
86 bouts of contact barks, 5 were by females with no
immature offspring in the group at the time. We omitted
these cases from analyses related to infant proximity.
Factors affecting the production of contact barks. The
majority of contact bark bouts could be accounted for by
females' proximity to the group or to their infants: Of the 81
bouts involving females with immature offspring, 74 oc-
curred when they were separated from the group (n = 21),
their infants (n = 19), or both (n = 34). Females did not
invariably call when separated from the group or their
infants. Females called approximately 1 time in every 6 that
they were separated from the group and 1 time in every 30
that they were separated from their infants. However, in only
7 cases did females call when they were both with the group
and in proximity to their infants. Thus, the rate of calling
was significantly higher when females were alone than when
they were with the group (Wilcoxon T = 4, p < .001, N =
17; see Figure 2), and despite the conservative effect of our
scoring of mother-infant separation during baseline condi-
tions versus episodes of calling, the rate of calling was also
higher when females were separated from their infants than
when they were in proximity to them (one-tailed Wilcoxon
T = 37, p = .0545, N = 16; see Figure 2).
Although separation from the group or from an infant
accounted for most instances of calling, the probability of
calling varied according to several other factors. For females
that called more than once, the production of contact barks
was inversely related to dominance rank (rs = -.60, p < .05,
N = 13). Thus, lower ranking females called more fre-
quently than higher ranking females, replicating an earlier
finding (Cheney et al., 1996). Call production also varied
according to the age of the infants, with mothers of older
infants calling more frequently than mothers of younger
infants, analysis of variance: F(1, 2934) = 4.25, p < .05. In
addition, females called at higher rates in peripheral as
opposed to central areas of the group's home range (Wil-
PROXIMATE FACTORS MEDIATING BABOON "CONTACT" CALLS 39
Figure 2. The mean (+SE) rate of calling by adult female
baboons as a function of their proximity to the group (with the
group or alone) and to their infants (in proximity or separated).
coxon T = 32,p < .05, N = 17), and this was especially true
when females became separated from the group. Thus, when
alone and on the periphery of their range, females called at
rates that were nearly double those when they were alone
and in the core of their range (Wilcoxon T = 11, p < .01,
N = 15).
Across females, call production also varied with habitat,
×2(3, N = 17) = 10.55, p < .05. The rate of calling was
highest in the island scrub habitat, where visibility was low,
and lowest on the floodplain, where visibility was compar-
atively high. Calling was also more likely to occur as the rate
of travel increased, F(1, 2934) = 18.19, p < .001; the
average travel rate during intervals in which females called
(554 m/hr) was almost double that of intervals in which no
calls were given (316 m/hr). These two factors were
interrelated: Analysis of variance revealed significant varia-
tion (p < .05) in the rate of travel across the four habitat
types for 17 of the 22 females. For 15 of these females, travel
rates were highest in island scrub habitats, where visibility
was low. In addition to affecting the probability that a female
would call, the rate of travel also affected the number of calls
a female produced in a given 15-min interval. For those
intervals in which a female called, the number of calls that
were produced was positively correlated with the rate of
travel (rs = .33, p < .01, N -- 86). The number of calls
produced did not vary systematically with any of the other
factors (e.g., habitat type, location in the range, dominance
To evaluate the relative importance of each of the above
factors in the production of contact barks, we used the data
on all females combined to conduct a stepwise logistic
regression with calling (yes or no) as the dependent variable
(see Table 1). Only proximity to the group and proximity to
one's infant made statistically significant (p < .05) contribu-
tions to explaining variation in calling. The rate of travel and
location in the range were nearly significant additional
predictors of calling. Noteworthy in this respect is the fact
that, in five of the seven instances of calling unaccounted for
by separation from the group or their infants, females were
at the time traveling rapidly (853 m/hr, or 2.5 times the
overall average rate of travel of 322 m/hr ) on the periphery
of their range.
Separation from the group. On average, females were
separated from the group approximately 10% of the time.
However, there was considerable variation among individu-
als in the percentage of time that they spent alone (ranging
from less than 4% to 22%), which was significantly corre-
lated with female rank: Low-ranking females were alone
significantly more often than were high-ranking females
(rs = .50, p < .01, N = 22). The likelihood of the females
becoming separated from the group also varied as a function
of travel rate: The rate of travel was significantly greater
during intervals in which the females were alone than during
intervals in which they were with the group (Wilcoxon T -~
6, p < .0001, N = 22). Of course, it is possible that the
increased rate of travel in intervals when females were
separated from the group was a consequence rather than a
cause of being separated, reflecting efforts to locate and
catch the group rather than being a factor contributing to the
separation. If true, then the rate of travel in the interval
immediately preceding that in which females were alone
should not have differed significantly from the baseline rate.
In fact, this was not true. Rates of travel in the interval
immediately preceding separation were significantly higher
than the baseline rate of travel for each of the 22 females
(Wilcoxon T = 0,p < .0001, N = 22), on average more than
double the baseline rate. Hence, high rates of travel are
better interpreted as a cause rather than an effect of
A female's proximity to the group also varied as a
function of her location within the group's home range: Of
the 22 females, 16 were alone more often in central than
peripheral areas of the range (sign test: p = .052, N = 22).
Finally, across females, habitat type also affected the likeli-
hood of becoming separated from the group. Females were
alone more than expected in island scrub habitats and less
than expected on the floodplain, X2(3, N = 22) = 34.00,
p < .0001.
Results of Stepwise Logistic Regression Using the
Production of Contact Barks (Yes or No) by Adult Female
Baboons (N = 22) as the Dependent Variable
Variable ×z p
Proximity to the group
Proximity to one's infant
Rate of travel
Location in the range
Overall stepwise model
Note. All variables that varied significantly with calling in
univadate tests were entered in the analysis. Variables are listed in
the order in which they were entered in the stepwise model.
adf = 1, N = 2,936. Ddf = 7, N = 2,936.
40 R~NOALL, CHENEY, AND SEYFARTH
Separation from infants.
approximately 53% of their time in proximity to (within a
distance of 5 m from) their mothers. Not surprisingly, the
percentage of time spent in proximity to mothers varied
considerably with the infants' age. For purposes of analysis,
we divided the infants into two age categories: younger
infants less than 6 months of age and older infants between 6
and 22 months of age. Averaged across individuals, younger
infants were in proximity to their mothers 91% of the time,
whereas older infants were in proximity to their mothers
only 40% of the time.
Younger infants were especially likely to be in proximity
to their mothers as the rate of travel increased: For these
infants, the rate of travel when in proximity to their mothers
(307 m/hr) was more than twice that when they were
separated (142 m/hr; Wilcoxon T = 1, p < .01, N = 11).
There was no significant effect of travel rate on the
likelihood that older infants would be in proximity to their
mothers (Wilcoxon T = 96, ns, N = 20). The likelihood that
these older infants would be in proximity to their mothers
did vary significantly according to their mothers' proximity
to the group. The percentage of time spent by older infants in
proximity to their mothers was greater when the mothers
were with the group than when the mothers were alone
(Wilcoxon T = 30, p < .01, N = 20). In other words, older
infants were more likely to be separated from their mothers
at times when their mothers were themselves separated from
the group. Younger infants were never separated from their
mothers when their mothers were separated from the group.
These results suggest that infants' age may be the best
predictor of the likelihood that they will be separated from
their mothers. Indeed, the data on calling tended to confirm
this relation. Only 1 bout of calling was recorded from a
mother separated from a younger infant, whereas 52 bouts of
calling were recorded from mothers separated from older
Infants in our sample spent
The results indicate that in adult female baboons, separa-
tion both from the group and from their infants affected the
production of contact barks. These two factors accounted for
more than 85% of calling bouts. This result does not imply
that females called invariably whenever they became sepa-
rated from the group or their infants. In fact, females called
approximately only 1 time in every 6 that they were
separated from the group and only 1 time in every 30 that
they were separated from their infants. This substantial
difference in calling rates between group and infant separa-
tion was probably largely artifactual, due to our conservative
scoring of mother-infant separation during baseline observa-
tions (>5 m). Thus, in many cases when females were
scored as being separated from their infants, they probably
could still see their infants or knew where they were located.
Hence, the rate at which females called when they were truly
separated from their infants was probably much higher than
our estimate reflects and was probably closer to the rate at
which they called when they were separated from the group.
Although females did not invariably call when they were
separated from the group or their infants, they seldom called
if they were not separated from one or the other.
The finding that separation from the group often appeared
to motivate calling supports an earlier study in which
females called most often when traveling alone or in the last
one third of the group progression (Cheney et al., 1996). The
production of contact barks was also affected by habitat
visibility, travel rate, and the caller's dominance rank.
However, these factors appeared to influence calling indi-
rectly by affecting the probability that females would
become separated from the group. Thus, the probability of
becoming separated was greater at higher rates of travel, the
rate of travel was highest in low-visibility habitats, females
became separated more often in low-visibility habitats, and
lower ranking females were more often separated than were
higher ranking females.
Females also called more in peripheral areas of their
range, particularly when they became separated from the
group. That females were separated from the group less
frequently in peripheral as opposed to central parts of their
range suggests that they may have actively avoided separa-
tion in these areas and that when they did become separated
in peripheral areas there was a lower threshold to calling.
This combination of findings might be explained by higher
risks associated with the females becoming separated from
the group in peripheral areas, where individuals are less
likely to reencounter their own group by chance, are
potentially more likely to encounter neighboring groups, and
are less familiar with predator habits and refuge sites.
Part 2: Processes Mediating Contact Between
Mothers and Infants
Behavioral observations indicated that female baboons'
contact barks were motivated in part by separation from
their infants, suggesting that these barks may function to
maintain contact between mothers and infants. However, it
is not clear how calling by mothers might facilitate such
contact. For example, do mothers and infants actively
exchange calls? Do mothers even recognize the calls of their
own infants? To examine these questions, we conducted
simultaneous observations and a playback experiment on
mothers and infants.
Simultaneous observations of mothers and infants. In a subset
of our behavioral observations of mothers, we conducted simulta-
neous observations of infants to evaluate the possibility that the
mothers and infants might exchange contact barks. While one
observer followed a particular focal female, a second observer
followed that female's infant. These observations focused on
6-18-month-old infants, because at that age, infants were fre-
quently separated from their mothers (see Part 1). Because this
subset of our obseivations was conducted using two observers, it
was possible for us to more accurately evaluate proximity between
mothers and infants. Proximity between mothers and infants was
evaluated at 3-rain intervals during 1-hr samples and scored in one
of five categories: 0-5 m, 5-15 m, 15-50 m, >50 m, or out of sight.
PROXIMATE FACTORS MEDIATING BABOON "CONTACT" CALLS 41
Observations of mothers were conducted as described in Part 1,
except that we now also recorded females' responses to all audible
contact barks by infants. We then synchronized each female's
record of responses with the simultaneous record of contact barks
produced by her infant.
Playback experiment. To more systematically examine moth-
ers' responses to contact barks, we designed a series of matched
playback trials in which females heard either their own infants'
contact barks or those of an unrelated infant of the same age and
sex. Playback trials were conducted in two different contexts: (a)
when the female was with the group and (b) when the female was
alone (i.e., more than 50 m from any other adult). The basic design
of the experiment, therefore, involved four conditions.
The contact barks used in the experiments were originally
recorded on Sony Type IV metal tapes using a Sony WM-D6C
Professional Walkman cassette recorder and a Sennheiser ME 80
directional microphone (with K3U powering module). Recordings
were digitized at 22.05 kHz on a laptop computer by using the
Canary software package (Version 1.2.1; Charif, Mitchell, & Clark,
1995). Playback stimuli were then constructed by using two calls
from the same infant recorded at the same age. Because playback
trials were conducted over a 9-month period, we updated our
stimuli for particular infants every 3 months to control for potential
developmental changes in the acoustic structure of their contact
Our set of playback stimuli consisted of 23 pairs of calls (46
different exemplars) originally recorded from 15 different infants.
Each playback stimulus consisted of 2 calls from the same infant
separated by a 1-s interval of silence. We chose this stimulus
arrangement for several reasons. First, in field playback experi-
ments, it is often difficult to predict short-term changes in ambient
noise created by environmental disturbances or conspecific (or
heterospecific) signaling. Therefore, we felt that a stimulus consist-
ing of 2 calls would better ensure that subjects would hear the
contact barks we were broadcasting. Furthermore, it was our
impression that rapidly repeated calls reflect greater distress on the
part of the caller (see below). A stimulus consisting of 2 closely
spaced calls, therefore, might be more likely to evoke a strong
response from listeners.
The choice of subject and experimental context for each trial was
determined by the ongoing regimen of behavioral sampling. The
order of "own" versus "unrelated" infant trials was balanced
across subjects. Trials were conducted only if there had been no
contact barks from anyone in the group in the preceding 30 min and
only if the mother's infant (and, in the case of control trials, the
unrelated infant whose calls were to be played) had not been seen
by observers for at least 10 min. The latter criterion was included to
ensure that the infant whose calls were to be played could
realistically be calling from some distance away. Because mothers
and infants were frequently separated from one another without
calling by either one, these conditions occurred regularly. Calls
were played through a Nagra DSM speaker positioned an average
distance of 30 m from the subject in the direction from which the
infant was last seen. Amplitude settings were the same for all
playback trials and were chosen to simulate an infant calling at a
distance of approximately 100 m.
We conducted 89 playback trials on 19 different females (that all
had an infant between 6 and 18 months old at the time) over a
9-month period. On average, a playback trial was conducted once
every 3 days, a rate far lower than the rate of natural call production
by infants (approximately two bouts per hour). Playback trials were
split roughly evenly between own-infant (n = 46 trials) and
unrelated-infant (n = 43 trials) conditions. Because females were
not often separated from the group, more trials were conducted in
the "with group" context (n = 56 trials) than in the "alone"
context (n = 33 trials). In some cases, a female served as a subject
for more than 1 trial in a given condition. In these cases, the subject
either heard the calls of a different infant (in the case of
unrelated-infant trials) or heard a different pair of calls from her
own infant (in the case of own-infant trials). In matched-
comparison tests of these data, females' responses were averaged
across the trials.
Simultaneous observations of mothers and infants.
conducted 230 one-hour simultaneous samples of 19 differ-
ent females and infants. Like adults, infants produced
contact barks in bouts of repeated calls. During follows of
focal females, we recorded 556 bouts of infant contact barks
that were clearly audible to the females themselves. Of
these, 93 bouts were produced by the focal females' own
infants. Calling infants were typically widely separated from
their mothers, with calling rate increasing as the distance
between infants and mothers increased, X2(3, N = 19) =
65.53,p < .001 (see Figure 3).
Mothers' responses to infants' contact barks varied consid-
erably. Mothers often showed no overt response, or they
simply oriented briefly in the direction of the caller. At other
times, however, mothers appeared to walt for calling infants
by sitting down and orienting in the direction of the caller or
remaining stationary even as the rest of the group continued
to move away. Occasionally, mothers responded more
dramatically by climbing to an elevated position and scan-
ning the area or approaching the caller. On a few occasions,
mothers retrieved infants that called persistently. Rarely,
however, did females respond with contact barks themselves.
For purposes of analysis, mothers' responses were grouped
into three categories: orient in the direction of the caller;
stop, wait, or move toward the caller; and produce a contact
bark in reply within 5 min. The latter 5-min criterion for a
vocal reply was generous but was used to facilitate compari-
son with an earlier study of females' responses to the contact
barks of adult kin (Cheney et al., 1996). Because the sample
of 93 bouts of contact barks by the focal females' own
infants was too small when it was distributed across 19
different mother-offspring pairs to permit analyses by
individual females, the data on females' responses were
Across mother-infant pairs, females were significantly
more likely to orient toward their own calling infants than
they were to other calling infants in the group, X2(1, N =
19) = 20.42, p < .001 (see Figure 4). Females also were
significantly more likely to stop, walt, or move toward their
own infants, ×2(1, N = 19) = 21.64, p < .001 (see Figure 4),
and the probability that they would do so was significantly
greater when the infants gave multiple calls (Mann-Whitney
U = 4,804, p < .001, N = 19). Females produced contact
barks within 5 rain of hearing infants call on only nine
occasions, two following calls by their own infants and
seven following calls by other infants. This sample of calling
responses was too small to permit any statistical test.
Mirroring the results from natu-
rally occurring bouts of calling, females' responses to
playbacks of contact barks varied considerably, although
42 RENDALL, CHENEY, AND SEYFARTH
Figure 3. The percentage of contact barks produced by infant baboons as a function of the distance
separating them from their mothers compared with the percentage of time infants spent at the same
distances from their mothers. OOS = out of sight (see the Method section).
their responses to playbacks were typically stronger than
their responses to naturally occurring calls. Females oriented
in 52 of 89 trials (58%). Their overall mean latency to orient
was 1.22 s, or within 0.25 s of hearing the second call in the
playback sequence. In 27 trials, females showed qualita-
tively exaggerated responses, abruptly abandoning their
current activity (foraging or moving) and scanning rapidly
and repeatedly in the direction of the speaker. In many trials
Figure 4. The percentage of naturally occurring infant baboon contact barks to which mothers
responded either by orienting toward the caller or by stopping, waiting, or moving toward the caller
(see the Results section for descriptions). Responses are plotted according to whether the caller was
the mother's own infant or another infant in the group.
PROXIMATE FACTORS MEDIATING BABOON "CON-TACT" CALLS 43
speaker; stood bipedally or climbed to a better vantage point
in a tree or on a stump to scan in the direction of the speaker;
or stopped in their tracks and sat down to wait, facing in the
direction of the speaker.
Females were clearly able to recognize the contact barks
of their own offspring. Subjects oriented in significantly
more own-infant trials than in unrelated-infant trials (Wil-
coxon T = 11.5, p < .05, N =
Furthermore, their orienting responses were both faster and
longer to the calls of their own infants as opposed to those of
unrelated infants. In the unmatched comparison across all
subjects and trials, the latency to orient was shorter (Mann-
Whitney U = 142, p < .01, N = 19) and the duration of
orientation longer (U = 399, p < .01) in own-infant trials
than in unrelated-infant trials (see Figure 6). Results were
the same in the within-subjects matched comparison, in
which again the latency to orient was shorter (W'dcoxon T =
16, p < .05, N = 14) and the duration of orientation was
longer (T = 19, p < .05) in own-infant trials as opposed to
unrelated-infant trials. Moreover, of the 27 trials in which
subjects showed qualitatively exaggerated responses, 20
were own-infant trials, ×2(1, N = 19) = 7.78, p < .01.
15), females moved--and even ran--toward the
14; see Figure 5).
Mothers responded by producing contact barks them-
selves in 16 of 89 trials (18%). However, the probability of
calling was unaffected by infant identity (Wilcoxon T =
24.5, ns, N = 14, including 6 ties; see Figure 5). Females
called in 9 own-infant trials and in 7 unrelated-infant trials.
Hence, females did not selectively "answer" the contact
barks of their own infants.
Instead, the probability of producing contact barks within
5 min of playback was strongly affected by the females' own
state of separation from the group. Females were signifi-
cantly more likely to call in trials in which they were alone
(n = 13) than in trials in which they were with the group
(n = 3; Wilcoxon T = 0, p < .01, N = 13; see Figure 5).
Across subjects and trials, there was no association between
orienting and calling responses, ×2(1, N = 19) = 0.86, ns,
and although a few females called within seconds of hearing
the contact barks of an infant, on average, the delay between
broadcast of infant calls and the production of contact barks
by females was 2 min, 25 s. These two findings suggest that
females' production of contact barks was independent of
their interest in the contact barks of infants (as manifested by
their orienting and moving responses).
Figure 5. The percentage of trials in which female baboons responded to experimental playback of
infant contact barks by either orienting toward the speaker playing the calls or producing a contact
bark themselves. Orienting and vocal responses are plotted as a function of the identity of the infant
whose calls were broadcast (one's own infant or an unrelated infant) and the female's proximity to the
group (with the group or alone). Orienting responses were scored from frame-by-frame analysis of
the videotape record and were defined as a change in the subject's orientation toward the playback
speaker within 10 s of stimulus presentation.
44 RENDALL, CHENEY, AND SEYFARTH
Figure 6. The mean (+SE) latency and duration of orienting responses by female baboons in
playback trials to the contact barks of their own or unrelated infants. Latency and duration measures
were calculated by using sound onset of the first call in the playback sequence.
Results of both simultaneous observations of mothers and
infants and the playback experiment indicate that mothers
recognized their infants' contact barks and were motivated
to locate the calling infants. In many playback trials
involving their own infants, females immediately abandoned
their current activity and either walked or ran toward the
speaker; climbed a tree to scan toward the speaker; or sat on
the ground facing the speaker, apparently waiting for their
Overall, females responded more strongly and more
frequently to the playback sequences than to naturally
occurring calls. This was probably due to the fact that the
sequences used for playback were deliberately designed to
mimic infant distress. Observations of naturally occurring
calls indicated that mothers were more likely to respond and
search for their infants when their infants gave multiple
calls. The increased incidence of calling by mothers follow-
ing playback trials probably occurred because many trials
were deliberately conducted when subjects were separated
from the group. Under natural conditions, females gave
more contact barks when they were separated from the
group than when they were in proximity to other adults.
Females were separated from the group during only 10% of
all behavioral observations; however, 37% of playback trials
were conducted in this condition.
Although mothers were clearly interested in the contact
barks of their infants and were motivated to locate them,
they did not exchange calls with their infants. Females
seldom responded to the contact barks of their infants by
producing calls themselves. Moreover, despite their strong
behavioral responses to the playback sequences, mothers
still did not answer their infants' calls. They were as likely to
give a contact bark in response to unrelated infants' calls as
they were to their own infants' calls, and their production of
calls seemed to be unrelated to their attempts to locate their
infants. In fact, mothers' call production appeared to be
motivated primarily by their own state of separation from
the group rather than that of their infants.
Females' responses to playbacks of their infants' contact
barks, therefore, were comparable to their responses when
the contact barks of adult female kin were played (Cheney et
al., 1996). In those trials, too, subjects gave vocal responses
primarily when they themselves were separated from the
group. Similar responses have also been reported for the
long-distance contact calls of other primate species. For
example, both Dittus (1988) and Robinson (1982), studying
toque macaques (Macaca sinica) and wedge-capped capu-
chins (Cebus nigrivittatus), respectively, found that group
members seldom answered the loud calls of isolated individu-
als, despite showing an interest in their calls by scanning and
sometimes moving in the direction of the caller. Likewise, in
PROXIMATE FACTORS MEDIATING BABOON "CONTACT" CALLS 45
a playback experiment on rhesus monkeys, Rendall and
colleagues found that adult females seldom called in re-
sponse to the contact calls ("coos") of their adult female
kin, despite orienting strongly to their calls and even
occasionally approaching the playback speaker (Rendall,
1996; Rendall et al., 1996).
We now briefly consider the two main issues raised at the
outset of this article: (a) What factors motivate the produc-
tion of contact calls? (b) What processes mediate the contact
function of calling? In answer to the first question, it is clear
that separation both from the group and from infants
motivates calling by adult females. In addition, a variety of
other social and ecological factors (e.g., age, social rank,
habitat type, travel rate, and location within the home range)
appear to indirectly influence call production by affecting
the probability that individuals will become separated from
the group or from particular social partners.
The second question is more difficult to answer. Our
results suggest that the contact function of loud barks by
adult female baboons and their infants is not mediated by a
system of selective call exchange. Mothers did sometimes
call when separated from their infants, and infants also
sometimes called when separated from their mothers. How-
ever, we found no evidence that calling by mothers and
infants was coordinated. Instead, contact between mothers
and infants appeared to be achieved by occasional retrieval
by mothers of calling infants.
There are at least two possible explanations for these
findings. First, the fact that mothers did not selectively
answer their infants' calls may indicate that vocal replies are
not the most effective means to achieve a reunion. When the
risks of predation and infanticide to unaccompanied infants
are high (Busse, 1980; Busse & Hamilton, 1981; Palombit et
al., 1997, in press; Tarara, 1987), it may be more adaptive for
mothers simply to retrieve their infants rather than to answer
them. This explanation is not entirely satisfactory, however,
because mothers did not invariably retrieve calling infants,
and when they did, it was only after persistent calling by the
infants, which would seem merely to advertise the infants'
vulnerability and therefore increase the likelihood of preda-
tion or infanticidal attack. Furthermore, this explanation
cannot account for the finding that adults in several species
also do not answer each other's calls (Cheney et al., 1995;
Dittus, 1988; Rendall, 1996; Rendall et al., 1996; Robinson,
An alternative explanation for the lack of selective vocal
replies in this study and the others noted above may be that
individuals do not appreciate the plight of others that are
separated or at least how calling in reply might help them to
rejoin the group. The ability to understand another's perspec-
tive and how one's own behavior (and communication) can
influence that of others is deemed to be an important
component of human social cognition and language (Grice,
1957). It has become a focal point of research on nonhuman
species because of the potential evolutionary implications.
Converging evidence from various aspects of behavior and
communication suggests that monkeys (though perhaps not
apes) do not fully appreciate the behavioral and mental
perspectives of others (reviewed in Cheney & Seyfarth,
1990, 1996; Povinelli, 1993; TomaseUo & Call, 1997). The
fact that mothers in this study failed to answer their infants'
calls despite calling when separated themselves is consistent
with these findings and suggests that call production may be
governed more by internal states associated with one's own
condition of separation than by cognitive evaluations of the
circumstances of others.
On the surface, a system like this that lacks vocal replies
seems to present an evolutionary challenge because calling
does not appear to provide obvious benefits to callers.
However, self-motivated calling could be beneficial if, on
many occasions, multiple animals were simultaneously at
risk of becoming separated and were therefore calling with
respect to their own state, thereby indirectly maintaining
contact with one another. Exactly this sort of synchroniza-
tion of activity and calling does seem to account for the
contact function of many forms of quiet contact call given at
high rates during dispersed foraging and travel in other
species. It is possible that loud calls given by isolated
individuals are an extension of this system, maintained in the
absence of significant cognitive mediation by the frequent
synchronization of activity and motivational states of others
and, in the case of infants, by occasional maternal retrieval.
However, this explanation of the proximate mechanisms
governing call production is not entirely satisfactory either.
First, evidence that mothers answered their infants' calls
would not by itself constitute proof that they understood
their infants' perspective or mental state. Some other simple,
contingency-based mechanism could also induce a vocal
response. Second, the fact that mothers often began to search
for their infants after hearing them call could be interpreted
as evidence that they were able to understand their infants'
perspective, at least to some degree.
On the basis of current evidence, then, it is not possible to
determine whether mothers do not answer their infants' calls
because (regardless of what they know about their infants'
perspective) occasional physical retrieval minimizes the
risks to infants or because females do not fully appreciate
their infants' perspective or the effect that answering calls
might have on achieving reunion. Given recent interest in
the cognitive mechanisms underlying the behavior and
communication of nonhuman species, future research would
profit from experimental methods that can distinguish be-
tween these alternatives under naturalistic conditions.
Altmann, J. (1974). Observational study of behavior. Behaviour,
Biben, M. (1993). Recognition of order effects in squirrel monkey
antiphonal sequences. American Journal of Primatology, 29,
Boinski, S. (1991). The coordination of spatial position: A field
study of the vocal behavior of adult female squirrel monkeys.
Animal Behaviour, 41, 89-102.
Boinski, S. (1993). Vocal coordination of troop movement among
RI~NI)ALL, CHENEY, AND SEYFARTH
white-faced capuchin monkeys, Cebus capuchinus. American
Journal of Primatology, 30, 85-100.
Bulger, J., & Hamilton, W. J. (1988). Inbreeding and reproductive
success in a natural chacma baboon, Papio cynocephalus
ursinus, population. Animal Behaviour, 36, 574-578.
Busse, C. (1980). Leopard and lion predation upon chacma
baboons in the Moremi Wildlife Reserve. Botswana Notes &
Records, 12, 15-21.
Busse, C., & Hamilton, W. J. (1981). Infant carrying by male
chacma baboons. Science, 212, 1281-1283.
Byme, R. W. (1981). Distance vocalizations of guinea baboons
(Papio papio) in Senegal: An analysis of function. Behaviour,
Charif, R. A., Mitchell, S., & Clark, C. W. (1995). Canary 1.2.1
user's manual. Ithaca, NY: Cornell University Laboratory of
Cheney, D. L., & Seyfarth, R. M. (1990). How monkeys see the
world. Chicago: University of Chicago Press.
Cheney, D. L., & Seyfarth, R. M. (1996). Function and intention in
the calls of non-human primates. Proceedings of the British
Academy, 88, 59-76.
Cheney, D. L., Seyfarth, R. M., & Palombit, R. A. (1996). The
function and mechanisms underlying baboon "contact" barks.
Animal Behaviour, 52, 507-518.
Cheney, D. L., Seyfarth, R. M., & Silk, J. B. (1995). The role of
grunts in reconciling opponents and facilitating interactions
among adult female baboons. Animal Behaviour, 50, 249-257.
Dittus, W. G. (1988). An analysis of toque macaque cohesion calls
from an ecological perspective. In D. Todt, P. Godeking, & D.
Symmes (Eds.), Primate vocal communication (pp. 31-50).
Berlin, Germany: Springer-Verlag.
Gautier, J. P., & Gautier, A. (1977). Communication in Old World
monkeys. In T. A. Sebeok (Ed.), How animals communicate (pp.
890-964). Bloomington: Indiana University Press.
Grice, H. P. (1957). Meaning. Philosophical Review, 66, 377-388.
Hall, K. R. L., & DeVore, I. (1965). Baboon social behavior. In I.
DeVore (Ed.), Primate behavior (pp. 53-110). New York: Holt,
Rinehart & Winston.
Hansen, E. W. (1976). Selective responding by recently separated
juvenile rhesus monkeys to the calls of their mothers. Develop-
mental Psychobiology, 9, 83-88.
Hintze, J. L. (1989). Number Cruncher Statistical System, NCSS,
Version 5.1, reference manual. Kayesville, UT: Author.
Itani, J. (1963). Vocal communication of the wild Japanese
monkey. Primates, 4, 11--66.
Marler, P., & Hobbett, L. (1975). Individuality in a long-range
vocalization of wild chimpanzees. Zeitschriftfur lierpsycholo-
gie, 38, 97-109.
Mitani, J. C., & Nishida, T. (1993). Contexts and social correlates
of long-distance calling by male chimpanzees. Animal Behav-
iour, 45, 735-746.
Palombit, R. A. (1992). A preliminary study of vocal communica-
tion in wild long-tail macaques (Macaca fascicularis): II.
Potential of calls to regulate intragroup spacing. International
Journal of Primatology, 13, 183-207.
Palombit, R. A., Cheney, D. L., Fischer, J., Johnson, S., Rendall, D.,
Seyfarth, R. M., & Silk, J. B. (in press). Male infanticide and
defense of infants in chacma baboons. In C. van Schaik & C. H.
Janson (Eds.), Male infanticide and its implications. Cambridge,
England: Cambridge University Press.
Palombit, R. A., Seyfarth, R. M., & Cheney, D. L. (1997). The
adaptive value of "friendships" to female baboons: Experimen-
tal and observational evidence. Animal Behaviour, 54, 599-614.
PovineUi, D. J. (1993). Reconstructing the evolution of mind.
American Psychologist, 48, 493-509.
Ransom, T. W. (1981). Beach troop of the Gombe. Lewisburg, PA:
Bucknell University Press.
Rendall, D. (1996). Social communication and vocal recognition of
free-ranging rhesus monkeys (Macaca mulatta). Unpublished
doctoral dissertation, University of California, Davis.
Rendall, D., Rodman, P. S., & Emond, R. E. (1996). Vocal
recognition of individuals and kin in free-ranging rhesus mon-
keys. Animal Behaviour, 51, 1007-1015.
Rendall, D., Seyfarth, R. M., Cheney, D. L., & Owren, M. J.
(1999). The meaning and function of grunt variants in baboons.
Animal Behaviour, 57, 583-592.
Robinson, J. G. (1982). Vocal systems regulating within-group
spacing. In C. T. Snowdon, C. H. Brown, & M. E. Petersen
(Eds.), Primate communication (pp. 94--116). Cambridge, En-
gland: Cambridge University Press.
Silk, J. B., Seyfarth, R. M., & Cheney, D. L. (1999). The structure
of social relationships among female savanna baboons in the
Moremi Reserve, Botswana. Behaviour, 136, 679-703.
Smith, J. H., Newman, J. D., & Symmes, D. (1982). Vocal
concomitants of affiliative behavior in squirrel monkeys. In C. T.
Snowdon, C. H. Brown, & M. E. Petersen (Eds.), Primate
communication (pp. 30-49). Cambridge, England: Cambridge
Snowdon, C. T. (1986). Vocal communication. In G. Mitchell & J.
Erwin (Eds.), Comparative primate biology, Vol. 2A: Behavior,
conservation, and ecology (pp. 495-530). New York: Liss.
Snowdon, C. T., & Hodun, A. (1981). Acoustic adaptations in
pygmy marmoset contact calls: Locational cues vary with
distances between conspecifics. Behavioral Ecology and Socio-
biology, 9, 295-300.
Sugiura, H., & Masataka, N. (1995). Temporal and acoustic
flexibility in vocal exchanges of coo calls in Japanese monkeys
(Macaca fuscata). In E. Zimmerman, J. D. Newman, & U.
Jurgens (Eds.), Current topics in primate vocal communication
(pp. 121-140). New York: Plenum Press.
Tarara, E. B. (1987). Infanticide in a chacma baboon troop.
Primates, 28, 267-270.
Tomasello, M., & Call, J. (1997). Primate cognition. Oxford,
England: Oxford University Press.
Wilkinson, L. (1992). SYSTAT, Inc., user's manual. (Available from
L. Wilkinson, 1800 Sherman Avenue, Evanston, IL 60201-3793)
Winter, P., Ploog, D., & Latta, J. (1966). Vocal repertoire of the
squirrel monkey (Saimiri sciureus), its analysis and significance.
Experimental Brain Research, 1,359-384.
Received October 23, 1998
Revision received June 21, 1999
Accepted June 22, 1999 •