Anim Cogn

DOI 10.1007/s10071-017-1105-7

ORIGINAL PAPER

Wild capuchin monkeys anticipate the amount of ripe fruit

in natural trees
Marı́a Paula Tujague1,2 • Charles H. Janson3

Received: 10 January 2017 / Revised: 7 June 2017 / Accepted: 10 June 2017

Ó Springer-Verlag GmbH Germany 2017

Abstract Tropical forests have a high diversity of tree did not explain variation in travel speed at any distance
species which have very low densities and vary across time from a focal tree. Our data imply that, after excluding
in their seasons of peak fruiting and maturation rates. As sensory cues, capuchins appear to anticipate time-varying
evidence of the ability of primates to track or anticipate ripe fruit quantity of natural resources, suggesting that they
changes in fruit production at individual trees, researchers use memory of tree location and anticipation of fruit
have used the increased speed of primate groups toward maturation. We also confirm that speed is a good measure
more rewarding food patches. We analyzed the speed of about expectations of resources, as has been shown in
approach to natural trees of wild capuchin monkeys under previous studies.
the effect of scramble competition, after excluding any
plausible visual, olfactory and auditory cues. We con- Keywords Spatial memory  Fruiting trees  Expectation 
ducted all-day group follows of three habituated capuchin Cebus  Sapajus nigritus
groups at Iguazú National Park, Argentina, collecting data
on ranging behavior and patterns of visits to fruit trees in
relation with their location and fruit availability. Travel Introduction
speed varied according to the expected reward at a feeding
tree, increasing as rewards increased from low values, but Tropical forests have a high diversity of tree species, most
decreasing again at very high values. Also, travel speed of which have very low densities and vary across time in
varied with time of day, decreasing from the time of first their seasons of peak fruiting and maturation rates (Milton
activity as the monkeys became less hungry, and increasing 1981). Fruit-feeding animals with a complex and season-
again toward late afternoon. Measures of unripe fruit cover ally varying diet face the problem of visiting several
resources in the course of each foraging trip (Noser and
Byrne 2010). The ability to remember tree features and to
Electronic supplementary material The online version of this travel efficiently among feeding patches would help them
article (doi:10.1007/s10071-017-1105-7) contains supplementary to minimize energy expenditure by reducing the time and
material, which is available to authorized users.
energy spent searching for food items in a random pattern
& Marı́a Paula Tujague (Janson 1998; Milton 2000).
[email protected] Speed can be a good measure of expectation about a
1
resource to be found (Pochron 2001; Janson and Byrne
Instituto de Biologı́a Subtropical (IBS), Consejo Nacional de
Investigaciones Cientı́ficas y Técnicas (CONICET),
2007). In terms of behavioral changes, researchers have
Universidad Nacional de Misiones (UNaM), Bertoni 85, used the increased directedness or speed of primate groups
CP3370, Puerto Iguazú, Misiones, Argentina toward more rewarding food patches as evidence of
2
Asociación Civil Centro de Investigaciones del Bosque anticipation of both the resource’s location and of its
Atlántico (CeIBA), Puerto Iguazú, Argentina expected value (Sigg and Stolba 1981; Janson 1998; Jan-
3
Division of Biological Sciences, The University of Montana, maat et al. 2006; Noser and Byrne 2007; Janmaat et al.
Missoula, USA 2014).

123
 Anim Cogn

Higher travel speed is expected along trajectories toward (a)

11
known resources than toward unknown ones (Janson and Di
10 Increased perceptual stimulus
Bitetti 1997; Janmaat et al. 2006; Noser and Byrne 2007).

Travel speed toward food patch

This higher speed could reflect a higher motivation to reach 9
a valuable resource (similar to higher rates of activity 8
approaching the time of the next reward in classical con- 7
ditioning, Gallistel and Gibbon 2000), or a reduced trade- 6
off between travel speed and searching ability when search
5
is not needed (e.g., Gendron and Staddon 1983). However,
4
the goal could be ‘known’ because the animals just per-
ceived it, rather than remembering its location from previ- 3

ous experience (Janson and Byrne 2007). If travel speed 2

reflects perception of goal value, we would expect travel 1
speed to increase uniformly with resource productivity 0
(Fig. 1a). Independent evidence is needed to show that the
change in movement speed started at distances that exceed low high
the range at which resources are typically perceived Food patch productivity
(Pochron 2001; Janmaat et al. 2006). (b)
11
Changes in travel speed toward a known resource might Scramble or indirect competition for food
10
also reflect individual strategies during spatial food com-

Travel speed toward food patch

9
petition. A common pattern of movements of capuchin
monkeys is that groups move relatively slowly away from 8

food patches and begin to speed up as they approach the 7

next resource (Janson and Di Bitetti 1997). This could be 6
interpreted as the outcome of scramble competition, as 5
individuals approaching a known goal race each other to be 4
the first to arrive to harvest it (Janson and Byrne 2007). A
3
group’s overall travel speed should increase as the number
2
of individuals participating in the ‘race’ increases. The
number of individuals that can benefit from early arrival at a 1

food source increases initially with its productivity, but will 0

eventually plateau or may even decline as food availability
low high
becomes large enough to allow all group members to feed
Food patch productivity
well. In this case, we would expect a hump-shaped (con-
cave-downward) relationship between travel speed and (c)
11
food availability of the targeted food patch (Fig. 1b). 10 Contest or direct competition for food
Capuchins also showed high levels of contest competi-
Travel speed toward food patch

9
tion (Janson 1985), and speed is expected to increase as the
8
group approaches known resources because subordinates
specifically can avoid aggressive competition within the 7

resource by arriving there before the dominants (e.g., Di 6

Bitetti and Janson 2001). If direct competition were the 5
only factor favoring higher travel speeds, we would expect 4
travel speed to be highest at small food sources (Janson 3
1985, 1996), where subordinates are most directly impac-
2
ted by the presence of dominants, and should decline as
1
resource size increases (Fig. 1c).
Many studies show that primates move toward food 0

patches in a manner that is consistent with goal-directed

low high
behavior implying spatial knowledge about food location
Food patch size (e.g. crown diameter)
(Menzel 1973, 1991; Milton 1981; Garber 1989;

123
 Anim Cogn

b Fig. 1 Hypothesized relationships between group travel speed Finally, using Janmaat’s observational method (2006),
toward food patches and food patch productivity or size. a Sensory Tujague et al. (2016) recorded visits by known capuchin
summation: speed toward a patch will increase in proportion with the
amount of ripe fruit, which is the source of the sensory stimulus; groups to pre-selected target trees at Iguazú National Park,
b scramble or indirect competition: speed will increase from low to Argentina, as a function of the tree’s amount of fruit.
medium values of ripe fruit as more individuals race each other to Tujague and co-workers found that capuchin monkeys
reach the patch early, but will decline at the highest productivities traveled faster to trees that carried fruit in comparison with
where scramble competition is weak; c contest or direct competition:
speed will decline from physically small to larger food patches, as empty ones, accelerating as they were getting closer to the
subordinates incur smaller competitive costs in larger patches and target tree. Results showed that monkeys had a higher
have less incentive to arrive early probability of visiting a tree when it had versus when it did
not have fruit. To analyze whether groups increased their
speed when traveling toward an out of sight tree, they
MacDonald and Wilkie 1990; Roberts et al. 1993, compared speed before, after and at the critical visual
MacDonald 1994; MacDonald et al. 1994; Garber and detection radius. They found that speed increased signifi-
Dolins 1996; Platt et al. 1996; Garber and Pacuilli 1997; cantly as the group approached the tree. These results
Janson and Di Bitetti 1997; Janson 1998; MacDonald and suggest that once monkeys had entered a critical detection
Agnes 1999; Gibeault and MacDonald 2000; Pochron radius around the tree, they had an expectation of whether
2001; Scheumann and Call 2006; Cunningham and Janson or not they would find fruit there. Preliminary analysis did
2007). Also, a few studies analyzed the quantity of reward not provide evidence that the monkeys responded to fruit
to be obtained during movements to food patches (Garber availability by faster travel at the critical detection distance
and Pacuilli 1997, Janson 1998) and some kind of temporal (Tujague et al. 2016).
knowledge about ripening states (Janmaat et al. 2006; The aim of this study is to expand the previous analysis
Cunningham and Janson 2007; Anzelc 2009; Janmaat et al. (Tujague et al. 2016) by looking at changes in group travel
2012; Janson 2016). speed for all distance categories as well as to examine the
There is limited evidence that primates are able to distin- alternative hypothesis that the monkeys might increase
guish between fruiting states of trees of the same versus dif- travel speed when there is less ripe fruit available because
ferent species (Janmaat et al. 2006; Cunningham and Janson of the greater importance of scramble competition when
2007, Janmaat et al. 2012; Ban et al. 2014; Suarez 2014; resource availability is low. We hypothesize that a monkey
Tujague et al. 2016). Janmaat et al. (2006) studied spatial group will travel more quickly during approaches to some
movements of two wild groups of Cercocebus atys atys and fruit trees according to fruit amount and quality. Capuchins
Lophocebus albigena johnstoni during daily visits to previ- might gain from faster travel because of the benefits for an
ously selected fruiting trees and concluded that monkeys used individual to arrive earlier at a food source than other group
information about previous visits to anticipate state and members. This behavior is the embodiment of indirect or
quantity of fruit in future visits. In another study about daily scramble competition (Janson and van Schaik 1988).
movements of Pithecia pithecia, Cunningham and Janson Scramble competition is expected to be especially benefi-
(2007) found that monkeys used a combination of knowledge cial when the available resource is limited relative to the
obtained through the repetitive use of their home range with metabolic needs of the foragers. Because capuchin mon-
information about location and condition of resources, taking keys at Iguazu keep track of the location, size and fruiting
more time to return to unripe fruiting trees. Janmaat and co- status (fruiting vs. non-fruiting) of their food trees (Janson
workers (2012) found that monkeys (Lophocebus albigena 1998, 2007; Tujague et al. 2016), they may be expected to
johnstoni) use synchronized fruiting patterns as an indicator of modulate their approach speed in anticipation of the
the presence of fruit on trees. amount of reward available at the goal. However, we do
In more recent studies, Ban et al. (2014) found that not expect a simple response to reward amount. If the
chimpanzees (Pan troglodytes verus) travelled longer amount of food is very small, then the total reward is small
distances to trees at which they had previously made even to the first forager arriving to the resource. A small
food grunts and had rejected fewer fruits, suggesting that benefit should not favor rapid travel in the face of its
individuals were able to anticipate the amount of fruit potentially large costs (see below). Conversely, when the
that they would find in the trees in future visits. Also, in total benefit is very large, little is gained by being the first
his study about the ecological factors that predicted to arrive as the most productive food trees often carry
foraging decisions of wild spider monkeys (Ateles enough fruit to satiate all group members (e.g., Janson
belzebuth), Suarez (2014) found that monkeys moved to 1988). Therefore, we expect that travel speed would have a
nearby, large canopy trees, in which previous feeding saturating or even possibly quadratic relationship with
success was high and which were visited after an interval expected food amount. In the context of other studies on
of 3.5 days. ecological cognition of capuchin monkeys, we were

123
 Anim Cogn

especially interested in whether they behaved differently 54°300 W). The site is a semi-deciduous forest at the
toward trees with different levels of fruit ripeness, although southern border of the Atlantic forest, characterized by a
we included other metrics of total fruit crop size, both a humid, subtropical climate (Brown and Zunino 1990). Fruit
priori (diameter at breast height, tree crown volume) and a productivity is seasonal, with lowest production between
posteriori measures (total group feeding minutes during a June and August, and highest between October and
particular visit). November (Janson and Di Bitetti 1997). Most of the tree
An assumption common to these arguments is that more species have short fruiting periods (a month or less) with
rapid travel is costly. Energetic costs of animal travel are productivity of species varying from one month to the next
typically expressed as cost per unit distance (e.g., for (Placci et al. 1994). Because critical feeding sites are
capuchins, these are estimated at 123.9 kJ per km in travel heterogeneously distributed (Di Bitetti 2001), remembering
costs, Janson 1988), but more rapid travel may have their location is important to improve foraging efficiency.
additional fitness costs for at least three reasons. First, The tufted capuchin monkey (Cebus = Sapajus nigri-
faster travel limits opportunities to detect potential food tus) is a small-bodied, diurnal primate and is the only
sources or predators (e.g., Gendron and Staddon 1983; primate species commonly found at Iguazú. At Iguazú,
Janson and Di Bitetti 1997). Second, faster travel may capuchins live in multi-male, multi-female groups with
increase the chance of monkeys falling from trees when typically 7–45 independent individuals and predominantly
they misjudge characteristics of their travel path (Janson, male dispersal (Di Bitetti 1997, Janson et al. 2012). About
personal observation). Third, faster individuals end up at 90% of their energy intake is from fruit, although 55% of
the front of a social group, where they may be more their daily activity budget is spent foraging for dispersed
exposed to predation risk as they enter recently unexplored arthropods and other food items (Janson 1990). Capuchins
areas where a predator may be waiting—in rainforests; it is show high variability in space and time in allocation of
not uncommon to find major aerial predators waiting in the feeding time, ranging from 81.2 to 53% of their feeding
vicinity of fruit trees (van Schaik and van Noordwijk time spent on ripe fruit (Chapman and Fedigan 1990).
1989). Thus, we expect that monkeys would travel more In the current study, data were collected for three
quickly only if there were greater benefits to be gained. habituated groups: Macuco (10–14 males, 17–18 females
The benefits of accessing food depend not only on the and a total of 27–32 individuals), Gundolf (7–9 adult
amount and quality of food consumed, but on how hungry the males, 13–14 females and a total of 20–23 individuals) and
animal is: The fitness value of food is higher when an animal Rita (5–12 males, 7 females and a total of 12–19 individ-
is hungrier (e.g., Stephens and Krebs 1983; Janson and Vogel uals), during September–December and March–April of
2006). All else being equal, we expect that animals would 2008–2011. All group members were individually recog-
compete more strongly soon after beginning activity in the nized by facial marks, color pattern and body size.
morning, as they have just finished a 10–12-h fast. Indeed,
fruit-feeding activity is highest in the morning, gradually Data collection
decreasing toward midday, at which time the monkeys
usually devote increasing time toward foraging for inverte- We conducted all-day group follows collecting data on
brates or to resting (during the hottest periods of the year). ranging behavior. Each group was followed for two con-
Regardless of the intervening activity, there is usually a tinuous periods, according to fruiting seasons: Macuco—
second minor peak of fruit-feeding in the late afternoon, two periods of 27 and 40 days, respectively, during
which may denote increased hunger then as well. Thus, we November to December 2008 and November to December
expected that there might be both linear (decreasing with 2009, Gundolf—two periods of 37 and 31 days, respec-
time of day) and quadratic (higher early and late) patterns of tively, during September to October 2009 and November to
travel speed toward fruit trees. Finally, scramble competition December 2010, and Rita—two periods of 35 and 34 days,
should be more intense in larger groups (Janson and Gold- respectively, during March to April 2010 and March to
smith 1995), so we expect that group travel speed would be April 2011. These long group follows were designed
higher in larger groups, all else being equal. specifically to allow the authors to analyze travel behavior
toward fruit patches that the monkeys had definitely visited
previously and for which the monkeys might therefore
Methods have memories of the tree’s size and ripening state (cf.
Janmaat et al. 2006).
Study site and subjects A modified form of the focal tree method (Janmaat et al.
2006) was used to assess the traits of focal trees that the
The study was conducted in Iguazú National Park, a groups visited; we also recorded traits of focal trees that
60,000 ha preserve in northeastern Argentina (25°400 S, were not visited, but they are not analyzed here (see

123
Anim Cogn

Tujague et al. 2016). Observers previously selected focal characteristic of that fruit when it increases in size (De
trees (FT) of 17 species without evident visual and olfac- Oliveira Oyama and De Souza 2011). With respect to Ficus
tory fruit ripeness cues (Azota caballo: Luehea divaricata; sp., ripe fruit size differs between individual trees (Tu-
Aguaı́: Chrysophyllum gonocarpum; Burro Caá: Casearia jague, personal observation) so it was necessary to analyze
sylvestris; Camboatá: Matayba eleagnoides; Cancharana: each tree as a separate case. We observed and touched ripe
Cabralea canjerana; Carayá bola: Guarea pohlii; Guaza- fruits that had dropped to the ground and took these as a
tumba: Casearia decandra; Guabira: Campomanesia xan- reference point to visually estimate ripe and unripe fruit on
thocarpa; Higuerón: Ficus sp; Laurel: Nectandra sp./ the tree. Monkeys have been seen to visually inspect and
Ocotea sp; Mborevı́ caá: Coussarea contracta; Mora touch Ficus sp. fruits before eating them (Tujague, per-
amarilla: Maclura tinctoria; Miconia sp; Ñandipá: Sorocea sonal observation). We also excluded FT that appeared to
bonplandii; Tala trepador: Celtis iguanaea; Ubajai: Euge- carry another kind of food inside (lianas or host trees that
nia piryformis and Yvaporoitı́: Plinia rivularis—see also carried fruit). In order to exclude the potential use of
Tujague et al. 2016 for a complete list of species con- auditory cues to infer fruit availability, we did not include
sumed). Patches of two vine species were also included in our analysis visits when any other frugivorous species
(Trichostigma octandrum and Dicella nucı́fera) since they was eating in the FT before the group arrived.
were considered to function as fruiting trees. Although Average group spread was estimated from previous
monkeys also approach other fruit trees, visits to these trees studies in this population that described a typical spread of
were excluded from the analysis because of their con- 57 9 42 m for a group of 15 individuals (Janson and Di
spicuous fruit size (e.g., Araticú, Rollinia sp., or citrus Bitetti 1997; Janson 1998). The average group spread of
species), coloration (e.g., Cocú, Allophyllus edulis, or capuchin groups at Iguazú during this study varied
citrus) and/or smell (e.g., Hovenia dulcis). Our selected according to group size and was observed to be similar to
studied trees were not notably odiferous, and human previously recorded estimates (Hirsch et al. 2013). To
observers at the study site were not able to smell fruits from ensure that every visit to each FT was independent, we
our focal trees until they are nearly under the tree. We also established a minimal distance between FT equal to twice
collected phenological data of each FT, as well as struc- the average group longitudinal spread during slow forage.
tural data: DBH = diameter at breast height and crown These minimum distances varied from 100 m (Rita 2010)
volume estimated from the formula of an ellipsoid (Janson to 240 m (Macuco 2009; see Tujague et al. 2016 for
1988, see Table 1 for details). We did not collect either details). We excluded from our sample any potential FT
structural measure for individuals of the two vine species. that was closer than the minimum distance (for each study
Two observers followed the group from the first group and sample year) from any FT already in the sample.
movement in the morning to the final sleeping site at dusk To be sure that the tree was considered an important
collecting data of group’s visits to FT. Ripe and unripe fruit resource for the monkeys, we only considered visits to trees
quantity was scored 0–4 for each FT according to the when at least one-third of the individuals of the capuchin
percentage of branches containing fruit (Janmaat et al. group entered the tree to consume fruit.
2006): 0 (no fruit), 1 (1–25%), 2 (26–50%), 3 (51–75%), 4 At Iguazú, the expected minimal detection distance
(76–100%). We visually divided the crop in four quadrants during slow forage (defined as the distance at which the
following the cardinal points and visually estimated the focal animal, generally at the edge of the group, detects a
average percentage of branches containing fruit. A tree resource while moving at 2 m/min) is 29.5 m (Janson and
with category 4 had all the branches containing fruit. When Di Bitetti 1997). We defined, for each group, a critical
a tree did not have branches in one of the quadrants, then radius (CR) around each tree as half the minimal distance
that quadrant was assigned as category zero. Every time it between FT described above (‘ the minimal distance
was possible to distinguish between ripe and unripe fruit, from the FT to the nearest other FT). We consider this
we counted number of branches separately and we aver- distance to be a conservative estimate of the distance of
aged the percentages across quadrants. Since trees could visual detection since it was higher than 29.5 m in every
change fruiting states and quantities inside the same indi- case (Tujague et al. 2016). This distance of detection
vidual tree as a result of the influence of sun and temper- during slow forage was calculated experimentally by
ature (Houle et al. 2007), this detailed methodology Janson and Di Bitetti (1997) using platforms provisioned
allowed some measurement of this variable. In the case of with tangerines that are easier to detect because of their
fruits that change color between ripe and unripe states, we size and color than natural fruiting trees. Although fruit-
use this feature to estimate ripe and unripe fruit quantities. ing trees are not the same as experimental platforms, our
In the case of Maclura tinctoria fruits, which remain green FT have an average height of 11.3 ± 3.6 meters SD,
at maturity, we considered the fruit as ripe when it had lost while the canopy at Iguazú reaches a height of about
trichomes and we started seeing the exterior creases 25–30 m (Tujague, personal observation). Thus, we

123
 Anim Cogn

Table 1 Variables and their descriptions

Variable Description

Group identity Identification of each studied group which varied in individual composition
Year Study period for each studied group
Tree species Tree categorization according to species selected for the study
Tree individual Identification of each individual tree in relation with its location along the study area
After CR (speed before versus after Categorical variable which took values of 0 to denote speed before and 1 to denote speed after the
CR) critical radius.
Observation Identification of each distinct approach to a focal tree. Treating the observation as a random effect
helps to control for the correlation between speed before and after critical radius in each approach.
Group size Number of individuals for each studied group
Time of day and time of day^2 Hour of the day at which the visit occurred. Both linear and quadratic (^2) terms
Ripe/unripe fruit cover and ripe/ During data analysis, we converted fruit scores to % fruit cover by using the midpoint of the range of
unripe fruit cover^2 percentage of fruit cover corresponding to each score category: category 0 = 0%; category
1 = 12.5%; category 2 = 37.5%, category 3 = 62.5% and category 4 = 87.5%. Both linear and
quadratic (^2) terms
Tree DBH Diameter at breast height
Crown volume Volume of the tree crown estimated from the formula of an ellipsoid: VOL = R2Dp4/3, where R is
the crown radius and D is the crown depth, form lowest to highest fruit-bearing branch
PC1 and PC1^2 First principal component as a result of the principal component analysis of ln(DBH) and
ln(CrownVolume) measures across the independent observations. Both linear and quadratic (^2)
terms
Group-minutes and Group-minutes^2 Total duration from the start to the end of the visit, subtracted any periods when there where no
monkeys in the patch, then multiplied the resulting time by the total number of individuals that
visited the patch, and divided by the total group size. Both linear and quadratic (^2) terms
Random effects are underlined; fixed effects in bold

believe it was not easy for a monkey feeding at the top of Speed before critical radius (BCR): group speed (m/
the tree to see the next focal tree (considering that the min) before entering a critical radius, calculated as
minimum distance between FT was at least 5 times the the distance in meters between the two UTM coor-
experimentally derived visual detection distance). We dinates just before entering the CR (latest UTM
used this CR to measure changes in speed of approach to coordinates position collected before approaching the
FT before and after entering the CR. CR) divided by the time elapsed between the two
To collect data on a group’s visits to FT (Vogel and measures.
Janson 2007), one observer was always in the front of the Speed after critical radius (ACR): group speed (m/
group and went to the FT that monkeys were expected to min) immediately after entering a critical radius
visit sensu Janmaat et al. (2006). Before the group arrived, calculated as the distance in meters between the last
the observer collected data on the presence/absence of UTM coordinate position collected before approach-
frugivores at the FT and scored the amount of ripe/unripe ing the CR and the first UTM coordinate position
fruit according to the categories described above. after entering the CR, divided by the time elapsed
While following the group, we recorded the UTM between the two measures.
coordinates for the spatial location of the group center
We defined a visit as an entry to a FT and/or tree patch
(Janson 1990) using a handheld Garmin 60Cx GPS (B10 m
in which at least one-third of the group had eaten at least
error) at 10-min intervals. When a GPS point could not be
once at the site. It did not include visits of isolated indi-
obtained directly from the center of the group, the coor-
viduals to isolated trees or patches. We defined a patch as a
dinates of the group center were projected using the
set of more than one tree of the same specie separated from
direction and distance from the observer. When the group
the other by no more than 10 m, a distance chosen to make
approached the CR for a given focal tree, UTM coordinates
it possible to see the entire group at the same time; we
were taken at least once prior to entering the CR, and once
considered a visit to a patch as a visit to one FT. For each
shortly after entering the CR, regardless of the time interval
visit, we defined a measure, Group-minutes, of time that
(Fig. 2). These data were used to estimate group speed
the group invested in a food patch (see Table 1 for details
during travel to FT (Noser and Byrne 2007) as follows:

123
Anim Cogn

Fig. 2 Speed (m/min) before

(BCR) and after (ACR) critical
radius during visits to target
trees. Continuous
circumference: critical radius;
arrow: direction of movement;
black square: approximate last
UTM position before
approaching a CR; black circle:
approximate UTM position
while approaching a CR; black
triangle: approximate UTM
position after crossing a CR;
black square and black circle:
involved in the estimation of
speed BCR; black circle and
black triangle: involved in the
estimation of speed ACR

of the defined measure). We treated successive visits to a To increase the robustness of the analysis, we took steps
single FT as independent visits only if the monkeys left the to reduce notably high pairwise correlations between the
FT and traveled a distance greater than the CR before fixed effect variables. Because DBH and crown volume are
returning. both measures of tree size and were moderately highly
correlated (r = 0.53), we carried out a principal compo-
Data analysis nent analysis of ln(DBH) and ln(CrownVolume) measures
across the independent observations. As a final measure of
An independent visit to a FT had a unique date and time of FT size for a given observation, we used the observation’s
the visit, along with a known tree species and individual, score on the first principal component. The first principal
and fruit cover values taken before the start of the visit. For component (PC1) explained 77% of the combined variation
each independent visit of a group to a FT, we treated the in the original variables. In addition, to avoid strong cor-
two estimates of travel speed (BCR and ACR) as distinct relations between the linear and quadratic terms for any
measures of travel speed, nested within a single ‘observa- fixed effect variable in the analysis, we centered each linear
tion,’ which was treated as a random effect. The ‘obser- variable so that the mean of the variable was zero; the
vation’ random effect was nested within a second random quadratic term was the square of the centered value of the
effect, the year of the sample. We also included tree spe- linear variable. Before centering, the pairwise correlations
cies as a random effect, because some species might be between the linear and quadratic terms of fixed effect
more highly preferred (and thus might encourage faster variables ranged from 0.77 for ln(group-minutes) to 0.99
travel). Initially, we also included tree individual (nested for time of day; after centering, they ranged from -0.12 for
within tree species) and group identity as random effects, ln(group-minutes) to 0.77 for ripe fruit cover. The com-
but these consistently had zero or near-zero variance plete model to analyze speed of approach (ESM1) had
components in the model and thus were excluded from three random effects (tree species, year and observation
further analyses. As independent fixed effect variables, we nested within year) and 12 fixed effects (see Table 1 for
used (1) a categorical variable for whether the speed was details).
measured ‘before’ versus ‘after’ entering the CR, (2) group We analyzed the complete model as a generalized linear
size, and (3) linear and quadratic terms for all the variables mixed model (GLMM) in R (R Development Core Team
describing potential feeding benefits: ripe fruit cover, 2011) using the function lmer in lme4 package version 1.1-
unripe fruit cover, group-minutes, DBH and crown volume 8 (Bates et al. 2015). For all inferential tests comparing any
(see Table 1 for details). To reduce notable skewness in two models, we refitted the models using ML (maximum
measures of group size, group-minutes, DBH and crown likelihood procedure) instead of REML (restricted maxi-
volume, we took the natural logarithm of the raw measures; mum likelihood procedure). Of particular interest for this
doing so reduced the effect on the analysis of large but rare study was the effect of phenological measures of fruit
values of these fixed effect variables. availability—if the best statistical model included

123
 Anim Cogn

phenological measures of fruit cover, such a result would Table 3, Fig. 3). Travel speed increased significantly
be consistent with the possibility that the monkeys can toward FT with greater ripe fruit cover, reaching peak
anticipate the tree’s phenological condition despite its speeds at intermediate cover values (*26–50%) and
variation across time. Because of strong directional declining thereafter (ripe fruit cover and ripe fruit cover^2,
expectations of the predicted trends for each variable, we Table 3, Fig. 4). The quadratic effect of ripe fruit cover on
used one-tailed probability values for all tests of fixed travel speed was significant (Table 3) in the expected
effect hypotheses. negative direction based on theory and prior evidence from
capuchin monkeys that scramble competition should be
relaxed at high food availabilities (Janson 1988). Travel
Results speeds were highest in the morning, declining significantly
toward midday before increasing again toward evening
The total number of visits to selected trees varied between (time of day and time of day^2 Table 3; Fig. 4). Travel
groups and years (Macuco: N 2008 = 52, N 2009 = 83; speed did not appear to depend on any measure of unripe
Gundolf: N 2009 = 48; N 2010 = 86; Rita: N 2010 = 59, fruit cover, nor on measures of either tree size (score on
N 2011 = 78). We excluded visits to trees that had other PC1 and PC1^2) or actual feeding success (group-minutes
frugivores eating before monkeys arrived (birds and/or and group-minutes^2; Tables 2, 3). The variances of the
coatis), to avoid the possibility that the group was attracted random effects tree species and year were generally small
by the sound of other animals feeding at the site (Macuco (absolute value and percentage of total variation:
2008 N = 2/2009 N = 6, Gundolf 2010 N = 2, Rita TreeSp = 0.01070, 2.52%; year = 0.04217, 9.93%).
2011 = 9). However, variation associated with observation (nested
There was considerable variation in fruit availability within year) was substantial (0.07515, = 17.7% of total
across time in individual tree crowns. For trees visited at variation), suggesting that variation in travel speeds among
least twice, the mean ripe fruit score was 1.312, while the observations depended on unknown factors in addition to
within-tree SD was 0.593, yielding a coefficient of varia- the tested variables. Finally, since speed BCR and ACR
tion (CV) of 45.2%. The corresponding value for unripe varied according to percentages of ripe and unripe fruit,
fruit scores was 1.273 and 0.905, for a CV of 71.1%. and all the possible combinations of those categories in
The GLMM model with all the fixed and random effects each tree, we refer to descriptive details of speed with
variables (Table 2) was significantly different from the mean and standard deviations to be found in Tujague
model with only random effects (all main effects, Table 3). (2013) and Tujague et al. (2016).
Travel speeds before and after passing the critical radius of We accounted for possible temporal variation or clus-
a FT were significantly different, with ACR being on tering of travel speeds that did not depend on the variables
average 52% greater than BCR (speed after vs. before CR, described in our hypotheses. First, we treated the year of

Table 2 Results of the full

Dependent variables Parameter estimate ±SE T
GLMM analysis
Intercept 1.249840 0.616784 2.026
After CR (speed after versus before CR) 0.459562 0.063126 7.280
Group size 0.010422 0.191508 0.054
Time of day 0.790644 0.269031 -2.939
Time of day^2 9.191018 1.974020 4.656
Ripe fruit cover 0.845644 0.406826 2.079
Ripe fruit cover^2 -2.023339 1.133070 -1.786
Unripe fruit cover -0.081789 0.225148 -0.363
Unripe fruit cover^2 0.445422 0.669139 0.666
PC1 0.012523 0.045700 0.274
PCI^2 0.006681 0.015190 0.440
Group-minutes 0.031464 0.035948 0.875
Group-minutes^2 -0.008876 0.013058 -0.680
The values are the estimated parameter, standard error (±SE) and the t-values (T) of the variable in the
complete model. ^2 = quadratic term of the variables. In bold are the fixed effects that were significant
during the hypothesis tests. Random effects are included in the statistical model but not reported here, as the
R program (lmer) does not report inferential statistical tests for random effects variables

123
Anim Cogn

Table 3 Summary of individual hypothesis test results

Fixed effects included N Change in -2*log-likelihood df Direction of effect P(X df), one-tailed

All main effects 221 79.21 12 NA 5.84E 2 12

After CR (speed after versus before CR) 221 41.545 1 Positive 0.57E 2 10
Group size 221 0.03 1 Positive 0.431
Time of day 221 9.2036 1 Negative 1.21E 2 03
Time of day^2 221 21.514 1 Positive 1.75E 2 06
Ripe fruit cover 221 4.7718 1 Positive 0.015
Ripe fruit cover^2 221 3.5923 1 Negative 0.029
Unripe fruit cover 221 0.2314 1 Negative 0.185
Unripe fruit cover^2 221 0.5885 1 Positive 0.279
PC1 221 0.1494 1 Positive 0.35
PCI^2 221 0.2787 1 Positive 0.20
Group-minutes 221 0.8282 1 Positive 0.1814
Group-minutes^2 221 0.484 1 Negative 0.2433
The ‘all main effects’ test compares the complete model (Table 2) to the model with only random effects included. In each other case, we
compared via ANOVA the complete model with all main effects to a reduced model that omits the fixed effect of interest. N number of
observations, -2*LL Chi-square values; df degrees of freedom; P(X df) probability values; ^2 quadratic term of the variables. In boldface are the
effects that were statistically significant

the sample as a random effect, to account for possible in cases of non-convergence of the full model, we tried a
large-scale temporal patterns (e.g., variation in climate or slightly simplified version of the random slopes model by
annual plant productivity) not captured by our hypotheses. disallowing a correlation between the intercept and slope.
Second, we tested for possible effects of sequential sam- However, this model also failed to converge after
pling on our results. We added the travel speed of the 5,000,000 iterations in lme4. Thus, no random slopes
previous observation as a fixed effect variable to the model preferred by Barr et al. (2013) was feasible with our
complete model given in Table 2. The contribution of prior data set.
travel speed to the expanded model was small and not
statistically significant (t = 0.23, P = 0.82), suggesting
that each observation of travel speed to a FT was condi- Discussion
tionally independent of previous observations.
We examined possible interactions between some of the Results from the present study provide evidence that wild
variables, to test whether the analytic model we used was capuchin monkeys changed their travel speed during
sufficient to describe the pattern of variation in speed of approaches to trees with different rewards without any
approach to a FT. First, it is possible that the effects of obvious perceptual knowledge of tree condition, given that
feeding benefit (fruit cover, tree or crop size) on travel all major sensory cues (visual, olfactory and auditory
speed would be more notable at large group sizes, in which information) were controlled or minimized.
scramble competition would be expected to be strongest. We predicted that travel speed would vary according to
However, we found no significant interactions of group the expected reward at a given feeding tree in a nonlinear
size with fruit patch variables either individually (ripe fruit way, increasing as rewards increased from low values, but
cover: t = -0.08, P = 0.94; unripe fruit cover: t = 0.09, plateauing or even possibly decreasing again at very high
P = 0.92; tree size (PC1): t = 1.04, P = 0.30; log of values. We also expected that travel speed would vary with
group-minutes: t = 0.24; P = 0.81) or collectively (com- time of day, decreasing from the time of first activity as the
plete model vs. extended model including all interactions: monkeys became less hungry, and possibly increasing
Chi-square = 1.37, df = 4, P = 0.85). Third, we attemp- again toward late afternoon. Both patterns were confirmed
ted to include all fixed effect by random effect interactions in our analyses (Table 3; Fig. 4), but the effect of time of
(‘random slopes’ model), according to the recommenda- day was the strongest. The predicted hump-shaped rela-
tions of Barr et al. (2013). Unfortunately, the full random tionship between travel speed and ripe fruit cover was
slope models did not converge in lme4 even after statistically significant. The studied groups increased travel
5,000,000 iterations. As recommended by Barr et al. (2013) speed toward trees with more ripe fruit only up to a score

123
 Anim Cogn

64.0

ln(Speed After Passing Critical Radius)

4
32.0
Travel speed after CR (m/min)

16.0
3
8.0

4.0
2
2.0

1.0
1
0.5
0.5 1.0 2.0 4.0 8.0 16.0 32.0
0 >0-0.25 0.25-0.5 0.5-0.75 0.75-1.0
Travel speed before CR (m/min)
Ripe Fruit Cover (Fraction of Tree Crown)
Fig. 3 Travel speed (meters per minute) before and after critical
radius (CR) is positively related (orthogonal regression with equal
variances for x- and y-axes, r = 0.44, ln(ACR) = 0.42 ? 5

ln(Speed After Passing Critical Radius)

0.994*ln(BCR), P \ 0.0001)

corresponding to 26–50% fruit cover, above which travel 4

speed declined. At the highest fruit cover values, monkeys
appear to have less urgency to be among the first to arrive
3
at the food source. This is consistent with the idea that
faster travel toward food patches is the result of scramble
competition to reach the resource (Fig. 1b), rather than a 2
simple response to greater food-related cues or larger
potential rewards (Fig. 1a). Measures of unripe fruit cover
did not explain variation in travel speed at any distance 1
from a focal tree. The fact that travel speed increased
notably at distances from a feeding tree that were at least 5
times the known detection distance of capuchin monkeys 0
6:00 9:00 12:00 15:00 18:00 21:00
for experimentally provided food sources (Janson and Di
Time of day
Bitetti 1997) suggests strongly that they were responding to
the memory of the resource characteristics from prior visits Fig. 4 Group movement speed (m/min) after passing the critical
rather than to immediate food-related cues. Although the radius toward a focal tree, plotted as a function of ripe fruit cover and
locations of UTM coordinates before and after entering the time. For ripe fruit cover, the travel data are shown for each value of
phenological cover score as box plots (median line, upper and lower
CR were highly variable, we, nonetheless, found a con-
box bounds of 25 and 75% of the data, whiskers at 10 and 90% of the
sistent pattern of speed related to ripe fruit cover, sup- data, plus individual points for all data outside the 90% limits).
porting the validity of our results. The detected changes in Regression lines are based on bivariate results only, not on the
travel speed suggest that the monkeys used memory of fruit complete model. Program used: R
quantity and ripening state. In at least some cases, this
memory was long term, lasting across several days since a did not change in relation to hunger levels (as assessed
prior visit to a given focal tree. by time of day). Yet, they did not test for a curvilinear
Our findings about speed in relation to time of the day effect of time, but instead compared only morning
support the results from Janson (2007) who found that (0700–1000 h) against the rest of the day. If we restrict
capuchin groups moved faster when they were hungry or our analysis of travel speeds in the same way as Janmaat
close to a goal, although his study was carried out using et al. (2006), we also find no effect of time of day for
experimental platforms and during winter when resources BCR (t test, P = 0.33) but do find a significant differ-
are scarce and platforms were the only concentrated ence for ACR (t test, P \ 0.005). Thus, it is unclear
resource available. However, Janmaat and co-workers whether the pattern we describe, in which travel speed
(2006) found that approach speed to trees by mangabeys increases when animals are likely to be hungry (Fig. 4),

123
Anim Cogn

is confined to this study site and species, or is more the animal’s behavior to estimate the value of food such as
general. fruit rejection ratios may prove beneficial for such studies
Other metrics of tree size or fruit amount eaten (DBH (Ban et al. 2014).
and crown volume; total group feeding minutes) had no Taking into account both quantity and fruiting state of
detectable effects on monkey’s travel speed toward trees, resources allow researchers to analyze primate’s spa-
after including phenological measures of ripe fruit avail- tiotemporal behavioral variation. How much they will eat
ability and a random effect for tree species. Other studies on a given visit will depend on their own past behavior
found that measures of tree size did predict the selection of (Janson 2000). Tujague and Janson (in press) suggested
trees by spider monkeys (Suarez 2014) as well as by that capuchins shortened their revisit intervals when a tree
capuchins in the same population using experimental had more unripe fruit relative to mature fruit, even though
feeding sites (Janson 1998, 2007). Our present results the monkeys did not consume the unripe fruit: the time
suggest that monkeys pay more attention to anticipated ripe until the next revisit for a given tree decreased indepen-
fruit cover than to whether a tree is large or small. How- dently with increasing amounts of both ripe and unripe
ever, all the focal trees in our analysis were ones at which fruits in the crown at the end of the visit. The presence of
at least one-third of the group members fed, thus biasing only ripe fruit (with no unripe fruit) promoted shorter
the sample toward more productive patches overall. Given return intervals than for trees bearing only unripe fruit,
other evidence from the same site that capuchins are able to regardless of fruit quantity (3.85 vs. 6.5 days, respec-
track elapsed time since a prior visit to each of multiple tively). Revisit intervals to trees that had no fruit (empty
feeding sites (Janson 2016), their ability to anticipate ripe trees, category 0) were the longest (11.8 days; Tujague and
fruit cover may depend on a similar cognitive capacity. Janson, unpublished data). The higher travel speeds that
The increase in travel speed toward trees with higher Tujague and co-workers (2016) found when an entry to a
ripe fruit cover (until it reaches the plateau) is consistent critical radius ended in a visit to trees that carried fruit
with the effect of indirect or scramble competition during support these conclusions. Capuchins increase speed while
the visits to fruiting trees (Fig. 1b). Although it is expected they were closer to a focal tree, indicating that they had an
that scramble competition should be greater in larger expectation about what they were going to find (presum-
groups (Janson and Goldsmith 1995), we did not find an ably based on their previous visit or visits). Monkeys
effect of group size on the approach speed toward fruit accelerated progressively while the resource was out of
trees. Perhaps variation in group size between studied sight to finally reach speeds close to their maximum
groups was simply not large enough to be detected by the velocity after entering the critical radius.
analysis. Finally, if increased direct competition over food Non-human and human primates show a very similar
were an additional factor favoring higher travel speeds, we pattern of discrimination in odor performance (Laska and
would have expected to find that speeds were higher when Freyer 1997) and the same range of olfactory performance
tree size was smaller, all else being equal. However, the with regard to sensitivity to aliphatic esters related to the
PC1 variable (which combines DBH and crown volume) degree of ripeness of fruit (Maarse 1991). Also, it has been
had no significant effect on travel speed. proved that functional olfactory genes or neuroanatomical
A cautionary note in interpreting our patterns is our olfactory features are poor indicators to predict olfactory
inability to estimate the true value of feeding trees (Ban performance since non-human primates species show dif-
et al. 2014). Many fruit species in our study do not change ferent functional genes and neuroanatomical features but
color during maturation and thus require elasticity mea- similar performances (Laska et al. 2004). Studies have
surements to assess ripeness/sweetness (Janmaat et al. concluded that olfactory cues are heavily used by nocturnal
2006); we were not able to take such measurements on primates and visual cues are relied on by diurnal primates
actual fruits in our rainforest trees that were up to 40 m (Bicca-Marques and Garber, 2004), including capuchin
high and contained more than a thousand fruits. Nutritional monkeys (Bolen and Green 1997). Although we cannot
value of fruits can vary tremendously within a tree but also exclude the possibility that the noses of capuchin monkeys
between trees of the same species (Houle et al. 2007). In might be more trained than those of humans or that odors
some cases, the variation in the nutritional values among can be transported over long distances by the wind, it
tree sites can be greater than the differences among tree seems unlikely that our studied groups relied on olfactory
species (Chapman et al. 2003). Also, monkeys are able to signals during approaches to trees. Given that the monkeys
regulate their nutritional state through selecting comple- were very unlikely to see the trees at the critical distance
mentary combinations of foods to achieve a nutritionally (Janson and Di Bitetti 1997), we conclude that speed is a
balanced diet (Felton et al. 2009). Thus, fruit cover may not good measure about expectation of resources, as was
always be an accurate guide to the nutritional value of a shown in previous studies for the same specie (Janson and
fruit crop to its consumer. Incorporating measures that use Di Bitetti, 1997; Janson 1998) and others (Papio

123
 Anim Cogn

hamadryas: Sigg and Stolba 1981; Papio hamadryas food availability or responses to food profitability? Folia
cynocephalus: Pochron 2001; Cercocebus atys atys and Primatol 54:177–186
Chapman CA, Chapman LJ, Rode KD, Hauck EM, McDowell LR
Lophocebus albigena johnstoni: Janmaat et al. 2006; Papio (2003) Variation in the nutritional value of primate foods: among
urcinus: Noser and Byrne 2007). trees, time periods, and areas. Am J Primatol 24:317–333
Cunningham E, Janson CH (2007) Integrating information about
Acknowledgements We wish to thank Administración de Parques location and value of resources in white-faced saki monkeys
Nacionales and Centro de Investigaciones Ecológicas Subtropicales (Phitecia phitecia). Anim Cognit 10:293–304
for granting permission to conduct this study. This work would not De Oliveira Oyama S, de Souza LA (2011) Morphology and anatomy
have been possible without field assistance from Alex Alcocer of the developing fruit of Maclura tinctoria, Moraceae. Rev Bras
Aldunate, Ana Alvarez, Marı́a Celia Baldovino, Marı́a Verónica Bot 34(2):187–195
Cifre, Lara Coleselli, Emanuel Galetto, Juan Garibaldi, Luis Molina, Di Bitetti MS (1997) Evidence for an important social role of
Natalia Gomez de Oliveira, Elizabeth Gonzáles Valentı́n, Rocı́o allogrooming in a platyrrhine primate. Anim Behav
Fernández Quintero, Daniel Lambruschi, Ricardo Melzew, Juan Pablo 54(1):199–211
Peretti, Noelia Rivas, Taylor Rubin, Clara Scarry and Fermino Silva Di Bitetti MS (2001) Home-range use by the tufted capuchin monkey
(Cebus apella nigritus) in a subtropical rainforest of Argentina.
Funding This study was funded by a dissertation fellowship from the J Zool 253:33–45
Argentine Research Council (CONICET) to M.P.T., as well as grants Di Bitetti MS, Janson CH (2001) Social foraging and the finder’s
from Idea Wild and the American Society of Mammalogists (no grant share in capuchin monkeys, Cebus apella. Anim Behav
numbers available). 62:47–56
Felton AM, Felton A, Lindenmayer DB, Foley WJ (2009) Nutritional
Compliance with ethical standards ecology: nutritional goals of wild primates. Funct Ecol 23:70–78
Gallistel CR, Gibbon J (2000) Time, rate and conditioning. Psychol
Conflict of interest The authors disclose no conflict of interest. Rev 107(2):289–344
Garber PA (1989) Role of spatial memory in primate foraging
Ethical approval All applicable international, national and/or insti- patterns: Saguinus mystax and Saguinus fuscicollis. Am J
tutional guidelines for the care and use of animals were followed. All Primatol 19:203–216
research reported in this manuscript adhered to the International Garber PA, Dolins FL (1996) Testing learning paradigms in the field:
Primatological Society and the American Society of Primatologists evidence for use of spatial and perceptual information and rule-
Principles for the Ethical treatment of Non-human Primates, and based foraging in wild moustached tamarins. In: Norconk MA,
complied with the protocols approved by the Animal Care and Use Rosenberger AL, Garber PA (eds) Adaptive radiations of
Committee and the Argentine Research Council (CONICET). It also Neotropical primates. Plenum Press, New York, pp 201–216
adhered to the legal requirements of Argentina. Garber P, Pacuilli LM (1997) Experimental field study of spatial
memory and learning in wild capuchin monkeys (Cebus
capuchinus). Folia Primatol 68:236–253
Gendron RP, Staddon JER (1983) Searching for cryptic prey: the
effect of search rate. Am Nat 121:172–186
References Gibeault S, MacDonald SE (2000) Spatial memory and foraging
competition in captive western lowland gorillas (Gorilla gorilla
Anzelc A (2009) The foraging and travel patterns of white-faced sakis gorilla). Primates 41(2):147–160
in Brownsberg Nature Park, Suriname: preliminary evidence for Hirsch BT, Tujague MP, Di Blanco YE, Di Bitetti MS, Janson CH
goal-directed foraging behavior. Master of Arts Thesis, Kent (2013) Comparing capuchins and coatis: causes and conse-
State University, pp 194. Unpublished quences of differing movement ecology in two sympatric
Ban SD, Boesch C, Janmaat KRL (2014) Taı̈ chimpanzees anticipate mammals. Anim Behav 86:331–338
revisiting high-valued fruit trees from further distances. Anim Houle A, Chapman CA, Vickery WL (2007) Intratree variation in
Cognit 17(6):1353–1364 fruit production and implications for primate foraging. Int J
Barr DJ, Levy R, Scheepers C, Tily HJ (2013) Random effects Primatol 28:1197–1217
structure for confirmatory hypothesis testing: keep it maximal. Janmaat KRL, Byrne RW, Zuberbühler K (2006) Evidence for a
J Mem Lang 68:255–278 spatial memory of fruiting states of rainforest trees in wild
Bates D, Maechler M, Bolker B, Walker S, Christensen RH B, mangabeys. Anim Behav 72:797–807
Singmann H, Dai B, Grothendieck G (2015) lme4: linear mixed- Janmaat KRL, Chapman CA, Meijer R, Zuberbühler K (2012) The
effects models using ‘Eigen’ and S4. https://cran.r-project.org/ use of fruiting synchrony by foraging mangabey monkeys: a
web/packages/lme4/index.html ‘simple tool’ to find fruit. Anim Cognit 15:83–96
Bicca-Marques JC, Garber PA (2004) Use of spatial, visual, and Janmaat KRL, Polanskya L, Ban SD, Boesch C (2014) Wild
olfactory information during foraging in wild nocturnal and chimpanzees plan their breakfast time, type, and location. Proc
diurnal anthropoids: a field experiment comparing Aotus, Natl Acad Sci USA 111(46):16343–16348
Callicebus, and Saguinus. Am J Primatol 62:171–187 Janson CH (1985) Aggressive competition and individual food
Bolen RM, Green SM (1997) Use of olfactory cues in foraging by owl consumption in wild brown capuchin monkeys (Cebus apella).
monkeys (Aotus nancymai) and capuchin monkeys (Cebus Behav Ecol Sociobiol 18:125–138
apella). J Comp Psychol 111(2):152–158 Janson CH (1988) Food competition in brown capuchin monkeys
Brown AD, Zunino GE (1990) Dietary variability in Cebus apella in (Cebus apella): quantitative effects of group size and tree
extreme habitats: evidence of adaptability. Folia Primatol productivity. Behaviour 105:53–76
54:187–195 Janson CH (1990) Social correlates of individual spatial choice in
Chapman CA, Fedigan LM (1990) Dietary differences between foraging groups of brown capuchin monkeys, Cebus apella.
neighboring groups Cebus capucinus groups: local traditions, Anim Behav 40:910–921

123
Anim Cogn

Janson CH (1998) Experimental evidence for spatial memory in PA (eds) On the move: How and why animals travel in groups.
foraging wild capuchin monkeys, Cebus apella. Anim Behav University of Chicago Press, Chicago, pp 375–417
55:1229–1243 Noser R, Byrne RW (2007) Travel routes and planning of visits to
Janson CH (2000) Spatial movement strategies: theory, evidences and out-of-sight resources in wild chacma baboons, Papio urcinus.
challenges. In: Boinski S, Garber PA (eds) On the move: how Anim Behav 73:257–266
and why animals travel in groups. University of Chicago Press, Noser R, Byrne RW (2010) How do wild baboons (Papio ursinus)
Chicago, pp 165–203 plan the routes? Travel among multiple high-quality food
Janson CH (2007) Experimental evidence for route integration and sources with inter-group competition. Anim Cognit 13:145–155
strategic planning in wild capuchin monkeys. Anim Cognit Placci LG, Arditi SI, Ciotek LE (1994) Productividad de hojas, flores
10(3):341–356 y frutos en el Parque Nacional Iguazú. Yvyrareta 5:49–56
Janson CH (2016) Capuchins, space, time and memory: an experi- Platt ML, Brannon EM, Briese TL, French JA (1996) Differences in
mental test of what-where-when memory in wild monkeys. Proc feeding ecology predict differences in performance between
R Soc B. doi:10.1098/rspb.2016.1432 golden lion tamarins (Leontopithecus rosalia) and Wied’s
Janson CH, Byrne R (2007) What wild primates know about marmosets (Callithrix kuhli) on spatial and visual memory tasks.
resources: opening up the black box. Anim Cognit 10:357–367 Anim Learn Behav 24(4):384–393
Janson CH, Di Bitetti MS (1997) Experimental analysis of food Pochron ST (2001) Can concurrent speed and directness of travel
detection in capuchin monkeys: effects of distance, travel speed indicate purposeful encounter in the yellow baboons (Papio
and resource size. Behav Ecol Sociobiol 41:17–24 hamadryas cynocephalus) of Ruaha National Park, Tanzania? Int
Janson CH, Goldsmith ML (1995) Predicting group size in primates: J Primatol 22(5):773–785
foraging costs and predation risks. Behav Ecol 6(3):326–336 R Development Core Team (2011) R: a language and environment for
Janson CH, van Schaik CP (1988) Recognizing the many faces of statistical computing. R Foundation for Statistical Computing,
primate food competition: methods. Behaviour 105:165–186 Vienna
Janson CH, Vogel ER (2006) Hunger and aggression in capuchin Roberts WA, Mitchell S, Phelps MT (1993) Foraging in laboratory
monkeys. In: Hohmann G, Boesch C, Nebel C (eds) Feeding trees: spatial memory in squirrel monkeys. In: Zendall TR (ed)
ecology in apes and other primates. Cambridge University Press, Animal cognition: A tribute to Donald A. Riley. Lawrence
Cambridge, pp 285–312 Erlbaum, Hillsdale, pp 131–151
Janson CH, Baldovino MC, Di Bitetti MS (2012) The group life cycle Scheumann M, Call J (2006) Sumatran orangutans and a yellow-
and demography of brown capuchin monkeys (Cebus [apella] cheeked crested gibbon know what is where. Int J Primatol
nigritus) in Iguazú National Park, Argentina. In: Kappeler P, 27(2):575–602
Watts D (eds) Long-term field studies of primates. Springer- Sigg J, Stolba J (1981) Home range and daily March in a hamadryas
Verlag, Heidelberg, pp 185–214 baboon troop. Folia Primatol 36:40–75
Laska M, Freyer D (1997) Olfactory discrimination ability for Stephens DW, Krebs JR (1983) Foraging theory. Princeton University
aliphatic esters in Squirrel monkeys and humans. Chem Senses Press, Princeton
22:457–465 Suarez SA (2014) Ecological factors predictive of wild Spider
Laska M, Wieser A, Rivas Bautista RM, Hernandez Salazar LT monkey (Ateles belzebuth) foraging decisions in Yasunı́,
(2004) Olfactory sensitivity for carboxylic acids in Spider Ecuador. Am J Primatol 76:1185–1195
monkeys and Pigtail macaques. Chem Senses 29:101–109 Tujague MP (2013) Comportamiento espacial y memoria en el mono
Maarse H (1991) Volatile Compounds in foods and beverages. Marcel caı́ (Cebus apella) en contexto de grupo: semejanzas entre
Dekker, New York primates no humanos y humanos. Ph.D. dissertation, University
MacDonald SE (1994) Gorillas’ (Gorilla gorilla gorilla) spatial of La Plata, Argentina, p 201. Unpublished
memory in a foraging task. J Comp Psychol 108:107–113 Tujague MP, Janson CH (in press). Foraging cognition in Neotropical
MacDonald SE, Agnes MM (1999) Orangutan (Pongo pygmaeus primates: integrating studies from wild and captivity. In
abelii) spatial memory and behavior in a foraging task. J Comp Kowalewski M, Oklander L (eds) Primatology in Argentina I.
Psychol 113:213–217 Editorial SAREM, Buenos Aires
MacDonald SE, Wilkie DM (1990) Yellow-nosed monkey’s (Cerco- Tujague MP, Bacigalupe MA, Lahitte HB, Janson CH (2016)
pithecus ascanius whitesidei) spatial memory in a simulated Memoria espacial en monos capuchinos de Argentina: un
foraging environment. J Comp Psychol 104:382–387 estudio observacional en vida silvestre. Revista Argentina de
MacDonald SE, Pang JC, Gibeault S (1994) Marmoset (Callithrix Antropologı́a Biológica 18(1):1–13
jacchus jacchus) spatial memory in a foraging task: win-stay van Schaik CP, van Noordwijk MA (1989) The special role of male
versus win-shift strategies. J Comp Psychol 108:328–334 Cebus monkeys in predation avoidance and its effect on group
Menzel EW (1973) Chimpanzee spatial memory organization. composition. Behav Ecol Sociobiol 24:265–276
Science 182(4115):943–945 Vogel ER (2005) Rank differences in energy intake rates in white-
Menzel CR (1991) Cognitive aspects of foraging in Japanese faced capuchin monkeys, Cebus capucinus: the effects of contest
monkeys. Anim Behav 41:397–402 competition. Behav Ecol Sociobiol 58:333–344
Milton K (1981) Distribution patterns of tropical plant foods as an Vogel ER, Janson CH (2007) Predicting the frequency of food-related
evolutionary stimulus to primate mental development. Am agonism in white-faced capuchin monkeys (Cebus capucinus),
Anthropol New Ser 83:534–548 using a novel focal-tree method. Am J Primatol 69:1–18
Milton K (2000) Quo vadis? Tactics of food search and group
movement in primates and other animals. In: Boinski S, Garber

123