Consequences of Joint-Territoriality in a Mixed-Species Group of Tamarin Monkeys

8/8/2019 Consequences of Joint-Territoriality in a Mixed-Species Group of Tamarin Monkeys

http://slidepdf.com/reader/full/consequences-of-joint-territoriality-in-a-mixed-species-group-of-tamarin-monkeys 1/28

Consequences of Joint-Territoriality in a Mixed-Species Group of Tamarin Monkeys

Author(s): Carlos A. PeresSource: Behaviour, Vol. 123, No. 3/4 (Dec., 1992), pp. 220-246Published by: BRILLStable URL: http://www.jstor.org/stable/4535071

Accessed: 20/11/2010 14:43

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at

http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless

you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you

may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained athttp://www.jstor.org/action/showPublisher?publisherCode=bap.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed

page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

BRILL is collaborating with JSTOR to digitize, preserve and extend access to Behaviour.

http://www.jstor.org

http://www.jstor.org/action/showPublisher?publisherCode=bap

http://www.jstor.org/stable/4535071?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp



http://www.jstor.org/page/info/about/policies/terms.jsp

http://www.jstor.org/stable/4535071?origin=JSTOR-pdf




Behaviour123 (3-4) 1992, E. J. Brill, Leiden

CONSEQUENCES OF JOINT-TERRITORIALITY IN AMIXED-SPECIES GROUP OF TAMARIN MONKEYS

by

CARLOS A. PERES1,2)

(Museu Goeldi, Departamento de Zoologia, Cx. Postal 399, Belem, Para 66.040, Brazil)

(With 5 Figures)

(Acc. 16-XII- 1992)

Summary

Groups of saddle-back (Saguinus uscicollis) as moustached tamarins (Saguinusmystax)in a

western Amazonian forest jointly defended home ranges larger than 100 ha, which were

held in common throughout the year. Resources were defended by direct exploitation in

extensive areas shared with other groups, or through intensive and frequent intergroupinteractions along territorial boundaries. These interactions were expressed primarily

during intergroup encounters, and affected the use of space, movements, time budget, and

foraging success of tamarins. During encounters, animals of both species spent more time inenergetically costly activities, such as rapid travel and intergroup chases, and less time in

energetically positive activities, such as feeding and foraging. In addition, foraging success

per unit of foraging effort within overlapping areas of the range periphery was lower than in

exclusive areas of the range centre, particularly for saddle-back tamarins. The time and

energy allocated by moustached tamarins to boundary contests was considerably greaterthan that of saddle-backs, despite the fact that only the latter species increased its foraging

efficiency by shifting from exclusive areas in the group's range to those shared by other

groups. This is probably because of saddle-back's greater use of depletable food supplies,such as small fruit patches and small microhabitats containing embedded prey items. These

benefits arelikely

to

justifythe substantial amount of time and

energyinvested in territorial

defence for both saddle-backs and moustached tamarins, but appeared to be highly asym-metric between species.

Introduction

Interspecific territoriality can occur between individuals of different spe-cies (heterospecifics) defending space against one another (e.g. ORIANS &

l) Present address: Center for Tropical Conservation, Duke University, Simons Building,3705-C Erwin Road, Durham, NC 27705-5015, U.S.A.

2) This study was supported by World Wildlife Fund-US grant 6199. I wish to thank the

Brazilian oil company (Petrobras) for logistical support in the upper Urucu. I am grateful to

Andrew JOHNS,Yarrow ROBERTSON,Robin DUNBAR,David CHIVERS, and one anonymousreviewer for providing helpful comments on earlier drafts of the manuscript. Mr. K. HAGAN

kindly wrote two Fortran programs used in the analysis. My studies at Cambridge Universitywere funded by the Brazilian Science and Technology Council (CNPq).



JOINT-TERRITORIALITY IN TAMARIN MONKEYS 221

WILLSON, 964; REED, 1982), or between mixed-species groups sharing a

territory, and jointly defending it against other similar mixed-species

groups (MUNN&

TERBORGH, 979).Under resource-limited

conditions,associated heterospecifics should segregate to a certain extent (e.g.MAY,

1974).Yetjoint-territorialityshould be expected to occur more frequentlyin stable mixed-species groups, because low interspecific overlaps in

spatialand resourcerequirementsshould increase the costs, and decrease

the benefits of forming and defending a common territory.It should also

be more likely to occur in species sharinga renewingfood supply(WASER,

1981), because in this case asynchronoususe of resources would decrease

foraging efficiency,even if the number of consumersremained

unchanged(CODY, 1971; DAVIES & HOUSTON, 1981).

That spatial priorityof access to resources,or exclusion of neighbour-

ing competitors,entails a trade-off,is a well-establishedand testedpropo-sition (e.g. GILL& WOLF,1975). Resources should not be defended unless

the benefits of doing so are greater than the costs (BROWN,1964; BROWN

& ORIANS,1970). Costs of territorial defence often equate to high ener-

getic expenditure and decreased foraging success associatedwith agonis-tic interactions with

non-residents,or

overexploitationof

peripheralfood-supplies shared by neighbours (PATON & CARPENTER, 1984; PERES, 1989).On the other hand, this frequently results in the exclusion of non-resi-

dents from an area, which denies them access to resources. This increasesthe local availability of food to territorialresidents, which constitutes a

common benefit of territoriality (GILL, 1978).It has been arguedthat the enhanced cooperativedefence of a common

territory is an important benefit of mixed-species associations, whether

they maybe between

primates (TERBORGH,983, 1984), birds (MUNN &TERBORGH, 1979), or coral-reef fish (ROBERTSON et al., 1976). For

instance, a permanent mixed-species group of guenons in Gabon pro-gressivelyincreased its range size over a 5-yearperiod to the detriment ofa neighbouring monospecific group (GAUTIER-HIONt al., 1983). How-

ever, studiesof mixed-species societies have implied, or failed to tests the

underlyingassumption,that differentspecies holding a common territoryequally share the costs and the benefits of territorialdefence (e.g.MUNN&

TERBORGH, 1979; TERBORGH,1983, 1984). Costs are likely to differ if, forinstance, differentspecies invest differentamounts of time and energy interritorialcontests, or if in doing so, one speciesbecomes morevulnerableto predators than others. The benefits derived by different species fromresource defence also need not be symmetric,particularly f spatial over-

lap between neighbouring mixed-species groups affects the distribution



CARLOS A. PERES

and availability of each species' food supply in different ways. For

instance, the consequences of range exclusion for two species using differ-

ent-sized food patches should differ, because small patches are more

limiting, and more depletable, than large patches.

In this paper I investigate the joint-territorial behaviour of two small-

bodied primate species the Avila-Pires saddle-back tamarin (Saguinus

fuscicollis avilapiresi)and the red-capped moustached tamarin (Saguinus

mystaxpileatus)-

resulting from their highly stable mixed-species associa-

tions (PERES,1991, 1992). Possible costs and benefits of resource defence

are examined in terms of (1) energetic investments allocated to territorial

interactions, and(2)

increased foraging success derived fromrange

exclu-

sion of neighbouring groups. Two mechanisms of resource competition

(PARK,1954) are identified. Interference competition occurs when contes-

tants decrease one another's access to resources via direct aggressive

interactions, which are often highly conspicuous. Exploitative competi-tion occurs when competitors, which need not take part in direct contests,

deny one another's access to a common resource supply by directly

depleting it (e.g. DAVIES & HOUSTON, 1981; PATON & CARPENTER, 1984;

PERES,1989). Mixed-species groups

of tamarinscompete

with other

groups by interference primarily during intergroup encounters, and by

exploitation when foraging within extensive home range areas shared byother groups. Comparisons between different contexts of intergroupinteractions are used to examine the effects of interference on tamarins'

time budgets, use of space, and foraging success. Comparison between

exclusive and overlapping areas in the home range are used to examine

the consequences of exploitative competition of these same parameters.

Methods

Studyarea and animals.

This study was conducted at an entirely undisturbed, low land (53-71 m a.s.l.) tropicalmoist forest, located 4 km inland from the upper Urucu river, Amazonas, Brazil (4°50'S,65°16'W). This site appears to be representative of vast expanses of yet inaccessible and

poorly known unflooded (terra firme) forests in remote regions, which account for 95% ofthe Amazon basin. Most of the 900-ha study plot of high forest was located within a refined115-km2 trail grid which had been created and then abandoned

by

a seismic company, buthabitat disturbance and selective hunting had never taken place (PERES,1991). Meanannual rainfall in this plot was 3256 ± 589 mm (N = 2 full years).

Thirteen primate species occur at this site, including the saddle-back tamarin (adultweight - 394 + 42 g, N = 17), and its larger-bodied congener, the moustached tamarin

(adult weight = 523 ± 70 g, N = 6). These two tamarin species consistently formed andmaintained highly stable heterospecific groups during 97% of their activity period through-out the year (PERES,1991, 1992). Monospecific groups occurred at a density of 1.78 groups/

222




km2, maintained a very stable membership throughout the year, and consistently associated

with the same group of the other species, with which they shared a perfectly congruenthome range of approximately 150 ha (PERES,1991).

Behaviouralsampling.

Systematic observations were conducted on habituated groups of both tamarin species

during 5-7 days in each of 14 consecutive months (August 1988-September 1989), when

groups were followed continuously, and located every 10 and 15 min throughout their daily

activity period. Data reported here are based largely on one study group (hereafter, main

group) of 5-8 saddle-backs (2-3 adult males, 1 adult female, 0-1 subadult male, 0-2 subadult

females, and 0-2 infants or juveniles) and 8-10 moustached tamarins (3-5 adult males, 1-2

adult females, 1-2 subadult males, 0-2 subadult females, and 0-2 infants or juveniles)observed during 106 days. On 76 of these days I was certain whether none, one, or more

intergroup encounters had occurred. Locations and activities of 5 other groups, whose

home ranges partially overlapped that of the main group, were also noted whenever

possible. These groups became individually recognizable primarily because of the distinc-

tive patterns of ear-scars of adult moustached tamarins.

Behavioural data consisted of scan samples (ALTMANN,974) obtained every 10 min

throughout the day from all visible animals of both species in the group, but I strived to

sample individuals of all age-sex classes during each group of scans regardless of their

behavioural differences. Behavioural patterns were divided into the following broad, mutu-

allyexclusive

categories: (1) move, (2) rest, (3)feed

(on plant-items), (4) forage (foranimal

prey), (5) within-group and (6) between-group social interactions. These data were recorded

in the same fashion during and outside intergroup encounters to allow behavioural compar-isons. The time encounters began and terminated, and thus their duration, was also noted.

The foraging success of each tamarin species was examined by considering the animal

component of their diet. Captures of animal prey, such as a wide range of orthopterans,involved clearly identifiable foraging patterns which were distinctive between species

(PERES, 1992). Scans of individuals performing a foraging maneuver (e.g.pounces, snatches,

lunges, swipes, manipulations) were prolonged for up to 10 sec to allow me to quantifywhether or not such capture attempts yielded a successful outcome. Capture success is then

defined per foraging effort (i.e. the number of prey items obtained per unit of foraging scan),

and per foraging maneuver (i.e. the proportion of prey capture attempts which actuallyresulted in successful captures). For these purposes, therefore, capture attempts which

resulted in undetermined outcomes were disregarded. The prey capture success of each

species were then compared between different contexts of intergroup interactions, and

between different parts of their home range.A combination of contiguous 0.92-ha quadrats used exclusively by the main group is

defined as an exclusive area, whereas overlapping areas refer to quadrats shared with other

groups. Encounter quadrats were those occupied during direct intergroup encounters.

These were located in the periphery of each group's home range, were not necessarilycentred around key food sources such as food trees, but were used consistently throughoutthe year by any two neighbouring groups. Approach quadrats are defined as those near the

range periphery, and were used primarily when groups moved towards, or withdrew from,one another shortly before, and shortly after a given encounter.

Centrality is defined as the straight-line distance between any given quarter-hour loca-

tion and the geometric centre of the group's home range. Centrality distances were

computed by a Fortran program for each 0.058-ha quadrat occupation record based on

their x, y intersection, and that of the home range centre. Because home range geometrycan affect the congruence between centrality distances and proximity to boundaries, the

average distance from the home range centre to territorial boundaries (c) was evaluated as a



224 CARLOS A. PERES

function of range eccentricity (E), which ranges from zero (in a perfect circle), to one (in an

infinite straight line). Home range shape has essentially no effect on 'c' unless it is highlyeccentric (E > 0.9: MITANI& RODMAN,1979). The study group's range eccentricity was,

however, very low (E = 0.33), suggesting that results obtained from centrality analysis wererobust with respect to range shape.

Statisticalanalysis.

Non-parametric statistics follow SIEGEL 1956). Two-sample Kolmogorov-Smirnov and

Mann-Whitney U-tests were used to examine differences between two observed distribu-

tions, and other continuous variables. Categorical data were analysed using log-likelihoodratio tests (G-tests: SOKAL& ROHLF,1981), and X2tests. I follow BROWN(1974) in examiningcell contributions to overall X2 values. Activities which contributed the most to x2 values

were examined by replacing observed numbersof

recordswith those

expectedfrom null

spatial or temporal contexts. Cells were not considered significant if the exchange of their

observed values resulted in a decrease of the overall X2to less than the critical X2value at p= .05 (e.g. BROWN,1974).

There appeared to be little temporal dependency in the data sets used here, partlybecause not all individuals of each species in the group were observed during single 10-min

intervals, and activities were not necessarily synchronous between conspecifics. For

instance, when comparing greatly asymmetric sample sizes, medium frequency scores of

activities from 10 randomly subsampled data sets, which approximately matched that of the

smaller data set, were not more than 1% different from their respective proportions in the

entire original samples. It was thus possible to control for unrealistically inflated signifi-

cance levels due to unequal sample sizes.Three-way contingency tables were based on log-linear models (SOKAL& ROHLF,1981),

and used to test for interactions between categorical variables, and their effects on a

frequency variable, such as the number of scans in different activities or successful prey

captures. For instance, the spatial heterogeneity of capture success was tested by examiningthe effects of location of capture attempts, and tamarin species attempting a capture, on

whether captures were successful or not. Because prey capture attempt could have only two

outcomes, frequencies of foraging scans and capture maneuvers were incorporated into the

model as dichotomous variables defined by whether they resulted in successful captures.Distance from capture location to the range centre was assigned to one of seven centrality

zones, and fitted as a

categoricalvariable.

Logistical regressionmodels were then used to

test for differences in estimated probabilities of prey captures by tamarins in different

centrality zones. Tamarin species and centrality zone were fitted as covariates of capturessuccess. Whether or not coefficients were equal to zero were then tested by the Wald

statistic (SOKAL & ROHLF, 1981), which follows a X2 distribution.

When comparing encounter days with days in which encounters did not occur, I used an

analysis of covariance (ANCOVA) on each dependent variable using one or more indepen-dent variables. In every case, the slope heterogeneity was tested before testing for differ-

ences in the intercept. All data were analysed using SPSSx (NIE et al., 1975).

Results

Patterns of intergroup interactions.

Two contexts of interactions between neighbouring mixed-species groupsare distinguished: long-range and face-to-face encounters. Long-rangeencounters were defined by intergroup distances of 40 m or more, and



JOINT-TERRITORIALITY IN TAMARIN MONKEYS

were characterized by frequent exchange of long-calls between groups.

The frequency of counter-calling increased as neighbouring groups

approached one another, which invariably escalated to a direct inter-

group contest, or face-to-face encounter. These disputes were defined by

short intergroup distances (< 40 m), when conspecifics of different groups

usually remained within sighting distance of one another, maintained

very high levels of aggression, and exchanged frequent long-calls, which

often fused into staccatos of short whistles, or 'chatter' calls (sensuPERES,

1986). Aggressive interactions and postures included piloerection,

arching displays (sensuRATHBUN, 1979), vigorous chases, and occasionally

actual fights. Chases often involved reciprocal displacements between

conspecifics of each group along boundaries, often resulting in rapid back

and forth travel by the entire group. Fights occurred when a chaser

caught up with its opponent, were characterized by explosive and vig-

orous wrestling, and could result in biting wounds. Aggressive inter-

specific interactions between members of neighbouring groups were never

observed. Encounters eventually terminated as neighbouring mixed-spe-

cies groups 'lost interest' and drifted apart towards the respective centres

of their home ranges.

Face-to-face encounters accounted for 9% of the main group's overall

time budget (N = 731 h). These tended to begin early in the day (0835 h 9

min, N = 39), last over one hour (74 ± 64 min, N = 39), and terminate by

mid-morning (0949 h ± 2 h 5 min, N = 39). On certain days, however,

tamarins allocated as much as 60% of their waking time (e.g. 356 min of a

589-min day) to long-range and face-to-face encounters. Although such

unusually prolonged encounters lasted until midday or early afternoon,

68.5% of timespent

in encounters took place before 0930 h (Fig. 1), and

those beginning after 0930 h were more likely to last for shorter periods (rs= -0.41, p = .01, N = 39).

All group members of both tamarin species were present in all face-to-

face encounters. Such encounters occurred almost every other day, and

were observed in 39 (51%) of 76 days spread equally over the 14 months

in which the occurrence of encounters was determined. The total number

of encounters in this period was actually 42 because double-encounters

with distinctneighbouring mixed-species groups

occurred in 3 of the 39

encounter-days. In an additional 8 days, long-range encounters led to

intergroup approaches within 150 m, which atypically failed to escalate

into actual face-to-face encounters. There were no between-month differ-

ences in the observed number of intergroup encounters. In fact, monthly

frequencies of encounters were very similar to that expected on the basis

225



CARLOS A. PERES

Jg 7 -n=/o aays

6 7 8 9 10 11 12 13 14 15 16

mixed-species group ofamar-s.

.80).

Patterns of intergroup spacing.

mixed-species group of tamarins.

of observation effort (e.g. number of complete days, X2 = 5.8, 12 df, p >

.80).

Patterns of intergroup spacing.

A total of 5 mixed-species groups of tamarins had home ranges contig-

uous or partially overlapping with that of the main group. There was a

large spatial overlap between the home range of all neighbouring groups

and that of the main group (Fig. 2); 163 (76.2%) of 214 quadrats of 0.92-

ha in the group's range were known to be used by at least one other

group. Despite this large overlap, the main group still retained an exclu-sive area of 50.7 ha.

Areas of overlap appeared to be stable throughout the study, and

encounters between any two groups appeared to occur consistently within

the same encounter quadrats. As expected, areas occupied during face-to-

face encounters were biased towards the range periphery (Fig. 2). Central-

ity distances during encounters averaged 602.8 + 133.8 m (N = 221), and

where thus significantly longer than those at other times (437.1 ± 179.3, N

= 2046; Kolmogorov-Smirnov, Z = 7.6, p < .001).Mixed-species groups of tamarins were never observed to change the

direction of their movements to avoid intergroup encounters. Rather,

they appeared to move towards the location of neighbouring groups as

soon as their long calls were heard. If movements of neighbours were

independent of one another, the probability of simultaneous visits to the

226

f I 21




Typeof quadrat

* Encounter

H Approach| Overlap

D[ Exclusive

Fig. 2. Home range of the main mixed-species group of tamarins divided into 0.92-haquadrats. Encounter and approach quadrats were those where face-to-face encounters, and

intergroup approaches took place, respectively. Overlapping quadrats were those shared byother groups, but where encounters did not take place. Exclusive quadrats were those used

only by the main group.

same encounter quadrats should be low. The observed tendency of

mixed-species groups to approach and collide with one another was thus

tested against that expected from WASER'S 1977) random movement

model (see also WASER& WILEY, 1979). This model generates an expectedfrequency of encounters on the assumption that groups move indepen-

dently of each other as in a 'two-dimensional gas'. Expected daily encoun-

ter frequency (/ is thus defined as:

f= 8 p a (d+s)/7,

227



CARLOS A. PERES

where 'p' is the density of mixed-species groups (per km2); a is the average

group velocity (km/day); 'd' is the criterion distance at which an encoun-

ter is said tooccur;

and 's' is theaverage spread

of thegroup.

Given a

density of 1.78 groups/km2, a mean group velocity of 1991 m/day, a

mean group spread of 20 m (PERES, 1991), and that face-to-face encoun-

ters were defined by intergroup distances ranging from 0 to 40 m,

expected rates of encounters ranged between 0.180 and 0.541 encounters

per day. The total observed frequency of 0.553 encounters per day (42encounters in 76 days) was thus higher than predicted to occur in the

same period (13.7 <f < 41.1 encounters).

Effects of defence on ranging patterns.

Mixed-species groups of tamarins in the Urucu jointly held very large

areas. The main group used a 141-ha home range, an estimate based on

0.23-ha quadrats and locations every 10 min, and two other groups

followed less systematically were known to use areas of at least 85 and 110

ha, respectively. The main group used most intensely those home range

areas shared by other groups, and spent nearly twice as much time inencounter quadrats than expected from their availability in the home

range (Table 1). Encounter and approach quadrats were used more often,

and those elsewhere less often, than expected by chance (X2= 406.2, 2 df,

p < .001).

Use of peripheral quadrats occurred primarily early in the morning,

following rapid boundary approaches. Tamarins then remained close to

boundaries, which resulted in centrality distances before 1200 h being

longer than those in the afternoon (Kolmogorov-Smirnov, Z = 5.21, p <.001, N = 2267; Fig. 3a). Centrality distances were on average 453 ± 182

m (N =2263), and skewed towards sub-terminal areas in the home range

(Fig. 3b); tamarins spent more time away from the range centre than

expected on the basis of range areas occupying 19 concentric zones with a

radial width of 50 m (G-test, G = 111.0, 18 df, p < .001). The group

usually approached, entered and remained within the range periphery

regardless of whether a neighbouring group had been detected. The

amount of time tamarins remained within these areas was thus not simplydetermined by the presence of other groups. Indeed, only a small fraction

of the time allocated by the group to encounter quadrats was actually

spent in encounters. Because tamarins always vocally advertised their

presence through long-calls when approaching encounter zones, theywere unlikely to go undetected unless their calls were outside a neigh-

228




TABLE 1. Number of group locations within encounter and approach

quadrats, and elsewhere in the group's range

Quadrats Number of group locations

Quadrat zone Freq. % Observed % Expected X2

Encounter 25 11.3 692 21.8 359.0 308.7*

Approach quadrats 21 9.5 391 12.3 301.6 26.5*

Other areas 175 79.2 2091 65.9 2513.3 71.0*

Total 221 100 3174 100 3174.0 406.2*

* 2 df, p<0.001. Expected number of locations were calculated on the basis of the area

available to the group within each type of quadrat.

bour's hearing range. This is supported by the fact that tamarins invaria-

bly discontinued their activities, initiated counter-calling, and moved

towards the other group, in response to long-calls of neighbours. They

then tended to remain in encounter quadrats, even if encounters failed to

escalate, at least until those calls were no longer heard.

The study group entered 16.6 ± 3.5 (N = 38) quadrats during complete

days in which encounters occurred, and 14.4 ± 4.9 (N = 30) quadratsduring days without encounters (Mann-Whitney U-test, Z = 1.64, p =

.10). This failed to result in a significant difference in the diurnal diversity

of quadrat use, as determined by the Shannon-Wiener index (encounter-

days: H' = 2.50 ± 0.25; non-encounter days: H' = 2.36 ± 0.42, Mann-

Whitney U-test, Z = 0.77, p = .44), even when differences in observation

time were controlled for. Both daily number of quadrats entered

(ANCOVA, F1,60 = 2.5, p = .12) and daily quadrat diversity (ANCOVA,

F1,60 = 1.2, p = .28) remained unchanged between encounter- and non-encounter days. This can be explained by the fact that the slightly greater

area used during encounter days was compensated for by the more even

allocation of time to different quadrats during days without encounters,

when tamarins were not forced to spend a discrepantly large amount of

time within encounter quadrats.

Whether or not encounters occurred on any given day affected the

distance travelled by the main group during complete days. The distance

covered on encounter days (2104 + 313 m, N = 38) was greater than thaton days without encounters (1835 ± 484 m, N = 30; Mann-Whitney

U-test, Z = 2.0, p = .04). This reflects differences in group velocity, which

was greater during encounter-days (234 ± 48 m/h, N = 38) than during

days without encounters (196 ± 56, N = 30, Mann-Whitney U-test, Z =

2.54, p = .01). Distances moved during intergroup approaches (when they

229



CARLOS A. PERES

._ n=2263 locations

C 600-.' -- lal

o 550--

80

o 450 --- -

400-

u 350-

'5 300- -4

250 .. . . . . . .

6 7 8 9 10 11 12 13 14 15 16

Time of day

350

n=2263300-

I ~~~~~~~*bi· 250-

200-

,150-

50

100-

0 100 200 300 400 500 600 700 800 900

Distance (m)from home rangecenter

Fig. 3. (a) Diurnal variation in group centrality during quarter-hour locations throughout the

study, showing (b) the overall distribution of time spent in different centrality zones. Numbers

of locations obtained (n) are shown for each figure.

reached theirgreatest velocities)

werehighly inconsistent, ranging

from

less than 100 m to 920 m, a distance almost as long as their home range

diameter.

MITANI & RODMAN'S(1979) ratio between day ranges and range size, or

'index of defendability' D (defined by D - d/d', where d' = /4A/a)

suggests that the main group was able to defend space by using a large

part of its range per unit time and frequently approach different points in

the range periphery. Given that the diameter of a circle (d') with an area

(A)equal to the main group's home range size was 1389 m, and that the

group moved a mean distance (d) of 1991 m per day, D was 1.43 the

average number of home ranges which could be crossed by the group on a

daily basis. This allowed tamarins regular visits to territorial boundaries,

regardless of how their movements were affected by the location of other

groups.

230




TABLE2. Proportion of time spent by each tamarin species, and both

species combined ("both sp."), in different activities during long-range

and face-to-face encounters, and outside encounters

Long-range Face-to-face Outside

encounters encouters encounters

Activity pattern S.f S.m. both sp. S.f S.m both sp. Sf S.m both sp.

Move 42.6 * 43.4 43.0 45.4 * 47.0 46.4 31.4 28.4 29.8

Rest 25.0 * 16.2 20.0 24.9 * 5.0 13.1 22.8 17.4 19.8

Feed 12.0 * 12.3 12.2 11.7 * 10.2 10.8 26.6 27.6 27.2

Forage 9.9 * 16.3 13.6 5.7 * 11.7 9.3 14.6 21.2 18.2

Intra-group social 1.7 2.5 2.2 1.1 * 1.2 1.2 3.0 3.3 3.1

Inter-group social 8.4 * 8.6 8.5 10.6 * 24.7 18.9 0.9 1.3 1.1

Other0.3 0.7 0.5 0.6 0.1 0.3 0.7 0.8 0.7

Number of scans 573 754 1327 615 895 1510 11230 13814 25044

X2values 434 445 870 775 4121 4778

*Log-linear models on interspecific comparisons: log-likelihood ratio tests, 3 df, p<0.0001.

Italic numbers indicate significant cell contributions to overall X2 values (BROWN,1974)based on species-specific comparisons of scan frequencies during each encounter context,

and those outside encounters. Log-linear models are based on 3-way interactions testing for

interspecific differences in scan frequencies of each activity between each encounter con-

text and that outside encounters.

Effects of defence on time budgets.

Time budgetsduring ntergroupncounters.

The time budget outside encounters of either tamarin species was signifi-

cantly different from those observed during long-range and face-to-face

encounters (Table 2). During face-to-face encounters, tamarins of both

species spent significantly more time moving and interacting with non-

group conspecifics, and significantly less time feeding and foraging, then

when they were outside encounters. Moustached tamarins spent signifi-

cantly less time resting during encounters, whereas saddle-backs did not.

The same was true for long-range encounters, except that the difference

in the amount of time spent foraging was not significant. Consideringboth species combined, standardized residuals were greatest for feedingand moving during long-range encounters, and feeding, intergroup social

interaction, and resting during face-to-face encounters, whereas other

activities failed to reach significance. At close quarters to other groups,

moving and intergroup interactions by both tamarin species accounted

for twice the amount of time of that outside encounters. Feeding, which

accounted for over one quarter of tamarins' null time-budget, declined to



232 CARLOS A. PERES

11-12% of their time during face-to-face and long-range encounters,

respectively.

Time budgets also differed between the two types of encounters (X2 =93.5, 6 df, p <.001), although only two activities - rest and intergroupsocial - were significant contributors to this difference. Animals

appeared to allocate a similar amount of time foraging for arthropodsbetween contexts outside encounters and those of long-range encounters.

This may be explained if animals anticipated long-range interactions to

escalate into direct confrontations, and increased their foraging effort to

compensate for the small amount of time which could be spent on

foraging during actual boundary contests.These differences although derived from lumping all age-sex classes,

did not appear to have resulted from oversampling behaviourally conspic-uous individuals during contexts of intergroup interactions. For instance,

observed frequencies of scans obtained from adult males - the most

active members of the group during territorial encounters - were not

significantly different from that expected based on their representation in

each monospecific group during either long-range (fuscicollis: X2= 0.07, 1

df, p = 0.79; mystax:X2 = 0.22, 1 df, p = 0.88) or face-to-face encounters(fuscicollis:x2 = 0.43, 1 df, p = 0.51; mystax:x2 = 2.17, 1 df, p = 0.14).

Interspecific ifferencesn timebudgets.

Although both tamarin species were involved in all encounters, their time

budgets differed significantly from one another during either face-to-face

(G- test, G = 170.6, 6 df, p <.001) or long-range encounters (G = 24.1, 6

df, p <.001), suggesting an asymmetric division of labour in their terri-torial defence effort (see Table 2). Interspecific differences in time budgetswere greater during either types of encounters than at other times, as

examined by 3-way matrices based on loglinear models, using scan fre-

quencies of all activities as a dependent variable of species identity and

intergroup interaction context (long-range encounters: G = 749.4, 18 df,

p <001; face-to-face encounters: 1770.0, 18 df, p <.001). Comparing the

time tamarins allocated to each activity against that to all others, inter-

specific differences remained highly significant for all activities duringboth types of encounters (log-linear models, 3 df, p <001 in all cases),

except for 'within-group social interactions' (G = 7.2, 3 df, p = .07) and

'other' (G = 2.5, 3 df, p = .48) during long-range contests (Table 2).Moustached tamarins usually dictated the velocity of intergroup

approaches and withdrawals, were the first animals to come into contact




with, and the last to retreat from, conspecifics of a neighbouring group,and led the course of face-to-face encounters. This species spent 25% of

their time in what appeared to be highly conspicuous and energetically-demanding intergroup agonistic interactions. Frequent chases and dis-

placements between moustached tamarins accounted for the bulk of the

intergroup aggression displayed by the entire mixed-species group. In

contrast, saddle-backs spent only 11% of their time during encounters

interacting with non-group conspecifics.The proportion of time spent resting during encounters by moustached

tamarins (5%) was less than a third of that outside encounters, whereas

saddle-backs spent slightly more time resting during encounters (25%)than at other times. During joint-territorial disputes, saddle-backs of all

age-sex classes were often found quietly resting at inconspicuous and

retreated positions as far as 40 m from their quarrelling heterospecifics.

Usually, however, they adopted a passive role of 'stand-by' spectators,

watching moustached tamarins chase one another, and only occasionally

engaging in their own squabbles with conspecifics of the other groups.Face-to-face encounters also comprised the only context in which saddle-

backs spent a greater (albeit not significantly different: G-test, G = 0.9, 1df, p = .34) amount of time feeding than moustached tamarins. This is

supported by the fact that saddle-backs spent more time feeding duringthe time of day when most encounters occurred (0700-1100 h) than at

other times (G = 16.2, 1 df, p <.001), whereas this difference was not

significant for moustached tamarins (G = 0.9, 1 df, p = .34).In addition, saddle-backs consistently 'lost interest' in encounters

sooner than moustached tamarins, and collectively moved away from the

encounter site. Their persistent long-calls during such premature with-drawals appeared to be targeted to moustached tamarins, and often

preceded the termination of encounters. Indeed, the only time the two

species were observed to drift apart took place towards the end of a

prolonged encounter, when saddle-backs clearly drifted off while mous-

tached tamarins still vigorously confronted their neighbours.

Timebudgets

within encounterquadrats.

In order to disentangle the potential effects of exploitative and inter-

ference competition on tamarins' activities, time budgets were examined

within and outside encounter quadrats, but outside contexts of actual

encounters. Time budgets outside encounter quadrats, in this case, were

considered as the null state. Only 14.3% of 10830 minutes allocated by



234 CARLOS A. PERES

TABLE 3. Proportion of time spent by each tamarin species, and both

species combined (both sp.), in different activities within and outside

encounter quadrats, excluding time spent in actual intergroup encountersInside encounter quadrats Outside encounter quadrats

Activity pattern S.f S. m both sp. S.f S. m both sp.

Move 28.6* 26.6* 27.5* 33.4 29.8 31.4

Rest 23.6 18.4 20.8 23.5 17.7 20.3

Feed 29.0* 27.9 28.4* 23.7 25.8 24.8

Forage 14.5 21.7 18.4 14.6 21.3 18.3

Intra-group social 2.6 3.3 3.0 3.2 3.3 3.2

Inter-group social 0.9 1.4 1.1 1.0 1.3 1.2

Other 0.7 0.7 0.7 0.6 0.8 0.7

Number of scans 3108 3796 6904 7806 9671 17477

*Significant cell contributions (p<.05) to overall X2values (BROWN, 974). X2tests are based

on species-specific comparisons between scan frequencies inside and outside encounter

quadrats (see methods): S.fuscicollis: X2 = 62.6; S. mystax:X2= 21.8; both species combined:

X2 = 71.0, 6 df, p<.001.

the group to approach and encounter quadrats was actually spent in

encounters. Thus groups appeared to be outside their neighbours' hearing

range during most of the time they spent in peripheral quadrats.There were significant differences between the time budgets of the two

species, both within (G-test, G = 80.7, 6 df, p < .001) and outside

encounter quadrats (G = 206.5, 6 df, p < .001); Table 3). Each species also

differed in its time budget between quadrat types, particularly in the

amount of time spent moving and feeding. The lower proportion of time

spent moving by both species, and higher proportion of time spent feeding

by saddle-backs, within encounter quadrats (but away from the other

groups) may be related to the high probability of intergroup encounters in

these areas. Tamarins frequently evoked responses from neighbouring

groups via long-calls, did not move away, and concentrated their sparetime around local food sources, which were clearly shared by other

groups. Again, interspecific differences in time budgets were a function of

the fact that moustached tamarins allocated a greater effort in assessingtheir neighbours' location. For instance, one or two adult males of this

species consistently drifted away from their core group-members, into the

adjacent territory, where they persistently called for, and apparentlylistened to other groups, clearly to the detriment of their own feedingsuccess.




Effects of centralityon timebudgets.

Distance from the home range centre affected the time budget of both

tamarin species, as there was a significant association between the

amount of time animals spent in different activities and group location in

one of seven centrality zones (G-tests, saddle-back: G = 298.8, 30 df, p <

.001; moustached: G = 408.0, 30 df, p <001; Fig. 4). There appeared to

be no significant differences between the amount of time each tamarin

species allocated to a given activity as they moved away from the rangecentre (Table 4). As centrality decreased, however, tamarins of either

species spent more time interacting with members of neighbouring

groups, and moustached tamarins spent less time in intragroup social

45 30

I-J- u S. fusdcollis* S.mystax

40 25

> 35 20

~o~~ V~o30 - 1 15

U) 25 'I 100 0 35 , 20

E (

30 - 16

c o

25 I12

0 0

*- ~~~~~~~u5- 200

8

C 0

0 6-) 8

0.0

0 0

Fig. 4. Variation in the proportion of time spent by each tamarin species in different activities

as a function of centrality zone.

0-200 200 300 400 500 600 >700 0-200 200 300 400 500' 600 >700

Distance (in) from the home range center

Fig. 4. Variation in the proportion of time spent by each tamarin species in different activities

as a function of centrality zone.



236 CARLOS A. PERES

TABLE4. Comparisons between each tamarin species' time budget in each

of 7 centrality zones, and correlations between centrality and amount of

time each species spent on different activities

Interspecific differencesa Spearman rank correlationsb

Activity pattern G p S.fuscicollis S. mystax

Moving 6.5 0.37 0.43 0.61

Resting 9.1 0.17 -0.61 -0.50

Plant-feeding 3.0 0.81 -0.43 -0.50

Foraging 11.3 0.08 0.54 0.64

Intra-group social 6.9 0.33 -0.50 -0.86**

Inter-group social 5.2 0.51 0.93** 0.75*

a

Log-likelihood ratio tests, 6 df;b n= 7;*

p<0.1;** p<0.05.

activities (Table 4). Other activities were not significantly correlated with

centrality. That centrality most obviously affected the amount of time

animals spent in intergroup social interactions is expected, since these,

with the exception of long-calls, could not occur away from rangeboundaries.

Effects of defence on foraging success.

Capture uccessduring ntergroupncounters.

Capture success of each tamarin species, considering both foraging time

and maneuvers, was calculated in different contexts of intergroup interac-tions (Table 5). These parameters within tamarin species were not signifi-

cantly different between either types of encounters and non-encounter

null states (G-tests, p > .10 in all cases). There were also no three-waydifferences in the effect of encounter context on rates of successful cap-tures by each tamarin species (log-linear models; prey items per capture

attempt: G = 0.91, 2 df, p = .64; prey items per foraging scans: G = 0.96,

2 df, p = .62). In real terms, however, few prey-captures took place during

encounters compared to those outside encounters because the amount oftime spent away from neighbours was substantially greater. Thus, inter-

group distance per se may not be an important variable in determiningdirect arthropod foraging costs and benefits of territorial defence. The

effects of spatial overlap with neighbouring groups on capture success,

independently of intergroup distance, is examined below.




TABLE5. Percentage of time spent foraging by each tamarin species, and

both species combined (both sp.), in different encounter contexts, showing

the proportion of foraging scans and prey capture attempts which resultedin successful captures

Long-range Face-to-face Outside

encounters encounters encounters

Percent of time Sf S.m both sp. Sf S.m both sp. Sf S.m both sp.

Stationary foraging 5.8 9.4 7.8 1.8 5.5 4.0 5.7 10.5 8.3

Foraging maneuvers 1.9 2.0 2.0 1.8 1.6 1.7 4.2 3.5 3.8

Capture success (%)

Per foraging scan 10.2 23.2 19.0 18.5 33.0 28.2 14.1 21.7 19.0Per capture maneuver 31.3 63.4 53.6 31.3 68.9 56.9 27.9 52.4 42.7

Number of scans 573 754 1327 615 895 1510 11230 13814 25044

Number of prey captures 5 26 31 5 31 36 182 536 718

Prey items were considered irrespectively of their size class.

TABLE6. Percentage of time spent foraging by each tamarin species, and

both species combined, within and outside encounter quadrats, showingthe proportion of foraging scans and prey capture attempts which resulted

in successful captures

Inside encounter quadrats Outside encounter

quadratsPercent of time S.f S.m both sp. Sf S.m both sp.

Stationary foraging 5.3 10.3 8.1 6.1 10.9 8.8

Foraging maneuvers 4.4 4.1 4.0 4.3 3.5 3.9

Capture success (%)Per foraging scan 13.3 21.8 18.9 12.5 20.5 17.7

Per capture maneuver 25.3 50.2 40.8 25.6 51.5 41.3

Number of scans 3108 3796 6904 7806 9671 1747

Number of prey captures 46 151 197 116 361 477

Prey items were considered irrespectively of their size class.

Capture uccesswithin encounterquadrats.

Use of encounter quadrats outside actual encounters determined no sig-nificant effect on the amount of time tamarins of either species spent in

stationary foraging, foraging maneuvers, foraging in general, and on

frequencies of prey captures (G-tests, p > .20 in all cases; Table 6).



238 CARLOS A. PERES

Capture success also remained unchanged for both species between these

two types of quadrats: there were no significant three-way interactions

when a log-linear model was applied to capture frequencies per capture

attempt (G = 0.03, 1 df, p = .87), or per foraging scan (G = 0.01, 1 df, p =

.99). Within encounter quadrats, both tamarin species combined

obtained 197 prey items during 18.9% of the time they spent foraging (N=

1041), and completed on average 40.8% of their capture attempts (N =

483). Outside encounter quadrats, the 477 observed captures resulted in a

slightly lower capture rate per unit of foraging time (17.8%, N = 2687),but slightly greater rate per unit of capture attempt (41.3%, N = 1155).

Because encounter quadrats represents an arbitrary measure of tam-

arins' fields of isolation, and accounted for only a small part of a group'stotal range area overlapping with those of other groups, foraging success

was further examined as a function of distance to range boundaries.

Effects of centralityon capture uccess.

The effects of group centrality on its foraging efficiency were examined

using

7

contiguous rings

of

activity,

concentric on the home

range

centre.

For these purposes, foraging data were pooled within the two most

central, and two most peripheral zones, because of inadequate samplesizes. Resources near the range centre are least likely to be depleted byother groups, and are subject to reduced intergroup exploitative competi-tion. This should allow resource availability and predictability to be

greater in these areas to the resident group. Residents' foraging success

should then increase with distance from range boundaries if other groupsare able to affect the distribution of a

particularresource

type.Capture success of tamarins was affected by the forager's species iden-

tity and its centrality zone (Fig. 5). These two variables, when incorpo-rated into a log-linear model, determined significant effects on the

number of foraging maneuvers yielding successful captures, whether theywere considered separately, or in a two-way interaction (Table 7). The

same effects applied to the number of captures when overall scan frequen-cies of all foraging activities were considered.

The number ofprey captures per foraging

scan for saddle-backs

decreased significantly as the group moved away from the range centre,into the range periphery (r = -0.78, 6 df, p = 0.04), but not for mous-

tached tamarins (r = -0.71, 6 df, p = .07; Fig. 5a). This effect was far more

pronounced within intermediate zones of 300-600 m from the home

range centre, the portion of the group's range subjected to the most




0.30

* mystax

.2-- fusdcollis0.25

0.20

(o

0 0.15

(a)

E ,________ ___

0.10

0.050-200 200 300 400 500 600 >700

E 0.6

(d 0.4 -

0.3

0.2

0.10-200 200 300 400 500 600 >700

Distance(m)from therangecenter

Fig. 5. Effects of group centrality on each species' prey capture success, defined as the number

of captures obtained (a) per foraging effort (time spent foraging), and (b) per foragingmaneuver (individual capture attempts).

intensive use by the main group and its neighbours. In this area alone,

centrality explained 98% (r = -0.99, 3 df, p = .01) of the variation in

capture rate per attempted capture by saddle-backs (which were largely

manipulative foragers), but only 9% (r = -0.30, 3 df, p = .70) of that by the

visually foraging moustached tamarins.

The frequency of successful prey captures per foraging maneuver byboth tamarin species also suffered a significant decrease with increasing

distance from the range centre (saddle-back: r = -0.74, 6df, p = .06;moustached: r = -0.84, 6 df, p = .02; Fig. 5b). Capture success per

foraging scan was again strongly correlated with distance to boundaries

within the four intermediate centrality zones, explaining 99% of this

parameter for saddle-backs (r = -0.10, 3 df, p = .004), but only 16% ofthat for moustached tamarins (r = -0.40, 3 df, p = .60).



CARLOS A. PERES

TABLE 7. Log-linear models showing the effect of tamarin species and

centrality on the capture success of animal prey items

Per foraging maneuver Per foraging scanSource G df p G df p

Tamarin species 26.3 12 0.01 31.1 12 0.002

Centrality zone 138.1 7 0.001 76.1 7 0.001

Species and centrality zone 12.1 6 0.06 23.4 6 0.001

TABLE8. Parameter estimates for a logistic regression model considering

centrality distance and tamarin species as independent covariates ofwhether prey captures were successful or not

Per foraging maneuver Per foraging scan

Independent variable Coeff.±SE Wald test Coeff.±SE Wald test

Tamarin species 0.582±0.054 116.1** 0.331±0.047 48.9**

Centrality zone 0.095±0.029 10.6** 0.044±0.024 3.4*

Constant 0.110±0.128 0.7 1.416±0.105 181.9**

** 1 df, p<.001; * p = .067.

These linear correlations are confirmed by a logistic regression model

where centrality zone was a significant predictor of frequency of prey

captures per foraging maneuver for both tamarin species (Table 8). Fre-

quency of prey captures per foraging scan also differed significantly

between tamarin species, but failed to reach significance between central-

ity zones.

Discussion

Resource defence by interference displayed by mixed-species groups of

Saguinus s similar to that documented for many other territorial species.

All wilds callitrichids, for instance, are typically neighbour-intolerant,

and use regular long-distance vocal signals to seek out, vigorously

approach, and confront other groups. Intense territorial interactions

occur along relatively narrow boundaries between groups of lion tamarins(RYLANDS, 1983; PERES, 1989), marmosets (RYLANDS, 1983; HUBRECHT,

1985), and other tamarins (TERBORGH, 1983, 1984; GARBER, 1988).

Intergroup vocal interactions, such as long-calls, occurred mostly early

in the morning, when conditions for sound transmission in tropical forests

are nearly optimal (WASER & WASER, 1977). Tamarins moved to bound-

240




aries at this time of day, a pattern similar to that of many other territorial

primates (e.g. gibbons: CHIVERS, 1974; titi monkeys: ROBINSON, 1979).

Intenselong-calling

whilstneighbouring groups approached

one another

then usually escalated into ritualized intergroup encounters, often within

relatively restricted, and apparently traditional sites. These encounters

occurred more frequently than expected if movements of groups were

independent, which markedly contrasts with many other forest primatesthat adopt mutual-avoidance spacing systems, and encounter neighboursat a rate lower than predicted by chance (e.g. WASER, 1976; WHITEHEAD,

1987).

Centripetal movements, particularlyafter territorial

interactions,occur

in many forest primates (WASER, 1976; ROBINSON, 1979), suggesting

avoidance of resource-depressed, peripheral areas. This may be possiblewhen defence by interference alone prevents intruder trespassing, effec-

tively excluding other groups from a resident's home range (e.g. ROBIN-

SON, 1979). In contrast, use of space by tamarins and lion tamarins is

biased towards the range periphery, particularly encounter zones, despitethe extensive range overlap between neighbouring groups (PERES,1989,this

study).If

range boundaries are defined by the location of intergroupencounters, then they did not coincide with the most peripheral rangeareas. This is supported by the fact that tamarins' aggression fields were

apparently highest in subterminal areas of their range: when approached

by other groups near the fringes of its home range, the main group often

rapidly retreated a short distance into more central encounter quadrats.This is also the case in golden lion tamarins, whose highly defended

territories overlap as much as 60% with those of neighbouring groups

(PERES, 1986). The proportion of time spent in peripheral, rather thanexclusive, areas was substantially greater than that expected by chance.

This facilitated detection of intruders, allowed resource defence by

exploitation, and depleted resources which otherwise would have been

lost to neighbours.Use of the range periphery when neighbouring groups are far apart

appears to complement defence by interference, particularly when the

latter alone cannot exclude competitors from large feeding territories.

This is likely to be the case in mixed-species groups of tamarins in theUrucu because their home ranges are extremely large compared to those

in other Amazonian sites, as well as other frugivore/insectivore primatesof equivalent group biomass (PERES,1991). Range defence thus appearedto be very costly, even though is failed to exclude other groups entirely.Defence by exploitation also occurs when residents are unable to directly



242 CARLOS A. PERES

challenge intruders, whether because these are larger-bodied (GILL &

WOLF, 1975; KODRIC-BROWN & BROWN, 1978), or in large flocks or

schools(MYERS

t al., 1981). Exploitative competition between tamarin

groups through peripheral bias in range use can thus be seen as an

energetically inexpensive form of resource defence. This seems to play a

major role in the territorial system of these and other callitrichid species

(PERES,1989), analogous to that shown for other territorial vertebrates

foraging primarily in the range periphery, such as reef fishes (ROBERTSON

et al., 1976), hummingbirds (KODRIC-BROWN& BROWN, 1978; PATON &

CARPENTER, 1984), sunbirds (GILL & WOLF, 1975), and shorebirds (MYERS

et al.,1981).It is clear that day-to-day territorial interactions occurred frequently,

and were often long-lasting, even though they provided no short-term

benefits to either species of tamarins. Intergroup interactions, particularlyface-to-face encounters, imposed strong energetic costs on tamarins, gen-

erally resulting in greater proportions of time spent in energetically

expensive or negative activities, such as moving to boundaries, displaying

to, and chasing non-groups members. They allowed little time for ener-

geticallyconservative or

positive activities,such as

resting, feedingand

foraging. Time budgets during encounters suggest that the energetic

burden, at least through interference mechanisms, fell mostly on mous-

tached tamarins. During encounters, tamarins of this species engaged in a

comparatively greater range and frequency of agonistic interactions with

conspecifics of other groups, to which they allocated significantly more

time than did saddle-back tamarins. Fights between non-group members

probably explain the ear-cuts and scars, which occurred in nearly half of

all adult moustached tamarins in 5 groups, but not in any saddle-backtamarin. Moustached tamarins also determined the timing and direction

of intergroup approaches and withdrawals, generally leading territorial

contests. Saddle-backs, on the other hand, usually remained stationary in

the vicinities of encounters between their heterospecifics, often groomingone another, and spending far less time in energetically demanding, or

hazardous activities.

Because home ranges of groups of each species were congruent, the

potential benefits of defence could be equally shared by the two species.These benefits are, however, likely to be greater to saddle-backs - the

species which contributed with the least defence effort -because of

differences in the nature and distribution of foods used by the two species.What lack of overlap there is in use of plant-food patches was largelyattributed to patch size (PERES, 1991): moustached tamarins, on average,




fed on larger patches than did saddle-backs. Small fruit patches are most

susceptible to higher rates of depletion than are large patches, which

increases the relative foraging benefits associated with exclusive use of

space (CHARNOVt al., 1976; WASER,1981). In addition, some 48% of the

prey items captured by saddle-back tamarins were non-mobile and

embedded in very small, and highly depletable microhabitats (PERES,

1991). The abundance of these items need not be evenly distributed in

space. Rather, they should track the distribution of the specific micro-

habitats in which they hide. Once probed, a microhabitat is rendered

useless to a subsequent forager before it is recolonized by new prey. Local

scramble competition between different groups is thus potentially higherfor manipulative foragers, such as saddle-back tamarins. In contrast, preyitems exploited by moustached tamarins - a leaf-gleaning predator

relying on embedded prey in only 4% of its captures- were highly

mobile and widely scattered in the midstorey foliage. This probably

explains the observed differences in the effects of centrality on the prey-

foraging efficiency of the two species. Capture success was sustained at

comparable levels for moustached tamarins, whereas for saddle-backs it

increasedsubstantially

asthey

moved fromperipheral

areas sharedbyother groups to exclusive areas near the range centre. The depression of

capture rates in the range periphery observed in saddle-backs is analogousto that of golden lion tamarins, which are also manipulative predators

relying heavily on embedded prey (PERES, 1989). This indicates that,

although contributing less to territorial defence, saddle-backs appeared to

derive relatively greater benefits from exclusive use of space than did

moustached tamarins. Saddle-back tamarins thus emerge as the net bene-

ficiaryof tamarins'

jointterritorial

system, enjoyinga

protectiveshadow

against resource depletion by other groups, which was provided primarily

by moustached tamarins' greater effort in range defence.

Given that joint-territorial benefits were strongly biased towards onlyone tamarin species, what then accounts for the investments and toler-

ance of the other in maintaining such remarkably stable associations?

Range defence should be considered in the wider context of other adap-tive advantages of heterospecific group living, such as division of labour in

locating different-sized plant-food patches (PERES,1991), foraging bene-fits associated with piracy and facilitated harvest of flushed prey items

(PERES,1992), and detection of different forms of predation threats held in

common by both species (PERES,n press a). In addition, what infrequent,but nevertheless decisive, competitive interactions which occurred

between the two species, such as physical displacements over small and



244 CARLOS A. PERES

spatially restricted food patches, invariably favoured the larger-bodiedand dominant moustached tamarin to the detriment of their smaller

heterospecifics (PERES,1991).In

any case,there is no

evidenceto

suggestthat the energetic costs of interference competition between neighbouringmoustached tamarin groups would have been reduced by the absence of

saddle-backs. Indeed, the costs of attending a mixed-species group

appears to be higher for the smaller-bodied saddle-back tamarin, mainlybecause of its invariably subordinate status, but negligible for moustached

tamarins. Benefits of mixed-species associations to the subordinate spe-cies were, however, in many respects greater than those of moustached

tamarins. Yet, should these two ecologically very similar species co-existat all, their joint-territorial system should be weighed against the no more

attractive alternative of their exploiting an almost identical set of

resources (PERES, n press b) asynchronously by holding overlapping

monospecific spacing systems independently of one another.

Finally, there are cases where a solitary territory holder will accept an

intruder because the feeding costs it imposes are outweighed by the

benefits of increased defence (DAVIES & HOUSTON, 1981). This switch

from solitary to social territoriality may approach the conditions underwhich group territoriality has evolved (BROWN,1964). Joint territoriality,in turn, may result in cases where large monospecific groups will enhance

resource defence by both interference or exploitative mechanisms, but

increases in group size are prevented by ecological or reproductivethresholds (PERES,1991). In such circumstances, joint-territoriality, par-

ticularly between ecologically similar species, may represent the best

solution to maintain a favourable feeding territory.

References

ALTMANN,J. (1974). Observational study of behavior: sampling methods. - Behaviour 49,

p. 227-267.

BROWN, J. L. (1964). The evolution and diversity in avian territorial systems.- Wilson

Bull. 6, p. 160-169.

- & ORIANS,G. H. (1970). Spacing patterns in mobile animals. - Ann. Rev. Ecol.

System. 1, p. 239-262.

BROWN, M. B. (1974). Identification of sources of significance in two-way contingency

tables.-

Appl. Statist. 23, p. 405-413.CHARNOV, E. L., ORIANS, G. H. & HYATT, K. (1976). The ecological implications of

resource depression.- Am. Nat. 110, p. 247-259.

CHIVERS, D.J. (1974). The siamang in Malaya.- S. Karger, Basel.

CODY, M. L. (1971). Finch flocks in the Mohave desert. - Theor. Popul. Biol. 2, p.142-158.

DAVIEs,N. B. & HOUSTON, . I. (1981). Owners and satellites: the economics of territorial

defence in the pied wagtail, Motacilla alba. -J. Anim. Ecol. 50, p. 157-180.




GARBER, . A. (1988). Diet, foraging patterns, and resource defense in a mixed species troopof Saguinus mystax and Saguinus uscicollis in Amazonian Peru. - Behaviour 105, p.18-34.

GAUTIER-HION, ., QURIS,R. & GAUTIER,J.P. (1983). Monospecific vs. polyspecific life: acomparative study of foraging and antipredatory tactics in a community of Cer-

cophitecusmonkeys.- Behav. Ecol. Sociobiol. 12, p. 325-335.

GILL, F. B. (1978). Proximate costs of competition for nectar. - Am. Zool. 18, p. 753-765.

-- & WOLF, L. L. (1975). Economics of feeding territoriality in the golden-wingedsunbird. -

Ecology 56, p. 333-345.

HUBRECHT, R. C. (1985). Home-range and use and territorial behavior in the common

marmoset, Callithrixjacchusjacchus, t Tapacura Field Station, Recife, Brazil. -Int.J.

Primatol. 6, p. 533-550.

KODRIC-BROWN, A. & BROWN,. L. (1978). Influence of economics, interspecific competi-

tion, and sexual dimorphism on territoriality of migrant rufous hummingbirds.

Ecology 59, p. 285-296.MAY, R. M. (1974). On the theory of niche ovelap.

- Theor. Pop. Biol. 5, p. 297-332.

MITANI, . C. & RODMAN, . S. (1979). Territoriality, the relation of ranging pattern and

home range size to defendability, with analysis of territoriality among primate species.- Behav. Ecol. Sociobiol. 5, p. 241-251.

MUNN,C. A. & TERBORGH,J.W. (1979). Multi-species territoriality in neotropical foragingflocks. - Condor 81, p. 338-347.

MYERS, J. P., CONNERS,P. G. & PITELKA,F. A. (1981). Optimal territory size and the

sanderling: Compromises in a variable environment. - In: Foraging behavior: eco-

logical, ethological and psychological approaches (A. C. KAMIL& T. D. SARGENT,

eds), Garland STPM Press, New York, p. 135-158.

NIE, N. H., HULL, C. H.,JENKINS,J. G., STEINBRENNER,K. & BENT, D. H. (1975). SPSS.

Statistical package for the social sciences. - McGraw-Hill, New York.

ORIANS,G. H. & WILLSON,M. F. (1964). Interspecific territories of birds. -Ecology 45, p.

736-745.

PARK,T. (1954). Experimental studies of interspecific competition in two species of Tri-

bolium.- Physiol. Zool. 27, p. 177-238.

PATON,D. C. & CARPENTER,. L. (1984). Peripheral foraging by territorial rufous hum-

mingbirds: defense by exploitation.- Ecology 65, p. 1808-1819.

PERES,C. A. (1986). Costs and benefits of territorial defense in golden lion tamarins,

Leontopithecusosalia. - Unpubl. MSc thesis, University of Florida.-

(1989).Costs and

benefitsof

territorial defensein

wild goldenlion

tamarins,Leon-

topithecus osalia. - Behav. Ecol. Sociobiol. 25, p. 227-233.--

(1991). Ecology of mixed-species groups of tamarins in Amazonian terra firme forests.

Unpubl. PhD thesis, University of Cambridge.-- (1992). Prey-capture benefits in a mixed-species group of Amazonian tamarins,

Saguinusfuscicollisand S. mystax.- Behav. Ecol. Sociobiol. 31, p. 339-347.

-- (in press a). Antipredation benefits in a mixed-species group of Amazonian tamarins.- Folia Primatol.

-- (in press b). Diet and feeding ecology of saddle-back and moustached tamarins in an

Amazonian terra firme forest. -J. Zool., London.

RATHBUN,C. D. (1979). Description and analysis of the arch display in the golden lion

tamarin, Leontopithecusosalia rosalia.-

Folia Primatol. 32, p. 125-148.REED,T. R. (1982). Interspecific territoriality in the chaffinch and great tit on islands and

the mainland of Scotland: playbacks and removal experiments.- Anim. Behav. 30,

p. 171-181.

ROBERTSON, . R., SWEATMAN, . P. A., FLETCHER, . A. & CLELAND,M. G. (1976).

Schooling as a mechanism for circumventing the territoriality of competitors. -

Ecology 57, p. 1208-1220.



246 CARLOS A. PERES

ROBINSON,J.G. (1979). Vocal regulation of use of space by groups of titi monkeys, Callicebus

moloch.- Behav. Ecol. Sociobiol. 5, p. 1-15.

RYLANDS,A. B. (1983). The behavioural ecology of the golden-headed lion tamarin, L.

chrysomelas,and the marmoset C. kuhlii (Callitrichidae, Primates). - Unpubl. PhD

thesis, University of Cambridge.

SIEGEL, S. (1956). Nonparametric statistics for the behavioral sciences. - McGraw-Hill,New York.

SOKAL, R. R. & ROHLF, F. J. (1981). Biometry, (2nded.).-

Freeman, San Francisco.

TERBORGH,J. (1983). Five new world primates: a study in comparative ecology.- Prince-

ton University Press, Princeton.

-- (1984). Cooperation is better than conflict: why saddle-back and emperor tamarins

share a territory.- In: The encyclopedia of mammals (D. MACDONALD,d.). Facts on

File, New York, p. 348-349.

WASER,P. M. (1976). Cercocebusalbigena:Site attachment, avoidance, and intergroup spac-

ing.-

Am. Nat. 110, p. 911-935.- (1977). Individual recognition, intragroup cohesion, and intergroup spacing: evi-

dence from sound playback to forest monkeys.- Behaviour 60, p. 28-74.

-- (1981). Sociality or territorial defense? The influence of resource renewal. - Behav.

Ecol. Sociobiol. 8, p. 231-237.-- & WASER,M. S. Z. (1977). Experimental studies of primate vocalizations: Specializa-

tions for long-distance propagation.- Z. Tierpsychol. 43, p. 239-263.

- & WILEY,R. H. (1979). Mechanisms and evolution of spacing in animals. - In:

Handbook of behavioral neurobiology, Vol. 3 (P. MARLER&J. G. VANDENBERGH,ds).Plenum Publ., New York, p. 159-223.

WHITEHEAD,J. M. (1987). Vocally mediated reciprocity between neighbouring groups of

mantled howling monkeys, Alouattapalliata palliata. - Anim. Behav. 35, p. 1615-1627.

Documents

Consequences of Joint-Territoriality in a Mixed-Species Group of Tamarin Monkeys