Content uploaded by Dorothy Munkenbeck Fragaszy
Author content
All content in this area was uploaded by Dorothy Munkenbeck Fragaszy on Mar 06, 2015
Content may be subject to copyright.
Determining the Value of Social
Companionship to Captive Tufted
Capuchin Monkeys (Cebus apella)
Elizabeth Dettmer and Dorothy Fragaszy
Department of Psychology
University of Georgia
This study used a method similar to one introduced by Dawkins in 1990 to assess the
magnitude of the psychological need for social companionship in pair-housed tufted
capuchin monkeys (Cebus apella). This method permits classification of commodi-
ties as necessities or luxuries. The study directly compared the commodity of social
companionship to the commodity of food, a known physiological necessity, in a se-
ries of preference tests following commodity deprivations. The majority of subjects
chose their social companion over food at baseline and persisted in this preference
evenafterseveralhoursoffooddeprivation.Inaddition,subjects’preferencesshifted
from 1 commodity to the other with manipulation of social and food deprivationlev-
els.Capuchinmonkeys perceived social companionshipasa necessity at alevelsimi
-
lar to that of food.
Recent animal use legislation (Animal Welfare Act, 1985, 1991) in the United
States requires that those responsible for the care of captive nonhuman primates
provide for their psychological well-being, including provisions to meet social
needs. Whereas physical well-being can be assessed as the absence of pathology
and on the basis of physical condition or physiological measures, the definition
of psychological well-being is more problematic. Although many parameters for
enrichment have been stipulated in regulations, no consensus on how to measure
psychological well-being presently exists (Committee on the Well-Being of
Nonhuman Primates, 1998; Dawkins, 1990; Novak & Suomi, 1988; Poole,
1992). Identification of basic needs might be a less elusive research goal than
identification of psychological well-being. If we can determine how to meet
JOURNAL OF APPLIED ANIMAL WELFARE SCIENCE, 3(4), 293–304
Copyright © 2000, Lawrence Erlbaum Associates, Inc.
Requests for reprints should be sent to Elizabeth Dettmer, Boston VAMC (116B), 150 South Hun
-
tington Avenue, Boston, MA 02130. E-mail: dettmer.elizabeth@boston.va.gov
needs, doing so will support physical and psychological well-being. We suggest
that a shift from the identification of psychological well-being to the identifica
-
tion of basic needs might bring us closer to current welfare goals.
How does one define basic needs? Maslow (1968) introduced a theory of basic
needs for Homo sapiens. He placed physiological needs such as hunger, thirst,
sleep, rest, exercise, and elimination at the base of the hierarchy. Individuals, he
postulated, must first meet these physiological needs before attending to needs of
safety, the next step in the hierarchy, or social needs, still another step higher.
Maslow’s hierarchy continues to levels of introspection that include self-esteem
needs and self-actualization and asserts thatwell individuals can fulfill all of these
needs—from nourishment to self-actualization: “If the essential core of the person
is denied or suppressed, he gets sick, sometimes in obvious ways, sometimes in
subtle ways, sometimes immediately, sometimes later” (p. 4). In free choice situa
-
tions, the need-deprived person gives basic needs priority over other satisfactions
(Maslow, 1968). This theory supports the use of choice tests to evaluate whether a
particular commodity fulfills a need.
Animal researchers have employed preference tests similar in theory to that
which Maslow (1968) termed choice situations. However, many of these tests
have been too simplistic in nature. An assumption, originating with Spencer
(1880), was made that animals experience subjective feelings of pleasure when
given access to reward or positive reinforcers. Therefore, many researchers have
presented subjects with a choice between two different environments, assuming
the animals would spend more time in the particular environments that they found
more reinforcing. Dawkins (1981) used such a method to demonstrate that bat-
tery-kept hens prefer larger cages to smaller ones and litter floors to wire floors.
When the two commodities were pitted against one another (the animal had to en
-
ter a smaller cage to gain access to litter or a wire cage to gain the larger area), litter
was given higher priority. According to Dawkins, battery hens “preferred the
smaller cage when this had a litter floor to the one four times larger with a wire
floor” (p. 255).
Duncan (1978) criticized such preference experiments for providing only rela
-
tive conclusions: “The fact thatA is always chosen in preferenceto B tells us noth
-
ing of the absolute properties of A and B” (p. 198). In addition, many choice
experiments are based on the assumption that
in the natural environment, proximate and ultimate needs will generally go hand in
hand. For example, animals that experience a proximate need for food and start
searching for it when their reserves are low will be fulfilling the ultimate need of
avoiding death by starvation. (Dawkins, 1983, p. 1197).
Duncan argued that animals in laboratory settings often choose according to
short-term gains that do not correspond to long term gains. In view of such criti
-
294
DETTMER AND FRAGASZY
cisms,Dawkins(1990)suggestedanoperanttrainingtechniquethatprovidesanob
-
jective wayfordeterminingwhatanimalsareexperiencing.Theparadigmrequires
subjects to demonstrate motivation toobtain certain rewards through operanttasks
for different commodities. An animal’s willingness to work for eachcommodity is
compared.Willingness toworkfor foodis usedtoanchor therelative measurement
allowing assignment of qualitative worth. If animals show that they regard being
without some commodity as aversive as being without food (a known basic need),
then one could argue that the animals suffer (as they surely would suffer without
food)withoutthatcommodity.Inotherwords,theneedforfoodisusedasayardstick
against which other less recognized needs are measured.
Demonstration of such commodity demand curves (demonstration of how hard
an animal will work for a commodity) requires a great deal of time. Researchers
have voiced concern about the large number of experiments that would need to be
done to plot demand functions of different animals, conditions, sexes, times of
year, and so on (Dawkins, 1990). In addition, the implementation of such a plan is
somewhat problematic in that operant responses such as pressing a lever are asso-
ciated preferentially with different stimuli. Animals seem to be prepared to make
certain associations and less prepared to make others. For example, Garcia and
Koelling (1966) showed long ago that rats readily associate an auditory–visual
stimulus with shock punishment but do not associate a taste stimulus with shock.
Such differential preparedness can alter performance and therefore complicate the
comparison of an animal’s motivation for different commodities with the use of
the same operant task.
This design, a simplification of Dawkins’s (1983, 1990) method, investigated
whether social companionship is a basic need for capuchin monkeys (Cebus
apella). Instead of comparing commodity demand curves, we directly compared
individuals’ needs for social companionship with the need for food.
METHODS
Design
Figure 1 illustrates the experimental design. Subjects were tested in a baseline
condition and two test phases. Baseline trials required subjects to choose be
-
tween food and social companionship without prior deprivation. Following de
-
termination of baseline preferences, subjects next were exposed to food depriva
-
tion trials (Phase 1, right arm of Figure 1). None of our subjects exhibited a
preference for food in baseline testing; therefore, the left side of Figure 1 was
not implemented in this study.
Phase 1 determined the duration of food deprivation required to produce a pref
-
erence for food.If the subject failed to choose food inat least 9 of 11 trials, the du
-
SOCIAL NEEDS IN CAPUCHINS 295
ration of food deprivation was increased and a new set of choice trials presented to
the subject. Once subjects switched to a food preference or reached 22 hr of food
deprivation, they moved to Phase 2. On advisement of the attending veterinarian,
no animal experienced more than 22 hr of food deprivation.
Phase 2 assessed the subjects’ need for socialcompanionshipin relation to their
need for food. Subjects were deprived of food for thedurationdetermined in Phase
1 to produce a preference for food.Concurrently, they also were deprived of social
companionship. The duration of social deprivation was increased until the subject
expressed a preference (9of 11 trials) for the social companion or until 24 hr of so
-
cial deprivation was implemented.
296
DETTMER AND FRAGASZY
FIGURE1 Thedesignof theexperiment.Subjectswere assignedtotheleft ortheright pathof
testingafter weevaluated theirpreference forfood orthe pairmatein achoice setting,under nor
-
mal conditions. In Phase 1, subjects experienced deprivation of the commodity they had ex
-
pressed preference for in the baseline condition. Testing in this phase was repeated with
increasing levels of deprivation until the subject switched its preference to the initially less pre
-
ferred commodity. Once its preference had shifted, the subject entered Phase 2. In thisphase of
testing,the subjectexperienceddeprivation ofboth commodities.The durationof deprivationof
the initially preferred commodity was increased until the subject reverted to its original prefer
-
ence.As no subject expressed apreference for food overthe social partner inthe baseline condi
-
tion, all subjects followed the path outlined on the right side of the figure.
Subjects
Seven capuchin monkeys, six males and one female, served as subjects. At the
onset of the procedures, these animals ranged in age from 19 to 88 months. The
subjects had lived as pairs in indoor pens (1.99 m × 1.07 m × 1.19 m) for at least
2 months prior to testing. All subjects but one were reared for at least 2 years in
species-typical groups. The only female subject, age 33 months, was paired with
the youngest male, 21 months. Both subjects were well below the age of sexual
maturity, estimated as 48 months (Fragaszy & Adams-Curtis, 1998). The ani
-
mals were maintained on Purina Monkey Chow 5045 (Richmond, IN) and daily
rations of fruit and vitamin supplements. Food was available ad libitum except
during testing periods. The light–dark cycle was set at 13:11 hr, with onset of
light at 6:30 a.m.
Materials
The testing apparatus was composed of three adjoining metal mesh chambers,
each 45.7 cm × 40.6 cm × 61.0 cm. The chambers were suspended 1.31 m from
the ground via a frame. Transparent doors provided access from the middle cage
SOCIAL NEEDS IN CAPUCHINS 297
FIGURE 2 The testing apparatus. (a) The subject sat in the center compartment of the three-
compartment apparatus. Transparent siding and doorsallowed the subject toview but nottouch
items in the outer two cages. (b) The monkey gained access to one of the adjoining cages by
pressing the appropriate knob, which caused displacement of a peg and allowed the separating
doorto drop. Theexperimenter then closedthe door tothe middle compartment,effectively per
-
mitting the subject a single choice.
to each of the outer cages (see Figure 2a). Each door was opened by a black
knob located in the middle cage to the right of the door. Pushing with a pressure
equivalent to 5 lb (2.3 kg) on either knob caused the corresponding door to drop
(see Figure 2b). Exertion of 5 lb of pressure was effortful but not difficult for the
subject to achieve.
Procedure
Training.
Subjects were placed individually into the testing apparatus for
15–30 min each session and hand fed. Agitation (vocalizations and intense watch
-
ingbehavior)droppedrapidlyafterapproximately11days.Throughaseriesofsuc
-
cessive approximations, the subjects were trained to operate the choice doors by
pushing on the appropriate knob. Subjects required from 7 to 17 (average = 12)
training sessions to master control of the doors.
The subjects then were presented with two types of choice trials. In the first
(Type A), more of one food item (fruits, nuts, and yogurt) was placed on one side
of the testingapparatusandlessofthesameitemontheother.Sideof placement of
the larger amount of food was randomized. Access doors were locked. Subjects
were released into the middle compartment where they could view the choices in
each adjoining compartment through Plexiglas doors. After 20 sec, a beeper sig-
naled the experimenter to unlock both choice doors and say the word “okay.” This
signaled the subject to make a choice. Pushing the knob to open a door provided
access to the corresponding compartment and any food item contained in it. The
experimenter immediately locked the opposite compartment. The subject was re-
quired to remain with the choice for 5 min, at which time the subject was removed
viaatransferbox.The animal remained in the box as the testing apparatuswaspre
-
pared for the next trial. Subjects underwent four trials of this type per training ses
-
sion. Subjects were assumed to have gained an understanding of the choice
situation when in 100% of trials in two concurrent testing sessions (a total of 8 tri
-
als) they chose the larger over the smaller amounts of food in Type A trials.
In one trial per session, the subjects were presented with the second type of
choice trial (Type B), a choice between two different food items instead of differ
-
ent amounts of the same food item. This procedure was implemented to give the
animals experience with the consequence of choice when choosing between two
different items.
Testing.
The first two trials of each testing session were designed to confirm
that subjects were willing to go to either side of the apparatus. The subject was pre
-
sented with a choice between 1 tsp of flavored sugar water and nothing. The place
-
mentofthe flavored sugar waterwasalternatedfrom side to side.Animalscontinued
with testing if on each of these trials they chose the side containing the sugar water
over the empty side. Only once did a subject refuse to choose both doors.
298
DETTMER AND FRAGASZY
The two sugar water trials were followed by one test trial in which the subjects
were presented with a choice between the home-cage partnerandtwo Purina Mon
-
key Chow biscuits (the basis of their normal diet). The side placement of these
stimuli was systematically varied using the Fellows (1967) testing sequence. The
subjects’ choice was followed by a 5-min waiting period where they remained in
the testing apparatus with the chosen object, the biscuits or social companion. If
the subject chose the social companion, the biscuits immediately were removed
fromtheroom. If thesubjectchose the biscuits, thesocialcompanion was removed
from the testing apparatus and placed on the floor in the same room. After 5 min,
the subjects were transferred to the home cage, where they had access only to the
chosen stimulus. If the subject chose food in the test trial, the home cage was sup
-
plied with food biscuits but the social partner remained outside the cage (although
still in view) for 5 min. If, instead, the subject chose the social partner during the
test trial, both subject and social partner were returned to the home cage together
but waited 5 min for access to food.
Test sessions were conducted in baseline, Phase 1, and Phase 2 conditions. In
the baseline, subjects were tested in the choice apparatuswith no prior deprivation
of food or social companionship for 11 trials (one per day). In Phase 1, the animals
were deprived of food initially for 3 hrbefore eachtrial.If the subject chose the so-
cial companion 3 times at that level of deprivation over an 11 trial sequence, the
subject’s next trial would be the first of another session of trials following a longer
period of food deprivation. Subjects were exposed to a different sequence of depri-
vation hours, dependent on their preceding performance. For example, Be began
Phase 1 trials following 3 hr of food deprivation. On 3 days, she chose her social
companion. She then wasdeprived of food for 5 hr and tested againuntil she chose
her social companion three times. She was then subjected to yet a higher depriva
-
tion level (7 hr). Subjects continued in this manner until they chose food on 9 of 11
SOCIAL NEEDS IN CAPUCHINS 299
TABLE 1
Testing Specifics
Subject Baseline Trials
Food
Deprivation
Trials
Range of Food
Deprivation
Hours
Food and
Social
Deprivation
Trials
Range of
Social
Deprivation
Hours
Be 11 (11) 38 (24) 3–17 (3–5) 21 (13) 17–24 (22–24)
Ch 11 30 3–13 22 22–23
Jo 11 20 3–15 11 17
Qu 11 27 3–13 32 22–24
Xa 11 61 3–22 11 22
Xe 11 41 3–18 11 22
Wi 11 44 13–19 37 22–24
Note. Be was tested twice. The values in parentheses are from her second testing sequence.
trials or reached 22 hr of food deprivation. See Table 1 forthe rangeof hoursof de
-
privation and the total number of trials each individual completed.
In Phase 2 trials, the level of food deprivation required to reach criterion in
Phase 1 was continued. Additionally,subjects were deprived of socialcompanion
-
ship for 17–22 hr. If the subjects did not chose their social companion on 9 of 11
trials, following the initial period of social deprivation, it was increased by 1 hr
(see Table 1). As soon as a significant preference for the social companion was
demonstrated or the animal failed to display a preference after 24 hr of social de
-
privation, testing was terminated. In Phases 1 and 2, subjects were tested on 5 con
-
secutive days, then given 2 days on their standard feeding schedule and without
social separation. Testing occurred over a 15-week period.
Welfare Implications
Subjects in the study were deprived of food for up to 22 hr. The attending veteri
-
narian advised that 22 hr of food deprivation was not life threatening and would
have no long-lasting effects on the animals. We monitored the subjects carefully
when feeding them after periods of food deprivation to make certain they did not
eat too quickly, which could contribute to bloating and stomach problems.
Subjects also were deprived of their social companions for up to 24 hr at a time.
Past research has shown that primates subjected to social separations are likely to
experience an increase in cortisol (Lyons et al., 1999). This physiological response
isa normal reaction ofthe body that facilitatescoping with stress andis expected to
occur in any new situation characterizedby uncertainty. Short-term stressors, such
as separation of juvenile rhesus from their natal social group and resulting in-
creased cortisol have been correlated to decrements in some immune parameters
(Gordonetal., 1992). However, it stillisnot clear if such changesoccurinall sepa
-
rations and how such specific changes, especially short-term changes, might affect
overall health. Chronic elevations of cortisol have been linked more clearly than
short-term elevations with detrimental health effects in past research. The separa
-
tions for this study were relatively brief (less than 24 hr). All subjects remained in
good health and exhibited normal behaviors during and after the study.
RESULTS
Baseline
Five of the seven subjects displayed a significant preference for their social
companion in baseline sessions, and all seven chose their social companions
more often than food in this phase (binomial probability of .008). Three subjects
chose their social companion on all 11 baseline trials (see Table 2).
300
DETTMER AND FRAGASZY
Phase 1
The duration of food deprivation required to produce a preference for food over
social companionship was 12–19 hr for six of the seven subjects. The seventh
animal (Xa), following 22 hr of food deprivation, still chose his social compan-
ion more often than food (7 out of 11 trials). The two animals who failed to
demonstrate a statistically significant preference for their social partners at base-
line required 19 and 12 hr of food deprivation before they exhibited a preference
for food.
Phase 2
Of the five animals who displayed a social preference at baseline, four again dis
-
playedapreferencefortheircompanioninPhase2whentheyweredeprivedbothof
food (12–22 hr) and social companionship (17–23 hr). The fifth animal, Be, never
exhibited a significant preference for social companionship in this phase. When24
hr of social deprivation was added to the 17 hr of food deprivation, she chose ran
-
domly between the two commodities. The two subjects who showedno significant
preferenceinbaselinechoserandomly betweenthetwocommoditieswhen 24hrof
social deprivation was coupled with food deprivation.
In sum, four different patterns of response were demonstrated:
1. Three animals displayedabaseline preference for social companionship fol
-
lowedbyafoodpreferenceafter12,15,or18hr offood deprivationonly, thena so
-
SOCIAL NEEDS IN CAPUCHINS 301
TABLE 2
Summary Table of the Results of Each Subject
Subject
Baseline: Social
Choices Out of
11 Trials
Did Animal
Display a
Baseline
Preference for
Social
Companion?
Hours of Food
Deprivation
Required to
Create Food
Preference
Did Animal
Switch Back to
Social
Preference With
Addition of
Social
Deprivation?
Social
Deprivation
Hours Required
to Shift
Preference Back
to Social
Be 11 yes 17 (5) no N/A
Ch 11 yes 12 yes 23
Jo 10 yes 15 yes 17
Qu 7 no 12 no N/A
Xa 9 yes > 22
a
yes 22
Xe 11 yes 18 yes 22
Wi 8 no 19 no N/A
a
Xafailedtoreachcriterionforfoodpreferenceevenat22hroffooddeprivation,choosingfoodononlyfourtrials.
cialpreferencefollowingfooddeprivation of12,15,or18 hrandsocialdeprivation
of 23, 17, or 22 hr (social–food–social).
2. One animal showed a baseline preference for social companionship and no
preferences after 22 hr of food deprivation. When social deprivation (22 hr) was
added to 22 hr of food deprivation, he switched back to his preference for social
companionship (social–neutral–social).
3. Two animals showed no baseline preference, a food preference after 12 or
19 hr offood deprivation (but no socialdeprivation), and no preference whenboth
food and socially deprived (neutral–food–neutral).
4. One animal showed a baseline preference for her social companion, a food
preferenceafter 17 hroffood deprivation, butnopreference when deprivedofboth
food (17 hr) and social companion (24 hr; social–food–neutral).
DISCUSSION
To measure the need for social companionship, subjects were asked—following
a host of commodity deprivations—to choose between two commodities: food
and social companionship. The only time subjects showed a food preference was
when they were provided with a social companion but deprived of food for at
least 12 hr prior to testing trials. None of the subjects showed an initial food
preference over social companionship, and four of the seven subjects chose the
social partner significantly more often than food, even following lengthy periods
(22 hr) of food deprivation.
These results show that social companionship and food are valued similarly by
tufted capuchins. This was demonstrated by the manipulation ofdeprivation levels
and the resulting preference shifts that indicate the “need” status of the two com
-
modities as very similar. Because capuchins treat social companionship as if it
were at the same need level as food, we propose that these monkeys can be consid
-
ered psychologically well only if they maintain their access to their companions.
Having said that, we also point out that our resultssay nothing of the scalar val
-
uesof the tested commodities(23 hr of socialdeprivation is notclaimedto be equal
to 17 hr of food deprivation). No parametric comparison of the amounts of food
and social companionship deprivation can be made. Moreover, the extent towhich
these findings may be generalized across species, ages, sexes, and housing condi
-
tions is unknown. For example, these subjects had a history of social housing. It
may be that animals housed singly from an early age would perform very differ
-
ently in this paradigm. As a second example, the quality of the relationship be
-
tween the subject animal and the social companion may be an important variable.
In this study, the subject’s familiar cage mate was used as the social companion. It
may be that only familiar individuals fulfill the basic need for companionship.
302
DETTMER AND FRAGASZY
CONCLUSIONS
Despite the caveats arising from our particular sample of subjects and our inabil
-
ity to address the scalar values for different needs, the results are clearly inter
-
pretable. Tufted capuchin monkeys value social companionship as they value
food: It is a necessity, not a luxury. This paradigm creates a conservative esti
-
mate of what animals deem as essential to their well-being as the physiological
necessity of food. It avoids time-consuming creation of demand curves and asso
-
ciated problems of equating different operant tasks that must be specific to each
commodity. Instead, commodities can be classified in one of two categories: ba
-
sic needs and luxuries, where no level of deprivation will offset deprivation of a
basic need. Future research using this paradigm could investigate relationships
or specific inanimate features of the environment such as those often targeted for
“enrichment.” This method can provide convergent validation for claims that
various alterations of housing or handling improve the psychological well-being
of animals. It can show that animals regard some object, environment, or event
as necessary to their well-being.
REFERENCES
Animal Welfare Act Amendments of 1985, Pub. L. No. 99–198 (1985).
Animal Welfare Act Amendments of 1991, 3.81(a) Fed. Reg. 56, No. 32, p. 6499 (1991).
Committee on the Well-Being of NonhumanPrimates. (1998). The psychologicalwell-being of nonhu-
man primates. Washington, DC: National Academy Press.
Dawkins, M. D. (1981). Priorities in the cage size and flooring preferences of domestic hens. British
Poultry Science, 22, 255–263.
Dawkins, M. D. (1983).Battery hens nametheir price:Consumer demandtheory andthe measurement
of ethological “needs.” Animal Behavior, 31, 1195–1205.
Dawkins, M. D. (1990). From an animal’s point of view: Motivation, fitness, and animal welfare. Be
-
havioral and Brain Sciences, 13, 1–9.
Duncan, I. J. H. (1978). The interpretation of preferencetests in animal behaviour. Applied Animal Be
-
haviour, 4, 197–200.
Fellows, B. J. (1967). Chance stimulus sequences for discrimination tasks. Psychological Bulletin, 67,
87–92.
Fragaszy, D., & Adams-Curtis, L.(1998). Growth andreproduction incaptive tufted capuchins(Cebus
apella). American Journal of Primatology, 44, 197–203.
Garcia, J., & Koelling, R. (1966). Relation of cue to consequence in avoidance learning. Psychonomic
Science, 4, 123–124.
Gordon, T. P., Gust, D. A., Wilson, M. E., Ahmed-Ansari, A., Brodie, A. R., & McClure, H. M. (1992).
Social separation and reunion affect immune system in juvenile rhesus monkeys. Physiology and
Behavior, 51, 467–472.
Lyons,D.M.,Wang,O.J.,Lindley,S.E., Levine,S., Kalin,N. H.,& Schatzber,A. F.(1999). Separation
induced changes in squirrel monkey hypothalamic–pituitary–adrenal physiology resemble aspects
of hypercortisolism in humans. Psychoneuroendocrinology, 24, 131–142.
Maslow, A. (1968). Toward a psychology of being (2nd ed.). New York: Van Nostrand Reinhold.
Novak, M. A., &Suomi, S.J. (1988).Psychological wellbeing ofprimates incaptivity. AmericanPsy
-
chologist, 43, 765–773.
SOCIAL NEEDS IN CAPUCHINS
303
Poole,T.(1992).Natureandevolutionofbehavioralneedsinmammals.AnimalWelfare,1,203–220.
Spencer, H. (1880). Principles of psychology (3rd ed.). London: Longman.
304 DETTMER AND FRAGASZY
