Behavioral Responses of Female Long-Tailed Macaques (Macaca

fascicularis) to Pair Formation

Scott W. Line, Kathleen N. Morgan, Hal Markowitz, Jeffrey A. Roberts, and Mike Riddell
California Regional Primate Research Center and San Francisco State University

While most primate species are social in nature, it has been standard practice in many laboratories to house
adult primates in individual cages. Single housing has several advantages from a husbandry standpoint,
including reduced frequency of trauma and disease transmission, improved access to individual animals for
experimental procedures, and increased safety for animal handlers. Considering the animals' perspective,
however, individual housing eliminates the opportunity for many normal behavior patterns.

The 1985 amendment to the Animal Welfare Act has led to renewed interest in group-housing as a means
of maintaining primates in research facilities. While it is not known what form the final regulations that
will be used to enforce the law will take, it is likely that they will emphasize increased use of social
housing. The last public proposal from the Department of Agriculture recommended social housing
whenever possible (USDA, 1989).

Pair-housing is one form of social housing that may be practical for replacing individual housing of
primates in many situations. Monkeys in pairs can perform a large number of social behaviors not available
to those that are singly caged, and expanding the behavioral repetoire is one way of improving primate
welfare (Line, 1987; Novak & Suomi, 1988). Pair-housing may also be accomplished relatively
economically by modifying existing single cages. We recently added side openings to several stainless steel
squeeze-back cages at a cost of approximately $150/cage. The openings are 50 cm wide by 23 cm high,
approximately 20% of the side wall. When the cages are placed next to each other, two monkeys housed in
them can move freely back and forth. If necessary, a solid partition can be locked in place to separate the
monkeys.

Despite the conventional wisdom that unfamiliar adult macaques are more likely to fight than to coexist
peacefully, Reinhardt and co-workers have shown that pair-housing can be successful for adult rhesus
monkeys (Macaca mulatta) that are unfamiliar with each other and have lived in single cages for prolonged
periods. They have paired juvenile rhesus with adults of both sexes (Reinhardt et al., 1987), adult females
(Reinhardt et al., 1988), and adult males (Reinhardt, 1989). Reinhardt (1990b) also reported that unrelated
stump-tailed macaques (M. arctoides) can be successfully pair-housed.

In this experiment we attempted to replicate these findings in another species of primate commonly used in
laboratory research, the long-tailed macaque (M. fascicularis). Beyond determining whether or not
compatible pairs can be formed with unfamiliar individuals, we also wanted to document general changes
in behavior associated with the move from single to pair-housing. Providing monkeys with the opportunity
for social interaction is only one step toward improving their welfare. We also wanted to obtain quantitative
data on whether the monkeys expressed more social behaviors.

Unlike Reinhardt (1990a), we feel that monitoring the occurrence of abnormal behavior is useful in
assessing primate welfare. While some abnormal behaviors have been observed among free-ranging and
group-housed primates, they are much less frequent among socially housed primates than among those that
are singly-housed. Still, there is no general consensus on the meaning and importance of abnormal behavior
patterns. Abnormal behaviors could be viewed as positive coping responses that enable an animal to adapt
to an unnatural environment. On the other hand, Dantzer (1986) argued that environments that require such
coping responses are inadequate. It is our view that reductions in the frequency or duration of abnormal
behaviors are evidence that welfare has improved. Consequently, we also measured abnormal behavior
before and after the move to pair-housing.
Methods

The subjects were 12 adult female long-tailed macaques (mean weight = 3.4 kg, mean estimated age = 8.8
years). Eleven were wild-caught as adults and had lived in single indoor cages in the colony for a mean of
3.5 years (range = 2.0 to 7.0 years). One was born in the colony in a multi-female, multi-male group, and
had been singly housed indoors for 6.4 years.

At the beginning of the experiment all subjects were housed singly in three different colony rooms
containing approximately 50 long-tailed macaques. They were kept in stainless steel squeeze-back cages
0.68 m deep, 0.79 m high, and 1.83 m wide (manufactured by Lab Products/Harford, Maywood, NJ). Each
cage was divided into three identically-sized sections (0.68 x 0.79 x 0.61 m) by removable stainless steel
partitions, and was mounted in a double tier on a wheeled rack. All subjects were housed in upper-row
cages, were fed a commercial diet twice daily, and had continuous access to automatic water dispensers.
Room lights were controlled automatically on a 12:12 light:dark cycle. All animal care procedures were
conducted according to standard laboratory protocols. The Primate Center is AAALAC-accredited.

Frequency and duration of behavior were recorded during 10-minute observation sessions using an
automatic wand and bar-code system with a laptop computer. Frequencies and durations of each behavior
were summarized individually, then combined to establish a general activity budget with the following
behavioral categories: agonistic behavior (hitting, grabbing, biting, presenting, grimacing, threatening, or
shaking the cage directed at other monkeys or the observer), cage manipulation (oral or manual exploration
of any part of the cage), abnormal postures (self-holding, saluting, and other bizarre postures), stereotypic
actions (drinking urine, self-sucking, head tossing, and rocking), stereotypic locomotion (pacing, circling,
swinging, somersaulting, and other repetitive actions with no apparent purpose), self-abuse (self-biting,
self-hitting, and plucking fur), foraging (holding, biting, sniffing, or manipulating pieces of food, and
drinking water), grooming, standing, non-stereotypic locomotion, and resting (passive contact, sleeping,
crouching, lying, and sitting while otherwise inactive). Abnormal postures, stereotypic actions, stereotypic
locomotion, and self-abuse were added together for a composite abnormal behavior duration. In addition,
we recorded the frequency of vocalizations, the time each subject spent in the front half of the cage and,
after pairing, the time pair-mates spent in proximity to one another (i.e. within the same half of the cage).
We also calculated the duration of observer-directed behavior from the amount of agonistic behavior
performed while looking at the observer.

Ten baseline observations were performed on each subject while they were housed singly in their original
cages. Ten additional observations were made on each monkey during the first two weeks after they were
paired according to the protocols described below. A total of 36 hours of data was collected by seven
observers. Inter-rater reliability exceeded 90%. Observations were conducted between 0800 and 1700
hours, primarily between 1000 and 1300. The time of day of observation was matched between the baseline
and paired conditions.

The first eight subjects were housed in Room 1, and were assigned to pairs after their baseline data were
analyzed. We made the assignments primarily on the rate of aggressive behavior observed during this
period. Two monkeys showed much higher rates than any of the others, and these two were assigned as
pair-mates to two that showed no aggressive behavior. The other two pairs were created by designating
monkeys of similar weight and age as pair-mates.

We then attempted to assess compatibility between these subjects prior to pair formation. We placed the
potential pair-mates into open-mesh stainless steel wire transfer boxes measuring 30 cm wide by 36 cm
high by 54 cm long. The boxes were then placed next to each other on the floor of the anteroom of Room 1
for 15 minutes while an observer recorded the behavior of each macaque. Each subject was returned to its
original cage after the test.
These subjects were subsequently paired by moving each monkey into an adjacent section of a new cage
and removing the partition. All eight were moved to a new location in Room 1 in an attempt to reduce
aggressive behavior caused by territoriality.

To test whether or not pair formation could be successfully accomplished with a minimum of labor, we did
not perform the pair compatibility test for the other four subjects. These monkeys were initially housed in
Rooms 2 and 3, and were assigned to pairs randomly. They were subsequently placed in adjacent cage
sections in Room 1, and we paired them by simply removing the partition.

Data were analyzed by either a one-tailed or two-tailed paired t-test, depending on whether or not we
expected a change in a particular direction for each behavior following pair formation. A p-value less than
0.05 was considered significant.

Results

Pair formation was successful for six of the first eight monkeys. In the unsuccessful case one subject
grabbed and bit the other repeatedly during the entire ten minutes they were together. The aggressor was
the animal that showed the highest rate of aggressive behavior in the baseline period. Her partner crouched
in a corner of the cage, vocalizing and grimacing, but this did not appear to deter the attacks. No serious
injuries occurred, but the frequency of attacks did not appear to be declining, and so the subjects were
separated. No aggressive behavior was seen during the initial interactions of the other three pairs.
Additional checks were made several times a day during the next week to assure continued compatibility.

Over the seven days following pair formation, four of the paired subjects sustained minor wounds, none of
which required veterinary treatment. No fighting was directly observed during this time, and no additional
wounds were detected after the first week.

The other four monkeys (which did not experience the pair compatibility test and were not moved prior to
pair formation) were also successfully paired. No injuries or fighting were noted in any of these subjects
during the first two weeks. Seven weeks after pair formation one monkey among these four sustained mild
trauma. While no fighting was directly observed, multiple small lacerations were discovered on her head
and neck. She was separated from her pair-mate and anesthetized to allow closer inspection and cleaning of
the wounds. The two were kept apart overnight, and re-paired the next morning by removing the partition.
They immediately moved together and groomed each other; no further trauma has been detected.

The five successful pairs have remained together for five to six months, and are still compatible. All
subjects in these pairs were observed to groom each other and show other affiliative behaviors within the
first ten minutes after pair formation. In two cases the monkeys moved together and grasped each other in a
ventral-ventral embrace immediately after the partition was removed.

Figure 1: Behavioral changes of 10 adult female long-tailed macaques following pair formation. Each bar
represents the mean (+SEM) % time engaged in a behavior while singly caged (open bars) and while paired
(shaded bars). P-values represent significance levels for paired t-tests.

A summary of the mean activity budget of the ten compatible pair-mates before and after pairing is
presented in Figure 1. The total time spent grooming increased significantly after pairing (one-tailed t-test,
p = 0.0097). Most of the grooming observed in the pair condition was allogrooming. The time spent
autogrooming decreased significantly from 24% to 10% after pairing (one-tailed t-test, p = 0.0053). The
amount of time devoted to rest and abnormal behavior also decreased significantly after pairing (one-tailed
t-tests, p = 0.0353 and p = 0.0028, respectively). Foraging also decreased significantly after pairing (two-
tailed t-test, p = 0.0249). No other activity budget categories changed significantly after pair formation.
Among the other behaviors recorded, time spent in observer-directed activity declined significantly from
2.9% to 0.7% (one-tailed t-test, p = 0.0333), vocalization frequency declined from 45.5/hour to 31.3/hour
(two-tailed t-test, p = 0.0241), and time spent in the front half of the cage decreased significantly from
75.5% to 40.5% (two-tailed t-test, p = 0.0216). Pair-mates spent 63.1% of the observation time in proximity
to one another.

Behavior Singly caged Pair-housed P-value pace 20,85 (14.4) 9.31 (6.96) 0.2419
abnormal posture 11.65 (8.32) 1.56 (1.01) 0.1281 stereotypical action 39.29 (16.92) 4.0 (2.25)
0.0221 self-abuse 4.48 (4.06) 0.0 (0) 0.1493 total 76.24 (21.71) 14.79 (9.12) 0.0028

Table 1: Influence of housing condition on abnormal behaviors among 10 adult female long-tailed
macaques. Each figure is the mean (+SEM) duration in seconds per 10 minute observation period. P-values
represent the significance levels for a comparison of the two conditions with a 1-tailed t-test.

The mean duration per 10 minutes of each of the four categories of abnormal behavior in singly-caged and
pair-housing conditions appear in Table 1. All four categories decreased in duration after the move to pair-
housing.

Discussion

The major findings of this experiment were: 1) pair-housing of unfamiliar adult female long-tailed
macaques was successful for five of six pairs, and 2) affiliative behaviors occurred frequently, while
abnormal behaviors decreased significantly among pair-housed subjects. Since the pairs have only been
housed together for six months or less, and the follow-up observations were performed in the first two
weeks after pair formation, these conclusions should be considered preliminary. While there is a continued
risk of fighting, we have seen no evidence of serious incompatibility between pair-mates thus far. We are
continuing to collect data to see if the behavioral changes observed are maintained beyond the initial period
of pair formation.

It is possible that the motivational systems involved in the expression of the different categories of
abnormal behavior are varied, and that these behaviors should be considered separately, rather than as a
composite response to the change in the social environment. While there was a decrease in the mean
duration of all four categories of abnormal behavior, only the decrease in stereotypic action was statistically
significant. There were decreases in the duration of each abnormal behavior for the majority of subjects,
however. A larger sample size may have resulted in more of the changes reaching statistical significance.
For example, self-abusive behaviors were recorded for five of the ten subjects when singly housed, but
were completely absent after pair formation.

We found that the minimal-labor protocol, in which monkeys were randomly assigned to one another and
paired by simply removing the partition between cage sections, was just as successful as the more detailed
procedure used for the first four pair attempts. We have subsequently created two additional compatible
pairs using the minimal-labor approach. Reinhardt (1990b) also reported success using a similar pairing
technique with stump-tailed macaques.

We did not predict a change in the amount of time foraging after pair formation. Examination of the
components of foraging revealed that the decline was entirely due to a decrease in the amount of time spent
drinking. All ten subjects showed a decline in drinking. One potential explanation is that the amount of
time drinking in the singly caged condition was unusually high. Although none of these subjects showed
clinical signs of polydipsia, psychogenic polydipsia has been documented in singly caged rhesus macaques
in our colony (Bicknese et al., 1989). It is possible that stress associated with single caging elicited
excessive water consumption, and that this pattern was changed by the move to pair-housing.

All but one of the subjects were wild-caught, so we cannot be completely certain whether or not all pairs
were unrelated, and unfamiliar with each other. We do know that they were not housed together while in
our colony (a period ranging from two to seven years). One pair was comprised of a wild-caught and a
colony-born animal. Two other pairs included wild-caught monkeys obtained from different suppliers 29
and 49 months apart. In the final two pairs both individuals were obtained from the same supplier, but at
times 8 months and 15 months apart from one another. Given this background, we feel it is reasonable to
assume they did not know each other prior to this experiment.

An alternative explanation for the changes in behavior following pair formation is that the subjects became
habituated to the presence of observers. This might explain the decreases in abnormal behaviors, resting,
observer-directed activity, and frequency of vocalizations. If this were the case, however, we would expect
to see a gradual decline in each of these behaviors over time, a pattern that did not occur. There were no
decreasing trends in mean level of each behavior over the course of the experiment, and repeated-measures
analyses of variance on each behavior within the baseline period were nonsignificant. It is more likely that
time interacting with the cage-mate replaced time devoted to these activities.

Pair-housing represents a compromise between keeping primates in large groups and keeping them in
individual cages. It increases some risks to the animals and personnel, including both risk of trauma and
disease transmission. It may also inconvenience some experimental protocols. Overall, however, it provides
substantial benefits to the animals. In a comparison of pair-mates, perches, and pieces of wood, Reinhardt
(1990c) concluded that pair-mates provided the most effective long-term stimulation for adult rhesus
macaques. While the absolute cost of increased use of pair-housing will not be small, it is likely to be one
of the least expensive and most effective alternatives for improving the welfare of nonhuman primates in
research facilities.

References

Bicknese, E. J., Eisele, P. H., & George, J. W. (1989). Psychogenic polydipsia in individually housed adult
rhesus macaques (Macaca mulatta). Laboratory Animal Science, 39, 476-477.

Dantzer, R. (1986). Behavioral, physiological and functional aspects of stereotyped behavior: A review and
a re-interpretation. Journal of Animal Science, 62, 1776-1786.

Line, S. W. (1987). Environmental enrichment for laboratory primates. Journal of the American Veterinary
Medical Association, 190, 854-859.

Novak, M. A., & Suomi, S. J. (1988). Psychological well-being of primates in captivity. American
Psychologist, 43, 765-773.

Reinhardt, V. (1989). Behavioral responses of unrelated adult male rhesus monkeys familiarized and paired
for the purpose of environmental enrichment. American Journal of Primatology, 17, 243-248.

Reinhardt, V. (1990a). Evaluating the effectiveness of environmental enrichment. Laboratory Primate

Reinhardt, V. (1990b). Environmental enrichment program for caged stump-tailed macaques (Macaca
arctoides). Laboratory Primate Newsletter, 29 [2], 10-11.

Reinhardt, V. (1990c). Time budget of caged rhesus monkeys exposed to a companion, a PVC perch, and a
piece of wood for an extended time. American Journal of Primatology, 20, 51-56.

Reinhardt, V., Houser, W. D., Eisele, S. G., & Champoux, M. (1987). Social enrichment of the
environment with infants for singly caged adult rhesus monkeys. Zoo Biology, 6, 365-371.
Reinhardt, V., Houser, W. D., Eisele, S. G., Cowley, D., & Vertein, R. (1988). Behavioral responses of
unrelated rhesus monkey females paired for the purpose of environmental enrichment. American Journal of
Primatology, 14, 135-140.

U. S. Department of Agriculture (1989). Animal welfare proposed rules. Federal Register, 54 [49], 10822-
10954.

-------------------------------------------------------------------

First author's address: California Regional Primate Research Center, Univ. of California, Davis, CA 95617-
8542.
This work was supported by NIH grant RR00169-28 to the CRPRC. The authors thank Carmel Stanko,
Julie Ferris, Astrid Paletzki, and Kevin Hall for assistance in data collection.