It is agreed by most behavioral investigators that a conspecific cage-mate can be a dynamic and enriching stimulus to the environment (Novak & Suomi 1988; Novak & Drewsen 1988; Stanton et al. 1985; Simpson 1984; Kraemer & McKinney 1979; Chamove et al. 1973; IPS 1988). It is recommended that social housing be considered one appropriate means of providing enrichment in cases where the ICD-ACUC approved animal study proposal does not prohibit it. Both pair and group housing conditions are categorized as social housing. It is recognized that group housing may limit access to the subject animals in many protocols, therefore, it is recommended that pair housing be considered for those studies requiring frequent investigator/subject contact.

This social housing may be full-time or part-time, again as the protocol permits. It should be noted that the formation of social units of animals should be done with care. Part- time pair housing may be applicable for protocols requiring food or water restriction or for studies which require routine removal of the animal from the home cage. The determination of the animal composition of the social housing unit should be flexible as recent data indicate that successful social housing can be accomplished with mixed or like ages (e.g., mother/infant dyads, peer groups, mixed age/unrelated groups or extended family units), and with mixed or same sex (Reinhardt et al. 1987a; Reinhardt et al. 1987b; Reinhardt 1989; Snowden 1985). As maternal and peer separation studies have clearly shown detrimental effects on nonhuman primate species-typical behaviors and reproductive capacity, it is recommended, unless the approved protocol requires otherwise, that infant nonhuman primates should be housed with the mother and/or with peers until a species- appropriate weaning time is reached (Ruppenthal et al. 1976; Suomi et al. 1976; Suomi et al. 1973; Suomi et al. 1971; Suomi et al. 1970).

Partial social contact may also occur outside of the home cage in a designated exercise area through which the animals are rotated. This exercise area may encompass an entire room in the animal facility or be a large pen in a separate room or in the middle of an occupied animal room (thereby providing social stimulation to the animals remaining in their home cages). In the case of exercise pens, the exercise area would be equipped with devices and "toys" to increase the activity and interest levels of the occupant(s). If a room is dedicated to this purpose, multiple pens could be placed in the room, thus increasing the frequency of rotation through the exercise area. As preliminary evidence (O'Neill 1988) indicates that different aged animals of some species (e.g., rhesus monkeys) have significantly different preferences for various types of enrichment (e.g., swinging apparatus for juveniles versus floor toys for adults), a multitude of strategies should be present in each pen, or different pens should be designed for various age groups. Special consideration should be given to the introduction of nonfamiliar animals in the exercise area, and provision should be made for only familiar pairs or groups of animals to be in the exercise area at the same time. Also, training of the animals to a transport cage may facilitate both introduction to and removal of the animals from the exercise area. Training programs have even been proposed for reducing stress levels in laboratory nonhuman primates (Layng & DauDelin 1988). Provision for the increased time and personnel needed for these methods of social contact must be made.

The time allotted to animals for socialization and/or exercise may occur between studies of the same protocol, or between protocols for the same investigator, or while animals are awaiting assignment to a different investigator. Rehabilitation through exposure to other animals can result in various degrees of social recovery (Huebner & King 1984; Dienske et al. 1980; Fritz & Fritz 1979; Novak 1979; Suomi et al. 1976; Novak & Harlow 1975; Suomi & Harlow 1972; Harlow et al. 1971; Harlow & Suomi 1971; Suomi 1973). Pairs of animals designated for socialization/exercise could be identified and then introduced to each other in a neutral environment. Again, it must be emphasized that caution must be exercised when introducing unfamiliar animals. Incidents of aggression can occur even between familiar animals, resulting in varying degrees of harm to the monkeys involved. The dyad could be composed of two naive animals or composed of one socially experienced animal and one naive animal. Alternatively, a group of naive animals could be formed and introduced to the neutral environment.

A serious consideration in social housing of nonhuman primates is the induction of separation anxiety. This phenomenon results when animals which have bonded together (mother/infant or peer/peer) are physically removed from each other (Suomi et al. 1970; Suomi et al 1973; Suomi et al. 1976). Typically the social bonding between two animals has occurred early in life. A considerable amount of variability in the impact of separation on the behavior of the animals involved has been reported (Suomi 1983). Suomi (1983) has also noted that the number of cases of disturbance resulting from separation (mother/infant, peer/peer) are outnumbered by cases where little disturbance has been observed. The incidence of separation anxiety in animals that were paired as adults will in all likelihood be less. Thus, if social units of nonhuman primates are to be formed, due consideration by the individuals forming the pairs must be made to the species, age and rearing history of the animals.

Another factor which must be closely monitored is the increase in health risks for animals in social units. Clearly, the possibility of a greater incidence of disease transmission and traumatic wounds must be taken into account when considering socially housing animals which are part of a research project (Novak & Drewsen 1988). A level of acceptable risk should be established by both the investigator and the attending veterinarian.

Another form of enrichment via social contact involves participation by humans. Socialization of dogs to improve handling and reduce behavioral abnormalities has been well documented (Vanderlip et al. 1985a; Vanderlip et al. 1985b). However, this area of environmental enhancement for nonhuman primates has to-date not been thoroughly investigated (see Cooper & Markowitz 1979 for the effect of a keeper on chimpanzees). In addition, questions of precise methodology and potential ramifications remain unanswered.

Other appropriate means of enriching the environment may occur in a social, but non-contact context. A viable solution to the problem of allowing social interaction between animals in a room whose physical configuration prohibits direct visualization of cage neighbors, is the provision of mirrors along an otherwise empty wall. This would allow the animals to visually communicate with each other via reflections (Gallup 1977; Gallup et al 1980; Gallup 1982). It has also been shown that many species of nonhuman primates will use social signals when viewing their own reflections in mirrors. This has been demonstrated in stumptail macaques (Anderson 1983), patas monkeys (Hall 1962), Japanese macaques (Itakura 1987), rhesus monkeys (Gallup et al. 1980), cynomolgus monkeys (Gallup 1977), and squirrel monkeys (MacLean 1964).

The importance of olfactory and visual signals for communication purposes in nonhuman primates has been clearly outlined for the various species (Zeller 1987). Single housed monkeys do experience a limited degree of social interaction with other animals in the room by auditory, olfactory and visual communication (Novak & Suomi 1988).

Social stimulation of several species of nonhuman primates can also be accomplished by means of nonthreatening, non- contact methodologies.

An increase in grooming behavior by the provision of an artificial fleece attached to the home cage has been examined at the NIH (Bayne et al., In press ).

Visual contact between what are generally considered "social" species (i.e., living in formal troops or more loosely structured groups) can often ease the stress from what would otherwise be an isolated environment if special attention is paid to the animal arrangement. It is probable that the maximum benefit from this arrangement comes from housing like species across from each other. The provision of visual contact between these animals can be accomplished by requiring that the animals in a room be housed across from each other as much as is possible (clearly an odd number of animals will result in one animal relying on diagonal social viewing). For rooms with a square rather than rectangular design, the physical arrangement of the cages can be planned to maximize visual contact between the animals.

Visual contact between animals can also be increased by providing the home cages with a mutual wall of translucent materials (for example, impact-resistant plastics such as polycarbonates). This would prevent injuries resulting from loosely woven metal cages and increase neighbor-viewing over the current cages with one solid wall. For those buildings with rooms too narrow to have cages along both walls, windows in the walls between such rooms (also constructed of impact resistant plastics) could be provided. As many species of nonhuman primates seem to be very curious about activities in the corridors outside their rooms, windows in the doors to the rooms provide a strong focal point for interest. The cages closest to the door, then, should not remain empty.

Much auditory communication occurs between the animals in a room. Communication between rooms might provide further stimulation for these animals (similar to one troop communicating with another). In some facilities, this may be accompliished by having a mesh component to the wall at the ceiling level. Other electronic solutions are conceivable as well. However, there is little available data in this area, and the effects of auditory communication on behavior need to be tested.