Although the number of rodents of different species and sexes was somewhat small, the consistency of the results enables a reasonably confident description of the effect of pairing rodents in an avoidance context. Considering first the situation in which both rodents were required to make the same manipulatory response to avoid or escape from electric shock: I. Naive rodents paired in this situation do not learn to avoid the shock by responding after the onset of a warning signal and before the shock. In the same situation, the rodents avoid 75-90 percent of the shocks if trained singly. This deficit is not the result of failing to notice the warning signal nor detecting its significance; the incidence of crouching, squealing and anticipatory foot-stomping indicate that the warning signal had acquired secondary aversive properties. 2. The presence of two rodents in the same avoidance situation does not materially affect their escape behavior when viewed at the molar level of the latency of escape responses. At a more molecular level, however, the escape response was almost invariably executed by the rat judged to be submissive when paired outside the avoidance context. When the shock was not quickly terminated, the dominant rat made distinct threat and aggressive responses such as foot-stomping and aborted attack until the submissive rat turned the wheel. 3. After the termination of the shock, the rodents engaged in various forms of interaction of an aggressive nature, including sparring, muzzling, over-and-under, and actual fighting. The most dramatic form observed was mounting with pelvic thrusting by the dominant rodent, even if it was a sexually naive female in the presence of another non-receptive female. 4. Essentially the same pattern of behavior was observed if either or both of the rodents had been pretrained singly in the apparatus either to escape or to avoid the shock. In the case of laboratory rats, which were observed to remain in close proximity to each other, the disappearance of avoidance behavior was more gradual than in their wild counterparts. In the case of the Florida packrats, which were observed to remain very far apart in the apparatus, avoidance behavior gradually reappeared. 5. The pairing of rodents in an avoidance situation had a dramatic effect upon their subsequent behavior when introduced into the situation singly. Even if they had previously learned the avoidance response individually, very little avoidance behavior was observed after a few sessions paired with another rodent. 6. All of the above observations are specific to the situation in which the same manipulatory response is required of both rodents. If a shuttle avoidance situation is employed,in which the necessity for confrontation in making the response is obviated, then avoidance behavior by both rodents continues when paired essentially at the same level as when run singly.