Interindividual behavior in a group of dogs reared together in a large field was studied to determine whether social and sexual interactions were more frequent in some pairs than in others. The experimental setting was designed to maximize the display of individual sympathies and antipathies. The basic test provided a situation in which behavioral interactions between a pair of animals was contingent on one animal approaching the other. One subject was restricted by a tether to a circular area with a diameter of approximately 12 feet while the other subject, the roving animal, was released into the outdoor field for 5 minutes, the duration of each test. Measures of association and a description of behavior were made and then the roles of each subject were reversed. Test combinations consisted of paired females, paired males and heterosexual pairs. Females were observed before, during, and after they were in behavioral estrus. Members of the opposite sex were most attracted to each other when the females were in estrus. However, all females were not equally attracted to all males. Three of four females showed preferences for some males over others. A preferred male enjoyed several advantages over nonpreferred males : (a) he was most frequently sought and solicited by the estrous female and she spent more time with him, (b) he was permitted to mount the female earlier in her estrous period, more frequently per day and, on more days, and (c) he was rarely prevented from investigating or mounting the female by being growled or barked at or by being bitten. Nonpreferred males were consistently rejected in this manner. One of the 5 males was readily accepted by all females. Otherwise, males preferred by one female were not uniformly preferred by the rest of them. In one case, a male preferred by one female was most frequently rejected by another. One female was equally receptive to all males. The preferences and aversions revealed by these subjects were evident the first time the females came into estrus and in subsequent estrous periods, whether natural or induced by exogenous hormones. These results were interpreted as indicating that sexual receptivity in the bitch is not completely dependent on ovarian hormones. Interactions between females were infrequent and brief as compared to those between two males or heterosexual pairs containing an estrous female. Strong individual associations between female pairs were not evident. Male-male interactions commonly involved elements of ritualized aggression while female-female interactions did not. The response of one male to another depended in part on which one had previously retreated in an aggressive encounter with the other. The submissive male in these stereotyped encounters was frequently visited, but he himself, rarely approached others. Some males preferred the company of some males to others. Members of a pair sympathetic to each other made frequent contacts of long duration and showed no aggression to each other.