The matching and oddity techniques offer a convenient means for surveying the discriminative capacities of higher animals (4,6,7,9, 10,11, 12,17, 20). These techniques also lend themselves readily to studies of delayed response (1,14,15,17), "categorizing" and "symbolic" behavior (18), and "conditional response" (2, 3,16, 19).The present paper describes an experiment in which seven chimpanzees were trained to respond to that one of two very different objects which matched a third object, the sample. After mastery of this habit, the animals were tested with a variety of entirely new objects to determine the extent to which the visual matching response had been generalized. Results of the generalization tests indicate that the similarity or difference of the choice-object in respect to the sample determined the discriminative response. Three possible mechanisms through which similarity-difference may be effective are proposed.
BASIC TRAINING IN MATCHINGSubjects, Apparatus, and Procedure Subjects of the experiment were four adult and three young chimpanzees whose names, sex, and ages are listed in table 1. Tom, Helene, Jojo, Scarf, and Banka were bom at the Yerkes Laboratories; Mimi was acquired in 1931, Shorty in 1942. Tom and Helene (twins) and Mimi previously had served in many laboratory experiments, including various discrimination problems. Tom had specific training in matching-from-sample in two earlier studies (1, 14), the more recent of these being done four years before the present experiment. Shorty had received a great deal of training as stage and screen performer, but had been a subject in only one laboratory study. Jojo, Scarf, and Banka are nursery-raised animals; they had been handled extensively for physical measurements but had limited experience in formal experiments. During the present study each of the subjects had its own living cage and the training was done by wheeling the apparatus up to the wire mesh wall of the cage; no direct handling of the animals, therefore, was necessary.The apparatus used is one designed for general utility in discrimination experiments. Its over-all dimensions are: width 36", depth 24", height 56". The stimulus-carrier contained three food-wells, 7-£-inches apart, each of which was covered by a stimulus-object. Baiting of the proper food-wells and placement of the objects was done behind an opaque screen. For response, the screen was raised and the carrier was rolled forwards on horizontal tracks to the wall of the cage. By reaching with its fingers through the two-inch mesh, the animal could 62