they were trained to run down a black strip A comparison of learning based on social or of cloth tape, 1 in. wide, which began at the front startbox and ended at one of the nonsocial discriminative stimuli four goal boxes containing .5 ml of water. The tape was alternated from goalbox to goal box in the same random fashion as the VAUGHN E. STIMBERT* leader Ss in the social group. This Memphis State University, Memphis, Tenn, 38111 arrangement produced nearly identical behavioral topographies in both groups. Two groups of six rats each were trained to run across an open-field maze with four Extinction sessions were identical to choice-point alternatives. For one group the correct response was based upon a social acquisition sessions, with the exception discriminative stimulus and for the other it was based upon a nonsocial discriminative that water was unavailable for any stimulus. Results indicated that social responses were learned faster and were more experimental S. Each experimental S durable under extinction conditions than were responses based on nonsocial stimuli, performing without error during one acquisition session (10 trials) was placed on Previous studies in the area of animal relays and timers to record the latencies of extinction the subsequent session. All Ss social learning (Nakamura, Smith, & experimental Ss leaving the start box and were run under 24-h water deprivation. Schwartz, 1963: Stimbert. 1970) have the running times of all Ss. Running times RESULTS suggested that responses based on social included both start times and time to enter The accuracy of rats learning a social cue cues may be acquired faster and be more the goal box. Other features of the vs an inanimate cue is displayed in Fig. I. durable than responses to nonsocial cues. apparatus and the training procedures have Ouring acquisition all 5s in the social group Although numerous investigations (Daniel, been previously described in detail learned to follow other Ss in a maximum 1942: Husted & McKenna, 1966; Miller, (5timbert, 1969). of 80 trials, whereas after 200 trials, the Banks, & Ogawa, 1962; Murphy, Miller, & PROCEDURE nonsocial group was running with only Mirsky, 1955: Skinner, 1962; Wiest, 1969) The social response to be acquired by 72% accuracy and only two Ss in the group have shown the feasibility of using social the social group in this study was had reached the 100% criterion level. Based cues in learning studies, apparently no following behavior. A leader S previously on number of trials to criterion, these direct comparisons have been made trained to a specific goalbox (three leader differences were statistically significant between social and nonsocial discriminative Ss per box) was released from the front (t = 269, df= 10, P < .001). Due to time stimuli utilizing identical apparatus and start box and simultaneously a follower 5 limitations, the acquisition phase was similar response topographies. Since a large was released from the rear startbox. The discon tinued after 200 trials and ex tinction part of the adaptive behaviors of animals follower was reinforced with .5 ml of water sessions were run on those Ss having requires responses based on social learning, for entry into the same goalbox as the reached criterion. it is of some concern to determine how this leader. The double-compartment goalboxes It can also be seen in Fig. I that during learning differs, if at all, from that based separated leader and follower Ss and extinction nonsocial Ss reached a chance on nonsocial stimuli. The present study precluded viewing of the other S being level (25%) of performance after 40 trials, was designed to determine if such reinforced. Leaders were alternated while social Ss were following above 40% differences exist, and no attempt was made randomly so that on anyone trial each after 130 trials. Based on trials to criterion to identify the specific components of goalbox had an equal probability of being for those Ss reaching an extinction social or inanimate stimuli that might correct. criterion of 0% following for one session, produce such differences. For the nonsocial group Ss were also these differences were statistically SUBJECTS released from the rear startbox; however, significant (t = 10.95, df= 2, p < .01). The experimental Ss were 12 experimentally naive Sprague-Dawley 100 female albino rats about 126 days old at 90 the beginning of the study. Six Ss were VI assigned randomly to the social group and III ~80 six to the nonsocial group. Twelve female o Sprague-Dawley rats, approximately 300 e; 70 days old, were used as stimulus (leader) Ss. III These rats had been used in two previous II' 60 studies and were highly proficient in I- ~ 50 traversing an open-field apparatus with II' other Ss present. ~40 APPARATUS V The apparatus was an open-field maze I- 30 Z with double startboxes in tandem on one ~ 20 Sociala-:l side and four double-compartment II' Nonsocial ...... goal boxes on the other side. Photocells ~ 10 were located 1 in. in front of the rear start box and at the entrance to each 0'----:-'--....L.-'---*'.....L..-'--' ...... 7:::- ...... __'_-'-~--'-..L-'-::' 5 10 15 20 5 10 goalbox. They were programmed through AcquiSition Extinction

*No\\" at the Child Development Center, BLOCKS OF TEN TRIALS l'njver~jty of Tennessee Medical Units, 711 Jefferson Ave., ~lemphis, Tenn. 38105. Fig. l. Percentage correct in acquisition and extinction by blocks of 10 trials.

Psychon. Sci., 1970, Vol. 20 (3) 185 Differences in running speeds during 1.3 acquisition and extinction are shown in 0 ~ 1.2 Fig. 2. Again, the performance of the social ...v III group is consistently superior to that of the "- 11 nonsocial group. In acquisition, group ...: ~ 1.0 }~• l·'!.~'0/I' differences are not as marked as in Q "~ ,.,,' '. ~.,/ ... .9 .' \ 0 /",6'" extinction, where, after 20 trials, the ... \, Q. " , '0 III . running speeds in the nonsocial group .8 I, " • "'2\ dropped to less than .2 ft/sec. On the other ~ I Z I •, hand, social Ss maintained speeds z .7 I , consistent with their terminal acquisition Z .6 , ;:) \ performance of about 1 ft/sec. co: \ .5 , DISCUSSION Z \ The present fmdings support the io( , is .4 \ hypothesis that a response to an animate, ... Cf • ~ .3 I \ social discriminative stimulus is learned • \ faster and is more durable than a response Z .2 , to an inanimate, nonsocial discriminative ...io( • ~ .1 Social 00000() stimulus. Even after reaching the same Nonsocial _. asymptotic level of performance in acquisition, differences between the groups 2 3 4 5 6 8 9 10 2 3 4 during extinction are striking and favor the Acquisition Extinction social response. Individual social Ss have BLOCKS OF FIVE TRIALS been observed to follow at an 80% level of accuracy after as many as 150 extinction Fig. 2. Running times in acquisition and extinction by blocks of five trials. trials. Informal comparisons of previous durability than extinction in rionsocial rats. Journal of Comparative , studies on following behavior in rats with situations. 1942,34,361-369. other studies involving nonsocial situations At this time it is difficult to identify the HUSTED, J. R., & McKENNA, F. S_ The use of lead to conclusions that are consistent with rats as discriminative stimuli. Journal of the specific factors responsible for the Experimental Analysis of Behavior, 1966, 9, the extinction data from this experiment. superiority of social discriminative stimuli. 677-679. For'example, it appears that approximately Even though the social stimulus is typically MILLER, R. E., BANKS, J. H., JR .. & OGAWA, 150 trials are required to reduce social available for viewing less than 2 sec, the N. Communication of affect in "cooperative following behavior to a chance (25%) level response is acquired more rapidly than one conditioning" of rhesus monkeys_ Journal of once acquired under continuous based on a nonsocial, fixed and, what Abnormal & , 1962, 64, reinforcement. With partial reinforcement would seem to be, a more discriminable 343-348. experience, 150 trials only reduces visual stimulus. Using present procedures, MURPHY, J. V., MILLER, R. E., & MIRSKY, I. following to 60% (Stimbert, 1970). Using assessment of the importance of olfactory~ A. Interanimal conditioning in the monkey. counterconditioning procedures during auditory, and movement cues in facilitating Journal of Comparative & Physiological extinction, whereby nonfollowing what is primarily a visually controlled Psychology, 1955,48,211-214. responses are reinforced, the number of response might be accomplished. For NAKAMURA, C. Y., SMITH, J. c., & extinction trials required to obtain a example, it may be productive to study SCHWARTZ, F. W. Establishment of a lasting chance performance is approximately 75 following behavior in anosmic rats or to discriminative stimulus in rats by competition (Stimbert, 1969) or one·half that of training. Journal of Comparative & compare social learning based on moving vs Physiological Psychology, 1963, 56, 852-856. extinction procedures making fixed social stimuli. Following behavior SKINNER, B. F. Two "synthetic social reinforcement unavailable for any could also be investi~ated bv svstematicallv relations." Journal of the Experimental response. In contrast, studies of extinction changing the leader S so as to include other Analysis of Behavior, 1962,5,531-533. in straight runways or T-mazes frequently STIMBERT, V. E. Effects of early experience on strains of rats or possibly a different social learning in rats. Journal of Com parative show Ss reaching criterion in 25 trials orIess. species. This methodology provides an & Physiological Psychology, 1969, 69, In fact, it is unusual to find extinction additional approach for determining t!).e 640-643. data reported on more than 30 trials, variables relevant to animal . STIMBERT, V. E. Partial reinforcement of social which suggests extinction is essentially behavior in rats. Psychological Reports, 1970, 26, 723-726. complete at that point. Thus, the fastest WIEST, W. M. Socially mediated stimulus control extinction obtained in studies of social REFERENCES in pigeons. Psychological Reports, 1969, 25, following behavior indicates more DANIEL, W. J. Cooperative problem solving in 139-148.

186 Psychon. Sci., 1970, Vol. 20 (3)