Shock to a conspecific as an aversive stimulus

Abstract

This experiment confirms and extends an earlier finding that a hungry rhesus monkey (0) will avoid securing food ifthis subjects anothe r monkey (SA) to e lectric shock. In the present series this "sacrificial" behavior was manifested in 6 of 10 animals independently of the r e lative positionofthe two animals in a dominance hierarchy. It was also found that while prior shock of the o resulted in inhibition of responding following the introduction of shock to the SA, this variable was not correlated with the fina l manifestation of a sacrificial pattern. Introduction In a pilot study itwasfoundthat 6 of 8 hungry rhesus monkeys (Operators, or Os) avoidedope rating a switc h for food when this caused another monkey (Stimulus Animal, or SA) of equal size to undergo elec-· tric shock. Inasmuch as animals frequently respond to shock with attack reactions (Ulrich & Azrin, 1962), and since the two Os that were not inhibited by shocked SAs of the same size did avoid shocking a much large r SA, the pre sent expe riment was designed to determine whether aversion to the shock of others would be maximized if the 0 wa s submissive to its SA. Data we r e also gathered on the effects of prior shock and the degre e of familiarity between the two animals. Subjeets The Ss were 6 male a nd 4 female sexually immature rhe sus macaques. For one year prior to and during the experiment, they were housed three to a cage as part of the larger colony. The apparatus was a 5 x 2 x 2 ft masonite and plexiglass box divided in the middle by a oneway screen that prevented vision from the right (SA) to the left (0) compartment. The latter was equipped with red and blue stimulus lights, two chains suspe nded from ceiling microswitches, a food tube, and a water cup. The SA compartment was bare exc ept for a floor grid attached to a Grason-Stadler E6070B shock source and scrambler. Training The animals we retrainedtosecure a .7 gm food pellet by pulling the left chain in response to a red light and the right chain after a blue one, each of 5 sec. duration, in the following steps: (a) Adaptation. The animals were given five daily 1 hr. ses sions in which to adapt to the 0 compartment and associate the feeder click with de li very offood. (b) Single lever training. Five animals were initially trained to pull the right chain during the blue s timulus and five were trained to pull the left chain during the red. The stimuli were pre sented at an average interval of 30 s ec. ; chain pulls between s timuli stopped the tape programmer controlling the stimulus pre sentation for 60 sec. Animals were run on alte rnate days afte r 22 hr. of food deprivation. Session length was s et at a maximum of 90 min., but stopped sooner if the Ptyclwn. Sci., 1964, Vol. 1. animal did not respond for 5 min. Following the attainment of criterial performance consisting of at least 90% correct trials and no more than 25 unsignalled chain pulls, the animals were trained on the other chain in the same manner and to the same criterion. (c) Discrimination training. The animals were presented with the r ed and blue light stimulus in random order at an average intertrial interval of 15 sec. IfanOshowed signs of positioning during this phase, i. e., consistently pulling one chain to the exclusion ofthe other, stimulus presentation for the preferred chain was halted until the 0 correctly responded to the other chain. Sessions were run daily and were terminate d when the 0 did not respond to five consecutive red and five consecutive blue stimuli. Animals were trained to a criterion of two consecutive sessions in which they responded corre ctly to 90% of the stimuli during a minimum of 50 trials and made fewer than 25 unsignalled r e sponses. ltominanC'e testing and pairing After all animals had attained criterion in discrimination learning they were paired against each other in another apparatus and required to compete for 100 grapes presented one at a time. In most cases dominance wa s quickly established, the dominant animal getting better than 90% of the grapes. On the basis of this relationship, the animals were divided into five pairs, with dominance, s ex and cagemate-noncagemate matchings balanced insofar as possible among the first and second members of the pairs. Testing The first m ember of e ach pair was used as the 0 and the second as the SA during the first test peri-ad , and the roles we re reversed during the second test period. The pairs were placed in their respe ctive compartments and the 0 was given a minimum of three days to adapt to the pre sence of the SA while continuing to perform at criteria l level. After this, one of the chains was programmed not only to operate the feeder for 0, but also to administer a 3 sec., 5 ma shock to the SA; whereas the other chain operated the feeder only. This programming was random, e xcept where 0 had manifested a position habit, in which case the prefe rred lever was wired for shock to the SA. The animals were tested daily until one of the following two criteria was reached: (a) two consecutive sessions in which one chain was sele cted more frequently than the other (binomial p < .01); or (b) two consecutive sessions in which relative response frequency was not significantly different at < .25 level. Sinc e the number of correct r e sponses per session at the end of training was almost never less than 50 , those with fewer we re not included in this calculation. After an 0 had reached criterion, the roles of 0 and SA were re-47