Intracranial stimulation in pigeons has previously been shown to be capable of sustaining extinction responding in hungry birds (Macphail, 1967). Experiment 1 was designed to test the hypothesis that this effect was due to the disruption by brain-shocks of the extinction process; it was found that brain-shocks had no effect on a task involving a series of reversals. The notion that brain-shocks have the effect of prolonging extinction through having acquired secondary reward value was shown to be untenable in Experiment 2. It was concluded that the brain-shocks have a positively reinforcing effect comparable to that responsible for self stimulation in mammals.In previous experiments it has been found that electrical stimulation of the brain (ESB) delivered to hungry pigeons for responses during extinction following food-rewarded sessions may maintain extinction responding (Macphail, 1967), but that most birds showing this effect do not work for ESB in the absence of prior food-rewarded sessions, despite the delivery of "priming" shocks (Macphail, 1966). The two experiments reported here were designed to test the plausibility of two possible explanations of the effects found, alternative to the view that the ESB is acting as a primary reinforcer.EXPERIMENT 1 Experiment I tests the possibility that the effect of ESB is in fact not to positively reinforce behavior, but rather to disrupt the extinction process.Method. Subjects were five adult pigeons (Columba livia) each having an implanted pair of bipolar electrodes in the forebrain; all had been shown to respond reliably during extinction periods where each response obtained an 0.4 sec train of ESB. The Ss were maintained at 80% ad lib weight throughout the experiment.The apparatus was a modified Skinner box, with two keys. In pretraining each key was exposed singly, and Ss learned to respond at FRIO for food reward.In test sessions, both keys were exposed, and the fust response, to whichever key, was rewarded; three food rewards were then delivered at FRIO on that key, responses to the other key being scored as errors and having no effect on rewards. After these three rewards, the reinforcement value of the keys was reversed, responses on the original key now counting as errors, and food being delivered at FRIO on the other key. The reinforcement value of the keys was reversed in this way, every 30 "correct" responses, until the bird had obtained 25 food rewards (i.e., had made 240 correct responses).In control sessions, no ESB was delivered at any stage, whereas in experimental sessions, an 0.4 sec train of ESB was delivered for every response, to whichever key. The intensity of ESB was set for each S at a level previously shown to maintain extinction responding in that S. There were two sessions per day, and five days testing. The first two sessions were control sessions, run to ensure reliable responding by the Ss, and error scores on those sessions were discarded. On the next four days, one session was experimental, and one, control; the fust session of the ...