Rats were trained to turn a wheel to avoid unsignaled shock. They then had Pavlovian conditioning in which a tone signaled shock (CS+ group) or time free shock (CS-group). During a "test" session presentation of the tone increased the mean rate of responding of the CS+ group and reduced the mean rate of responding of the CS-group. In a second experiment, the tone was first contingent on an increase, then, later a decrease, in the avoidance response rate. The distribution of avoidance responses of the CS-group was successfully modified by both reinforcement contingencies. The results demonstrated that CS-, a conditioned stimulus for safe periods, established independently of any avoidance situation, can reinforce ongoing avoidance behavior. Rescorla and LoLordo (1965) have examined the discriminative properties of Pavlovian conditioned stimuli superimposed on unsignaled avoidance (Sidman, 1953) extinction. In their experiments, dogs were trained in a Sidman shuttleavoidance task, then subjected to Pavlovian conditioning in which one tone, CS+, signaled shock, while another tone, CS-, signaled shock-free or safe periods. Both CS+ and CS-were later presented during Sidman avoidance. Responding increased over the base line during CS+, and declined below the base line during CS-. In ordinary terms, CS+ elicited fear and CS-elicited relaxation.The present experiments examined the discriminative and reinforcement functions of inhibitory, or safe, signals established during independent Pavlovian conditioning sessions. According to theories of avoidance behavior advanced by Soltysik (1963) and by Weisman, Denny, and Zerbolio (1967), CS-should not only elicit conditioned inhibition or relaxation, but also when made contingent upon an instrumental response, CS-should function as a positive conditioned reinforcer for instrumental behavior in much

In Experiment 1, rats were trained to turn a wheel to avoid unsignaled shook, then received Pavlovian differential conditioning with either a .5-, 1.0-, 2.0-, or 5.0-min. intertrial interval. Decreased responding to CS-, but not increased responding to CS+, was a function of intertrial interval duration. In Experiment 2, Pavlovian differential inhibitory conditioning, with two conditioned inhibitors, CIi and CI», each signaling a different shock-free interval, was administered. Decreased responding to CIi, that always signaled at 200sec. period free from shock, was a function of the duration of the shock-free (10-, 110-, 200-, 290-, or 390-sec.) interval correlated with CL. The results suggested possible relationships between the role of the intertrial, or shockfree, interval in Pavlovian inhibitory conditioning and in avoidance training.

No abstract

stripes. He also showed that the transfers were a genuine cansequence of discrimination training and could not be explained in tenns of either novelty or Two squi"el monkeys (Samiri scuireus) initial preferences between "rows," were trained to discriminate between two "columns," and stripes. The same effect patterns that were identical, except that has been demonstrated in squirrels and is their orientations differed by 90 deg. In sufficiently surprising 10 make it worth subsequent transfer tests (a) the fmding out wh ether or not it occurs in a training-positive shape )Ws shown with a primate visual system. novel shape having the same "positive" METHOD orientation, and (b) the training-negative Two young squirrel monkeys (Samiri shape )Ws shown with the same novel scuireus) were used, maintained on a 23-h shape in the "negative" orientation. food-deprivation schedule, being fed daily Transfer choices were anomalaus; the novel after the experimental session. The shape)Ws prefe"ed to the training-positive discrimination apparatus was a shape, and the training-negative shape was modification of the suspended monkey pref~ed to the novel "negative" shape. restraint platfonn described by Thompson, These results are held to be a function of Seal, & Bloom (1959). The S faced a black the contour-coding processes of the plywood panel, in the center of which were mammalion visual system. two 3 x 2~ in. elear Plexiglas panels.Nonnally, one expects an organism that has been trained to discrirninate between two stimuli to choose the training-positive stimulus if it is shown in a subsequent transfer pair or to avoid the training·negative stimulus if it is shown. Relational learning, or transpositional discrimination, may be an exception to this generalization under some conditions, but in these cases, the dimension a10ng which the transposition occurs is usually unitary and easily identified. Kreehevsky (J938) discovered an anomalous transfer situation, in which a novel shape was preferred to the training-positive shape but where the relevant stimulus dimension (if any) was not obvious. He trained rats to discriminate between square arrays of small squares arranged in "rows" or in "columns," as shown in Fig. 1, i.e., the horizontal center-to-center distance between squares is smaller than the vertical center-to-center distance in the "row" array, and vice versa for "columns." A group of rats trained with "rows" positive was subsequently presented with a choice between rows and an equally bright pattern of smooth horizontal stripes, and reliably preferred the stripes. Another group trained with "columns" positive subsequently preferred smooth vertical stripes to "columns." Dodwell (1965) replicated this fmding (also with rats) and, moreover, showed that transfer preferences were anomalous with respect to the negative shape; rats trained with "rows" negative subsequently preferred "rows" to horizontal stripes and similarly for "columns" and vertical Psychon. Sei., 1970, Vol. 18 (I) Stimuli were back-projected onto a milk-glass...

Ether anesthesia administered either 100 or 3160 sec after footshock significantly improved discriminated avoidance responding and reduced response latency only at the longer interval. Ether anesthesia as a highly aversive stimulus accounts for the results better than the hypothesis that ether disrupts memory consolidation.