Separate groups of gerbils received 100%, 50%, or 0% reinforcement during avoidance acquisition. A reinforced trial was one in wh ich the avoidance contingency was in effect, while a nonreinforced trial was a CS-US pairing. Groups trained under either partial or continuous reinforcement were subdivided into two extinction conditions, 100% c1assical CS-US trials or regular avoidance extinction. In acquisition, CR frequency and percent reinforcement were positively related. Significantly more rapid extinction obtained under the CS-US pairing procedure relative to regular avoidance extinction, and animals trained under partial reinforcement were more resistant to extinction than animals trained under continuous reinforcement. These results were interpreted in terms of the effects of delayed punishment on nonreinforced trials in acquisition and extinction. The traditional discrirninative avoidance extinction procedure is one in wh ich the CS is presented alone and the US omitted, although the avoidance response remains effective in terminating the CS. Recently, Davenport & Olson (1968) have redefined extinction of avoidance so that the operations are more symmetrical with respect to those employed following appetitive instrumental training, Le., the reinforcement contingency for the instrumental response is eliminated. The redefined extinction trial is essentially a classical conditioning trial in which the CS is inevitably followed by the US irrespective of the S's behavior, although the US must then be response terminated_ This procedure effectively elirninates reinforcement via CS termination and/or US omission while maintaining motivation via the US presentations. When the traditionaI avoidance extinction procedure is employed, extreme persistence of the avoidance response is often observed (e.g., Solomon & Wynne, 1954). Davenport & Olson (1968), however, employing the redefined extinction procedure, found rather abrupt extinction of the leverpress avoidance response in rats. Unfortunately, these investigators did not include groups extinguished under traditional procedures for comparison. Gilbert (1970), on the other hand, found that wheel-turning avoidance in rabbits was much more persistent in a

A series of studies of shuttlebox-avoidance learning in the gerbil evaluated the efficacy of an exteroceptive feedback stimulus (FS). Experiment 1 assessed the relative effectiveness of a FS at 30-and 90-sec intertrial intervals (ITls), and found that the FS and warning signal termination contingencies were additive sources of avoidance reinforcement; i.e., they produced "supernormal acquisition" at the short ITI, but not at the 90-sec ITI. The effectiveness of a FS at the 30·sec ITI was further explored in Experiments 2 and 3, in which FS duration was varied in delayed and trace avoidance conditioning, respectively. In both studies, a FS facilitated acquisition but FS duration was not a critical determinant of performance. These results were interpreted in terms of an expectancy account of the informational value of a FS, and the problem of experimentally distinguishing between cognitive and inhibition-of-fear accounts of avoidance learning was discussed.There presently exists a considerable literature demonstrating the, efficacy of an exteroceptive feedback stimulus (FS) on avoidance acquisition (Bolles & Grossen, 1969, 1970 Bower, Starr, & Lazarovitz, 1965; D'Amato, Fazarro, & Etkin, 1968;Galvani, 1977;Keehn & Nakkash, 1959). The typical finding has been that, in the absence of response-produced warning signal termination (WST), a responseproduced FS results in avoidance learning equivalent to that produced by the traditional WST contingency. Recent formulations of avoidance learning have proposed that the FS reinforces avoidance behavior because it functions as a safety signal (Bolles & Grossen, 1970) or an inhibitor of fear (Morris, 1974; Weisman & Litner, 1972). Regardless of the terms employed in conceptualizing the functional significance of feedback in controlling avoidance responses, there seems to be agreement that the efficacy of a FS can be quantified in terms of the duration of the shock-free interval that is contingent upon FS presentation (Bolles & Grossen, 1970;Morris, 1974).In the discriminative avoidance paradigm, the contingency between the avoidance response and FS insures that the FS consistently precedes the shockSupported by National ScienceFoundation Grant B043556-001 awarded to the first author. This paper is based, in part, upon Michael Twitty's master's thesis, directed by P. , 1970). Indeed, relative to the baseline provided by animals lacking both a WS and FS, evidence of learning was at best marginal. Moreover, Morris (1974, Experiment 3) has provided indirect evidence suggesting that a FS might not be an effective reinforcer at a short ITI. In a transfer-of-training test, he failed to obtain evidence of the development of inhibition to a FS at a 3D-sec ITI, although conditioned inhibition to a FS was evidenced at an ISO-sec ITI. Taken together, these results raise serious questions about the effectiveness of feedback at a short ITI.Accordingly, the purpose of Experiment I was to assess the efficacy of feedback at a short (3D-sec) compared to a long (9O-sec) ITI in the Mongol...

No abstract