M. E. Bitterman

Resistance to Extinction in the Goldfish as a Function of Frequency and Amount of Reward

The results of early experiments on partial reinforcement (PRE) in the fish suggested that fish and rat might be differently affected by inconsistency of reward.1 More recent results suggest that fish and rat perform differently in such experiments because they are differently affected, not by inconsistency of reward, but by frequency and by amount of reward. A difference in the effect of frequency of reward is indicated by the fact that the fish shows the PRE more readily in experiments with equated reinforcements than in experiments with equated trials-when trials are equated, the partial Ss have fewer reinforcements-while the rat seems to show the PRE as readily in either kind of experiment.2 Evidence of a difference in the effect of amount of reward comes from experiments on the PRE as a function of amount of reward. While the resistance to ex-

Classical appetitive conditioning in the pigeon

Pigeons were trained in an appetitive conditioning situation (CS paired with the presentation of food). The CS-US interval was 1 sec. for one group and 10 sec. for another. Stabilimetric measures of anticipatory activity showed better conditioning at the longer interval.

Resistance to Extinction in the Pigeon as a Function of Secondary Reinforcement and Pattern of Partial Reinforcement

Suppose that rats are partially reinforced in a runway with differentiated end-boxes (e.g. black on reinforced trials, white on unreinforced trials) which cannot be seen until the instrumental response has been made, and then extinguished-one group with the end-box used on reinforced training trials, and a second group with the end-box used on unreinforced training trials. The principle of secondary reinforcement suggests that resistance to extinction will be greater in the first group than in the second; yet exactly the opposite results were obtained in two experiments reported about 10 years ago, and the results of other experiments, while to some extent inconsistent, certainly do not yield readily to a simple secondary-reinforcement interpretation. Hulse and Stanley obtained fragmentary evidence of greater resistance in the group extinguished under end-box conditions like those of reinforced training trials, but the design of their experiment was unbalanced2 Lawrence and Festinger,3 like Freides,4 and Notterman,5 found no significant difference between the groups. Using two different negative end-boxes, Marx found no significant differences between groups extinguished with one or both of the negative end-boxes and a group extinguished with the posi-

Resistance to Extinction in Normal and Extensively Decorticated Rats as a Function of Amount and Percentage of Reinforcement

In the first of a series of neurosurgical experiments patterned after those which differentiate the normal rat and the (pre-cortical) fish, habitreversal and probability-learning were studied in adult rats with extensive cortical lesions made in infancy.1 The operates behaved like normals in spatial problems, i.e. they showed progressive improvement in habit-reversal and tended to maximize in probability-learning. Only in visual problems (simultaneous brightness-discriminations) did the role of the cortex become evident. In these problems, the operates behaved like fish, i.e. they failed to show improvement in habit-reversal and showed random matching in probability-learning.