The present experiment was designed to determine whether abnormal fixations varied in amount and strength with an increase in frequency of punishment. A modified jumping-stand apparatus was used (1). Subjects were 44 naive albino and hooded rats from 70 to 90 days of age. Main procedure was the same as one used by Maier and Feldman (7).
All animals were first trained to jump at cards and then were confronted with an insoluble problems (condition A). Animals were divided into four groups in terms of frequency of punishment relative to reinforcement, i. e., 0%, 30%, 50%, and 70%, respectively. After a consistent position response persisted for more than 95% of 160 trails, all rats were given a card discrimination problem for 200 trials (condition B). In this condition the guidance technique was alternated by the ordinary trial-and-error procedure. Under both conditions an electric shock from the floor of the jumping-stand was used to force the animal to respond when it refused to jump within 30 seconds.
The results were as follows :
Condition A 1) Of 44 animals, 41 readily developed position responses. 2) The amount and the intensity of forcing electric shocks increased with frequency of punishment (Table 1). 3) Although the mean response latency in each of three punished groups showed a tendency to decrease as the trials progressed, it was uncertain whether the machanism of avoidance learning operated there (Fig. 1). 4) Wolpe's interpretation based on primary escape drive (14) may not be adequate for the present results.
Condition B 1) All rats abandoned position responses and all but one formed discrimination response within 200 trials. This means that guidance is effective for breaking a response. 2) The rigidity of position responses for each group positively correlated with frequency of punishment in condition A (Table 2). 3) The concept of similarity of two situations proposed by Hilgard (2) would not be appropriate for an interpretation of persistences of position responses, but rather for the speed of learning after rats abandoned position responses (Table 2). 4) The number of fixated rats increased with frequency of punishment in condition A. But no significant differences were found among the mean strengths of fixations (Table 3). 5) All punished rat's scores showed a tendency to distribute bimodally (Fig. 2). 6) The number of errors for fixated rats in additional trials was significantly great than for nonfixated rats (Table 4). 7) There was no significant difference in the amount and intensity of electric shocks in condition A between fixated and nonfixated rats.
From these results the following conclusions are derived :
1) The increment of punishment ratio in an insoluble situation increases the number of fixated animals, but is not related to the strength of fixated response. 2) The results of the present experiment seem to favour Maier's frustration theory rather than the learning theories of Mowrer, Wolpe, and Hilgard.