Winfred F. Hill

Reward magnitude and number of training trials as joint factors in extinction

Reward magnitudes of 1, 2, 4 and 8 pellets (.045 gm) were combined factorially with 8, 16, 32, 64, and 128 reinforced training trials for a total of 300 rats in a straight runway. Extinction speed was a monotonic increasing function of number of training trials and also tended to increase with reward magnitude, but the latter effect was not significant. There was no interaction between reward magnitude and number of trials and no indication of nonmonotonicity for either variable.

Spontaneous Alternation with and Without Reward in Hooded and Albino Rats

Spontaneous alternation was investigated with and without food reward in hooded and albino rats by means of a 2 × 2 factorial design with 18 Ss per cell. Neither the main effects of strain and reward nor the trend over successive trials were statistically significant. However, the interaction of strain × reward was significant, reflecting significantly more alternations by rewarded Ss within the hooded strain but a nonsignificant difference in the opposite direction within the albinos. A significant strain × training interaction probably reflected a greater decline in alternations with training by the albinos.

Differential accrual of frequency in verbal-discrimination learning

The notion that Ss build up a different number of frequency units for correct and incorrect items of a verbal-discrimination task was tested. Following verbal-discrimination learning, Ss were given a recognition retention test. It was found that they correctly identified more of the correct than incorrect members of pairs. This is in support of the tested notion.

Effect of Unconsumed Reward on Subsequent Alternation of Choice

The quantity of food reward initially available and the quantity remaining in the goal box of either chosen discriminandum were investigated as possible determinants of subsequent choice behavior, i.e., subsequent repetition or alternation of choice. In two similar experiments, 108 rats were divided equally among 3 experimental conditions: Ss in Group C+ and Group C were identical in terms of the number of food pellets consumed on each trial but differed in that Ss in Group C+ always had a considerable quantity remaining upon their removal from the chosen alternative. Ss in Group One were always presented with 1 pellet, which they consumed. No differences in alternation behavior were found among these groups.

The effects of sensory and activity deprivation on choice of an incentive

Twelve female albino rats, subjected for 24-hr, periods to (1) sensory stimulation coupled with activity deprivation; (2) unrestricted activity accompanied by sensory deprivation; or (3) a normal laboratory group condition, were given a choice between incentives of sensory experience and a Wahmann activity wheel immediately following each period of confinement. While there was no significant effect of conditions upon initial choice, over 42 days with one 8-min.irial daily, Ss spent more time at the sensory incentive and less time at the activity incentive following condition 2 than following the other 2 conditions.

The effect of varying numbers of rewarded and nonrewarded trials on subsequent performance in a straight runway

Sixty male albino rats were assigned randomly to groups in a design in which one variable was whether rewarded or nonrewarded (latent learning) sessions began the experiment, and the second variable was whether one, two, or four sessions preceded a final rewarded session and extinction session. Extinction speed in a straight runway increased with number of sessions and with initial rewarded rather than nonrewarded sessions. Latent learning sessions facilitated performance on the first rewarded session, the facilitation increasing with number of latent learning sessions.

Extinction under nine food deprivation levels1

One-hundred-eighty Ss received 60 acquisition trials under 22 h food deprivation and 30 extinction trials with nine different food drive levels for different groups. A regular increase in extinction running speed with drive was observed early in extinction.

An Attempt to Canalize Activity Drive

Murphy ( 1947) has suggested that an important factor in the development of motives is the canalization of generalized drives into certain specific modes of expression. Whenever a certain type of consummatory behavior reduces the drive, the probability that this behavior will be used again to reduce the drive is presumed to increase. The present study is an attempt to canalize activity drive (Hill, 1956) to different degrees into a particular expression, wheel-running, by varying the availability of alternative forms of activity. Ss were 36 male albino rats of Sprague-Dawley strain. At age 25 days they were placed in individual mesh cages of two kinds: small (10 in. by 6 in. and 5 in. high) and large (18 in. by 12 in. and 1 2 in. high). Each rat had continuous access to a Wahmann activity wheel. After living for 21 days in these cages (cmlization phase), all Ss were transferred to large cages (different specific cages in the case of those already in large cages). Each rat continued to have access to the same wheel as before. This test phase lasted seven days. Ss were run in nine shifts of four rats each, two in each condition, with each wheel used alternately for a large-cage and a small-cage S in successive shifts. One rat died and one was eliminated because of Es error. Since the large cages permitted extensive locomotion and climbing, while the small cages were quite confining, it was predicted that large-cage animals would show higher wheel scores in both phases and a relative increase during the canalization phase. For the last 18 days of the canalization phase, the mean number of wheel revolutions was 793.5 for the rats in large cages and 680.5 for those in small cages ( t = .91) . For the last four days of the test phase the means were 1219.2 for the large-cage and 1003.6 for the small-cage rats ( t = .80). The large-cage group was much more variable, the ratio of variances yielding an F of 256.2 for the canalization phase and 7.87 for the test phase, both significant at the .001 level. In each group, eight Ss showed an increase in activity from the first half of the canalization phase to the last half and nine Ss showed a decrease. These findings, negative except for the unexpected difference in variability, suggest some reservations about either the generality or the strength of the canalization process.