In a pre-training four groups of rats ran a straight alley for 11 days. Group H (Hungry) and group HT (hungry and thirsty) received dry food as reward while group T (thirsty) and group TH (thirsty and hungry) were given water. For group TH the reward was shifted from water to food on the 9th day. The mean log running times are shown in Fig. 2 and thier statistical comparisons in Table 2. Drive discrimination learning followed the pre-training for the next 7 days using a T-maze. In this discrimination learning a specical procedure was divised and the drive was suddenly shifted from hunger to thirst or from thirst to hunger. This was done in the following manner : the shift from hunrer to thirst was made by having the animal eat dry food only : in the same way the shift from thirst to hunger was by letting the animal drink water only. With the method, it took two hours hunger to thirst and a half hour from thirst to hunger. In this discrimination learning correct responses of all groups reached 80 per cent on the 5th day (Fig. 3 and 4). This result is superior to those of previous drive discrimination learning experiments under traditional procedures. Group TH showed a significantly better performance than groups T and TH. However, since the effect of “irrelevant drives” was not found, Kendler's “selective principle”was not confirmed so far as this training period was concerned. The effect of pre-training on the drive was shown for the first two days (Fig. 5). Groups H and HT which were given food in pre-training, showed more correct responses to food than to water, conversely groups T and TH more to water than to food. The difference between group TH and group HT whose drive state was approxmately equal, was statistically significant (Table 4). It was concluded that the effect of or pre-training on the drive was due only the reinforcing agent and not to the physiological drive conditions. This effect was found to be related to the acquired drive.
It has been long since the fact was noticed that figure look smaller, paler and dimmer after prolonged inspection period. It goes without saying that we need to analyse the changing process of such phenomena with the increase of the stimulating time, in order to clarify the significance of the temporal factor in visual phenomena. We made it clear in the former report that there is growing process under the short stimulating time. And, we expect that after the growing process is over, the decaying process may begin. We tried to make clear by quantitative measurements of the perceived length and size in connection with prolonged inspection in this study. The results were as follows : 1. The perceived length of a line ends its growing process at about ten seconds, and then the decaying begins. And, it seems that with some subjects the decay will not reach its maximum even after four minutes. 2. Up to the stimulating time of 5 seconds, the overstimation coccurs. It is conceivable that this happened because we used the variable stimuli which were exposed for one second, as the measure. 3. The same occurred when the variable stimuli were exposed for two seconds, as the measure. This fact seems to support our supposition mentioned above, as the degree of the overstimation in the latter case is less than that in the former case. 4. But, we could hardly find any difference between the two exposure, with regard to the tendency of increasing understimation rate in the stimulating time of over ten seconds. This indicates that we must consider the interaction between the standard and the variable figure which might produce some effect on the judgments. 5. Although the measures are dependent upon the various experimental conditions, it has been made clear that the perceived length of a line ends its growing process in the stimulation time about five seconds, and then its decaying process begins and the underestimation rate increases. 6. There is no difference in the tendency between the two stimulus intensities of 5 and 20 radlux, excepting that the degree of the underestimation in the former is less than that in the latter, and that in the former the underestimation rate gets to the maximum in the stimulating time of four minutes with all subjects. 7. The decaying process of the perceived size of a square is not different from that of a line. 8. But, the underestimation rate gets to the maximum in 2 minutes. The reason seems to be becoause the stimulus intensity is only 5 radlux. We must consider that corresponding to the temporal process of visual phenomena, the psycho-physical field, constructed by us, goes through the decaying process after the forming process has passed. Supposing that the current distribution corresponding to the strength of the field as Köhler did, we may understand the forming process of the field as the temporal process of the current distribution owing to the specific nature of the medium in the nervous center. And, likewaise, we may understand the decaying process of the field as the results of such current process and the inhibitory process caused by the increase of polarization.
Two groups of rats were run a T-maze with reward in both goal-boxes. Group LD under three hr. food deprivation was given two series of 11 trails a day for six days. The inter-trial interval used was 20″ for one series and 60″ for the other. The order fo the two series was counterbalanced from day to day. Group HD was the same as Group LD except that the deprivation was 23 hrs. Per cent alternation as a joint function of drive levels, inter-trial intervals and the nember of trail days was shown in Table 1 (see Text) and the analysis of it in Table 2. The effect of drive strength itself was lacking but the triple interaction between drive, intervals and days was significant. A closer analysis of the data (Fig. 1) revealed that the alternation tendency in Group LD was little affected by the interval but rather by the number of trials, while both effects were observed in Group HD. As the intra-day series of trials were spaced one hr. and the first daily runs were separated 24 hr., alternation percentages for the inter-trial intervals of one hr. and 24 hrs. could be computed (Tables 4, and 5, and Fig. 2). Decay of the alternation tendency appeared but the effect of drive was again nil. As the decay of alternation does not necessarily indicate the growth of position preference, number of turns to preferred side was measured throughout 12 series (Fig. 3). The preferred side was defined as the side which the rat chose more than half of the total trials. The growth of position preference was clearly shown although drive strength was ineffective (Table 6). The presrnt results did not agree with either Elliott nor Fujita both of which proved the negative relationship between hunger drive strength and alternation behavior. However the weakness in experimental design of the two studies was pointed. That is, five choice-alleys of Elliott's apparatus would not be behaviorally equivalent in that less energy was needed for turns to the middle alleys compared with others ; the time interval between 22 massed trials was not the same among the three groups of different drive lavels in Fujita's study. These points could explain their results irrespective of drive differential. Previous inhibition theories of alternation behavior expect the growth of alternation with the increase of discriminative ability in the organism. As the hungrier rat would have a greater ability to discriminate, its alternation tendency would be greater. However the same rat would acquire a greater SER increment with a reinforcement and this in turn would reduce alternation. These two factors would work against each other and thus the effect of drive level on alternation would not be realized. However as the high-drive rat would be more sensitive to the inter-trial interval difference than the low-drive rat because of its greater ability to discriminate, the effect of the time interval on alternation would be greater in the former. These were what was obtained in the present study. The decay of alternation throughout trials could not be explained by previous theories because the daily blocks of trials were spaced as long as 23 hrs. The author agreed with the conditioned-inhibition theory of Calvin et al. but believes SIR is a learned negative drive. It was hypothesized that negative drive decrease alternation and facilitates the pre-learned or innate position preference in the organism. This could explain not only the present results but also less than 50% alternation which was often observed, and especially Maier's response fixation. The fact that negative drive of electric shock decreased alternetion as shown by Iwahara, Soeda and Iwahara, was in complete accordance with the hypothesis.