It is generally stated that the ability to reproduce previously learned material decreases with the passing of time and that the most rapid forgetting occurs immediately after learning. In most cases a negatively accelerated curve has been obtained when efficiency of retention is plotted against successive increments of time. Experimental studies, however, have yielded a significant number of deviating cases in which an initial rise in the curve of retention precedes a negatively accelerated section.Ebbinghaus (5), in 1885, made the first experimental attempt to plot the memorial history of an acquired material by taking retention tests at successive points in time and plotting the amount retained against the time unit. The curve of retention, as traced by Ebbinghaus, shows a descent which is initially rapid and which then becomes less and less rapid as time elapses. Ebbinghaus obtained his curve with the use of only one subject, with only one type of material, nonsense syllables, and with only the relearning method of measuring the amounts retained. The general trend of the curve, however, has received experimental substantiation by many investigators who have varied the conditions extensively. Radossawljewitsch ( 14), Finkenbinder (6), Luh (10), Bean (2), and Strong (17), varied the experimental conditions both in regard to the character of the material for which retention was tested and in the methods in terms of which retention was measured. They made use of nonsense syllables, poetry, consonants, abstract words and typewriting, and they measured retention by the methods of anticipation, recognition, 385

The preceding studies in this series of investigations of the transfer-time question have attempted to determine the extent to which the positive transfer effect from one learning situation to another is a function of the interval of time between them. Training on, one problem may however interfere with the later learning of a similar problem rather than facilitate it, and the question investigated in the present study is whether or not the negative transfer effect shown in the learning of the second problem is a function of the amount of time elapsing before the second is learned.Bergstrom (1) studied interference in card-sorting tasks and found that the interference effects was quite marked when only a few seconds elapsed between the two sortings, but decreased as the interval was made longer and disappeared completely after an interval of 24 hours. The related studies in the animal field have reported a two-phase result in the amount of positive transfer as a function of time. The amount apparently increases at first and then declines gradually as the interval of time between the two tests is made longer. This result has been found in one instance when the first task was completely mastered (4), and in another maze study when the first problem was only partly mastered (7). These results, taken together with the data on the retention of a maze problem for short as well as long intervals of time (3,11), indicate that apparently an increase in the efficiency of performance follows the lapse of a short interval of time 189

It was not thought necessary to present the directions here. They were similar to the directions ordinarily employed in maze experiments with human adults..

The familiar admonition to "get more sleep" reflects a presumption that loss of sleep is deleterious. This presumption is substantiated by experimental investigations which have shown that prolonged, uninterrupted deprivation of sleep leads, in human beings, to hallucinations and to other evidences of mental disorganization (1,9,13,14,16). Animal subjects have, indeed, been kept awake until they collapsed and died (3,6,10,17).

No abstract