DISCUSSION

The method here used for measuring the fixational pause of the eyes has certain disadvantages and certain advantages as compared with the methods previously used by others. One disadvantage lies in its dependence upon the verbal report of an observation which Subject 3 (one of the authors), at least, found to be somewhat difficult and felt to be somewhat vague in the outcome. The method falls, perhaps, into the class sometimes called "subjective,” as compared with one in which a graphic record is made for subsequent leisurely study and measurement. Further, the actual time-interval within which a given number, n, of spots may be uncovered, cannot be much more accurately defined than to say that it is more than the time corresponding to n - I intervals on the circle, and less than that corresponding to n + I.

As to the advantages of this method, the one which appears to be most important is the fact that it makes possible the accumulation of a large number of results in a short time, without the risk and discomfort of making a mechanical attachment to the eyeball each time a subject comes to the experiment, and without the use of the expensive and troublesome photographic recording apparatus. In either case, the subsequent tedious work of measuring from the records and tabulating the measurements is avoided. Incidentally, the number of measurements actually involved in the averages of Table I is close to 1,900.

A further possible error to be mentioned is that of the selection of the results in the course of observation. One of the authors, Subject 3, found the work difficult and uncertain and he was unable to make report on many of his trials. In this case, we believe the results are selected, and the fact that this subject's average times are shorter than those of any others would seem to bear this out. On the other hand, the distribution of his results showed no definite asymmetry. Other subjects in the group, notably No. 1, were able to make, almost without exception, a definite report upon each trial.

A question of interest concerning the fixational pause of

the eyes is whether it may become shortened in the course of prolonged practice of a routine operation. The present results were examined with this in mind by the comparison of the averages within each group of four or five series for identical conditions. Thus in a series of five averages there are ten possible pairs of averages for comparison; within a group of four averages, six comparisons are possible. All told, the data yielded 358 such comparisons. In 154 of these, the later result was smaller than the earlier one, in 12 cases equal and in 192 cases the later result indicated a longer fixational pause than the earlier one. Thus these results, rather than showing a shortening of the pause with practice, actually showed a tendency for it to become longer as time went on. It must not be forgotten, however, that in no case was the opportunity for practice very great; no subject went through more than 25 series of ten separate observations each, and four of the subjects showed results contrary to the aggregate result just mentioned. The possibility that the time could be shortened by practice is still open. Indeed it would seem to be quite probable. In the study of the separate observations it appears that, while the average time was about 150, some few results were as low as 710 and values of 830, while not frequent, occurred in considerable numbers. It is not unreasonable to suppose that the result of practice might in a measure prove to be that the occasional best performance of the unpracticed subject becomes in the course of time his usual performance.

SUMMARY

I. A total of nearly 1,900 measurements, made with nine subjects, would indicate that the average fixational pause of the eyes, between two 10° sweeps, is close to 1500.

2. The distribution of the separate measurements is not far from normal. Very few were as low as 710, and few as high as 2500. The mean variation of the average from a series of ten observations runs as low as 80 for one subject and as high as 240 for two others. Selection of the results by the subject in the course of the work was not excluded by

the method used, and undoubtedly took place in the former In other cases it was probably negligible.

case.

3. While the duration of the fixational pause tended on the whole to lengthen rather than to shorten as the experimental work proceeded, the possibility of materially shortening the average time by practice is by no means excluded.

REFERENCES

1. HESS & HARRISON. Trans. Illum. Eng. Soc., 1923, 18, 787-794.

2. COBB, P. W. J. Exper. Psychol., 1923, 6, 138–160; 1925, 8, 77-108; 1926, 9, 95-109.

3. FERREE, C. E., & RAND, G. Trans. Illum. Eng. Soc., 1922, 17, 75-77.

4. HUEY, E. Amer. J. Psychol., 1900, 11, 283–302.

5. DODGE, R. Psychol. Rev. Monog. Suppl., 1907, no. 35, 95; Zsch. f. Psychol., 1909, 52, 321-422.

6. MILES & SHEN. J. Exper. Psychol., 1925, 8, 344-362.

A STUDY OF THE PHENOMENON OF

REMINISCENCE

BY OSBORNE WILLIAMS

Psychological Laboratory, University of Chicago

HISTORICAL INTRODUCTION

The purpose of this investigation is to study the phenomenon of reminiscence as a function of the age of the subjects and of the character of the material used. The nature of the study can best be stated in terms of its historical setting.

In 1885, Ebbinghaus (4), showed that the ability to recall nonsense material decreases with the passing of time and that the most rapid forgetting occurs immediately after learning. These results were later confirmed by Radossawljewitsch (9), Finkenbinder (5), Strong (10), Bean (2), and others, who found the relative amount forgotten, however, to be less than Ebbinghaus reported.

The original Ebbinghaus curve was obtained with one adult subject, with nonsense syllables, and with the relearning method of measuring the amount retained. The later experimenters employed various experimental conditions. They made use of poetry, abstract words, consonants and typewriting, in addition to nonsense materials. The methods of anticipation, recognition, selection and reconstruction, and written reproduction were also employed.

In spite of the variety of conditions the typical Ebbinghaus curve was invariably obtained. This curve was, therefore, generally assumed to represent a universal law of forgetting, although it was admitted that differences in the degree of learning, distribution of effort, length or amount of material, method of presentation, method of measuring the amount of retention, etc., might affect the amount retained.

In 1913, however, Ballard (1) secured certain curves which differed from those of the Ebbinghaus type. The characteristic feature of them was the initial rise which occurred during the first two or three days after learning where the Ebbinghaus curve indicated a rapid decrease in retention. Ballard called this phenomenon of improvement in memory without further learning "reminiscence," as opposed to "obliviscence," or the process of decline in the capacity to remember.

This use of the term is perhaps unfortunate inasmuch as reminiscence usually refers to any act of recall. We have decided to retain the term, however, because of its adoption by other writers.

Ballard experimented most on twelve-year-old children, and his percentages of retention for this age-level are shown in our Table A. Exhibit 1 was obtained from a 20 per cent. degree of learning of some verses from "The Wreck of the Hesperus;" Exhibit 2 represents 40 per cent. degree of learning of parts of "The Ancient Mariner;" while Exhibit 3 is for 20 per cent. degree learning of nonsense verses. The amount remembered immediately after learning was taken as the basis on which the percentage later reproduced was calculated.

These peculiar findings of Ballard have since been verified by Huguenin (6), Brown (3), and Nicolai (8). Reminiscence must thus be regarded as a valid phe

nomenon.

This difference of results in the two sets of experiments is presumably due to some difference in the experimental conditions, which may be compared in four respects as follows.

1. The age of the subjects. While Ballard worked both with adults and children, he obtained the phenomenon of reminiscence only with children. Hence he assumed that reminiscence is a function of age. This conclusion is not justified, first because Ballard employed too few adult subjects and secondly because he varied the kind of material, the method of presentation, the degree of learning, and the method of measuring the amount of recall. Both Brown and Nicolai obtained reminiscence with adult subjects, and Radossawljewitsch secured the Ebbinghaus type of curve with children. Thus the results are not at all decisive.

2. The nature of the material. The experimental data furnish no evidence that reminiscence is a function of the kind of material employed. While Ballard utilized both connected and disconnected materials, and secured reminiscence only with such connected materials as ballad poetry and nonsense verse, yet Brown and Nicolai obtained the phenomenon with such disconnected materials as lists of abstract words, geographical names and familiar objects. Likewise, the Ebbinghaus type of curve has been secured with the use of poetry, as well as with nonsense syllables, abstract words and consonants.

3. The degree of learning. On a priori grounds it would seem that partial learning is an essential condition of the phenomenon, for reminiscence would hardly be possible with complete learning. This assumption is apparently justified by the data, for Ballard, Huguenin, Brown and Nicolai secured their results when the materials were only partially learned; while the previous investigators required their subjects to memorize the materials perfectly. But Ballard also obtained the Ebbinghaus type of curve when the materials were incompletely learned. Evidently some other coöperative factor is essential to the phenomenon.