measurement of attention by means of the detraction procedure; that is, they could not be used in the way I have used irregular preparatory intervals.

The results of Table XX show that the degree of attention of adults is very much superior to that of children. Further, by assuming à priori that attention increases with maturity, the results of Table XX lead to the same conclusion as those of Chapter III, namely, that the absolute detraction effect exerted by unfavorable preparatory intervals varies inversely with the degree of attention detracted from. This same conclusion was also reached in Chapter IV from the results of Table XV. The results of Chapter III and those of Table XX, taken together, show that the detraction effect produced by the unfavorable preparatory intervals is affected in the same manner by a decrease in the intensity of the reaction stimulus as by a decrease in the maturity of the subject. The results in both cases make it clear that only the absolute detraction effect of the unfavorable preparatory intervals, and not the relative, can be used as a measure of attention.

GENERAL SUMMARY

It is not my intention to restate here all the conclusions that may be drawn from the experimental work reported in the preceding pages. The most important of these conclusions have already been summarized at the end of each chapter. In conclusion, however, I shall give a general account of the method for measuring attention that has been worked out in the preceding pages. This account may, perhaps, best be made by basing it upon an outline of the distraction method, inasmuch as the distraction method is one with which every psychologist is familiar, and one that has been generally regarded as one of the most hopeful of the methods by which the measurement of attention has hitherto been attempted.

There are a number of ways in which the distraction method may be formulated, but the following schematic statement will, I think, be found fairly just. First, the efficiency of the subject in some work is to be determined under maximally favorable conditions. This work may be any kind of work which requires attention, as, for instance, the cancellation of certain letters of a printed page or the discrimination of a small difference in brightness. Second, the efficiency of the subject in this work is to be determined while he is working under definitely specified distracting conditions, or distractors, all other conditions remaining the same. A distractor is some condition which takes away a part of the subject's attention from the main task in hand. For instance, if a subject were to cancel the A's in a page of printing and at the same time to count the beats of a metronome, the counting of the metronome beats would constitute the distractor. Or, in case the subject were not required to count the metronome beats, then merely the sound produced by these beats might be called the distractor. After the decrease in efficiency resulting from a given distractor has been determined,

this decrease is somehow to be utilized as a measure of attention, presumably on the assumption that the decrease in efficiency produced by a given distractor varies inversely as the degree of attention against which the distractor acts. In short, attention is to be measured by measuring the ability of the subject to resist distraction.

I believe the general principles of the distraction method are sound. Practically, however, this method has been a failure. One of the chief reasons for this failure is that it has been found impossible to obtain a satisfactory distractor. A satisfactory distractor is one that, among other things, is constant or uniform in its effect. It must not greatly decrease the subject's efficiency at one time, and then hardly affect it at all at another time. No distractor which is constant in its effect has yet been discovered. Those so far used are not only irregular in their action but, for the most part, soon lose their efficiency.

One of the chief reasons for the failure of distractors is that they require divided attention, though this division may, perhaps, be thought of rather as an oscillation of attention. When the subject's attention is divided, it is impossible for the experimenter to control the proportion between the two parts. I have, therefore, undertaken to work out a method which would preserve the fundamental principle of the distraction method, but would not involve divided attention upon the part of the subject. The fundamental principle of the distraction method is not to distract, that is, to rend attention asunder, but merely to increase the difficulty of attention, to introduce a definite resistance. Now it is possible to increase the difficulty of attention without asking the subject to divide his attention. An important paragraph in any textbook of psychology is entitled the Conditions of Attention. Under this heading we have given a number of conditions which bring about a high degree of attention. Many, and probably all, of these conditions may be varied, that is, they are subject to quantitative gradation. Take for instance the so-called objective condition, intensity of stimulus. The intensity of the stimulus may be increased or decreased. And when we speak of intensity of stimulus as a condition of attention, what we mean is

that the more intense the stimulus, the more attention it attracts. Other things equal, a bright light attracts more attention than a dim one, and a loud sound attracts more attention than a soft one. Evidently, then, by weakening the intensity of a stimulus, we make a change which is unfavorable to a maximal degree of attention.

There are many other conditions of attention than intensity. An unfavorable state of any of these conditions is unfavorable to the highest degree of attention, but does not necessarily result in division of attention, and, therefore, cannot be spoken of as a distractor. It is proposed to use the term detractor to designate an unfavorable state of any of the conditions of attention. A distractor tears attention apart, a detractor merely reduces the degree of attention-no matter how. A detraction method of measuring attention, then, is the same as a distraction method, except that the distractor of the distraction mehod is replaced by a detractor, i.e., an unfavorable state of one of the conditions of attention, and replaced by a detractor which does not require division of attention.

In the detraction method described in the preceding pages the act the efficiency of which is measured is the act of reaction. When the subject is instructed to react as quickly as possible, the subject's efficiency is measured by the reaction time. In reaction time work it has always been customary to have a warning signal precede the stimulus by an interval of about 2 secs. This warning signal is given in order to allow the subject time to get ready to react as quickly as possible, so that the interval between the signal and the stimulus is called the preparatory interval. With a preparatory interval of 2 secs., especially after it has been repeated a few times, the subject knows exactly when to expect the stimulus, and so reacts with maximal adaptation of attention. Now when the procedure is varied so that the subject does not know just when to expect the stimulus the reaction time is much longer. For instance, if the reaction time to a visual stimulus with a regularly repeated 2 sec. preparatory interval is 200, with a set of intervals varying from 4 to 20 secs. in length, given in entirely irregular order, it may be, perhaps, 3000. In

the case of the reaction experiment, therefore, one of the conditions of maximal attention is a preparatory interval of 2 secs. Longer and irregularly mixed intervals on the other hand bring about a lower degree of attention. They decrease the degree of attention under which the reaction occurs, so that as regards the attention involved in a reaction, they constitute a detractor. The method, then, stated briefly, consists in determining first an individual's reaction time with a regularly repeated 2 sec. preparatory interval, and second, the prolongation produced in this time by substituting for the regularly repeated 2 sec. interval a set of irregularly mixed intervals of widely different length. This prolongation remains, on the average, constant in the case of a given individual, and does not decrease with moderate practice. This fact not only demonstrates the constancy of the action of the unfavorable preparatory intervals as detractors of attention, but means that the measurement of attention by this method does not vary with practice. This latter fact differentiates this measurement from most, if indeed not all, mental measurements so far in use.

A question of fundamental importance which remains to be answered is, just how is the prolongation produced by unfavorable preparatory intervals to be used as a measure of attention? Shall we, for instance, use the reciprocal of the absolute prolongation or the reciprocal of the relative prolongation? This problem was solved by artificially producing different degrees of attention, and applying the detractor to these different degrees of attention. In order to obtain the different degrees of attention, among other methods, use was made of one of the conditions of attention already referred to, namely, intensity of stimulus. Four different intensities of stimulus were used, on the assumption that the weaker the intensity the lower would be the degree of attention. It is true that weakening the intensity of stimulus brings about a lengthening of reaction time as the result of its effect upon other factors involved in the reaction than attention. It is not a pure detractor of attention. This fact would clearly prevent the use of weak intensity as a de