cal importance of the question and the fact that most attempts to repeat Scripture's work had led to negative results invited a new attack. The plan which we undertook may be illustrated by the following scheme, though the actual execution of the experiment was carried out with greater refinement and in a different way as to details.

The observer is presented with a triple series of meaningless syllables, as in Group I below, and is required to read series b a certain number of times and if possible learn it. Series a and c are of course all the time before his eyes though not involved in his task. After reading b the required number of times, his knowledge of it is tested by the "Treffer method,' and his success in giving the required syllables recorded. Then after a brief interval he is presented with Group II of which the middle series is the same as one of the side series in Group I, e. g., series a, and he is required to read (and learn) series a in the same manner in which he has just read (and learned) series b.

If a is on the average learned with greater ease or completeness than b, the inference is that the previous presentation of a in indirect vision has somehow been helpful-directly by rendering the syllables individually more familiar, or indirectly through their association with the syllables of series b which have in the learning been associated with each other. If a is not learned on the average more easily or perfectly than b the inference is either that no assistance is gained by the "unconscious' perception of a or that the gain is not of sufficient amount to be determined by this method of experimentation.

Such experiments were carried out on two trained observers through a considerable number of days, but it may be said at once that the results were on the whole negative. There was no clear evidence of any advantage. The presumption is that the assistance gained is small in amount-too small to be deter

mined by this method. Later experiments undertaken expressly to determine the delicacy of the method showed that one reading of the a series with full attention had no beneficial effect upon the learning of the a series after an interval of ten minutes during which the b series had been learned.

Though the results of this series of experiments must therefore be set down as inconclusive, they may have, perhaps, a certain value in other connections and are therefore given in the Appendix of this paper.

2. The Effect of Attention and Distraction on the Formation of the Motor "Set" (Motorische Einstellung)

EXPERIMENT 2. The purpose of the second series of experiments was to find the effect on the "Motorische Einstellung' of attention and distraction. The term Motorische Einstellung'' indicates the effect which repeated lifting of a heavy weight has in making subsequent lighter weights seem too light. It is probably due to a temporary habit of the nervous system. The problem in our case was to discover whether a neural habit of this sort, of whose existence the subject was unaware, would be more readily formed when he was attending to the lifting of the heavy weight than when he was inattentive to it.

The phenomenon of "Motorische Einstellung'' was first reported by Müller and Schumann. They lifted a moderate weight of, say, 600 grams and, after it, lifted a heavier weight of 2,400 grams to an equal height a certain number of times, in a definite rhythm. Then a weight of 800 grams was lifted and found to seem lighter than the 600 grams, lifted before the training with the weight of 2,400 grams. They explain the illusion by saying that the 800 grams, which is lifted with an unusually powerful impulse after the work with the weight of 2,400 grams, rises with unusual speed and therefore seems lighter than the first weight, because we are apt to judge as lighter a weight which raises more quickly. The repetition of the lifting of the heavy weight has set up a tendency in certain sub-cortical centres to discharge automatically with a somewhat extra intensity. Experimentation of this kind was carried further by Steffens.2

The apparatus used is pictured in the accompanying cut. Two boards measuring about eighteen inches long were clamped to the sides of the bottom of a chair so that the ends

'Müller u. Schumann: Ueber die psychologischen Grundlagen der Vergleichung gehobener Gewichte, Pflüger's Archiv, XLV, 1889, 37-122. 2Steffens: Ueber die motorische Einstellung. Zeits. f. Psy., Bd. 23, S. 240-308.

extended about seven inches beyond the front edge. Holes were bored near the forward ends of the boards and through these were passed the ends of two handles by which the weights were lifted. The upper parts of the handles were made of wood and were provided with grooves into which fitted the fingers of the observer, enabling him to hold the handles firmly and in the same way each time he lifted. The handles below the board consisted of brass rods having at their lower ends disks of wood, on which the weights rested. An iron needle was passed through each brass rod in the middle, making it possible to raise the handles only a given distance. To prevent the needles hitting against the boards with a jar, a string was fastened in front of the chair, by means of two iron standards clamped to the table, at such a height that the observer's hands would touch the string before the needles came in contact with the boards; as soon as the hand touched the string the weight was lowered. A disk of cork was used on each handle to prevent the clinking of the weights against each other. The entire weight of each handle with the cork disk was 100 grams.

The chair stood on one of the large laboratory tables. As far as possible from the observer a metronome was placed, its noise being deadened by a cloth pad between it and the table. The experimenter sat at the side of the table to the observer's left, and changed the weights as the experiment required. These were flat and circular in form with a rather large slit so that they would slip on and off the handles easily. The method of the experiment was this: The right-hand weight was always the standard, and was always kept at 300 grams, i. e., a 200 gr. weight plus the weight of the handles. By trying different weights a weight was found for the left hand which usually seemed equal to the right-hand weight. Owing to the difference in strength between the right and left hands this was actually a weight much smaller than the standard. Since practice was apt to increase the strength of the left hand, it was necessary to determine what this weight was before every experiment; and doing this counteracted also any influence which might have been carried over from lifting heavy weights in the experiment of a previous day. After determining the apparently equal weight, twenty judgments were made, upon weights offered for comparison with the standard (300 grams in the right hand), four with the weight which had been judged equal, and four each with weights ten and twenty grams above and ten and twenty grams below the "equal" weight. If the judgments were perfect the results would of course show four judgments "equal," eight "heavier'' and eight "lighter." As a matter of fact they