■152 



SCIENCE. 



[Vol. XII. No. 295 



the marks by which the impression is identified. Dr. Lehmann at- 

 tempted to bring the problem to a simple issue by a study of the 

 power of recognizing various shades of gray produced by the rota- 

 tion of a disk partly white and partly black. Two such rotating 

 disks were used, either of which could be disclosed without the 

 other. First a ' normal ' disk was shown ; after an interval, either 

 a darker or a lighter disk was shown ; and the subject was to de- 

 cide whether the second disk was or was not the same as the first. 

 With this judgment, the direction of the difference, whether lighter 

 ■or darker, is always noticed. In the first series, the normal disk 

 was composed of equal parts of black and white ; the lighter disk 

 (only two disks were used) varied in the number of degrees of 

 white it contained as in the table ; the interval between the ap- 

 pearance of the two disks was thirty seconds, and the average 

 number of correct answers in a set of thirty judgments for each of 

 two observers was as follows : — 



We see at once that the power of recognizing the identity of the 

 ■disks decreases as the difference in shading between the normal 

 and the light disk diminishes ; fifteen answers of each thirty being 

 right by the action of chance. The difference between the normal 

 disk (180 degrees) and the disk of i88 degrees white is hardly 

 recognized at all. 



If we next complicate the matter by introducing a darker disk that 

 shall always contain as much more black than the normal as the 

 lighter contains more white, the power of recognizing the normal 

 disk when it appears, though still the same act, is made more dif- 

 ficult, as the following table, based on a series of experiments 

 precisely similar to those just cited, indicates : — 



Here mere guess-vifork would make ten answers in thirty correct. 

 The proportion of correct answers is smaller than in the former 

 case, markedly so when the disks approach one another in shade. 

 An increase in the number of possible impressions out of which 

 we are to select a particular one increases the difficulty of the act. 

 Again : the interval between the viewing of the two disks acts in 

 a similar way. When but two disks were used, always differing 

 by 12 degrees in the amount of white they contained (and the nor- 

 mal disk varying from 172 to 176 degrees of white), the average 

 number of right answers per set of thirty, with the interval varying 

 from five to a hundred and twenty seconds, was as follows : — 



In explanation of the individual differences, we recall that it has 

 been clearly shown that our memories are not equally retentive for 

 all kinds of impressions, but that here personal inclination and 

 talent find free play. Observer B is a student of art, with an in- 

 terest for such color-distinctions, and thus naturally surpasses his 

 <;ompanion. On the other hand, the effect of practice, tested by 

 comparing the first half of the experiments with the last, is equally 

 evident in both observers. It is to be noted, too, that after about 

 thirty seconds the memory-images have been about equally ob- 

 scured in both observers, B then making as many errors as L. 



Dr. Lehmann holds, that, if we recognize color-shades by simi- 

 larity, we should (when. three disks are employed) less often fail 

 to recognize the normal than the other disks, because we see it 



oftener than either of the others, it being shown every time. On 

 the other hand, the adjacency hypothesis would not expect such 

 a difference if the original shade v^'as clearly held in memory. If 

 the remembrance is indistinct, the extremes, the lighter and darker 

 disks, would be less apt to be wrongly identified than the normal 

 disk. The results speak in favor of the latter view. B, with a 

 clear memory, made 107 confusions, in 55 of which he miscalled 

 the normal disk, and in 52 of which he confused either the lighter 

 or darker disk with the normal one or with one another. L, with 

 a vague memory, made 165 confusions. 109 of which belong to the 

 former class, and 56 to the latter. 



For another experiment three scales, proceeding from black to 

 white by five, by six, and by nine equal gradations, were prepared ; 

 and after viewing this scale, any one of the five, of the six, of the 

 nine shades of gray was separately shown, and the attempt made 

 to assign to it its place in the scale. With the scale of five grades 

 (five observers and sixty observations), 96.7 per cent of all identifi- 

 cations were correct ; with the scale of six shades (three observers, 

 thirty-four observations), 70.6 per cent; with the scale of nine 

 shades (four observers, fifty observations), 46 per cent. Dr. Leh- 

 mann believes that the five shades are so well identified because 

 they have names attached to them, — black, white, dark gray, light 

 gray, and medium gray, — and supports this by the fact that he 

 succeeded in identifying the shades of nine-shade scale 75 per cent 

 of all times by simply associating a number with each shade. This 

 argues for the association through details or by adjacency in op- 

 position to association by similarity. 



If this adjacent mark is really the clew to recognition, then it 

 ought to be a great help to have the two disks (in the first-men- 

 tioned experiments) shown side by side before shown separately 

 for recognition. Furthermore, if the difference between the two 

 disks is great enough to be clearly perceived, it ought to be no 

 more likely to be overlooked than when it is much more marked. 

 A comparison of the following with the first table supports both 

 these inferences : — 



Degrees of white in light disk. 



Observe 

 Observe 



Furthermore, so slight an interval as from five to a hundred 

 and twenty seconds ought not seriously to affect the clearness of 

 this distinction. A comparison of the following with the third 

 table bears evidence to the truth of this deduction (the normal 

 disk contains 180, the lighter 215, degrees of white) : — 



To more completely show that this retention was not due to the 

 persistence of a memory after-image, an interval of a hundred 

 and twenty seconds was used, within which another recognition- 

 experiment was made. It was found that this interruption did not 

 seriously decrease the number of correct recognitions. 



Individual differences ought likewise not to be so marked with 

 this process, and a reference to the tables shows that they are slight. 

 Practice has little effect. Four observers made correct identifica- 

 tions ' through details ' S3.8, 85. S, 85.8 and 85.4 per cent of all 

 cases, while two observers differed in the percentage of correct 

 ' simple identifications' by as much as 83.5 and 67.6 per cent. 



Dr. Lehmann thus concludes that the associative law that best 

 explains the facts is the law of adjacency in opposition to the law 

 of similarity. 



The Magnet and Hypnotism. — The claim that the magnet 

 influences hypnotic phenomena is strongly upheld by Parisian ex- 

 perimenters, while others as boldly deny this influence. The effects 

 attributed to the magnet can be explained as due to unconscious 

 suggestion from the operators and the bystanders. Dr. Tanzi 



