April 30, 1920] 



SCIENCE 



443 



mental ability or in physical well-being. For ex- 

 ample, the grades of students who sit at the peri- 

 phery ef a group are appreciably lower than those 

 of students who sit in the center. Again, grades 

 at the rear of a room ^ow greater variation than 

 do those at the front. In general, the grades ob- 

 tained by a given student are dependent par- 

 tially upon such factors as his mental ability and 

 physical condition, but partially also upon his posi- 

 tion with reference to the rest of the group to 

 which he belongs. The disadvantages arising from 

 an unfavorable position in the group can not be 

 wholly attributed to the size of the lecture-rooms, 

 or to idiosyncrasies of the speaker. It is over- 

 come, in part, during the course of the semester, 

 and it may also be offset by the addition of fre- 

 quent small sectional meetings; it is increased by 

 such factors as intervening aisles and by unoccu- 

 pied seats. The disadvantage has been found inci- 

 dentally to rest upon variaitions in certain per- 

 ceptual and attentional factors and upon differ- 

 ences in the type of self-instruction under which 

 the individual works; but essentially to rest upon 

 the varying degrees of social integration which are 

 always present among the members of an assembled 

 group. 



', Speech and hrain patterns: L. W. Cole. Asso- 

 ciation experiments with nonsense syllables indi- 

 cate that verbal recalls are due to the presence of 

 brain patterns in which each syUalble is under the 

 influence of one branch of the pattern. The inter- 

 weaving of these patterns accounts for the con- 

 tinued suggestion by similarity of one idea by 

 another, or, in other words, it gives a neural basis 

 for association by similarity. It also gives a rea- 

 son for verbal lapses of memory in which there is 

 recall of part of one word with part of another 

 when the word sought for is partially forgotten. 

 The theory is merely an extension of Sherrington 's 

 conception of reflex patterns and it would replace 

 with a definite meaning such vague terms as 

 "mode" of impression, retention and recall, which 

 aTe used by many writers for the lack of a more 

 definite term. Finally the experiments with non- 

 sense syllables show that rhythm is the most per- 

 sistent and permanent element of a verbal impres- 

 sion. 



, A learning curve starting at approximately zero: 

 E. K. Strong, Jr. A boy of 5 years has been 

 given two minutes drill on addition combinations 

 a day for 150 days. At the start he knew nothing 

 of additions except that one and one made two and 

 that one and two made three and that he could 



count orally to twenity-five. The learning curve ob- 

 tained in this case does not follow the usual course 

 but runs nearly parallel to the base line for many 

 daya and then rises with a positive acceleration. 

 At the end of 158 days it had not suggested a 

 change from positive to negative acceleration. 



Methods of error elimination in a mental maze: 

 T. Peter.son. The mental maze method attempts 

 to study maze learning devoid of all the disturbing 

 spatial factors characterizing the usual mazes. The 

 experimenter has before him a picture of a cir- 

 cular maze, with the various parts lettered in a 

 random order. Sitting behind a screen, he calls 

 out to the subject pairs of letters representing bi- 

 furcations in the maze and the subject chooses 

 without seeing the maze. Whether the eorreot let- 

 ter is called first or last is a matter that is de- 

 termined by chance. The subject is instructed to 

 get to the goal with as few errors as possible, and 

 is told the number of errors each time on reaching 

 the goal, but he must find out for himself where 

 the errors are. Subject is also timed. Results 

 show backward elimination of errors of entrance 

 to blinds, and relatively early elimination of re- 

 turn ' ' runs, ' ' thus substantiating results obtained 

 by the author on rats in different forms of mazes. 

 ,The tendency to return to the starting place in the 

 maze at first greatly exceeds that expected on the 

 law of probability, but this tendency rapidly 

 yields to that of keeping the forward direction. 

 ' ' Coefficients of learning ' ' for the runs past the 

 several blinds are worked out statistically, each 

 coefficient representing the ratio of probable runs 

 past to probable runs into the blind. These coeffi- 

 cients are found to increase toward the goal end 

 of the maze, thus accounting for the backward 

 elimination of errors; and the advantage for learn- 

 ing at the goal-end of the maze over the entrance- 

 end is shown to be greater than in mazes with 

 many than in those with few blinds. Moreover, 

 this advantage is greater in the first trial than in 

 subsequent trials by any subject; it decreases with 

 successive trials, thus favoring more rapid learn- 

 ing in early trials. Statistical calculations as to 

 the number of errors in each part of the maze on 

 the expectations of chance laws, lead to the con- 

 clusion that, independently of the backward elimi- 

 nation tendency, learning progresses more rapidly, 

 in proportion to exercise, in the first and in the 

 last part of the maze than between the extremes. 



The development and functioning of a concept in 

 problem-solving: 3. C. Peterson. An objective 

 study is made of the reactions of adults to a nmn- 



