644 



SCIENCE. 



[N. S. Vol. XX. No. 515. 



strata of our atmosphere. And at this 

 threshold it will be well for us to pause and 

 defer further exploration to a future time. 



L. A. Bauer. 

 U. S. Coast and Geodetic Soevey, 

 Washington, D. C. 



SCIENTIFIC B00K8. 

 Educational Psychology. By Edward L. 

 Thoendike, adjunct professor of genetic 

 psychology in Teachers College, Columbia 

 University. New York, Lemclie and 

 Buechner. 1903. 173 pp. 

 This book 'attempts to apply to a number 

 of educational problems the methods of exact 

 science ' and to place at the service of students 

 the incoherent mass of psychological knowl- 

 edge which constitutes ' the beginnings of what 

 we may call a general dynamic psychology.' 

 That the facts already known " have thus far 

 gone without systematic and convenient ex- 

 position is due to the complexity of the prob- 

 lems involved, not to any doubt concerning 

 their practical importance. What we think 

 and what we do about education is certainly 

 influenced by our opinions about sucli matters 

 as individual differences in children, inborn 

 traits, heredity, sex differences, the specializa- 

 tion of mental abilities, their interrelations, 

 the relation between them and physical en- 

 dowments, normal mental growth, its period- 

 icities, and the method of action and relative 

 importance of various environmental in- 

 fluences." These then are the topics with 

 which the book is concerned. 



The first chapters deal with the measurement 

 of mental traits. They explain the statistical 

 principles by which we can apply exact meas- 

 urements to groups of variable objects taken 

 as a whole, and then proceed to show that in 

 the matter of mental traits, as of physical, 

 human beings of the same sex and approxi- 

 mately the same race, age and experience con- 

 stitute a true group — that for every mental 

 trait there is a mean or ' center of gravity ' 

 from which slight variations are frequent and 

 great variations rare, according to the law 

 of the distribution of chance events; so that 

 measurements in terms of percentile grades 

 or other relative standing can be interpreted 



fairly accurately in terms of absolute amount, 

 and causal agencies acting unequally on dif- 

 ferent parts of the group can be detected by 

 the analysis of curves of distribution. 



In the fourth chapter the author explains 

 the statistical principles by which we can 

 measure the correlation between different char- 

 . acteristics in a group of individuals, and then 

 applies Pearson's coefiicient to show that in 

 the matter of mental traits this correlation is 

 amazingly small. This means that within a 

 given group of individuals goodness or bad- 

 ness in one psychological function (as shown 

 by school grades or the results of special 

 tests) is not particularly likely to be accom- 

 panied by a similar amount of goodness or 

 badness in another, even though the two ap- 

 pear to be only slig'htly different. If any one 

 is disposed to deny this, let him first go over 

 the large amount of evidence (much of it 

 based on his own incessant investigations) 

 which Dr. Thorndike puts together in this 

 chapter. 



Ever since the appearance of Professor 

 James's chapter on memory the doctrine of 

 general mental functions which can be cul- 

 tivated by appropriate exercise and then 

 turned to any use (a survival of the old 

 'faculty' psychology) has been more and more 

 discredited amongst psychologists, though col- 

 lege presidents stiU make use of it to pro- 

 claim their wares and teachers in the public 

 schools have never dreamed of anything else. 

 But this doctrine can not be reconciled with 

 the small coefficients of correlation shown in 

 chapter IV. If there is such a thing as 

 training or neglecting ' the memory ' we 

 should expect to find goodness or badness in 

 remembering words accompanied by something 

 like the same amount of goodness or badness 

 in remembering numbers, to say nothing of 

 colors, forms and tastes. . But as a matter of 

 fact the correlation is 'slight and variable'; 

 and so with other functions. 



In chapter VIII., on 'The Influence of 

 Special Forms of Training upon More Gen- 

 eral Abilities,' the author presents evidence 

 bearing still more directly upon this important 

 educational doctrine, and gives a powerful 

 argument to those who would adapt school 



