33^ 'Journal of Comparative Neurology and Psychology. 



regards the persistence of the habits acquired, the white-black discrimination habit, 

 it is concluded, "may persist during an interval of from two to eight weeks of dis- 

 use," but "is seldom perfect after more than four weeks." The color discrimina- 

 tion habits never persisted more than two weeks. The white-black discrimination 

 was re-learned, after all traces of it had disappeared, in a shorter time than had 

 been required for its original acquisition, thus suggesting "the existence of two kinds 

 or aspects of organic modification in connection with training; those which consti- 

 tute the basis of a definite form of motor activity, and those which constitute 

 bases or dispositions for the acquirement of certain types of behavior." As 

 Rouse found to be the case with the pigeon, experience with one form of laby- 

 rinth made the learning of another form easier. 



3. The suggestions on method which the book contains are among its most 

 valuable features. In the first place. Dr. Yerkes finds that the best motive to 

 employ in studying the learning processes of mammals is not hunger, which is 

 variable in intensity, unfavorable, in its extreme degrees, to the exercise of the ani- 

 mal's full powers, and inhumane; but the punishment of mistakes by slight electric 

 shocks, given through wires on the floor of the discrimination box or labyrinth. So 

 far as my recollection serves, the author first used this method in his experiments 

 on the frog. It has certainly a decided advantage over any method where the 

 motive is constituted by a continuous state in the animal, such as hunger or the 

 desire to escape from confinement. Any continuous state is likely to vary in inten- 

 sity, and discomfort under confinement is likely to diminish as the animal becomes 

 used to its surroundings. An intermittent stimulus, given when mistakes are made, 

 has a much better chance of producing a constant effect. 



Another interesting point in method concerns the evidence which was obtained 

 that apparent color discrimination was really, in some measure at least, brightness 

 discrimination. This evidence consisted in the fact that mice which had been 

 trained to choose a compartment illuminated by green light in preference to one 

 illuminated by red light, on being offered a choice between black and white com- 

 partments, chose the white, although before the green-red training they had shown 

 no such preference. Thus it looked as if they had been choosing the green partly, 

 at least, as the lighter of the two visual impressions. This method might well be 

 given a wider scope in the study of the sensory discriminations of animals. It is 

 puzzling, by the way, in view of the evidence that red is much darker to the mouse 

 than to the human observer, to find that a tendency to choose red rather than yellow 

 in tests with colored cards is explained as the result of previous training to goto the 

 brighter compartment in white-black tests. 



The results of training tests are throughout stated in terms not of the time re- 

 quired to perform the act, but of either the number of errors made or the number 

 of tests required before the formation of a perfect habit. The time required in 

 traversing a labyrinth, for example, is in the author's opinion a poor index of the 

 perfection of the habit. In this respect his position is the opposite of that taken by 

 Watson, who states the results of his tests of the white rat wholly in terms of time. 



On the whole, the most favorable number of tests per day in the discrimination 

 series, taking into account economy of time and fatigue for both experimenter and 

 animal, was found to be ten. 



For labyrinth experiments, the author recommends the use of a standard maze in 

 which "errors by turning to the right, to the left, and by moving forward should 



