358 THE AMERICAN NATURALIST [Vol. XLII 



While it appears conceivable that the running of the maze, 

 after it is learned could be exclusively a matter of the inner 

 sensations, so that the rat might follow the path even if the maze 

 were absent, it is extremely difficult to see how this could be the 

 case with the first learning of the maze. Our proposed ex- 

 periment of removing all source of ' ' extraorganic " sensation by 

 doing away with the walls of the maze seems a reductio ad ab- 

 surdum; the rat certainly would not learn the typical maze path 

 under such conditions. How does the rat learn that it. must 

 put forth so much effort in a certain direction, then turn in a 

 certain way? It would seem that for this, certain data from 

 outside,— due either to feeling the wall with vibrissa?, or run- 

 ning squarely against it, or the like— are necessary. Even when 

 the rat runs full tilt against a wall, that gives some sort of an 

 extrinsic sensation that would seem to require consideration. 



This is a point which the author does not make clear, though 

 he seems to argue decidedly that it is possible to explain the 

 entire behavior, both in learning and in finding the way after it 

 has been learned, from the purely inner sensations. Does he 

 perhaps hold that the only effect sensed by the rat, when it runs 

 against a wall, is one of restraint, of prevention of further effort 

 in that direction? Such a view would seem possible, if at all, 

 only for actual collision at full speed, while cases where the rat 

 merely feels the wall with its vibrissa? seem hardly open to this 

 interpretation. It would have been helpful to certain readers if 

 the author had so far recognized their obtuseness as to take up 

 this point explicitly. In passing it may be remarked that the 

 entire paper is written in a curiously confused and careless man- 

 ner, as if it were a first draft which the author had not found 

 time to revise. So thorough and important an investigation de- 

 served better treatment. 



A further development of the work set forth in the paper of 

 Watson is presented in a later paper by Carr and Watson. 2 

 If the rat's method of threading the maze is merely to go a 

 certain distance (as measured by the amount of effort put forth) 

 then to turn in a certain direction, etc., without any data from 

 outside itself, then evidently it would be greatly disturbed by 

 altering the lengths and proportions of the passageways. Or if 

 it is set down not at the entrance to the maze, but in the middle 



*Carr, Harvey, and Watson, J. B. Orientation in the White Rat. 

 Journ. Comp. Neurol, and Psychol, 1908, 18, 27-44. 



