Holmes, Random Movements. 109 



tended to cover the reactions of an animal which orients itself 

 according to the usual scheme. If an earthworm when illum- 

 inated more on one side of the body than another simply turn- 

 ed away, by a direct reflex, from the greater stimulus and kept 

 on doing so until the body was brought parallel with the rays 

 we would have a case of orientation according to the commonly 

 accepted theory of tropisms. The creature is forced into line 

 by unequal stimulation of the two sides of its body. When in 

 crawling one side of the body comes to be presented to the 

 light more than the other deviation is corrected by a direct 

 movement away from the stimulus. Getting out of line may 

 be regarded as an error which brings about a certain reflex that 

 sets the animal again upon a straight course. The worm keeps 

 in an approximately straight path because those movements 

 (errors) which bring it out of a straight course are counteracted 

 by a definite reflex, while those movements (successful trials) 

 which bring the creature away from the light are followed up. 

 It is because these errors are corrected that the orientation even 

 according to the ordinary theory, is maintained. These devia- 

 tions are as much errors as the random movements towards 

 the light in the method of orientation that is actually followed. 

 Tlie stimuli in ordinary tropisms may, however, serve to 

 prevent errors as well as to correct them after they are made. 

 Take an organism that orients itself by the direct method. As 

 it swings out of line it is doing so against the influence of the 

 unilateral stimulus which tends to turn it back during, as well 

 as after, the trial. But the mechanism of preventing the move- 

 ment and of reversing after it is made is, in this case, the same; 

 only in proportion as the checking predominates does the ran- 

 dom character of the movements becomes reduced. Between 

 the behavior of an organism like an earthworm that jerks back 

 and turns to one side when the anterior end is stimulated and 

 that of one which turns directly when the side is presented to 

 the stimulus there is to be sure a maked difference in behavior, 

 but there is an underlying basis of similarity in the two cases in 

 that in both errors are made, although they are corrected in dif- 

 ferent ways. The end result of both methods is the same, i. e. 



