100 "Journal of Comparative Neurology and Psychology. 



and rarely until it has actually executed a coiled reaction. Fur- 

 thermore the S reaction is ordinarily first performed by a reversal 

 of the head from an extended general flexure or a coiled reaction 

 before the original flexure is completed in the caudal part of the 

 trunk. This reversed movement of the head, in early stage of the 

 embryo, may simply progress caudal till it reverses completely 

 the original flexure ; but when the movement attains its typical form 

 it is a relatively short, quick movement, and, when performed in 

 series, it becomes the normal swimming movement. 



The occurrence of the S reaction in series has its origin, evidently, 

 in a mode of response which appears very early in the course of de- 

 velopment. It may be designated as the ''secondary reaction." This 

 secondary reaction is a movement that is made during the phase of 

 relaxation from a direct response to an external stimulus. It is 

 caused, probably, by a rhythmic process in the motor cells, or, possi- 

 bly, by stimuli from the proprio-ceptive field. It may be of greater 

 or less extent than the original flexure. It may, for instance, ad- 

 vance a general flexure into a coiled reaction. It is a conspicuous 

 feature in the behavior up to the time when the S reaction appears. 



Now, it is obvious that when the head is once reversed from a 

 flexure into an S reaction, the secondary reaction would explain the 

 second reversal, which is simply repetition of the initial movement. 

 The successive reversals of the head may, then, be initiated as second- 

 ary reactions and the progression of the successive flexures caudad, 

 in the form of S reactions, propels the animal forward. 



Locomotion, therefore, in the amphibian embrj^o is dependent upon 

 the progression of the flexure cephalo-caudad, and the cephalo-candal 

 progression of the individual movement is further correlated with a 

 similar progression in the ontogenetic development of the reaction. 

 Furthermore, it is clear that this order of development of function 

 is correlated with the order of structural development of the central 

 nervous system, as illustrated, for instance, in the order of closure 

 of the neural tube. These correlations naturally suggest, further, 

 that the necessity of locomotion may have been an important phylo- 

 genetic factor in determining the order of development of the parts 

 of the nervous system in vertebrates. 



