The Reaction to Tactile Stimuli 
259 
special cutaneous receptors and conductors, but such a cephaliza- 
tion would certainly favor the development of such systems, for, as 
already suggested, their peripheral conductors hold essentially the 
same relation to the cephalic part of the central system as do the 
most primitive central conductors from the trunk. 
It should be noted here that a certain amount of locomotion 
may be acyfuked by an amphibian embryo by other movements 
than the S' Reaction as described above. The body may be flexed, 
for instance, and straightened by a series of secondary, vibratory 
movements. Such a reaction propels the animal on its side in a 
circle or spiral path. Also, a rapid succession of reversed flexures, 
in which no S reaction can be detected, may give swimming in a 
zigzag, erratic course. But normal, upright swimming in a direct 
course is, according to my observations, attained only through the 
perfecting of the S reaction and its performance in series. 
As already suggested, this development of the swimming move- 
ment is of interest from the point of view of animal behavior. We 
now see that swimming, which may be regarded as instinctive in 
these forms, arises as the elaboration of the simplest known reflex 
in the embryo, the contraction of the most cephalic trunk muscles. 
I Certain forms of the flexure, such as the U reaction and the coiled 
reaction, do not seem to be in the direct line of the development 
of the swimming movement, being simply intensive or tetanic forms 
i of the ordinary flexures. On the other hand, the other types of 
I flexure develop in a regular order and in a remarkably constant 
1 manner into the movements of locomotion. Now none of these 
I simple flexures can be regarded as having any value as trials, since 
1 the Diemyctylus swims perfectly upon leaving the egg membranes 
in the normal course of development, and within them it can gain 
no practical experience for swimming out of movements of any 
sort. Instinctive swimming, therefore, and the simplest reflex 
alike, are inherent in the neuro-muscular system of the embryo, 
and while 'the former develops in a regular order out of the latter 
I the movements themselves, which conform to this order, can have 
i no selective value. The question naturally follows whether in 
I forms which do not admit of such early experiments, such as birds, 
many quadrupeds and primates, the various forms of locomotion, 
I as well as other forms of behavior, which, in a greater or less degree, 
[ appear to develop out of a series of trials, may not conform to the 
I same law. It seems altogether possible that in such cases, also. 
