BULLETIN OF THE UNITED STATES FISH COMMISSION. 
268 
Terminal buds do not occur in the outer skin of all fishes; in fact, they are prob- 
ably lacking here in the greater number of species. But whenever they do occur 
they tend to be arranged according to one general plan. This is particularly true of 
their nerve supply, for, though the details of the peripheral nerves of fishes are 
exceedingly diverse, yet the main communis branches for terminal buds, when such 
occur, are substantially similar from the Siluridae to the Gadidse. There are, how- 
ever, striking resemblances in detail between the siluroids and the cyprinoids, which 
are much more significant of close relationship. Both groups are characterized by 
an extreme development of the system, reaching generally over the whole body 
surface; in both cases the peripheral communis nerves correspond to the general 
teleostean type, though with a remarkable modification of the recurrent branch of 
the facialis in the case of the cyprinoids, and finally the communis centers in the 
medulla oblongata differ from those of all other teleosts in that there is developed a 
facial lobe as well as a vagal lobe in the primary central gustatory center. The 
facial lobe (the so-called lobus trigemini of siluroids and the “tuberculum impar” of 
cyprinoids) in both cases receives by way of the communis root of the facialis the 
nerve fibers from all of the terminal buds of the outer skin, while the vagal lobe is 
reserved for those from the mouth and viscera. This emphasizes from a new point 
of view the close relationship between these two gronps of fishes as recognized by 
the systematists generally. 
Though the Ostariophysi may have had a different origin from that of the other 
teleostean orders, yet the resemblances in general plan of the terminal bud system 
of sense organs in this group and in the other orders make it improbable that this 
system of organs has arisen independently and followed a paralleled development in 
the two groups of fishes. Its phylogenetic origin must therefore be sought among 
the ganoids, and until we have much more exact information concerning the nerve 
components and sense organs of these fishes further speculation in this direction 
is idle. 
This study has been directed primarily toward the solution of a simple physio- 
logical problem; but in a purely incidental way some points of interest to compara- 
tive psychology have come up. We have seen that in the cat-fish, hake, and tom cod 
the reflex of seizing food is normally set off by a combined stimulus of tactile and 
gustatory end organs. At first the fish will react similarly to a pure tactile stimulus 
and to the tactile plus the gustatory. After brief training, however, he acquires the 
ability to discriminate between the former, which is never followed by satisfaction, 
and the latter, which is followed by the pleasure of feeding. Clearly the fish learns 
by experience. We find also some differences between the different species of fishes 
in this respect, depending on the relative importance of the tactile and gustatory 
elements of the sensation complex in the normal reflex life of the fish. 
It would be interesting to inquire the part played by memory in these reactions. 
In the ease of Arneiurus , where the tactile and gustatory elements of the reflex of 
seizing food can be experimentally isolated by training, it would doubtless be possible 
to measure quantitative^ the duration of the persistence of this acquired discrimi- 
nation. I have made no accurate observations on this point, but can say in general 
that the memory of these fishes seems to be fairly good. (B} t the term memory I 
do not mean to prejudice the question of the part played by consciousness here. 
