404 'Journal of Comparative Neurology and Psychology. 



been shown (Herrick, '04) to find their food ordinarily by the 

 simultaneous use of both taste and touch, though either sense 

 alone may, upon occasion, call forth the feeding reflexes and the 

 gustatory and tactile factors may be experimentally isolated by 

 training. The primary cerebral centers for both taste and touch 

 in these fishes are now accurately known. They belong, as we 

 have just seen, to entirely different systems of neurones, visceral 

 sensory and somatic sensory, respectively, which throughout the 

 central nervous system are anatomically very distinct. It becomes , 

 then, a question of some interest, where and in what way the 

 peripheral neurones of taste and touch from the outer skin come 

 into physiological relation to serve the ordinary reflex feeding 

 movements of these fishes. 



Since both systems of cutaneous nerves are highly developed in 

 Ameiurus, our inquiry will begin with this type, using for control 

 other species in which one or the other member is better isolated. 

 This study is based upon an extensive series of sections of brains 

 of various teleostian fishes, stained with Delafield's haematoxy- 

 lin, and the methods of Weigert and Golgi. The following types 

 have proven especially helpful: Ameiurus nebulosus, which ex- 

 hibits a high development of both gustatory and tactile centers; 

 Cyprinus carpio, where the gustatory system attains maximal devel- 

 opment; Conger conger where the gustatory system is feebly devel- 

 oped and the tactile system strong; and Prionotus carolinus, where 

 the gustatory system is moderately developed and the spinal tactile 

 system is greatly enlarged. 



This paper will be devoted to a description of these centers and 

 their connections in the siluroid fish, Ameiurus. The facts here 

 presented have been controlled by a study of the other types men- 

 tioned, but the detailed discussion of the comparative data will 

 be deferred. 



The spinal cord and region of the funicular nuclei of fishes have 

 never been satisfactorily analyzed, and such analysis is exceedingly 

 difficult if attention is directed solely to fishes of the ordinary type 

 with typical or average development of all of the elements. But 

 by comparing such fishes with various other types showing diverse 

 specialization of the several functional systems, a tolerably precise 

 result can be obtained without resort to the degeneration methods, 

 whose application to fishes proves to be very uncertain. 



