420 Journal of Comparative Neurology. 



In the smooth dog fish, Miistehis canis, the proximal rela- 

 tions are much clearer (Fig. 3). Here the acustico-lateral, gen- 

 eral cutaneous and communis ganglia of the trigemino-facial 

 complex can be quite sharply analyzed by simple dissection. 

 The acustico-lateral system lies dorsally to the others so as to 

 obscure the relations somewhat, but the figure shows the roots 

 and ganglia in their normal relations save that the general cuta- 

 neous branch, cut. VII, for the hyomandibular trunk has been 

 displaced laterally to expose the geniculate, or communis gan- 

 glion of the facialis. The pre-trematic ramus and chorda tym- 

 pani are seen to arise together from this ganglion independently 

 of the palatine ramus. Mustelus, it will be recalled, lacks a 

 spiracular pseudobranch, but the pre-trematic ramus, after giv- 

 ing off several twigs for the mucosa of the anterior wall of the 

 spiracle, terminates in the exact region where it would be if it 

 were present. The chorda curves around the anterior wall of the 

 spiracle to the point of divarication of hyoid and mandibular 

 arches just as in Carcharis. 



It will also be noted in this dissection that the geniculate 

 ganglion clearly sends a branch backward into the hyomandib- 

 ular trunk, showing that here, as in Squalus acanthias, this 

 nerve contains visceral sensory fibers in addition to the lateralis 

 and motor components, while the branch, cut. VII, referred to 

 above, seems to carry general cutaneous fibers from the Gas- 

 serian ganglion into it also. Whether the geniculate ganglion 

 contributes to the maxillary or mandibular rami of the trigem- 

 inus could not be determined. 



The spiracular nerves of the common skate. Raja erinacea, 

 and of Torpedo occellata, were also examined, but in these 

 cases the nerves corresponding to what we have termed the 

 chorda were less perfectly developed than in the other species 

 and they could not be traced far down toward the ventral 

 surface. 



It is obvious that the facialis of fishes departs from the 

 typical arrangement of branchial nerves by reason of the loss 

 or reduction of the hemibranchs primitively innervated by its 

 post- and pre-trematic branches. The post-trematic, or hyo- 



J 



