— 208 — 



all the mail-cheeked fishes examined, excepting Cottus, and there appears as a jugular groove on the 

 outer surface of the prepared skull. In Cottus the outer wall of the space is of bone and the inner 

 wall largely of membrane, the space there appearing as a recess on the inner surface of the prepared 

 skull. The outer wall of the space, primarily of membrane, is invaded to a different extent, in different 

 fishes, by the neighbouring bones, thus giving rise to greatly varying conditions. 



16. The myodome of fishes is primarily a subpituitary space that is connected with the 

 orbit of either side, or with the orbital region, by a canal that transmits the pituitary vein. 

 Secondarily this subpituitary space acquires a wide communication with the orbits, the primal 

 cause of this secondarily acquired communication apparently being a deepening of the hind ends 

 of the orbits due to a marked enlargement of the eyeballs. Following this deepening of the 

 orbits, certain of the eye-muscles of either side enter the pituitary canal and, enlarging that 

 canal, finally wholly break down the wall that separates the orbits from the subpituitary space. 



It seems possible that the subpituitary space may represent the conical depression on the 

 anterior surface of the body of a vertebra, the later acquired, posterior, or basioccipital extension 

 of the myodome then possibly being due to the assimilation of similar depressions in more 

 posterior vertebrae. 



17. The peripheral course of each of the cranial nerv es is so constant that it would seem 

 to be of greater value for the determination of the segmental position of the nerve than the 

 apparent centers of origin of the fibers composing the nerve; it seeming more reasonable to 

 postulate varying composition, and condensations of ganglia or of centers of origin, than variations 

 in peripheral course. This being so, a nomenclature based on peripheral course seems much 

 preferable to one based on the character of the component fibers of the nerve or on the apparent 

 centers of origin of those fibers. 



18. In all the Loricati examined, the ramus palatinus facialis either perforates the proötic 

 bridge or adjoining portions of the side wall of the proötic, and so passes from the cranial cavity 

 directly into the myodome. In all of the Craniomi examined this nerve first passes from the cranial 

 cavity into the trigemino-facialis Chamber and then traverses that Chamber to issue by its trigeminus 

 opening and so enter the myodome. 



19. In all of the mail-cheeked fishes examined, excepting Dactylopterus, the nervus abducens, 

 in passing from the cranial cavity to the muscle it innervates, either passes over the anterior 

 edge of the postpituitary portion of the proötic bridge or perforates that bridge near its anterior 

 edge. In Dactylopterus the nerve first passes from the cranial cavity into the trigemino-facialis 

 Chamber and then issues from that Chamber through its trigeminus opening to reach the muscle 

 it innervates. In Lepidosteus also the nerve has this latter course. 



20. Cottus differs markedly, in many more or less important anatomical features, from the 

 other Loricati, and Dactylopterus differs still more markedly from the other craniomi; and if 

 craniological characteristics are of any especial value in phylogenetic determinations, it would 

 seem as if these two fishes could not be closely related to the other fishes of the mail- 

 cheeked group. 



Palais de Carnoles, Menton, France. 

 February 27th. 1907. 



