ITS ONTOGENY, MORPHOLOGY, PHYLOGENY, AND FUNCTION. 403 



was the first to establish the fact that it is a true nerve centre. Catois (:01) fol- 

 lows Sala in his ideas of the torus. 



Johnston (:01, pp. 48, 145) has described in Acipenser certain small fusiform 

 cells "bordering on the cavity near the dorsal decussation" which he designates as 

 "Type C: cells of the torus longitudinalis Halleri," and which he believes are homol- 

 ogous with the cells of the torus described by Sala ('95). I have elsewhere shown 

 (Sargent, :03) that the cells of the "nucleus magnocellularis " of Johnston are homol- 

 ogous with Sala's torus cells. Johnston's "Type C" cells are more properly elements 

 of the tectum opticum. Johnston draws some erroneous generalizations from this 

 incorrect homology, and follows Rabl-Ri'ickhard ('87) in believing the torus to be 

 represented in higher vertebrates by the ependymal thickening. 



II. ONTOGENY. 



The torus longitudinalis is developed from the mesencephalon as a longitudinal 

 thickening of its roof, as was pointed out by Rabl-Rtickhard ('87). More exactly 

 each lateral lobe of the torus is differentiated from the mesal edge of the tectum 

 of the corresponding side. The precise mode of development, however, differs some- 

 what in the different groups of teleosts. 



In the primitive Siluridse and degenerate Amblyopsidae the median roof of the 

 mesencephalon is very thin, almost membranous. As a result of the rapid develop- 

 ment of the cells in the mesal edges of the tectum, they become differentiated from 

 it and partially constricted off from the tectum, by the formation of two longitudinal 

 grooves on either side of the median plane (Figs. 8, 9). The two incipient torus lobes 

 thus formed are widely separated and connected only by the thin median roof (Fig. 8). 

 As they increase in size they become more closely associated, but never unite except at 

 their anterior extremity. This development is earliest at the anterior end of the 

 mesencephalon and continues progressively posterior. The torus longitudinalis is 

 late in development in the Silurid<e and Ambly opsidte ; and in the adult it is of small 

 size and simple structure. In Amiurus nebulosus the torus is not developed as 

 an independent structure until the larvae have attained a length of 15 millimetres. 



In most of the other groups of teleosts where the two halves of the tectum are 

 closely appressed and the median fissure narrow or closed, the rapidly developing 

 cells of the mesal edges of the tectum crowd into the median plane and there form 

 a longitudinal thickening (Figs. 1, 2) which presses down into the median fissure 

 between the halves of the tectum (Figs. 2, 4). Usually, as in the sculpin (Fig. 5), 



