ITS ONTOGENY, MORPHOLOGY, PHYLOGENY, AND FUNCTION. 411 



difference is not easy to explain, but I would at least suggest that there is no constant 

 relation between the size of the cell and the length of its axon, as has been supposed. 

 Rabl-Ruckhard ('94, p. 703) described in reptiles a peculiar growth of ependyma 

 ("Ependymwucherung"), under the posterior commissure, horseshoe-shaped in 

 transverse section. It had previously been noticed by Herrick ('93, p. 97) as follows: 

 " The posterior commissure is underlaid by a remarkable thickening of the epithelium 

 that can be traced far ventrad." Rabl-Ruckhard ('87) had found this occurring 

 in other groups of vertebrates, amphibia, reptiles, and birds, and believed it to repre- 

 sent a rudiment of the torus longitudinalis. This erroneous interpretation has been 

 accepted by Gage ('93), Johnston (:01, :03), and others, and has led to considerable 

 confusion. Aside from the fact that the torus is the homologue of the "Dachkern," 

 this homology fails to hold, as I have elsewhere shown (Sargent, :03 a ), for (1) this 

 ependymal thickening lies morphologically cephalad of the commissure posterior 

 (Figs. 15, 24, eras, e'end.), while the torus is caudad of it; (2) the ependymal thick- 

 ening occurs in a modified form in teleosts. In a later paper I shall treat of the inter- 

 esting relations of this ependymal thickening more fully and present a theory as to 

 its function. 



VII. PHYLOGENY. 



The nidulus of cells which gives rise to Reissner's fibre is one of the most archaic 

 elements of the vertebrate brain, occurring in the Cyclostomes and in a rudimentary 

 way in Amphioxus. The torus longitudinalis as an independent structure, however, 

 has its beginnings phylogenetically in the ganoids. With the increasing complexity 

 of the adult ganoid brain, and the development of the dorsal decussation of the 

 tectum, the cells of the "nucleus magnocellularis " are crowded downward so as to 

 form two longitudinal ridges on either side of the median plane extending backward 

 for a short distance from the posterior commissure. These ridges are scarcely apparent 

 in Amia, but in adult Lepidosteus and Acipenser they are quite pronounced. 



The Siluridse naturally supply the necessary connection, in the phylogeny of the 

 torus, between the primitive condition in ganoids and the highly differentiated 

 state in the higher teleosts. The torus of Amiurus is intermediate both in its form 

 (Fig. 8) and in the large size of its cells. In Amblyopsis the torus as a result of 

 arrested development has reverted to the primitive ganoid condition (Fig. 9). In the 

 Salmonidse and related forms the torus is still of small size and its cells relatively 

 large (Fig. 3). The climax of teleostean aberrancy is reached in the Scisenidae, where 

 the torus becomes of enormous size (Fig. 18) and its cells very small. 



