Eyes of MoUuscs and Arthropods. 589 



a nerve libre, after following tlie cell wall for some distance, giving ofif at 

 irregulär iutervals smaller, lateral branclies, finally becomes so minute 

 as to disappear. Many small nerve branclies, instead of penetrating the 

 pseudo-membrane, unite with each other to form a network offibrillae,lodg- 

 ed in the circular Spaces formed by the flanged walls of the retinophorae. 



There are still other nerve fibres, surrounding the attenuated outer 

 ends of the retinophorae, which could not have originated from either 

 of these ganglionic layers. They probably arise from the angle of the 

 ganglionic nerve (PI. 29, fig. 19), and, entering the periphery of the re- 

 tina, follow the general direction of the axial nerves to the inner ends 

 of the retinophorae, where they are lost among the other nerve fibres 

 from the ganglionic layers. It is well nigh impossible to follow a 

 single fibre any distance from its source of origin, but we can say with 

 certainty that most, if not all, of the large fibres from the inner gang- 

 lionic cells extend directly inward, over the surface of the rods, while 

 others, from the same cells, extend in the opposite direction to the gang- 

 lionic nerve. The inwardly directed fibres may terminate on the wall 

 of the retinophorae, or be continued beyond the pseudo-membrane to the 

 surface of the rods. Their outwardly directed ends may vary from thick, 

 celi prolongations, terminating in many fine fibres, to attenuated pro- 

 cesses, themselves no larger than fibres, and which are then continued 

 outward, without branching, to the ganglionic nerve. 



There are stili other kinds of nerves found in the retina, which, 

 instead of originating from one of the layers of ganglionic cells, may arise 

 from the ganglionic branch of the optic nerve, and proceed directly to the 

 retinal layer, without the Intervention of a ganglionic celi (PI. 32, fig. 140). 



The fibrous layer has been described by Carrière, as a thin layer 

 ofnucleated, columnar cells ; Hensen belle ved itwas produced by shrink- 

 age and a consequent drawing out of the nerve fibres from the septum. 



Many of the inner cells of the onter, ganglionic layer present a 

 most interesting and instructive sequence of forms. Their inner ends are 

 drawn out into several fibres, while the outer end shows ali stages in 

 the transformation of a thick celi body into one or two nerve fibres. One 

 often finds parts of cells, not quite reduced to a nerve fibre, 

 but giving off lateral branches, which unite with the 

 neighboring cells (PI. 29, fig. 33, PI. 32, fig. 140 gc.^). I have 

 also found minute, inner, ganglionic cells, to whose surface was attached 

 a nerve fibre ending in a manner exactly similar to those upon the sur- 

 face of the retinophorae (fig. 140 x). The ganglionic layers, therefore, 

 contain cells in ali stages of metamorphosis, from ordinary cells, to the 



