STUDIES IN THE EETJNA. 307 



the lower forms^ e. g. in the Amphibia^ most of the ganglionic 

 nuclei either have no cytoplasm at all, and are simply nuclei, 

 somehow suspended in larger or smaller meshes of the retinal 

 syncytium, or else they have a mere trace of it, usually on 

 their distal sides, and trailing away into the inner reticular 

 layer (see, for instance, figs. 2 a, 3 d, 18 d, the top nucleus). In 

 every case in which cytoplasm does occur in connection with 

 these nuclei it is impossible that it could be overlooked ; 

 whatever its real constitution may be, it appears under the 

 mici'oscope as a refractive granular mass, and I shall always 

 refer to it as granular cytoplasm to distinguish it from the 

 smooth, glassy strands of the syncytial supporting framework 

 of the retina. In the vertebrate retina, then, the granular 

 cytoplasm, so essential a factor in the old diagram, is found 

 associated with the ganglionic nucleus in very varying quan- 

 tities ; it shows almost every stage, from the condition in 

 which it renders a reticular chamber in the retinal syncytium 

 turgid, as represented in the diagram, to that in which it is 

 present in the merest traces, or even completely absent. Figs. 

 1 to 3 and 5 to 12 represent ganglionic " cells " of different 

 vertebrates; such figures might be multiplied indefinitely; 

 those here given are but a very small selection from the 

 number I have drawn under the microscope. Figs. 2 b, c, 

 3 a, h, c, and 8 show ganglionic nuclei without any cytoplasm, 

 while, in fig. 12, three nuclei have a certain amount of cyto- 

 plasm and the rest not a trace. 



Almost at once, then, we are called upon to modify our 

 conception of the retinal " ganglionic cells," as obtained from 

 the familiar diagram, and we must describe them as the inner- 

 most layer of the retinal nuclei, which may or may not have a 

 certain quantity of associated cytoplasm. Thus it is obvious 

 that (1) in tliose cases in which there is no cytoplasm there 

 can be no axis-cylinder process such as that required by the 

 diagram; (2) in those cases in which the cytoplasm is entirely 

 confiued to the distal side of the nucleus, the diagrammatic 

 cytoplasmic axis-cylinder process, continuous with a fibre of 

 the optic nerve, is equally out of the question; and (3) even in 



