STUDIES IN THE KETINA. 331 



among the strands of the inner reticular layer may perhaps 

 give rise to some of the "dendrites " of these " cells/' such as 

 are revealed by Golgi's method. The streams of pigmented 

 matter such as I figured and described in Part V, are, how- 

 ever, I believe, the explanation of most of them. In neither 

 case, whether as streams of cytoplasm or of pigmented 

 matter, do they supply us with structures of any morpho- 

 logical value. The forms assumed are mere passing physio- 

 logical phases, except, perhaps, in old eyes, in which the 

 Miiller's fibres may have acquired some stability (see Part V, 

 PL 5, fig. 32 a). 



All the appearances are as if the granular cytoplasm were 

 purely nutritive and had only a passing association with the 

 nuclei. The observations I have been able to make point in this 

 direction. A number of salamanders accidentally neglected 

 for nearly six months were found alive but emaciated. Not 

 the faintest trace of the cytoplasm could be found associated 

 with any of the nuclei of the innermost layer in any one of 

 them. This observation, however, is not very conclusive, 

 because very little indeed can be found in normally fed 

 salamanders, or, indeed, in the Amphibia as a rule. My 

 experience is that it is rare to find a nucleus actually enve- 

 loped in cytoplasm in this group, and that anything even 

 approaching the typical diagrammatic ganglionic cell must be 

 looked for among higher vertebrates ; they seem to be most 

 common in the higher mammals, especially in man. Hence 

 it is of significance to note that in several sets of sections of 

 the retinas of mice which had been tired out by long (three 

 hours) exposure to the light of an arc lamp, every trace of 

 granular cytoplasm had vanished, while, in other retinas from 

 wild, albino, and other mice, which had not been put to this 

 strain, most of the ganglionic nuclei had at least a tongue of 

 this matter trailing off into the inner reticular layer, though 

 very few were completely surrounded by it. Lastly, in the 

 sections of the human retina to which we have frequently 

 referred above, the majority of the nuclei of the innermost 

 layer, at least near the yellow spot, are nearly, if not quite. 



