28 H. M. BERNARD. 



reticular layer towards the membrana limitans externa. The 

 only possible escapes from the assumption that when the 

 original supply in the outer nuclear layer has been exhausted, 

 fresh supplies migrate outwards from the middle nuclear 

 layer are two : (1) if it could be shown that the layer of rods 

 and cones with their nuclei grows only at the edges ; and (2) 

 if it could be shown that the nuclei for the new rods are pro- 

 duced by the division of those already composing, the outer 

 nuclear layer. 



With regard to the former of these alternatives, it is 

 certainly true that the retina as a whole does grow mainly at 

 the edges. I am not, indeed, now inclined to lay very much 

 stress upon the argument used in Part J, that if growth 

 took place only at the sides, the eye would not keep its 

 shape, for growth at the sides alone would, I thoughf, 

 merely carry up those sides, and the eyes would be funnel- 

 shaped rather than round cup-shaped. This argument wonld 

 perhaps hold if no other factors were present which could 

 help to keep the retina hemispherical. There is, however, 

 another traceable factor, the full force of which I did not 

 then see. I refer to the vitreous humour which, as a collec- 

 tion of semi-fluid matter in the hollow of the eye, Avould, if 

 the supply is kept up at any pressure, compel its flexible 

 walls to adopt the normal shape. But though this is a 

 possible factor in keeping the growing retina round, the 

 argument which refers the persistence of its shape to its own 

 growth-processes can hardly be put altogether on one side, 

 for it is a fact that cones, i.e. new rods, can be seen forming 

 over the whole Amphibian retina at all stages of its 

 growth, and even in the eyes of adults. 



Secondly, the suggestion that the nuclei for new rods 

 might be supplied by the division of those already present 

 can be met by a decided negative. In embryonic eyes 

 (Mammalia), or in amphibian eyes before they are functional 

 (see fig. 5), i. e. before any rods are formed, and only the 

 merest traces of vesicular pi'otrnsions are to be seen, 

 divisions of nuclei occur over the whole retina in the outer- 



