STUDIES IN THE RETINA. 329 



tion for this persistence of the ectoplasm, I may note that the 

 delicate membranous walls of the rods can now and then actually 

 be seen to develop from the radial cytoplasmic attachments 

 of the spindle-shaped cells above mentioned found at the rim 

 of the retina (see figs. 23 and 24). Even though we have 

 no direct evidence, it is a natural assumption that these 

 attachments are parts of the walls of the original columnar 

 cells, and therefore ectoplasraic. 



If this be so — that is, if the retinal framework is actually 

 composed of the more resistent outer layers (ectoplasm) of the 

 original row of single columnar cells which persists in the par s 

 ciliaris retinae, — then, whenever a chamber of the retinal 

 ectoplasmic framework containing a nucleus also fills with 

 granular cytoplasm, there results a kind of secondary "cell" 

 formation. At any rate, it is quite possible that the question 

 as to what is a "celP' and what is a syncytium may 

 ultimately turn upon the arrangement of the hyaline ecto- 

 plasmic substance. 



The syncytial framework, though much more stable than 

 the granular cytoplasm, is nevertheless subject to great 

 modifications; it is sometimes abundant, at others very 

 scarce. And here, in passing, we may note a second function 

 performed by the syncytial framework. In addition to being 

 a support, its strands and membranes form paths along 

 which streams of pigmentary matter, absorbed during the 

 functional activity of the rods, escape through the retina. 

 When these streams of matter are abundant and of old 

 standing, they seem to effect a rearrangement of parts of the 

 usually tangeutially arranged syncytial strands, so as to form 

 striking radial structures, the familiar Muller's fibres (see 

 Part V). 



A variable unstable syncytial framework supporting a 

 continuous protomitomic system presents no difficulty. It 

 leads on naturally to the more stable frameworks which 

 culminate in cartilage and bone. I may as well here call 

 attention to the fact that, just as the protomitomic fila- 

 ments must pass through the membranes of the retinal 



