ORIGIN OF THE TEABEOULiE OP THE OPTIC NERVE. 487 



numerous^ are so comparatively narrow^ and the nerve-fibres 

 so densely packed around the nuclei, that the pull exerted on 

 a nucleus by each fibril of the trabecules^ can only result in 

 the production of a short cone. 



VII. Formation of the Pial Sheath, and its Relation to 



THE Membkana Limitans Externa. 



I have figured the formation of the pial sheath from the 

 earliest stages to show how the mesoblastic cells that enter 

 into the formation of its connective-tissue layer, gradually 

 unite with the external membrane of the stalk or its later 

 representatives — the ends of the epiblastic trabecules. 



But the union is only apparent, for a regular system of 

 lymph spaces is formed between the ends of the trabecula3 

 and the layer of connective tissue, which may be separated 

 in sectiouising (fig. 17). 



I have referred to the scantiness of its vascularity in the 

 frog in a preceding section^ p. 484. 



In rat embryos of 8 mm. in length Robinson (7) found the 

 peripheral boundary of the stalk clearly defiued_, but was 

 unable to demonstrate a distinct external limiting membrane. 



In the frog there never is any doubt about the external 

 limits of the stalk, though the boundary is naturally more 

 delicate in a 6 mm. tadpole than in those of succeeding 

 stages. 



The optic stalk of the chick, containing a very great 

 number of cells, shows a well-defined external limitinar 

 membrane, supported by numerous mesoblastic cells, when 

 the nerve-fibres begin to grow in, on the fourth day of 

 incubation. 



VIII. Some Variations op Epiblastic Trabecul.e met with in 

 THE Developing Optic Nerve of the Mouse, the Trout, 

 the Dog-fish, and the Chick. 



Although I have selected the tadpole for tracing the com- 

 plete development of the epiblastic trabeculas, still we can 



