316 n. M. BERNARD. 



centre of fig. 18 h is not a nucleus ; it will bo referred to 

 later on in this paper (see p. 339). 



Figs. 20 c, e, f, are froui the outer nuclear layers of 

 different retinas, c of a thirteen-year-old cat, e and / of two 

 different mice. As is usual in the higher vertebrates, the rods 

 are very tliin, and their nuclei have thus to be disposed in 

 many rows. The tilaments joining together the rod-nuclei in 

 well-fixed retinas are frequently for some reason very con- 

 spicuous. The figures show the nuclei iu different phases : in 

 e the filaments radiate directly from a nearly homogeneously 

 stained nuclear mass; in c and / there is a membrane sur- 

 rounding the centrally placed chromatin. Similar figures 

 could be given from other retinas (see, for instance, fig. 27 d). 



The stellate character ol the rod-nuclei lias already been 

 noted and figured by Gustav Mann^ in the eyes of rabbits, 

 but, being on another quest, he apparently did not follow the 

 matter up and discover tliat the rays of adjacent nuclei meet 

 and foi'm connecting filaments. 



Evidence for the Existence of the Inter nuclear 

 Filaments between the Nuclei of the Inner 

 Nuclear or '^ Ganglionic Cell" Layer ; and again 

 between these and the Nuclei of the Middle 

 Layer. 



The junctions between the nuclei in this layer show tlie 

 same variations in visibility. At times no traces are to be 

 seen ; they are either ruptured or obscured by the tangle 

 of the syncytial reticulum. In other slides, liowever, they 

 come to view in a way which makes it astonishing that 

 they have not been seen before. Figs. 8 and 12 are two 

 drawings that might be multiplied indefinitely. Not only are 

 the filaments themselves visible ahnost under a dry lens, but 

 their connection with the intra-nuclear network of these 

 large vesicular nuclei is especially clear. Other details shown 

 in these figures will be referred to lower down. 



' ' Joiuii. Aiiat. and Pli^s.,' xxix (1895), p. 100. 



