ON THE ORIGIN OF VERTEBRATES. 245 



Parker's argument, another argument based on the pigmentation 

 points to the same conclusion ; in the adult eye the retina 

 proper, with its rhabdomes and retinulas, is strongly pigmented, 

 and the pigment extends slightly into the layer of terminal 

 nerve fibres, so that the limit of the true retina is defined in a 

 general way by the limit of pigment. In the stage reached in 

 Eeichenbach's fig. 224, the pigment is well marked, as is shown 

 in his fig. 225, and does not extend into the outer wall of the 

 optic fold as defined by lleichenbach, from w^hich T would con- 

 clude with Parker that this outer wall does not give rise to the 

 rhabdomes and retinulse, but is a part of the retinal ganglion in 

 close connection with the visual rod layer, the double band of 

 large, lightly-staining cells separating two parts of the retinal 

 ganglion. In the adult retina, according to Berger, the nuclear 

 layer of the retinal ganglion is distinctly divided into two parts, 

 n.l.r.g.o. and n.l.r.g.i., (fig. 5), and it is possible that this separation 

 is the outcome in the adult eye of the division so apparent in the 

 younger stage. I reproduce here Berger's picture of the retina of 

 Astacus, but have omitted the blood-vessels. It is impossible to 

 speak with certainty on this point until the fate of this peculiar 

 double band of large cells so vividly depicted in Eeichenbach's 

 paper has been decided. It is, however, very suggestive that 

 according to Parker's latest paper there is a difference of opinion 

 as to the nature of the cells in these two layers (n.l.r.g.o. and 

 n.l.r.g.i.) in Astacus, of the same character as already described 

 in the case of the two lavers in Ammocoetes ; thus Parker^ 

 states that " in the first optic ganglion close to the retina two 

 rows of cells are found, separated from each other by a 

 fibrous layer. Of these two layers the distal nuclear layer 

 (n.l.r.g.o. fig. 5) consists of a number of large oval or roundish 

 nuclei imbedded in what is apparently an irregularly fibrous 

 mass." He says that Viallanes in Palinurus considered them 

 to be true ganglionic nuclei, while Berger in Squilla and 

 Grenacher in Mysis suggest that they are the nuclei of support- 

 ing tissue cells. In Golgi preparations " the retinal fibres can 

 be seen passing between these nuclei without entering into 

 definite relations with them. The fact that these cells do not 

 send processes into the Punktsubstanz, as well as their lack of 



^Op. elf., 1895, p. 40. 



