History of a transient nervous apparatus in certain Icbthyopsida. 379 



polar when fully developed. This is sufficiently evident from hori- 

 zontal sections as depicted in Figs. 82, 84 and 91, where the cells lie 

 in a close meshwork of their own processes. The crossing of fibres 

 from side to side is often very apparent in transverse sections as in 

 Fig. 76, plate 26. At first the ganglion- cells are quite naked and 

 destitute of capsule-cells. Such covering cells make their appearance 

 at a fairly early period, taking their origin from neighbouring cells 

 situate at the dorsal part of the cord. 



The protoplasm of the ganglion-cell is granular and does not 

 stain deeply, the nucleus is large and rounded, and apparently contains 

 several large nucleoli. 



The nerve-processes of these centrally-lying ganglion-cells will be 

 described presently. 



The portion of the apparatus at first developed, extending, as 

 described, over some 25 trunk somites, or, in other words, beginning 

 at the 6th and ending in the region of the 31st trunk-somite, is at a 

 later stage somewhat reinforced by cells arising further caudalwards. 

 These later formed cells constitute a "thin" continuation of the ap- 

 paratus right to the end of the tail. This portion of the system, so 

 far as it has been studied, appears to be but a feeble representation 

 of the first formed part ^). It would have been a great sacrifice of 



1) A little consideration seems to warrant the conclusion that to 

 the presence of transient ganglion-cells in the posterior part of the 

 trunk region beyond the yolk sac and in the tail no great morphological 

 importance should be attached. In other parts of this memoir it is 

 insisted that the transient system does not belong to the embryo- 

 proper, in other words, to the sexual generation. It is an appendage 

 of the larval or asexual generation. As indicated further on, the two 

 generations are somewhat intimately interbleuded in the development, 

 but, properly speaking, the larva coincides with the blastoderm on which 

 the embryo begins to develop. As the embryo grows in a backward 

 direction by the addition of new parts, it naturally carries with it some 

 parts of the larva, and thus these, which ought to lie entirely above 

 the yolk-sac, get more and more displaced into a region really foreign 

 to them. In the skate these portions are thereby scattered over a 

 somewhat extensive area owing to the great length of tail (which, of 

 course, began to be formed in the larval region, the blastoderm). In 

 none of the other forms to be described does the tail region ever attain 

 anything approaching the extreme length of that of the skate, and in 

 this way, as in Scyllium, a greater concentration of the transient nervous 

 apparatus in the caudal region is brought about. 



The same reasoning accounts, perhaps, for the absence of the system 



