202 
studying these macro-ganglion cells and their nerves in Raja, it is 
impossible to conclude otherwise than that nerve-forming nuclei or 
cells are transformed ganglion cells. 
This matter is one about which I would say more elsewhere. 
Before leaving it, let me add that I find myself unable, after all that 
I have seen while investigating this apparatus, to support Donkrn’s 
view that a nerve-fibril would never appear to be a process of a ganglion 
cell. As stated above I have observed many cases of what was termed 
the spinning of axis-cylinders by the ganglion cells themselves. On 
the other hand, when a fibril has attained certain dimensions, or when 
other reasons come into play, it would appear to be undoubtedly true 
that the axis-cylinder for the greater part, or even for the whole of 
its length, may be developed apart from the ganglion cell, and only 
applied to the latter. Of contact between axis-cylinders and ganglion 
cells plenty of cases may be observed in Raja, even among these 
larval ganglion cells. My attention was directed to this question by 
the researches of Rerzius and others, long before the appearance of 
Dourn’s memoir (Studie 17). 
Fig. 8. Transverse 
section from the same 
embryo as fig. 7. (Com- 
bination drawing. i. e. re- 
construction.) 
In their essentials the last two figures (7 and 8) bear great re- 
semblance to figs. 3 and 5. They are given as typical of the motor 
portion of the larval apparatus in its highest development as met 
with in an embryo of about 43 mm. Both figures are reconstructions 
of several transverse sections. In the first of the two (fig. 7) three 
large ganglion cells (gl. c) lie in the dorsal portion of the cord, and 
from them a bundle of three non-medullated axis-cylinders (ma. n.), 
with numerous applied nerve-forming nuclei, passes to the myotome. 
The nerve „trunk“ thus formed perforates the apex of the latter and 
runs, in close application thereto, along its outer surface. A more 
definite form of termination was not made out in this particular in- 
