- 156 - 
to affirm that a direct combination behveen the ganglion cells, by direci 
anastomosis of the protoplasmic processes does not exist. 
A thing which I take to be rather puzzling is, that processes 
which in their aspect can not be distinguished from common pro- 
toplasmic processes, are often seen to pass towards the periphery 
of the nerve-cord ; arrived there, they do not, however, terminate, 
but run along the periphery and partly back into the white sub- 
stance (vide fig. 102, 2^r, br^) I have not found the terminations of such 
processes. Some other processes, with quite similar aspect, may 
be traced for very long distances, even crossing the ventral com- 
missures and passing over into the other side of the spinal nerve- 
cord (vide fig. 102, br^). A great many processes which I have 
observed to have a similar course are, near their origin in the celi, 
very thick and large, and the}^ give off branches which terminate 
in the periphery, under the sheath, quite in the same way as 
common protoplasmic processes, whilst at the same time, the main 
stem of the process runs along, and takes a course as above de- 
scribed (vide fig. 93, mp and 94, 7np). It has been possible to trace 
similar processes not only in chromo-silver stained preparations but 
I have also traced them. for considerable distances in preparations 
stained with carmine or Heidenhains hæmatoxylin only (vide fig. 
93 and 95). If, now, these processes are protoplasmic processes, 
and if the protoplasmic processes have a nutritive function, which 
makes it necessary for them to terminate under the sheath enve- 
loping the spinal cord, it is, indeed, very difficult to understand why 
they pass through the nervous system for so long distances without 
terminating; why they pass to the other side; and why they do not 
terminate, but return again, when they are in the periphery.^ These 
are questions of so serious a nature that they oblige us to assume 
that either these processes are not wholly protoplasmic processes 
though they have their aspect, or that the function of the proto- 
plasmic processes is not only a nutritive one. I will not decide 
between these alternatives at present, though I find the former one 
most probable. It can not be doubted, in my opinion, that some 
anastomosis is really present. If there is an anastomosis, however, it may be ac- 
counted for as a remnant of a cell-division not yet finished, as the bridge between 
the cells is very thick and short. 
Mr. W. S. Colman (1. c. 1884) has actually observed anastomoses of gangUon 
cells in the spinal cord of a human foetus; but there is, in my opinion, little doubt 
but that they are results cf divisions of ganglion cells; in two of the cells there 
were even two nuclei, what obviously indicates that the cells are dividing. 
