— 148 — 
neurium enveloping the nervous system (e. g. vide fig. 50, ppr). 
In Patella I have even found ganglion celis with similar processes 
in peripheral nerves (e. g. vide fig. 20, gc, the nervous process 
of this cell is not distinctly seen in the illustration), Whether the 
protoplasmic processes, as a rule, extend to the perineurium, like 
what is the case in the spinal cord of Myxine (as will be sub- 
sequently described) I do not know, as they are generally lost in 
the neuroglia reticulation extending between the ganglion cells (e. g. 
vide fig. 47, 50, 71, 73). In the brain of th.^ Ascidia, as well as in 
the nervous system of Patella, I have found bipolar and multipolar 
ganglion cells situated in the dotted substance, and sending their 
protoplasmic processes towards the periphery of the nervous system 
Like Pr of GoLGi I belive the function of the protoplasmic pro- 
cesses to be a mitritive one; when the ganglion cells can not get 
sufficient nutrition in their neigbourhood, they have to send processes 
towards the periphery of the nervous system, or out into the loose 
neuroglia reticulation, where there is sufficient nutritive fluid for the 
processes to absorb.') This is the reason why the protoplasmic pro- 
cesses have generally a peripheral direction, and why they chiefly 
occur in cells which are somewhat deeply situated in the nervous 
system, i. e. at some distance from the periphery. The ganglion 
cells situated near the periphery do not need any special organs to 
absorb their nutrition from the fluid surrounding them, neither do 
the ganglion cells situated in a loose neuroglia-reticulation need any 
(compare the ganglion cells of the lobster). 
Though we have of course no proof of the real function of 
the protoplasmic processes, I think this theory of their nutritive 
function is very probable, and explains a great deal which, else, 
may seem inexplicable. What other function could we suppose them 
to have when they terminate under the perineurium, what they 
actually do, at all events, in a great many cases .^^ — Indeed, I do 
not, at present, see any acceptable explanation besides this one 
suggested by GoLGI. Subsequently, we will have an opportunity to 
state some other facts which will still further add to its probability. 
The comh in at ion of the ganglion cells. — Having thus 
elucidated that no combination between the cells can be produced 
by the protoplasmic processes, and as it is very improbable that 
any combination between them can be produced by the neuro- 
glia, I suppose it to be most probable, if not certain, that if any 
^) Where there are blood-vessels in the central nerve-system (e. g. in the 
Vertebrata) the protoplasmic processes can often be seen to be directed towards 
the connective tissue surrounding the blood-vessels. 
