— 147 — 
But we must helieve in it We need, for instance, only go to the 
nervous systems of the lobster, and we shall be quite convinced of 
the existence of real unipolar cells. If those large ganglion cells of 
the lobster have several processes, it is strange that we should not 
be able to observe one of them by our present modes of investigation. 
We are thus obliged to reckon with real unipolar cells, indeed 
we are obliged to reckon with nervous systems containing almost 
entirely unipolar ganglion cells, and how could direct combination 
between the cells be present where there are no processes to pro- 
duce the combination. We thus see that also for that reason we 
are obliged to abandon the theory of the direct combination between 
the ganglion cells. 
Two questions which will now present themselves are : 
How is the combination between the ganglion cells produced? 
and what is the function of those processes which were previously 
thought to produce the direct combination between the cells 
We will first examine the latter question. As before described, 
the processes of the ganglion cells are of two kinds: nervous pro- 
cesses and protoplasmic processes. Of nervous processes each cell 
(unipolar or multipolar) has always one and never more ; the nervous 
processes are always directed towards the dotted substance, or in 
a few cases they may pass directly into peripheral nerves (cf p. 143, 
note i). It was not, however, these processes which were gener- 
ally supposed to produce the combination between the cells, but it 
was the protoplasmic processes. 
The protoplasmic processes. — When a ganglion cell is 
bipolar or multipolar, then the processes it posesses, besides the 
nervous process, are protoplasmic processes. These protoplasmic 
processes are not directed towards the dotted substance, but gener- 
ally have a peripheral direction towards the external layers of the 
central nervous system. In a great many cases I have been able 
to trace such processes just to their communication with the peri- 
according to the reality, some people can, at all events; as the mountain would 
not walk to Muhamed, Muhamed had to walk to the mountain. In chapter 7 we 
will have an opportunit)"- to discuss this subject somewhat more circumstantially. 
(Regarding unipolar cells etc. vide also my quotation from Rawitz's paper 
p. 66.) 
As subsequently may be seen, it is, in my opinion, of no importance as to 
the combination of the cells whether they are unipolar or not. I do not at all 
doubt the existence of the slender processes Fritsch describes in his ganglion 
cells from the ganglion gasseri of Lophins, but I suppose them to be proto- 
plasmic processes, i. e. nutritive processes. 
10* 
