— 159 — 
I am, however, at all events, entitled to say that, the nerve-tubes 
of the dorsal (posterior) nerve-roots have no direct origin from gang- 
lion cells. 
The nerv e-tub es of the ventral (anterior) nerve-roots 
do not sub divide after they have entered the spinal nerve- 
cprd, according to my observations. I have been able lo trace 
them for considerable distances into the nerve-cord, in chromo-silver 
stained preparations (vide fig. 105) as well as in carmine and 
hæmatoxylin preparations (vide fig. 93 and 95) but no subdivisions 
have been discovered, though the tubes are thick and easy to ob- 
serve; slender side-branches are, however, given oft' (vide fig. 93, w&r; 
105, sbr). 
In a few cases only have I been able to trace their direct com- 
bination with ganglion cells (vide fig. 106) and the tubes which in 
these cases had a direct combination were comparatively very thin. 
From what is seen, there can, however, be no objection to the 
assumption, that all the nerve-tubes of the ventral nerve-roots are 
directly connected with ganglion cells, indeed, I think this is all but 
certain, seeing that these tubes have no subdivisions; but it is pro- 
bable that the thick ones pass for some distance through the white 
and grey substance before they united with the cells. A great 
many of them seem to pass over into the other side of the spinal 
cord Crossing the ventral transverse commissure. 
I will not omit to mention a section which I once obtained 
of a chromo-silver stained preparation. A part of this section is 
illustrated in fig. 102 which is drawn under the camera lucida and 
is an exact representation of what was seen in the preparation. 
The nervous process [mpri) or perhaps »mixed process« (cf above) 
issuing from the ganglion cell gc, is seen to give off a thick branch 
(&rj), which passes towards the ventral (anterior) nerve-root (vnr). 
This branch becomes very thin towards the periphery of the white 
substance and can not be traced entirely into the ventral nerve-root, 
though there only wants an extremely short distance. Whether 
this branch was really connected with the ventral nerve-root (was 
transformed into a peripheral nerve-tube), and it was only owing to 
an incomplete staining that it could not be traced into the nerve- 
root was very difficult to decide. In its whole appearance this branch 
was not at all like common protoplasmic branches, which I certainly 
very often have observed to be directed towards the nerve-roots 
(vide fig. io<^,ppr\ log, ppr] 93, gc^), but it had that smooth and black 
staining which is caracteristic for nervous processes. What is strange 
