— 154 — 
ced by a transsections of the primitive tubes of which the proto- 
plasm chiefly consists. These primitive tubes extend into the pro- 
cesses and give them a longitudinally striated appearance, the dark 
lines of the striation being the spongioplasmic walls of the primi- 
tive tubes. 
In sections of a great many, especially large, ganglion cells, 
there are in the protoplasm circular or oblong lightly staining areas 
resembling vacuoles (vide fig. 96, v) which are very similar to 
those, previously mentioned, in the ganglion cells of Homarus 
(cf p. 103 & fig. 25, 27). These areas are transsected tubes of a 
lightly staining substance, probably bundles of primitive tubes, which 
circulate in the protoplasm and run into the processes. They are 
often obliquely or longitudinally transsected and produce then 
areas of an elongated shape (vide fig. 96, v'). I have observed very 
similar tubes or bundles in the cells of the spinal ganglia of Myxine, 
as will be subsequently described, and I have also observed similar 
structures in the ganglion cells of higher vertebrates (Mammalia). 
It seems thus to be a relation of rather general nature, and it is 
therefore, in my opinion, strange that it has not previously been 
described. We know of course yet too little of these structures to 
attempt to say anything of their significance. 
The sheath enveloping the ganglion cells. — The gang- 
lion cells are always enveloped by a membran or sheath (vide fig. 
96, cm), which also extends into the process and forms their sheaths 
(cf fig. 96). This sheath is, I think, a product of the neuroglia, or 
has, at all events, the same origin as the latter. I have not, how- 
ever, hitherto been able to observe any nuclei situated in the sheath 
or adhering to it. Between this sheath and the protoplasm of the 
ganglion cells there is, generally, a cavity ^) (fig. 96, a) filled with a 
loose, apparently reticular substance, which is extremy lightly stain- 
ing; this substance is somewhat delicate, and is seen in well pre- 
served sections only. 
The pr otoplasmic processes of the ganglion cells have al- 
ways a more or less peripheral direction; they subdivide, generally 
dichotomically a great many times, and the numerous extremely 
slender, branches resulting from these divisions run, between the 
longitudinal nerve-tubes of the white substance, towards the peri- 
phery of the spinal cord, where they terminate in small thickenings. 
^) It is probably the same cavities which are described by ILey and HetziuS 
(1. c, 1876) and which these authors supposes to be filled with lymphoid fluid. 
