204 
reached, i. e. over some 15 segments, and in early stages terminates 
about the 31 trunk segment. Absent or undeveloped in early stages 
posterior to this region, it is found in older embryos extending along 
the whole length of the tail. 
3) The apparatus is largely motor, that is, in the skate examined 
the motor part was found to be invariably developed. In that its 
nerves supply apices of myotomes, i. e. segmental structures, the 
system must also be looked upon as metameric in character, although 
in other respects a segmentation is not one of its very prominent 
features. 
4) Its nerves may be either simple (spun) axis-cylinders, or com- 
pound non-medullated ones, possessing a neurilemma, and formed by 
chains of nerve-forming nuclei or cells. Cases of ,,contact between 
nerve and ganglion cell, as also between two or more of the latter are 
found to occur, but do not appear to be invariable. 
5) Ganglionic motorial end-plates, derived from wandering ganglion 
cells, and possibly occasionally from cells of the outer epithelium of 
the myotome, may occur very frequently in the latter. In later stages, 
i. e. in embryos of about 43 mm. in length, muscle is present around 
the motorial endings. 
6) These macro- or transient nerves, and other nerves also, are 
merely transformations of ganglion cells. All stages in this metamor- 
phosis can be noted !). The capsule cells arise from epiblast cells of 
the cord lying just beneath and around the macro-ganglion cells ?). 
They, too, are to be looked upon as modified ganglion cells, which 
have lost their specifically ganglionic functions *), 
7) Soon after the 45 mm. stage is reached, it becomes difficult, 
owing to the great increase in the formative tissue of the mesoblast (Bil- 
dungsgewebe of GÖTTE and ZIEGLER), to follow the transient nerves. 
I have observations, which lead to the conclusion that they become sub - 
stituted or replaced by branches springing from the anterior or 
motor spinal roots. The involution of the transient nerves would 
1) Nerve-forming nuclei always rest on the fibril, unless the latter 
is „spun“ by a ganglion cell. 
2) Horizontal sections show the ganglion cells to be unipolar, bipolar 
or multipolar, at present no rule as to their polarity can be laid down. 
3) Observation led me some 4 years ago to the conclusion that the 
otoliths of the auditory organ were products of degenerated sensory cells. 
Vide the Lepidosteus paper already cited p. 112, where the process 
is briefly described but no interpretation of its meaning given. 
