2^10 
EDWIN S. GOODRICH. 
But the interpretation of the occipital region is complicated 
by the well-established fact that one or more myotomes 
immediately behind the auditory capsule may undergo de- 
generation during ontogeny. These are the myotomes 
corresponding to the glosso-pharyngeal, the first, and the 
subsequent vagus dorsal roots, and supplied by the ventral 
spino-occipital roots. They tend to disappear in development 
from before backwarks, and their nerves go with them. It 
might be thought that here at last we have a case of excala- 
tion.^ Indeed, Furbringer himself, if I understand him rightly, 
seems to believe that the most posterior spino-occipital nerve, 
z, shifts forwards by the disappearance in phylogeny of the 
more anterior roots s, t, u, etc. 
The process of ontogenetic degeneration of myotomes in the 
occipital region must not be confused with excalation. The 
glosso-pharyngeal supplying the first branchial slit, and the 
first root of the vagus supplying the second slit, form a well- 
differentiated nerve complex which can be homologised 
throughout the craniate vertebrates. If the shifting of the 
hind limit of the head is to be attributed to the disappearance 
of segments behind the auditory capsule, then it is clear that 
(unless we sacrifice the homology of the glosso-pharyngeal and 
first vagus and their slits for the sake of saving that of the 
occipital segment) we must suppose that the segments have 
vanished, not at the anterior end of the metaotic series, but 
in the middle region — that is to say, between the first vagus 
root and the condyle. This supposition is directly against the 
evidence of embryology, which plainly shows that the more 
anterior segments are more degenerate than the posterior. 
AVe should also be compelled to assume that the change of 
position of the occipital condyle, being due to the disappearance 
of segments, could take place in one direction only — towards 
the anterior end. It would follow that the forms with an 
occipital region containing few segments have always been 
derived from those with many, which is in the highest degree 
improbable. 
‘ See footnote, p. 241. 
