4 GEORGE L. STREETER 
30 mm. It is also well known that the remodeling which takes 
place in the gill region completes the obliteration of the gill bars 
before the embryo is 20 mm. long. One well might expect these 
processes of dedifferentiation and redifferentiation to be more 
active in the earlier stages. They are not confined, however, to 
- this period, for in the case of the ear capsule they were found to 
be very active throughout fetal life. In the case of the spinal 
cord dedifferentiation is well demonstrated in the period repre- 
sented by embryos between 11 and 30 mm. long. A comparison 
of these two stages can be made in figure 1. It will be noted in 
the first place that the spinal ganglia show a regression varying 
from arrest in development to complete disappearance. All but 
two of the coceygeal ganglia have disappeared in the 30-mm. 
specimen, and the remaining two are of about the same size as — 
the same two ganglia in the 11.5-mm. specimen.’ 
As for the cord itself, the changes are equally marked. In the 
younger stage (embryo 11.5 mm. long) the extreme caudal end 
of the spinal cord, the part belonging to the non-vertebrated tail, 
shows little differentiation, consisting only of indifferent cells 
resembling embryonic ependyma. In the coccygeal region, how- 
ever, the development is more advanced. Opposite the five 
coccygeal ganglia the wall of the cord is differentiated into dis- 
tinct ependymal, mantle, and marginal zones, with well-developed 
rootlets entering into it from the first two ganglia. Sections 
through it show nothing to indicate that this region is not going 
on to complete its differentiation into the adult condition. When, 
for comparison, one examines the very same region in the older 
specimens (fig. 1, embryo 30 mm. long) it is found that its con+ 
dition, relative to the remainder of the cord, has undergone a 
marked change. While the precoccygeal cord has continued to 
increase in the thickness of its walls and in the elaboration of the 
mantle and marginal zones, the coccygeal region is less advanced 
3 Throughout this paper the twenty-fifth to the twenty-ninth segments have 
been uniformly regarded as sacral. The slight variation which is known to exist 
in this respect is too small to be taken into account in our genéral conclusions, 
and for convenience the regional terms, lumbar, sacral, and coccygeal, will be 
used, upon the assumption that the specimen concerned has the usual regional 
distribution of its segments. 
