518 EMBRYOLOGY OF THE TURTLE. Part III. 
to establish its independence of fecundation within a year. But, since the self 
division of the mesoblast was not remarked as occurring in eges just about to drop 
from the ovary into the pavilion of the oviduct, it must be considered, without 
doubt, as an intra-uterme phase, commencing before the completion or even before 
the first appearance of the shell, since the latter was still quite soft and thin in 
a fecundated egg in which the segmentation of the yolk had but just separated 
a small area of the surface of the yolk into eight portions, and since it was 
altogether absent in several eggs whose embryonic disc was well marked. 
Moreover, from what we have seen in the yolk of the thin shelled ege just 
mentioned, and considering also that the segmentation of the yolk was not far 
advanced, it may be safely inferred that the selfdivision of the mesoblast begins 
before the other process. Judging from the heaps of already minute and numer- 
ous mesoblasts (Pl. 9a, fig. 55) in the segmenting mass, at the period when seg- 
mentation begins, (while but a single mesoblast existed before in each ectoblast,) 
and also from their presence all over the superficies (fig. 35a) of the ege, (their 
parent envelope, the ectoblast, having disappeared in both cases,) we may further 
say, that the selfdivision of the mesoblast is in fact a forerunner of the segmen- 
tation of the yolk, wherever this occurs, whether it be at the blastoderm, or over 
the whole surface of the yolk mass, as we shall presently attempt to show. 
As we have already mentioned, the earliest period, at which selfdivision of 
the mesoblast has been vbserved in an egg fully fecundated, belongs to that age 
when the embryonic area is divided into but eight parts, (Pl. 10, fig. 1, 2, 3,) 
and heaps of numerous mesoblasts exist, (Pl. 9a, fig. 35a,) which, we can safely 

affirm,—although we ascertained that in this instance they had lost their parent 
cell, while its presence (fig. 34, 4) around those belonging to a little older 
embryonic dise (Pl. 11, fig. 5) and to that portion of the germinal layer exte- 
rior to this was satisfactorily made out,—had evidently originated from a frequent 
repetition of that same process which at first, im the more internal portion of the 
yolk mass, simply doubled (Pl. 9a, fig. 35, 56-56c, 57-57d) the single mesoblastic 
bodies, as exhibited in the figures here referred to. In such instances, the wall 
of the ectoblast, which in the case of undivided mesoblasts is very obscure, (PI. 
Ya, fig. 35a, 39a, 39b, 39d, 39e,) was rendered very conspicuous, as it bridged over 
the constricted portion; but again became more or less indistinct where the meso- 
blastic masses had multiplied considerably, as may readily be seen in eges scarcely 
older than this (PI. 9a, fig. 7, 9, a, 38, a, 38a, 38b, a, 38c). These last eggs pre- 
sented abundant materials for the investigation of the self-division of the meso- 
blast, from its beginning, through all degrees of multiplication, until the mesoblasts 
have become very numerous. We will, however, refer at the same time to figures 
illustrating this subject im younger, and in some much older, stages of growth. 
