Cap. II. TRANSFORMATIONS OF THE YOLK. 519 
At first the mesoblast constricts gently, so as to leave a broad sinus between 
the separating portions, (Pl. 9a, fig. 37, 40,) then a little later its constriction 
becomes narrower as it grows deeper, (fig. 40a, 40b, 40c, 40d,) till finally it 
divides into two (Pl. 9a, fig. 56-36c, 37a—57d, 38, 40e, 40f, 40g); then each of 
these again doubles itself Sometimes the one division begins before the other, so 
that there is a triple mesoblast (fig. 40h, 401); or, in the case of the next phase, 
one of four doubles previous to the others, (fig. 58b, 40j,) thus producing a quin- 
tuple mesoblast; and so on, again and again, almost to infinity, we might say, when 
we consider the innumerable quantity of these bodies (PI. 9a, fig. 9, a, 4) in each 
ectoblast, at the time they have just entered the boundaries of the embryonic disc 
and become part and parcel of the embryo. In this connection we may express 
the opinion, that it is very probable that the entoblast also segments, since in 
some instances (Pl. 9a, fig. 40d, ¢) it is so large as to make it almost impossible 
that it should enter entire into either of the two portions of a duplicated mesoblast. 
In several other cases the entoblast has been observed in the same position (fig. 40, 
40a, c, 40c, ¢) as in the former, directly in the line of the approaching constric- 
tion; but im these cells it might, considering its size, be forced into either seg- 
ment by the narrowing strangulation. 
We have good reason to believe, that the phenomenon of self-division of the 
mesoblast obtains throughout the period of incubation of the animal; at least, it was 
observed in an ege two months old, (Pl. 9a, fig. 40-40b,) which is half the time 
required to develop the Turtle; and, on the day the animal left the shell, the 
still pendent yolk sac contained cells of the largest size, having each but a sin- 
gle undivided mesoblast (Pl. 18, fig. 4a); so that, at best, it cannot be said that 
the self-division of the mesoblast had pervaded the whole yolk at the time of 
hatching. Again: this process does not go on uniformly throughout the yolk at 
one time, but has a centripetal aim, provision for the embryonic dise (Pl. 9a, fig. 
7, 35) and the germinal layer (fig. 5, 35a) all round the yolk sphere, being first 
made, previous to segmentation; then the next more interior portions become the 
seat of action, and so on, deeper and deeper.’ 
By referring to cells (Pl. 9a, fi 
and from the germinal layer (fig. 3) after segmentation, we may gain abundant 
e. 7, Ta, 9, a, 6) taken from the embryonic disc 
evidence, that, even at this period, the selfdivision of the mesoblast has not fin- 
ished its part. This may be confirmed by resorting to an embryonic disc a little 
older, where the cephalic hood is just about to be formed, the depression all 
1 Such cells were observed from the centre of the to be quite numerous; those, divided into two being 
yolk mass, from the surface, and midway between by far the most frequent. Both the mesoblast and 
these points, and found everywhere alike, but in entoblast, throughout the yolk mass, are very faintly 
greater proportion near the surface, where they seemed yellow. 
