69 
runs backwards and downwards, and in front of it the out- 
line of the body is rounded and bears the velum. The time 
when the valves of the shell make their appearance "varies 
slightly, but I have seen them when the transverse groove 
was as well marked as it is in Figure 35, and this fact/ as well 
as the similarity in the outline of Figures 32 and 36, does hot 
seem to leave room to doubt that the shell occupies the posi- 
tion of the blastopore. At the stage shown in Figure 36, the 
endoderm is a pretty compact mass, separated, around its en- 
tire circumference from the body wall, and with no traces of 
a central cavity, although it is perfectly possible that a small 
cavity may be present. 
Figure 37 is an embryo a few hours older, viewed from the 
right side, with its dorsal surface uppermost. The shell s is 
much larger than it was at the preceding stage, and is usually 
quite irregular in outline, although a few embryos w'ere found 
at this as well as at the stage 36, in which each valve was per- 
fectly regular in outline, pear-shaped, and placed with the 
narrow end nearest the middle line. A little posterior to the 
shell is the anal papilla a , which now carries a few short, stiff 
cilia or setae. The relative positions of the shell, anal papilla 
and velum in the preceding stages seem to show with satis- 
factory clearness that the side which is uppermost in this 
figure is that which is below in most of the preceding figures, 
and which I have called dorsal. 
The digestive tract is now much larger than at the pre- 
ceding stage, and its centre is occupied by a distinct cav- 
ity, the wall of which is ciliated. At a point nearly op- 
posite the shell this cavity opens externally by a new opening, 
m, and small particles of food now find their way into the 
central cavity, where they are kept in rotation by the cilia. 
At this stage the margins of the opening m can be protruded 
so as to form a projecting sucking disk, by which the embryos 
adhere to each other and to foreign bodies. 
The series of stages which I have figured seems to show 
that this new opening is almost directly opposite the position 
which the blastopore occupied at stage 32. 
