ies BULLETIN OF THE 
probably the same phenomenon that Salensky (’71) described as a pro- 
cess of invagination, but to what extent it is comparable to a true in- 
vagination I am not at present able to say. Although a direct connection 
between this depression and the primitive cumulus has not yet been 
traced, it is certain that in point of time the depression is the forerunner 
of the cumulus, and the circumstantial evidence of their similar positions 
on the egg indicates a connection between the two. 
The external feature just spoken of as the primitive cumulus origi- 
nates as a thickening of the blastoderm, at one end of the flattened sur- 
face of the egg, and usually terminates in the production of a low conical 
elevation. In surface aspect the cumulus is ovoid, with its more pointed 
end directed towards the centre of the flattened surface, and it often 
shows a tendency to elongate in that direction. This patch of cells being 
rather opaque, appears whitish by reflected light, and dark by transmit- 
ted light. In some specimens it is considerably elevated above the sur- 
face of the egg, but in other cases it is only slightly or not at all raised. 
Upon hardened eggs the surface of the cumulus is often depressed by a 
median longitudinal furrow from which two or three smaller irregular 
furrows radiate towards its margin (Pl. I. fig. 4). 
A second thickening, which I shall call the caudal thickening (¢ dn. 
ca.), soon makes its appearance on the flattened surface of the egg, at a 
distance of about 80° from the cumulus (Fig. 2.) It increases rapidly 
in size, spreading out most in the direction of the cumulus, and ulti- 
mately becomes shield-shaped. In the region between these two struc- 
tures the ventral plate is gradually formed by a blastodermic thickening, 
which is not at first continuous with the two terminal thickenings. In- 
dications of the existence of a ventral-plate thickening, which appears 
lighter by reflected light, are to be seen in the surface view shown in 
Wie, 2Pl: i. 
Immediately following the stage just described, the whole ventral 
surface of the egg becomes divided by a series of transverse ridges and 
furrows into protozonites (PI. II. fig. 6). I have not the material to 
determine all the steps in the process, for the time involved in passing 
from the stage of the primitive cumulus to the protozonite stage is a com- 
paratively short one. The earliest condition in the latter stage which I 
have examined shows three zonites and the cephalic plate. At this time 
the latter is only faintly outlined. It isa broad thickening, rounded 
towards the dorsal region of the egg, and fading into the protozonites on 
the ventral surface. The caudal plate does not become visible until two 
or three more zonites are established. It is similar in outline to the 
