GEGENBAUR, ON SAGITTA. 49 
line membrane, which divides the yelk into two equal hemi- 
spheres. This groove or depression, over which the vitelline 
membrane is stretched like a bridge, penetrates more and more 
deeply into the vitelline substance, until a complete separation 
is effected, between the two portions which simply remain in 
contact. The next division separates each hemisphere into 
two segments, the yelk now appearing to be composed of 
four segments of a sphere. 
Where the four segments meet in the centre of the vitel- 
lus, a minute hollow space may be observed, formed by the 
rounding off of the contiguous angles of the four segments, 
This central space is of no small importance in the further 
stages of the process. 
Whilst these changes are going on, the molecules above 
described as surrounding the germinal vesicle, and which 
are subsequently ageregated around the nucleus of each 
segment derived from the division of the original vesicle, 
assume a peculiar disposition. They become more densely 
packed around the nucleus, from which they extend in ra- 
diating lines towards the periphery. 
Each of the four vitelline segments is now again sub- 
divided into two equal portions in a direction perpendicular 
to the middle of its longitudinal axis, so that the whole yelk 
is constituted of eight equal segments. By a continuance of 
a similar divisional process the number of pyramidal segments 
goes on increasing, each pyramid having the apex directed to- 
wards the centre of the vitellus. An arrangement owing to 
which the segmentation of the ovum of Sagitta appears under 
a very remarkable type. Even in these latter periods of the 
process of segmentation, the segments never assume a sphe- 
rical form as in the Mollusca and Annelids, as well as in 
the Vertebrata, nor constitute the aggregate masses so well 
known in their mulberry-like form. On the contrary, in the 
present case, owing to the circumstance that certain condi- 
tions, in other instances occurring only in the earliest stages 
of segmentation appear to remain persistent, it happens that 
a single segmentation-cell extends from the centre of the 
vitellus to the surface, and at the same time, whilst multi- 
plying, the contiguous segments are closely in contact by 
their corresponding surfaces, in consequence of which they 
become mutually flattened. Pyramidal segments of this 
kind do not, so far as Gegenbaur is aware, occur in any 
other case. The summit of each of these four- five- or six- 
sided pyramids is truncated, and contributes to the forma- 
tion of the boundaries of the vacant central space in the 
vitellus above described. The basis of the pyramid does not 
