88 SIDNEY F. HARMER. 
seen in an early stage in fig. 5, where the trumpet-shaped 
mouth is not yet formed. Transparent preparations (fig. 12) 
show that the trumpet is developed in the immediate neigh- 
bourhood of the occluded zocwcium. The appearance of the 
fully formed trumpet is shown in fig. 7. 
The occlusion of a fertile zoecium may or may not result in 
the obliteration of all signs of its presence. In fig. 13 the 
principal spike of z? is still visible externally, in spite of the 
occlusion of that part of the orifice. In other cases, a careful 
examination of a colony may result in the detection of the 
ends of two or three spikes in the immediate neighbourhood 
of the aperture of the ovicell, which clearly belong to a zoecium, 
and from their relation to the trumpet must have belonged to 
the occluded fertile zoccium. What is probably part of the 
orifice of z*? is seen as a zigzag line below the aperture of 
the ovicell in fig. 7. 
The study of sections shows certainly that z? is not neces- 
sarily the fertile zoecium, although the cases in which either 
this or z> is fertile very largely preponderate over all other 
cases put together. The study of entire colonies amply con- 
firms this, if we may for the present assume that the conjunc- 
tion of an occluded zocecium with a trumpet-shaped aperture 
belonging to the ovicell is a strong reason for believing the 
zoecium to have been fertile. In many cases where the open- 
ing of the ovicell is formed in connection with z®, it occupies a 
very characteristic position with regard to that zoecium, being 
placed immediately on the distal side of the original orifice of 
2°, and with its trumpet-like mouth directed towards the disc 
from which the colony started its growth. The opening of 
the ovicell less commonly has a similar relation to 2”. 
It will be remembered that some few of the oldest zoccia of 
each colony form a group which diverge from the centre of 
the discoidal colony. These zocecia are connected with one 
another by vertical septa which do not necessarily reach the 
roof of the ovicell. In other words, a part of the roof which 
starts at one vertical septum does not always become connected 
with the adjacent septum, but may form an arch extending 
