168 . Dr. W. Marshall on the 
stones, in such a manner that their convex sides are directed 
outwards. These spicules are not quite irregularly arranged ; 
they have a tendency to radiate from certain points in a stellate 
arrangement, and, on close examination, for example in sec- 
tions, we observe that at these points the inner surface of the 
capsule is in contact with the true envelope of the gemmule. 
The latter, in fact, is not uniformly smooth, but furnished 
with numerous, irregularly distributed, conical prominences, 
which pass to the inner surface of the exterior capsule, but 
are only loosely connected therewith. If the gemmules are 
allowed to swell up in warm water within the capsules, which 
is a rather long operation, it is seen that in course of time the 
exterior capsules are first ruptured at these points of contact. 
This may also occur normally when the capsule with its still 
living contents has again got under water; as the capsule has 
not, like the true gemmule-envelope, a special aperture of 
issue, the germ, newly awakened into life, could hardly other- 
wise be set free. 
The true envelope of the gemmule is not thick, shows a 
simple granular structure, and is only covered superficially 
with round siliceous shields, which, on the outside, rise into 
short spiniform knobs and are slightly hollowed within, 
These shields lie so closely over the whole surface that their 
margins overlap; only the conical prominences are quite free 
from them. ‘Their margins are imbedded in the base-sub- 
stance of the envelope, while their little points are exposed. 
The round exit-aperture for the germ is here furnished with a 
closing apparatus. 
The significance of this complicated investiture of the gem- 
mules of Parmula is not hard to understand. The exterior 
capsule, as Carter has already pointed out, is composed of 
true skeletal elements specially modified. These naviculiform 
spicules exactly resemble those with which the stronger trains 
of spicules of the framework of the sponge are coated; the 
capsules usually pass into this coating, and are continuously 
connected with it; frequently also two or more capsules are 
intimately united by bridges of such spicules, as Bowerbank 
has already shown in his figures. It is clear that in this way 
the gemmules are very firmly attached to the rest of the 
framework ; and it often happens that, when one wishes to 
separate such a gemmule-capsule from the dried sponge with 
the forceps, a portion of the adhering spicule-train of the 
skeleton is broken away with it. This intimate union is 
certainly of some importance to the gemmule, and I believe 
that we come upon the track of it if we consider the mode of 
lite of the species of Tarmula. Bates mentions that Spongilla 
