Gy 
STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 151 
viz. that of the dermal cortex, that of the gastral cortex, and 
that of the chamber-containing layer of the sponge-wall 
between the two. To these may be added the skeleton of the 
peristome, but this is hardly of sufficient importance to deserve 
special consideration. The principal part of the skeleton is, of 
course, that of the chamber layer, which occupies nearly the 
entire thickness of the sponge-wall, and it is here that we meet 
with the most surprising peculiarities in structure. 
Skeleton of the Dermal Cortex.—The dermal cortex 
is very thin, and its proper skeleton consists of a thin, confused 
layer of rather small, slender-rayed, normal sagittal triradiates, 
lying parallel with the dermal surface, together with an im- 
mense number of the minute oxeote “ mortar-spicules.” Mr. 
Carter notes the occurrence of small quadriradiates on the 
dermal surface, but these I have not detected in the small 
piece which I have examined. They certainly do not play any 
important part in the formation of the skeleton, and, unlike 
the large subdermal quadriradiates of the Amphoriscidz, cannot 
be regarded as of any systematic importance. The huge 
oxeote spicules take no part in the formation of the dermal 
cortex proper, although many of them lie just beneathit. Here 
and there the dermal cortex is pierced by very stout and 
densely packed bundles of the long slender oxeotes (inter- 
mingled with “tuning-fork” spicules. These bundles (fig. 3, 
d. t.) constitute the expanded (but densely packed) ends of 
some of the spicular fibres of the chamber layer, which just 
pierce the dermal cortex and give rise to the characteristic 
ridges on the outer surface of the sponge. The outermost 
ends of the spicules are commonly broken off short. 
Skeleton of the Gastral Cortex.—The gastral cortex 
is very much more strongly developed than the dermal, having 
a thickness of about 0°2 mm. It is composed almost entirely 
of very densely packed, laterally extended sagittal triradiates, 
with very long oral rays and very short basals. These spicules 
are arranged with their longest axes parallel to the gastral 
surface, but otherwise in the greatest confusion, with the 
short basal rays pointing in all directions instead of constantly 
