Prof. W. Thomson on the ‘‘ Vitreous” Sponges. 129 
the normal'form. I am not quite satisfied on this point even 
now. As I am precluded from using either of Dr. Gray’s 
names, I substitute one which I had. in MS. before I saw Dr. 
‘Gray’s paper. 
H. corbicula, Valenciennes (sp.). Pl. IV. fig. 1. 
Aleyoncellum corbicula, Val. (in Paris Museum); Bowerbank, British 
Spongiadee, vol. i. p. 176. 
Heterotella corbicula, Gray, Proc. Zool. Soc. 1867, p. 531. 
_ The Sponge is tubular, shaped somewhat like a wine-glass, 
about 4 inches high and 2 inches wide across the lip, and tapers 
downwards somewhat irregularly to a diameter of 14 inch at 
the base. The wall of the cylinder is formed of a rather thick 
irregular network of delicate siliceous spicules, from ‘01 to °5 
inch in length, loosely arranged in fascicles which cross one 
another at low angles, and are loosely connected and com- 
bined by a small quantity of very soft mucilaginous sarcode, 
which, in the dried sponge, remains as a thin yellowish film. 
These cords curve upwards and downwards and round, anasto- 
mosing in all directions, and leaving between them rounded 
openings of various sizes, and show no tendency whatever to 
a regular longitudinal and transverse direction. All the long 
‘spicules are formed on the hexradiate type; but the four 
secondary rays are usually abortive, being represented merely 
by four tubercles at right angles.to one another, about the 
middle of the main shaft, which is somewhat enlarged and 
tuberculated at each end (PI. IV. fig. lc). The spicules 
have, according to the ordinary plan of sponge-structure, a 
delicate centre canal, which sends off four short radii to the 
four secondary tubercles. The walls of the spicules consist of 
concentric layers of silica separated by films of sarcode, which 
can be readily shown discoloured by burning in a spirit-lamp. 
This structure can be best studied in the long spicules of Hya- 
lonema, which are in every respect, except in size, essentially 
the same. 
Scattered among the long spicules of the skeleton, there are 
many fully developed hexradiate spicules. Some of these are 
perfectly regular in form, the rays smooth and nearly equal 
(Pl. IV. fig. 16); but many of them are irregular, the rays 
are distorted and bent (fig.1a@), and in some cases two or 
_ More are irregularly united together. I have little doubt that 
__ these latter indicate the first stage, as it were, to the formation 
of a continuous network such as we find in Huplectella and 
Aphrocallistes. In Habrodictyon, however, the coalescence 
“never occurs to any extent, and the network remains perfectly 
flexible and without a trace of the raised filigree ridges which 
