TRANSACTIONS OF SECTION D. ri 
characters which are maintained throughout the interior of the sponge. A section 
through it shows that it is largely made up of cells united into a reticulum with 
vacuolar spaces of various sizes forming the meshes. The jelly which is so con- 
spicuous in the surface layers appears to be scanty or altogether absent here. 
Besides the smaller branched protoplasmic masses with nuclei 1°5—2y in diameter, 
which make up the greater part of the reticulum, there are larger and more 
circumscribed cells with larger nuclei (2—3 diam.) Scattered through the 
reticulum, and with their walls apparently formed by portions of it, are the 
ciliated chambers and the ramifying branches of the canal system. 
Ciliated Chambers.—These are round or oval chambers of minute but fairly 
uniform size, the larger measuring 18 by 11, the smaller 10 by 8u. (The sections 
of smaller diameter are doubtless in many cases transverse to the long axis of larger 
chambers.) Their walls are formed by cells which send out processes laterally, and 
these, joining with one another, bound the chamber, others extending away from the 
chamber are continuous with other cells of the reticulum, while a third set of 
processes project into the cavity of the chamber, and each, tapering gradually from 
its base, forms a flagellum which may extend across to the other side of the 
chamber. There is no indication of a collar, or of the abrupt truncated termination 
of the cell body at the base of the flagellum, characters which are usually seen in 
choanocytes. A well-marked nucleus is situated at the base of the flagellum at its 
junction with the body of the cell. The flagella project from about half the 
concave inner surface of the chamber, some four or five commonly appearing in 
section, and their tips thus converge, and are often seen to have become entangled. 
The remainder of the inner wall of the chamber is smooth. The cavity opens into 
a branch of the canal system by a narrow passage about 4 in diameter. 
Although the features of the ciliated chambers above described differ in many 
points from those usually met with in sponges, they are so distinctly seen as to 
give the impression that they are approximately those of the living tissue. 
The larger branches of the canal system, with flattened nuclei scattered over 
their walls, are easily recognised, but in many cases it is difficult to decide whether 
a particular space met with in a section is to be referred to the canal system or to 
a vacuolar mesh of the reticulum, if indeed the two are distinct. 
In some of the skeletal canals the tissue above described is abundant, in others 
only a thin lining of loose vacuolated tissue is found, but all intermediate con- 
ditions in the character of the soft tissue lying between the skeleton and the main 
trunks of the canal system are met with. 
The existence of ciliated chambeys implies the existence of afferent and 
efferent channels. I have not been able to recognise, from the structure, a 
differentiation of the canals into these two categories, but it seems clear that 
the efferent system is not collected into one or a few large channels, such as 
are commonly found in sponges. My impression is that there are a large number 
not only of afferent but efferent trunks to the canal system, supplying and draining 
approximately equal areas. 
Reproduction.—Each of the three specimens which I examined by sections 
contained large eggs or embryos. They are found near the orifices of the larger 
canals, separated from the skeletal wall by a thin layer of soft tissue. I have not 
been able to recognise eggs in a young stage of growth. An advanced ovwm in 
one specimen measures 0:1 mm. in length, and has a thick-walled nucleus 25u 
in diameter, and a well-marked germinal spot. An embryo in the same specimen 
has a superficial layer of nuclei, and the protoplasm about them is disposed in 
columns perpendicular to the surface. Internally the columns are merged in the 
granular protoplasm which occupies the interior of the embryo, and this is 
obscurely divided into irregular masses, but not, so far as I can detect, containing 
nuclei. None of the embryos have a segmentation cavity. It appears, then, that 
the development leads to the formation of a larva of a parenchymula type, rather 
than an amphiblastula. I have not been able to recognise any stage in the 
formation of spermatozoa. 
Tn the course of growth the soft tissues appear to be withdrawn from the 
