STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 135 
finds its way out of the sponge through the wide, terminal os- 
culum (fig. 1, osc.). 
The histology of the sponge, so far as the condition of the 
specimen will permit of investigation, offers no features of 
special interest, and appears to agree with that of other He- 
terocela. Beyond the transparent gelatinous ground-substance 
of the mesoderm, the contracted collared cells, and the nuclei 
of the pavement epithelium lining the canals, I have not been 
able to make out any details. 
III. Revationsuirs oF LELAPIA. 
A. Relationships to other recent Heterocela. 
The canal system of Lelapia australis, as already pointed 
out, offers no features of peculiar interest and, as regards its 
probable derivation from the more primitive Syconoid type, 
stands on exactly the same footing as the canal system of any 
other Leuconoid Heterocele. As the probable mode of deriva- 
tion of the Leuconoid from the Syconoid type has already been 
discussed in my memoir on the structure and classification of 
the Heteroccela (1), it is unnecessary to enter further into the 
question in this place. 
The skeleton, however, is very peculiar, and, at first sight, 
may seem to place great difficulties in the way of believing 
in the Syconoid ancestry of Lelapia. These difficulties, how- 
ever, disappear upon closer examination. 
The peculiar form of the “tuning-fork” spicule is not, in 
itself, of much significance, and, as already pointed out by 
other writers, it is paralleled more or less closely in Haeckel’s 
Leucandra (Leucortis) pulvinar and L. (Leucetta) 
pandora (4), both of which are recent species ; while it is also 
met with in the fossil Sestrostomella rugosa and S. 
clavata described by Dr. Hinde (5). In none of these, how- 
ever, does it appear to attain to anything like the degree of 
development met with in Lelapia australis. The tendency 
of the triradiate to vary is well known, and we meet with a 
