OSCULA AND ANATOMY OF LEUCOSOLENIA CLATHRUS. 489 
some days. The sponge contracted completely when the cold 
weather began, and has not expanded since; but it is still 
perfectly healthy, as shown both by the histology of sections 
from portions of it, and by its continuing to grow and send 
out processes. 
If we compare fig. 5 with figs. 1 and 2, it is obvious that the 
tubes have shrunk to about one eighth of their former dia- 
meter. Imagine now a Leucosolenia tube, with its walls 
composed of ectoderm externally, jelly containing a single 
layer of spicules and a few cells, and most internally a con- 
tinuous, closely packed lining of collared cells. If this tube 
contracts greatly what must be the result? There can be no 
longer room for the collar cells to form a single layer, and the 
spicules will also be closely packed, probably into several 
layers. 
Figs. 15 a@ and e¢ represent two sections from a series 
through some partly contracted tubes. The spicules now 
form at least two layers in the much-thickened mesoderm, 
and the collar cells are arranged in a stratified epithelium, of 
which the uppermost only bear flagella. In some places the 
endoderm is thrown into folds (fig. 15 c). In other words, we 
have before us Haeckel’s variety Ascetta meandrina. 
Mr. Bidder, in his recent review of Dendy’s ‘ Monograph of 
the Victorian Sponges,’! has written (p. 628), “In these Aus- 
tralian sponges (Calcarea Homoceela) there appears to 
occur none with a many-layered endoderm. This structure, 
observed by Haeckel, and since universally discredited, cer- 
tainly appears in Ascetta clathrus.” I must say that a 
many-layered endoderm as a normal feature of sponge ana- 
tomy is to me as inconceivable as that a sponge should be 
permanently without an osculum. In every preparation I 
have made of this sponge in the expanded condition I find a 
single-layered endoderm. On the other hand, if the sponge 
be sufficiently contracted, a many-layered endoderm does and 
must occur. One usually finds it in preparations made from 
1 * Quart. Journ. Mier. Sci.,’ vol. xxxii, part 4, October, 1891, pp. 625— 
632. 
VOL. XXXIII, PART IV.—NEW SER. LL 
