16 GEORGE BIDDER. 
aiisseren glatten Bautheiles vorkommen dass die benachbarten 
Collare sich fast beruhren.” In S. compressum such con- 
tact does not occur, either in healthy life or in any of the 
morbid conditions I was able to investigate. 
Sollas’s membrane occurs, on the other hand, in paraffin 
sections of S. compressum (v. figs. 16—18, 20), preserved 
by any delicate method except the very best; careful examina- 
showing that it is always associated with great distortion of 
the cells, and that this is also the case in the drawings by other 
authors. Where there is no distortion (fig. 15) the membrane 
is not present. 
Dendy was right in saying that cells showing the membrane 
may also possess flagella, though generally this is not the case. 
And the phrase “ portions of flagella and collars irregularly 
sticking together” (19) is not descriptive of this very definite 
structure as it occurs in many sections. But these same 
sections have been prepared with great care (all being osmic 
acid preparations) from a sponge which I know in life had all 
its collars disunited and normally cylindrical; and in five 
cases the same individual was examined partly by a living 
section, partly by paraffin sections (cf. figs. 19 and 20). It is 
not disproved that union of the collars may occur in some 
living sponges—more probably in some dying sponges. But the 
evidence of ordinary paraffin sections for its existence must 
now be considered valueless, and, with exception of the ob- 
servations quoted and explained above, there was no other 
evidence for its existence. There remains no reason to believe 
that it occurs in nature at all, and I must thank Dr. Vosmaer, 
my old friend and master, for yet another lesson in sponge lore. 
Some measurements will be found in the note on distortion 
of cells at the end of this paper. 
It is worth mentioning that in the living larva of S. rapha-~ 
nus (Naples, June) I found that the transparent ends of 
the flagellate cells, lettered by Barrois (5) as “ collier,” are 
solid and refractile, as faithfully figured by Schulze, x 5; 
the convex distal surfaces are correctly shown by both authors, 
