304. E, A. MINCHIN. 
versally true, at least for the primary monaxons of Calcarea, 
that each such spicule is produced by two formative cells. 
This removes the only difficulty in the way of deriving the 
triradiate systems from a fusion of originally separate 
monaxons. ; 
The species of Leucosolenia studied by me are L. com- 
plicata and L. variabilis. I have elsewhere (1905 [2]) 
given a detailed description of the characters, and especially 
of the spiculation of these two species. The material was 
obtained from Plymouth, and I have to thank the authorities 
of the Marine Biological Association for the trouble they have 
taken in collecting and preserving these sponges according 
to my directions. The technique used by me was the same 
as in the case of Clathrina, namely, preservation fresh from 
the sea in } per cent. osmic acid, staining with picrocarmine 
(either Ranvier’s or Weigert’s picrocarmine, supplied by 
Griibler and Co.), and examination in neutral glycerine or 
Canada balsam. This method is excellent for all proto- 
plasmic structures, and does not corrode the spicules in the 
least; but it does not show details of nuclear structure, 
though the nuclei as a whole are well shown and clearly 
differentiated from the cell-body. For nuclear structures 
acid fixatives are indispensable, but then, of course, spicular 
structures are destroyed. 
I have studied the stages of spicule- formation principally 
in surface-views of the thin wall of the sponge laid out flat, 
and rely upon sections only for confirming and supplementing 
the results so obtained. But I have abandoned the method 
which I formerly practised of brushing the collar-cells away 
with a paint-brush, as I now believe this procedure to be a 
dangerous source of error. The young stages of the spicules 
are found so close to the gastral epithelium (see especially 
Pl. 18, figs. 50—52, 54, 55, etc.) that I am now convinced 
that brushing may remove spicule-forming cells as well as 
collar-cells, and that some of my former figures (1898, 
Pl. 38, figs. 9, 12) are, in consequence, erroneous and incom- 
plete, as I shall have occasion to point out further below. I 
