544 
H. V. WILSON 
Acetic acid. Pieces were fixed in glacial acetic for a few min- 
utes (5-10) and then transferred to water. The dermal mem- 
brane was peeled and cut from the choanosome, and was then 
cleaned of the underlying sponge parenchyma which was picked 
away with forceps and needles. The pieces were then stained, 
some in methyl green, others in acetic carmine or in haemalum, 
and were mounted in glycerine. Preparations so made give re- 
sults similar to the foregoing. But they are not as transparent 
as balsam preparations and do not disclose the detailed structure 
of the internuclear sheet. 
Osmic acid. Pieces were fixed in one-half per cent osmic for 
10-15 minutes, washed in running water, and hardened in Muller's 
fluid 12 hours. They were run up very gradually through the 
alcohols. Sections and strips were made, stained in haemalum, 
and mounted in balsam. The preparations very frequently 
exhibited interesting artefacts. At first sight an epithelium seemed 
to be marked out in the clearest way. Perfectly clear channels 
of considerable width cut up the surface layer into areas that 
were often polygonal. Examination with an immersion objec- 
tive showed that these areas were not cells. They sometimes 
have nuclei and sometimes not, and the channels between the 
areas exhibit peculiarities in their course which clearly indicate 
them to be cracks. The whole appearance must be due to the 
cracking of the very delicate epidermal sheet. The fixative 
probably makes the sheet brittle, and it later cracks perhaps 
during the washing. 
Sublimate. Pieces were fixed for a few minutes in saturated 
corrosive subhmate and washed in iodised 70 per cent alcohol in 
the usual way. Tangential sections and strips of epidermis were 
prepared and stained in haemalum and congo red. Such prepara- 
tions frequently exhibit artefacts similar to those produced by 
osmic. The surface layer is broken up into thin and irregularly 
polygonal pieces that are widely separated by perfectly clear chan- 
nels. The latter are crossed in some places by a few slender 
protoplasmic filaments. Careful examination shows that the 
pieces are certainly not cells. Some are without nuclei, others 
with a nucleus or sometimes with two. They often include one 
