EPITHELIOID MEMBRANES IN MONAXONID SPONGES 561 
nucleoplasm, and chromatin in the shape of granules or short 
pieces (doubtless a reticulum exists). 
The term ^^cell" and the idea expressed by it are not altogether 
appropriate to the nucleated areas present in this membrane. 
The areas are everywhere united b}^ abundant intercellular con- 
nectives, and merge ver}- gradually into these, the cytoplasm 
often thinning away into reticulated films which then pass into 
the intercellular strands. Where the margin of the cell area is 
densely granular, the distinction between cell and connective 
is sharper (a figure inevitably represents the contrast between 
cell body and connectives as sharper than it exists in nature). 
The nucleated areas are frequently directl}^ confluent, so that 
one and the same area may contain two nuclei (fig. 20) . The mem- 
brane is actually of course a syncytium, but it is one in which the 
component cells permanently remain in a state of imperfect 
union. The regenerating epidermis passes through an essentially 
similar stage (fig. 7). The canal lining thus remains in a con- 
dition not so far removed from the mesenchyme as is the epidermis. 
Like the epidermis it may be regenerated from the mesenchyme, 
probably as Weltner ('07) maintains, largely frojn the granular 
amoebocytes. 
When the membrane is examined with a comparatively low 
power, the nucleated cell areas appear to have distinct boundaries 
and to be independent cells separated by wide spaces. Fixation 
with sublimate may lead to the same erroneous conclusion. In fig. 
21 the canal epithelium is represented as it appears when pre- 
pared from sublimate material. The cells are widely separated and 
have good sharp boundaries. The cytoplasm is finely granular 
and fairly dense. Almost no intercellular connectives are present. 
The absence of the connectives and the uniform dense granular 
appearance of the cells when comparison is made with a good 
alcoholic preparation such as that from which fig. 20 was made, 
must be regarded as artefacts due to the sublimate treatment. 
There is good indication that the lining epithelial cells are con- 
tractile and of themselves bring about the diminution in bore of 
the canal. The canals certainly do diminish in bore, and greatly at 
times. Round the canals there are no fibre-like mesenchyme cells 
