32 GEORGE BIDDER. 
intra-choanal area, however difficult it may be to understand 
how the food is brought there. It is also obvious that where 
there is an interstitial substance the water cannot pass over 
the surface of the cell, as I formerly supposed. Therefore 
until direct evidence is obtained we must consider it probable 
that the pupil of the iris is the aperture both of ingestion and 
egestion. I have never witnessed in life anything suggesting 
pseudopodial action of the collar (except possibly change of 
length), but it is difficult otherwise to see how cells can ingest 
through the intra-choanal area starch-grains as wide as them- 
selves. 
It is commonly stated that sponges can be easily starved by 
filtering the water. Fig. 3 represents collared cells from S. 
raphanus which had been four days in water passed entirely 
through filter-paper ; there was no difference apparent from 
sponges which had been detached on the same day and re- 
placed in the water from which they had been gathered. 
In L. aspera and 8S. raphanus the current is not stopped 
by the application of carmine,—which, as stated above, is in- 
gested from the first. The current was stopped (L. aspera) 
after a few minutes by the carminate of alumina employed, 
which may have had with it some soluble poison producing 
this effect; but the sponge was preserved within fourteen 
minutes from first administration, and the collar-cells were 
found to have ingested the carminate freely. Far from the 
dermal pores closing for hours against suspended matter, 
powdered charcoal (L. aspera), and starch (S. raphanus), 
in sponges killed after seven minutes and five minutes res- 
pectively, were found solidly filling the afferent canals. With 
the starch the prosopyles were also filled, and widely open, 
and there was starch free in the flagellated chambers and even 
in the cloaca; the starch grains (and still more the particles 
of charcoal) were too large for easy ingestion, but they were 
adhering to and certainly occasionally ingested by the collar- 
cells. 
Topsent (9) finds that with the parasitic Cliona “ méme d’y 
