THE COLLAR-CELLS OF HETEROCGLA. 29 
the persistent union of nutritive with motor functions in the 
cells lining these chambers, since the flagella have their highest 
efficiency where the velocity is least. The healthy nutrition 
of a sponge (excepting lamellar forms) depends on the energy 
of the current from the osculum being high; the economy of 
its motor apparatus depends on the velocity of the water in its 
chambers being low. All transition from more to less primi- 
tive canal-systems exhibits an increase in the ratio between 
these quantities.” 
The mechanism of filtration we now know to have nothing 
to do with Sollas’s membrane; the cardinal fact of filtration 
was very striking, and remains to be explained.! 
As to the locality of ingestion and digestion, my permanent 
preparations available are in all from five specimens of S. 
raphanus, eight of Leucandra aspera, and one of Leu- 
cosolenia clathrus. The intervals between the first applica- 
tion of suspended particles (carmine, starch, &c., rubbed up in 
the sea-water), and that of the preserving fluid were respec- 
tively 5, 10, 10, 11, 14, 21, 27, 50, 60, 77 minutes, 44 hours, 
18 hours, 22 hours, and 3 days. Most of the sponges were 
placed in clear sea water for various periods before killing ; 
but the accumulations on the spicules, &c., render this of 
doubtful value. 
Re-examining anew all these preparations very carefully 
with Zeiss’s apochromatic immersion lens, I can support my 
old conclusions, and make some additions. Ingestion com- 
mences freely at once; on the whole, evidence is in favour of 
it taking place within the collar of the cell. After twenty 
minutes the foreign particles are often found enclosed in a 
vacuole, and they are more generally in the basal parts of the 
collar-cells. 
Carmine is found here in S. raphanus which had been 
in pure sea water eighteen hours, after feeding for twenty 
minutes; only very fine particles are present, in the bases of 
1 [ should warn anyone repeating the experiment that carmine is often 
soluble to a considerable extent in sea water. That which I used at Naples 
in 1887 was not soluble in the sea water of the aquarium. 
