30 GEORGE BIDDER. 
the cells, and mostly in vacuoles. The cells containing foreign 
particles do not lose their collars, and the column-and-plinth 
appearance occurs independently of the amount of carmine 
contained. Nor do the collar-cells show any tendency to 
migration, even after being fed (L. aspera) for four and a half 
hours, when many are filled to their very outlines with carmine. 
In the sponge here referred to about 1 per cent. or fewer of 
the glandiform ectocytes contain a grain or two of carmine. 
This may be excretion, but there is no evidence against it 
being casual ingestion. In most recently fed preparations 
there are one or two canal ectocytes containing a grain of 
carmine. 
Examination confirmed the statement (18) that there are a 
number of gonocytes connected by processes or pseudopodia 
with the basal surfaces of the collar-cells, and containing, in 
both body and process, spherules precisely resembling the 
basal spherules of these cells. I still believe, therefore, that 
the gonocytes nourish themselves on the basal spherules at the 
expense of the collar-cells; and in the hypothesis (which I 
think I owe to an oral suggestion of Miss Greenwood in 1888) 
that these spherules are stores of digested food. The prepara- 
tions mainly examined are of the 8S. raphanus eighteen hours 
after feeding, where the carmine lies among the basal spherules. 
A large number of the gonocytes are in contact with collar- 
cells which contain plentiful carmine; in only two of them I 
found carmine- grains, and it is tempting to deduce that vacuoles 
and undigested food do not pass into the gonocyte. 
In L. aspera and S. raphanus migration of the collar-cells 
into the parenchym certainly does not take place after satiation 
to any degree for any period with carmine; nor in L. aspera 
when a large proportion of the collar-cells contain completely 
ingested starch grains;! nor after fourteen minutes’ feeding with 
carminate of alumina, freely ingested; nor after one hour’s 
feeding with Indian ink, freely ingested. There is one clear 
1 The use of the polariscope for recognising starch grains is easily 
practicable with the highest powers. Without it vacuoles of the same size 
are often difficult to discriminate. 
