THE COLLAR-CELLS OF HETEROCGLA. 31 
case (S. raphanus, eighteen hours after feeding) of carmine 
in the parenchym jelly among similar sized brownish particles, 
giving vividly the impression that they have been discharged 
from the collar-cell above. There is one apparently certain 
case (Li. aspera) of a starch grain apparently enclosed between 
mesocytes in the parenchym near an afferent canal. I have 
seen no other instances, and there is nothing which leads me to 
suppose that as a rule undigested food ever passes into the 
parenchym, nor have I any observations which indicate the 
means of nutrition of the parenchym otherwise than as con- 
cerns the gonocytes. And it is worth stating that the few 
carmine-grains observed in ectocytes were never enclosed in 
vacuoles. 
Though there are many cells containing carmine in S. ra- 
phanus after eighteen hours’ feeding, the particles are fine 
and the mass small. lL. aspera, twenty-one and a half hours 
after twenty-one minutes’ feeding, shows no carmine. 
As to the natural food and feeding of the sponge, S. com- 
pressum killed directly from the sea shows in the protoplasm 
of its collar-cells, besides and among the basal spherules, 
numerous minute irregular particles, often highly refractive ; 
sometimes three or four in a vacuole-like structure (cf. cut, a), 
Many appear to be bacilli, being rod-like bodies 1p to 1°84 
long by ‘1uto*2 broad. In another specimen there are lying 
freely in the chambers several specimens of what appears to be 
an alga, one a sphere of four cells, one probably of sixteen; 
also lying inside the collars of different collar-cells are several 
isolated spheres, of about the same size as the individual cells 
of the larger spheres, and similarly stained. In this preparation, 
and another of L. aspera, there are in the chambers several 
larger nucleate cells, possibly Protozoa, partly enveloped by 
the distended collars, sometimes more than one cell converging 
on them. I have not hitherto witnessed any similar phenomena 
in life, nor do I know of any such being recorded. 
In several instances in the carmine preparations there are 
grains inside a collar, as Vosmaer and Pekelharing describe, 
and the evidence certainly so far points to ingestion by the 
