RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 299 
The secretion of gelatinous membranes is of interest. Other 
noteworthy points are as follows :—(1) The size of the spherules 
produced varies within considerable limits. Those produced by 
a Single mass might be approximately equal, or of very different 
sizes. (2) I at first thought that the phenomenon was deter- 
mined by the proportion of dermal cells present, subdivision 
continuing until enough surface was formed to accommodate all 
the dermal cells in the state of simple epithelium. Appearances 
like that of fig. 1, b, however, seem to negative this, for there 
the epithelium surrounding the spherules is cuboidal, and quite 
unlike any dermal cells seen by me. This epithelium seems to 
consist of the healthiest choanocytes present. The difference 
of colour between them and the cells of the inner masses, how- 
ever, 1s to be noted, and it is possible that they represent 
a dedifferentiated condition of the dermal cells. If so, they 
would resemble the cuboidal form of the ectoderm cells seen 
in dedifferentiated stolons, &e., of the Ascidians, Perophora 
and Clavellina-cells which are normally as flattened as the 
normal dermal cells of Sycon. At all events the phenomenon 
must be determmed by some surface-volume relation, the cells 
not being able to cohere in large masses when in certain con- 
ditions. 
In any case, the spontaneous segmentation of the masses 
into regularly-arranged portions of smaller size is of interest. 
This phenomenon never occurs, as far as is known, in the normal 
life-history of Sycon; yet the process is regular, and at first 
sight would be taken for a normal occurrence. It ig an example 
of the determination of physiological processes by the direct 
action of external circumstances, without any modification 
by way of heredity. A somewhat similar phenomenon was 
found by Miiller (10) in restitution-bodies of Spongilla, but it 
was not so regular, nor, since it only occurred in large masses, 
does it seem to have been due to identical causes. 
(3) The separation of the clearer cells from each other, 
apparently when the circumstances have become slightly more un- 
favourable, is also of interest. In Perophora and in Hydroids, a 
slight concentration of toxic substances starts dedifferentiation 
