RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SyYcon 813 
dedifferentiation in various calcareous sponges, including 
a heterocoelous form (Sycandra). He also describes degenera- 
tion and phagocytosis of the collar-cells in late stages of the 
process. 
In view of my work (see discussion below), it would appear 
that in both Calearea and Monaxonida the choanocytes are 
more susceptible than the amoebocytes, and will degenerate 
in certain conditions. In Calcarea, however, this difference in 
susceptibility is less marked, and the choanocytes will remain 
capable of maintaining existence in dissociation-masses, while 
this is not possible for those of Spongilla. 
My own work (7) on Sycon indicated that the conclusions 
of Wilson as to the fate of the cells in restitution do not apply 
in the case of Sycon. On dissociation the tissue elements all 
become deditferentiated morphologically, e.g. the choanocytes 
lose both collar and flagellum and become rounded, the dermal 
cells lose their extended flat shape for a spheroidal one; but 
this dedifferentiation is not complete in the sense that the 
various kinds of cells become physiologically similar, or acquire 
the same potentialities of development. After this dedifferen- 
tiation caused by shock the cells redifferentiate in appropriate 
directions, the dermal cells producing an external epithelium 
round a central choanocyte mass, which in its turn becomes 
hollow with epithelial walls. The normal form of the post- 
larval sponge is thus produced by a process exactly the reverse 
of that envisaged by Driesch and Wilson. The fate of the cells 
is not a function of their position, but their eventual position is 
a function of their constitutional differences. The development 
of a restitution-body is primarily a process of sorting-out of 
different kinds of cells, followed by a redifferentiation of the 
individual types of cells. We have thus to distinguish sharply 
between two types of cellular dedifferentiation: (1) that 
which leads to complete loss of the character of the tissue 
to which the cell belongs, and a return to a totipotent, or at 
least, if I may coin a new word, to a pluripotent condition. 
This may be called ultra-typical (or pluripotent) 
dedifferentiation; (2) that which leads to a temporary 
