314 JULIAN S. HUXLEY 
suppression of the characters of the cell, also with the assump- 
tion of a simple spheroidal form. Redifferentiation, however, 
is only possible in the direction of the original form, and the 
cell has not acquired pluripotency by dedifferentiating. ‘This 
may be called intra-typical (or unipotent) dedifferentia- 
tion.! 
The existence of pluripotent dedifferentiation is rendered 
probable by various observations which cannot be entered into 
here. It has frequently been assumed, however, on insufficient 
evidence ; and in view of its theoretical importance, and the 
difficulty of proof, very thorough investigation is required to 
establish its existence in any particular case. 
Further evidence against its occurring in Sycon was afforded 
by the artificial production of masses composed entirely, or 
almost entirely, of collar-cells. These, though they lived 
healthily for a number of weeks, never produced a dermal 
epithelium or spicules. This is paralleled by the failure of 
endoderm or ectoderm alone to regenerate in Hydra, as has 
been shown by various observers. 
In a later paper (Huxley, 8) attention was drawn to the 
fact that masses composed only of collar-cells were less viable 
than those containing dermal cells also, although both were 
kept under identical conditions, and although the collar-cells 
are the organs of nutrition. 
In the present paper, the converse of the collar-cell blow-outs 
is shown to occur in the form of masses with an excess of 
dermal cells. These form blown-out vesicles exactly as do the 
choanocytes when they are in excess. 
It is thus clear that, in Sycon at least, the form and composi- 
tion of the restitution-mass depends (apart from questions of 
size) upon the proportions of the different types of cells which 
entered into its Composition. 
It is clear from the observations of Wilson that some process 
of deditferentiation does occur in restitution-masses of Hydroids. 
1 Since writing the above, I find that a very similar classification of 
types of dedifferentiation from the point of view of tumour-growth has 
been adopted by Adami and McCrae (1, p. 324. See also pp. 318-22). 
