512 JULIAN 8. HUXLEY 
of a cell is a function of its position. He finally concludes that, 
since the restitution-masses of sponges are so like those of 
Hydroids, the processes occurring in them are of the same 
nature. 
He further stated that the cells in the early stages of the 
restitution-mass formed a syncytium, few or no cell-boundaries 
being distinguishable. 
De Morgan and Drew, in their later work on restitution-masses 
in other Hydroids, while confirming Wilson in some points, 
differ from him in others. In the first place, although restitu- 
tion-bodies with perisare, ectoderm, and typical endodermic 
coenosarcal tubules were produced and lived for as long as 
sixty days, no hydranths were formed. As Orton (12) suggests, 
this may be due to the fact that de Morgan and Drew’s expert- 
ments were performed from December to March, when the 
srowth of Hydroids appears to be at a standstill, while Wilson 
worked in the summer. 
In the second place, although they describe a syncytial 
phase, their figures do not show any such complete cell-fusion 
as Wilson’s, and they mention that a small portion of endoderm 
cells are always to be recognized as such. They do not pro- 
nounce definitively one way or the other as to whether dedifferen- 
tiation of all cells to a ‘ totipotent ’ condition occurs. 
Miiller (10) also believes that collar-cells do not take part 
in the redifferentiation of restitution-masses in Spongilla, but 
that the amoebocytes and thesoeytes form the new gastral 
cells. In view of the great réle played by the amoebocytes in 
monaxonid sponges, and the specialization, small size, and 
relatively small number of the choanocytes, this is not sur- 
prising. In gemmule development, for instance, the flagellated 
chambers arise from archaeocytes. The same author (11), 
in describing dedifferentiation in Spongillidae, notes that the 
choanocytes early dedifferentiate and disappear, apparently 
ingested by amoebocytes. It would appear that they cannot 
maintain themselves as such in unfavourable conditions. In 
this connexion, mention may be made of the work of Maas (9), 
who found that slow deprivation of calcium led to similar 
