422 BICKFORD. [Vou. IX. 
nature and arrangement of the cells was taking place. The 
whole end became more deeply pigmented, — which may have 
been due to the increase of endodermal cells forming the folds 
of the hypostomal region—and two rows of pigmented spots 
appeared ; one near the end, the other a short distance below. 
This lower row soon became elongated to pigment bands, see 
Fig. 4a, which were later found to be the tentacles. Allman 
figures a stage similar to this in his work. Up to this period 
all changes had taken place within the perisarc, with no visible 
increase in the size or length of the regenerating coenosarc. 
This appears to me to be a point of much interest, as it 
apparently indicates that a large part of the changes in the 
formation of new hydranths are not alone due to increase in 
bulk by new growth, z.c. a development from a few indifferent 
cells, as Lang describes for budding hydroids, nor to the resto- 
ration of each part, as Weismann says occurs in the regenerating 
hydra. (In the case of this hydroid at least, the regeneration 
appears to be largely a direct transformation of a stem portion 
over into the body portion of the new hydranth) This may 
imply that the endoderm cells of the stem portion have in 
some way undergone a differentiation or transformation, fitting 
them more definitely for the various functions of the digestive 
processes, which take place in the hydranth. Since finishing 
the experiments described here, I have been unable to make 
out these histological details of the cell-changes of this form, 
for lack of material. I have, however, investigated this point 
in my work on Cordylophora and have been able to obtain 
sections showing different stages in the transformation of the 
more cubical endodermal cells of the stem, into the more 
elongated cells of the regenerating hydranths. 
It is worth while to notice it was observed, in connection 
with this work, that regeneration does not appear to start from 
a few undifferentiated cells, as in the case of budding in 
hydroids, nor from owe cell, as in the case of embryonic devel- 
‘ opment, but that a// the cells in the regenerating end take part 
directly in this process. What is still more suggestive is that 
this fact admits of the interpretation that possibly the histo- 
logical and physiological qualities of the cells of an adult tissue 
