26 GEORGE BIDDER. 
observed. The latter were, however, always exposed to a con- 
siderable mass of air, and respiration may be supposed to have 
been still possible; under the cover-slip this was of course not 
the case. On the other hand, the radial chambers under the 
cover-slip each contained the excretory products of at the 
most two hours, for which time only they had been deprived 
of food; in the case of the tidal changes, nothing but gaseous 
matters could have been either received or eliminated for one 
or two days. It seems, therefore, plausible to suggest that the 
characteristic appearance results, in the suffocation changes, 
from want of oxygen or presence of carbonic acid ; in the tidal 
changes, from starvation or the presence of non-gaseous excreta. 
The local suffocation transparency appears to be mere segrega- 
tion, the tidal transparency may be due to starvation. It may 
possibly be important that the metamorphosis here attributed 
to lack of oxygen results in a maximum surface, that attributed 
to presence of poisonous products results in a minimum surface. 
Tidal changes were never observed to originate under the 
cover-slip, nor on the other hand did. cells so metamorphosed 
ever give rise to suffocation forms. The elongated suffocation- 
cells died extended, the hemispherical tidal cells died hemi- 
spherical, neither modification showing any signs of giving rise 
to the other. Only in one section (of a sponge twenty-seven 
hours out of the water) I found, after an hour on the slide, a 
chamber lined with the usual low, round, collarless cells (as in 
fig. 13), but with two collared-cells of the extreme suffocation 
form (as in fig. 12), 30 long, stretching almost across the 
chamber. The contrast was very striking, and seemed to hint 
that accompanying the loss of the collar is some change, perhaps 
of the lateral walls, which means the loss of power of exten- 
sion. These two cells had escaped the tidal modification, and 
therefore were able to respond to the stimulus of suffocation. 
All appearances suggest that the extension under suffocation 
is due to constriction of the lateral wall—whether it be a con- 
traction set up by these conditions, or a normal tone which the 
enfeebled cell-contents can no longer overcome. 
Apparent migration of collar-cells into the parenchym 
