No. 396.] HYDROMEDUSA, GONIONEMUS VERTENS. 949 
medusze the new manubrium develops always from the cut end 
of the radial canal. It often lay far to one side of the center of 
the new bell. I tried in another experiment, cutting in two the 
radial canal in the one-fourth piece at the middle of the canal, 
i.e. at a point that would correspond approximately to the 
middle of the new bell. Under these conditions it was possible 
that a new manubrium might develop at this point, rather than 
at the other proximal end of the tube. This result did not 
follow, however, since, at the cut, the ends came together and 
fused. The experiment might succeed if the cut ends could 
be kept apart, or if a short piece were cut out of the canal 
at some point. 
Finally I cut off the rim with its tentacles from three of the 
four quadrants, leaving only one quadrant entire. The experi- 
ment was made to see if, when a part of the rim was left, 
it might give rise at its free ends to the missing part of the 
ring, or possibly, under these conditions, the injured part of the 
rim might more readily produce new tentacles. The medusa 
closed in so that the opening into the subumbrella cavity be- 
came quite small. The opening was surrounded for the greater 
part of its extent by the tentacles of the uninjured quadrant. 
Over a small region no tentacles were present at first, but after 
five days a few new ones appeared. They were not more num- 
erous near the ends of the uninjured part of the ring than 
elsewhere. The result is similar to that in which the entire 
margin is cut off (Fig. 12), although in this case the new tenta- 
cles were better developed. 
The results of these experiments show, for Gonionemus at 
least, that Haeckel’s statement, that even the smallest pieces 
may make new medusa, is not correct ; for although pieces 
somewhat smaller than one-eighth of the medusa may make 
new individuals having the medusa-form, yet these small indi- 
viduals, as well as larger ones, lack the most essential features 
of the medusa. The remodeling extends only to the form of 
the entire piece and does not include the internal organs. 
It is puzzling to determine whether the medusa-form assumed 
by the pieces is simply the form that necessarily results after 
the fusion of the cut edges, or whether the process includes 
