948 THE AMERICAN NATURALIST. [VoL. XXXIII. 
a week or more after the knobs appeared, they failed to de- 
velop. Considering the large number of experiments that I 
have made, I am inclined to think that Hargitt overestimated 
the capabilities of pieces of this sort. The result, it seems 
to me, follows only when the cut has been made near the 
margin of the medusa. The peripheral rings that had been 
cut off in the preceding experiments, fifty-three in number, 
were kept alive to see if, as stated by Hargitt, they too would 
give rise to medusee. All those pieces from which a large part 
of the bell had been removed failed entirely to close in, and 
died after a few days. It seems that under very favorable cir- 
cumstances a piece from which only a small proximal part 
has been removed may again develop into a complete medusa, 
and in fact, in one or two cases, this seemed very nearly ac- 
complished. In one case a new stomach formed, and a very 
small manubrium ; in another case the proximal ends of two of 
the canals united, and a small manubrium developed, but lay 
somewhat eccentric in position. The difficulty seems to be in 
the closing in of the ring rather than in the regeneration of a 
new stomach and manubrium. This is shown by the following 
experiment : 
In several cases I cut out from the oral side the entire 
stomach and its attached manubrium, as well as the immediate 
proximal ends of the radial canals. Care was taken to remove 
completely all the endoderm in this region, but the cut did not 
pass through the jelly of the bell. I hoped to see if, under 
these circumstances, a new stomach and manubrium would 
develop, or if from each of the proximal ends of the radial 
canals a single manubrium would sprout forth. The latter pos- 
sibility would seem to exist in the light of the experiments on 
the one-half and one-fourth pieces. If, on the other hand, a 
new stomach and a single manubrium developed, this would 
seem to indicate some sort of interrelation of the parts with 
one other (the canals are, of course, connected by a plate of 
endoderm). The endoderm grew forward over the region pre- 
viously occupied by the stomach, and out of it was formed a 
stomach from which a new manubrium grew out. 
It has been shown that in the one-half and one-fourth 
