8 



HARGITT. 



the same time, one being formed a day or two earlier than the 

 other. The presence of these rings is not uncommon in speci- 

 mens otherwise normal, as noted by Hargitt, 1901. As has 

 already been noted, in the fourth set the upper half of the bell 

 was removed. When the wound had healed the bell was flat- 

 tened and elongated. It was therefore impossible for the canals 

 to unite in the usual way. The two canals at the ends of the 

 elongated bell united and between these two pairs of canals a 



FIG. 8. 



FIG. 9. 



straight connecting canal then regenerated, thus completing the 

 chymiferous system. The manubrium then regenerated from 

 these straight connecting canals, but not till the canal was com- 

 pletely formed and functional. If the manubrium formed at the 

 union of the connecting canal with one of the pairs, it was tri- 

 angular in shape and three-lobed as in Fig. 9. If it formed in 

 the center of the connecting canal it was two- or four-lobed. 



III. EXPERIMENTS UPON THE REGENERATION OF THE RADIAL 



CANALS. 



The medusse were first cut in a vertical interradial plane. 

 Thus each piece had two canals and half the tentacles. The cut 

 sometimes passed through the manubrium and sometimes to one 

 side. Thus some pieces would have a part of the manubrium, 

 others the entire manubrium and in the rest the manubrium 

 would be entirely lacking. The cut edges gradually approached 

 each other and finally met and fused, and the typical medusa- 

 form was thus assumed. Occasionally the manubrium projected 

 between the cut edges and delayed the healing. Usually, how- 



