578 bulletin: museum of comparative zoology. 



The relation between the invagination and the nerve cord was 

 suggestive. Two slender prolongations of the fibrous region of the 

 cord occuired at its anterior end. The length, position, and shape 

 of branch a (Fig. 5) shows that it was probably composed of regen- 

 erated material. This branch lay between the digestive tube and the 

 invagination. The irregular contour of the anterior end of the cord 

 suggests that branch c may have been produced by a partial degenera- 

 tion of this end, though the possibihty of its being a regenerated 

 structure is not excluded. No regenerated brain was present. 



The conditions in the other four worms closely resembled those in 

 number 433, except in the following details. In two of the four cases 

 the slender regenerated end of the nerve cord turned upward toward 

 the invagination, but along one side of the digestive tube instead of 

 in front of it. In one case the invagination branched twice dicho- 

 tomously. In another worm there were two invaginations from the 

 scar tissue on the dorsal side, exclusive of the one whose inner region 

 was closely associated with the ends of the nerve cord and digestive 

 tube — three invaginations in all. 



These pouches were not due primarily to abnormalities in healing 

 at the anterior end of the body, for in one of the five cases the worm 

 healed completely and normally. Moreover, they were altogether 

 too long and too slender to have been produced by muscular contrac- 

 tions. 



3. Discussion and Conclusions. 



The observations in the second group of experiments will be dis- 

 cussed first. 



The upward regeneration of the nerve cord and the downward 

 development of the ectodermal pouch strongly suggest that these 

 invaginations are stomodea representing abortive attempts to re- 

 generate a head on the dorsal side. Alluring as this hypothesis may 

 be, there are several valid objections to it. (1) The brain is absent 

 in all cases and the connectives can not be demonstrated conclusively. 

 (2) None of the invaginations have the slit-like form which I observed 

 in normal stomodea. (3) The invaginations did not open into the 

 alimentary canal, though in three of the worms they were long enough 

 to do so. (4) One of the pouches branched twice dichotomously. 

 (5) In one worm several invaginations were present in the cicatrix 

 on the dorsal side, but normally only one stomodeum forms in a 

 regenerating worm. 



