102 BULLETIN OF THE 



1 have already stated (page G5) that the region from which the re- 

 generating buds of Cheilostomes arise, although one of flattened epithe- 

 lium, is one in which many more nuclei persist than elsewhere in the 

 adult (cf. again Fig. 71). This fact, coupled with the constancy of 

 position of regenerating buds with reference to the .degenerated polypide, 

 is to my mind evidence against the assertion that buds arise here from 

 "histologically very definitely differentiated tissue." 



As for regeneration in Endoprocta, no one is more competent to speak 

 than Seeliger himself. I am the more surprised, therefore, to find that 

 in Ascopodaria macropus, which is quite closely allied to the species 

 studied by Seeliger ('89), the cells at the part of the stalk immediately 

 below the " head," from which regenerated buds arise, are, as Elilers's 

 magniticent Pedicellina work shows, very large and cuboidal (Ehlers, 

 '90, Taf. 11. Figs. 26-33). I think one may conclude that a similar 

 condition obtains ii\ some cases in Pedicellina, even judging from Seeli- 

 gei*'s own drawings, although they are drawn to a scale that is not quite 

 large enough to allow of settling this point (Seeligei*, '89, Taf. X. Fig. 

 35, a, Fig. 41, etc.). 



If the increase in size of the flattened cells, and their subsequent rapid 

 division and invagination to form a bud, are due to their more active noui'- 

 ishment, it would be difficult to see why certain cells of any region should 

 quickly undergo this modification, while the adjacent cells apparently 

 as favorably situated with reference to the acquirement of food retain 

 their flattened, quiescent condition, if we assumed such favorable situation 

 to be the only requisite. Still less satisfactorily would such an assump- 

 tion explain the regular position of regenerating buds. It is taking only 

 one step farther back, but, to my mind, a helpful step, to assert that cell 

 proliferation in any region which produces invagination depends upon 

 the capacity of the cells of thM region to become better nourished than 

 their fellows. This may evidently be effected by a diminution in the 

 feeding capacity of the surrounding cells, or by an increase in this respect 

 in the growing cells. 



IV. Relationships of Endoprocta and Ectoprocta. 



I discussed this topic in my earlier paper (Davenport, '90, pp. 132, 

 133). I have only to add, that later studies have confirmed my 

 opinion of Nitsche's correctness in placing these two groups close to- 

 gether, and in regarding the Endoprocta as nearer the ancestral types. 

 The stages of Figures 25 (Plate III.) and 77 (Plate IX.) probably rep- 



