MUSEUM OF COMPAKATIVE ZOOLOGY. 89 



the germ layers in the polypide bud may be subdivided into two: What 

 is the significance of the outer layer of the bud? and, What is the sig- 

 nificance of the inner 1 



The outer layer of the hud is derived from the coelomic epithelium. 

 The views of those who have studied the formation of this inner layer 

 of the cystid in Phylactolsemata may be classed in two categox'ies : 

 (1) those in which it is regarded as entoderm, and the process of its 

 formation as gastrulation ; and (2) those in which it is regarded as 

 mesoderm. To the former class belong the views of Reinhard ('80, 

 p. 208), KorotnefF ('89, p. 403), and JuUien ('90, p. 19) ; to the latter, 

 those of Kraepelin ('83, p. 601) and Braem ('90, p. 116), and in this 

 class the views of Barrois ('86, p. 68) and Haddon ('83, p. 543), founded 

 on a priori considerations, must be placed. 



It seems to me that, since, as Barrois has demonstrated, there is a great 

 similarity between the Phylactolsematous and Gymnolsematous larvse, 

 and especially since the former show evident signs of degeneration, we 

 are bound to study the phenomena they exhibit in the light of our 

 knowledge of the ontogeny of Gymnolaemata. 



But first it is necessary to give reasons for believing that the larva of 

 Phylactolsemata is to be regarded as homologous with that of Gymnolse- 

 mata ; and to do this I will first name the points of similarity in the two 

 larvie, and then try to show that the differences which exist are not suf- 

 ficient to invalidate the attempt to establish a homology. And, first of 

 all, it may be said that, since the adult Phylactolsemata and Gymnolse- 

 mata are strikingly similar to each other, and since no one doubts their 

 close relationship, we should expect a priori that their larvse would be 

 homologous, especially since the larvse of Gymnolsemata are admitted to 

 belong to the trochosphere type, of whose ancient origin there can be 

 little doubt. In the second place, the very existence of a larval stage 

 in Phylactolsemata is indicative of its inheritance from an earlier condi- 

 tion, for two reasons : (a) because in general fresh-water life tends to 

 eliminate larval stages from species which have inherited them from ma- 

 rine ancestors, and tends little to form them de novo (Hydi-a, fresh-water 

 Turbellarians, Rotifera, Oligochseta, Hirudinea, Astacus, and fresh-water 

 Mollusca); and (6) because, specifically, the early stages of development 

 of Phylactolsemata are pasf?ed within a uterus-like sac, from which the 

 embryo is released only when a colony is already well established. In 

 the third place, the Phylactolsematous larva possesses, in common with 

 all Gymnolsematous larvse, the following characteristics. The primaiy 

 polypides arise in both at a pole, and this pole is in both a prominent disk, 



