90 BULLETIN OF THE 



surrounded by a circular fold, — the so-called mantle fold, — from which 

 it is separated by a circular groove, — the so-called mantle cavity. This 

 organ has a similar origin and fate in the two groups, as shown by Barrois. 



The following points of difference, however, must be recognized. 

 First, the absence of a definite ciliated ring, couronne (Barrois), of an 

 internal sac, and of a pyrifoi-m organ. But, as Barrois ('86, p. G7) has 

 shown, these are absent, or at least (OstroumofF, '87, pp. 182, 183) little 

 developed, in Cyclostomatous Bryozoa. The ciliated ring and pyriform 

 organ are doubtless organs connected with a free locomotive larval life, 

 which is greatly abbreviated in P])ylactola;mata. A second difference 

 exists in the fact that, while most Gymnolaematous larvse possess either, 

 rarely, (1) a functional alimentary tract, or (2) a mass of loose tissue 

 lying inside of the ectoderm, the PhylactoUiemata possess (3) a central 

 space lined by an epithelium placed next to the ectoderm. However 

 great the difference between the first and third conditions mentioned 

 above, it is to a large extent bridged over by the widespread existence of 

 the second. In some Cyclostomes, moreover, a similar condition to that 

 in Phylactolsemata seems to exist. Compare Metschnikoflf ('82, p. 310, 

 Taf. XX. Fig. 62). Lastly, the origin of two primary polypides, instead 

 of one, at the aboral pole, upon which Barrois has laid some stress, can- 

 not be considered a very strong objection to the homology, because in 

 reality the two polypides do not arise at the same time even in Pluma- 

 tella, and in Cristatella this difference is still more pronounced. In fact, 

 it is not the formation of tivo polypides which requires explanation, but 

 that of a young stock before hatching. 



There remains, therefore, to my mind, no serious objection to regard- 

 ing the larvse of Phylactolsemata and Gymnolsemata as having been 

 derived from some common ancestral larva, possessing, of course, more 

 points of resemblance to the Gymnola^raatous than to the Phylactola^ma- 

 tous type ; and therefore it is perfectly justifiable to interpret the latter 

 by aid of the former. 



Admitting the larvae to be homologous, we should expect the process 

 of gastrulation to be comparable throughout Ectoprocta. As a matter 

 of fact, we do find a great similarity in the earliest stages. Thus, the 

 first indication of the inner layer is the ingression of four cells at one 

 pole, which by multiplication give rise to a layer of cells lying inside of 

 the ectodei'm.* It is to the comparative study of the fate of this inner 



1 This has been shown for Membranipora (Tendra) by Repiachoff (78, pp. 

 416-420); for Alcyonidium polyoum by Harmer ('87, pp. 445, 446); for Bugula 

 by Vigelius ('86, p. 519) ; and for Cristatella in the present paper (page 68). 



