MUSEUM OF COMPxVRATIVE ZOOLOGY. 23 



tlie two groups can be harmonized, and at the same time the physiologi- 

 cal nature of the differences indicated, by putting the statement thus : — 

 In all Bryozoa, the formation of the alimentary tract begins at that end 

 which is turned towards the gemmiparous region (cf. page 28). 



The problem of the difference in the method of development of the 

 alimentary tract in Ectoprocta and Endoprocta, is the same as that of 

 the differences in the development of the alimentary tract of Triploblas- 

 tica in general. 



As is well known, the midgut in Triploblastica is produced by an in- 

 vagination whose mouth — the blastopore — comes to be in some cases 

 at the anterior part of the tract, in others at its posterior part. This 

 variation in the method of formation has been explained by the hypo- 

 thesis that the blastopore represents the opening into the gastro-vascular 

 cavity of Coelenterates, which is functionally both mouth and anus ; and 

 that as we find a physiological separation of the opening in many Coelente- 

 rates, so a morphological separation of the gastrula-opening into mouth 

 and anus by concrescence of the lips of the blastopore in the mid-oral line, 

 has occurred in the ontogeny of Triploblastica. In some cases both mouth 

 and anus arise by this process, in other cases only one organ, the other 

 arising secondarily, or (preferably) later. The part which arises later 

 might be regarded as a new formation, or, following Caldwell ('85, p. 23), 

 as derived from a part of the entoderm which had become separated 

 from the greater part in the separation of the two extremities of the 

 elongated lip of the blastopore to permit the placing of mouth and anus 

 at opposite poles of an elongated animal. 



The application of these facts and their explanations to the facts of 

 the formation of the alimentary tract in the Bryozoan polypide is evi- 

 dent. Gastrulation takes place not in the act of first invagination of the 

 inner layer of the bud, but in a secondary invagination from the bottom 

 of the first formed sac. The blastopore does not lie on the surface of 

 the body wall, but has been carried below the surface, and its position is 

 indicated by the plane of separation of alimentary tract and atrium, 

 where the roof of the gut and the floor of the atrium have been produced^ 

 by concrescence of the lips of the blastopore. The atrium then is in 

 no way lined by entoderm ; it is merely a precociously developed, pro- 

 tecting pocket of the body wall, which occurs in that region in which 

 invagination of the entoderm is to take place. The primary atrial open- 

 ing is not at all the blastopore, as some authors have called it. 



1 Compare Seeliger, '89* pp. 181. 182 ; so also in Paludicella, see my earlier 

 paper, '91, p. 19, and probablj' in Pliylactolsemata, cf. Kraepelin, '92, p. 33. 



