CLOSURE OF TUBULARIAN STEMS. 157 



a complete membrane. If the cut edge extends just above a 

 ridge, when the membrane begins to stretch across, a cavity is 

 left between the closing membrane and the ridge, which in some 

 cases looks like a canal within the edge of the closing membrane, 

 such as Godlewski describes (Fig. G). It maybe that his canals 

 are formed in this way and his figures might be so interpreted. 

 In the sections of Tnbnlaria crocca, I have found no canals formed 

 by the disintegration of endodermal cells in the edge of the closing 

 membrane. In the living material, no circulation has been ob- 

 served within the edges of the closing membrane indicating the 

 presence of canals, yet the circulation in the bottom of the open 

 cylinder was easily seen, so that the circulation is not something 

 peculiar to the canals of Godlewski. 



The pieces may be cut off in such a way that a ridge is left in 

 the middle of the half-cylinders. Sometimes it appears as though 

 the growing edge of one side joined such a ridge, but this appear- 

 ance was seen in but a few consecutive sections, showing that it 

 would not have interfered with natural closing, in further develop- 

 ment. If the piece is rather small, or the ridge large, the closing 

 membrane is prevented from bending in as much as usual. But 

 this middle ridge makes still greater complications in small pieces, 

 as here it may fill up the cavity formed by the closing membrane, 

 and cause an appearance of solidity. 



In many of the smaller pieces, the sections, at first sight, look 

 solid as though the endoderm cells had remained inactive, and 

 lay in the same position that they were in when the piece was 

 cut. It looks as though the ectoderm has grown over this solid 

 mass, thus making a solid core of endoderm with a thin covering 

 of ectoderm as Godlewski describes. But it is natural to expect 

 that if the endoderm cells lag only a little behind the ectoderm 

 cells in the closure of a large piece of a tubularian stem, they 

 would do the same in a small piece of the same stem. If the 

 work had been entirely with large and small pieces coming from 

 different stems or different portions of the same stem, this expecta- 

 tion might not be justifiable, but the pieces used were obliquely 

 split, large at one end and small at the other, the conditions for 

 both ends thus being nearly alike. A contraction, if the closing 

 is due to this, would be along the entire free edge regardless of 



