EXCRETORY ORGAN OF FRESH- WATER POLYZOA. 347 



of the last cavity. It might be remarked here, further, that the 

 lophophornl cavity is separated from the rest of the body-cavity by a 

 sort of diaphragm on the oral side, while on the anal side this 

 boundary is by no means so sharp. The dotted lines in woodcut 1 

 show the limit of the diaphragm. An anal diaphragm, such as was 

 described by CoRi, seems not to exist in this genus. 



Thus a renewed investigfation convinces me of the correctness of 

 my former statement, as well as that of I^raem, that the cavity of the 

 ciliated tubes stands in open communicaticm with the lophophoral 

 cîivity. I reproduce here those passages that show briefly how I then 

 represented this connection. " The wall of these tubes are continua- 

 tions of the epithelial lining of the invaginalile portion of the endocyst. 

 They open below into the body-cavity by funnel-shaped openings." 

 " The shape of the fannel-like openings may l^e compared most appro- 

 priately with the obliquely cut end of a hollow^ tube." " As the two 

 tubes deviate from e;ich other below, a part of the perigastric space 

 appears l^etween them. This is the jjassage by which the cavity of 

 the epistome communicates with the perigastric space ' I can not 

 understand why CoRi describes the nephrostomes as opening into the 

 ' Metasomhöhle,' for the communication between the lophophond 

 cavity and the cavities of the ciliated tubes is too obvious to escape 

 attention. However, this difference in our statements mi übt have 

 arisen from the fact that the boundary between the lophophoral cavity 

 and the rest of the body-cavity on the anal side is necessarily arbitrary 

 and we may have fixed on different levels for it. 



After having thus determined the form and extension of the 

 body-cavity, we may now proceed to the discussion of the position 

 which the ciliated tubes occupy in a polypide. This can best be done 

 by referring to fig. 1 PI. XXXIV, a schematic drawing which has 

 been deduced from a study of cross and longitudinal sections. It repre- 



