MKDUR^. 19 



longitudinal canal-liko tubes, which run along the stomach, are probably the four 

 corners of an ordinary (luadrilutoral stomach which has become nearly blocked up 

 owing to the growth of the gonads. I do not think that they are permanent canals, 

 like the radial canals, but rather spaces in the stomach which have not become filled up 

 with cndoderm. At the same time it is possible that they might be used as channels 

 for the conveyance of nutriment to the developing gonads. 



Although the gonads have every appearance of being next to the endoderm, 

 without the intervention of a layer of mcsoglaea, and without a trace of ectoderm, still 

 there is no evidence that they are of endodermal origin. The gonads are shut off from 

 the endoderm by a very thin delicate membrane which may be a layer of mesoorljea. 

 As the gonads are fully ripe they have probably in the course of development absorbed 

 all the adjacent ectoderm cells. The position of the ripe gonads is certainly peculiar, 

 and a few young and intermediate stages were much wanted for tracing the 

 development. 



The sections of the stomach belonging to the specimen which has shed all its 

 gonads are also of interest. The positions of the shed gonads are marked out by 

 spaces, which are either straight simple cavities or tubular cavities more or less curved 

 (fig. 5). These cavities are lined with a well-marked ectoderm which has apparently 

 developed after the shedding of the gonads. The new ectoderm is continuous with the 

 old ectoderm on the outside of the stomach. 



The specimen wliich has quadrangular holes along.side the perradial ridges has 

 somewhat the appearance of having its gonads arranged in short, transverse folds, as 

 described in the other species of the genus. But it is after all only an external 

 resemblance. 



The four perradial canals are in direct communication with the stomach through 

 the interior of the "mesenteries" upim which the stomach is suspended, and which 

 form its cruciform base. The four interradial canals have no direct communication 

 with the stomach. They run nearly the whole length oi the sub-umbrellar cavity. 

 Some terminate at their proximal end, i.e., nearest to the top of the uml)rellar 

 cavity, eitlicr in a .straight point, or in slight diverticula (fig. 2), witliout any 

 communication with the stomach or the perradial canals. In this condition thev 

 have every resemblance to long centripetal canals wliich develop direct from the 

 circular canal. 



A few of the interradial canals at their proximal ends do communicate with the 

 perradial canals by means of irregular branches. Tlicre is cither a single braiicli 

 running to one of the adjacent perradial canals, or two oppo.site branches running to 

 both the adjacent canals. The union with the perradial canals is at the point where 

 the " mesenteries " are about to becomes radial canals. In one specimen none fif the 

 interradial canals show any connection with the perradial canals. But in another 

 specimen three of the interradial canals have a connection by means of an irregular 

 branch or branches. 



