240 



bulletin: museum of comparative zoology. 



later), is always found to continue into the horizontal membrane, the 

 line of rupture would have to be supposed to be at the point x in the sec- 

 tion rej^resented by Figure 2. It will be seen that the layers to be 

 ruptured in this position would be the ectoderm and the test, or the line 

 at which the test terminates within the branchial orifice. But the extreme 

 delicacy of the ectoderm has already been pointed out ; hence one may 

 readily believe not only that rupture here would occur with ease, but also 

 that the severed edges of the ruptured laj^ers might be observed with 

 difficulty. As a matter of fact, one of the chief obstacles that T have 

 found to this interpretation is the absence of local evidence of rupture. 



.f -ec. :m 





:0y. a 



Fig. 2. — A schematic section of the animal corresponding to that shown in Fig. 1, 

 to show my interpretation of the relation of the several parts, as contrasted with Herdman's. 



Attention may be called to the circumstance that the wall within the 

 branchial orifice opposite the suj^jposed rupture is intact, as the figure 

 indicates ; and tliat consequently this rupture is restricted to the region 

 corresponding to the muscle bands, ?». b." 



My interpretation of the horizontal membrane finds strong confirma- 

 tion, it seems to me, in these statements by Moseley : " Tlie membrane 

 was observed to be attached to the inner surface of the test wall at the 

 intervals between the conical processes ; but the specimen was too much 

 injured to allow of the investigation of the extent and manner of its 

 attachment within the conical processes. It appeared to be attached 

 laterally on either side to the inner walls of these processes, and is prob- 

 ably reflected so as to line their cavities." And again: "No reflection 

 of the membrane over tlie inner siu-faces of the upper and lower walls of 

 the test was observed." 



Moseley speaks repeatedly of the more thickened central portion of 

 the horizontal membrane, tlie portion thus characterized being in general, 

 one is led to infer, the distinctly octagonal area shown in his figures of 



