238 BULLETIN : MUSEUM OF COMPAEATIVE ZOOLOGY. 



chial from the perihranchial, or ventral chamber should be composed of 

 aa epithelium constitutiug each surface layer, with something of the 

 mantle (blood-spaces at least) between. But here, likewise, examina- 

 tion of microtome sections fails to reveal such structure as would be 

 expected. The tissues are so highly specialized in the adult state of the 

 animal tliat it is difficult if not impossible to say definitely just what we 

 have before us. Herdman has described a squamous ejDithelium as 

 extending over the "general surface of the membrane," without, how- 

 ever, specifying which surface is thus covered. Such an epithelium is 

 undoubtedly present, but on one surface of the membrane only, and that 

 the ventral or deeper surface. The relation of tlie different elements 

 entering into the structure of the membrane can be particularly well 

 made out by examining a flat preparation from the portion in the base of 

 the arms where the muscle fibres are v\^ell developed. Seen from the 

 dorsal surface the squamous epithelium is found at a deeper level con- 

 siderably than the fibres ; and from the level corresponding to the fibres 

 down to that of the epithelium, numerous cellular elements not consti- 

 tuting a uniform layer, but composed of several kinds of cells, some 

 lai'ge and spherical, others smaller and more or less spindle-shaped, are 

 present. Microtome sections show the epithelial layer to be exceedingh'- 

 thin, and give the impression that the layer is interrupted in places. 

 This latter is probably not in reality true. The flat preparations exam- 

 ined give no intimation of such a state. Sections of the membrane show 

 in addition to its cellular constituents a considerable quantity of more 

 or less homogeneous, or somewhat fibrous material strongly resembling 

 test. This occupies in general the surface opposite the epithelium, i, e. 

 the dorsal surface. 



My interpretation of the membrane is that in life it was closely ad- 

 herent throughout by its dorsal surface to the test wall of the oral disc, 

 and became separated from the latter only on the death of the animal. 

 This would mean that the large dorsal chamber of the oral disc, 7'egarded 

 by Moseley and Herdman as branchial, is an artifact. On this view the 

 homogeneous test-like material mentioned above as entering into the 

 structure of the membrane would be accounted for by supposing that 

 the rupture plane was within the test for a short distance. The i-agged 

 character of the dorsal surface of the membrane, as seen on the sections, 

 confirms this interpretation. If this is right, the ectoderm of the region 

 involved should be present in the membrane dorsal to the muscle fibres 

 where these exist. I am unable to recognize anything that can with 

 certainty be regarded as such a layer, either in this membrane or in the 



