774 
ROBERT PAYNE BIGELOW 
pitted epithelium which extends centrally into very large lateral 
pockets, where it covers only the floor of the pocket. In the dis- 
tribution of the pitted epithelium and the form of the lateral 
pockets the Chrysaora-stage of Dactylometra differs from Chry- 
saora of the same size. 
9. Sections of the rhopalia of all of these forms fail to afford 
any evidence for Hesse's theory of the intercellular origin of the 
otoliths. On the contrary, they appear to be formed within the 
cells, instead of between them. It is true, as Hesse says, that 
there is an unbroken gradation between the otolith cells and the 
columnar, endodermal epithelium o^f the rhopalial canal. The 
same gradation is found between the chorda-cells andt he colum- 
nar epithelium at the base of the solid tentacles from which the 
rhopalia are developed in the scyphistoma (see Bigelow, '00, 
pi. 37, figs. 53-56), but it is not evident what bearing this fact has 
upon the probable origin of the otoliths. 
10. The Pelagidae show no layer of nerve fibers underlying 
the endodermal epithelium and nerve fibers of the dorsal sensory 
groove do not appear to penetrate the mesogloea. In this region 
the mesogloea of Dactylometra appears to be quite structureless, 
although fibers are abundant in it elsewhere. There may be a 
nervous communication between the nerve centers in the sensory 
niche and the endoderm through the endodermal lamella, but if 
present its demonstration will require special methods of stain- 
ing, which were not employed in the present investigation. 
11. In considering these results due allowance must be made, 
of course, for individual variation and for liability to error due 
to shrinkage and distortion of specimens. There is also a possi- 
bility of confusing larval forms when they are not reared from 
the egg. Still I think that in all essential particulars the results 
given above will be confirmed by future investigations. 
