262 McMURRICH. [Vol. II. 



described and figured frequently, and reference may be made 

 to the figures given by von Heider ('95), Haddon and Shackle- 

 ton ('91), and myself ('89). The epithelium of this part of the 

 filament forms a rounded or crescentric layer resting upon a 

 somewhat T-shaped enlargement of the edge of the mesogloea. 

 The tips of the crescent extend to about the tips of the trans- 

 verse limb of the T, the outer surface of the limb being cov- 

 ered by a very different kind of epithelium, generally admitted 

 to be endodermal. The general surface of the mesogloea of 

 the mesentery immediately external to the attachment of the 

 /"-shaped enlargement is covered by a thick endodermal epithe- 

 lium, which, traced outwards, gradually diminishes in thickness 

 to pass into the ordinary epithelium of the mesentery. 



In Z. flos-mariuHS I found ('89) in this thickened epithelium 

 numerous foreign bodies, and suggested that it was a special 

 region for intracellular digestion. Haddon and Shackleton ('9la) 

 have described the thickening in Z. macgillivrayi, and von 

 Heider in Z. cJiiercJiiae, — the latter, however, taking exception 

 to my interpretation of its function, believing it to be the area 

 in which the reproductive elements will develop. This idea is 

 readily disproved by the examination of specimens in which the 

 gonads are developed. For instance, I have preparations of 

 Z. nymphaeus which show the thickened endoderm very dis- 

 tinctly crowded with foreign bodies, and, quite externally to 

 this, in the region where the endoderm has assumed its usual 

 low form, the sexual cells are found. On the other hand, my 

 interpretation is confirmed by Willem ('95), for it is in exactly 

 the region of the thickening that he finds an abundant inges- 

 tion of carmine particles in the Hexactiniae, these forms also 

 presenting thickenings of the endoderm, usually less pro- 

 nounced than in the zoanthids, immediately external to the 

 glandular streaks of the mesenterial filaments. 



It is very generally believed that the epithelium of the mid- 

 dle lobe of the ciliated bands is in direct continuity with the 

 epithelium of the glandular streak ; indeed, it has generally 

 been regarded as the upper part of the glandular streak. It 

 appeared that there was a marked difference in Z. sociatns 

 between this median epithelium and that of the glandular 



