No. 6.] THE MORPHOLOGY OF THE ACTINOZOA. 26 1 



ance of the lamellae noted by nearly all observers, and it is 

 interesting to note that Thorell in 1858, with his usual accu- 

 racy, described the arrangement of the ciliated cells in trans- 

 verse streaks in the ciliated bands of Metridiu7n diantJnis. 



The intermediate area cannot then be regarded as consisting 

 wholly of endoderm, since it is generally admitted that the cili- 

 ated cells are ectodermal. Is then the intermediate epithelium 

 endodermal and digestive.' This is a question difficult to 

 answer, but it may be said that the epithelium is certainly con- 

 tinuous with the stomatodaeal ectoderm above, and not with the 

 endoderm. I do not find it essentially glandular in Z. sociatus; 

 indeed, it contains relatively few gland cells in comparison with 

 the epithelium of the median lobe. Those which do occur, 

 however, are very different from the usual stomatodaeal glands 

 with clear contents, since they stain deeply and are packed 

 with granules. It is possible that such glands are digestive in 

 function ; they are especially abundant, as will be seen later, 

 in the glandular streak of the filament, but they also occur 

 here and there in the stomatodaeal ectoderm, and their occur- 

 rence in the intermediate epithelium cannot be accepted as 

 evidence of its endodermal nature. From the evidence at my 

 disposal, I am inclined to regard the intermediate epithelium 

 as being ectodermal, as is the rest of the ciliated band epithe- 

 lium, and think it erroneous to homologize it with the digest- 

 ive, or rather ingestive, area described by Willem, which is 

 unquestionably endodermal. I think, however, that intracel- 

 lular digestion does occur in this epithelium, as I have seen 

 imbedded in it particles which were neither zooxanthellae nor 

 normal constituents of the tissue, but which may have been 

 ingested food particles. 



2. T/ic Glandular Streak. 



Following a series of sections downwards below the level 

 indicated in Fig. 2A, one finds the lamellae of the ciliated 

 bands extending downwards for some distance, but they finally 

 disappear, the median lobe of the filament alone persisting. 

 The general appearance of the glandular streak has been 



