174 I. IJIMA : HEXACTINELLIDA. I. 



breaking up of the spherules into irregular granules, which 

 finally disappear, leaving behind pale-looldng, collapsed relics of 

 the cells with plainly visible nuclei. 



Of much the same general appearance as the thesocytes 

 known to me, seems to be the corresponding cells (' Knollen ' 

 cells) of the Arctic forms studied by Schulze Çgga; ig'a). In 

 respect of microchemical reactions of the inclosed reserve-matter 

 and also in the changes undergone by this, there exists a tolerably 

 close agreement not only among the Hexactinellidan thesocytes 

 themselves but also between them and those of Thenea as des- 

 cribed by SoLLAs ('88, pp. xxxix-xl). 



In E. marshaUi the thesocytes occur partly solitarily and 

 partly in massive groups. 



The solitarily lying thesocytes are present in a quite sparing 

 quantity, being found only here and there without regularity 

 upon both the external and the internal trabeculse (PI. V, fig. 

 36, th.). On the dermal as well as the gastral membrane there 

 are occasionally found irregularly shaped cells containing a 

 variable number of coarse and loose granules in the cell-body. 

 One such cell is represented in PL IV, fig. 23, th. I believe 

 Ave have here to do with old thesocytes in which the resorption 

 of the fat-like bodies has advanced to a great degree. 



On pp. 87-88, I have described irregularly shaped spots of 

 a deep orange-yellow color, visible in the choanosome of fresh 

 E. marshaUi. ]\Iicroscopical examination showed that these spots 

 consisted of thesocytes closely packed together. I consider these 

 cells as identical with those found isolatedly on the trabeculîB 

 not only because of the agreement in appearance, but also on 

 account of the fact that I have found signs of the former loosen- 



