192 I. IJIMA : HEXACTINELLIDA. I. 



13. Spermatozoa have not been observed. In a single 

 specimen tlieie were observed comparatively large, 

 ovum-like cells which seemed to arise from isolated 

 archjeocytes on all parts of the trabecular system, 

 liiit not from the archaeocyte-congeries. 



Development of Hexasters. 



Here will be recorded some fragmentary observations in 

 regard to certain developmental stages of hexasters, chiefly of the 

 floricome and the graphiocome in F. marnhalU. As illustrations 

 will serve PI. V, figs. 20-31 for the former, and figs. 32-3o 

 for the latter. 



Both these hexasters in different stages of development are 

 met with not uncommonly among the trabeculre of the subdermal 

 region, outside the choanosome. The frequency of their occur- 

 rence even in old specimens is explained by the fact that both 

 kinds of the hexasters are being constantly lost, or rather used 

 up, after full development and thus require to be replaced by 

 new formations (see p. o3). Young stages of the graphiocome 

 are by fiir the more common. 



The earliest stage observed by me in the hexaster develop- 

 ment was already in possession of the six principals in the 

 usual regular arrangement (fig. 29), but the terminals were not 

 yet even indicated. The said principals were of the same length 

 as in the fully developed state, but appreciably thinner. 



In any hexaster, so much of it as consists of the six principals 

 is undoubtedly identical with the ordinary hexactiii, the funda- 

 mental type of the Hexactinellidan spicules. This is clearly 

 shown by the invariable presence and the extent of the charac- 



