198 I. IJIMA : HEXACTINELLIDA. I. 



some of the terminals, sometimes lying in direct contact with, 

 these. — Among the floricomes which have reached, in their 

 emigration, the tips of dermal hilt-rays, it is quite exceptional 

 to meet with one still retaining the scleroblast-mass, though a 

 number of isolated nuclei are usually found attached to the 

 different parts (fig. 36). 



With regard to the graphiocome development, the stage 

 shown in fig. 32 is one of the earliest I have seen. It is at 

 once referable to that form of the hexasters, and not to the 

 floricome, by the somewhat larger size of the scleroblast-mass, by 

 the greater number of nuclei in that mass, by the stouter prin- 

 cipals and finally by the planoconvex disc at the outer end of 

 each principal. The outer convex surface of the disc just 

 mentioned is roughend by minute tubercles, which I hold to be 

 the very beginnings of the formation of the rhaphidial termi- 

 nals. — In a slightly more advanced stage, the microtubercles are 

 developed into spiny processes. Then, the spicule is scarcely 

 distinguishable in form from the portion of an old graphiocome 

 remaining after the loss of the rhaphides (see pp. 53, 77, 101 ; 

 PL IV, fig. 20), except for the fact that in the young state there 

 is invariably present the scleroblast-mass, which is always want- 

 ing to old graphiocomes. 



Figs. 33 and 34 represent two stages in which the elonga- 

 tion of the terminal sheaf has advanced to different desirees — to 

 about 18 //. length in the one case and to about half the full 

 length in the other. A number of other cases of growing termi- 

 nal sheaves were measured and I may say I have found these 

 in all grades of length, from only a few /^ up to 100," and more. 

 The berry-like cluster of scleroblasts as well as the capsule re- 

 main visible all the while the graphiocome is growing ; afterwards, 



