102 I. IJIMA : HEXACTINELLIDA. I. 



should here mention the rosette I have figured on PI. IV, fig. 21. 

 It occurs quite rarely and solitarily but apparently is constant 

 alike in both young and old specimens. So far as I know, it 

 seems to be peculiar to the present species. The diameter meas- 

 ures 40-45 ,«. The principal rays resemble those of graphio- 

 comes ; the discs at their ends are lens-like or prominently con- 

 vex on the outer surface. The terminals are exceedingly fine, 

 about as long as or somewhat longer than the principals ; they 

 are pointed at the outer end and arise closely together from all parts 

 of the outer disc-surfiice. The peripherally situated terminals in 

 each tuft are slightly but distinctly flaring, so that the tuft may 

 be said to be campanulate. The entire rosette looks not unlike 

 a plumicome or graphiocome in an early stage of its development 

 in which the terminals are still very short. However, it differs 

 from the former in that the terminals in a tuft are all nearly 

 equally long ; and from the latter it differs in having the termi- 

 nal tuft expanded into a bell-like form. The rosette above 

 described is found only at such long intervals that it requires a 

 close study of preparations in order to come across one. 



The spiculation of the sieve-plate deserves special notice in 

 respect of a few points. Unlike E. imperialls and many other 

 species, the predominant elements of the sieve-plate parenchymalia 

 are thetactins, instead of diactius. The tlietactins furnish both 

 the principalia and the accessoria, the latter also containing 

 elongate hexactins, pentactins, &c. The parenchymalia, in forming 

 the beams, are disposed in several loose or compact strands, 

 between which are left sufficient spaces for the location of the 

 much folded chamber-layer and of small excurrent canals opening 

 on the i;astral side. All of the four kinds of rosettes found in 



