196 I. IJIMA : HEXACTINELLIDA. I. 



stages in the growth of the terminals, — that each growing ter- 

 minal is completely invested by an extremely thin protoplasmic 

 layer, specialized physiologically at least as the secretive matrix 

 and standing in direct continuity with the scleroblast-mass. Such 

 a layer however could never be clearly demonstrated. Never- 

 theless, from what I have said, the assumption seems not to be 

 unwarranted that the centrally situated scleroblast-mass is res- 

 ponsible for the development of the entire hexaster. The nuclei 

 in that mass should superintend, as it were, not only the initial 

 formation of the spicule but also the finishing up of the terminals. 



Here may be introduced the mention of a thin wall, sur- 

 rounding the scleroblast-mass but separated from it by a space 

 of some width (see figs.). For the sake of reference I may call 

 it * the capsule.' It is at first roundish or irregular in shape. 

 As the terminal perianths or sheaves grow in length, it is pushed 

 out by these, soon to become broken through, so that the greater 

 part of them comes to lie without the capsular wall and among 

 the trabeculse. 



The capsular wall, as seen in optic section, appears as an 

 irregular line, much interrupted by breaks in its course. Its 

 substance is of just the same appearance as the trabecule. A 

 few nuclei occur on or in it at quite irregular intervals. And, 

 by focussing the microscope so as to view the wall face on, it 

 can be demonstrated that we have to do not with a continuous 

 membrane, much less with an epithelium, but with a thin layer 

 of an irregularly meshed cobweb. Therefore I take the capsule to 

 be only a specially adapted part of the general trabecular system. 



A number of trabecule join the capsule on the outside and 

 keep it suspended in position. On the inside, a trabecula or 



