44 



I. IJIMA. 



The power of opening secondary oscular or of budding out daughter 

 persons seems to be widely spread, although neither of them are ever 

 formed in any great number. A daughter person first arises as a 

 cœcumdike outbulging of the wall, eventually to open an osculum at the 

 summit. 



The principal parenchymalia are exclusively long diactins of well- 

 known nature and arrangement. The intermedia are always of three 

 kinds : discoctasters, oxyhexasters, and microdiscohexasters. 



With respect to discoctasters, it is to be noted that the central 

 node contains a typical triaxial cross, which seems not to have been 

 observed before in this kind of hexactinellidan spicule. It is made plain- 

 ly visible when examined in glycerine or in any other medium of similar 

 refractive power. The six points of the cross are turned towards the 

 middle of the protuberant or otherwise somewhat concave space sur- 

 rounded by every four secondary principals. The terminals are always, 

 though often obsoletely, rough. The minute terminal discs are either 

 simply shaped like pin-head or toothed at the margin. It frequently 

 happens that deeply situated discoctasters are considerably larger than 

 those m the periphery. 



The oxyhexasters, which are the most abundant of all parenchymal 

 intermedia, have very short and sometimes almost entirely atrophied 

 principals. The terminals are smooth or more frequently rough. The 

 roughness may develop at their basal parts into prickles with centrally 

 turned points. Not unfrequently the oxyhexasters situated near the 

 dermal surface differ from those more deeply situated in having longer 

 principals and more slender and more numerous terminals, although 

 they seem to be connected by transitional forms. The most usual 

 number of terminals to a principal is two, and it very frequently happens 

 in such a bifurcated ray that one of the terminals is but very little deve- 

 loped or entirely absent. Thus I have seen cases in which a small rudi- 

 mentary and a normally developed terminal stood together on a princi- 

 pal. Then there are cases to be met with often enough, in which a 

 single terminal is joined to a principal by its crooked basal end. The 



