162 I. IJIMA : HEXACTINELLIDA. I. 



tliem. It accounts in part at least for the pronounced state of 

 protoplasmic continuity shown by the connective-tissue cells, for 

 there exists no other tissue to participate in bringing about the 

 connection of parts. It also makes intelligible the extension of 

 the lacunar spaces to such an extreme degree that nearly the 

 entire external surface of the chamber- wall is directly exposed 

 to the incurrent water, which fact affords a necessary condition 

 for the opening of prosopyles at all available places between 

 the choanocytes, without however aifecting the continuity 

 of these as a layer. Further, may it not be that the con- 

 nective-tissue matrix, when developed, requires an epithelial 

 limitation against the external world, and that by the absence 

 of the former, the differentiation of the latter is dispensed with 

 in the Hexactinellida ? I make this of course as a mere sugges- 

 tion. 



A knowledge of the developmental history of different parts 

 in the Hexactinellid body will be felt by all to be a great 

 desideratum. All that I know at present of the embryology is 

 connected with the larva of two Rossellid species which I have des- 

 cribed under the names of VUrolluIa fertile and Leucopsacus ortho- 

 dociis (Ijima, '98,). The larva will be described in details under 

 these species, accompanied with figures. Here let it be mentioned 

 that in its early developmental stage, it is spherical, covered exter- 

 nally by a flagellated cell-layer, and contains internally a mass of 

 cells. Remarkable is the fact that the first spicules, which arise 

 already in such a stage in the periphery of the internal cell-mass, 

 are stauractins. The fuller developed larva is spindle-shaped, 

 being thickest towards one end. I tbink this observation suffici- 

 ently warrants me in making the general statement that 

 the Hexactinellidan larva, as regards the arrangement of its 



