IX]. OF THE NASSELLARIAN SKELETON 715 



of hexagons. In the beautiful form which Haeckel calls Archi- 

 scenium the boundary edges disappear, the, four edges converging 

 on the median point are^ intensified, and only three of the six 

 convergent facets are retained; but, much as the two differ in 

 appearance, the geometry of this and of Callimitra remain essentially 

 the same. 



We learned also from Plateau that, just as a tetrahedral bubble 

 can be inserted within the tetrahedral skeleton or cage, so may a 

 cubical bubble be introduced within a cubical cage; and the edges 

 of the inn^er cube will be just so curved as to give the Maraldi angles 

 at the corners. We find among Haeckel's Radiolaria one (he calls 

 it Lithocubus geometricus) which precisely corresponds to the skeleton 

 A B 



Fig. 331. A, bubble suspended within i cubical cage. 

 B, Lithocubus geometricus Hkl. 



of this inner cubical bubble; and the httle spokes or spikes which 

 project from the corners are parts of the edges which once joined 

 the corners of the enclosing figure to those of the bubble within 

 (Fig. 331). 



Again, if we construct a cage in the form of an equilateral 

 triangular prism, and proceed as before, we shall probably see a 

 vertical edge in the centre of the prism connecting two nodes near 

 either end, in each of which the Maraldi figure is displayed. But 

 if we gradually shorten our prism there comes a point where the 

 two nodes disappear, a plane curvilinear triangle appears horizon- 

 tally in the middle of the figure, and at each of its three corners four 

 curved edges meet at the famihar angle. Here again we may insert 

 a central bubble, which will now take the form of a curvihnear 



