720 ON CONCRETIONS, SPICULES, ETC. [ch. 



so tends to assume in miniature the very same shape as the tetra- 

 hedron to whose corner it is attached. Out of these bourrelets, then, 

 the cusps at the four corners of our httle skeleton are formed. 



A large and curiously beautiful family of radiolarian (or "poly- 

 cystine") skeletons look, in a general way, Uke tiny hehnets or 

 Pickelhauben, with spike above, and three (or sometimes six) curved 

 lobes, Hke helmet-straps, below. We recognise a family likeness, 

 even a mathematical identity, between this figure and the last, for 

 both alike are based on a tetrahedral symmetry: the body of the 

 helmet corresponding to the inner vesicle of Callimitra, and the 

 spike and the three straps to the four edges which ran out from the 

 inner to the outer tetrahedron. In the one case an inner vesicle 



is surrounded by a tetrahedral figure whose outer walls, indeed, are 

 absent, but its edges remain, and so do the walls connecting the 

 outer and inner vesicles. In the other case the outer edges are 

 gone, and so are the filmy partition- walls, save parts which corre- 

 spond to the four internal edges between them*. There are apt 

 to be two slight discrepancies. The helmet is often of somewhat 

 complicated form, easily explained as due to the presence of two 

 superposed bubbles instead of one. The other apparent anomaly 

 is that the three helmet-straps are curved, while the corresponding 

 edges are straight in Plateau's figure of the regular tetrahedron. 

 But it is a paramount necessity (as we well know) for each set of 

 four edges in a system of fluid films to meet in a point two and two 

 at the Maraldi angle \ just as it is necessary for the faces to meet 



* Looking through Haeckel's very numerous figures, we see that now and then 

 something more is left than the mere edges of the partition-walls. 



