The Skeleton and Classification of Calcareous Sponges. 67 



cells of all sponges, of placing themselves in the direction where the 

 alternate thrust and tension, caused by expansion and contraction 

 of the sponge, is maximal.* The carbonate of lime crystallised as 

 calcite, in three-rayed stars, the optic axis being radial to the sponge, 

 in the line of greatest pressure, and Professor Lewis has shown me an 

 example of calcite from Freiberg where the rhombohedra are etched 

 and infiltrated in three-rayed stars, whose rays bisect the rhombo- 

 hedral surfaces (fig. 8). These crystalline triradiates being them- 



FlG. 8 (from a drawing by Professor Lewis) .Crystal of calcite infiltrated by 

 opaque black matter (probably a metallic oxide), which forms a three-rayed 

 star, shown by the strong short lines in the diagram. The rays bisect the 

 angles made by the edges of the faces e, e', e", and therefore lie in the planes 

 of symmetry of the crystal. The black matter is in part underneath the faces, 

 so that the reflecting surface can be seen to be uninterrupted over it. The 

 specimen in the Cambridge Collection contains upwards of 120 crystals in 

 which the star is conspicuously seen. 



selves of the form advantageous for skeletal elements, the formative 

 cells lost their directive instinct, and are now passively carried on 

 the point of the -growing crystal, showing no trace of individual 

 character except in mere stalactitic modification of the crystal- 

 line contours. In the ancestors of Leucosulenia, Sycon, and Leu- 

 cai/dra, crystallisation appears to have taken place while as yet only 

 one cell (following Minchin) was concerned with the formation of 

 the calcareous element. As in the reticulate Ascons, the optic axis 

 lay in the line of greatest pressure, that is, vertical to the surface of 

 the flagellate gastral cavity ; but in the former case the principle of 



* [The first combination of the primitively isolated spicu'es would be to form 

 longitudinal sp'cule-flbres ; since, on account of the open osculuin, the surface- 

 tension parallel to the axis of the cloaca is only in a slight degree opposed by in- 

 ternal pressure. To oppose the transverse strain, which would follow on variation 

 in flagellar activity, such fibres must be united ; this I conceive to have been effected 

 in the simplest manner by two spicules, instead of one, applying themselves to the 

 end of a fibre, and diverging to meet the circumferential stress. Tims would be 

 produced longitudinal reticulations of branching spicule-fibre, resembling the 

 skeleton which appears in many thin- walled Moraxonida ; on this conception the 

 triradiate represents phylogenetically the fork of a pair of branches. July 23, 

 1898.] 



