144 



and with transverse flaws, having the matrix intervening 

 abruptly. Its fractures are longitudinal and splintery. The 

 columns are four-sided, with faces replacing the edges in the 

 centre of the angles : on one, two, or more sides, the ends 

 approach towards a pyramid (in such as I have seen) with 

 four rhomboidal faces. Among a tolerable quantity, I found 

 very few with crystallized terminations, as figured : the faces 

 however are very distinct. 



We find this fossil has been taken for a rubellite, and 

 Kirwan's description in a great measure accords with that 

 idea. See Kirw. v. 1. 288. but in many respects it has been 

 confounded with the titanite of Kirwan. See his description. 

 May the radiating variety be the substance of which Mac- 

 quart says the garnets are formed ? He describes it as con- 

 sisting of straight fibres diverging from a common centre. 

 See Kirw. v. 1. 261. Its common appearance resembles 

 garnet much, but it is not fusible by the blowpipe, whereas 

 garnet is fusible into a black enamel. 



Kirwan mentions red schorl, p. 271, and says rubellites 

 are also so called. Another substance resembling this, 

 according to the short description of Mr. Kirwan, was found 

 by Morveau in Poitou, v. 1. 336, which he presumed to be 

 adamantine spar. Again, as Haiiy observes, another men- 

 tioned by M. Morveau, found in Le Forez, resembles it 

 greatly, and which is of great hardness. See Kirw. 337. 



Hardness of ours nearly the same as that of spinelle. We 

 found that the harder spinelles would scratch it ; but the softer 

 ones are scratched by it. This seems undoubtedly the 

 ee Spath adamantin d'un rouge violet" ofBournon, which 

 he described in the year 1789 from specimens found in 

 Le Forez, (Journal de Physique 453.) and now considers as 

 a variety of corundum. Other authors have had a similar 

 idea. We here subjoin a part of his description: see 



