M. Schiel on the Atomic Weight of Silicon. 517 



periclase, and the other consisted of lustrous octahedra. They 

 were found on analysis to have the composition Fe 2 3 MgO, 

 and are true spinelles. 



Hausmannite. — By heating red oxide of manganese in hydro- 

 chloric acid gas, small dimetric octahedra of Hausmannite, 

 Mn 2 O 3 MnO, were obtained. 



Protoxide of Manganese. — This was prepared by reducing any 

 oxide of manganese in hydrogen, and heating it in the apparatus 

 with a little hydrogen and a few bubbles of hydrochloric acid 

 gas succeeding each other at long intervals. The small quantity 

 of the latter gas required is truly surprising, and it escapes from 

 the apparatus unaltered. The crystals obtained have a remark- 

 able lustre ; their colour is emerald-green, and they appear to be 

 highly refringent, but exercise no action on polarized light. 

 Their form is that of the cube-octahedron. 



The same chemist has made the following observations*. 

 When fluoride of silicon is passed over oxide of zinc at a high 

 temperature, a mixture is formed of silicate and fluoride of zinc, 

 which dissolve each other. The latter being volatile, qn being 

 heated the silicate is left in hexagonal prisms large enough to 

 be readily measured, by which, and by their analysis, they were 

 identified with native Willemite, 3ZnO, SiO 3 . 



Fluoride of zinc acting upon silica gave the same products ; so 

 that a small quantity of fluoride of silicon could mineralize an 

 indefinite quantity of silica and oxide of zinc. 



Daubree had stated that he had obtained Willemite and zircon 

 by the action of chloride of silicon upon oxide of zinc or of zir- 

 conia. Deville, who has repeated these experiments with care, 

 has found that neither Willemite nor zircon is formed; in fact 

 by passing chloride of zircon over Willemite this substance is 

 destroyed. This result might be expected; for the chlorides of 

 silicon, in acting upon mineral oxides, do so not only by their 

 chlorine, but by the metalloid, which exerts a powerful reducing 

 action ; and as the metallic chlorides formed under the influence 

 of the chloride of silicon never dissolve the silicates formed, there 

 is no reason why they should crystallize. The reverse is the case 

 with fluoride of silicon ; it is to the solvent effects of this sub- 

 stance on the silicates that its mineralizing properties are due. 

 A series of experiments on the action of chloride of silicon on 

 various metallic oxides, and which led to negative results, proved 

 the correctness of these views. 



Schiel has communicatedt a determination of the atomic 

 weight of silicon. 



* Comptes Rendus, June 1861. 

 f Liebig's Annalen, October 1861. 



