Minerals of the Composition Mg 'Si 3 . 39T 



volatilizes and is absorbed by the thermoelements.* This con- 

 tamination produces a drop in the electromotive force of the 

 element which with continuous usage may result in a large 

 aggregate error. An unglazed porcelain tube is quite per- 

 meable to the vapor. For exact measurements above 1000°, 

 therefore, the thermoelement must be protected by an enclos- 

 ing tube of pure platinum in the hot portion of the furnace, 

 whenever a furnace coil containing iridium is used. 



The maximum difference of any two determinations is 2*5°. 

 The absolute value of the melting temperature is based on 

 extrapolation from the copper melting point. 



II. Orthorhombic Pyroxene. Enstatite. — So far as our 

 knowledge goes, pure enstatite, free from other polymorphic 

 forms, has never been prepared. Ebelmen and Hautefeuille 

 obtained some true enstatite which was afterwards proved to be 

 mixed with the monoclinic variety. The crystals of Meunier,f 

 who believed that he had synthetized enstatite, were shown by 

 Fouque and Levy to be monoclinic. Daubree's product,;}; 

 which was obtained by melting portions of meteorites, was, 

 according to his own statement, too opaque for optical examina- 

 tion, so that he could not say whether his crystals were ortho- 

 rhombic or monoclinic. 



The same experimenter also melted olivine with 15 per cent 

 silica, and obtained a crystalline cake, the interior of which con- 

 sisted of a fibrous mass which was un attacked by acids, and 

 "which had the properties of enstatite."§ Not only is no proof 

 of the orthorhombic nature of these fibers offered, but we 

 know from our own experiments that Daubree's conditions 

 must have yielded a product which consisted in the main of 

 the monoclinic form. Fouque and Levy, in the effort to repro- 

 duce meteorites by artificial means, succeeded in crystallizing 

 enstatite intermixed with a small amount of monoclinic pyrox- 

 ene in the following way : 12 grams silica, 3 grains mag- 

 nesia, and 5 to 5 - 5 grams ferric oxide, were melted and rapidly 

 cooled. The cake showed arborescent crystals which by pro- 

 longed reheating at a temperature a little above the melting 

 point of copper, developed into larger needles showing paral- 

 lel extinction. 



As previously stated, we also obtained enstatite in small 

 but variable quantities from melts of magnesium silicate, 

 except when.it was slowly cooled. The crystals of enstatite 

 thus formed were intercalated between the fibers of mono- 

 clinic pyroxene and could only be distinguished from the latter 



*W. P. White, " The Constancy of Platinum Thermoelements and Other 

 Thermoelement Problems," Phys. Eev., vol. xxii, p. 872, 1906. 

 f Synthese des Mineraux et des Roches, pp. 109, 111. 

 {Compt. Rendus, Ixii, 202, 1866. §Ibid., lxii, 374, 1866. 



