Bowen and Andersen — Binary System MgO-SiO^. 497 



the liquidus is very steep close to the eutectic. It would then 

 be very flat close to silica, if 1625° is the melting point of cris- 

 tobalite. 



Previous Work on the System Mg^SiO^MgSiO^. — A recent 

 paper by Deleano presents the results of an investigation of the 

 system Mg 2 Si0 4 -MgSi0 3 .* Deleano found that all the composi- 

 tions close to MgSi0 3 soften at the same temperature (within 

 the error of measurement). These results are in accord with 

 our determination that there is no eutectic between MgSi0 3 and 

 Mg„Si0 4 but a dissociation-temperature at which all composi- 

 tions between D and forsterite (fig. 1) break up into liquid 

 and forsterite, though Deleano presents a diagram showing a 

 eutectic. 



Deleano concludes, on the basis of finding periclase in a melt 

 of forsterite composition, that forsterite is unstable at the melt- 

 ing-point. It is to be noted that the mixtures used were made 

 up by mixing silica and a calculated weight of MgC0 3 .3H 2 0. 

 This latter compound is very difficult to obtain free from basic 

 carbonate, the presence of which would result in an excess of 

 magnesia in the product. 



We made the forsterite mixture, as we did our other mix- 

 tures, by mixing silica and magnesia in calculated proportions, 

 the magnesia being obtained by igniting the precipitated car- 

 bonate to constant weight. When the mixture is so made and 

 when perfect mixing is accomplished by repeated grinding and 

 heating, the product may then be melted and crystallized 

 repeatedly and the crystalline product is always pure forsterite 

 without a trace of periclase. 



Optical Study. 



The product of each quenching experiment was examined in 

 powder form under the microscope and the optical properties 

 of the various phases determined. In this manner, not only 

 the nature of the crystalline phases is noted, but, at times, the 

 composition of the liquid phase may be obtained by determin- 

 ing the refractive index of the glass. The various optical con- 

 stants were measured on suitable material and are recorded 

 below. 



Periclase. — Periclase was the only form of MgO noted. It 

 occurs in the preparations, when embedded in glass, as rounded 

 grains of isotropic character and index 1*73. We have also 

 examined periclase made by the Norton Company by fusing 

 magnesite. It is in coarse granular lumps showing the excel- 

 lent cubic cleavage and the optical properties of periclase. The 

 hardness of this material is about equal to that of apatite (5) 

 and distinctly less than that of orthoclase (6). 



*Zs. anorg. Chem., lxxxiv, 423, 1914. 



