314 Shepherd, JRanhin, Wright — Binary Systems of 



methods, we have made some preliminary fusions in order to 

 ascertain whether or not compounds formed between them. 

 It was necessary to use a furnace made of the purest artificial 

 graphite, the cone being supported on a graphite block which 

 did not form a part of the furnace resistance. 



It was found that in all fused mixtures of CaO and MgO the 

 two oxides crystallized out side by side, showing no evidences of 

 combination, from which we conclude that there is no compound 

 between the two. These temperatures are beyond the working 

 range of the iridium furnace. Even with the purest graphite, 

 small amounts of silica will get into the cones, giving a small 

 amount of birefracting material which is identical in all deter- 

 minable properties with Mg 2 Si0 4 . This would seem to account 

 for the birefracting material sometimes found when these 

 oxides are fused in the arc. The amount of this birefracting 

 substance was independent of the nominal composition. 



The location of the eutectic was rendered uncertain because 

 of the attack on the lime by the graphite. 



'Magnesia- Alumina. — This series, melted in graphite, gave 

 one well-formed compound, MgO.Al 2 3 , similar to the lime- 

 alumina series. There was no other compound formed. 

 Between per cent and 71*6 per cent A1 2 3 , the solid phases 

 are MgO and MgO.Al 2 3 . Between 71*6 per cent A1 2 3 and 

 100 per cent A1 2 3 , MgO.Al 2 3 and A1 2 3 are the solid phases. 

 The melting temperature of the MgO— MgO.Al 2 3 eutectic 

 is at about 1950°. 



These last two series are given only for the purpose of 

 guidance in calculating the possible effect of small amounts of 

 magnesia, which in commercial work are always present in the 

 lime-alumina-silica mixtures. In time, some one may devise a 

 furnace and containing vessel which will allow their more 

 precise examination. 



The results of the present thermal study may be summarized 

 as follows : 



1. There is but one compound (Al 2 Si0 5 ) of alumina and 

 silica stable in contact with the melt. This is the mineral 

 sillimanite. The two minerals andalusite and cyanite pass 

 slowly into sillimanite on being heated above 1300° C. 



2. There are four definite compounds of lime with alumina, 

 namelv, 3CaO.Al 2 3 ; 5Ca0.3Al 2 3 , melting point 1387° C. ; 

 CaO.Al 2 3 , melting point 1587° C. ; 3Ca0.oAl 3 . 



3. 3CaO.Al 2 3 and 3Ca0.5Al 2 3 have no true melting 

 point, but the former will be completely melted at about 1550° 

 and the latter at about 1725° C. 



I. Two of these compounds, 5Ca0.3Al 2 3 and 3Ca0.5Al 2 3 , 

 have an unstable form each, while 3Ca0.Al 2 3 , and probably 

 3Ca0.5Al 2 3 , are unstable at the melting point, i. e., do not 

 jDroduce a maximum on the liquidus. 



