to Calcium and Magnesium Metasilicates. 3 



(magnesia, calcium carbonate, quartz), its loss would be dispro- 

 portionate to its weight in the mixture. But the analysis of 

 the 60 per cent MgSi0 3 mixture, in which the magnesia was 

 slightly high, shows that the suspicion was unfounded. 



Found Cal. 



1. 10#MgSiO 3 ._ Si0 2 52-75 52-67 



90$ CaSi0 3 Caa ._ 43-52 43-33 



MgO--- -- 4-06 4-00 



Fe 2 3 etc '11 



100-44 100-00 



2. 60$MgSiO 3 Si0 2 56*80 56'73 



40$ CaSi0 3 CaO 19-12 19-26 



MgO 24-11 24-01 



Fe 2 3 etc. '13 



100-16 100-00 



Thermal Study. — The thermal study was carried on, as in 

 other published work from this laboratory, by Frankenheim's 

 method, melting point curves being used exclusively. Freez- 

 ing-point curves are unreliable on account of the undercooling, 

 which, even in substances which crystallize as readily as the 

 metasilicates of calcium and magnesium, is often considerable. 

 The mixture under investigation was heated in a platinum 

 crucible, in an electric resistance furnace fed from a storage 

 battery, the temperatures were read to tenths of a degree by a 

 Le Chatelier thermoelement and a potentiometer, and evalu- 

 ated in terms of a standard saturated cadmium cell. 



Since the work here described continued over a considerable 

 period of time, the methods used varied somewhat as succes- 

 sive improvements were introduced. In the earlier portion of 

 it, charges of about 25 grams were used, and the thermoelement 

 was protected by a porcelain tube with a platinum jacket. A 

 control element was also employed to detect any irregularity in 

 the heat supply.* Soon after the work was begun, the practice 

 was also instituted of comparing the thermoelements with 

 standards in order to correct for their deterioration, f due to 

 contamination with iridium from the platinum crucibles and 

 furnace coil. This has now been greatly reduced by using 

 specially pure platinum within the furnace, but occasional com- 

 parisons with standard elements are still necessary 4 Fre- 

 quently the thermoelements were inserted together in the 

 porcelain tube and used to read the same melting point. The 



*This Journal, xxi, 94, 1906. 



•{•The Constancy of Thermoelements, W. P. White. Phys. Zeitschr., viii, 

 325, 1907. 

 %W. P. White, Phys. Rev., xxv, 336, 1907. 



