318 



kanolt: melting points of eefractory oxides 



melting point of the lime might be changed by the presence of 

 the tungsten. To confirm the results another method was used. 

 A tube made of lime was inserted into the furnace, being supported 

 by the upper end, which was at a low temperature, while the 

 lower end extended into the hot part of the furnace. The tube 

 was kept free from smoke by a gentle cm-rent of hydrogen, which 

 passed out thru a small oblique opening at the bottom. The 

 pyrometer was sighted down the tube and a heating curve was 

 plotted. Determinations made in this way were in close agree- 

 ment with those made in tungsten crucibles. It is believed that 

 this method will be very useful with other substances for which 

 it is difficult to find crucible materials. 



Alumina can be melted in a vacumn high enough to avoid all 

 smoke without vaporizing so rapidly as to cause serious trouble. 

 Determinations were made in this way with both tungsten and 

 graphite crucibles. TVlien the alumina was melted in graphite 

 it did not form appreciable carbide until after it had completely 

 melted and the melting point had been determined; and then 

 the contamination was not rapid. The results with graphite cru- 

 cibles and tungsten crucibles were in agreement. Determinations 

 were also made at atmospheric pressure with the apparatus con- 

 structed for magnesia, the smoke being removed by a current of 

 gas. These results agree with the others. 



Chromium oxide was melted in tungsten crucibles in a vacuum. 

 When melted in graphite it is rapidly reduced. 



The results are summarized in the following table: 



