488 White — Melting Point Methods at High Temperatures. 



It is in determinations of this sort that the neutral body is 

 particularly advantageous, 



If the inversion takes place promptly at a constant tempera- 

 ture (as for instance, with the quartz inversion at 575°) its 

 detection is evidently much easier. For the best results a 

 temperature reading should come immediately after the inver- 

 sion occurs, and before its effect has been dissipated by tem- 

 perature exchange with the furnace. The interval between 

 readings should be as short as possible with a rate about fast 

 enough to supply all the latent heat in one interval. A rather 

 small fraction of a calorie could probably be detected in this 

 way. 



Allen, Wright, and Clement* confronted this problem of 

 detecting small energy changes in their study of the inversions 

 of MgSi0 3 and finding the Frankenheim method inadequate 

 they developed a new one, whose essential feature was the 

 production of a very rapid rise of temperature in the substance 

 under examination by dropping it into a previously heated 

 furnace. The method here described is still the Frankenheim 

 method but with some improvements due to accumulated 

 experience. It is now several times as sensitive as is necessary 

 to detect one of the MgSi0 3 inversions. 



"V. Residual meltings. — The eutectic melting produced in 

 diopside by 2 per cent of pseudowollastonite is easily measur- 

 able on the smooth curve which precedes the melting point. 

 At the end of melting, however, as the last portions of solid 

 core or crust disappear in liquid, the temperature distribution 

 and variation in the crucible is irregular, and a rapid temper- 

 ature rise occurs. Hence, in the common case of eutectic mix- 

 tures containing a very small excess of one component, its detec- 

 tion may be impossible on an ordinary melting-point curve. 

 The accurate location of the eutectic, however, requires the 

 detection of these small residues as the composition of the 

 melt is varied in the adjacent region. For this purpose two 

 like crucibles may be placed symmetrically within the inner 

 tube, containing either two mixtures of slightly different com- 

 position whose relative nearness to the eutectic we wish to 

 determine, or two substances one of which is supposed to con- 

 tain and the other known not to contain the small residual 

 whose melting is being investigated. The temperature of 

 these is slowly and carefully raised until the (previously 

 measured) main melting point is just passed, an operation 

 requiring a little patience, but presenting no particular diffi- 

 culty. The furnace will then be stationary. The tempera- 

 tures are read, the heating current of the furnace is suddenly 

 increased and alternate readings of the two crucibles made every 

 15 seconds as long as necessary. After the suspected region has 

 been traversed, the furnace is brought to a (higher) constant 

 * This Journal, xxii, 415, 1906. 



