W. P. White — Melting Point Determination. 453 



Art. XLIV. — Melting Point Determination; by Walter 



P. White. 



[Introductory. — The recent great advances in pyrometry, to- 

 gether with the development of the electric furnace, have 

 given to many physical and physico-chemical determinations 

 at high temperatures almost the ease and certainty attainable 

 at ordinary temperatures. The appropriate special technic, 

 however, being of very recent development, is not yet gen- 

 erally familiar ; the importance, also, of the whole new and 

 fertile high-temperature field is still growing in appreciation. 

 It has therefore seemed wise that the methods developed at the 

 Geophysical Laboratory be published from time to time for 

 general information, aside from their immediate application 

 to our own work. In pursuance of this idea, two special papers 

 have already been published, treating of furnace construction 

 and of temperature measurement up to 1600° C* The present 

 two papers deal with the application 

 of such measurements to those 

 methods of physico-chemical thermal 

 analysis which, best known through 

 their revelations of the constitution 

 of metallic alloys, are now being 

 applied with equal success to the 

 minerals and allied compounds. 

 The first treats of melting phe- 

 nomena in general, and its conclu- 

 sions are not restricted to the high- 

 temperature field. The second 

 describes the furnace technic used 

 to realise the conditions treated in 

 the first.] 



Fig. 1. 



Fig. 1. Melting curves. A, 

 ideal ; B and C, actual curves 

 for silver and anorthite, reduced 



ism and Thermal Properties of 

 the Feldspars. 



The preeminent value of melting 

 ice as a temperature standard has 

 made familiar the great^ constancy ™ m 

 of the ideal melting point and its 

 independence of external tempera- 

 tures. The great majority of actual 



melting point determinations, however, fail to show this ideal 

 constancy, and display a melting interval, rather than a point. 

 If the temperature-time curve, A (fig. 1), represents an ideal 



*Day and Allen, Phys. Rev., xix, 177, 1904; W. P. White, Phys. Rev., 

 xxv, 334, 1907. Other papers from this laboratory incidentally treat of 

 methods, but in this respect are largely summarized (as well as supplemented) 

 by the present papers. 



