156 A. L. DAY MINERAL RELATIONS FROM LABORATORY VIEWPOINT 



these types shall be examined in the vicinity of their formation tempera- 

 ture. Here the experience of the laboratory points out two situations 

 Avhich lead to confusion. The first is the fact, not hitherto mentioned, 

 that the addition of impurity, even of minute proportions, to a pure min- 

 eral not only has the effect of displacing its melting point, but also of 

 stretching it out along the melting curve in such a way as to conceal its 

 exact location upon the temperature record of the thermometer. The 

 second is the frequent difficulty in establishing equilibrium during the 

 operation under investigation. 



Melting ^Toints'^ and melting Intervals 



Among the somewhat voluminous records of mineral melting points 

 which have come out of European laboratories in recent years, one feature 

 has reappeared persistently, namely, the inability of the observer to fix 

 upon a single temperature for the melting point of a mineral under observa- 

 tion ; instead, he almost invariably finds and attempts to fix the boundaries 

 of a melting interval, often of 30 or 40 degrees (actually 50 degrees in the 

 adjacent feldspar curves (figure 2) measured in this laboratory), through- 

 out which the phenomenon of change of state appears more or less con- 

 tinuous. Physical and physico-chemical experience offers reasons for 

 this behavior which may in part explain, but which do not always help to 

 remedy the difficulty. If the material under investigation is a solid solu- 

 tion or isomorphous mixture (as in the feldspars), we know that the phe- 

 nomenon of melting includes a continuous change of composition which 

 is actually distributed over a considerable range of temperature, so that 

 the appearance of slow melting is not here misleading, but is the correct 

 record of its actual progress. The true relations here require to be ascer- 

 tained by establishing, either through chemical analysis or the micro- 

 scopic examination of a thin section of the material after solidification, 

 the order in which the crystals of different composition have formed and 

 the limits of variation in their composition.^* 



If, however, the substance under investigation is a chemically pure 

 compound the composition of which does not change during the change of 

 state, the reasons for the change being distributed over a temperature 

 interval are of a more accidental character, and should in most cases be 

 capable of elimination by careful study. Suppose, for example, we ex- 



" The general cases of this kind (see also page 166), together with various exceptions 

 encountered in the study of the feldspars, were discussed, both from the practical and 

 theoretical standpoint, in the record of the first investigation made in the Geophysical 

 Laboratory, from which figure 2 is taken. (The Isomorphism and Thermal Properties 

 of the Feldspars, Publication No. 31, Carnegie Institution of Washington.) The discus- 

 sion is too long and too severely physico-chemical for reproduction here. 



