158 A. L. DAY MINERAL RELATIONS FROM LABORATORY VIEWPOINT 



on the appearance of the exposed surface in the crucible, and are working 

 with a closed furnace and a sensitive thermometer imbedded in the charge 

 which will record the temperature of the portion of tlie charge in contact 

 with it; melting will then begin in the outermost layers (figure 3), 

 and will proceed gradually toward the thermometer at the center. Here 

 the temperature record will show heat absorption from the beginning to 

 the end of the melting process, but only the melting of the innermost 

 layer which h in immediate contact with the thermometer gives the true 

 melting temperature of the substance. Such an observation gives the 

 appearance of a melting interval where none really exists. It can be 

 avoided by simply omitting the outer layers — that is, by using a much 



smaller quantity of material in the form 

 of a narrow cylindrical charge. If, in- 

 stead of using a crucible of conventiona] 

 form, we reduce the amount of material 

 to 1 or 2 grams in a crucible about the 

 size and shape of a lead pencil, and of 

 perhaps 1 inch in length (figure 4), the 

 same melting point, which appears much 

 displaced and distorted upon its curve in 

 the large crucible, will appear and re- 

 appear with the most satisfactory sharp- 

 ness with the small one, with the added 

 advantage that the experimental opera- 

 tions all become much simpler. By this 

 simple expedient measurements of the 

 melting temperature of pure diopside 

 have been repeatedly made which agree 

 within 1 degree at 1,391 degrees. 

 Substantially the same effect is produced by the presence of a small 

 amount of impurity. Traces of heat absorption, indicating the beginning 

 of the change of state, appear on the heating curve far below the melting 

 temperature of the pure substance, owing to the solvent action of the im- 

 purity and consequent change in the specific heat of the mixture. This 

 disturbing effect may be serious, or it may be insignificant, depending on 

 the thermal properties of the substances involved, their amount, and mu- 

 tual solubility. If the amount of impurity is very small (a few tenths 

 of a per cent), the absorption of the heat of fusion will proceed rapidly 

 to a reasonably constant maximum, the interpretation of which will not 

 be difficult after a few trials with widely different rates of heating. Very 

 rapid heating rninimizes the apparent disturbance, but may (in viscous 



Heating 



Figure 3. — An Electric Furnace of 

 usual Type with Crucible and 

 Thermoelement in Position 



