456 W. P. White — Melting Point Determination. 



The principal causes of oblique melting curves appear to be 

 the following : 



A. Primary, i. e., inherent in the substance itself. 



I. Time lag in the melting process (with very viscous 

 substances). 

 II. The presence of impurities. 



B. Secondary, i„ e., due to a failure of the apparatus to register 



truly the behavior of the substance. 



III. Inconstant heat supply. 



IV. The normal temperature gradient between the outside 



and inside of the melting charge. 

 V. Accidental irregularities in temperature distribution. 

 VI. Flow of heat along the thermoelement. 



VII. Electrical conductivity of the charge (in case the 



thermoelement is used bare). 



VIII. Inhomogeneity of the thermoelement. 



IX. Differentiation of the charge in crystallizing. 

 X. Radiation through the melting substance. 



I. Viscosity. — Day and Allen* discovered that albite and 

 orthoclase exhibit a sort of hysteresis in their melting, and the 

 same thing has since been found true of quartz. It is con- 

 nected with the great viscosity of the melted substance near its 

 melting point. The change of state takes place gradually, even 

 at temperatures many degrees (100 or more) above that at which 

 it will also occur, if time enough is given. Theoretically, this 

 effect must probably be regarded as characteristic of all sub- 

 stances, though of course it is too small to be perceived in 

 most. That it might generally be large in silicates, however, 

 did not at first seem at all unlikely, but such is not the case. 

 Although considerable portions of a charge of albite or ortho- 

 clase may remain unmelted after several hours exposure at a 

 temperature 150° or more above the point where melting begins, 

 diopside (to take one instance) has been heated three times at 

 rates varying from 4° to 13° per minute with results agreeing 

 to '1°. (The distinguishing sign of the presence of this phe- 

 nomenon is of course the variability of the melting tempera- 

 ture with the rate of heating.) Hysteresis in melting, there- 

 fore, is only an occasional cause of obliquity in melting curves.f 



II. Impurity. — The effect of impurity in diminishing tKe 

 sharpness of melting points is fairly familiar, especially to 

 organic chemists. But the character of the effect on the tem- 

 perature-time curve seems to have been little appreciated, and 

 its importance and magnitude have often been underestimated. 



* Isomorphism and Thermal Properties of the Feldspars, Publication No. 

 31, Carnegie Institution of Washington, pp. 50-54 ; this Journal (4), xix, 

 pp. 119-125, 1905. 



f If the hysteresis is large, however, the melting point procedure evidently 

 needs radical modification. See the next article, p. 483. 



