W. P. White — Melting Point Determination. 455 



In the middle of the charge is the thermometer, which in our 

 case has always been a thermoelement of some kind. This is 

 sometimes used bare, sometimes surrounded by porcelain and 

 platinum jackets, as here represented. The crucible is heated 

 by an electric resistance furnace, for which a storage battery 

 furnishes a very regular source of heat. An additional control 

 element, C, indicates the furnace temperature, and also, if 

 desired, permits of regulating it. 



Fig. 2. 



Fig. 2. Sectional view of ordinary melting point apparatus (half size). 

 C, control element. 



It is a familiar fact to workers in this field that freezing 

 points are often sharper and show much better agreement than 

 melting points. The reasons for this difference occupy much 

 of the present paper. One result of it is that freezing point 

 determinations are generally preferred, and most of the litera- 

 ture of the subject relates to observations made with falling 

 furnace temperature. Freezing points, however, are uncertain 

 or useless in substances where undercooling is marked or 

 crystallization sluggish. Since this is nearly always the case 

 in silicates, the melting point, in spite of its greater experi- 

 mental difficulty, is the only one used in our silicate work, and 

 is the point mainly in view in this paper. The special disad- 

 vantages of the melting curve would no doubt be largely over- 

 come by stirring, but effective stirring is difficult in many 

 cases, impossible in the rest, and has never been attempted 

 here. 



