I06 Ri;PORTS OF INVESTIGATIONS AND PROJECTS. 



melting temperatures. In the cases of bismuth and antimony the formation 

 of the monoxide was estabhshed by separation and analysis. The physical 

 behavior of aluminium indicated the formation of the monoxide of it also, but 

 it was not found possible to separate it from the metal by methods then 

 available. 



The essential experimental feature of the problem was the development of 

 a pressure-bomb in which temperatures as high as 700° could be maintained 

 for short intervals at these immense pressures. Several such bombs were 

 finally made upon the successful design which held perfectly tight under these 

 conditions and which may prove of great value in the study of those mineral 

 formations in which pressure is supposed to have played the principal role. 



(12) Melting-point determination. Walter P. White. Am. Journ. Sci. (4), vol. 28, 453. 

 1909. 



Actual melting- and freezing-point curves are nearly always oblique — that 

 is, they show, not the constant temperature called for by elementary theory, 

 but instead, an interval within which the temperature continuously rises or 

 falls. The prime cause of obliquity in melting curves is the obliquity of the 

 melting itself, due to impurity. The true melting-point is the high end of the 

 oblique melting interval. The melting hysteresis of some very viscous sub- 

 stances (mostly compounds of boron and silicon) is also an occasional cause 

 of obliquity. 



A number of causes of obliquity lie in the experimental determination of 

 the behavior of the melting and freezing substance. The determination of a 

 melting-curve necessarily involves two factors — temperature rise and heat 

 supply ; the latter depends on the temperature difference of furnace and melt- 

 ing charge ; if this varies, the curve is distorted in a way striking but easy to 

 correct. The most conspicuous example is where the furnace temperature is 

 allowed to rise or fall continuously, while the substance, melting or freezing, 

 remains nearly stationary. The freezing-point, coming at the beginning (in 

 time) of the interval, where temperature distribution in the charge is rela- 

 tively uniform, is easier to observe than the melting-point, but is inadmissible 

 in substances where undercooling is marked. 



The melting-point, coming at the end (in time) of the interval, is liable, 

 where stirring is not practiced, to obliquities resulting from uneven tem- 

 perature distribution : ( i ) due to the inevitable temperature difference be- 

 tween inside and outside of the charge, troublesome with large charges, neg- 

 ligible with small; (2) due to various irregularities in heat flow, less with 

 narrow charges and small therm o-elements, hardly ever over a degree or two ; 

 (3) due to conduction of heat down the thermo-element, also less with nar- 

 row charges and small thermo-elements, for which it is usually negligible, 

 but possibly amounting to several degrees with inclosed elements. 



Electrical conductivity in the melt produces an error in the reading of bare 

 thermo-elements, thus far negligible in small charges of salts. Contaminated 

 elements, besides reading false, read so as to increase obliquity. Differentia- 

 tion and diathermancy of the charge probably increase obliquity. Meltings 

 have been made above 800° agreeing with each other to 0.05°. In most 

 cases an experimental obliquity remains of from 0.5° to 1.5° whose cause is 

 still to be definitely determined. 



