4:78 White — Melting Point Methods at High Temperatures. 



slowly in order that the liquid may have time to yield, and 

 not deform the metal, which has very little strength at high 

 temperatures. 



Each of these types of crucible has its advantages. With 

 the larger, the thermoelements may be changed or compared 

 during an observation, the melted charge may be prodded or 

 sowed with crystals, or a sample may be taken, before heating, 

 of the very mass of crystals whose melting-point is to be 

 observed. With the smaller : (1) Melting points are much 

 sharper, as pointed out in the preceding paper. (2) A far 

 more uniform and steady distribution of the external tem- 

 perature can be secured (by using shielding tubes around the 

 crucible). (3) The expense for material, and especially for 

 platinum, is far less — an important point with silicates, whose 

 removal after cooling rapidly uses up crucibles. (4) The cruci- 

 ble and charge can be removed and replaced again without 

 cooling the furnace. (5) By quickly inserting the charge in 

 the furnace or raising or lowering it when there, it can be sud- 

 denly brought to almost any desired temperature of crystalli- 

 zation. In particular, by removing the melted charge and 

 chilling it to glass, crystallization at some other temperature 

 may then be secured with certainty. 



In some early work on wollastonite* w r e considered our- 

 selves fortunate when once a large crucible, cooled in the fur- 

 nace, happened to cool below the inversion temperature before 

 crystallizing and thus to give a cake of solid wollastonite. 

 With the small crucibles this condition has since been repeated 

 at will and rapidly. 



Bare thermoelements dipping. into crucibles of about* 20 gr. 

 capacity have also been used where the heat absorption was 

 small and difficult of detection. Otherwise, their advantages 

 will be sufficiently clear from the foregoing. 



The special element described by Dr. Dayf has, as far as 

 its original purpose is concerned, been superseded by the small 

 crucible with the bare element. 



' III. The control element. — Since the thermal behavior of 

 a charge depends both on the charge and on the heat supplied to 

 it, some knowledge of the source of heat is always necessary. 

 To obtain it in delicate determinations, a "neutral body" has 

 often been placed in the furnace with the charge, whose behavior 

 is designed to show the effect produced by the furnace alone in 

 the absence of the peculiarities presented by the charge. The 

 control element as used in this laboratory differs from this 

 device in giving, not the effect of the furnace temperature on 

 a third body, but that temperature itself. The difference is 



* Allen, White and Wright, this Journal (4), xxi, 89, 1906. 

 f Day and Allen, loc. cit. 



