476 White — Melting Point Methods at High Temperatures. 



somewhat dependent on the body itself. Even in a horizontal 

 direction, considerable differences of temperature occur. For 

 instance, a difference of over 4° has been found between two 

 points side by side within a small porcelain tube of only 8 mm 

 internal diameter in a furnace held steady at 1000°. In this 

 extreme case the tube was near the furnace wall, so that one side 

 was shielded from the cooling effect of the ends. With the 

 same tube in the center of the furnace the difference would not 

 reach 1°, though horizontal differences of 2° or so, often vari- 

 able from day to day, can hardly be prevented with large cruci- 

 bles, even when symmetrically located. The vertical differences 

 of course are greater. The furnace just mentioned was of our 

 "general utility" type, with working chamber 6 cm w T ide and 

 14 cm high. In a furnace built specially for uniform tempera- 

 ture with a higher and narrower cavity (4*5 x 23 cm ), fitted with 

 four horizontal partitions to shield off the cooling effect of the 

 ends, the temperature variation over a crucible 4 cm high was 

 20° at 500°, falling to 6° at 1500°. (Temperature differences 

 inside the furnace usually decrease with increase of tempera- 

 ture above 500°.) A small porcelain tube (4 cm .dianieterXl0 cm 

 long) in the center of the furnace chamber can be used to 

 reduce irregular vertical and horizontal inequalities of tempera- 

 ture when the crucible is small enough to go inside it. The 

 uniformity of temperature about the crucible is then, sufficient 

 for most melting-point determinations. For some other kinds 

 of work, notably gas-thermometry, where greater uniformity is 

 needed, other types of furnace have been developed. Con- 

 siderable success has been obtained by the use of auxiliary 

 coils'* and with a combination of furnace and bath, but these 

 are outside the field of the present paper. 



For melting-point determinations with substances liable to 

 be altered by oxygen, the air has been excluded by a glazed 

 porcelain tube running through the furnace from top to bot- 

 tom. With a tube 4 cm in inside diameter and 5*0 mrn thick, the 

 melting of nickel (about 1453°) has been readily and satisfac- 

 torily observed. f The temperature distribution inside such a 

 tube has not been studied. A platinum tube closed at one end 

 has been used at various times to exclude air and also a closed 

 porcelain cylinder.:): The method of controlling the atmos- 

 phere about the charge by the use of an air-tight inclosure 

 inside the furnace has, after trial of both, proved more effec- 



* Day and Clement, loc. cit., p. 412. See also, for a further account, a 

 forthcoming paper in this Journal on the gas thermometer, by Arthur L. Day 

 and E. B. Sosman. 



f Further details of this method will also appear in the forthcoming paper 

 on the gas-thermometer. 



% The Role of Water in Tremolite and Certain Other Minerals, E. T. Allen 

 and J. K. Clement, this Journal, xxvi, 103, 1908. 



