482 



Mr. J. E. Petavel. An Experimental 



As far as I am aware only one research has been carried out with a 

 view of repeating the Violle standard on anything like the scale 

 originally used. This research was made at the Reichsanstalt* in 

 Berlin. The work proved unsuccessful, and a detailed account of the 

 experiments has never been published. This renders it very difficult 

 to suggest a reason why reliable results were not obtained. 



In the case of molten platinum, owing to the high rate at which heat 

 is being dissipated, it is quite possible for one part of the mass to be 

 liquid while another part only a few millimetres distant is considerably 

 below the temperature of solidification. This source of error should be 

 minimised by the choice of suitable experimental conditions, and by 

 reducing the results by some method similar to the one given below. 

 Again, all the magnesia or lime bricks I have been able to secure con- 

 tain a certain proportion of silica. This, as will be pointed out later, 

 is sufficient to render them useless for the purpose in view. In the 

 account of the Reichsanstalt experiments this difficulty is not men- 

 tioned. 



In common with all other pure substances, platinum has a constant 

 freezing point ; the length of time, however, during which the con- 

 stancy of temperature will be maintained, depends mainly on three 

 factors. 



If we take D to represent the quantity of heat dissipated by the 

 platinum per unit time when at its temperature of solidification, H the 

 heat supplied to the metal per unit time, and L the total latent heat of 

 the mass, and supposing for the present that the thermal conductivity 

 is very great, the time during which the temperature will remain con- 

 stant will be — 



For the object we have in view the time of constancy must be made 

 as great as possible. 



One method of increasing the time t is to make H very nearly equal 

 to D, or, in, other words, to supply heat to the molten metal at nearly 

 the same rate as the metal is losing it. We may supply the necessary 

 quantity of heat H by means of an electric current ; or, using a large 

 surface of metal, we may keep the blow-pipe going on one part of the 

 surface while the observations are being made on some other portion. 



If we abandon the idea of supplying heat during the time the obser- 

 vations are being taken, the formula becomes — 



* ' Thatigskeitsberichte der Keichsanstalt,' 1890 and 1892—1894, or ' Zeit- 

 schrift fur Instrumentenkunde,' vol. 11, p. 149, 1891, and toI. 14, p. 266, 1894. 



