222 Mr. Calleridar on Experiments with a Platinum 



heat be very gradually and uniformly applied ; but it was 

 found that the more uniform and gradual the heating, the 

 more closely the melting-point agreed with the freezing-point. 

 In one case the whole of the metal was melted within half a 

 degree above the freezing-point. 



An experiment was then tried in order to ascertain the 

 effect of adding aluminium to the gold contaminated with 

 silver. The gold-silver alloy was raised to a temperature 

 about 10° above its melting-point, and 0*5 per cent, of Al 

 was added to it. and stirred in with a red-hot pipe-clay rod. 

 Considerable heat was evolved in the combination. As 

 the mass cooled, portions of the metal began to solidify at 

 the bottom at a temperature very little lower than the 

 freezing-point of gold, but there was no well-defined arrest 

 in the fall of temperature at any point. A layer of metal at 

 the top remained in a gritty and pasty condition, as tested by 

 the stirrer, and did not ultimately solidify till the mass had 

 cooled down to a temperature of about 600° C. 



The pyrometer was then melted out and tested for change 

 of zero by comparing it with a mercury-thermometer in a 

 vessel of cold water. No change of zero could be detected. 



An experiment was next tried on the freezing-point of a 

 nearly pure specimen of silver (99*97 per cent.). About 

 7 ounces of the metal were melted as before in the oxygen 

 furnace. The temperature of the mass did not become quite 

 steady during solidification, but fell slowly from p£ = 829"7 to 

 jtf=829-0. 



This result was somewhat unexpected. In my own expe- 

 riments above mentioned I had used some 27 ounces of silver, 

 the fineness of which was about 99'9 per cent. The freezing- 

 point had always been sharply marked within a tenth of a 

 degree ofp£=830 o, with this pyrometer. With the purer 

 specimen now used we expected to find a higher result. On 

 repeating the experiment, however, the same effect was 

 observed. It was noticed that the silver began to " spit " 

 violently about 8° below its freezing-point, and that its tem- 

 perature fell some 20° very rapidly while the spitting con- 

 tinued*. In my previous experiments the silver had been 

 melted in a gas-furnace with an ordinary air-blast and a 

 reducing-flame, and I had noticed very little spitting. Prof. 

 Roberts-Austen suggested that the lowering of the freezing- 



* The fall of temperature was, at this stage, too rapid to allow accurate 

 observations to be taken with the bridge-wire. Subsequent experiments 

 by a different method show that " spitting " does not usuall} T begin before 

 814° Pt., and that the fall of temperature is more rapid if there is no 

 spitting. 



