Pyrometer on the Melting-points of Gold and Silver. 223 



point in the present case was probably due to the occlusion of 

 oxygen. 



The silver was therefore remelted cautiously with a smaller 

 supply of oxygen, and care was taken not to heat it many 

 degrees above its melting-point. The freezing-point was now 

 much more sharply defined at^ = 829°*8, and there was far 

 less spitting. 



The experiment was again repeated, keeping a small supply 

 of coal-gas on during cooling to prevent the absorption of 

 oxygen. The freezing-point was now found to be quite 

 clearly marked at p£ = 830°*l. 



The above observations show that the influence of the 

 occluded oxygen cannot be neglected, and, further, that it has 

 the effect, like many other impurities, of obliterating the 

 sharpness of the freezing-point, as well as of lowering it. 

 The experiments show, however, that it is possible to obtain 

 very consistent results if precautions are taken to prevent 

 the absorption of oxygen. 



The silver was subsequently remelted and 0*5 per cent, of 

 lead was added. The metal soon became covered with a film 

 of molten oxide, although some gas was left on to retard the 

 oxidation. The freezing-point of the alloy was fairly well 

 defined at pt = 827" 8. On taking out the pyrometer it was 

 noticed that the lead oxide had attacked the glaze on the 

 porcelain tube. It was therefore again tested in cold water, 

 but no change of zero could be detected. 



The temperatures so far have all been expressed on the 

 platinum scale. The reduction to the air-thermometer scale 

 is still, unfortunately, subject to considerable uncertainty. 



If we assume the empirical formula 



t- P t=8{tim\ 2 -t/ioo} 7 . . . . (d) 



which holds very accurately at temperatures up to 500° C, 

 we find for the freezing-point of silver the value £ = 981°"6 0. 

 according to this pyrometer (S=1'751). 



Becquerel and liiemsdyk, however, have both shown that 

 the melting-point of silver is nearly the same as the boiling- 

 point of zinc. The latter point has been determined by 

 Deville and Troost* with a porcelain air-thermometer with 

 great care. They find the figure 942° C. as the mean of 

 twenty-seven observations varying between 954° and 929°. 



* Comptes Rendus, xc. (1880), p. 773. Using hydrogen instead of air 

 they found values between 910° and 925° ; using CO.,, values between 

 1067° and 1079°. 



