148 Holbom and Day — Melting Point of Gold. 



point somewhat above the melting point and coming back to 

 it again afterward. The stirring produced by the gas bubbling 

 through the molten metal is not the cause of this phenomenon, 

 for the cases where carbonic acid was introduced under like 

 conditions do not show it. 



Furthermore, where the oxygen atmosphere was introduced 

 the time-curves are always irregular and often even with 

 ordinary air. This suggests the possibility that in these cases 

 the melting metal takes on some oxygen. The effect of this 

 irregularity upon the melting temperature, however, is small 

 and scarcely exceeds the errors of observation. 



From the same gold which had served for the measurements 

 in the crucible, several grams were afterward drawn out into 

 thin wire 0*25 mm in diameter, and used for a series of control 

 observations by the wire method which it would be so desirable 

 to be able to use in the case of gold. The same thermo-element 

 was used for these also, and great care was taken that the two 

 wires of the element enclosing the short gold wire be as free 

 from strain as possible to avoid rupturing the gold before its 

 proper melting point. 



Five observations yielded the following values for the melt- 

 ing temperature: 



Sept. 26. 10206 MV 1064'3° 



10197 1063-6 

 10203 1064-1 

 10199 1063-7 



10198 1063-6 



The mean value, 1063'9°, differs only 0*4° from that obtained 

 by the crucible method and only 0*1° from the earlier results 

 obtained under the same conditions (see former paper, p. 190) 

 in which other gold was used. 



For the calibration of thermo-elements then, the determina- 

 tion of the melting point of gold by the wire method is per- 

 fectly trustworthy, and only about 0*03 gr. of gold are necessary 

 for the determination. 



If for any reason, however, the crucible method be preferred, 

 the neighboring melting point of copper in air is well adapted 

 for calibrations. Our determination of it was 1064*9. This 

 point, aside from the diminished expense, is more convenient 

 to determine than the gold-melting point by the crucible method, 

 on account of the greater latent heat of copper. 



Table II contains a time-curve for comparison with the gold, 

 which was observed on Oct. 1 with 370 gr. of copper in a 

 thin porcelain crucible. The current strength i in the heating 

 coil amounted to 5*2 amp., and was smaller than in any experi- 

 ment made with gold. 

 Charlottenburg, November, 1900. 



