100 PROCEEDINGS OF THE AMERICAN ACADEMY 



IX. 



NOTE ON THE RELATIVE POSITION OF HIGH TEM- 

 PERATURE MELTING AND BOILING POINTS. 



By Carl Barus. 



Presented May 24, 1892. 



In my work on the measurement of high temperatures,* I largely- 

 availed myself of boiling points for calibration purposes. With 

 the aid of suitable apparatus, ebullition can be maintained for an 

 indefinite length of time, and the boiling point is probably less 

 influenced by the impurities of the substance. Apart from this, 

 one of them — the boiling point of zinc — has been so frequently re- 

 determined by different observers (Becquerel, Deville and Troost, 

 Violle, and myself) as to have become a veritable landmark in the 

 region of high temperature, and guaranteed with an accuracy of 

 about one per cent by actual air thermometry. 



Melting points, however, have as a rule been found indirectly, 

 and those most relied upon to-day are due to the unique calori- 

 metric work of Violle. I am not aware that anybody has recently 

 endeavored to compare melting and boiling points with the same 

 pyrometer, and with especial reference to the boiling point of zinc. 

 This I have had occasion to do, and the results of the survey made 

 are unfortunately disappointing. 



I made use of a platinum-iridioplatinum thermocouple, which 

 bad been compared with the porcelain air thermometer between 

 300° and 1,200°. The couple was thus found to be quite free from 

 anomalies, and to be quite regular in its thermal variations. It 

 placed the boiling point of zinc at 930°, or at the lowest estimate 

 at 925°, agreeing well with the classic results for this datum 

 referred to. 



Now, on using these same couples to determine the melting 

 points of a number of elements, among which silver, gold, and 

 copper need here alone be cited, I found that if Zn boils at 925°, 



* Bulletin U. S. Geological Survey, No. 54, 1889, Chap. II. 



