284 DAY AND SOSMAN : EXPANSION COEFFICIENT OF GRAPHITE 



series of thermoelectric measurements and the most extensive 

 recent series of resistance thermometer measurements (Bureau of 

 Standards) on the one hand with the recent gas-thermometer 

 determinations made in this laboratory on the other, serves there- 

 fore as an efficient and independent check upon the trustworthi- 

 ness of the present gas thermometer scale between 0° and 1100°. 



PHYSICS. — The expansion coefficient of graphite. Arthur L. 

 Day and Robert B. Sosman, Geophysical Laboratory. 

 To appear in the Journal of Industrial and Engineering Chem- 

 istry. 



As a basis for the measurement of the specific volumes of sili- 

 cates at high temperatures we found it necessary to know approx- 

 imately the expansion coefficient of Acheson artificial graphite up 

 to 1600°. The published data on graphite were so inconsistent 

 and covered so small a temperature range, that we redetermined 

 this constant over the range from 0° to 1500°; the results obtained 

 from the material of the present paper. 



The expansion was measured directly on a bar 3 mm. by 5 mm, 

 by 700 mm. long, with a comparator which has been used in this 

 laboratory for measuring the expansion coefficient of alloys of 

 platinum with rhodium and iridium. 1 The comparator consists 

 of two micrometer microscopes held at a fixed distance of 500 mm. 

 apart by invar bars, and having mounted upon the system another 

 invar bar by which any changes in this fixed distance amounting 

 to 0.001 mm. or more can be detected. 



Two points on the graphite bar, 500 mm. apart, were marked by flat- 

 tened pieces of platinum wire bound around it, each having a fine line 

 cut at the middle of the wire. The bar was heated by sending a heavy 

 current through it, and a uniform temperature along its length was 

 obtained by slightly adjusting its thickness. It was surrounded by a 

 tube of Marquardt porcelain, and this again by a wide glass tube, each 

 free to expand independently of the bar. The space within the tube was 

 kept filled with carbon monoxide. 



In the reducing atmosphere a thermoelement could not be used. The 

 temperature was, therefore, found by noting the melting point of small 



1 Day and Sosman, Am. J. Sci. 26: 425-436. 1908. 29: 111-114. 1910. Car- 

 negie Pub. No. 157; 27-36, 61-63. 1911. 



