370 



HEXRY A. KOWLAXD 



obtained by using the coefficient- of expansion of glass -0000264: at 

 100, or a -0036698, using the coefficient -0000287. 



The thermometers Nos. 6163, 6165, 6166, were always taken out of 

 the bath when the temperature of 40 was reached, except on Novem- 

 ber 14, when they remained in throughout the whole experiment. 



The thermometer readings are reduced to volumes by the tables of 

 calibration. 



TABLE X. IST SERIES, Nov. 14, 1877. 



The first four series, Tables X to XIII, were made with one bulb 

 to the air thermometer. A new bulb was now made, whose capacity 



was 192-0 c. cm., that of the old being 201-98 c. cm. The value of L. 



for the new bulb was -0058. 

 follows : 



June 8th 

 June 22d 

 June 25th 



]\Iean 



The values of li' and a were obtained as 



00366790 

 00366977 

 00366779 



0036685 



ft' 



753-876 

 753-805 

 753-837 



753-84 



This value of is calculated with the old coefficient for glass. The 

 new would have given -0036717. 



It now remains to determine from these experiments the most prob- 

 able values of the constants in the formula, comparing the air with 

 the mercurial thermometer. The formula is, as we have found, 



but I have generally used it in the following form: 



t=CV-f mt (100 /) (1 n (100 -f #)) , 



