134 PROCEEDINGS OP THE AMERICAN ACADEMY. 



In preparing the bulb for the introduction of the gas, it is of great 

 importance to remove as far as possible all other adsorbed gases, of 

 which water vapor probably forms the most dangerous component. After 

 dry air had been passed through the bulb for several hours, it was ex- 

 hausted by means of a Sprengel pump and the outside was heated with 

 hot water to about 70° C. Dry gas was next admitted, and then the bulb 

 was again exhausted. This process was repeated three times, after which 

 dry gas was allowed to stream through the bulb from two to three hours. 

 The mercury used in the bulb was dried by heating it to 150° C, and dry 

 air was drawn through the rubber tube which connected the lower part 

 of the bulb with the levelling bulb for five hours. 



Calculation of Results. 



The value sought in the work is the coefficient of expansion of each 

 gas under constaut pressure ; but because it was impracticable to keep 

 the pressure absolutely constant a slight correction had to be applied 

 for changes. The consideration below shows that Boyle's law may safely 

 be used in referring these changes to changing volume. The actual vari- 

 ation of pressure in the case of hydrogen was about 2.5 mm., and Reg- 

 nault found that for this gas the product of pressure and volume varied 

 from constancy by less than one part in two thousand for an increase of 

 an atmosphere in pressure. This correction would amount, then, to an 

 error in volume of not over 0.0005 c.c. for a change of 3^ atmosphere 

 in these experiments, a quantity which is beyond their limit of accuracy. 

 Hence, for this very small range of pressure Boyle's law may be used as 



exact. 



Accordingly, in the previous paper, the equation used for calculating 



the results was as follows : 



(P + p -K) (V+ v) - PV 



« = " pyl 



in which a equals the expansion per degree Centigrade of unit volume at 

 0° C.; V equals the volume at 0° C; v, the increase in volume; t, the dif- 

 ference in temperature between the two baths ; P, the original pressure 

 at 0° ; p, the small increase of pressure in the gauge on the reading at 

 the higher temperature; and K, the constant value to be subtracted from 

 p due to the greater distance apart of the platinum points in the side tube 

 over the distance apart of those in the bulb. 



In this equation no account is taken of the fact that the mercury en- 

 closing the gas has a higher vapor tension at 32.4° than at 0°. This 





