107 



and without to the same level, before he registered the vo- 

 lume of the included air. 



On the next day, after the apparatus was mounted, and 

 the four following ones, the volume of the dry air, its tem- 

 perature, and the existing pressure were accurately noted. 

 This pressure, which was measured by a portable barometer 

 of Newman's, having undergone a variety of corrections, for 

 the capacity of the cistern compared to that of the tube, for 

 the excess of the temperature of the quicksilver over 32°, for 

 capillarity, and for a constant error by which I found my 

 barometer affected, when compared with the standard instru- 

 ment in the Observatory of Trinity College, I reduced by 

 calculation in each instance the observed volume of air to 

 what it would be at 32°, and under a pressure of 30, using 

 for the expansion of air the corrected coefficient ^^, which 

 has resulted from the experiments of Rudberg, and thus 

 obtained the following numbers, which, it will be observed, 

 differ very httle from each other. 



1 91M1 



2 911-85 



3 910-21 



4 913-30 



5 911-72 



911-64, therefore, the mean of the five observations, may 

 be assumed as the true volume of the included dry air, at 

 32°, and under a pressure of 30. 



The volume of the dry air being determined, the next 

 step was to charge it with moisture. In order to accom- 

 plish this, the air vessel was lifted by means of the string, 

 so as that the mercury within should be about an inch 

 higher than the external mercury, and distilled water was 

 then poured into the upper cavity of the stopcock, so as 

 completely to fill it. The stopcock was now cautiously 

 turned, so as to admit the entrance of the moisture guttatim; 



