and elastic Fluids f and on the Measurement oJTemperahires. 379 



all the precautions which we have mentioned; and when we had 

 attained the stationary temperature and had noted the baro- 

 metrical height, we conveyed under the inferior extremity of the 

 vertical tube a ca})sule full of very dry mercury : the tui)e was al- 

 lowed to cool until the oil had resumed nearly the temperature of 

 the air. During the continuance of this cooling, the mercury 

 ascends into the vertical tube, and does not stop until the air con- 

 tained in the tube is completely cooled. The elastic force of this 

 air is then equal to the external pressure of the atmosphere di- 

 minished by the height of the column raised up : that of the warm 

 air was equal to the barometrical height observed at the instant 

 when the temperature was stationary. We might therefore con- 

 clude, by means of the law of Mariotte, what would have been the 

 dilatation of the air if it could have been dilated. In order to 

 render this process completelv exact, we ought to have kept an 

 account of the capillarv depression which the mercurv undergoes 

 in the straight tube into which it had risen. This depression had 

 been measured beforehand, and care had been taken to make 

 choice of a tube of such a calibre as not to admit of its varying 

 sensibly. 



In the second place, the volume of air did not remain exactly 

 constant ; the portion comprised in the vertical tube re-entered 

 in part into the great tube, in proportion as the column of 

 mercury rose, and this portion of air did not sensibly change its 

 temperature. It became necessary to calculate the influence of 

 these two causes, and make our observations undergo the correc- 

 tion flowing from it. This correction depended on the relation of 

 the capacity of the large tube with that of the small ; the calcu- 

 lation which gives it is besides too simple to render it necessary 

 to indicate it. We shall now give the results of one of the series 

 of experiments made by tiiis new process : all of them agree with 

 the observations made bv the first means, and they have entered 

 into the determination of the mean measurements formerly re- 

 lated. 



, , , , Milliinrtrcs.-| Capillary Correction. 



Length of the great tube . . 62 S. millimetres 

 Length of the small tube . . 57 J 4'5. 



