of Vapours to Mariotte and Gay-Lussac's Law. 289 



means of a gradual rarefaction of the air in the intermediate 

 part of the apparatus) does not recommend itself. We should 

 then run the risk of individual particles of fluid adhering to the 

 glass, without evaporating, perhaps much longer than would 

 correspond with the particular temperature and rarefaction of the 

 air. However, before any measurement was taken, we kept the 

 vapour for a long time dilated to such a volume that it obeyed 

 Mariotte's law at the denned temperature, and then allowed it 

 to proceed to a smaller volume by the gradual introduction of the 

 air, whereby a longer time was allowed for the acquisition of a con- 

 stant condition before each measurement of the vapour, so that we 

 had more confidence that we were observing circumstances which 

 actually corresponded to the external pressure and temperature. 



The determination of the pressure by the measurement of the 

 six mercury-levels could be made very accurately with the above- 

 named cathetometer. The cathetometer itself, which stood on a 

 strong fixed table, was daily corrected. 



Through the telescope of the cathetometer we could clearly 

 read off the volume of the vapour found in the calibrated tube to 

 the tenth part of a cubic centimetre. Having measured the 

 volume and pressure coexisting at the given temperature, we 

 then subtracted from the latter the pressure exercised under these 

 circumstances by the air-bubble, which had been determined 

 first of all. For each temperature, the volume v and the pres- 

 sure p of the vapour were measured from the maximum of ten- 

 sion, i. e. from the saturation of the vapour, to such a distance 

 from saturation that the vapour followed Mariotte's law. The 

 commencement of this latter was manifested by the constancy of 

 the product pv, which up to this time had been always increasing. 



§3. 



One word here as to the accuracy of the numbers thus arrived 

 at. Neither the apparatus nor the method of investigation can 

 admit of errors from any other source than the two usual ones, 

 which cannot be quite avoided, viz. slight variations of tempera- 

 ture in the bath, and slight irregularities in placing the catheto- 

 meter on the six quicksilver-levels. As to the first, it has been 

 already remarked that the variations of temperature arising du- 

 ring the measurement did not amount to "05 of a degree. The 

 error arising hence in the estimation of the tension (which was not 

 necessarily in strict accordance with the same temperature, yet at 

 most could vary from the specified temperature on either side to 

 the extent of 0*5 of a degree) is greater or smaller as the variation 

 of the tension with the temperature is greater or smaller. The 

 extreme case must be that of the maximum tension. Taking the 

 maximum tension of alcohol at 69° as 537*63, a variation of tem- 

 perature of 0°*05 would correspond to about 1 millim. However 

 that the errors which actually occurred never reached these amounts 



