§11 OF ARTS AND SCIENCES. 4OI 



By inspection of the figures given, it will be seen that a 

 combined variation of about twenty per cent, in these two 

 pressures would annihilate completely the external pres- 

 sure; hence if the observed values lie between zero and 

 those calculated, the total error due to disturbing causes 

 cannot exceed twenty per cent. 



For the critical temperature, Z^, we have from §2, I. 

 (see § 6, init.) the formula, 



pr ^ pr fll_ (/q — n _T_\ 



° [p il-V) Tj 



Hence 





= 0.09566 



or, finally, 





'^V (i.5302)^(.5302) ^^ 



p: ^ ^^ ' ^^^ ' L—r 



o 



■^ = ■''''' {t) ^ 7-7' 



an equation by which the value of the critical temperature 

 may be determined when we know at any temperature T^ 

 the ratio of /' to 4. We have seen reason to expect that 

 our formulae are to be relied upon within ten or twenty 

 per cent, and shall hope to find, accordingly, later on, 

 that the values of the critical temperature, calculated by 

 this formula, are sufficiently close to those estimated by 

 CaEcniard de la Tour. 



§ II. The surface tension of a liquid is very easily 

 determined by means of the height to which it will rise 

 in a capillary tube, of known diameter, which it thoroughly 

 wets. It is believed that the surface tends to contract 

 with a perfectly measurable force, not depending upon 

 the depth of the liquid, but being the same for the thin- 



