DETERMINATION OF ATOMIC WEIGHTS 509 



Compressibilities. — The determination of the values of Ao, A" . . . 

 is a problem that at the present time cannot be regarded as 

 solved in a perfectly satisfactory manner, except in the case of 



a few gases. From the definition of Aj^ given in equation 

 (2), it is clear that the value of A„ cannot be obtained directly 

 from compressibility measurements but must involve an extra- 

 polation from the lower pressure to zero pressure. The 

 uncertainty attaching to this process will be diminished in 

 proportion as the lowest pressure at which experimental 

 observations are made approaches zero ; but a limit is set to 

 the extent to which p a may be diminished by the fact that 



cCfr.MJ 



since the absolute error in measuring a pressure is inversely 

 proportional to the magnitude of the latter, it eventually becomes 

 so great that experiments at lower pressures are worthless 

 owing to experimental errors. 



The most direct and satisfactory method of determining A l 

 is to realise experimentally the o°C. isothermal of the gas for 

 pressures starting at one atmosphere and diminishing as far 

 as is consistent with trustworthy results. The results should 

 be expressed by stating the product pv as a function of the 

 pressure p and extrapolated to zero pressure for the value of 

 p Vo. The simplest plan is to extrapolate graphically. 



Unfortunately, accurate data of this character are limited to 

 the cases of oxygen and hydrogen chloride, for which gases 

 the admirable experiments of Gray and Burt (20) are available. 



