THE EXACT DETERMINATION OF ATOMIC 

 WEIGHTS BY PHYSICAL METHODS 



By H. F. V. LITTLE, A.R.C.S., B.Sc. 



The atomic weights of the elements are usually arrived at by 

 measuring their combining weights as precisely as refined 

 methods of chemical analysis or synthesis will allow and then 

 selecting those multiples which most nearly approach the 

 approximate atomic weights deduced with the aid of Avogadro's 

 theorem. This may be called the chemical method. 



There is, however, an alternative method of arriving at exact 

 atomic weights, namely, to develop processes for the accurate 

 determination of molecular weights. This may be called 

 the physical method. The method has been developed during 

 the last twenty years ; the present article is devoted to the 

 consideration of the results that have been obtained. 



With few exceptions, the only substances of which the 

 densities have been determined with a high degree of accuracy 

 are those which exist as gases at ordinary temperatures and 

 pressures ; these alone will be considered in the present article. 

 From the work of Rayleigh, Leduc, Morley, Gray and Guye 

 and his collaborators, it may be concluded that the methods of 

 preparing gases have been rendered so efficient and Regnault's 

 method of determining the densities of gases has been so 

 improved, that the densities of the commoner gases are now 

 known with an error not exceeding i part in 10,000. In 

 deducing molecular weights from these results, it is not 

 sufficient to assume the truth of Avogadro's hypothesis in its 

 primitive form ; the fact that Boyle's Law does not accurately 

 express the isothermal relationship between pressure and 

 volume in the case of any known gas and that the coefficients 

 oi expansion of gases are not exactly alike is proof that even 

 if, at some particular temperature and pressure, the relative 

 densities of gases were accurately proportional to their mole- 

 cular weights, at any other temperature and pressure this 



504 



