604 



SCIENCE PROGRESS 



which are equi-diffusive when present in equal weights, that of 

 greater molecular weight is the more diffusible, as the salts 

 should be dissolved in the ratio of 74/5 to 58*5 not that of equal 

 weights to be present in equal molecular proportions. Similarly, 

 sodium chloride and nitrate, instead of being about equi- 

 diffusive, as Graham thought, differ considerably, as the 

 diffusivity of 85 parts of the nitrate should be contrasted with 

 that of 58*5 of chloride. Sodium and magnesium sulphate also, 

 from this point of view, are more diffusive than sodium chloride. 

 It must be confessed that much of the work on diffusion 

 has not been carried out in a way that is calculated to give the 

 information required for the comparison of the behaviour of 

 substances under comparable conditions. As the water plays 

 its part in the process as well as the dissolved substance and 

 must also be thought of as diffusing, it should be used in some 

 constant molecular proportion, if we are to ascertain the 

 behaviour of a variety of substances relatively to water. No 

 measurements have been made hitherto which satisfy this 

 condition. The nearest approach are those of Long, who 

 determined the diffusivity of a considerable number of salts 

 from solutions containing one equivalent of the salt per litre. 

 The results quoted in the following table show the relative 

 number of molecules of the different salts that diffuse under 

 equal conditions, according to Long's determinations ; it is to be 

 noted, however, that the observations are not always in agree- 

 ment with those made by Graham : 



Relative Molecular Diffusivities of Salts 



These values are significant as showing that liquid diffusion, 

 unlike gaseous diffusion, is not a process in which the molecular 

 weights of the dissolved substance are determinative of the 

 rate of translation of the molecules. But a clue to the 



