424 



Dr. J. H. Long on the 



employed by me rests in the fact that it allows any variation 

 in the rate of diffusion, as in the above case, to be followed 

 from day to day. For the present I must leave the cause of 

 these variations unexplained. 



The question now presents itself, Do the phenomena of dif- 

 fusion bear any relation to other known physical pheno- 

 mena ? or, in other words, are the relations which molecules 

 of different substances bear to each other as regards diffusion 

 analogous to those depending on other molecular actions ? 

 From the Tables given above no simple relation is recogni- 

 zable ; but if instead of giving the results in grams of sub- 

 stance diffused, they be given in number of molecules diffused, 

 several very interesting relations will appear. Of course we 

 have no means of knowing the absolute number of molecules 

 which undergo a change of place ; but it is quite easy to ex- 

 press the relative numbers diffused in any given length of 

 time. 



In the Tables above the salt diffused is expressed in grams 

 per 24 hours ; now if each one of these values be divided 

 by the molecular weight of the salt, the quotients thus ob- 

 tained will be proportional to the numbers of molecules dif- 

 fused from the solutions in question in equal times. 



It will be remembered that the solutions submitted to 

 diffusion were normal, half-normal, and quarter-normal — 

 that is, were made to contain a definite number of mole- 

 cules per constant volume ; and it will be interesting to see 

 how many molecules diffused in equal times from solutions 

 which contained originally the same number. This is pre- 

 sented in the following Table. In order to have these values 

 in the form of whole numbers, the quotients as obtained above 

 have been simply multiplied by 1,000,000, which of course 

 does not alter their relative values. The results obtained from 

 salts diffused in half- or quarter-normal solution have been mul- 

 tiplied by two or four to make the comparison easier. This is 

 not absolutely accurate, as the amounts diffused are not quite 

 proportional to the concentrations ; but the error will not be 

 large. 



KC1 .. 



. 803 



CaCl ... 



429 



KI 



823 



SrN 2 6 ... 



552 



NH 4 C1.. 



. 689 



MgCf 2 ... 



392 



Nal 



672 



(NH 4 ) 2 S0 4 



724 



NaCl .. 



. 600 



CoCl 2 ... 



306 



NH 4 N0 3 



680 



Na o S0 4 ... 



678 



LiCl .. 



. 541 



NiCl, ... 



304 



KN0 3 ... 



607 



MgS0 4 ... 



348 



KCy .. 



. 767 



KBr" ... 



811 



NaN0 3 ... 



524 



ZnS0 4 ... 



332 



SrCl 2 .. 



. 432 



NH 4 Br... 



629 



LiN0 3 ... 



512 



CuS0 4 ... 



316 



BaCl 2 .. 



. 450 



NaBr ... 



509 



BaN 2 6 . . 



656 



MnS0 4 ... 



298 



This arrangement of the results shows several very inter- 



