Theory of 



Fractional Precipitation, 



Solution. 



Precipitate. 







NiS0 4 . 



MnS0 4 . 



MS0 4 . 



MnS0 4 . 



Temp. 



e' 



B. 



A. 



y- 



X. 







grm. 



•1 



grm. 



•0953 



•5850 



12-9 



291 



•2 



•8 



•1852 



•4616 



13-6 



303 



•3 



•7 



•2799 



•3766 



12-5 



3-50 



•4 



•6 



•3588 



•2976 



13 



3-31 



•5 



•5 



•4305 



•2450 



13-6 



2-93 



•6 



•4 



•4788 



•1536 



12-8 



3-30 



•7 



•3 



•4991 



•1089 



17 



2-76 



•8 



•2 



•5584 



•0722 



17 



2-68 



•9 



•1 



•5841 



•0363 



15-2 



2-32 



6 



Mean valnfi of ratio -: ... 





-2-97 









e 







125 



These experiments are not all made at the same temperature, 

 and it is seen that generally — decreases with increase in the 

 temperature. 



Taking the mean value of - as being nearest the truth, it 



would appear that MnS0 4 resists the decomposing action of 

 sodic carbonate with a force 2*97 times greater than does 

 NiS0 4 , when both salts are simultaneously subjected to the 

 action of this reagent. If a solution had been employed con- 

 taining these salts in the ratio MnS0 4 : NiS0 4 = 2*97 : 1, and 

 an extremely small fraction precipitated, the ratio of the two in 

 the precipitate would have been 1:1; or, if the solution had 

 contained equal weights of the two sulphates, the small preci- 

 pitate would have consisted of MnS0 4 : NiS0 4 = l : 2*97; the 

 larger the precipitate in this latter case, the more nearly would 

 this ratio equal unity. In the experiment in the foregoing 

 table, in which equal masses (*5 grm.) were employed, the 

 fraction precipitated being large, the NiS0 4 is a little less than 

 twice the amount of MnS0 4 precipitated. 



Another interesting series of experiments by Mills and 

 Smith* relating to fractional precipitation, in which sulphates 

 of nickel and cobalt were employed, sodic hydrate being the 

 precipitant, are given in the following table. In applying 

 these results to the formulae, the total amount of cobalt preci- 

 pitated is taken, it being for this purpose immaterial whether 

 or no the precipitate consists of a basic salt. 



* Proc. Rov. Soc. vol. xxix. 



