584 HEPORT— 1 886. 



upon some cliemical reaction in which there is the most likelihood of a difference 

 in the behaviour of the elements under treatment, and peiforming it in an incom- 

 plete manner, so that only a certain fraction of the total bases present is separated, 

 the object being to get part of the material in the insoluble, and the rest in the 

 soluble, state. The operation must take place slowly, so as to allow the affinities 

 — which, by the nature of the case, are almost equally balanced — time to have free 

 play. Let us suppose that two earths are present, almost identical in chemical 

 properties, but diii'ering by an almost imperceptible variation in basicity. Add to 

 the very dilute solution dilute ammonia in such amount that it can only precipi- 

 tate half the bases present. The dihition must be such that a considerable time 

 elapses before the liquid begins to show turbidity, and several hours will have to 

 elapse before the full effect of the ammonia is complete. On filtering we have the 

 earths divided into two parts, and we can easily imagine that now there is a 

 slight difference in the basic value of the two portions of earth ; that in solution 

 being, by an almost imperceptible amount, more basic than that which the ammonia 

 has precipitated. This minute difference is made to accumulate by a systematic 

 process until it becomes perceptible by a chemical or physical test. 



In most methods of fractionation a rough sort of balance of affinities is arrived 

 at before long, beyond which further separation by the same method is difficult. 

 I have long noticed this action when fractionating with ammonia, with oxalic acid 

 and nitric and with formic acid. One valuable point which renders this fact 

 noteworthy is that the balance of affinities revealed by fractionation is not the 

 same with each method. It was in consequence of the experience gained in these 

 different methods of fractionation that in my paper read before the Royal Society, 

 June 10 last ('Chemical News,' vol. liv. p. 13), after stating that I had not been 

 able to separate didymium into Dr. Auer's two earths, I said, ' Probably didymium ' 

 will be found to split up in more than one direction according to the method 

 adopted.' 



The process adopted must vary according to the bodies under treatment. Frac- 

 tional crystallisation has yielded new results with didymium in the hands of Ur. 

 Auer von Welsbach ; and precipitation with formic acid, with ammonia, with 

 ammonium oxalate, crystallisation of the oxalates from strong nitric acid, and 

 fusing the nitrates and chlorides have all given me good results. 



Working with the samarskite earths, fractional precipitation with oxalic acid 

 separates first erbia, holmia, and thulia, then terbia, and lastly yttria. This is the 

 only method which is applicable for the separation of small quantities of terbia 

 from yttria. 



Fusing the nitrates separates ytterbia, erbia, holmia, and thulia from yttria. It 

 is not so applicable when terbia is present, and is most useful in purifying the 

 gadolinite earths. This process is the only one known for separating ytterbia 

 from yttria. 



The formic acid process is best for separating terbia, as terbic formate is diffi- 

 cultly soluble in water, the other formates being easily soluble. 



Selection must be made of these methods according to the mixture of earths 

 under treatment, changing the method as one earth or the other becomes concen- 

 trated on one side or thro^ii out on the other. Each operation must be repeated 

 very many times before even approximate purity is attained. The operations are 

 more analogous to the separation of members of homologous series of hydrocarbons 

 by fractional distillation than to the separations in mineral chemistry as ordinarily 

 adopted in the laboratory. 



When the balance of affinities, of which I spoke above, seems to be established, 

 the earths appear in the same proportion in the precipitate and the solution ; they 

 are thrown down by ammonia, and the precipitated earths are worked up by some 

 other process so as to alter the ratio between them, when the previous operation 

 can be again employed. 



Fractional precipitation by ammonia is the process generally adopted, for 

 although in some cases it is not so powerful as other processes it is more generally 

 applicable. 



The best plan is to add half the equivalent amount of precipitant to the liquid. 



