THE SEPARATION OF THE RARE EARTHS 569 



Place in a beaker at least three inches in diameter and examine with a 

 spectroscope of low power for absorption bands. Probably the bands of 

 neodymium and " erbium " will appear. Separate the first by sulphate 

 of sodium as usual, and the last by ferrocyanide of potassium from an 

 acid solution as above. The filtrate will then contain the pure yttrium 

 a, whose calcined oxalate will be pure white without trace of yellow. 

 After separation of iron, calcium, and possibly manganese, the earth will 

 be a pure element as far as I can tell spectroscopically. However, like 

 Zr, Fe and many other substances, the addition of Na or K to the elec- 

 tric arc while obtaining the spectrum will change the intensity of cer- 

 tain lines of the spectrum, while others are unchanged. If this is con- 

 sidered as evidence of the existence of two elements, then the same evi- 

 dence will apply to Fe and Zr. The reason for believing that the sub- 

 stance thus found is an element is based on the fact that its spectrum 

 remains unaltered in all minerals and after all chemical operations that 

 I have been able to devise. Furthermore, I believe that the new pro- 

 cess is not only more easy than any other, but also that it has given a 

 single element for the first time, as it eliminates the element d. The 

 yield will of course depend on the amount of purity required. From the 

 earths of gadolinite about one-tenth of quite pure yttrium (a) can be ob- 

 tained and about one-twentieth of very pure. 



I have determined spectroscopically that when, by the above process, 

 the absorption band of "erbium" at last disappears from 3 in. of strong 

 solution, all the other elements have also disappeared. 



By taking the first precipitate several times by ferrocyanide of potas- 

 sium from an acid solution, a mixture of many elements is obtained 

 which contains much of that element to which the so-called "erbium" 

 band is due. By dissolving a weighed quantity of this mixture in nitric 

 acid and water and examining the band spectrum, I have determined the 

 limit when the band can no longer be seen. Thus I have proved that 

 when the band vanishes from 3 inches of concentrated syrupy solution 

 of yttrium there cannot exist in it more than -| per cent of the mixed 

 element as compared with the yttrium, and there is probably less. 



I have not found ferrocyanide of potassium useful in the further 

 separation of the elements, but only in separating out a from the others. 



When the neodymium band has disappeared by use of sulphate of 

 sodium, all the other elements of the cerium group have disappeared. 

 The element thorium is sometimes present in the crude earths, but dis- 

 appears after a while from the purified earths. The conditions for its 

 disappearance I have not determined. 



