April \, 1889] 



NATURE 



539 



used for this operation may be summarized under the 

 name of fractionation, whether they be fractional pre- 

 cipitations, crystallizations, or decompositions. The 

 essential principles of this process were so fully discussed 

 on the last occasion when I had the honour of addressing 

 you that I need not further allude to them. 



The Didymium Group. 



A combination of such delicate and prolonged chemical 

 processes with spectroscopic examination applied to 

 bodies showing absorption spectra soon led to important 

 discoveries. When in that year the didymium from 

 samarskite was examined by M. Delafontaine {Comptes 

 rcndus, vol. Ixxxvii. p. 632 ; Chemical N'ezvs, vol. xxxviii. 

 p. 223), he found it to differ somewhat from ordinary 

 ■didymium as extracted from cerite and gadolinite, and by 

 -a series of chemical fractionations he succeeded in separ- 

 ating from it an earth which he called decipium, giving 

 at least three absorption bands, one having a wave-length 

 of 416 (i/X- 578) ; another narrower and stronger, at 

 wave-length 478 (i/X- 438), and a very faint "minimum 

 ■of transmission " near the limit of the blue and green. 

 Nine months later, M. Lecoq de Boisbaudran {Comptes 

 rendus, vol. Ixxxix. p. 212 ; Chemical Neivs, vol. xl. p. 99) 

 announced the discovery of samarium as a constituent of 

 the didymium from samarskite. 



Still didymium was not reduced to its ultimate sim- 

 plicity. In 1885, Dr. Auer von Welsbach {Monatsh. 

 Chem., vi. 477), by fractionally crystallizing the mixed 

 nitrates of ammonium, didymium, and lanthanum, 

 showed it was thus possible to cleave didymium in a 

 certain direction and separate it into two other bodies, 

 one giving green salts and the other pink salts. Each of 

 these has a characteristic absorption spectrum, the sum 

 of the two sets of bands approximating to the old didy- 

 mium spectrum. These bodies the discoverer has named 

 respectively praseodymium and neodymium. The neo- 

 dymium spectrum, according to Dr. Auer, consists of the 

 whole of the bands in the red, with part of the large one 

 in the yellow ; it then misses all the green and blue, and 

 takes in the second line in the vioiet. The spectrum of 

 praseodymium takes the other part of the yellow band 

 and all the green and blue, except the second blue, which 

 belongs to neodymium. Subtracting these two spectra 

 from the old didymium spectrum, there are still two bands 

 left at X 462 and 475 (i/X- 465 and 443). Assuming that 

 the argument from absorption spectra is a legitimate 

 one — and all recent research tends to show that if not 

 •quite trustworthy it is at all events a weighty one — the 

 inference I draw from these results is that the old didy- 

 mium still contains a third body distinct from neo- and 

 praseodymium, to which one or both of these extra bands 

 is due. 



I must venture to lay especial emphasis on the words 

 £n a certain direction. Didymium in my own laboratory 

 has undergone other cleavages, and I have not yet de- 

 cided whether we shall have to recognize further decom- 

 positions of neodymium and praseodymium, or whether 

 the original didymium is capable of being resolved 

 differently according to the manner in which it is treated. 

 Keeping the band in the orange always of the same 

 strength, in many of the fractions of didymium from 

 different sources the other bands of neo- and praseo- 

 dymium are seen to vary from very strong almost to 

 obliteration {Chemical News, vol. liv. p. 27). In this 

 way I have worked on the spectra of didymium from 

 allanite, cerite, euxenite, fluocerite, gadolinite, hielmite, 

 samarskite, yttrotitanite, &c., and the further I carry the 

 examination the more the conclusion is forced upon me 

 that didymium must not be regarded as compounded of 

 two elements only, but rather as an aggregation of many 

 closely allied bodies. Later researches of Messrs. Kriiss 

 and Nilson have led them to the same conclusion. 



By examining the absorption spectra of solutions of 



rare earths obtained from widely different sources, MM. 

 Kriiss and Nilson {Berichte der deutsch. chem. Gesellschaft, 

 vol. XX. Part 12, p. 2134; and Chem. News, vol. Ivi. 

 pp. 74, 85, 135, 145, 154, 165, 172) came to the conclusion 

 that the elements giving absorption spectra, and known 

 as didymium, samarium, holmium, thulium, erbium, and 

 dysprosium, were not homogeneous, but that each one 

 contained almost as many separate components as it 

 produced bands of absorption. 



They have discovered that in didymium obtained from 

 some minerals one of the fainter lines of the normal 

 didymium spectrum is strong, while others usually stronger 

 are almost or quite absent ; results to which I shall pre- 

 sently refer will show that this cannot be explained by 

 dilution or concentration. In this way, by examining a 

 great number of minerals, they found anomalies occurred 

 in the case of almost each of the old didymium lines, and 

 therefore decided, as above mentioned, that it is a com- 

 pound body, capable of resolution into at least nine 

 separate components. 



Identical arguments are brought forward to prove that 

 each of the other so-called elements, samarium, erbium, 

 holmium, thulium, dysprosium, &c., are compounds of 

 many closely allied bodies. 



Messrs. Kriiss and Nilson, I believe, are pushing their 

 investigation with the object of isolating the separate 

 components of these different earths. They, however, 

 question the possibility of resolving the erbia and didymia 

 earths into their several ultimate constituents by a 

 fractionated decomposition of the nitrates. In fact, they 

 assert that by means of the methods of separation at 

 present known it would be almost impossible to com- 

 pletely isolate any single constituent of the mixed earths. 

 They therefore propose, as I had previously done,' a 

 method by which we may certainly arrive nearer to the 

 mark and dispense with much tedious fractionation. If 

 we examine the minerals which contain these rare earths, 

 we find they occur in very different states of mixture or 

 combination. Sometimes many of the constituents which 

 we wish to separate are conjointly present, and sometimes 

 but few. The desired differentiation, in fact, has already 

 been commenced by Nature. Kriiss and Nilson, there- 

 fore, whichever ingredient they wish to separate, propose 

 to operate upon a mineral which contains that ingredient 

 as far as possible in a state of isolation. In other words, 

 they will take advantage of the work that Nature has 

 already begun, and endeavour by refined chemical means 

 to put the last finishing touches to her work. Thus they 

 will be able to work with smaller quantities of primary 

 material, — no small consideration in the case of some 

 minerals,— and to obtain results in a shorter time. How 

 widely the composition of one and the same mineral, as 

 judged by our searching physical tests, may vary, will 

 be seen from the following instances. Fergusonite from 

 Arendal shows six of the bands of holmium, fergusonite 

 from Ytterby four, and that from Hittero only three. 

 Moreover, the ingredient provisionally called Xa is to be 

 found in the fergusonite from Ytterby, but not in that of 

 Arendal and Hittero. 



The foundation for thus firmly declaring what I had 

 previously ventured to infer, is the striking differences 

 in the spectra given by several specimens of one earth, 

 say didymium, when obtained from different sources. 



We are anxiously waiting the results of this investi- 

 gation, but although the paper quoted was published in 

 July 1887, no further communication has come from these 

 illustrious workers. 



Chemists recently have stated, as proof of the existence 

 of new elements, the fact that certain bands of absorption 

 as seen in various fractions " follow the same variations 



'Address to the Chemical Section of the British Association, Birmingham 

 Meeting, Chem. Nervs, vol. liv. p. 113- "On the Fractionation of 

 Yttrla," Chem. News, vol. liv. p. 157- ^'■<«'- *<V« S<"> wL xl. {1886), 

 p. 505. 



