542 



NATURE 



[Aprils, 1889 



vestigation in the nature of the elements themselves. 

 Alumina is especially active in inducing new spectra 

 when mixed with rare earths. I have given more than a 

 .twelvemonth to the exclusive study of alumina phos- 

 phorescence, and still the research is incomplete. But I 

 have obtained some remarkable results. A moderate 

 amount of fractionation has enabled me to penetrate 

 below the surface of the red glow common to crude 

 alumina, and to see traces of a most complicated sharp 

 line spectrum. By pushing one particular process of 

 fractionation to a considerable extent I have obtained 

 evidence of a body which is the cause of some of these 

 lines. The spectrum, described by me in 1887 {Chem. 

 News, vol. Ivi. pp. 62, 72), is one of great beauty. The new 

 body is probably one of the rare elements or meta- elements 

 closely connected with decipia, for I have reproduced the 

 spectrum very fairly by adding decipia to alumina. Before 

 arriving at definite conclusions much time must be de- 

 voted to the subject. Certain it is that this new earth is not 

 yttria, erbia, samaria, didymia, lanthana, holmia, thulia, 

 gadolinia, or ytterbia, the spectrum of each of these when 

 mixed with alumina being very beautiful, but differing 

 entirely from the decipia-aiumina spectrum. 



M. DE BOISBAUDRAN'S REVERSION SPECTRA. 



Another modification of the phosphorescence process 

 is afforded by the " reversion spectra " of M. Lecoq de 

 Boisbaudran. 



The following is the description of this process by M. 

 Lecoq de Boisbaudran read before the Academy of 

 Sciences on June 8, 1885 :— "When the electric spectrum 

 of a solution luith a inetaUic base is produced, it is cus- 

 tomary to make the outside platinum wire (whence the 

 induction spark strikes) positive, the liquid consequently 

 forming the negative pole. If the direction of the current 

 is reversed, the metallic rays (due to the free metal or to 

 one of its compounds) are scarcely or not at all visible ; 

 at all events, so long as the exterior platinum wire now 

 forming the negative pole is not coated with a deposit." 



M. de Boisbaudran continues : — " Having again taken 

 up last year my researches on the rare earths be- 

 longing to the didymium and yttrium family, I had 

 occasion to observe with many of my preparations 

 the formation of spectrum bands, nebulous, but some- 

 times tolerably brilliant, having their origin in a thin 

 layer of a beautiful green colour, which appeared at the 

 surface of the liquid (a solution of a chloride) when it was 

 rendered positive." 



M. de Boisbaudran further adds : — " The production of 

 my reversion spectrum appears to be analogous physic- 

 ally with the formation of the phosphorescence spectra 

 obtained by Mr. Crookes at the negative pole in his high 

 vacuum tubes containing certain compounds of yttria. 

 The conditions of the two experiments are, however, 

 practically speaking, very different." 



By this method M. de Boisbaudran has discovered 

 phosphorescent spectra, which he considers due to the 

 presence of two earths, one of which, supposed to be new, 

 he has provisionally named Za, and another, also thought 

 at first to be new, and therefore called Z,3, but since ad- 

 mitted by him to be terbia {Compies rendus, vol. cviii. 

 p. 167, January 28, 1889). In the hands of so skilful an 

 experimentahst as my accomplished friend, this method 

 may give trustworthy indications, but the test is really 

 beyond the range of practical analysis, owing to the diffi- 

 culty of eliciting the phenomena. Unless the strength of 

 the spark, the concentration and acidity of the solution, 

 and the dispersive and magnifying power of the spectro- 

 scope bear a certain proportion to each other, the 

 observer is likely to fail in seeing a spectrum even in 

 solutions of earths which contain considerable quantities 

 of Za and terbia. 



The Phosphorescence of Alumina. 



I now wish to draw attention to some recent researches 

 on the phosphorescence spectrum given by alumina. So 

 far back as 1859, Becquerel examined in his phosphoro- 

 scope pure alumina carefully prepared, and described it 

 as glowing with a splendid red colour. He rendered his 

 specimens phosphorescent by exposure to the sun, and 

 made no use of the induction spark. As described by 

 Becquerel {Annales de Chemie et de Physique, vol. Ivii', 

 1859, p. 50), the spectrum of the red light emitted from 

 alumina agrees with that of the ruby when submitted to 

 the radiant matter test. It displays one intensely red line 

 a little below the fixed line B in the spectrum, having a 

 wave-length of about 689*5. There is a continuous 

 spectrum beginning at about B and a few fainter lines 

 beyond it, but in comparison with this red line the faint 

 ones are so dim that they may be neglected. My latest 

 observations in the vacuum tube prove this line to be 

 double, the distance apart of the components being about 

 half the distance separating the D lines (Roy. Soc. Proc, 

 vol. xlii. p. 26, December 30, 1886), their respective wave- 

 lengths being 694*2 and 6937 (i/X.- 207*5 and 207*8). 



The red phosphorescence of this alumina is exceedingly 

 characteristic. M. de Boisbaudran {Comptes rendus, vol. 

 ciii. p. 1107 ; vol. civ. pp. 330,478, 554, 824) contends, 

 however, that this red phosphorescence is due, not to the 

 alumina itself, but to an accompanying trace of chromium, 

 i/iioo part of chromium being sufficient to give a splen- 

 did red phosphorescence, whilst even i part of chromic 

 oxide in 10,000 will produce a very distinct rose colour. 

 In testing this view I have purified alumina most 

 carefully, so as to secure the absence of chromium, 

 and on examining it in the radiant-matter tube I 

 have still obtained the characteristic phosphorescence 

 and spectrum. I have then added to my purified 

 alumina chromium in known varying proportions, but 

 without finding any increase in the intensity of the 

 phosphorescence. I fractionated my purified alumina by 

 different methods, and found that the substance which 

 forms the crimson line becomes concentrated towards one 

 end of the fractionations, whilst chromium concentrates 

 at the other end. I have suggested four possible 

 explanations of the phenomena — 



(i) The crimson line belongs to alumina, but it is 

 liable to be masked or extinguished by some other earth, 

 which accumulates towards one end of the fractionations. 



(2) The crimson line is not due to alumina, but to 

 the presence of an accompanying earth which accumulates 

 towards the other end of the fractionations. 



(3) The crimson line belongs to alumina, but its develop- 

 ment requires certain precautions to be taken in the 

 duration and intensity of the ignition, and absolute 

 freedom from alkaline and other bodies carried down 

 by precipitated alumina, and difficult of removal by 

 washing. 



(4) The earth alumina is a compound molecule, one 

 only of its component sub-molecules giving the crimson 

 one. If this hypothesis is correct, alumina must admit 

 of being split up in a manner analogous to yttria. 



Conclusions. 



During the course of the investigations — whose results 

 are briefly summarized in the foregoing pages, — I have 

 repeatedly had recourse to the balance, to ascertain how 

 the atomic weights of the earths under treatment were 

 varying. An atomic weight determination is valuable in 

 telling when a stable molecular grouping is arrived at. 

 During a fractionation, the atomic weight of the earth 

 slowly rises or falls until it becomes stationary, after which 

 no further fractionation of that lot by the same process 

 ma.kes it vary. Usually a result of this kind has been 

 relied on as proof that the elementary stage has been 



