1887.] on Genesis of the Elements. 45 



When the vapour of an element is rendered incandescent by the 

 electric spark, the characteristic system of lines in its sj^ectrum is 

 regarded as unalterable, and is looked upon as a certain proof tliat 

 this special element is under examination. However much chemical 

 or other tests fail to show the presence of a given element, the 

 indications of its lines in the spectroscope are regarded as infallible. 

 Spectrum analysis is the court of final apj)eal, whose decision no 

 chemist has yet had the hardihood to dispute. 



By way of illustration I will project on the screen the very 

 characteristic system of lines given by yttrium when ignited by the 

 electric spark, — a system, be it remembered, having no connection 

 "whatever with the peculiar phosphorescent spectrum yielded by 

 yttrium. The coloured diagram gives as accurate a representation 

 of the spark spectrum of yttrium as can be drawn by hand. Omitting 

 minor lines you will notice two very strong groups of lines in the red 

 and orange. These lines have been always regarded as the charac- 

 teristic test for yttrium ; the presence of these groups proves the 

 presence of yttrium, and tlieir absence proves its absence. 



I now project the electric spark spectrum of GS as pure as I have 

 been able to prepare it. GS is one of the bodies which by long and 

 tedious fractionation I have separated from yttrium ; it occurs at one 

 extreme end of the fractioning, and dififers not only from the parent 

 yttrium in its phosphorescent spectrum, but by virtue of the process 

 adopted for its isolation, it must likewise diifer in chemical properties. 

 But what tale does the spectrum tell ? It tells us there is absolutely 

 no difference between this spectrum and that given by old yttrium. 



I now pass to the other end of the fractionation of yttrium, v^here 

 a body, Grj, concentrates giving a totally different phosjDhorescent 

 spectrum to that given by G8. And it also differs chemically from 

 old yttrium, and in a more marked manner from its brother, GS, at 

 the other extremity of the fractionation. Look at its spark spectrum ! 

 It is perfectly identical both with old yttrium and with G8, and 

 when I examine these three spectra in my laboratory with all the 

 appliances for exact measurement, the whole system of lines is still 

 identical. 



What inference can be drawn from these results ? Is discredit to 

 be thrown on spectrum analysis ? Is the superstructure which has 

 been so laboriously raised upon its indications to fall to the ground ? 

 By no means. Spectrum analysis and its grand generalisations are 

 on as firm a foundation as ever. I see two possible explanations of 

 the facts I have brought before you. According to one hypothesis 

 research has somewhat enlarged the field lying between the indications 

 given by ordinary coarse chemistry and the searching scrutiny of the 

 prism. Our notions of a chemical element have expanded. Hitherto 

 the molecule has been regarded as an aggregate of two or more atoms, 

 and no account has been taken of the architectural design on which 

 these atoms have been joined. We may consider that the structure 

 of a chemical clement is more complicated than has hitherto been 



