268 



NATURE 



\yuly 21, 1 88 1 



reflected on to that other half by means of the prism. You will 

 see in a moment, therefore, that it was quite easy by thi^ 



method to see in his observing telescope no longer the speclrum 

 of the sun alone, but the spectrum of the sun together with the 



Fig. 5. — Path of light through comparison prism, d/o, prism, /, light 

 source; r, [oint of reflecticn ; s, slit; F, light source in front of slit. 



spectrum proc'uced by each of the chemical sulstances which he 

 cho e to experiment upon. 



Again, anticipating matters somewhat, I can show you 

 .something like what Kirchhoff then saw — only again I give you 

 a pholograjih, and therefore we have a part of the .■-pectrum with 

 which he did not begin his work-. In the lower half of this 

 slide we have ihe solar spectrum. There are ihe two lines II and 

 K, and in the upper portions we have ihe bright lines given us 

 by a metallic e'ement — in this ca e cerium. You see when 

 some of the metallic elemenls are Irea'ed in this v\ay they have 

 a trick of giving us very complicated spectra. 

 (The photograph was projected on the screen.) 

 I will show you the iron spectrum which Kiiclihoff worked 

 upon. The point was to determine which of the bri-ht line; corre- 

 si-onded with the dark Fraunhofer Hues. Over the whole visible 

 reach of the spectrum Kirchhoff ma) ped the re: id's, for iron ; I 

 will give one or two extracts from his paper. He say=,' " It is 

 specially remarkable that coincident with the positions of the bright 

 lines w hich I have observed [that is tire bright lines from the vapour 

 of iron, using two iron poles with an induction coil] definite dark 

 lilies cccur in the solar spectrum. By the help of a very delicate 

 method of observation which I hr^ve employed, I believe that 

 each coincidence ob; erved by me between the iron lines and the 

 lines of the so'ar spectrum, may be considered to 1 e at least as 

 well establihed as the coincidence of the sodium lines." Then 

 he show, limiting hisatfenlion to sixty of the most defined iron 

 lines in the region included in his map, that the betting that 

 there was iron in the sun was about three trillions to one, dealing 

 alone with the absolute matching of the positions of the lines 

 recorded in the solar spectrum. Then he goes on to show that 

 this probability of three trillions to one was rendered still greater 

 by the fact that Ihe brighter a given iron line is seen to be ihe 

 darker as a rule — and I beg you to mark those words "as a 

 rtile" — does the corresponding solar line appear. Hence th s 

 coincidence must be produced by some cause, and a cause can 1 c 

 assigned which affords a very perfect explanation of the pheno- 

 n:cnon. He then gives the cause, which has already been 

 stated by Prof. Stokes. 



Fig. 6. — Coincidence of some of the bright lii 



Now before I go further I must point out that there is ^ 

 considerable assumption here. It is quite easy in an electri<^ 

 lamp to produce the vapour of a meteorite or of any of ouf 

 terrestrial rocks, to throw their spectra on the screen, and to 

 map them with considerable minuteness ; and we say we have 

 ihe spectrum of such and such a meteoiite, or of such and such 

 a rock. Similarly we can get the spectrum of iron term'nals, 

 and serve that in the same way, and we are considerably 

 astonished at the wonderful simila; ity of the results thus ob- 

 tained. Now chemistry has advanced to a certain stage, and 

 low temperature chemisti-y comes in and shows us that this 

 meteorite or rock may be an excessively complicated substance. 

 The same chemistry ppjilied to iron shows that nothing can be 

 done with it. But to say that iron cannot be broken up because 

 low temperature chemistry fails to break it up is, you will see, 

 an assumption, for as we undoubtedly get the lines of the con- 

 stituents of the rock, or of the meteorite, recorded in the spectrum, 

 we may also be registering the lines of the constituents of iron ; 

 and it is fair to say this, because we know that in the electric arc 

 we have a stage of heat at which at present no experiment what- 

 ever has been made. 



Passing on from that poii.t, however, I will ask you to consider 

 somew hat more in detail that part of Kirchhoff's work which 



deals with the connection between the solar spectium and the 

 spectra of the chemical elements.* 



Confining his observati'ins to the region of the solar spectrum 

 between F and P, Kirchhofif found the following coincidences 



Hofmann' continued these researchts on both sides if the 

 region observed by Kirchh iff as far as A on one side and G on 

 the other, and in addition investigated the spectra of the follow- 

 ing metals : — Potassium, rubidium, lithium, ceiium, laiahauuni, 

 didymium, platinum, palladium, and an alloy of iridium and 

 ruthenium. Hofmann added the following coincidences between 



^ " Researches on the Solar Spei 

 18. 



■■' Kirchhoff's " Researches," tram 

 3 lb.. Part II , Appendix. 



Ri 



t"s translation, Port I , 

 e, Part I , Supplement. 



