ON OUR KNOWLEDGE OF SPECTRUM ANALYSIS. 129 



duced by substances in -which three pairs of carbon atoms are doubly 

 linked together, as in the benzine ring. 



In most of the cases which we have hitherto discussed, the charac- 

 teristic absorption of the substance under examination extended over a 

 considerable range ; the substance either blocked out altogether a large 

 part of the spectrum, or at least showed absorption bands which were 

 broad and increased considerably in width with increased concentration. 

 When, however, absorption bands become narrower and more definite, 

 so that they can be examined under high dispersive powers, their 

 behaviour under different circumstances becomes more interesting, for we 

 can trace smaller differences and more minute changes. 



It was Bunsen ^ who first showed that such small changes do occur, 

 and he thereby led the way in a line of research which promises to be of 

 great importance. While examining the absorption spectra of different 

 didymium salts, he found that though all the salts showed spectra so 

 nearly identical that with the ordinary one prism spectroscope they could 

 easily be mistaken for each other, higher dispersive powers revealed some 

 very interesting and characteristic changes. His conclusions are best 

 quoted in his own words : — 



' Very remarkable and noteworthy are the small alterations in position 

 which occur in the minima of brightness in the didymium spectrum, 

 dependent upon the nature of the compound in which the metal occurs. 

 These changes are too minute to be seen with the small, though seen 

 with the large instrument. I have as yet only investigated them com- 

 pletely in the case of three didymium salts, viz., the chloride, sulphate, 

 and acetate. It is, however, more than probable that the same phe- 

 nomena will also be found to occur with other solutions, and with the 

 absorption spectra of other crystals of didymium salts, and perhaps may 

 be exhibited with the luminous spectra of the oxide and other compounds 



of didymium The atomic weight of didymium chloride is 95-9, 



and that of the anhydrous acetate is 106-9. It will be noticed that all 

 the groups of bands in the case of salts under examination approach 

 the red end of the spectrum in the order of their increasing atomic 

 weights. 



' These differences here noticed in the absorption spectra of different 

 didymium compounds cannot in our present complete state of ignorance 

 of any general theory for the absorption of light in absorptive media be 

 connected with other phenomena. They remind one of the slight and 

 gradual alterations in pitch which the notes from a vibrating elastic rod 

 undergo when the rod is weighted, or of the change of tone which an 

 organ-pipe exhibits when the tube is lengthened.' 



The increased lowering of the vibrations with increasing atomic 

 weight of substance combined with the didymium is no doubt very sug- 

 gestive, but we cannot at present assign any definite law regulating the 

 displacement in different cases. Thus the difference in the wave-length 

 between the bands of the chloride and acetate is nearly the same for all 

 four bands, but the difference in the wave-length between the bands of 

 the chloride and acetate decreases rapidly with decreasing wave-length, 

 so that the yellow band is displaced about twice as much as the green 

 band, and about three times as much as the bands in the blue. It follows 

 from this difference in the behaviour that while the effect of the acetate 

 on the yellow band is about seven times as large as that of the sulphate, 



' Phil. Maq. [41 xxxii. p. 177. 



1882. K 



