iv "BUNSENIANA" 71 



As well as I recollect what passed between Thomson and 

 myself about the lines was something of this nature. I 

 mentioned to him the repetition by Miller of Cambridge of 

 Frauenhofer's observations of the co-incidence of the dark 

 line D of the solar spectrum with the bright line D of certain 

 artificial flames, for example a spirit lamp with a salted wick. 

 Miller had used such an extended spectrum that the 2 lines 

 of D were seen widely apart, with 6 intermediate lines, and 

 had made the observation with the greatest care, and had 

 found the most perfect co-incidence. Thomson remarked 

 that such a co-incidence could not be fortuitous, and asked 

 me how I accounted for it. I used the mechanical illustration 

 of vibrating strings which I recently published in the Phil. 

 Mag. in connection with Foucault's experiment. Knowing 

 that the bright line D was specially characteristic of Soda, 

 and knowing too what an almost infinitesimal amount suffices 

 to give the bright line, I always, I think, connected it with 

 soda. I told Thomson I believed there was vapour of 

 Sodium in the sun's atmosphere. What led me to think it 

 was sodium rather than soda, chloride of sodium, &c., was 

 the knowledge that gases that absorb (so far as my experi- 

 ence went) yield solutions that absorb in the same general 

 way, but without the rapid alternations of transparency and 

 opacity. Now if the absorption were due to vapour of 

 chloride of sodium we should expect that chloride of sodium 

 and its solution would exercise a general absorption of the 

 yellow part of the spectrum, which is not the case. Thomson 

 asked if there were any other instances of the co-incidence of 

 bright and dark lines, and I referred to an observation of 

 Brewster's relative to the co-incidence of certain red lines in 

 the spectrum of burning potassium and the lines of the group 

 a of Frauenhofer. I am nearly sure this is in a volume of 

 the reports of the British Association being analogous to but 

 not identical with Brewster's obs n in the report for 1842, pt. 2, 

 p. 15. (Since I wrote this I have looked through the indices 

 of the vols. of the reports of the British Association and do 

 not find it there.) Thomson with his usual eagerness said, oh 

 then we must find what metals produce bright lines agreeing 

 in position with the fixed dark lines of the spectrum, or 

 something to that effect. I was, I believe, rather disposed 

 to rein him in as going too fast, knowing that there were 

 terrestrial lines (seen when the sun is low) which evidently 

 take their origin in terrestrial atmospheric absorption where 

 metals are out of the question, and thinking it probable that 

 a large number of lines in the solar spectrum might owe their 



