6 



Lieut. J. H. Hennessey on the [June 16, 



Roscoe's Spectrum Analysis, fig. 57. The scale adopted in this diagram 

 is, however, only about half that of Kirchhoff's, and the lines are so very 

 numerous that identification is exceedingly difficult. It is evident, how- 

 ever, that my 812, a notably variable line, is identical with his 254. 



18. The same, continued. — Comparing with the lines numerically alluded 

 to by KirchhofY in his appendix as atmospheric lines, the following coinci- 

 dences (nearly) appear : — 



My map at sunset. Kirchhoff's list of atmospheric lines. 



711. 711-4 



955 954-2 



961 96 I • 



967 | 9657 } 



19687 J 



971 970-5 



973 .972-1 



977 977-4 



987 988-9 



990 989-6 



In assuming the foregoing identities, I have been guided necessarily merely 

 by numerical coincidences. All the foregoing, in the proposed nomencla- 

 ture, are variable lines, not air-lines. 



19. Discussion of air-lines and bands. — Again, comparing my map sun 

 high with that at sunset, and beginning from the red end, the first group 

 met with are the air-bands 262 to 302. These bands, so far as I am 

 aware, have not been placed before, if seen. I have seen and taken mea- 

 sures to them repeatedly ; but without the interposition of a blue glass 

 they were invisible. The slit, too, must be sensibly widened. The whole 

 group is nearly at the end of the luminous rays, where, at sun high, 

 almost complete darkness prevails ; and after careful watching of this 

 group, I cannot help entertaining the belief that the range of the luminous 

 rays increases towards the red end of the spectrum as the sun approaches 

 the horizon. With the sun high, hardly any light reaches so far as 262 

 of the scale. When, however, the sun is setting, this part of the spectrum 

 lights up sensibly. I am even inclined to assert that other bands exist still 

 further removed from A than the group under notice ; they, however, ap- 

 peared too faint for intersection. May it not be anticipated that, at con- 

 siderable terrestrial altitudes, an extension of the spectrum will occur both 

 at the violet and red ends ? 



20. The same, continued. — Between 508 and 570 a very marked suc- 

 cession of air-lines and bands occur. Indeed, so far as I have been able 

 to glance at the spectrum in general, the most refrangible rays appear the 

 most liable to absorption. The air-bands about C and D are worthy of 

 notice, as also are the air-lines 724 and 726. 



21. Discussion of variable lines — Both A and B appear as lines of this 



