Prismatic Spectra of the August Meteors. 169 



tion (that is to say the latter portion) of the time that they 

 continue visible. Their condition is then exactly that of a 

 flame of gas in a Bunsen J s burner, freely charged with the 

 vapour of burning sodium ; or of the flame of a spirit-lamp 

 newly trimmed, and largely dosed with a supply of moistened 

 salt. 



It is difficult to believe that the vapour of the metal 

 sodium exists, in such considerable quantities, at the confines of 

 the atmosphere. It is much more reasonable to suppose that 

 it is brought into the atmosphere by the meteors themselves, 

 so as to be deposited in the luminous trains that mark their 

 course. The material of the August meteors is, therefore, 

 probably a mineral substance, in which sodium is one of the 

 chemical ingredients. Such is the rather satisfactory termi- 

 nation of an experiment, which it will be very easy to repeat, 

 whenever an abundance of meteors on the night of the 10th 

 of August offers an equally favourable opportunity for examin- 

 ing their spectra by the aid of the meteor spectroscope. 



The connection believed by adherents of Ohladni to 

 exist between shooting-stars and aerolites is now shown, at 

 least in August, to extend itself in some measure to their 

 chemical composition. The meteorites of Aumale, which fell 

 on the 25th of August, 1865, were found, on analysis by Mr. 

 Daubree, to contain soluble salts (chloride and carbonate) of 

 sodium. A circumstance so uncommon in the composition of 

 aerolites, allies the meteorites of Aumale very closely with 

 the sodium-bearing streaks of the meteors of the 10th of 

 August. 



In this manner each new acquisition of knowledge, how- 

 ever unforeseen may be its origin, tends to support the theory 

 of Chladni, and to confirm the belief that shower-meteors and 

 shooting-stars are actually aerolites of small dimensions. In 

 whatever manner aerolites and shooting-stars are related to 

 each other in their astronomical and other peculiarities, they 

 will certainly require a vast number of further experiments 

 to unfold their real source. 



EXPLANATION OF THE PLATE. 



The meteor spectra are numbered in the plate, after the 

 description in the text, in the order of their appearance ; a, 

 represents the spectra of the nucleus, and b, the spectra of the 

 meteor trains. Double white lines indicate the boundaries of 

 the field of view. 



Figs. 1, 10, represent the persistent luminous impressions 

 of the nucleus only, and the meteors left no streak. 1, a 

 meteor in Gygnus ; 10, a meteor spectrum, conveying a faint 



