24 MEMOIR OF DR. DALTON, AND 



" The light of the aurora has been accounted for on three 

 or more different suppositions : — 1. It has been supposed to 

 be a flame arising from a chymical effervescence of com- 

 bustible exhalations from the earth. 2. It has been supposed 

 to be inflammable air, fired by electricity. 3. It has been 

 supposed electric light itself." 



" The first of these suppositions I pass by as utterly 

 inadequate to account for the phenomena. The second is 

 pressed with a great difficulty how to account for the existence 

 oi strata of inflammable air in the atmosphere, since it ap- 

 pears that the different elastic fluids intimately mix with each 

 other; and even admitting the existence of these strata^ it seems 

 unnecessary to introduce them in the case, because w^e know 

 that discharges of the electric fluid in the atmosphere do 

 exhibit light, from the phenomenon of lightning. From these 

 and other reasons, some of which shall be mentioned hereafter, 

 I consider it almost beyond doubt that the light of the aurora 

 borealis, as well as that of falling stars and the larger meteors, 

 is electric light solely, and that there is nothing of combustion 

 in any of these phenomena. Air, and all elastic fluids, are 

 reckoned amongst the non-conductors of electricity. There 

 seems, however, a difference amongst them in this respect : 

 dry air is known to conduct worse than moist air, or air 

 saturated with vapour. Thunder usually takes place in sum- 

 mer, and at such times as the air is highly charged with 

 vapour ; when it happens in winter, the barometer is low, and 

 consequently, according to our theory of the variation of the 

 barometer, there is then much vapourized air ; from all which 

 it seems probable, that air highly vapourized becomes an 

 imperfect conductor, and, of course, a discharge made along a 

 stratum of it will exhibit light, which I suppose to be the 

 general cause of thunder and lightning." 



" Now, from the conclusions in the preceding sections, we are 

 under the necessity of considering the beams of the aurora 

 borealis of a ferruginous nature, because nothing else is 



