﻿M. H. Wild on the Absorption of Light by the Air. 301 



duration of the same series of observations by shortening them as 

 much as possible, and choosing days when the cloudiness did not 

 much vary. The position of neutralization was observed in only 

 one quadrant, and first of all when one tube was exhausted and 

 the other filled with air, and then after air had been allowed 

 rapidly to enter again. The results to be obtained from these 

 observations are therefore only to be regarded as first approxi- 

 mations. This comparison, moreover, of the ratio of the lumi- 

 nous intensity when the tube was full in one case and in the 

 other exhausted, merely gives us the ratio of the coefficient of 

 transparency of rarefied air, and of air of ordinary density. 



The series of observations on the 29th of August gave as the 

 mean angular difference of the two positions on the photometer 

 21 r , which corresponds to a ratio of the coefficients of transpa- 

 rency of the air for pressures of 35 and 720 millims., referred to 

 a length of 1 millim., 



-^- 5 =1-01023. 



« 720 



This result, however, is uncertain, because on this day during 

 the time of observation the clouds, and therewith the illumina- 

 tion of our disk, greatly varied. 



On the 31st of August, when the sky was almost cloudless 

 and the illumination almost constant, for four successive series of 

 observations the mean angular difference in the two positions 

 of the photometer was 8 f, 5, the pressures being 715 and 100 

 millims. From this we get 



^= 1-00413. 



^715 



It is here assumed that the increase in the luminosity when 

 the tube was exhausted was exclusively due to diminished ab- 

 sorption, and was not due to a simultaneous diminution of the 

 enfeeblement of the light owing to its passage through the en- 

 closing plates. FresneFs formulae of intensity show in fact that 

 transmitted light is not materially altered in its intensity when 

 the glass plates are bounded on one side by a vacuum instead of 

 on both sides by air. 



The above numbers may also be used to calculate the coeffi- 

 cients of transparency of air of ordinary density. 



The coefficient of transparency of air whose density is - that 



of ordinary air is obviously 



i_ 



n 



