﻿M. H. Wild on the Absorption of Light by the Air. 299 



to its frame. The sky must also be almost cloudless ; at any rate 

 there must be no mist in the neighbourhood of the sun. Owing 

 to small clouds which passed over the sun, I had frequently to 

 break off observations which had been commenced * for irregular 

 and sometimes very considerable changes ensued in the relative 

 brightness of the two screens. It was also impossible to observe 

 when the wind was at all powerful ; for the larger screen more 

 especially was altered in its position and in its illumination. 

 Reflection from adjacent objects, too, which affects the two 

 screens differently, must be carefully avoided. 



The above values, when reduced by Beer's formula, and by my 

 modification of it, to the unit of 1000 Paris feet, and, on the 

 other hand, to the unit I have adopted of 1 metre, give us the 

 comparison — 



For 1000 Paris feet. For 1 metre. 

 Schlagintweit-Beer . a = 0*9029 a = 0*9997 

 Schlagintweit-Wild . a = 07525 a = 0-9990 

 Wild a=0-2801 0=0*9961 



Our comparatively much smaller coefficient of transparency of 

 the air is partially to be explained by the circumstance that in 

 our observations all those influences which are usually consi- 

 dered to favour absorption were represented. On the days of 

 observation a north-east wind prevailed which at times was pow- 

 erful • the air was very dry and its temperature high, so that 

 undoubtedly local ascending currents set in, and abundant dust 

 and vegetable germs were met with in the lower layers. 



How great is the influence of dust in transparent liquids on 

 their power of absorption has been already shown by my deter- 

 minations of the absorption of water*. According as the water 

 was more or less freed from admixed dust by filtration through 

 various kinds of paper, were different values obtained for its 

 coefficient of transparency. Reduced to 1 metre as unit of path, 

 we get from these determinations for the coefficient of transpa- 

 rency of water after filtration through 



Coarse filtering-paper . . 0*5368 



Medium fine 0*6491 



Finest 0*7978 



These numbers differ, therefore, comparatively far more than 

 those for air for the same unit of path. It must, on the other 

 hand, be taken into account that in our determinations of the 

 coefficients of transparency of air in the open the particles 

 of dust are also illuminated, and thus not only act as absorbents 



* Pogg. Ann. vol. xcix. p. 272. 



