no. 3 



ATMOSPHERIC SCATTERING OF LIGHT FOWLE 



Over the region plotted k may be considered equal to i — a„ x , that 

 is the scale of ordinates of figure i represents approximately the 

 fractional absorption of energy by i cm. of precipitable water in the 

 form of atmospheric vapor. The data for wave-lengths to the right 

 of the region shown (wave-length less than 0.35 /*) and to the left 

 (greater than 0.60 fx) are very inaccurate, the first because of spectro- 

 scopic field-light and very small measurable quantities, and the 

 second because of selective absorption. The accuracy with which the 

 observations fall on a straight line is beyond expectation. Within 

 the wave-length limits just named the average departures from a 

 straight line for the different years correspond in absorption as 

 follows : 



(1913, omitting poor points.) 



The mean value of N tp obtained from the atmospheric aqueous 

 water vapor is about 



P 



whereas assuming 

 value of 



5 XI °" Ki+aO/60. 

 Avogadro's law applicable to water vapor, a 



2 / XI ° 19 l^ + a076o 



or about 50 times as great would be expected. This anomalous 

 result, already noted in an earlier communication, 1 is therefore con- 

 firmed by the results of subsequent years. 



There appears to be associated with water vapor what has else- 

 where been denoted a " wet haziness " producing a uniform absorp- 

 tion over all the wave-lengths considered and giving a value for D 

 averaging about 0.017 which corresponds to a 2 per cent loss. 



There is an apparent peculiarity of the formula for N tp in that the 

 more opaque the vapor, that is the greater k, the smaller the number 

 of molecules per cm. 3 This formula is based upon Rayleigh's 



, 327T 3 , N . 



(1 



1 Astrophysical Journal, 38, p. 392, 191 3. 



