8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 69 



which is derived from 



SttW (D'-D) 2 7* 

 k ~ 3 D 2 A 4 » t 2 ) 



by substituting from 



NT (D'-D ,. 



where D and Z7 are now the original and the altered densities of the 

 medium and T the volume of the disturbing particle. That is n, 

 the index of refraction, is a function of N, the number of molecules 

 per cm. 3 In the present case the value of n cannot be observed. A 

 preliminary use of the formula for Nt P leads to the suspicion of 

 something abnormal in the condition of atmospheric aqueous vapor. 

 For instance, is it in some colloidal state resulting from some form 

 of ionization of the air? Wilson, 1 for instance, has shown that 

 under the influence of ultra-violet light, in moist dust-free air, 

 nuclei are formed and may grow " till they become large enough to 

 scatter ordinary light." By careful laboratory researches he has 

 shown that oxygen and water vapor alone are necessary for their 

 production ; that water vapor is necessary ; that saturated vapor is 

 not necessary ; that these nuclei persist for some time after their 

 formation ; that they are different from ions since they carry no 

 electric charge; that they are probably due to some combination, 

 H 2 2 , which by decreasing the vapor pressure allows drops of water 

 containing one of them to form and grow where pure water drops 

 would evaporate. Bieber ' has since shown that H 2 O a is formed by 

 the action of ultra-violet light. Although the ultra-violet energy 

 in sunlight is too weak at the surface of the earth to be very efficient 

 in the formation of these nuclei, in the clear air above Mount Wilson 

 it may well be very active. In such nuclei, dependent directly upon 

 the presence of water vapor, there seems a possible explanation of 

 the increased absorption. Or, is it possibly due to some emanation 

 from the sun producing some change in the condition of the water 

 vapor ? 



Reverting now to formula (2) it is to be noted that if the molecules 

 cluster together because of some ionization phenomena or other- 

 wise and in such state each cluster acts as a whole in scattering light 

 as ordinarily a single molecule does, then, neglecting for the moment 

 the effect of the factor (D' — D) 2 /D 2 , the intensity of the scattering 



1 Philosophical Transactions of the Royal Society, 192, p. 403, li 



2 Annalen der Physik, 39, p. 1313, 1912. 



