242 HUMPHREYS: DUST LAYERS AND POLARIZATION ■ 



equally distant neutral points, Brewster's point below it, and 

 Babinet's above, the polarization plane is parallel to the horizon, 

 as it also is between the antisolar point and its companion, Arago's 

 neutral point. Between Babinet's point and Arago's point how- 

 ever, the plane of polarization along the sun's vertical is per- 

 pendicular to the horizon. Hence, when observations are con- 

 fined to the sun's vertical, as they often are, the neutral points 

 are those points of this vertical at which the two polarizations, 

 vertical and horizontal, are equal. 



The vertical polarization is due almost entirely to the primary 

 scattering of solar light by the dust in and the molecules of the 

 atmosphere, while the horizontal polarization appears to be due 

 essentially to secondary scattering. And as both the relative 

 and the absolute intensities of these two quantities of light, as 

 seen by an observer at the surface of the earth, are functions 

 of the amount and distribution of dust in the atmosphere, it 

 follows that the positions of the neutral points must also be func- 

 tions of the dust in the atmosphere and its distribution; and, as 

 a matter of fact, observations indicate the existence of distinct 

 dust layers with fairly defined upper boundaries. 



The first of these, effective in its action on the position of the 

 neutral points when the sun is only about 1° below the horizon, 

 is only that lower and relatively dense layer of dust, seldom 

 more than 1 kilometer thick, that is so frequently seen from 

 mountain tops and from balloons. It is essentially a dry weather 

 condition and is due to dust caught up from the surface of the 

 earth by winds at times when there is but little vertical con- 

 vection. 



The second dust layer, effective when the sun is roughly 3? 5 

 below the horizon, and therefore extending to an elevation of 

 about 4 kilometers, is due to that great quantity of dust that is 

 distributed thru the atmosphere up to this level at the times of 

 rather strong vertical convection, or at the times when cumulus 

 clouds prevail. It is well known that 4 kilometers is one of the 

 levels of maximum cloud formation — the level of the cumulus 

 cloud. That is, it is the ordinary limit of vertical convection 

 during clear weather. Hence, as a result of this considerable 



