65 



tint and extending it in altitude, this having existed since 

 1880. The glimmer is frequently bluish green to southward, 

 with somewhat northern glow. In these cases there is cross- 

 polarization in the glow and radial polarization and selenite 

 tints in the glimmer, brightest latest. 



How is the fact of cross-polarization to be explained? 

 Supposing the red tints to be a species of fluorescence. This 

 could give the phenomenon of cross-polarization, for in 

 the cross position we should only have the fluorescent tint 

 and the residual tint of the polarised sky light, and in the 

 parallel position the tint from the admixture of the two lights. 

 If this were true we should get a gradual change of tints in 

 passing from one position to the other. That is, a change in 

 intensity of the respective tints with the important fact that 

 one hue would be brighter than the other. Thus, in the cross 

 position the non-polarised light plus the residual would be less 

 bright than the mixed tint of the two lights in parallel position. 

 In this case the cross position would always give the most 

 decided difference of tint. 



If dust were present in an unusual degree and light were 

 polarised by reflection from it, the result would be that 

 one tint w^ould be observed, for the plane of polarization 

 would be parallel to that of the sky. Thus the tint would 

 be brightest in parallel and darkest in cross position. If 

 dust were present and sky somewhat uniform, whatever 

 occurred in sky-light would occur in dust-light polariza- 

 tion. Supposing the residual or unpolarised light of the 

 sky bluish, the residual light of the dust reddish, and that 

 both were proportionately polarised. Then, in different 

 2Jarts of the sky the hue of cross position should be the same 

 with only its difference of brightness due to varying complete- 

 ness of polarization in various parts of the sky. Supposing the 

 residual light of dust greatest. Then in positions of most com- 

 plete sky polarization the red tint in cross-position would 

 be brightest ; that is, more decided, and the spectroscope with 

 nicol-prism would reveal it at any time in the day. Xext, the 

 difference of tints as the neutral point were approached would 

 diminish until the hue of ordinary sky tint at the time would 

 be seen. The polarization of the sky is most perfect when the 

 sun is above the horizon. "When the sun is above the cross- 

 polarization is rarely apparent, and at the point of best 

 polarised sky-light the residual colour is bluish grey. Hence, 

 cross-polarization cannot be due to this cause, for the residual 

 light is that light which is unpolarised, and at the neutral 

 point the sky-light is the residual light. But throughout my 

 scrutiny this point is rarely reddish, and is never tinted with any 

 comparison to the sky that is farther away still from the sun side; 



