64 



and this tint in tlie east disappears just when the red tint 

 appears in tlie west. In cloud-patched sky the area is more 

 extended and undefined, the tint more carmine, and duration 

 longer. The radial polarization of the remainder of sky is 

 not constant. When there is a red glimmer in the sun side the 

 phenomenon of cross-polarization is apparent in the vicinity of 

 this glow. Thus, when the long diagonal of the nicol is about 

 perpendicular to the plane passing through the prism, the point 

 of observation and the sun, the red tint is brightest. The posi- 

 tion of maximum cross-polarization is continually varying in dis- 

 tance from the sun. It is not symmetrical. In this colony.. 

 from all the points I got intelligence, it is generally more 

 perfect on the north of glow poiut. When there is a bluish 

 green glimmer the cross-polarization is not observable in it. 

 The cross-polarization varies with the kind of glow. It is fre- 

 quently prominent beyond the limits of the red. It is best in 

 w^eather when there are clouds about, and when the dew point is 

 high. Sometimes it is visible in bright blue sky, the cross position 

 of the nice] showing a fine red. Sometimes it is visible beneath the- 

 clouds when they are of large extent and the air seems hazy. It 

 is not always necessary to turn the nicol 90° from the radial 

 position, or from the position showing best blue, the red- 

 reaching its niaximum in a very small angle. This is most 

 marked when there are rain clouds not Tery heavy and dense,, 

 and especially when the spectroscope shows a heavy vapour 

 band for a day or two previously. The clouds are generally 

 low level bands, though it is also observable when the same- 

 kind of clouds are rising. When the cross-polarization is 

 prominent the selenite colours are very fine, and seem to be 

 most splendid at an angle of about 70^ to 80° from the sun. 

 KoTE. — This angle should be the angle of maximum polariza- 

 tion if it were due to water particles whose dimensions were 

 not insignificant in comparison with light waves. 



Placing a nicol-prism between the slit and the battery of a 

 spectroscope, I examined the cross-position, and the result 

 of a number of evening and morning scrutinies seems to show 

 that when the nicol is in cross-position the spectrum is 

 brighter at the less refrangible end, and that the principal 

 absorption takes place between E and F. That when in the 

 early part of the glow the lines in the red are broader when, 

 the nicol is radial and sharper when the nicol is crossed. 



In examining the halo or corona I have tried to find the 

 phenomenon of cross-polarization, but with very questionable 

 success in the daytime. It seems to be present when there is- 

 panther-hue background and clouds obscure the sun. Cross- 

 polarization also exists in variable degree when orange banks 

 extend far from the sun, the cross-polarization deepening their 



