Polarization of Sky Light. 99 



be remembered that at Davos the sun was much higher (57° 

 instead of 32°), and the ground consequently more than half 

 as bright again. 



Anj marked deviation of the plane of polarization from the 

 plane through the sun was shown in my instrument by the 

 black cross gradually being transformed into the white one 

 without even disappearing. While the point nearest neutra- 

 lization was being passed, the black cross was rotating through 

 forty-five degrees and changing its appearance, so that the 

 four arms of the black cross became the dark spaces between 

 the arms of the white cross. I believe that with proper pre- 

 cautions very accurate measurements of the position of the 

 plane of polarization of partially polarized light might be 

 made on this principle. I found that, with large clouds 

 about, the deviation might reach, perhaps, three or four degrees; 

 but in a clear sky, e. g. in all the measurements I have recorded, 

 it was insensible. It was curious to find that it was easy to 

 poini: out the position of the sun behind a mountain with 

 considerable accuracy by applying the same principle. It 

 was done thus. First putting the plane/ (i? e (fig. 2) vertical, 

 I turned the instrument round the bolt K till there was no 

 rotation of the black cross on neutralization. This insured 

 the plane /<^ e passing through the sun. Then, having turned 

 the box about G till a point near the horizon was under 

 observation, I opened or shut the hinge H, till fde again 

 passed through the sun. Then the sun lay in the direction 

 ed. 



In the observations cited above I have measured the average 

 polarization of all the different colours which go to make up 

 the light of the sky. There remains the question how the 

 polarization varies with the wave-length of the light. Since 

 the imperfection of the polarization is due to light reaching the 

 particles from the sky and earth as well as from the sun, we 

 should expect the polarization to increase with the wave-length. 

 For sky light is blue and earth light can hardly be redder than 

 sunlight. I tried at Thusis the effect of interposing orange 

 glass, which, according to my estimate, should have increased 

 the reading (r by several degrees, but did not find any decided 

 change. I cannot account for this result. Though I hope to 

 repeat the observations, I should like to call the attention of 

 other workers to this point. The effect ought to be greater 

 at sea-level on account of the increased brightness of the sky. 

 Another theoretical result, which might be easily tested by any 

 one with a suitable polariscope, is that the position of the 

 neutral point should change considerably with the wave-length 

 of the light under examination. 



H2 



