Polarization of Shj Light. 95 



weakening of the polarization. I found later that the same 

 thing had been noticed by Rubenson. On October 20, at 

 10 A.M., at the highest point in a cloudless sky, the reading 

 was 49°, while, judging from the two succeeding days, it 

 should have been 51^°. By 11.45 A.m. the sky was nearly 

 covered with cirrus, but through a clear rift near the highest 

 point I obtained the reading iS^''. It seems probable that 

 the ice particles of the cirrus were present in considerable 

 numbers at 10 A.M., but, though as thick or thicker than a 

 wave-length, they were too small and too few to be visible as 

 cloud. 



Generally speaking the polarization was much the same on 

 two successive cloudless days, unless there had been a mani- 

 fest change in the whiteness of the ground. This is very 

 different from Rubenson^s experience, but is only what might 

 be looked for in the clear atmosphere of the high Alps. 



Observations at Thusis and Davos. 



During the spring other work intervened, and, indeed, 

 0T\-ing to the bad weather, there were few opportunities, so 

 that 1 did not take any more readings till after leaving St. 

 Moritz early in April for Thusis. During the journey the 

 Iceland spar unfortunately got sliifted, but I was able to 

 measure the consequent alteration in the angle of incidence 

 on the three fixed plates in the following manner : — At St. 

 Moritz the average difference of the readings on the right and 

 left sides of the zero was about 2°. At Thusis it was about 

 6°. To produce this change the axis of the crystal must have 

 been shifted through 2°, and the angle of incidence on the 

 fixed plates must have been diminished by 2°. It is true that 

 for the journey I dismounted the movable pile ; but, owing to 

 the method of attachment, it must have been replaced in prac- 

 tically the same position with respect to the pointer of the 

 graduated circle. Further on I shall explain how I made 

 allowance for this change of 2°. Since the observations at 

 Thusis nothing in the polarimeter has been disturbed. 



The village of Thusis (2450 feet above sea-level) lies at the 

 southern end of a broad open valley running due north. To 

 the south there are high wooded cliffs separated by the narrow 

 cleft of the Via Mala. Here I had another opportunity of 

 testing the effect of snow on the ground. Two or three days 

 of hea\y snow were succeeded by a glorious day. At first it 

 was cloudless, but by 11.30 a.m. a few small clouds appeared 

 to the west. 



