98 Mr. J. 0. McConnel on the 



Nov. 27, 11.47 A.M. 451° -285 -80 

 This was a few days after twenty inches of snow had fallen, 

 and the covering was fairly complete except for the bare trees. 

 There were a few clouds about, but none near the point of 

 observation. Q r s 



Dec. 2, 12.5 a.m. 44^° -295 -84 Sky almost cloudless. 

 This was after some fresh snow. The covering was rather 

 more complete and the trees were slightly sprinkled with 

 snow. 



January 2, 1889. 

 G. r. s. 



12.42 P.M. 53° -220 -57. Sky cloudless. 



At this time there was unusually little snow for the season 

 of the year. The north side of the valley was in great part 

 bare even up to the mountain-tops. Though the snow- 

 covering was more complete than on May 12, the value of s 

 was less than half that on the former date. This of course 

 was due to the low altitude of the sun. 



Comparing the values of s on Aug. 9 and Dec. 2, we see 

 that the subsidiary light is nearly doubled by the replacement 

 of the green grass by the white snow, notwithstanding the 

 much brighter light on the former. It seems fair to conclude 

 that the greater part, at least three quarters, of the subsidiary 

 light in December comes from the snow. The ratio of the 

 subsidiary light to the whole being 84 : 184, we deduce that, 

 of the light from the sky near the point of maximum polari- 

 zation, at least one third has been pxcviously reflected from 

 the snow on the ground. 



It must be only too obvious to a reader of my observations 

 that I was much hampered by the limited power of my instru- 

 ment. It was barely able to deal with the polarization in the 

 middle of the day at St. Moritz and Davos, so that we are 

 still in ignorance of the strength of the polarization at sun- 

 rise or sunset. This is unfortunate, for the most striking 

 difference between the polarization at high altitudes and at 

 the sea-level would naturally be when the earth-light was 

 weakest. I am taking steps, therefore, to get a more powerful 

 instrument. Meanwhile we may notice that, even at noon, 

 the diiference is in favour of the high altitudes when the 

 ground is nearly free from snow. The strongest mid-day 

 polarization found by Rubenson at Rome was r='20 on 

 Nov. 5. Compare this with r = *185 at Davos on Aug. 9. 

 I should not like to lay much stress on the slight difference 

 between these numbers ; but, even if they were equal, it is to 



