ON THE POLARISATION OF THE ATMOSPHERE. 221 



In the normal state of the atmosphere, as represented in the Map, Plate 

 XIL, namely, when the sun is in the horizon, Arago's neutral point is about 

 18^° above the horizon, or above the antisolar point ; but when the sun is 11° or 

 12° above the horizon, and the antisolar point as much below it, the neutral 

 point is in the horizon, and consequently only 11° or 12° above the antisolar 

 point.* 



As the sun descends to the horizon, and the antisolar point rises, the distance 

 of the neutral point from the latter gradually increases ; and when the sun 

 reaches the horizon, the neutral point is 18^° above it, and therefore 18^° distant 

 from the antisolar point. 



After the sun has set, the distance of the neutral point from the antisolar 

 point increases ; that is, it rises faster than the sun descends, and its maximum 

 distance, when the twilight is very faint, is about 25°. 



When the sun is advancing to the meridian, and the light of the sky is in- 

 creasing, the distance of the neutral point from the antisolar point diminishes, 

 as shown in the morning observations on the 14th, 20th, and 21st November 

 1842. 



On a Secondary Neutral Point accompanying Arago's Neutral Point. 



"When the sea horizon was terminated by a dark purple belt about 1^° above 

 it, I observed that the vertical bands of the polariscope became brighter over 

 that belt. 



The same phenomenon was seen, but less distinctly, over the land horizon. 

 It was difficult to measure the amount of this new polarising influence, but it 

 was obvious that we should observe it separately when the neutral point came 

 above the belt. In this case, it would eclipse, as it were, the neutral point, which 

 would recover itself when it emerged from the belt. It was obvious also, that 

 when the negative or oppositely polarised bands came over the belt, the new 

 polarising influence would extinguish them where they had the same polarising 

 force, and form a secondary neutral point, the primary one being then out of 

 the belt. 



On the 8th of June 1841, at 5' 50°" p.m., when the polarised bands were 

 strongest, both on the land and sea horizon, I watched the rise of the neutral 

 point, which, as I had foreseen, did not appear first in the horizon, but about 1|° 

 above it, the compensation taking place where the vertical or positive polarisa- 

 tion was weaker than in the horizon. We had now the singular phenomenon of a 

 neutral point with positive polarisation on each side of it. When this phenome- 

 non was more fully developed under a favourable state of the horizon, the positive 



* In abnormal circumstances, sometimes only 7", 8°, 9°, or 10°, as in 1842, February 15 

 and 16. 



