Effect of Snow on the Polarization of the Sky

Abstract

THE polarization of the sky has been shown experimentally by Tyndall and theoretically by Lord Rayleigh to be due to fine particles suspended in the atmosphere. According to both, the sunlight scattered at right angles to its original direction by very small particles is completely polarized in a plane through the sun. In observation, however, we find the light from a region of the sky distant 90° from the sun is only partially polarized. This is due to that portion of the atmosphere being illuminated not merely by the sun, but also by the rest of the sky and the surface of the earth, and partly also no doubt to some of the particles not being sufficiently small compared with a wave-length. From these considerations we may expect that a fall of snow would cause a considerable diminution of the polarization. This expectation has been fulfilled in some recent observations of mine here at 6000 feet above sea-level. My polarimeter consists essentially of two piles of glass plates to depolarize the sky light; and a crystal and Nicol prism to test the depolarization. Owing to the strength of the polarization at this attitude, I find it necessary to use two piles of glass plates separated by two or three inches. This arrangement diminishes the number of double internal reflections, and so is a much more powerful polarizer or depolarizer than the same number of plates combined into one pile. As I have not seen this important practical consideration noticed before, I may point out that, in addition to the light refracted directly through the pile, there are a number of portions twice reflected. One of these for instance is reflected first at the second surface of the last plate, and secondly at the first surface of the last plate. The number of such twice-reflected portions for n plates is n (2n - 1). When, as in my instrument, the fixed pile is much inclined, no light can reach the edge after being reflected first by one pile and then by the other. If the two piles were combined into one, I should have 120 portions twice reflected; as it is, I have only 60. This increases the polarizing power of the instrument by at least one-third.