Here it is only necessary to distinguish halos from coronae.
That Newton must have begun early to make careful observations of natural phenomena is sufficiently testified by the following remarks about halos, which appear in his Optics, book ii.
Luminous arcs (T), tangential to the upper and lower parts of each halo, also occur, and in the case of the inner halo, the arcs may be prolonged to form a quasi-elliptic halo.1 The physical explanation of halos originated with Rene Descartes, who ascribed their formation to the presence of icecrystals in the atmosphere.
The impurity of the colours (due partly to the sun's diameter, but still more to oblique refraction) is more marked in halos than in rainbows; in fact, only the red is at all pure, and as a rule, only a mere trace of green or blue is seen, the external portion of each halo being nearly white.
The two halos are the only phenomena which admit of explanation without assigning any particular distribution to the ice-crystals.