SUPPLEMENTARY REPORT ON METEOROLOGY. 133 
(4). The important fact ascertained by M. Arago, that the 
light of halos is polarized by refraction and not by re- 
flexion *. 
283. But other evidences arise from the application of the 
same principles to the phenomena which accompany the first 
halo, and many of which Mariotte pointed out so clearly as to 
supersede much which has since been written on the subject. 
284. The halo of 46° may be ascribed to two refractions with 
the minimum deviation ti:rough two successive prisms of 63°, 
or as Cavendish supposes f, to the refracting angle of 90° formed 
by perpendicular terminations of the ice-prism. Such termina- 
tions, Dr. Young observes, are rather doubtful, and in his later 
writings { he inclines to the doctrine of successive refraction 
through two prisms. Fraunhofer § supposes pyramids with an- 
gles of 88° to be the cause; and he assigns as an explanation, 
which appears to him perfectly satisfactory, of Hevelius’s great 
circle of 90°, the limiting angle of total reflewion in six-sided 
prisms, which is 89° 56’. Hence this circle is white. 
285. Parhelia«.—Mariotte accounted for this phenomenon by 
the preponderance of ice-prisms in a vertical direction (which 
he attributed to their being heavier at one end,—Dr. Young, 
more justly, to the resistance of the air ||), which therefore will 
form a brighter coloured image of the sun when the plane of re- 
fraction is horizontal than in any other plane. Mariotte had 
even the acuteness to see, that as the sun rose above the horizon, 
the plane of refraction being no longer accurately horizontal in 
order to reach the eye, the virtual refracting angle of these ver- 
tical prisms would necessarily be increased, and the deviation 
being greater, the parhelia would stand out beyond the limit of 
the halo, a fact remarkably coinciding with experience. Thus 
an officer of Sir Edward Parry’s Expedition§, observed par- 
helia distant 24° 40’ from the sun, the halo being at 22° 30). 
The parhelia are coloured, the red edge being nearest to the 
sun. 
286. The horizontal or parhelic circle is not accounted for 
by Mariotte, because (he says) he had not an accurate descrip- 
tion of it. His commentator, Dr. Young, ascribes the horizon- 
* Bulletin Universel (Ferussac), 1825, Sci. Math. iii. 304. 
t Young, ut supra, 
} See the article Chromatics in the Encyclopedia Britannica. But the same 
doctrine is stated in his earliest papers, to which nothing of any consequence 
has been added in elucidation of this subject, except M. Arago’s important ob- 
servation of the plane of polarization. 
§ Schumacher’s Abhandlungen, iii. 77. 
|| Journal R. Inst. and Lectures, ii. 307 (1807). 
1 Quoted in art. Meteorology, Encyclopedia Metropolitana, p. *169, 
