134 REPORT— 1840. 
tal circle to ‘‘ the reflexion or even the repeated refraction of 
the vertical facets*’’, an explanation which seems entirely satis- 
factory ; for though amorphous vertical fibres would produce the 
same effect (as Fraunhofer showed), yet it is much more natu- 
ral to ascribe the horizontal circle to the same cause with the 
parhelia which occur within it, and which are, in fact, merely its 
most notable points. Young attributes the mock suns occa- 
sionally observed at an elongation of about 142° to two refrac- 
tions and one reflexion in the same ice crystals+. 
287. Fraunhofer, in his Memoir already cited{, ascribes the 
horizontal circle to the superposition of diffraction-spectra, 
which produce an excessively elongated image of a body viewed 
by reflexion from a striated surface, like that of glass smeared 
with grease, and then cleaned by rubbing it in one direction, 
when a whitish reflexion will take place perpendicular to the 
streaks. A similar view has been given more lately by M. Ba- 
binet §, who compares the horizontal reflexion to that observed 
from fibrous crystals, such as topaz and gypsum. 
288. Contact-arches||.—The only remaining phenomenon 
which seems fairly accounted for, is that of inverted ares of lu- 
minous circles touching the halos, usually at their vertical dia- 
meter and accompanied by a parhelion, so that Dr. Young de- 
scribes it as “a bright parhelion immediately over the sun, with 
an appearance of wings cr horns diverging upwards from the 
parhelion.”’ This Dr. Young has ascribed with great inge- 
nuity and probability to very short triangular prisms, which 
from their flatness fall with their axes and refracting edges in 
a horizontal position, so that the plane of refraction is vertical. 
The abundance of such prisms (compared to those which fall 
obliquely and form the halo) give rise to the vertical parhelia 
(which Fraunhofer has, I think very unsatisfactorily, explained 
by diffraction**), Horizontal prisms parallel to the former, 
lying to the right or left of a vertical plane, passing through the 
observer and the sun, will evidently refract the solar image in a 
plane not perpendicular to the axis of the prism (because not 
vertical), and for which the refracting angle being greater, the so- 
lar image (formed always at the angle of minimum deviation,) 
will appear more elevated as the obliquity of refraction is greater, 
that is, as we proceed to the right or left from a line vertically 
above the sun. Dr. Young has confirmed his view by actual cal- 
culation Tf. 
* Lectures, i. 444. + Chromatics, sect. ii. paPasye 
§ Comptes Rendus ; ut sup. || See art. 277. (3). 
“| Lect. ii. 807. col. i. See the figure of Hevelius’s halos, and others. 
** Schumacher, wt supra, p. 78. tt Lect. ii. 308. col. i. 
