lxxiv 
the house, I found the solar haloes, and they continued unchanged, 
essentially, until a few minutes previous to their waning. - The 
sky was clear, except that a solitary mare’s tail, a lucid and gauzy 
cloud, was, almost imperceptibly, creeping eastwards, across the 
haloes, so that an indiscriminating eye might readily have 
associated it with them. Lest some trees at the back of my 
garden, and the large Infirmary buildings beyond, might conceal 
a part of the phenomena from my sight, I started hastily, armed 
with only a black-lead pencil and a large sheet of paper, to place 
myself westward of these obstacles. 
I made an unpremeditated sketch, which I present to the In- 
stitution; but I regret that, from my being unprovided with a 
sextant, the angular distances indicated must be regarded as only 
approximately correct. 
The time is 6 Greenwich, or 5.40 C 
local. The sun S, an hour before sunset, 
is 15° above the horizon H H. s and n 
are two short oblong vertical patches 
of light, south and north respectively 
of the sun, but slightly further from 
the horizon, each manifesting, though * 
only in a blurred manner, the prismatic SS eA 
colours withthe leastrefrangible nearest R R 
to, and the most so furthest from, the ; 
Sun. Above was a bow, whose plane was 
parallel, or thereabouts, to the hori- 
zontal, exactly like the ordinary rain- 
bow, with, as in the patches of light, 
but very distinctly, the Red R BR next, 
and the Violet V V most remote from 
the Sun. Speaking of the objects 
already described as if limited to their 
middle lines or points (as the case may 
require) the bow constituted about one- 
sixth of a circle, whose centre C H H 
plainly lay in the azimuthal plane Reduced from Sketch. 
passing through the Sun, though I could not settle in my mind 
whether C was coincident with my zenith, or lay slightly to the 
west of it, so difficult was it to decide from so small an are. I 
estimated that the imaginary lines CG and SG were equal, 
and either of them twice as long as Ss or Sz. I also thought 
it probable that a circle described from G as a centre, and of radius 
GS, might pass through n and s. Thuss and 2 would be regarded 
as coloured parhelia, at the intersections of these two imaginary, 
though to me invisible, circles. 
Be ee 
C 
