Water, be placed in the beams of 
' be two circles of coloured 
the vertex was 46°. The applica- 
tion of this to explain the effet of the {un’s light on fmall 
globular particles of fogs is obvious ; and he follows it up 
with fome pertinent remarks upon the greater dilution of th. 
‘colours in halos than in rainbows, and fhews, that in favour- 
able circumftances, a halo of 46° diameter fhould. always 
‘be formed. Thefe obfervations, made prior te t i 
weries of Ni 
hifcor ewton, certainl 
‘author's fagacity. Hay 
halos ; 
$y eo 
10n to 
the phenomenon. ‘This explana- 
this conjeGture by an e 
i by r ion, in a veffel of ftagnant water, 
enree | r rings of colours, about the fun, which have 
abou! 
5 the 2 ae 3 the diameter of the fecond 
about 12°. He 
ions, and entitled to confideration. nomena 
les of saucy 7 — 
is 2, 0. 
fituated, as to be incapable of tranfmitti 
larly by two fucceffive refractions, direGted 
and of the remaining two-fourths, the one 
the ig within thefe three degrees, and the other would . 
difperfe the light in a {pace of between 20° and 30° beyond 
them. In the fam may imagine an immenfe 
number of prifmatic particles of fnow to be difpofed in all 
poflible direGtions, and a confiderable Proportion of them to 
of their t i 
and the {peGator. 
half of thefe only will appear ihinineadt nadeeieg ; 
i itted 
water in the atmaf , before. 
astoform acloud. Thus the flame of a candle 
through that has been 
breathed upon, or through a receiver « g its exhauttic 
or after the re-admiflion of air, exhibits the appearance 
halo. Alio water thrown up againit the fun, and difperfed 
in fmall drops, fhews the colours of the rainbow. It has b 
obferved, that large halos do not appear concentric to the 
un or moon, b si et ee “aes 4 : 
a ope ot 
