15 
Troil says, thut each saw around the shadow of his own head, 
though not around the shadows of the heads of others, a circle of 
almost the same colours which compose the rainbow, and around 
this another circle. There are many curious instances of such 
pheenoniena. I have known boys to imitate the rainbow by simply 
sjju-ting- water from their mouths, in the sunshine, &c. to imitate 
fine rain, producing the rainbow tints extremely well. 
Somethmg like this may account for the number of circles or seo^- 
ments of circles in the various substances in which they happened 
to form, even the common air. Thus we may conceive certain un- 
dulations, not unaptly represented in the common crown glass of 
our windows (which it appears is made by taking a largi>h ball of 
metal* and spreading it by turning the instrument as when a mop 
is whirled or trundled to dry it), when the arrangement of the par- 
ticles of water in the air, like the rings in the glass, varying, gives 
each a similar eH'ect, or forms circles Now it will be found that the 
eye cannot see those circles, but by the proportion of light and shadow 
with the reflexions : and betwixt every swelling, or other cause of light 
and shadow, there always appear the coloured rings when looked at 
in the sun, or when the sun shines upon the plate so formed, or 
obliquely through it on a wall, &c. I'he eye is often so circum- 
stanced, that things not at all forming real circles have yet the co- 
loured appearance of them, as a shining waxed table, a greasy ap- 
pearance on glass, or the almost imperceptible dust that may be on 
some smooth surfaces, the most minute particles of damp^ breathin"- 
on glass and copper, or the slime of snails, and many other objects 
sufficiently well known, either requiring the prism to refract the 
tints, or forming them to the unassisted eye. These colours fre- 
quently obtain permanency, and we may trace them till they seem 
to approach solidity in heated or oxidized steel or other metals, 
which begin with yellow: this is followed bv orange, then the red, 
passing to crimson or violet, and thence to blue, as the heat is con- 
» In the glass fiarnace, the melted glass is called metal by the workmen-. 
