196 Incident Light not decomposable into the prismatic Colours. 
as I am also certain that science and liberality always go hand in 
hand, I rest my ideas on experimental inquiry. 
To show that the decomposition of light takes place only at 
the prismatic angles, and arises entirely from those fringes of re- 
flected light, I made the following experiment: When the sun 
was shining very powerfully, I placed my prism on a table at the 
open window; and having formed a spectrum on a sheet of white 
paper, I slowly turned the instrument on its axis, until I separated 
the red and yellow from the blue, and in place of green, white 
or undecomposed light passed through between the angles. I 
now ascertained that the red and yellow rays passed through the 
upper and thin angular edge, by intercepting them with my fin- 
ger placed on it; and by running my finger along the middle an- 
gle, I intercepted the blue rays; and by pasting a narrow strip 
of paper between those two angles I made two spectra. But to 
place the fact beyond the possibility of doubt, standing at a little 
distance, I looked by means of another prism at the light passing 
through, and perceived three beautiful fringes hanging from the 
angles. Indeed it is surprising that those fringes, so easily 
proved, so evident to the eye, and so highly important in their 
consequences, should have escaped the observation of such able 
and accurate experimenters as those already mentioned. I shall 
conclude this paper with the following deductions. 
Ist. That incident light has never yet been decomposed ; and 
that Sir Isaac Newton and other philosophers only decomposed 
light refiected from opaque substances, or fringes of blue, red, and 
ellow. 
2dly. That there are but three primary colours, blue, red, and 
yellow, by the mixture of which, either by the prism or the painter, 
all the others are formed. 
3dly. That Herschel, Leslie, Davy, Englefield, and other phi- 
losophers, drew their conclusions, relative to the heating power of 
the prismatic colours, from erroneous data, viz. from experiments 
on reflected light, whose heat must in a great measure depend 
on the reflecting media, and also on the thickness or thinness 
of those parts of the prism through which the fringes pass. Thus 
the red and yellow rays passing through the very thin upper an- 
gle, must be accompanied by more radiant caloric than the blue 
rays, which pass through the thickest. 
The following diagram will demonstrate my opinions; and as 
I am at present engaged in a series of experiments, to prove 
that the prismatic colours have similar heating powers, I shall 
not anticipate. 
Let s represent the sun (fig. 3. Plate III.) d. o,f. rays of un- 
decomposed light, impinging on the angles A, B, C of the prism. 
These 
