spectrum became yellow, and returned
instantly again to green, as often as the hands were applied to cover the
eyelids, or removed from them: for the retina being now insensible to red
light, the yellow rays passing through the eyelids in greater quantity than
the other colours, induced a yellow spectrum; whereas if the spectrum was
thrown on white paper, with the eyes open, it became only a lighter green.

Though a certain quantity of light facilitates the formation of the reverse
spectrum, a greater quantity prevents its formation, as the more powerful
stimulus excites even the fatigued parts of the eye into action; otherwise
we should see the spectrum of the last viewed object as often as we turn
our eyes. Hence the reverse spectra are best seen by gradually approaching
the hand near the closed eyelids to a certain distance only, which must be
varied with the brightness of the day, or the energy of the spectrum. Add
to this, that all dark spectra, as black, blue, or green, if light be
admitted through the eyelids, after they have been some time covered, give
reddish spectra, for the reasons given in Sect. III. Exp. 1.

From these circumstances of the extraneous light coinciding with the
spontaneous efforts of the fatigued retina to produce a reverse spectrum,
as was observed before, it is not easy to gain a direct spectrum, except of
objects brighter than the ambient light; such as a candle in the night, the
setting sun, or viewing a bright object through an opake tube; and then the
reverse spectrum is instantaneously produced by the admission of some
external light; and is as instantly converted again to the direct spectrum
by the exclusion of it. Thus, on looking at the setting sun, on closing the
eyes, and covering them, a yellow spectrum is seen, which is the direct
spectrum of the setting sun; but on opening the eyes on the sky, the yellow
spectrum is immediately changed into a blue one, which is the reverse
spectrum of the yellow sun, or the direct spectrum of the blue sky, or a
combination of both. And this is again transformed into a yellow one on
closing the eyes, and so reciprocally, as quick as the motions of the
opening and closing eyelids. Hence, when Mr. Melvill observed the
scintillations of the star Sirius to be sometimes coloured, these were
probably the direct spectrum of the blue sky on the parts of the retina
fatigued by the white light of the star. (Essays Physical and Literary, p.
81. V. 2.)