LUMINOUS IMPRESSIONS ON THE EYE. 587 
ard 
2 2 
Ceres 
6, ant x 
d,? 
It has here been assumed that the brightness of the gas-flame remains con- 
stant during the experiment, a condition which is not fulfilled in practice, owing to 
the variable pressure of the gas in the pipes. It is probable, however, that the 
brightness of both flames will vary nearly in the same proportion, so that the 
distances necessary to equalise the apparent brightness of the apertures in the 
screens will remain almost unaltered ; and it is obvious also, that any residual 
error, arising from a gradual change in the brightness of the flames, will be nearly 
eliminated by taking the mean of a series of observations immediately succeeding 
each other, and conducted in the manner now described. 
Il. Proof of the Gradual Action of Light on the Eye. 
If a disc, with a sector of a small angle, is made to revolve between the eye 
and a luminous object, a flash of light is seen at each revolution ; but as the velo- 
city of rotation increases, the brightness of the flashes diminishes, which shews 
that the apparent brightness of a luminous object diminishes as the time during 
which it is visible becomes shorter. A similar result is obtained by placing two 
discs with sectors of different angles before the screens of the selaometer, and 
observing the relative intensity of the simultaneous flashes of light when the discs 
are made to revolve. It will always be found that although the apertures in the 
screens are equally bright, when seen by uninterrupted vision, the disc whose 
sector has the greater angle produces the brighter flash. Now as both discs re- 
volve with the same velocity, the length of the luminous impressions will be pro- 
portional to the angles of the sectors; so that, by this experiment also, the ap- 
parent brightness of the light is shewn to increase with the time during which it 
continues to act upon the eye. A convenient selascope, which exhibits this phe- 
nomenon in a striking manner may be made by causing a disc with a sector of 
the form EF GH (Fig. 4), to revolve before a luminous aperture. The flash pro- 
duced by the wide part of the sector E F, greatly exceeds in brightness that pro- 
duced by the narrow part G H. 
The experiment of causing a disc, with a sector cut in it, to revolve before a 
luminous aperture, also affords a simple proof of the duration of luminous impres- 
sions on the retina. At about seven revolutions in a second the luminous impres- 
sion becomes continuous, so that the aperture always appears visible, even during 
the dark interval between the successive passages of the sector between it and the 
eye; but a considerably higher velocity, about twenty revolutions in a second, 
is required in order to produce a sensibly uniform impression. These velocities 
are not meant to be stated with great exactness. They were obtained by experi- 
