opening the eye I constantly revived the image, so that it did not quite disappear till seven minutes had elapsed.
Future observers may find these periods shorter or longer as their eyes may be stronger or weaker (23), but it would be very remarkable if, notwithstanding such variations, a corresponding proportion as to relative duration should be found to exist.
42.
But this remarkable phenomenon no sooner excites our attention than we observe a new modification of it.
If we receive the impression of the bright circle as before, and then look on a light grey surface in a moderately lighted room, an image again floats before us; but in this instance a dark one: by degrees it is encircled by a green border that gradually spreads inwards over the whole circle, as the red did in the former instance. As soon as this has taken place a dingy yellow appears, and, filling the space as the blue did before, is finally lost in a negative shade.
43.
These two experiments may be combined by placing a black and a white plane surface next each other in a moderately lighted room, and then looking alternately on one and the other as long as the impression of the light circle lasts: the spectator will then perceive at first a red and green image alternately, and afterwards the other changes. After a little practice the two opposite colours may be perceived at once, by causing the floating image to fall on the junction of the two planes. This can