without any intervening paper, no coloured border could be seen, owing, as it seemed, to the glare.
Experiment III.
The aperture in the metal plate was again covered with white paper, having a strip of tinfoil across it, and the plate was fixed before the window in the box, as in Experiment I ; a 16-candle power lamp was placed immediately behind it. When the lamp was switched on, the red border w7as distinctly seen to be backed with greenish-blue, the red itself being much less evident than when the lamp was 18 in. ( 4 5 cm.) behind the aperture.
I have hitherto failed to detect any greenish-blue near the border when the disk was suddenly illuminated by the shutter method of Experiment I, instead of by switching on the lamp.*
Experiment IV.
The object of this experiment was to ascertain whether the red border could be produced by the sudden accession of light which contained no red constituent. Ten different coloured glasses were successively interposed between the lamp and the aperture with the shutter. In every case Avhen the spectroscope showed that the glass transmitted red light, the tinfoil strip became red, but never otherwise. For example, it reddened with a dark blue cobalt glass, but not with a blue glass Avhich transmitted much more light, but intercepted the red end of the spectrum.
Experiment V.
The momentary redness around the edge of the suddenly illuminated disk and along the tinfoil strip, as described in the account of the previous experiments, can only be seen by a practised observer. By a different method, hoAvever, it can be made quite evident to almost any person w'hose vision is normal.
The paper-covered aperture in the box was arranged as before, but the shutter was not used. An incandescent lamp was placed inside the box, and a second lamp outside, at a distance of a few inches from the aperture, the observer’s eyes being shaded from it by a screen. The tinfoil strip was on the interior side of the paper, and nothing was seen of it from outside, except when the lamp in the box was alight.
A rotating commutator was constructed, by means of which
- [Since this was written, I have found that the greenish-blue may be shown by
the shutter method without difficulty if the distance of the lamp from the aperture is suitably adjusted.—Dec. 19.]
