The last curiosity which I will notice is connected with the fact already mentioned, that when the illumination is strong, the transient border-colours are nearly reversed, greenish-blue appearing in place of red, and brick-red in place of blue. I was for a long time quite unable to imagine any reasonably probable explanation of this circumstance, but a clue was finally obtained from consideration of the fact that greenish-blue is the complementary colour to red, and in a subsequent memoir (Proc. Roy. Soc., vol. 61, p. 269) some experiments were described which show, as I believe conclusively, that the greenish-blue borders seen in a strong light are simply negative after-images of the usual red one. These negative after-images are of the familiar kind that are observed after one has gazed for some time at a bright coloured object. If a red "wafer" lying upon a sheet of white or grey paper is looked at steadily for about half a minute, and the gaze is then suddenly transferred to some other part of the paper, a greenish-blue ghost of the wafer will be seen. The portion of the retina upon which the red image at first falls becomes fatigued and partially insensible to red light; it is therefore unable to appreciate the red component of the white light afterwards reflected to it by the paper, and the sensation of the complementary colour consequently predominates; hence the greenish-blue ghost, which is called the negative after-image of the wafer. The new experiments show that, if a certain condition is fulfilled, the usual prolonged stare becomes unnecessary, a momentary glance sufficing to produce a strong but fugitive after-image. The condition is that the part of the retina where the image is to be formed, shall have been darkened immediately before excitation by the bright object. The retinal nerves, when in darkness, rapidly acquire a state of sensitiveness far exceeding the normal average in the light, but quickly diminishing again under the influence of illumination. This peculiar sensitiveness may, indeed, be both gained and lost in a small fraction of a second, and is therefore very favourable for the rapid generation of negative after-images. Once more making use of the black and white screens depicted in Fig. 46, let the black screen first cover the paper upon which the wafer is lying; this will darken a portion of the retina, and render it sensitive. Then let the screens be quickly moved sideways, so that the wafer, after having been seen for a moment through the opening, may be immediately covered by the white screen. A bright but evanescent greenish-blue ghost will succeed the red impression.