But the most curious thing is that if the illumination is strong, and the screens are moved at the proper speed, no trace of red will be seen at all; it will appear exactly as if the actual colour of the wafer seen through the gap were greenish-blue. I am informed that analogous phenomena have been observed in other branches of physiology; a well-defined reaction sometimes occurs when no direct evidence can be detected of the existence of the excitation to which the reaction must be due. As in the former experiments, the effect may be shown continuously by means of a rotating disk with an open sector. The annexed diagram (Fig. 50) indicates a convenient apparatus for the purpose. The disk is made of thin metal, and properly balanced; the dark portion of the surface is covered with black velvet, and the light portion with unglazed grey or buff paper. It should turn some six or eight times in a second, while its front is well illuminated either by bright diffused daylight or by a powerful lamp. A red card placed behind the rotating disk is made to appear green, a green card pink, and a blue one yellow, while a black patch painted upon a white ground appears lighter than the ground itself. I have prepared some designs which demonstrate the phenomenon in a very striking manner. One of these is a picture of a lady with indigo-blue hair, an emerald-green face, and a scarlet gown, who is represented as admiring a violet sunflower with purple leaves. Seen through the disk, the lady's tresses appear flaxen, her complexion a delicate pink, and her dress a light peacock-blue; the petals of the sunflower also become yellow, and its foliage green. Other designs show equally remarkable transformations of colour. [Illustration: _Fig. 50.--Disk for transforming Colours._] I have mentioned only a few among many curious phenomena which have presented themselves in the course of these investigations. It is not improbable that a careful study of the subjective effects produced by intermittent illumination would lead to results tending to clear up several doubtful points in the theory of colour vision. William Byles & Sons, Printers, 129, Fleet Street, London, and Bradford. FOOTNOTES: [1] It should be clearly understood that the length of each wave of a series is measured by the distance between the crests of two successive waves. The length of water-waves which break upon a sea shore is not the length along the crest of a single wave measured in a direction parallel to the shore, as the uninitiated are apt to suppose. The true wave-length, or distance from crest to crest of successive waves, can be well observed from the top of a cliff. [2] In practice, wave-lengths are expressed in ten-millionths of a millimetre. The wave-lengths of the lines A and H of the solar spectrum, which approximately coincide with the limits of visibility, are 7594 and 3968 ten-millionths of a millimetre.