ABSTRACT
By the Purkinje phenomenon is usually meant the curious fact that blues and especially greens look abnormally bright when the illumination is diminished, and that yellows and especially reds look abnormally dull. The name commends itself as being also a proper designation for the still more curious fact that when a red and a blue-green are combined so as to give a colourless mixture, and also blue and yellow, these two mixtures, if made equally bright for an ordinary illumination, differ very much from each other, and also from an undecomposed colourless light, when the illumination is faint. This is in spite of the fact that, as far as the undiscriminating human eye can make out, they are all three identically the same thing. Tschermak, in a recent paper, 1 does not make use of this term, but it is evidently the right name for the fact, and its use will facilitate the report of Tschermak’s observations. It is the Purkinje phenomenon only in so far as it deals with these colourless mixtures that is the subject of the present paper. The fact itself is evidently in direct contradiction with Newton’s law of colour mixtures. That law states that if there are two pairs of indistinguishable light-mixtures, the double mixtures formed by uniting them two and two will also be indistinguishable; as a particular case, one particular pair of light-mixtures may be the same as the other pair, and hence the law covers as well the supposed constant equivalence of two light-mixtures (and in particular of two colourless light-mixtures) under all variations of objective intensity of illumination. The difficulty which besets the search for truth by experimental means, even in well-appointed laboratories, is well illustrated in the historical sketch with which Tschermak begins his paper. Newton’s law was tested experimentally and proved to be valid by Maxwell and by Aubert by means of the colour-wheel. As Hering pointed out, this is very inadequate proof, so far as colourless mixtures are concerned, because greys made out of different coloured papers are very likely to be, in the end, physically exactly alike. [This objection does not hold for a disproof of the law; for it is evident, even in the absence of a spectroscopic examination, that a difference in sensation-effect could only be due to a difference in light-ray composition.] v. Kries and Brauneck (1885) tested the law by spectral lights (colourless equations being included) and declared it to be valid. At the same time Hering published the results of his own investigation of the question, both by papers and by a good colour-mixing apparatus, and declared the “complete constancy” of colourless equations under changing illumination.
