values for gas- and solar light. The curves of the two observers agree, when differences of absorption in the macula lutea are allowed for. Their correctness is controlled by the comparing of the color complementary to a terminal color with the light corresponding to the point of intersection of two elementary-sensation curves. (b) Anomalous trichromatic systems. § 18. The color equations: their immediate results; and the calculation of the elementary-sensation curves R', G', and V'. § 19. Survey and transformation of the results: testing by complementary colors. § 20. Comparison with normal trichromatic systems. The anomalous curves run less evenly than the normal; but this is to be ascribed to errors of observation only. As re- gards the R' curve: its maximum coincides with that of R, but the forms of the curves differ somewhat. As regards G': the curve differs greatly from G. Its maximum (in both gas-light spectra) is shifted towards the red end of the spectrum, and its form is inter- mediate between those of R and G. In the solar (diffraction) spectrum, the maximum nearly coincides with that of G, but the form of the curves is different. As regards V': there is no differ- ence between it and V.
(5) THE GROUND-SENSATIONS. – § 21. Definition of ground-sensation, and its relation to elementary sensation. A ground-sensation is a sensation to which corresponds a simple peripheral process of the N. opticus. (Donders' fundamental color). There cannot be fewer ground-sensations than elementary sensations; there is no proof that there are more; we assume that there is an equal number of both. The intensities of the ground-sensations of a color system are homo- geneous linear functions of the intensities of its elementary sensa- tions. § 22. The mutual relations of the different color systems. The most obvious assumption is, that the ground-sensations of monochro- matic and dichromatic systems are identical with ground-sensations of trichromatic systems; or at least, that this holds for monochro- matic and dichromatic systems. The assumption can be tested experimentally and calculatively. Experimentally, tri- and dichro- matic color equations must be valid for monochromates; trichromatic for dichromates. The converse need only occur as an exception. Calculatively, a similar result must be gained by aid of the element- ary-sensation curves and color equations already obtained. (a) As regards monochromatic systems: the assumption is not valid. Connate monochromatism cannot be explained as arising by the elimination of one or two of the ground-sensations of the dichromatic and trichromatic systems hitherto investigated. This fact refutes
