ON examining the data contained in Mr. Pillsbury's interesting and valuable resume on colour measurements in the United States, by means of ordinates and abscissae for the various colours on squared paper, it became at once evident from the parallelism of the diagonals which could be drawn that, although previously hidden, there was a numerical relation underlying them, and that probably the measurements would furnish an affirmative answer to the question printed at the head of this letter. Can it possibly be that those compound colours which occur with such profusion in nature are the result of simple colours being combined in definite multiple proportions? Can there be a law of multiple proportions here, similar to that which holds good in the domain of chemistry?