This paper reports on a simple experiment which enables splitting incident light into two different modes, each having a colour exactly complementary to the other. A brief historical development of colour theories and differences in a physicist's point of view with respect to an artist's one is discussed. An experimental system for producing colours and their physically exact complements using cellophane is presented. The origin of the colours lies in the transmission of polarized light through the birefringent cellophane, and therefore the optics of birefringent materials is briefly presented. A set-up which will be described in the following can be used in a laboratory experiment at an undergraduate level. (Contains 6 footnotes and 12 figures.)