It is shown that the linearity of the photometric scale of spectrophotometers is not self-evident. The widely used Lambert-Beer test does not give a conclusive proof of linearity. Light addition methods do not show this disadvantage. In the form of the supplementary light method, as used here, they are suited for testing commercial instruments. The arrangement for testing the Zeiss DMR 21 recording spectrophotometer is shown, and the results do not reveal a nonlinearity exceeding 0.1% T. The accuracy of the method is limited only by the precision of the instrument under test.