The aim of this study was to compare the relative therapeutic efficacies of three different light sources for treating winter depression. A balanced incomplete block crossover design was employed, whereby all patients (n = 18) were randomly assigned to two out of the three treatment conditions: white, red and blue light. The degree of depression was assessed by the 21-item Hamilton Depression Rating Scale. The data suggest that at a photon density of 2.3 X 10(15) photons/s/cm2, white light has greater therapeutic benefit than red or blue light. It is clear that a larger sample population should be tested to confirm this result. This preliminary finding indicates that light sources currently in use for phototherapy could not be improved by narrowing the wavelengths provided and shifting them towards either end of the visible spectrum.