Colour constancy was studied by asymmetrical matching procedure under 14 different illuminations. Two kinds of colour stimuli were used: real Munsell chips and simulated equivalent color patches presented on a Barco system. Eight subjects took part in the experiment. One of the 40 Munsell chips (T) under test illumination was presented for a short time - 3 s. After displaying of test (T) chip under the test illumination the subject had chose the such Munscll chip which color perceived under C illumination should be the same as the color of test (T) chip under test illumination. The variability of appearance of particular hues under different illuminations was calculated by measuring Brunswik ratio. It was found that the color constancy were maximal for chips which hues coincide with illumination hue and minimal for chips which hues are complementary for illumination hue. The pcrecived color depends both on a perceived color of background and on color difference betweenbackground and chip. The drift of background colour can not be explained by the adaptation (the observed drift was made quicker than adaptation process).