The Stability of Human Colour Memory and Distribution of Recalled Colours

Uniform colour stimuli were presented to a user on a screen. The user memorised a target colour and then reproduced their remembered colour in tests lasting five minutes and in untimed tests. In timed tests, it was found that colour memory is disrupted to a large extent by misrepresentations of the target. The memory stability appeared greater over longer time periods. This shows that the misrepresentation memory decays more rapidly than that of the target. In untimed tests, the distribution of remembered colours from targets within the redgreen spectrum was uniform whereas for targets varying in hue, this hypothesis was rejected. An alternative model was proposed based on the quantisation of the colour range into regions verbally described by a single colour term. This is due to the use of colour descriptions in the memorisation of colours. Remembered colours for targets within each region are normally distributed. The distribution widths were calculated from the proportion of remembered colours lying within the region. This proportion was found to increase for basic colour terms that are more accurately defined and to decrease for colours that are not aesthetically pleasing.