The ability of colour defective observers to recognise an optimised set of red, green and white signal lights

Restricted red and green colour domains have been recommended by the CIE in order to assist colour defective observers identify coloured signal lights. However, the benefits, if any, of using such a restricted domain with point source signals have never been demonstrated. Sloan and Habel showed that if the colours were relatively large and bright then only protanopes will have difficulty although other studies conclude that some congenital colour defective groups cannot reliably recognise colour signal lights even if their colours lie within the CIE restricted domains. The purpose of this study is to compare the ability of 20 congenitally colour defective observers and 4 normal controls in identifying an optimised and well practised set of point-source coloured signal lights with the recognition of a more variable set of colours chosen from the CIE restricted boundaries. If substantial improvements in performance are found then there may be an ergonomic alternative to excluding colour vision defectives from occupations that require the recognition of coloured signal lights. Our results indicate that optimising signal colours and allowing training in their recognition does not afford reliable recognition by colour defectives. Protanopes appear to be particularly disadvantaged regardless of the colours used. We note that the present yellow limit for red signals may be too liberal and that longer wavelength reds afford more reliable recognition by all observers. We note large individual variability in performance although the lantern test appears to separate those observers who are able to perform the task from those who cannot.