Abstract
Purpose:
Long-wavelength light that looks yellow in a black surround looks brown in a white surround. No other unique hue (blue, green, red) shows such distinct hue change when seen on white surrounds: they simply appear darker. We determined whether cone and rod weights differ for yellow vs. brown, and what differences occur for bright vs. dark versions of other hues.
Methods:
Observers saw a 1-s duration 2°-diameter foveal disk presented in a full-field surround that was either white (141 cd/m2, CIE 10° x,y .29,.29) or black, and adjusted the disk hue around the RGB-phosphor-mixture triangle of the CRT display at 21 cd/m2 in both surrounds. To assess cone balances, observers adjusted the hue to (a) red-green (RG) null for yellow, brown, and blue disks and (b) blue-yellow (BY) null for red and green disks. To assess rod hue bias, we reduced light levels by 2 log units to mesopic level, and compared hue balances under dark adapted and bleached conditions.
Results:
Cone balances. Yellow/brown: All 5 observers showed different RG balances for yellow and brown, with balanced-brown chromaticity shifted toward red compared to unique yellow at both photopic and mesopic levels. Blue: 3 of 5 observers showed the same direction of shift for bright/dark blue as for yellow/brown. Red and green: Shifts in BY hue balances between bright/dark versions were small or inconsistent across observers. Rod hue biases. Yellow/brown: Rods exerted a green bias on yellow for all 4 observers but a red bias on brown for 3 of 4 observers. Blue: Rods also exerted a green bias on bright blue for 3 of 4 observers. 2 observers showed a red bias for both dark blue and brown. Red and green: Rod effects were inconsistent across observers.
Conclusions:
Both RG balance and rod hue biases differ when the same test light appears yellow vs. brown. The different RG balances suggest M cones have stronger weighting relative to L cones in the portions of midget pathways mediating RG balance for brown compared to those for yellow. The opposite rod hue biases can be explained by rod and cone signals mixing with same sign in ON and OFF midget pathways. While some observers showed similar effects for light vs. dark blue, results were inconsistent for light vs. dark green and red. Thus cone and rod input differences were not apparent in S-cone pathways for light vs. dark colors, and may be exclusive to midget pathways.
Keywords: 471 color vision •
641 perception