Ten light-adapted observers viewed six narrow-band wavelength displays and a white display through two artificially turbid and illuminated water media. The two water samples were intended to simulate the ambient underwater environment of a near-shore ocean at 20 meters depth and an inshore harbor or bay at 5 meters depth when surface illumination was full sunlight, 10, 000 ft-candles. The Ocean simulation was composed of a low concentration of relatively large-sized suspensoids illuminated by 17 ft-candles of light spectrally restricted to the green region, 540 nm peak wavelength. The Harbor simulation was composed of a high concentration of relatively small-sized particles in suspension and illuminated by 214 ft-candles of light spectrally restricted in the yellow region, 600 nm peak wavelength. Against these two backgrounds, observers viewed 2-digit, self-luminous displays which varied in size and color. Observers controlled display luminance to three magnitudes defining three levels of display legibility Minimum, Clear and Limit. At each legibility criterion, observers described the color appearance of the display using a forced-choice, color-naming technique and a restricted set of color-name alternatives blue, green, yellow, red and white.