Psychophysics of reading. IV. Wavelength effects in normal and low vision.

Does the color of text influence its legibility? There are reasons why it may do so for specific groups of low-vision observers. We used psychophysical methods to measure the effects of wavelength on the reading performance of four normal observers, two dichromats, and twenty-five low-vision observers. Reading rates were measured for text scanned across the face of a television (TV) monitor. We compared performance under four luminance-matched conditions in which sets of neutral-density and Wratten color filters were placed in front of the TV screen--blue (lambda max = 430 nm), green (lambda max = 550 nm), red (lambda max = 650 nm), and gray. Under photopic conditions, the reading rates of normal subjects were independent of wavelength, with the exception of characters near the acuity limit. At lower luminances, wavelength effects could be explained by the shift from photopic to scotopic vision. It was hypothesized that light scatter or absorption in eyes with cloudy ocular media would result in depressed performance in the blue. Only one of seven subjects demonstrated this effect, which we traced to wavelength-specific absorption. Observers with advanced photoreceptor disorders tended to read blue text faster than red text. This could not be explained on the basis of photopic spectral sensitivities alone. Finally, the presence of central or peripheral field loss was not predictive of wavelength-specific effects in reading. On the whole, wavelength only occasionally plays a significant role in reading. When it does, performance tends to be depressed either in the red or the blue and to be nearly optimal for green or gray.