36 



HOW WE SEE 



background, so that there is some luminance 

 contrast present also. 



One'set of results obtained on this function 

 is illustrated in Fig. 25. These curves are 

 taken from a study by Ferree and Rand 

 (25, 26)^and represent average data for 

 three subjects. In one-half of the experi- 

 ment, the visual acuity test was illuminated 



100 



50 



025 



EQUAL IN SATURATION 



UNEQUAL IN SATURATION 



450 500 6 00 700 



WAVELENGTH IN m)j. 



Fig. 25. Visual acuity as a function of the color 

 of illuminance. (Data from Ferree and Rand, 

 25, 26) 



with spectrally pure light of the wavelengths 

 shown. Spectrum lights, however, differ 

 markedly in both dominant wavelength 

 and colorimetric purity (see Appendix) and 

 in their psychological equivalents, hue and 

 saturation. To isolate the effect of hue 

 alone in the second part of the experiment, 

 these investigators equated the saturations 

 of the illuminants by mixing enough white 

 light with the spectrum lights until they 

 appeared equally saturated. The results 



from both parts of the experiment agree in 

 showing that visual acuity is best when the 

 illuminant is yellow in color. Visual acuity 

 drops off markedly when the colors are 

 selected from the ends of the spectrum. 

 Almost identical results were obtained in a 

 study by Luckiesh (56), also done apparently 

 on three subjects. 



It is difficult to get very much intensity 

 from most optical systems used to produce 

 spectral lights. For this reason, the illumi- 

 nances used in the study by Ferree and 

 Rand were very low: 0.075 and 0.31 foot- 

 candles. Since the reflectance of the white 

 background was about 85 percent, the 

 background luminances were about 0.064 

 and 0.26 ft.-L, respectively. In the study by 

 Luckiesh, the illuminance was 4.2 foot- 

 candles but the background luminance 

 cannot be computed from the data given. 

 This is less light than is ordinarily used for 

 reading and other visual tasks. It would be 

 informative to have similar e.xperiments 

 done with much higher light values. 



Two other findings of the Ferree-Rand 

 study are worth noting, (a) Measurements 

 of speed of vision (time needed to recognize a 

 test object) and the power to sustain clear 

 seeing (the length of time an observer could 

 focus the pattern "li" clearly) gave results 

 essentially identical to those obtained with 

 the acuity measurements. For all three 

 visual functions, yellow light gave the best 

 results, (b) Light from a Macbeth daylight 

 lamp (roughly equivalent to illuminant C) 

 gave better results than any of the spectrum 

 lights even when they were all equated in 

 illuminance. Light from a Mazda lamp — 

 which is considerably yellower than a day- 

 light lamp — gave results almost the same as 

 those obtained with spectral 3-ellow. 



Visual Acuity as a Function of Color of 

 Illumination at Rod Levels. Two experi- 

 ments of a slightly different sort are those 

 reported by Shlaer et al. (82) and Luckiesh 

 and Taylor (59). Shlaer and his col- 

 laborators studied visual acuit}^ over a wide 

 range of luminances, but with only two 



