VISIBILITY 



39 



without having to recognize its shape or 

 form. In this sense it is, perhaps, synony- 

 mous with detectabihty. This is not the 

 sense in which Luckiesh (57) uses the term 

 visibihty, but it is more in accord with 

 recent usage (2). In general, there should 

 be some kind of relation between visual 

 acuity and visibility, and we might expect 

 visibility to be low order visual acuity. 

 This is essentially correct. In addition, we 

 find that the same factors that affect visual 

 acuity — luminance contrast, time, size, and 

 luminance — also affect visibility. It should 

 also be clear that much of the data presented 

 in the section on rod and cone vision are 

 visibility data. This section, however, will 

 consider in slightly more detail some basic 

 visibility data which we shall make use of 

 in Chapters 3 and 6. 



Visibility of Lights 



Colored Signal Lights. Many light-sig- 

 nalling devices require us to recognize the 

 colors of lights. In a general way, we 

 know that if we are to do this, the amount of 

 light reaching the eye must be greater than 

 is required for merely telling the presence 

 of the Hght. Further, the more complicated 

 the color system is, the higher must be the 

 amount of hght for each color. If one 

 must recognize three different colors in a 

 light signalling system, more light is required 

 than one would need to recognize two. 

 Although these generaUzations are un- 

 doubtedly valid, exact quantitative data for 

 putting them to practical use are not 

 available. 



There are available, however, some very 

 informative experiments by Hill (40) on 

 this problem. He used a point source of 

 light, against a background equal to a 

 starlit sky, with exposures of 1 .5 seconds, and 

 with two intensities of light — one which 

 produced one mile-candle of illuminance 

 at the eye, and another which produced two 

 mile-candles. The observers knew where 

 the light was and looked in that direction. 

 Over 30,000 observations were made with 



73 different colored lights which the ob- 

 servers had to identify as red, yellow, orange, 

 white, green, or blue. 



The results of the experiment show that 

 blue and green (and yellow and orange) are 

 indistinguishable when they are far away. 

 Red and green (or blue) were the easiest to 

 recognize, then came white and last yellow 

 (or orange). A portion of the quantitative 

 data obtained in this study are shown in 



__g 99 PERCENT 

 KWl RECOQNITION 



_™, 90 PERCENT 

 l^^:^ RECOGNITION 



. , 80 PERCENT 



' ' RECOGNITION 



*^YELLOW 



0.8 



Fig. 29. The most recognizable colors for sig- 

 nalling at night. This plot employs the I.C.I. 

 chromaticity diagram which shows the location of 

 individual colors in a color triangle in terms of the 

 percentage of red (x) and green (y) needed in a 

 matching red-green-blue mixture. For details, 

 see Appendix description. (After Hill, 40) 



Fig. 29. The two intensities of light gave 

 pretty much the same result and Fig. 29 

 contains the data for only the two mile- 

 candle light. The double-hatched areas 

 are those colors which could be recognized 

 correctly 99 percent of the time. The 

 single-hatched areas show those colors which 

 could be recognized correctly 90 percent of 

 the time. There was no yellow or orange 

 which could be correctly recognized 90 

 percent of the time, but the chart shows 

 those yellows which could be recognized 

 correctly 80 percent of the time. 



