COLOR-VISION RESEARCH PROCEDURE 471 



animal to a color versus a medium gray, substituting other grays from a 

 finely-graded series after the animal has become thoroughly trained to 

 the color. If no gray elicits confusion between it and the positive stim- 

 ulus, then the latter is being seen qualitatively. Several publishers here 

 and abroad offer long series of colored papers, and gray ones ranging 

 from pure white to dead black. In some of these gray series, no two 

 adjacent samples can be told apart by the human, and it is not likely that 

 many animals (except birds) have any finer capacity for brightness dis- 

 criminations. Some of your critics, even so, are sure to say that if you 

 had used more shades of gray, one of them would have confused the 

 woodchuck which, you insist, can see blue. To silence all critics, one 

 simply must use colored lights, whose intensities can be very gradually 

 regulated. It will help though, if you establish that a woodchuck trained 

 positive to a medium gray cannot identify it alongside of neighboring 

 shades of gray. 



Having established that your animals distinguish the hues of red and 

 blue, you have made but a bare beginning. You now re-train them to 

 other pairs of colors and try to find out how many hues they can dis- 

 criminate as compared with man's 160 or so. By training to darkness 

 versus a red wavelength, and increasing that wavelength slowly, the limit 

 of the animal's spectrum can be found at that wavelength which, how- 

 ever intense physically, is invisible — at the border of woodchuck infra- 

 red. So, also, the violet spectral limit can be located. In the woodchuck, 

 it will not be as low a wavelength as that of man, because of the strongly 

 yellow coloration of the lens of the animal. Systematic pursuit between 

 the animal's spectral limits, using pairs of wavelengths which are fairly 

 close together in the spectrum, with the negative hue (after training) 

 being gradually approximated to the positive one, will enable you to plot 

 a graph of the rate-of-change-of-hue against wavelength. Comparison of 

 this with the human curve will be interesting, and may be extremely 

 valuable to color-vision theories and theorists. Similar examinations of a 

 series of red-blue mixtures will disclose whether the color circle is closed 

 by purples for the woodchuck, as it is for man. Perhaps you will find it 

 possible to detect chromatic after-imagery in the woodchuck — by training 

 him positive to blue-green, for instance, then seeing whether he gives the 

 positive reaction to a neutral gray after being fatigued with red. And 

 of course you will wish to determine his brightness-threshold for various 

 colors, and to ascertain how color equations hold for him — what mixture 

 of red and green looks the same to him as a given orange, what mixtures 



