1897.] NEW YORK ACADEMY OF SCIENCES. 193 



others when the two sources of light differ in color. Under this 

 condition it is impossible to obtain a uniform field. 



The accuracy of the photometric methods, so far considered, 

 is dependent upon the sensitiveness of the eye to slight inequal- 

 ities of illumination. Various experiments * for the purpose of 

 determining the degree of sensitiveness show that for white light 

 the eye can detect differences in luminosity of from 3 per cent, 

 to about 6 per cent, of the total luminosity of the field, the 

 sensitiveness being less for high and low illuminations than for 

 a field of medium luminosity. If, however, the surfaces compared 

 are of different colors, a judgment concerning the equality of 

 their luminosities is subject to a very much greater error. In- 

 deed Helmholtz f remarks : 



" I must explain that personally I put no confidence in my 

 judgment concerning the equalitj^ in luminosity of differently 

 colored surfaces. I admit, however, that of two differently 

 colored fields one can be so much darkened that there remains 

 no doubt that the other is brighter. 



" For myself I have the impression that in the comparison of 

 the luminosities of differently colored fields I am not dealing 

 with a single magnitude but with the effect of two, luminosity 

 and color, for which I do not know how to form a simple sum, 

 and which I cannot scientifically define." 



There have, nevertheless, been many attempts made to com- 

 pare directly the luminosities of differently colored surfaces. 

 These attempts have, in general, led to slight modifications in 

 the use or construction of the photometers belonging to the 

 types already described. 



The diflSculty of judging when two differently colored surfaces 

 are of the same luminosity, a difficulty encountered in the use of 

 photometers belonging to any of the classes above described, is, 

 to some extent, obviated by taking the mean of two measure- 

 ments. A measurement is taken when one of the surfaces is 

 evidently more luminous than the other, another measurement 

 when it is less luminous by the same amount. The average of 

 the two measurements gives a determination of the luminosity. 

 This method was employed by Professor 0. N. Rood in measur- 

 ing the luminosities of colored discs.J 



Another method very commonly used is to continually vary 

 the luminosity of one of the surfaces so that it will be alternately 

 darker and lighter than the other, the amounts of the variation 

 being gradually diminished. By this means the observer can 



* Grundziige der Physiologishen Optik von Hermann Aubert. Page 488. 1876. 

 ■f-Handbuch der Physiologishen Optik. Second edition, page 440. 

 J American Journal of Science, 1878. 



Tkansactions N. Y. Acad. Sci., Vol. XVI., Sig. 13, April 24, 1897. 



