COLOR PHOTOGRAPHY. 183 



should exceed 3. In Chapter XI it is, however, shown that there are 

 other processes in play which would cause a slight color displacement 

 in the prism experiment with increasing intensity of light. The second 

 plate had, indeed, a stronger exposure than the first. Besides this, a 

 small absolute error in determining the displacement would cause a 

 larger one in the computation of the refractive index. The observations 

 make no claim to great accuracy. They were not originally intended 

 for the measurement of n 2 , but only to show the approximate amount 

 of the color displacement. 



The magnitude of the displacement rendered it probable that it could 

 be observed in air without the use of a prism by simply changing the 

 angle of incidence. Indeed, in the case of the second plate, a shifting 

 of the middle of the bright strip in the sodium flame, amounting to 0.36 

 millimeters, was observed on changing the angle of incidence from 0° 

 to 45°, from which it follows that/i = 0.98. 



But/i may be calculated from n v and n 2 by equation (1), if one substi- 

 tutes for n 2 the the above-mentioned value 3.1. In this way/i is found 

 to be 0.97, which agrees with the value observed within the limits of 

 experimental error. 



The possibility of the recognition of a color change with a wave- 

 length relation 0.98 permits the determination of the limits of the safe 

 application of the prism experiment. The question arises, How great 

 the index of refraction of a film may be and still permit the recognition 

 of interference colors in it by the use of a prism ? If one com pares the 

 color at direct incidence in air with that at 45° in the prism, we have 

 substituting n x = 1.75 and / p = 0.98 in equation (2) : n z = 6.2. If the 

 comparison is restricted to 45° incidence in both air and prism, it 

 follows with/p! = 0.98 by equation (3) : % = 5.2. 



So far as I know there have been no greater indices of refraction 

 than this observed for the D line. That of molecular silver is, accord- 

 ing to Wernicke, 1 based on the computation of Drude, 2 equal to 4. 

 Checking his calculation by the molecular refraction from the known 

 refraction equivalents of a haloid and a haloid compound of silver, a 

 value less than 3 is obtained. 



Thus it is possible by the prism experiment, for example, to test the 

 assumption recently made by Wernicke, that the colors of silver 

 observed by Carey Lea are only interference phenomena of molecular 

 silver. With n 2 = 4 and n x = 1.75, / p becomes equal to 0.95 ; and thus a 

 silver plate appearing in air golden yellow (A. = 5S9piju) should in the 

 prism look distinctly greenish yellow (A, = 560/i/<). If such a color 

 change is not to be observed, then a case of body colors is at hand, and 

 Carey Lea was right is assuming particular modifications of silver. 



Still more certainly can one distinguish between interference and 

 body colors in any chlorine compounds intermediate between silver 



1 Wernicke, Annalen der Physik unci Ckemie, 52 : page 527, 1894. 



2 Drude, ibid., 51: page 98, 1894. 



