212 Dr. R. H. Wright on the Photometry of the 



colours. The regular reflexion begins sooner for red than for 

 yellow, green, &c. This fact alone, I believe, can scarcely be 

 a sufficient explanation, because Fresnel f and Hankel f used 

 angles of 79° as a minimum, whilst here the phenomenon 

 can be noticed already at 20°. 



The law for the angle of emission having been found to 

 be independent of the colour, I believe to have proved the 

 truth of von Lommers hypothesis, that the so-called coefficient 

 of diffusion is independent of the wave-length for particles of 

 the said dimensions. Contrary to this the dependence upon 

 colour has been shown by Christiansen % in the case of very 

 small suspended particles. This cannot, however, be used as 

 an argument against my assertion, because the size of the 

 particles is so very different, and the situation altogether 

 another one ; especially as the ratio of the refractive indices 

 is very nearly unity by Christiansen's experiments. 



If we compare the mean of the complementary colours 

 (tables 7 a and 10 a) with the values found without the use 

 of a prism, we find deviations which exceed the accuracy of 

 the measurements very much. In the first case we combine 

 the angles (a) of the nicol in the following way : 



tan 2 a red + tan 2 <x green 



In the second case we read an angle a', of whose relation to 

 the angles for a red , a green , &c, we are in total ignorance, 

 as physiological questions arise here. Hence we see theo- 

 retically that the values of both tables are probably different, 

 and the experiments confirm this also. Thus one is not able 

 to establish a law for white light, after having made observa- 

 tions of the plate at different parts of the spectrum only. 

 And it is also not permissible to draw an inference from 

 Messerschmitt's results (p. 884). They only show that a 

 deviation exists ; but its value is absolutely unknown. 



Part II. 



If the variations between the results and the theory are 

 given, as in the case of other observers, only the law as a whole 

 can be considered. It is, however, a great advantage to 

 compare the single elements of the law with the reality. 

 Hence I append the following tables : 



1. Deviation from the Law of Incidence. — The figures are 



* A . Fresnel, CEuvres completes, i. p. 225. 



T W. Hankel, Pogg. Ann. c. p. 302^ (1857). 



\ C. Christiansen, Wied. Ann. xxiii. & xxiv. 1881 & 1885. 



