PHYSICS. 239 



ated than the red. Using a hollow prism of glass, placed 

 inside a hot-air bath, the author has successfully measured 

 the refractive indices for the blue and red rays, and finds 

 them as follows : for the blue 1.019, and for the red 1.0205. 

 This gives 0.06 for the negative dispersion of iodine vapor, 

 which is very near the positive dispersion of flint-glass. As- 

 suming that the refractive power of a substance is indepen- 

 dent of its physical state, the calculated indices for solid 

 iodine would be for the red 1.89, and for the blue 1.83, the 

 values obtained from measurements of the refractive power 

 of a solution of iodine in carbon disulphide being 2.07 and 

 1.98 respectively. 



Thompson has described a rainbow phenomenon, seen 

 chiefly in Switzerland, in which radial streaks of light de- 

 void of color are observed within the primary and without 

 the secondary bow. He explains it by supposing that the 

 wedge-shaped radial streaks are beams of sunlight, which be- 

 come visible by diffuse reflection from particles of matter in 

 their path, just as the apparently divergent beams of sun- 

 rise or sunset become visible. Being practically parallel, 

 they appear to converge in the point exactly opposite to 

 the sun by perspective, just as the parallel beams of the sun 

 appear divergent. Since the rainbow has for its centre the 

 point opposite to the sun, such beams must have positions 

 radial with respect to the bow. They have never been ob- 

 served crossing the dark span between the primary and sec- 

 ondary bows. The phenomenon is a frequent one in the 

 Rocky Mountains. 



Rosenstiehl has made use of rotating disks for the pur- 

 pose of studying the phenomena of the sensations produced 

 by colored light, using a method employed so successfully 

 already by Rood. With regard to the chromatic circles of 

 Chevreul, he has proved that what Chevreul calls equidis- 

 tance between colors is really the result of a mixture of col- 

 or sensations according to an arithmetical progression. It 

 has not been possible to recognize the relation between the 

 tones of a scale when the color is modified by white, in 

 Chevreul's circles. For most of the scales each tone has 

 another complementary to it, and there is no common meas- 

 ure between them. 



Lommel has communicated to the Physical Society of Er- 



