BLACKWELL. — DISPERSION IN ELECTRIC DOUBLE REFRACTION. 649 



refraction " varies exactly, or very nearly " inversely as the square-root 

 of the wave-length. The only other observer who noticed dispersion is 

 Schmidt. In 1901 he published observations of the amount of electric 

 double refraction in a number of organic liquids compared with that in 

 carbon bisulphide as standard. His observations were made with blue 

 and with red light obtained with color screens (light of wave-lengths 

 " 490/t/t " and " 680 /x/a "), and with what he called the " blue-violet 

 transition tint (A = 580 fifi), " obtained apparently by rotating the Nicols. 

 He expressed the electric double refraction of the substance under inves- 

 tigation in terms of the effect in carbon bisulphide. It thus does not 

 appear whether his data indicate dispersion in the liquid under obser- 

 vation, or in the carbon bisulphide used for comparison, but merely that 

 in one or other, or both, there is dispersion. The reduction of his 

 observations to absolute value would depend on the determinations of 

 " B " in carbon bisulphide made by Quincke and by Lemoine, neither 

 of whom noted the variation of " B " with variation of wave-length. 



In short, Kerr's measurements on the variation of electric double 

 refraction with variation in wave-length furnish the only suggestion thus 

 far as to the law of dispersion. 



New Kesults. 



In the eight experiments whose results are given below, the tempera- 

 ture remained constant near 11° in five, and near 24° in the other 

 three. The field-strength during each of the experiments was also very 

 nearly constant, and differed little from experiment to experiment. 

 The results are given for a potential gradient of 56,000 volts per centi- 

 meter. The observed double refraction, or separation 8, is proportional 

 to the difference between the ordinary and electrically produced extraor- 

 dinary indices of the liquid, (jt t — p^). This quantity is the separation 

 in centimeters introduced between the mutually perpendicular compo- 

 nents of a wave of plane-polarized light in traversing one centimeter of 

 the substance in question, — in this case carbon bisulphide, — at a 

 certain temperature, in a field of certain strength. The double re- 

 fraction of natural uniaxial crystals is usually similarly expressed as 

 the difference between the two indices. Accordingly, the present obser- 

 vations are expressed in terms of this characteristic difference. Its 

 values are given in Table I, and plotted in Figure 1, in per cent of 

 its value at 24°. at wave-length 5890. This value, the factor reduc- 

 ing the values of Table I to absolute index of refraction, is found in 

 these experiments to be about 7.04 • 10 -7 . 



