I 



Dispersion and Selective Ahsorption of Fuclisin. 219 



there is a shift of the bands due to the Literal displacement 

 of the beam of li^ht, so it is usual to use a double compen- 

 sator, one piece in each path, thus overcoming the difficulty 

 by moving both pieces together and giving an equal lateral 

 displacement to each beam. The compensator that was used 

 could be turned through 40 degrees, producing a retardation 

 of three or four bands before any effect due to this lateral 

 displacement could be observed, and as the greatest shift that 

 had to be measured was less than a band, it was decided to 

 use a single compensator for the sake of convenience. 



Using the above formula the values of i that were observed 

 for one, two, and three bands displacement were found to be 

 consistent. Assuming fju to be constant and equal to 1*58, 

 the value of t was found from the formula, and then by sub- 

 stituting back this value the various values of n corresponding 

 to the different values of i were found, and a curve of n\ and 

 i plotted. From this the fraction of a band displacement 

 corresponding to any value of i for any part of the spectrum 

 could be determined. There would be a slight error due to 

 the variation of the index of refraction of the mica for light 

 of different wave-lengths^ but as this was less than one per 

 cent, from one end of the spectrum to the other, the com- 

 pensator could be calibrated for one part of the spectrum, 

 and a linear correction made in any other part. As a matter 

 of fact, it was calibrated in the red end of the spectrum, and 

 as the index of refraction of fuchsin is nearly unity at the 

 violet end of the spectrum, and as (//< — 1) is the quantity that 

 is measured it does not matter very much whether the cor- 

 rection is made or not, considering that there are much 

 graver errors which are unavoidable. 



Knowing nX the retardation produced by the fuchsin in 

 light of any wave-length, we may determine yu the index of 

 refraction of fuchsin for light of that wave-length by the 

 formula 



nX = {iJi — l)t^ 



provided we know t the thickness of the fuchsin-film. This 

 w^as determined by a method which wil] be described later. 

 The measurements were made upon a film whose thickness 

 was 580 /x//., with the exception of those values falling between 

 X = 460/A/A and X = 590 yLt/x, which were made upon a film 

 192 yLt/^ thick. 



The results are plotted in fig. 1, PL X., and in the following 

 table they are compared with the values obtained by Pfliiger^ 



* PhiL Mag. i. p. 43 (1901). 



