MAGNUSSON — ANOMALOUS DISPERSION OF CTANIN. 281 



placement between the two sets of fringes varies when the wave 

 length of the light used is changed. The light passing through 

 the slit, D, was changed systematically for the successive ex- 

 posures, so that in the series of photographs shown on plates 21 

 and 22 a complete record of the retardation throughout the spec- 

 trum is secured. 



The film (No. I), used in preparing the series shown in fig- 

 ure 1 of plate 21, was very thin (about .00012 mm.) and per- 

 fectly transparent throughout the absorption band. The white 

 mark on the sides of the fringe systems was made later to show 

 which pair of fringes belong together. The number given be- 

 low each exposure is the wave length in fi ft units of the light 

 used in each case. 



The series given in figure 2 of plate 21 shows the retardation 

 produced by a somewhat thicker film (about .00017 mm.) not 

 quite as transparent in the region of the absorption band as the 

 first. The third film (No. Ill), used in the preparation of 

 the first figure of plate 22, was much thicker (about .00071 

 mm.) and quite opaque for yellow or yellowish green light. 

 Consequently, no fringes were formed in this part of the spec- 

 trum by the light passing through the cyanin film. 



To continue further into the ultra-violet a Rowland grating 

 was substituted for the spectroscope. Using film No. II, a new 

 series was taken on the blue, violet, and ultra-violet, the results 

 being shown in the second figure of plate 22. Repeating th& 

 same series and using film No. Ill, the photograph shown in 

 figure 3 of the same plate was obtained. 



In these photographs it is to be noticed that beyond 

 A = 345.6 no fringes appear. This is due to the absorption, 

 of shorter waves by the glass. Further, below A = 374.2, the 

 intensity of the fringes in both halves of each exposure is prac- 

 tically the same, showing that the cyanin film in this part of the 

 spectrum was perfectly transparent. Between these lines, 

 X = 374.2 and A = 345.6 that part of the light which had the- 

 cyanin film in its path is much fainter for the larger waves and 

 disappears altogether before reaching the absorption band in 

 the glass. This corroborates previous observation and proves 

 3 



