278 BULLETIN OF THE UNIVERSITY OF WISCONSIN. 



The light of an arc lamp, A, was focused on the slit of the spec- 

 trometer, C, and parallel rajs coming from the collimator passed 

 through a water prism, T, and falling on an achromatic lens, F, 

 of long focus (a telescope lens of six feet focus) formed' a 

 sharply denned spectrum, V — E, on the photographic plate, H. 

 In front of the objective, F, was placed a cyanin prism, E, with 

 its refracting edge horizontal and turned down. This prism 

 deviated all the rays upwards before they entered the lens, F, 

 thus elevating the image on the photograph. Since the devia- 

 tion is due to the fact that light is transmitted slower in cyanin 

 than in air, and that the velocity of transmission is a function 

 of the wave length, therefore, the elevation of light of any wave 

 length will be proportional to ^ — 1 where /j. is the index of re- 

 fraction for the given wave length. Consequently, if the spec- 

 trum produced by the water prism is considered to be normal 

 then the crossed prism spectra produced by the cyanin prism 

 as described above, is a curve identical in every respect with 

 the curve of dispersion calculated by the direct spectrometer 

 method and shown graphically in the curve of plate 16. Ery- 

 thro plates were used and the more intense action of the blue 

 and violet was modified by the introduction of an Aurantia 

 color screen or filter, D. In this way fully exposed photographs 

 were obtained of the dispersion curve from the extreme red to 

 the violet. The results are seen in figures 1-6 of plate 20. For 

 figure 1 the lower straight band was made by a direct exposure, 

 while the curved strips above are parts of the same spectrum 

 after having passed through the cyanin prism. The angle of 

 the prism was 17' 1.0", and the time of exposure, by the direct 

 path, was twenty seconds, while through the cyanin it was one 

 hour. The yellow part of the spectrum is absent, the prism 

 absorbing all the energy of this wave length. For figures 2 and 

 3, the same method of procedure was pursued, and these photo- 

 graphs are inserted to show the effect of the Aurantia color 

 screen, D. The angle of the cyanin prism used for these photo- 

 graphs was 17' 3.8". For figure 3 no screen was employed ; for 

 figure 1 the light was passed through a screen of medium thick- 

 ness, while for figure 2 a rather dense filter was used. The haze 



