MAGNUSSON — ANOMALOUS DISPERSION OP CYANIN. 269 



a deep plum color and of the consistency of a rather thick syrup. 

 The crystals when heated to about 80° C change color from a 

 bright green to a golden brown. After fusing the cyanin 

 rapidly evaporates and leaves a solid residue. 



To secure good prisms the heating of the cyanin must be done 

 rapidly, the prism formed, pressed, and cooled quicxly and at 

 the right temperature. If too cool, the fused mass can not be 

 pressed into the required thinness ; and if too hot, a number of 

 small pin holes will appear, due to the gas bubbles formed. The 

 pressure applied is also a factor tjiat must be carefully adjusted 

 if good results are to follow. This becomes very important 

 when thin prisms are desired. For work in the transparent 

 parts of the spectrum, prisms of 10' to 20' angle can be used, 

 but in and near the absorption band they must be exceedingly 

 thin to let any light through. 



Cyanin being found to be the only dye from which wholly 

 satisfactory prisms could be formed it was decided to limit the 

 work to an examination of the dispersion of this single sub- 

 stance. 



Four distinct methods have been employed, namely : 



(a) Direct spectrometer readings in the visible spectrum, 

 the slit of the instrument being illuminated with monochro- 

 matic light. 



(b) Photographic records of the deviation of a system of 

 monochromatic rays from a Rowland concave grating illumin- 

 nated by an iron arc. 



(c) A qualitative method based on Newton's principle of 

 crossed prisms. 



(d) Photographic records of the displacement of the fringes 

 in the Michelson interferometer produced by thin films of 

 cyanin. 



(a) The visible part of the spectrum. The first method was 

 employed by Wood: 1 A plan of the apparatus is shown in 

 figure 3. Sunlight reflected from a heliostat (A) is concen- 



*Phil. Mag. XLVI (1898), p. 380. 



