MAGNUSSON — ANOMALOUS DISPERSION OF CYANIN. 271 



ble. Some line of known wave length was chosen and by nar- 

 rowing the slit the spectrum was gradually cut down until only 

 light in the immediate vicinity of the line remained. In those 

 parts of the spectrum where no conspicuous and easily recog- 

 nized line existed a different method was employed. Remov- 

 ing the prism and placing a small mirror in its place, the light 

 was thrown on a Brashear reflection grating spectrometer (See 

 H, Fig. 3). This instrument had previously been graduated 

 and the readings for the principal Fraunhofer lines been taken 

 in its fixed position, and, hence, the wave length of any light 

 used could be determined. 



The angles of the cyanin prisms were measured by the Geneva 

 Society Spectrometer. The prism was placed with its refract- 

 ing edge vertical and at the center of the circular table. Par- 

 allel rays (a b, Fig. 4) coming from the collimator are reflected 

 from the glass plate in the direction b d and from the surface of 



Fig. 4. 



the prism along b c. The angle e between the reflected rays is 

 twice the angle of the prism. For prisms of small angle both 

 images of the slit were in the field of the telescope and the dis- 

 tance between them was measured by a filar micrometer. These 

 readings were then reduced to the circular scale. For larger 

 angles the readings were taken on the circular scale direct. 



If A represents the angle of the prism and d the deviation, 

 both in circular measure, the index of refraction is found by 

 the usual formula: 



sin Yo, (A + d) 

 M = ainy 2 A (30). 



