in the Absorption-spectrum of Didymium. 179 



The crystal (Plate II. fig. 11) employed in these experiments 

 exhibited a strongly-marked tabular habitus, owing to the growth 

 of the surfaces P. It was cemented between two thin micro- 

 scopic glasses with Canada balsam, and was strongly coloured 

 and perfectly clear and transparent. 



The plane of polarization of the ordinary and extraordinary 

 rays which pass perpendicular to the surface P makes with the 

 klinodiagonal and with the orthodiagonal an angle of 20°. 

 Which is the plane of polarization of the ordinary, and which of 

 the extraordinary ray, cannot be settled, as the positions of the 

 optic axes have not been determined. In order to distinguish 

 the spectra of the two rays, a NicoPs prism was placed in front 

 of the crystal, which was so arranged that the light was trans- 

 mitted in a direction perpendicular to the surface P; and the 

 crystal so turned round that the plane of polarization of the 

 Nicol coincided with one or other of the two planes. The posi- 

 tion of the crystal when the plane of polarization of the Nicol 

 made an angle of nearly 20° with the orthodiagonal I shall call 

 the orthodiagonal position ; the other I call the klinodiagonal 

 position. 



In the investigation of the spectra of different salts of didy- 

 mium, a difficulty occurs, in the fact that absorption-spectra, 

 otherwise perfectly similar, assume a different appearance accord- 

 ing to the degree of intensity, the breadth of the absorption- 

 bands varying with the thickness and with the proportion of salt 

 contained in the absorbing medium. Such comparative obser- 

 vations, therefore, can only be of value when, in all the compari- 

 sons, the light has been acted upon by the same quantities of 

 the absorbent body in passing through the absorbing medium. 

 This condition is fulfilled when the amount of didymium con- 

 tained in the absorbing medium is inversely proportional to the 

 length of the column through which the light passes. In the 

 comparison of the crystallized and dissolved sulphate with the 

 solutions of other didymium salts this condition was most carefully 

 taken into consideration. 



Let the thickness of the layer of crystal through which the 

 light passes be /, the thickness of the column of solution l { ; let 

 the amount of didymium oxide contained in the unit of volume 

 of the crystal be d, and that in the unit of volume of the solution 

 be d x ; then in all the experiments d x was so chosen that 



d^i = dig 



The crystals used in the experiments had a specific gravity of 

 2*7153 at 8° C. as a mean of two well-agreeing determinations. 

 1 cubic centim. of the crystalline mass therefore contains 1'2563 

 grm. of didymium oxide. The layer of crvstal lying between 



N2 



