WOOD. — ANOMALOUS DISPERSION OP SODIUM VAPOR. 383 



to 100 of helium. The shift of wave-length 579 with respect to helium 

 is given by the proportion 5 : 6 = 14 : a:, from which we get x = 168., 

 a value which fell exactly on the curve obtained with the interferometer. 

 The same thing was done in the case of the values obtained by break- 

 ing the helium line up into a Zeeman doublet, 11 fringes for A = 587487, 

 11.5 fringes for A = 587513. The values for the range of wave-lengths 

 comprised between the helium line and Dj were determined by the 

 method of crossed prisms, the measurements being made with the 

 reticulated mesh in the focus of the piece. For light of wave-length 

 midway between Dg and Do the deviation is twice as great as for Dg, 

 consequently we enter this iu the table as 200. 



The Ultra-violet Dispersion. 



A qualitative study of the dispersion in the ultra-violet was first made 

 by the method of crossed prisms. 



A tube of hard Jena glass, 40 cms. long and 1.5 cm. internal diameter, 

 was fitted with end plates of quartz, loaded with sodium, and exhausted. 

 The arrangement of the apparatus is shown in Figure 5. 



Arc 



I 



c*^ 



rtcrlxo*TAt 

 I Slit 



\^B 



Figure 5. 



As the method depends on forming a sharp image of a brightly illu- 

 minated horizontal slit upon the vertical slit of the spectrograph, quartz 

 lenses could not be used, owing to their chromatic aberration. Con- 

 cave silvered mirrors are free from this defect, but have abnormally 

 low reflecting power in the ultra-violet region. Very satisfactory photo- 

 graphs were obtained with them, however, in the preliminary work, 

 though they were replaced with magnalium reflectors later on. The 

 light from the crater of an arc-lamp was focussed on the horizontal 

 slit, and collimated with one of the concave mirrors. It then passed 



