the Refractive Index and Dispersion of Glass. 117 



The ordinary soda-flame produced by holding a piece of 

 common salt in a Bunsen flame will be found to give 

 sufficient light to enable accurate observations to be taken 

 on the sodium line. If more light should be required for 

 any purpose, an apparatus designed to produce a very 

 intense source of sodium light and described by H. E. 

 Armstrong in the Proc. Roy. Soc. vol. lxxxi. (1908) is 

 recommended. It will be found convenient to provide two 

 sources of light, one to be used for obtaining the equality of 

 the indices, and the other to be used in conjunction with the 

 prism-condenser of the refractometer, which is of course 

 swung out of action whilst the first source is being used, to 

 determine the refractive index of the liquid as soon as the 

 equality has been obtained. It is a matter for regret that it 

 has not hitherto been found possible to determine the 

 refractive indices of the glass for the other standard refracto- 

 metrie lines C, F, and Gr' to the same satisfactory order of 

 accuracy that is possible with the D line. The intensity 

 of the Gr' line in the hydrogen spectrum given by the usual 

 H-tubes used in refractometrv is uot great, and the line is by 

 no means an ideal one to use at the best of times. Moreover, 

 as the curve shown in fig. 3 indicates, the immersion fluid 

 itself transmits only a negligible percentage of light of this 

 wave-length. This line therefore must be regarded as quite 

 useless for our present purposes. The position as regards 

 the G and F lines is not quite so unsatisfactory, but even in 

 this case the intensity of the light given by the hydrogen 

 tube is relatively feeble and is not calculated to give results 

 with any certainty. It should be remembered that the 

 adjustment for equality of indices is really a photometric 

 one, and consequently far more light is required than in the 

 case of the usual refractometer readings where the cross- 

 lines in the eyepiece of the refractometer telescope are 

 brought into coincidence with the edge of the line under 

 examination, and the setting is rather of the nature of a 

 geometrical or positional one. The separation of the C and 

 F lines is effected by means of coloured filters, the red filter 

 employed transmitting about 89 per cent, of the C line, and 

 the blue filter about 56 per cent, of the F line. The most 

 satisfactory sources to be used for measurements on the 

 partial dispersions would appear to be the mercury-vapour 

 lamp for a green line \ = 546//<yu,, and the recently de- 

 scribed cadmium arc (H. J. S. Sand, Phvs. Soc. Proc. 

 xxviii., 1916) for the red cadmium line X^bll^a/x. The 

 Phil. Maa. S. 6, Vol. 32. No. 190. Oct. 1916. 2 F 



