GASEOUS REFRACTIVE INDICES 277 



caused by the free gaseous atom. The forces which compel atoms 

 of different elements to combine and to assume the liquid or 

 solid states, whatever they may be, are evidently very powerful, 

 and it is unreasonable to expect that causes which can modify 

 other attributes of matter in ways and to an extent which we are 

 unable to predict should have no effect, or always the same 

 effect, on its power to retard light. 



It is then to the study of gaseous refractive indices that it is 

 necessary to turn in the hope of obtaining really accurate 

 information with regard to the optical properties of matter. 

 In this field the prospects of success a priori are all favourable. 

 We know that a gas consists of discrete particles, moving about 

 with velocities which are small relatively to the velocity of light, 

 and separated by distances which are great in relation to their 

 size. The time spent in traversing the free path is long com- 

 pared with that spent in collisions. We may, therefore, expect 

 that the intrinsic power of each atom or molecule to retard 

 light will be freely exerted without modification by the forces of 

 neighbouring atoms, and that the measurement will represent 

 some fundamental attribute of matter in an intelligible form. 

 At the same time the simplicity of the gas laws greatly 

 facilitates the reduction of the observed index to standard 

 conditions. 



The causes which have retarded the expansion of our 

 knowledge in this direction have been the lack of materials 

 and the experimental difficulties. Very few elements and 

 comparatively few compounds are gaseous at ordinary tempera- 

 tures, and these were quickly disposed of by the early pioneers, 

 Biot and Dulong. It was not till 1861 that Le Roux made a 

 gallant attempt to determine the indices of iodine, mercury, 

 sulphur, arsenic, and phosphorus at high temperatures. He 

 was succeeded by Mascart, who in 1878 greatly enlarged 

 our data by observations on a large number of vapours of 

 liquids boiling below ioo°. 



Other researches of the same nature followed, and more 

 recently means have been found to overcome the experimental 

 difficulties and to measure indices at temperatures as high 

 as 8oo° C. with what is believed to be a fair amount of 

 accuracy. 



Before turning to an examination of the results obtained, 

 it will be convenient to give a very brief sketch of the methods 



