GASEOUS REFRACTIVE INDICES 281 



It was evident that nothing could be made out of such scanty 

 material. 



In 1861 Le Roux attempted to enlarge the available data by 

 attacking iodine, sulphur, phosphorus, arsenic, and mercury 

 by heating them in a hollow glass prism ; but as his figures 

 have not been confirmed by subsequent investigation, they need 

 not be given. 



In 1874 M. Mascart added bromine to the list, finding for it 

 the high value of 1,125, an< 3 in 1877 M. Hurion published two 

 values for the index of iodine, finding for the red rays a re- 

 fractivity of 2,050 and for the violet 1,920, numbers which derive 

 an additional interest from the fact that they showed, for the 

 first time, a case of anomalous dispersion in a gas. 



Thus at the close of the nineteenth century we were in 

 possession of the gaseous refractive indices of only six elements, 

 hydrogen, nitrogen, oxygen, chlorine, bromine, and iodine. The 

 first three showed little or no connection, while of the last three 

 no more definite statement could be made than that the re- 

 fractivity increased concurrently with the atomic weight. 



With the commencement of the new century a considerable 

 advance was made. In 1901 Ramsay and Travers published 

 the results of their investigations on the five inert gases of the 

 atmosphere. These elements — helium, neon, argon, krypton, 

 and xenon — are of peculiar importance in the study of optical 

 constants. They are all gases at the ordinary temperature, and 

 they are all members of the same chemical group, of zero 

 valency. In all, too, the molecule is monatomic. It was to 

 be hoped that, if any discoverable relation existed between the 

 refractivities of the elements, it would show itself in this 

 group. 



The result showed that this expectation was not unfounded. 

 The refractivities of these five gases proved to be almost exactly 

 in the ratios of 1, 2, 8, 12, 20. 



Since the original measurements were made four out of the 

 five indices have been redetermined with larger quantities of 

 gas and with better methods of purification, and the approxi- 

 mation to simple integral ratios is now even closer, whether 

 we consider the index for infinite waves or that for the D line, 

 as the table below shows. 



The question immediately suggested itself whether any other 

 group of elements showed similar relations ; and, fortunately, 



