320 iMESSRS. C. CUTHBERTSON AND E. B. K. PRIDEAUX 



been published,* but the following details are added in order that the value of the 

 experiment may be criticised. 



Table I. exhibits the results obtained in five experiments, performed on two 



occasions. 



TABLE I. 



These figures require some explanation and comment. The refractivity of fluorine 

 is certainly much lower than those of oxygen and nitrogen, while that of all other 

 elements (except hydrogen, helium, and neon, the presence of which need not be 

 suspected) and, tt fortiori, of all compounds,! is higher. Consequently, when air is 

 displaced from a prism by fluorine, the lower the index observed the nearer do we 

 approach to that of the latter. 



The first experiment recorded in Table I. may be discarded. During its progress 

 an unaccountable change of zeros took place, which makes it doubtful whether the 

 reading given above, or a lower one ('215), should be accepted. The balance of 

 probability is in favour of the higher value. 



When the prism had been swept out with dry air the second experiment was 

 performed, and gave two trustworthy readings of 228 and 22G for the refractivity of 

 the contents of the prism. 



After making some improvements in the stability of the apparatus and substituting 

 tubes of finer bore (about '2 millims.) for the old leads, a second attempt was made on 

 the 20th January. 



On this occasion a very trustworthy experiment gave a refractivity of 243, or 16 

 points worse than that of the second experiment of the 13th January; while a second 

 trial, made after recovering the zero by sweeping out the prism with air, gave an 

 almost identical result, 241. 



It was then suspected that the electrolytic tube had developed a leak, and a new 

 one was substituted. 



The fifth experiment, performed with this apparatus, at once gave a refractivity of 

 227, which is nearly identical with the best experiment of the 13th January. 



* 'Phil. Trans.,' A, vol. 204, 1905, p. 323. 



t A molecule of HF probably retards light less than a molecule of fluorine, but since the molecule of 

 this vapour, under normal conditions, is at least as complex as HoF-j, its presence in an atmosphere of 

 fluorine would probably raise the refractivity of the mixture. 



