328 



MESSES. C. CUTHBERTSON AND E. B. R. PRIDEAUX 



The following table shows the results obtained by this method. With the exception 

 of the second experiment, which was discordant, probably owing to the presence of a 

 slight leak afterwards detected, the coincidence is as close as we can hope to attain, 

 having regard to the difficulties of the inquiry. 



We believe that the refractivity of fluorine for the D line lies in the neighbourhood 

 of 195, most probably within 2 per cent, of that number. 



TABLE III. Refractive Index of Gaseous Fluorine. 



The principal difficulties and sources of error involved in the method are as 

 follows : 



(1) In disconnecting the refractometer tube and connecting with the measuring 



apparatus a few seconds are spent, and a small proportion of the contents of 

 the tube may be lost by diffusion. To minimise this the leads were made 

 about 20 centims. long, of platinum tube of less than 2 millims. bore. 



(2) The volume of fluorine present is measured by contraction during the com- 



bination of the fluorine with the lead filings. The success of the method, 

 therefore, depends absolutely on the assumption that no gaseous compounds 

 are formed, or that if formed (e.g., SiF 4 ) they are allowed for : and, secondly, 

 on the absence of leaks. 



