:',:! 1 



Ml; CLIVE CUTHBKRTSON ON Till: 



TABLE IV. llef'ractive Index of Mercury. Preliminary Experiments. 



An investigation into the cause of these discrepant values led to the discovery that, 

 owing to the position of the thermometers, which were inserted in holes pierced 

 through the electric furnace, the ordinary formula for stem correction does not give 

 the true temperature of the tube, which was a degree or two higher than the 

 calculated value. Since the vapour density of mercury increases very fast between 

 200 C. and 300 C. such an error is fatal. Thus at 230 C. a mistake of 1 C. 

 corresponds to an error of nearly 3 per cent, in the result ; and at 250 C. to 2i per 

 cent. 



It was also found that sufficient allowance had not been made for the lag of the 

 temperature in the tube behind that shown by the thermometer. 



In view of the difficulties involved in eliminating these sources of error, difficulties 

 which would have been increased in dealing with substances whose vapour-pressure 

 at high temperatures has never yet been determined, it was decided to abandon the 

 method for that of the balance. A small glass tube, about 5 millims. in length and 

 5 millim. in internal diameter, closed at one end, was weighed before and after a 

 drop of mercury had been introduced into it, and carefully placed in the refractometer 

 tube, which was then evacuated, sealed, and heated till the whole of the substance 

 had evaporated. The amount of mercury required to produce a convenient number 

 of bands was found to be between '01 and '02 gramme, a quantity which could be 

 weighed without difficulty to one part in five hundred. 



The results obtained by this method showed a most satisfactory agreement. 

 Below are given the details of every experiment performed, except those which were 

 obviously spoiled by the cracking of a tube, or other disaster. 



