" Law " in Physical Optics. 



333 



This table shows that the difference between the specific 

 refraction of a liquid and its vapour as calculated by the 

 Gladstone expression may be 10 per cent, of the whole ; also 

 that if the value of ft as calculated from the liquid be em- 

 ployed in the modification now proposed, the difference 

 between the liquid and vapour falls to less than 1 per cent. 

 With a more accurate estimate of ft this difference might be 

 still further reduced. The modified expression gives practi- 

 cally a constant for both liquid and vapour. 



One remark ought to be made. The value of ft for carbon 

 disulphide is much greater than that deduced from Ketteler's 

 figures already quoted; a fact which is difficult to explain, 

 though difference of specimen may have something to do 

 with it. 



There is another set of observations by which ft can be 

 tested. Prytz has given data for nine liquids and their 

 vapours ; but as he only took one observation on the liquid 

 his work is only available for calculating ft from the liquid 

 and vapour values. This can be compared with the average 

 values previously obtained. 



Substance. 



/3 calculated from Liquid 

 and Gas. 



Methyl Alcohol 



0-09 



0-116 



0-085 



0-091 



0-0655 



0-093 



0-112 



0-096 



0-16 



Ethylene Chloride 



Ethedene ,, 



Acetone 



Methyl Acetate 



Ethyl Formate 



Propyl Iodide 



Methyl Propionate 







These values are slightly higher than those deduced from 

 two observations on the liquid, but the difference is not great. 



A prima facie case having been shown for attaching some 

 physical significance to the symbol ft, it appeared probable 

 that the values for different substances would show definite 

 relationships between themselves ; or that the value of ft for 

 a given substance would be related to the value of v at the 

 freezing, boiling, or critical temperature. 



Of these the boiling-point is the one for which data are 

 mostly available, and a little examination suggested a close 

 relation between the absolute temperature of the boiling- 

 point and the two volumes v and ft. It may conveniently be 



written 



ftot 





