322 Mr. W. Hibbert on the Gladstone 



Hitherto no light has been thrown on these anomalies, but 

 this present paper attempts to explain the exceptions in the 

 ease of the Gladstone expression, and, incidentally, introduces 

 a physical magnitude which is conveniently described as the 

 " actual volume occupied by one gramme of the substance/' 



The suggestion of the present argument arose out of a 

 number of experiments carried out by Dr. Gladstone and 

 myself during the last two years. We have examined a 

 great variety of salt-solutions, whose strength varied from 

 saturation down to about 5 per cent. — a dilution limit imposed 

 by experimental error when the hollow-prism method is 

 adopted for obtaining the refractive index. - 



Our main conclusion, to the effect that the value of — -. — 



a 



for the salt is practically constant, has received special con- 

 firmation by the publication of a paper by Kohlrausch and 

 Hallwachs *, in which they extend the work on about 10 salts 

 to extremely dilute solutions. They used an interference 

 method for determining yu,, and a special apparatus for getting 

 d with the necessary accuracy, so that the experimental error 

 is not larger than usual, even when the amount of salt present 

 is not greater than 0'05 per cent, of the solution. 



An examination of this body of evidence proved to me that 

 the constancy of p — l/d was much greater for an ordinary 

 dissolved salt than for a simple liquid whose density is 

 changed by heat. For example : — 



When heptane is heated from 12° to 88° its density changes 

 by about 10 per cent., and the value of fi — 1/d by 0*55 per 

 cent. 



Sodium chloride, diluted from 13 per cent, down to 0'03 

 per cent, (a concentration change of 400 to 1), has its value 

 for fi, — l/d changed by 0'22 per cent. only. 



In speculating on this far greater constancy in the case of 

 a dissolved salt, it seemed not unreasonable to regard the 

 solution in the light of Van't HorFs theory, attributing a 

 pseudo-gaseous condition to the dissolved salt. This at once 

 suggested other comparisons, and led to the following summary 

 statement of all experimental results on the subject : — 



1. The Gladstone expression is true in the case of gases 

 whose density is changed by varying pressure f» 



* Wied. Ann. 1894, liii. p. 1. 



+ Biot and Arago, Mem. d. VInst. vii. p. 301 (1806), and Cliappuis and 

 Riviere, Compt. Rend. ciii. Biot and Arago express their results 

 (V 2 — l)y = const. But ft 2 is only slightly greater than unity, and there- 

 fore {\x — l)v must also be constant. 



