138 Dr. C. H. Lees -on the Viscosities of 



viscosity would then diminish as the size of each element of 

 the pattern of fig. 3 or 4 was decreased, and we should 

 have some approach to the known fact that the viscosity in 

 general decreases as we go from liquids with larger to liquids 

 with smaller molecules. If we may take the molecular 

 volumes of the liquids of the preceding table as indications 

 of the magnitude of the component parts of the pattern of 

 figs. 3 and 4, the tables show no evidence of this decrease of 

 viscosity for mixtures of liquids with small molecular volumes, 

 in fact one of the most marked cases of small viscosity is that 

 of turpentine and toluene, both with comparatively large 

 molecular volumes ; and in neither case would it be legitimate 

 to assume that the molecules had dissociated. 



Those cases in which the observed viscosities are higher 

 than the normal can be more easily accounted for by some 

 kind of loose chemical combination or " association " of the 

 two constituents. The case of acetic ether and nitrobenzine 

 may be instanced. Here there is a contraction of '8 per cent, 

 for a 50 per cent, by volume mixture, which shows that some 

 action has taken place, and the viscosity is greater than in a 

 normal mixture. 



The experiments of Arrhenius *, which enabled him to 

 establish the logarithmic law as an empirical law, were made 

 on aqueous solutions of alcohols and other substances of 

 concentrations not exceeding about 10 per cent., and Reyer f 

 found that the law held for aqueous solutions of acids. 

 From what precedes it will be seen that in these cases, if the 

 observed value of the viscosity of a 10 per cent, mixture is 

 used to calculate by means of the formula that of a 5 per 

 cent, mixture, and this value compared with the one found 

 by experiment, the range over which the law is tested is too 

 limited to furnish a sufficient test of its generality, and the 

 deviations from it found in the above tables would not make 

 themselves evident. 



For salt-solutions in which the concentration does not 

 exceed 10 per cent., Arrhenius {, Reyer §, Wagner ||, Lau- 

 enstein % and Kanitz ** have found that the logarithmic 

 formula applies. Bruckner f f has found that the formula 

 which makes the viscosity a linear function of the concen- 

 tration applies, and Mutzel J J uses a formula which introduces 

 an additional unknown constant for each mixture. 



* Arrhenius, Zeitsch.fiirphijs. Chem. i. p. 285 (1887). 



f Reyer, ibid. ii. p. 744 (1888). J Loc. fit. § Loc. cit. 



|| Zeitsch. fihrphy*. Chem. v. p. 31 (1890). 



IF Ibid. ix. p. 417 (1892). ** Ibid. xxii. p. 336 (1897), 



tf Ann. der. Phys. xlii. p. 287 (1891). \1 Ibid, xliii, p. 15 (1891). 



