Isometrics relative to Viscosity. 95 



Pressure = atru. 500 atm. 1000 atm. 



Increment of pressure = 67 atm. 256 atm. 880 atm. 



and so on. Thus the relative inefficiency of pressure as com- 

 pared with temperature is apparent, though to make the com- 

 parison just, both agencies should be taken per unit of volume 

 increment. Cf. § 11. 



11. Digression, Logarithmic isothermals. — Believing that 

 the error due to slipping increases with pressure, i. e. in pro- 

 portion as the charge becomes more solid, and noting the 

 tendency (§ 7) of isothermals for high viscosity to slope up- 

 ward, I thought it worth while to compute the isothermals on 

 the supposition that log t] = a' + b'p, as an extreme case. Nec- 

 essarily, marked violence is thus done to the observations, and 

 b' obtained from high pressures must be smaller than V from 

 low pressures. Preferring the latter, I found, for instance 



?7 /l 9 = -69 5-9 4-7 2-9 -64 



fr'XlO 5 = 81 73 93 73 74 



As before a dependence of b' on ^ does not appear and b' = 

 •00078 may be taken as the mean value. 



The interest which attaches to this case is its bearing on the 

 isometrics, which now appear as straight lines. For if 



G = Vophu, lo g C + B6 = b'p, and (dp/d0) = B/b' = 210 



In other words 210 atm. would annul the decrement of viscos- 

 ity produced by a rise of temperature of 1° C, at all temper- 

 atures and pressures. 



Seeing that in an elegant research of Ramsay and Young,* 

 and in high pressure workf of my own, the volume isometrics 

 of liquids appear as straight lines, the present considerations 

 may possibly claim more than passing comment. 



The immediate object of the present paragraph, however, is 

 to give warrant to the statement, that in high pressure phe- 

 nomena at least 200 atm. must be allowed per degree Centi- 

 grade, in order that there may be no change of viscosity. 



12. Maxwell's theory.— rlf for the sake of definiteness, vis- 

 cosity {rj) be defined as proportional to the ratio (JY—n)/n, of 

 the number of stable configurations (iV — «), to the number of 

 unstable configurations (n), in a given volume, then the above 

 expressions may easily be translated into the language of 

 Maxwell's theory of viscosity.;}; I shall therefore withhold 

 further remarks here. The conditions are simplified since for 



* Ramsay and Young : Phil. Mag., xxiii, p. 435, 1887; xxiv, p. 196, 1887. 



•j-Barus: Phil. Mag., xxx, p. 338, 1890. 



% This was done in my note in this Journal for September, p. 255. — In the 

 series, atom, molecule, viscous configuration, the last can not be as sharply 

 defined as the other two, and only the former as yet admits of generic classifica- 

 tion (periodic law). Cf Am. Chem. Journ , xiv, pp. 197-201. 



