36 THE TRUE VALUE OE a OE VAN DEK WAALS' 



( ƒ ) . a = 3 ^ 2/3 T ( P N^ Mj{d — D) {T r — TfK From Verschaf- 

 felt's surface tension measurements by capillary rise 

 method at — 24.3°: s = 9.21 dynes; ^=1.005; D = 



0.047; T c — T=55.6 a = 3.55>(10 12 



Mean of Verschaffelt and Eötvös computation 4.08 X 10' 2 . 

 (^r). « = V*d, à being the constant of Ramsay and 

 Young's equation expressing variations of 

 pressure with changing temperature while 

 volume is constant. dPjdT = 1.72 at- 

 mospheres (Amagat). d c = 0.4G4 a = 4.11 )< I0 12 



These results are very consistent with the exception of Eötvös' 

 and Verschaffelt's determinations of the surface tension. I think in 

 general that Eötvös' method of determining the surface tension gives 

 a result which is higher than any other method. On the other hand 

 the capillary rise method often yields too low a result, so a 

 determined from Verschaffelt's observations is too low and that 

 from Eötvös is somewhat high. The mean of all these is 4.23 X 10" 

 but I believe the actual value may be a little higher, say 4.29 X 10 12 . 

 This is exactly the value which van der Waals x ) assigned to it 

 in his work on the Continuity of Liquids and Gases. He there 

 gives the value of .0.0874 for a. This computed for dynes and 

 gram mol quantities is 4.29 X 10 12 . If we now compare this value 

 with that computed from the molecular weight and number of 

 valences, Ave find an exact agreement if C0 2 is given 5 valences to the 

 molecule; but if we take the number of valences usually assigned, 

 namely 4 for the carbon and two each for each oxygen atom ma- 

 king S in all, the value of a would be 5.86 X 10 12 , which is 

 too high. Carbon monoxide, oxygen gas and carbon dioxide all 

 appear to be exceptional. For these gases to agree with the general 

 law of the dependence of cohesion on the valence electrons it is 

 necessary that fewer valences be in the molecule than is generally 

 believed. I think it probable that C0 2 and N 2 really only have 

 5 valence electrons in their molecules and that in C0 2 and CO the 

 carbon has lost three of the valence electrons it usually has. If 

 there were 4 valence electrons in the molecule d would be com- 

 puted as 3.69 X 10 12 and this is obviously too low. C0 2 is there- 

 fore an exception to the law I am attempting to establish, unless 

 it can be shown that it really has 5 valence electrons. The 

 value van Laar assigns to a is 3.50 X 10 12 which again is seen 

 to be far too low, 20°/ lower than 4.29 which is the true value. 



x ) Vax der Waals: La Continuité des Etats Gazeux et Liquides. Paris, 1894. p. 118, 



