62 THE TRUE VALUE OF a OP VAN DER WAALS' 



compute closely similar figures, namely, 18.70, 18.64, 19.20, 19.34, 

 18.87, 19.30, and 18.52. Or if we take an aliphatic compound, 

 one of the best investigated, «-pentane, from the gravitation and 

 valences we compute 20.52 X 10 1 ' 2 dynes, and in these other ways 

 of determining a we have: 20.68, 20.90, 21.05, 20.69, 20.88. 

 This close agreement is not chance. Tn the case of an ester, a com- 

 pound of still another type, methyl propionate, from the gravitation 

 and valences we compute the value of a of 21.46 X l^ 12 dynes. 

 And in these other ways of direct determination we have: 21.32, 

 20.68, 23.27, 22 42, 21,11. We can pass even to a totally dif- 

 ferent type of compound, an inorganic compound having entirely 

 different chemical elements in it, stannic chloride. From the valence 

 and molecular weight we compute a as 30.40 X 10 12 dynes; and 

 from the direct and indirect measurements we have the values: 

 29.48, 30.07, 30.69, 30.12, 30.59. Does anyone think that it is 

 chance which brings such an agreement to pass? Van Laar reproa- 

 ches me for having taken chlorine to be trivalent, but the trivalence 

 of chlorine is better established than any other valence number and 

 this valence alone will agree with the computation in all chlorine 

 compounds of which the critical data are accurately known. The 

 independent investigation of the valence of chlorine by a totally 

 different method by Pascal shows chlorine to be trivalent. The 

 refractivity of chlorine compounds also shows that chlorine is 

 more than univalent. So in choosing three valences for the number 

 for chlorine 1 have chosen the best substantiated valence number. 

 Even in so complex a substance as diisoamyl with 62 valences 

 in its molecule and a molecular weight of 142 we compute for 

 a the value of 50.20 X I0 1i ; and from the surface tension it is 

 found to be 52.79 X 10 12 . The product of the molecular weight 

 and the number of valences in this substance is 8.804. Now let 

 us turn from this most complex of the substances examined and 

 go straightaway to the lightest and simplest substance, elemental 

 hydrogen. Its critical temperature is 32° Abs, whereas that of diiso- 

 amyl is computed to be 675°. The product of the molecular weight 

 and the number of valences is 4, as contrasted with 8.804. Now 

 if we calculate a of hydrogen from the valence and molecular weight 

 we have 2.98 X 10 11 dynes. From the surface tension of liquid 

 hydrogen we compute its cohesion to be as a minimum 2.54 X l^ 11 

 dynes and from the critical data by the formula a = 6.5 P,./ 7 ,. 2 

 it is found to be 3.19 X 10 11 . Chance does not produce results 

 like this. It is true that here and there there are deviations from 

 the expected values amounting to as much as 8 to 9 °/ lü tne mos t 



