EQUATION AND THE NATURE OE COHESION. 65 



„polar" substances in the way suggested by G. N. Lewis, in dis- 

 tinction from such compounds as methane or the hydrocarbons. It 

 is highly significant that all of these exceptional substances contain 

 oxygen. It is true that in the case of the esters, in which we also 

 have one carbon atom in union- with oxygen, these substances do 

 not appear to be exceptional. That is: in the case of the esters we 

 have about the correct value of a if we consider each carbon atom 

 to be tetravalent and the oxygen divalent. But even in these cases 

 if we examine column 3 in which a is determined by the formula 

 a = 6.5 P c Vc we see that in general the values in column 3 are 

 somewhat lower than those of column 8 which have been computed 

 on the basis of the larger number of valences. The difference is 

 not very marked, and the other methods of computing a in the 

 esters bring the average value up to the expected value. I formerly 

 thought that the esters had more valences than they should have 

 and that this was due to the oxygen being in one atom tetravalent. 

 This conclusion was based ou the earlier computations of the value 

 of a and it is not borne out by this work if we take the mean 

 value of a as determined in all the methods. Since the mean values 

 of a agree with the expectation on the basis that all the carbons 

 are tetravalent and the oxygen bivalent we cannot consider them 

 exceptional. At the same time I do not feel sure that when we 

 have a more accurate method of determining a from the latent heat 

 of vaporization, it will not be found that there are two valences 

 less in the molecule than is indicated in table 2. With this digres- 

 sion we may now return to 2 , CO and C0 2 . It may be stated 

 in the first place that there is no theoretical objection to these 

 compounds having fewer valences than chemists have usually ascribed 

 to them. There is no other way than this, so far as I know, of 

 accurately determining how many valences there are in their mole- 

 cules. So they may have the number of valences, that is the number 

 of ^valence electrons, which will make them no longer exceptional. 

 There is no evidence against this assumption. In 2 the oxygen 

 may indeed be monovalent; in CO both the carbon and oxygen 

 may be monovalent; there is no evidence that they are not. In fact 

 in ozone, in peroxides and in oxygen we may be dealing with mono- 

 valent oxygen. As I have suggested elsewhere it may be for this 

 very reason that oxygen in these substances, or at least in some 

 of them, has different magnetic properties than oxygen in other 

 compounds. It is well known that elemental oxygen is strongly 

 magnetic, whereas compounds containing oxygen, with the possible 

 exception of CO, are not magnetic but diamagnetic. 



Verhand. der Kon. Akad. v. Wetensch. (I* Sectie) Dl. XII N°. 4. D 5 



