730 
The Philippine Journal of Science 
1921 
compounds. A discussion of some compounds which may be 
exceptions to this rule is the subject of the present paper. 4 
ODD MOLECULES 
Lewis has called molecules which contain an odd number of 
electrons odd molecules, pointing out that they constitute an ex- 
ceptional class. He has further observed 5 that such a molecule 
“will contain at least one atom with an uneven number of elec- 
trons in the shell. This may be called an odd atom.” Literally 
construed, this means that only an even number of electrons 
can be shared, and CIO 2 , for example, is O^Cl^O, with a va- 
I • 
cancy in the shell of the chlorine atom. There is nothing in 
Lewis’s fundamental postulates, however, against the supposition 
that an odd number of electrons can be shared, and the writer is 
inclined to think that, in certain rather exceptional cases, the 
facts warrant this view rather than Langmuir’s postulate 11. 
Admitting that there is a strong tendency for electrons to be 
shared in even numbers, we also know that there is a strong 
tendency for octets to be completed. Why, then, should odd 
molecules exist? In an odd molecule which of these tendencies 
is the stronger? Is the sharing of odd numbers of electrons 
confined to odd molecules? Let us consider a particular case. 
CHLORINE DIOXIDE 
In this molecule nineteen shell electrons are to be accounted 
for. The following possibilities present themselves: 
1. The total number of shell electrons becomes even by 
transfer of one or more electrons to or from the kernels. 
2. One of the shells in the molecule does not become com- 
pleted to an octet, but to some odd number of electrons. 
3. Two of the shells share, not an even number, but an odd 
number of electrons. 
Explanation 1 is contrary to the strong evidence of the 
definiteness and stability of the kernel pair in He, Li, Be, Bo, 
* The three-electron bond is not here presented as in any sense a new 
idea. It is entirely consistent with Lewis’s postulates, but not with the 
more specific and remarkably successful postulates of Langmuir. Various 
suggestions are found in recent literature of three-electron bonds in C10 2 , 
CeH 8 , etc., cf. J. J. Thompson, Phil. Mag. 41 (1921) 535. 
“Lewis, G. N., Journ. Am. Chem. Soc. 38 (1916) 771. 
