EQUATION AND THE NATURE OF COHESION. 67 



one and the oxygen one. Perhaps it is for the reason that the oxygen 

 here as in 2 has but one negative electron that it unites with the 

 hemoglobin of the blood to form the carbon monoxide hemoglobin, 

 just as oxygen unites to make oxy-hemoglobin, which has so similar 

 a spectrum. In C0 2 we find by this method an agreement between 

 the values of a computed from the surface tension and in other 

 ways with that computed from the weight and valences, if we 

 assume 5 valences that is 5 electrons in the molecule. In this case 

 as in CO we would assume that one valence electron remained 

 attached to the carbon atom and that each oxygen atom was reduced, 

 that is each had its full complement of two negative valence elec- 

 trons. Carbon dioxide does not unite with hemoglobin in the way 

 oxygen does or as CO does. The oxygen is evidently in a different 

 state; moreover, altho we have here two atoms of oxygen, just as 

 we do in 2 , yet this compound is diamagnetic and not at all para- 

 magnetic as is 2 . This again would suggest a difference in the 

 state of the oxygen in the two cases. Similarly in S0 2 we have 

 also too few negative electrons if the sulphur be considered hexa- 

 valent. It is an interesting fact that in all other compounds 

 of sulphur for which I have computed the cohesion, with the 

 single possible exception of carbonyl sulphide, sulphur is found 

 to be hexavalent. This is in entire agreement with the sugges- 

 tion of Lewis that the maximum valence of sulphur is six, and 

 there are six: valence electrons. In S0 2 we find only 8 valence 

 electrons. Evidently when the sulphur was oxidized by the oxy- 

 gen the latter relieved it of 2 valence electrons. It is in this 

 way, perhaps, that polar compounds are produced. Here we would 

 have sulphur with 4 valence electrons, and each of the oxygen 

 atoms with 2. 



We may represent these compounds graphically in the following- 

 way arranging the electrons somewhat in the manner suggested by 

 Lewis (see Figure 2, next page). 



I have put the electrons representing the valences beneath the 

 atoms, instead of between them as Lewis does, to suggest that the 

 atoms may place themselves in such a way that the valence electrons 

 are rotating in parallel orbits and do not necessarily lie between 

 the atoms. 



A further study of other so-called polar compounds will be of 

 great interest and will perhaps show the truth or erroneous cha- 

 racter of these suggestions, but this will have to be deferred until 



uu ? 



a shall have been determined for other such substances. It will 

 however be remarked that there is no evidence of any such oxida- 



