Physical Units of Nature. 387 



It must not be assumed that the atom is always the semi- 

 molecule. In some cases it is found to be the entire molecule, 

 and in other cases the quarter molecule. Thus the mercuric 

 compounds of mercury give vapours of the same bulk as the 

 vapour of the mercury they contain, and indicate an atom of 

 mercury equal to its molecule ; while the other volatile com- 

 pounds of mercury contain more than one molecule of mercury 

 in each molecule of the compound, and therefore do not disturb 

 this conclusion. Again, a litre of phosphuretted hydrogen 

 yields a quarter of a litre of the vapour of phosphorus and one 

 and a half litre of hydrogen, indicating that the quarter mole- 

 cule of phosphorus is its atom. The same is true of arsenic. 



A similar treatment of marsh-gas furnishes 12 as the mass 

 of an atom of carbon, although carbon is not sufficiently vola- 

 tile to enable us to ascertain the relation of its atom to its 

 gaseous molecule. 



By extending this method to all the available cases, we may 

 deduce from the fundamental properties of gases a demonstra- 

 tion of a great part of the modern Table of atomic weights and 

 of the doctrine of atomicity which depends on it. Thus, two 

 bonds* are necessary to connect the group S0 4 with the two 

 atoms of hydrogen that are united to it in sulphuric acid, while 

 one bond is sufficient to join the atoms of hydrogen and chlo- 

 rine in an atom of hydrochloric acid, and so in other cases. 



8. Now the whole of the quantitative facts of electrolysis 

 may be summed up in the statement that A definite quan- 

 tity of eiectricity teayeeses the solution foe each bond 

 that is sepaeated. Thus, if a current pass in succession 

 through vessels containing solutions of sulphuric acid and 

 hydrochloric acid, two atoms of hydrochloric acid will be de- 

 composed in the one vessel for each atom of sulphuric acid 

 that is decomposed in the other ; but the number of bonds 

 separated will be the same in both vessels. 



9. It is the quantity of electricity that passes per bond 

 separated that we have now to determine, and this may be 

 done approximately in the following manner. Several in- 

 quiries (see Prof. J. Loschmidt, " Zur Grosse der Luftmole- 

 ctile," Academy of Yienna, Oct. 1865 ; G. Johnstone Stoney, 

 " On the Internal Motions of Gases," Phil. Mag. August 

 1868 ; and Sir William Thomson, " On the Size of Atoms," 

 ' Nature,' March 31, 1870) have led up to the conclusion that 

 the number of molecules in each cubic millimetre of a gas at 



* The word bond is here used of the connexions between atoms when 

 they enter into combination. When we use this, which seems the proper 

 signification of the word, the bonds are to be distinguished from the hands 

 or feelers which each atom has, and which, by grappling with the hands 

 or feelers of other atoms, establish bonds between them. 



