V.
��REVIEW OF THE RESULTS OBTAINED.
��molecules of the elements were always formed by a certain number of atoms, quite independent of the state of aggrega- tion in which they exist.
Compound molecules behave in the same way — that is to say, the molecular weight is the same in the gaseous and in the dissolved condition. In so far as the composition of the dissolved molecules is concerned, there frequently exists a alight difference between two solvents, as, for instance, between water and benzene, in the same way as there is a difference between these liquids and a vacuum, which may be conceived as a solvent for gases.
A further conclusion drawn &om the doctrine that no free valencies could occur in molecules, was that the valency of an element could only change by an even number. ThuSy gold could be monovalent or trivalent; but the assumption was made that gold is trivalent, and two of the valencies may saturate each other and so allow the gold atom to appear monovalent. Auric and aureus chlo- rides were assumed to have the con- stitutions shown. Nevertheless, several cases were knoMm in which the number of valencies changes by an uneven number, as, for example, in ferrous chloride (FeCla) and ferric chloride (FeCls). In order to explain this anomaly it was assumed that iron is di- and tetra-valent in these compounds which are constituted as shown in the diagram given.
So long as the molecular weights of these substances could not be determined such assumptions appeared quit« valid. But it is much more difficult to explain the exceptions formed by the series of nitrogen oxides. Nitrogen is pentavalent, and accordingly the only compounds which should be capable of existence are N2O, NaOs, and NgOs. Besides these, however, the oxides NO and NOa are also known, and their gas densities
��Au^
�Au-Cl
�Cl
�Anrooi
�Auric
�chloride.
�chloride.
� �^ci
� �„ x'
�.---s
�Fe-Cl
�/
� �Cl
�Fe Cl
� � �\
�Cl
�Fe-('l
� �^Cl
�Ferrous
�Ferric
�chloride.
�(
�^loride.
�� �
