, in their phosphorus
compounds, upon water, when the halogen takes the place of the
hydroxyl group, but never that of the hydrogen.

Now as to some logical deductions from the foregoing considerations.
Hydrogen is by many looked upon as a true metal. This theory cannot be
directly proved by the above, but it is certainly greatly strengthened
thereby. To compare. Hydrogen is a powerful reducing agent; it is
similarly affected by the halogens, the hydroxyl group, the acid
radicals, oxygen and sulphur; hydrogen and members of the univalent
alkali metals group are readily interchangeable; it forms superoxides
analogous to the metals; its analogy to the alkali metals as exhibited
in the following:

H H(OH) HCl HNO_{3} H_{2}SO_{4} H_{2}S H_{2}O_{2}
K K(OH) KCl KNO_{3} Na_{2}SO_{4} Na_{2}S K_{2}O

But if we consider hydrogen as a gasiform metal, we naturally arrive
at the conclusion that _water is the hydroxide of this gasiform
metal_, that is _hydrogen hydroxide_, while gaseous hydrochloric and
hydrosulphuric acids would be looked upon as respectively the chloride
and the sulphide of the metal hydrogen. This would then lead to
curious conclusions concerning the hydroxyl group. This group would,
by this theory, become an oxygenated metal radical similar to the
hypothetical bismuthyl and uranyl, and yet one in which the metallic
character has disappeared as completely as in the ferrocyanic group.

An entirely new light is shed by this view upon the composition of
hydrogen peroxide, which would be looked at as two free hydroxyl
groups joined together thus: (OH)--(OH), analogous to our di-ethyl,
diphenyl, dicyanogen, etc. Considered as dihydroxyl, it would explain
the instability of this compound.

The ethers proper would also be placed in a new light by this new
conception of the constitution of the water molecule. The hydrogen in
the hydroxyl group, as is known, may be substituted by an alkyl group.
For instance, an alkyl may be substituted for the hydroxyl hydrogen in
an alcohol molecule, when an ether results. According to the new
theory this ether will no longer be considered as two alkyl groups
connected by an oxygen atom, but as a compound built up on the type of
water by the union of an alkyl group and an alkoxyl group. Thus
ethylic ether would not be represented by

C_{2}H_{5}
> O,
C_{2}H_{5}

as heretofore, but by the formula C_{2}H_{5}(OC_{2}H_{5}),