180 
ON THE MOLECULE OF WATER. 
sented by a tetrahydric model, and that of ammonia by a trihydric model, then the mole¬ 
cule of water must be represented by a dihydric model; thus :— 
NallO" Na,0" 
When water is acted upon by a metallic sodium, one equivalent of sodium expels one 
equivalent of hydrogen, to form the well-defined body hydrate of sodium, and the action 
proceeds no further. But when melted hydrate of sodium, at a temperature approaching 
that of redness, is acted upon by metallic sodium, a second equivalent of sodium turns out 
a second equivalent of hydrogen, to produce oxide of sodium. This replacement of the 
hydrogen of water by sodium at two successive stages,—the first stage of replacement 
taking place at ordinary temperatures, and the second stage at a dull red-heat,—is even 
better defined than the replacement of the hydrogen of ammonia by ethyl at three succes¬ 
sive stages, and then the replacement of the hydrogen of marsh-gas by chlorine at four 
successive stages. 
But the formation of ether from water by Williamson’s process, affords a still closer 
resemblance to Hofmann’s formation of triethylamine from ammonia. Water, H 2 0", when 
acted upon by potassium, yields hydrate of potassium, orpotassic water, KHO". Hydrate 
of potassium acted upon by iodide of ethyl yields alcohol, or ethylated water, EtHO". 
Alcohol acted upon by potassium yields potassium-alcohol, or ethylated potassic water, 
EtKO". Potassium-alcohol acted on by iodide of methyl yields the first mixed ether, or 
methylated ethyl-water, EtMeO"; and when acted upon by iodide of ethyl yields common 
ether, or ethylated ethyl-water, Et 2 0". None of the intermediate bodies can be repre¬ 
sented save with sixteen parts of oxygen, and hence the terminal bodies being members 
of the same series, must be also represented with sixteen parts of oxygen. The intimate 
relations and analogies of ethylated methyl-water, EtMeO", and diethylated water, Et 2 0", 
as regards modes of formation, modes of decomposition, boiling-point, vapour-density, etc., 
will not allow the molecule of the mixed ether to be represented with sixteen parts, and 
that of the common ether with only eight parts of oxygen; any more than the relations 
and analogies of ethyl-methyl-phenylamine, EtMePhN ,,/ , and triethylamine, Et 3 N m , will 
allow the molecule of the former compound to be represented with fourteen parts, and 
that of the latter with only 4‘7 parts of nitrogen. 
With regard to chlorhydric acid, there is no intermediate stage in the replacement of its 
hydrogen; the substitution taking place at once or not at all. We have, for instance, 
chloride of hydrogen, HC1, chloride of ethyl, EtCl, and chloride of sodium, NaCl, but no 
intermediate bodies ; whence the representation of chlorhydric acid by the simple mon- 
hydric model, thus:— 
HC1 NaCl 
In addition to the above-mentioned chemical parallel, there is one very remarkable 
physical relation subsisting between marsh-gas, ammonia, water, and chlorhydric acid, 
namely, that all four molecules occupy the same bulk. Two cubic feet, for instance, of 
chlorhydric acid-gas, HC1, yield one cubic foot of hydrogen and one cubic foot of chlorine. 
Two cubic feet of steam, H 2 0, yield two cubic feet of hydrogen and one cubic foot of 
oxygen. Two cubic feet of ammonia, H ;t N, yield three cubic feet of hydrogen and one 
cubic foot of nitrogen ; while two cubic feet of marsh-gas, H 4 C, yield four cubic feet of 
hydrogen and a quantity of carbon of which the gaseous volume has not been and cannot 
be ascertained. With equal volumes of all four gases, the marsh-gas contains four times 
as much, the ammonia three times as much, and the steam twice as much hydrogen as 
the chlorhydric acid. 
Many other arguments may be adduced to show that the molecular weight of water is 
eighteen, corresponding to the formula H 2 0", and not nine, corresponding to the formula 
HO'. Thus water is concerned as the agent or product of some chemical reaction more 
frequently than any other body with which chemists are acquainted; but in no well-de¬ 
fined reaction do we ever find the reacting or resulting water expressible by nine parts, 
