ON TIIE MOLECULE OF WATER. 
177 
silicated hydrogen, and its prototype marsh-gas, as each containing four combining pro¬ 
portions of hydrogen ? For all are agreed that in the molecules of fluorhydric acid and 
chlorhydric acid there is one part of hydrogen united with 19 parts of fluorine, and 35*5 
parts of chlorine respectively ; that in the molecules of ammonia and phosphoretted hy¬ 
drogen there are three parts of hydrogen, united with 14 parts of nitrogen and 31 parts 
of phosphorus respectively; and that in the molecules of marsh-gas and silicated hy¬ 
drogen there are four parts of hydrogen, united with 12 parts of carbon and 28 parts of 
silicon respectively. 
The reasons by which chemists are induced to be so seemingly inconsequent are very 
numerous, and some of them very recondite; but the reason of greatest weight and 
most obvious character is drawn from the phenomena of substitution ; by which is meant 
the change effected in the composition of various bodies, by the abstraction of certain of 
their constituent elements, and the introduction of certain other elements or group¬ 
ings in their stead. Thus we have sodium-alcohol, in which a portion of the hy¬ 
drogen of common alcohol is substituted by its equivalent of sodium; bromaniline, in 
which a portion of the hydrogen of common aniline is substituted by its equivalent of 
bromine ; nitrophenol, in which a portion of the hydrogen of common phenol is substi¬ 
tuted by its equivalent of peroxide of nitrogen ; and so in very many other instances. 
Now the conclusions deducible from the phenomena of substitution, whether elementary 
or compound, direct or indirect, obvious or latent, are, as will presently appear, of the 
greatest importance in determining the molecular formulae of compound bodies. 
It is evident, for instance, that the molecule of marsh-gas must contain four propor¬ 
tions of hydrogen, because we are able to replace ■£, ■£, f, and ± of its hydrogen by four 
successive substitutions, to produce a series of bodies differing from one another by a re¬ 
gular gradation of properties. Thus the final product of the action of chlorine upon 
marsh-gas is a compound in which all the hydrogen of marsh-gas is replaced by its equi¬ 
valent of chlorine; so that if we consider the molecule of marsh-gas to consist of one 
combining proportion of hydrogen united with 3 parts by weight of carbon, then the 
molecule of its chloro-derivative will consist of one combining proportion of chlorine 
united with 3 parts by weight of carbon ; or if we represent the molecule of marsh-gas 
to consist of four combining proportions of hydrogen united with 12 parts by weight of 
carbon, then the molecule of its chloro-derivative must consist of four combining propor¬ 
tions of chlorine united with 12 parts by weight of carbon ; as shown in the following 
table:—■ 
C' = 3 
(N 
rH 
II 
Q 
(3) IIC' 
II 4 C"" (12) 
(12) 3II C'. Cl C f 
Cl II 3 C"" (12) 
(G) HC' C1C' 
C1oH 2 C"" (12) 
(12) HC'.3C1C' 
C1,H C"" (12) 
(3) Cl C' 
C1 4 C"" (12) 
Now exactly intermediate between the original hydride and the final chloride is a body 
which contains hydrogen and chlorine in equivalent proportions, and which consequently 
cannot be represented with less than 6 parts by weight of carbon, and may of course be 
represented with 12 parts ; while intermediate between this body and marsh-gas on the 
one hand, and between it and chloride of carbon on the other, are two additional com¬ 
pounds, the one containing three combining proportions of hydrogen and one combining 
proportion of chlorine, the other containing three combining proportions of chlorine 
and one combining proportion of hydrogen, and both of them consequently incapable of 
being represented with less than 12 parts by weight of carbon. There are thus two sets 
of formulae presented for selection ; the one representing the above-mentioned compounds 
with unequal quantities of carbon by the simplest possible individual formulae, the other 
representing them with an equal quantity of constituent carbon by the simplest possible 
series of formulae ; and a little consideration has led chemists to the unanimous opinion 
that the formulae in the second column do, while those in the first column do not, express 
the actual correlations of the bodies represented. For these bodies manifest in every re¬ 
spect a regular sedation of properties, such as necessarily would be the case if their 
molecules differed from one another only by a gradually increasing substitution of 
chlorine for hydrogen ; but such as could not be the case if the constituent carbon of their 
