178 
ON THE MOLECULE OF WATER. 
respective molecules varied in the ratio of 3 to 12, or even of 3 to C. Moreover, by adopt¬ 
ing the formula H 4 C' W for marsh-gas we perceive at once why in its chloro-derivatives 
the hydrogen and chlorine should always stand to one another in the relation of fourths 
But if marsh-gas is to be represented by the formula HC', and its chloro-derivatives re¬ 
garded as compounds of HC' with C1C', there is no reason why we should not have com¬ 
pounds of 1HC'with 2 or 5, or 6 of C1C', or of 2HC' with 3C1C', etc., in which the 
hydrogen would stand to the chlorine in the relation of a third, or of two-thirds, or of a 
fifth, or of a sixth, etc. But with the tetrahydric model, corresponding to the formula 
H 4 C"", and represented below, the successive displacement of the hydrogen of marsh-gas 
in fourths, and in no other proportion, becomes perfectly intelligible, thus :— 
H 4 C 
C1H 3 C C1JLC ClgHC 
C1 4 C 
By the action of chlorine upon marsh-gas, there are produced in succession monoclilo- 
romethene, in which one-fourth of the hydrogen is displaced by chlorine; then dichlo- 
romethene, in which two-fourths of the hydrogen are displaced by chlorine; then tr>chlo- 
romethene or chloroform, in which three-fourths of the hydrogen are displaced by chlo¬ 
rine ; and lastly, tetrachloromethene, in which all four-fourths of the hydrogen are dis¬ 
placed by chlorine. If, then, we take one relative part of hydrogen as the least indivisible 
proportion of hydrogen that can enter into a combination, it is evident that the mole¬ 
cule of marsh-gas must contain four such parts, or four combining proportions, or, in 
common parlance, four atoms of hydrogen. 
The case of ammonia is even more striking than that of marsh-gas. When we effect 
the substitution of some element or grouping for the hydrogen of ammonia, the substi¬ 
tuted and remaining hydrogen are not related to one another in fourths, as happens with 
marsh-gas, but always in thirds. When, for instance, potassium is heated in ammonia, 
there is produced the compound known as potassamide, which consists of one combining 
proportion of potassium, and two combining proportions of hydrogen—one-third of potas¬ 
sium and two-thirds of hydrogen—united with fourteen parts by weight of nitrogen. 
Again, when ammonia is acted on by iodine, there is produced the very explosive body 
known as diniodamide, which consists of two combining proportions of iodine and one 
combining proportion of hydrogen—two-thirds of iodine and one-third of hydrogen— 
united with fourteen parts by weight of nitrogen; while if we act upon ammonia by 
excess of chlorine, we obtain the highly dangerous compound known as trichloramide or 
chloride of nitrogen, in which all the hydrogen of ammonia is replaced by chlorine. But 
the most satisfactory evidence of the trihydric character of ammonia is afforded by Hof¬ 
mann’s experiments on the volatile organic bases, in which he succeeded in replacing 
one-third, two-thirds, and three-thirds of the hydrogen of ammonia by one and the same 
radicle, and by a continuation of one and the same process; thus:— 
N' = 4-7 
(4-7) HN' 
(14 ) 2 HN'. EtN' 
(14 ) HN'. 2EtN' 
(4-7) Et N' 
N'" = 14. 
H 3 N'" (14) 
Et H 2 N'" (14) 
Et 2 H N'" (14) 
Et 3 N'" (14) 
The composition of ethylainine and diethylamine cannot possibly be expressed with 
less than fourteen parts by weight of nitrogen, whilst that of ammonia and triethylamine 
might be represented with one-third of that quantity, or with 4*7 parts. But all four 
bodies present such a marked resemblance to one another, and such a regular gradation 
of properties, as could not be the case unless their respective molecules were strictly com¬ 
parable in constitution. For instance, their boiling-points and vapour-densities increase 
by a regular progression, according to the regular increase in the substitution of ethyl for 
hydrogen; whereas it is certain that the boiling-point and vapour-density of a body con- 
