bstances may have precisely the same chemical
constitution--the same number and kind of atoms--and yet differ utterly
in physical properties. The word isomerism was coined by Berzelius to
express this anomalous condition of things, which seemed to negative the
most fundamental truths of chemistry. Naming the condition by no means
explained it, but the fact was made clear that something besides the
mere number and kind of atoms is important in the architecture of a
molecule. It became certain that atoms are not thrown together haphazard
to build a molecule, any more than bricks are thrown together at random
to form a house.

How delicate may be the gradations of architectural design in building
a molecule was well illustrated about 1850, when Pasteur discovered that
some carbon compounds--as certain sugars--can only be distinguished
from one another, when in solution, by the fact of their twisting or
polarizing a ray of light to the left or to the right, respectively. But
no inkling of an explanation of these strange variations of molecular
structure came until the discovery of the law of valency. Then much of
the mystery was cleared away; for it was plain that since each atom in a
molecule can hold to itself only a fixed number of other atoms, complex
molecules must have their atoms linked in definite chains or groups. And
it is equally plain that where the atoms are numerous, the exact plan of
grouping may sometimes be susceptible of change without doing violence
to the law of valency. It is in such cases that isomerism is observed to
occur.

By paying constant heed to this matter of the affinities, chemists are
able to make diagrammatic pictures of the plan of architecture of any
molecule whose composition is known. In the simple molecule of water
(H2O), for example, the two hydrogen atoms must have released each
other before they could join the oxygen, and the manner of linking must
apparently be that represented in the graphic formula H--O--H.
With molecules composed of a large number of atoms, such graphic
representation of the scheme of linking is of course increasingly
difficult, yet, with the affinities for a guide, it is always possible.
Of course no one supposes that such a formula, written in a single
plane, can possibly represent the true architecture of the molecule:
it is at best suggestive or diagrammatic rather than pictorial.
Nevertheless, it affords hints as to the structure of the molecule such
as t