46 HISTORY OF CHEMISTRY [LECT. in.
sible, whilst the latter assumed that substances combine in a
few definite proportions only. We call a substance a com-
pound when it contains its constituents in invariable pro-
portions.
I do not know whether the difference between the two
conceptions has been made clear. In order to appreciate
the full bearing of the question, a person must himself have
required to decide as to whether he had a mixture or a com-
pound in his hands. Even yet we are without a definition
which shall suffice for every case ; that is, without such a
definition as Berthollet repeatedly demanded from Proust.
It is true that we have certain means of judging with respect
to chemical compounds, as, for instance, capacity for crystal-
lising, and invariable melting point in the case of solids, and
constant boiling point in the case of liquids. Yet these are
frequently insufficient. I need only recall the phenomena
of isomorphism, when we must admit that mixtures also can
crystallise. I mention the solutions of hydrochloric acid,
hydriodic acid, etc., in Avater, regarding which Roscoe has
proved that they are only mixtures (solutions), and we must
admit that these likewise can possess a constant boiling
point. In short, this distinction forms one of the most dif-
ficult and most important problems, and in point of fact it is
often insufficiently attended to. In the study of chemical
papers opportunity is often afforded of observing how errors
have arisen through neglect of this very matter. How often
have formulae been advanced for substances and theoretical
conclusions based upon their existence, before their com-
pound character has been conclusively settled ! The pur-
pose of the foregoing remarks is to serve as a warning against
any such error, and I therefore hope to be excused for
having, for a short time, quitted my proper theme.