THE OLD AND NEW CHEMISTRY. 
33 
The new system seeks for as wide a basis as possible on 
which to rest the fabric of chemical science ; the old system, by 
restricting itself to what it calls facts, narrows that basis so 
much as to render the fall of the whole structure not very 
improbable. 
' I have said that the modern system of notation is based upon 
two-volume formulae, checked by specific heat determinations 
and by the study of crystalline forms : in many instances the 
specific heat and isomorphism methods are alone applicable for 
the determination of combining numbers. The numbers thus 
obtained express the (relative) amounts in which the elements 
enter into combination. We are justified in saying that oxygen 
always enters into combination in the proportion of 16 parts by 
weight, or a multiple of this number. It has been urged, by a 
distinguished chemist, that such a statement as this is not jus- 
tified by the facts ; that the amounts of oxygen occurring in two 
volumes of series of compounds are very nearly, but sometimes 
not exactly, multiples of 16 ; that the so-called law of multiple 
proportions has no existence apart from the atomic theory. 
Unless we found our notation on the theory that matter has an 
atomic structure and that chemical combination takes place 
between atoms, we have no right, it is argued, to say that the 
elements combine in multiples of certain fixed numbers. If 
the truth of the atomic theory be assumed, the 44 law of multiple 
proportions ” must necessarily hold good. But it has been stated 
that we have no rigid experimental proof of the 64 law ; ” hence 
the conclusion is drawn that the atomic theory is a true theory. 
From the data I should rather conclude that the atomic theory 
is not a true theory. But I deny the data. We have sufficient 
experimental proof in favour of the generalization that the 
elements combine in certain fixed proportions by weight. The 
modern system of notation does not, I am convinced, necessi- 
tate the adoption of the atomic theory. In the preceding 
parts of this paper I have endeavoured to elucidate the main 
points of the old and new systems without referring to the 
atomic theory or using language which suggests such a theory. 
I do not believe that we can develop a system of chemical 
notation (nor indeed any scientific system) which shall be abso- 
lutely free from hypotheses. The 44 law of multiple proportions ” 
assumes that the apparent small divergences from the general 
statement which sometimes seem to occur are really due to 
errors of experiment. We must be content to make use of hypo- 
theses in science ; only we must test these by frequent appeals 
to facts, and we must always remember that they are hypotheses. 
The combining numbers of the elements, then, generalize a 
great amount of knowledge concerning these bodies. But one 
naturally asks why do the elements combine in fixed proportions ? 
NEW SERIES, VOL. II. NO. V. D 
