EXPERIMENT. This simply means that we turn again to the outside world and examine it to see whether these deduced results do actually obtain in practice.
If they do we argue that our hypothesis is, probably, a correct one and we retain it until it is shown that if it be correct some result must inevitably occur which in fact does not.
There is a difference between a valid hypothesis and a true one—or, as the latter is commonly termed, a Law.
Any hypothesis is valid which explains the observed facts or at least explains some of them and contradicts none. But the epithet "true" can only properly be applied when it has been shown that all necessary deductions are invariably borne out in practice. As a matter of fact we can never say this with absolute certainty for it is always conceivable that some exception may some day be found which would necessitate the remoulding of the hypothesis.
The most we can say is that certain hypotheses have stood the test in such a very large number of cases without a single failure that there is a very high degree of probability that they are really true.
The hypothesis that the Chemical "Atom" was the ultimate and indivisible unit of matter was a perfectly valid one in the light of the facts that had been observed at the time of its formation and of its apparent proof by Lavoisier and others.
It is only the facts which have been elicited by the study of Ionisation, of Radio-active substances and similar phenomena that have proved it to be untenable and necessitated the substitution of the electronic theory.
Again the Corpuscular theory of light affords a very pertinent illustration of the point I wish to make.