[October, 
488 On Scientific Method. 
luminous whole. But he does not stop here ; he again 
appeals to faCts. He says “ If my hypothesis be correct, 
this or that faCt must follow.” Then he tries experiment. 
Is the predicted faCt really a faCt ? By the test of experi- 
ment he is content to abide ; he knows that Nature — how- 
ever hard sometimes it is to make her answer at all — never 
answers except truly. 
Hypotheses must be used in Science, but hypotheses may 
be abused. What, then, are the marks of a good hypothesis ? 
An hypothesis must be workable ; it must not go against 
any well-established scientific generalisation, and it must 
be ready to submit to have its predictions proved by 
stridtly experimental methods.* 
A good scientific hypothesis must be workable ; that is, it 
must allow us to make determinate predictions — predictions 
which can be proved or disproved by experiment. A vague 
generalisation, which does not allow of definite deductive 
reasoning, can have no place as a scientific hypothesis. 
A good scientific hypothesis must not be opposed to any 
well-established generalisation of Science. This statement 
may probably be called in question by many. It is no un- 
common thing to find people talking of the way in which 
Science sweeps aside all preconceived ideas, all Old World 
notions, all long-cherished delusions. And this is very true; 
only these people, I am afraid, forget the other side of the 
case ; they forget that did Science present us with nothing 
but change succeeding change, doctrines swept away and re- 
placed by others to be themselves removed as Science 
advances, Science would have no claim on our acceptance. 
It is because Science is at once prolific of changes, and con- 
servative in the extreme, that she has accomplished her work 
in the world. 
We know so little of Nature that we must be ready at 
any moment to give up that which we had supposed we did 
know ; and yet we have such trust in the stability of Nature 
that we must cling to those theories which have been gained 
by slow accumulation of faCts, until there is absolute experi- 
mental proof of their falsity. 
Such a theory as that of the Conservation of Energy is 
the general expression of a vast number of faCts : it explains 
these faCts ; it is a well-established generalisation of Science. 
If we are seeking to explain a number of newly-discovered 
fadts, it is evidently our duty to frame an hypothesis which 
shall not be itself out of keeping with this theory of the 
* Principles of Science, vol. ii., p. 139. 
