THE EXPLANATION OF LAWS 87 



So far the truth of a theory has been based on two 

 grounds ; first, that the laws to be explained can be 

 deduced from it ; second, that it really explains in the 

 sense that has been indicated. But actually there is in 

 addition, a third tost of the truth of a theory, which is 

 of great importance ; a true theory will not only explain 

 adequately the laws that it was introduced to explain ; 

 it will also predict and explain in advance laws which were 

 unknown before. All the chief theories in science (or 

 at least in physics) have satisfied this test ; they have 

 all led directly to the discovery 7 of new laws which were 

 unsuspected before the theory was proposed. 



It is easy to see how a theory may predict new laws. 



The theory, if it is worthy of consideration at all, will 



be such that the old laws can be deduced from it. It may 



.y be found on examination that not only these laws, 



but others also can be deduced from it ; so far as the 



theory is concerned, these others differ in no way from the 



and if the theory is to be true, these laws 



that are consequences of it must be true. As a matter of 



it is very seldom that a theory, exactly in its original 



form, predicts any laws except those that it was proposed 



to explain ; but a very small and extremely natural 



lopment of it may make it predict new laws. Thus, 



iple, in order to explain the laws (Boyle's 



and Gay-Lussac's) to which the theory was originally 



unnecessary to make any assumption about 



ize of the molecules ; those laws can be deduced fn >m 



C that size (so long as it is below a 



and the assumption was at fu>t made for 



lie molecul< mathematical points 



1 all But obviously it was n 



i.-iu^h extra 

 . ' havt: id once that assumpt 



1 If a drop of wat the 



molecules v 



