THE LAWS OF SCIENCE 43 



No more difficult question could be asked, and I cannot 

 pretend to answer it completely, even for that small 

 branch of science which is my special study. The reason 

 for the difficulty is interesting, and we must examine it. 

 Let us return to our first " law," namely that steel 

 will rust if exposed to damp air. I said that the use of 

 the word steel implied an invariable association of 

 properties which is asserted by the more elementary law : 

 There is such a thing as steel. But if we look at the 

 matter closely we shall see that this is not really a law. 

 For there are many kinds of steel ; the substances, all 

 of which the man in the street would call equally " steel," 

 are divided by the fitter into mild steel, tool steel, high- 

 eel, and so on. And the scientific metallurgist 

 Id go further than the fitter in sub-division ; he would 

 recognize many varieties of tool steel, with slightly 

 different chemical compositions and subjected to slightly 

 different heat-treatments, which might be all very much 

 -.line thing for the purposes of the fitter. But if we 

 say that there are several kinds of steel, we are in effect 

 saying that the association of the properties of steel is 

 not invariable, that there can be many substances which, 

 though they agree in some of their properties, differ in 

 is. Thus everything anybody would call steel con- 

 I, according to the chemist, two elements, iron and 

 carbon ; but most steel contains some other element 



as these two, and these other elements cl 

 from one steel to another ; one contains manganese, 

 tungsten, and so on. It is not a law that every 

 substanr >n and carbon (and has cer- 



1 properties of steel) c 

 substan Tee in all the-' respects, 



but iiing nickel in place of mangan 



and iysical propc are not 



There mav out of the difficulty 



