DEVELOPMENT BY NATURAL SELECTION. 113 



chemistry have to appeal to facts in their own depart- 

 ment of knowledge. 



But this attitude leaves us in face of an apparent 

 absolute contradiction, which crops up at every point 

 and paralyses investigation. We must evidently probe 

 deeper, in the faith which is the presupposition of all 

 knowledge, that no real contradiction exists. Let us, 

 then, examine the physical and chemical data with a 

 view to discovering how far we are dealing with 

 established fact, and how far with hypothesis. 



I think the discussion of this point will be more 

 intelligible if I give it a concrete form, and I will take as 

 a starting-point the kinetic theory of gases. By starting 

 from the assumption that a gas is an assemblage of 

 immense numbers of free molecules each of which is 

 perfectly elastic, so that it repels anything else on 

 contact, and that each molecule is in very rapid motion, 

 the source of this motion being heat, Waterston, Clausius, 

 and Clerk Maxwell showed that we can predict a number 

 of facts, and in particular those hitherto known only 

 empirically as the so-called gas-laws discovered by 

 Boyle, Charles, and Avogadro, and embodied in the 

 well-known equation PV = RT. 



When, however, we go back to the actual facts, we 



find that although for many gases PV is within certain 



common ranges of pressure and temperature nearly 



constant at constant temperature, the constancy is 



only approximate. When P is sufficiently high, or T 



sufficiently low, the law breaks down in every case, 



and to different extents and in different directions for 



different gases. This failure is not merely due to the 



fact that the molecules themselves have a certain 



8 



