152 THE ANATOMY OF SCIENCE 



of the twenty balls I had taken a million grains 

 of white and black sand the mixing would have 

 appeared to be more thorough, but the fluctua- 

 tions, although not so evident, would still be 

 there, and if we should take an enormously 

 greater number of particles we would have a 

 complete analogy to the mixing of the mole- 

 cules of oxygen and nitrogen. And here again 

 we must assume the existence of fluctuations 

 from the mean, sometimes small, sometimes 

 great, and after the lapse of a vast period of 

 time even so great as to give once more pure 

 oxygen and pure nitrogen. Here again if there 

 were a way of recording these fluctuations be- 

 tween the original distribution of pure oxygen 

 and pure nitrogen until we arrive once more at 

 the same distribution, there would be nothing 

 in all these fluctuations which would not be 

 symmetrical with respect to past and future. 



Are all cases of irreversible phenomena as 

 simple as this? For many years I have found 

 this one of the most perplexing and tantalizing 

 of scientific questions. As long as it was sup- 

 posed that radiation, even from a single atom, 

 goes out in a way which can never be entirely 

 reversed, there seemed an insurmountable obsta- 

 cle to the view that all irreversible processes are 



