NEW CONCEPTIONS IN SCIENCE 



the weight of a drop about the thickness of a pen- 

 knife's blade. A millionth of a milligram of sodi- 

 um in a lamp-flame suffices to reveal distinctly 

 the characteristic sodium coloration in the rain- 

 bow-like spectrum of the flame. To do this, this 

 millionth of a milligram must have been vaporized 

 into a gas containing millions of distinct particles. 

 Another striking example might be drawn from 

 the phenomena of smell ; a grain of musk will scent 

 a room for years. All this time, if the present 

 theory of smell is correct, it must have been raining 

 its small particles in every direction, incessantly. 

 Yet it seems to lose little or nothing in weight. 



There would be very little point in cataloguing 

 these instances of extreme divisibility by them- 

 selves if they did not afford some clew to the 

 structure of matter. From the days of the old 

 Greeks, and, doubtless, tens of thousands of for- 

 gotten and unstoried years before, curious-minded 

 man has puzzled over the problem of whether this 

 divisibility has any limit or not. The old Greeks 

 could only guess; they had no tools for the brain 

 to work with. The whole progress of science, and, 

 for that matter, of the human mind, is conditioned 

 on mechanical appliances. The only difference 

 between a savage and a Lord Kelvin is that the 

 savage must depend on his unaided hands, his 



114 



