408 BELL SYSTEM TECHNICAL JOURNAL 



which it is found expedient to ascribe to the atoms. These experi- 

 ments are enormously important, for they iinest the atom with a 

 reality which nothing else could give it. To some they have given 

 the hope that all the properties of the atom may one day be demon- 

 strated unquestionably by direct evidence. There is little reason 

 to expect that we shall see that day. The atom is no longer entirely 

 a product of the scientific imagination; but neither is it entirely an 

 object of experience. Most of its properties are invented, not dis- 

 covered. Whether this in\ented and imagined entity is "real" is a 

 difificult question. Perhaps it is best to evade such a question by 

 asking the questioner what he means by "reality". As a matter of 

 fact, it is not pcjssible to discuss atomic theories thoroughly without 

 raising and settling such formidable questions as, what is a theory? 

 and, what is an explanation? and e\-en, what is reality? perhaps 

 eventually, what is truth? I do not aspire to answer these questions. 

 But there are some common misconceptions about atoms which it is 

 prudent to clear away at the beginning. 



In the first place, one does not utter an atomic theory by saying 

 that a substance is made up of small pieces, each exactly like a large 

 piece of the substance in every respect except size. We should 

 achieve nothing by saying that iron is made of black lustrous con- 

 ductive magnetic atoms, or that glass is built of colorless transparent 

 brittle insulating atoms, or that an apple consists of white soft 

 sweet juicy atoms. The atoms must be endowed with fewer and 

 simpler properties than the substance they are meant to compose, 

 else they are futile. One must select some of the properties of the 

 substance to be attributed to its atoms, and st-t tlu' others aside 

 to be explained by those. 



Again, it is not obvious which properties should be selected for 

 the atom; these depend largely on the fantasy of the atom-builder. 

 However, certain qualities such as viscosity and plasticity, con- 

 ductance for heat and conductance for electricity, opacity and trans- 

 parency and lustre, warmth and flavor and fragrance, are not usually 

 assigned to atoms. In general, the more a quality of a substance 

 varies with its state, the less it is suited to be made an atomic quality. 

 Ferromagnetism is a quality which one would assign almost in- 

 stinctively to the iron atom; but it is possiljle to deprive iron alto- 

 gether of this quality by a simple heat treatment, and hence it is not 

 generally supposed to be a feature of the atom. But the rule is not 

 an absolute one. The visible radiations from gaseous iron are sup- 

 posed to be characteristic above all other things of the atom itself, 

 yet they cease when the iron is condensed. It is supposed that in the 



