308 BELL SYSTEM TECHNICAL JOURNAL 



Standard; I must leave it to some chemist to say which is the Hkelier 

 alternative. 



Despite these particular cases, it is a general rule that the masses 

 of the atoms of an element cannot be ascertained, unless its isotopes 

 are separated from each other and separately measured. Indeed, the 

 exceptions to the rule are more apparent than real. One cannot be 

 quite sure that any element is an exception, without performing upon 

 it such an experiment as would separate its isotopes if there were more 

 than one existing in a sensible amount. It is true that there are 

 different radioactive isotopes of one and the same element, which 

 come into being from different sources and therefore are not mixed 

 with one another; but these are generally so scanty in amount that 

 their atomic weights have not been measured at all. Thus every 

 valid measurement of what can properly be called the mass or the 

 weight of an atom requires an "isotope analysis" of the element in 

 question. 



The way of separating isotopes and the way of measuring the masses 

 of their atoms are happily the same, although of course the latter aim 

 demands a great refinement of the method over what is needed for 

 the former. One sends a stream of ions of the element through a 

 sequence of electric and magnetic fields, the first of which accelerates 

 them to a considerable speed, while in the remaining field or fields they 

 are deflected. The deflection depends upon the mass, so that ions of 

 equal charges and different masses — and thus, ionized atoms of the 

 different isotopes of a single element — arrive at different points of the 

 photographic plate which receives them and registers their presence. 

 When the scheme was introduced by J. J. Thomson, he considered it 

 a method of chemical analysis: it was applied to the ions found in 

 electric discharges in ordinary gases and mixtures of gases, and he 

 expected to observe — and did observe — ionized molecules of com- 

 pounds too unstable to be durable. Unexpectedly it turned out to 

 be a method of ultra-chemical analysis, for when applied to the ions 

 of a discharge in neon, it disclosed two kinds instead of one. Efforts 

 were made to identify one of the two as something else than neon, but 

 when they all failed, neon was registered as the first of the elements to 

 be separated into isotopes. 



This discovery was made in 1912, and then occurred the great 

 hiatus of the war. The later story will be an easy matter for historians 

 to trace, at least as far as 1933; for despite its obvious importance, this 

 subject of research invited incredibly few workers. I cannot guess 

 why, in times when many physicists were looking for experimental 

 problems, it was so seldom chosen. There are just three names to be 



