The conception of the elementary atom as a particle which had not been decomposed followed from many observed facts besides those concerning elements, and has been justified by the usefulness of it as an instrument for forwarding accurate knowledge.

Investigations proved radio-activity to be a property of the very minute particles of certain substances, and each radio-active substance to have characteristic properties, among which were certain of those that belong to elements, and to some extent are characteristic of elements.

Evidently, the simplest way for a chemist to think about radio-activity was to think of it as an atomic property; hence, as atomic properties had always been regarded, in the last analysis, as properties of elements, it was natural to place the radio-active substances in the class _elements_, provided that one forgot for the time that these substances have not the class-mark _have-not-been-decomposed_.
As the facts of radio-activity led to the conclusion that some of the minute particles of radio-active substances are constantly disintegrating, and as these substances had been labelled _elements_, it seemed probable, or at least possible, that the other bodies which chemists have long called elements are not true elements, but are merely more stable collocations of particles than the substances which are classed as compounds.

As compounds can be changed into certain other compounds, although not into any other compounds, a way seemed to be opening which might lead to the transformation of some elements into some other elements.
The probability that one element might be changed into another was increased by the demonstration of the connexions between uranium and radium.

The metal uranium has been classed with the elements since it was isolated in 1840.