its radio-activity. Taking the above estimate of the life of radium, the value of λ is 5·4 × 10-4, with a year as the unit of time. A mass of radium left to itself should be half transformed in 1300 years and only one-millionth part would remain after 26,000 years. Thus supposing, for illustration, that the earth was originally composed of pure radium, its activity per gram 26,000 years later would not be greater than the activity observed to-day in a good specimen of pitchblende. Even supposing this estimate of the life of radium is too small, the time required for the radium practically to disappear is short compared with the probable age of the earth. We are thus forced to the conclusion that radium is being continuously produced in the earth, unless the very improbable assumption is made, that radium was in some way suddenly formed at a date recent in comparison with the age of the earth. It was early suggested by Rutherford and Soddy19 that radium might be a disintegration product of one of the radio-elements found in pitchblende. Both uranium and thorium fulfil the conditions required in a possible source of production of radium. Both are present in pitchblende, have atomic weights greater than that of radium, and have rates of change which are slow compared with that of radium. In some respects, uranium fulfils the conditions required better than thorium; for it has not been observed that minerals rich in thorium contain much radium, while on the other hand, the pitchblendes containing the most radium contain a large proportion of uranium.
If radium is not produced from uranium, it is certainly a remarkable coincidence that the greatest activity of pitchblende yet observed is about five or six times that of uranium. Since radium has a life short compared with that of uranium, the amount of radium produced should reach a maximum value after a few thousand years, when the rate of production of fresh radium—which is also a measure of the rate of change of uranium—balances the rate of change of that product. In this respect the process would be exactly analogous to the production of the emanation by radium, with the difference that the radium changes much more slowly than the emanation. But since radium itself in its disintegration gives rise to at least five changes with the corresponding production of α rays, the activity due to the radium (measured by the α rays), when in a state of radio-active equilibrium with uranium, should be about five times that of the uranium that produces it; for it has been shown that only one change has so far been observed in uranium in which α rays are expelled. Taking into account the presence of actinium in pitchblende, the activity observed in the best pitchblende is about the same as would be expected if the radium were a disintegration product of uranium. If this hypothesis is correct, the amount of radium in any pitchblende should be proportional to the amount of uranium present, provided the radium is not removed from the mineral by percolating water.
This question has been experimentally attacked by Boltwood20, McCoy21 and Strutt22. McCoy measured the relative activities of different minerals in the form of powder by means of an electroscope, and determined the amount of uranium present by chemical analysis. His results indicated that the activity observed in the minerals was very approximately proportional to their content of uranium. Since actinium is present as well as uranium and its products, this would indicate that the amount of radium and actinium taken together is proportional to the amount of uranium. This problem has been attacked more