412 Mr. Frederick Soddy [March 15, 



The question arose whether by any means an upper Hmit, or 

 maximum value, for the period of ionium could be assigned. By 

 the law already discussed there must be many times as much ionium 

 as radium in uranium minerals, and if the actual ratio were known 

 the period of ionium could at once be found. For example, if the 

 period were 100,000 years, there should be 12*5 grm. of pure ionium 

 per ton of uranium. Auer von Welsbach, in a masterly chemical 

 separation of the rare-earth fraction from ;50 tons of Joachimsthal 

 pitchblende, separated a preparation, which he described as thorium 

 oxide, containing ionium, the activity of which was measured by 

 Meyer and von Schweidler. To obtain a maximum estimate for the 

 period of ionium, 1 assumed that Welsbach's preparation was in 

 reality pure ionium oxide (which it certainly was not, as it gave the 

 thorium emanation), and so I obtained the period of a million years as 

 the upper possible limit. In proportion as the percentage of ionium 

 oxide present is less than 100 %, this period must be reduced.* Thus 

 we have fixed the period of ionium as between 10"^ and 10"^ years, if 

 ionium is the only intervening long-lived member. 



Quite recently a method has been devised for calculating the 

 period of ionium from the range of its a-particles, which is based 

 upon an empirical mathematical relation holding between this range 

 and the periods of the substances giving a-rays in the case of the 

 other members of the series.f The most recent estimate by this 

 method is about 200,000 years, which may be accepted provisionally 

 as the most probable at the present time. If this is correct, there 

 should be 25 grams of ionium per ton of uranium in minerals. A 

 variety of evidence thus leads to the conclusion that to detect the 

 growth of radium from uranium either still larger quantities of 

 uranium or still longer time is necessary. Even after ten years, that 

 is at the end of 1916, if the period of ionium is as estimated, 

 the uranium in No. Ill preparation should only have produced 

 12 X 10'^^ grm. of radium, which is less rather than half the amount 

 that will then have been formed by the ionium initially present. Nos. I 

 and II preparations are very much less favourable. But it is interest- 

 ing to consider No. IV preparation, which, though only 2 • 6 years old, 

 has over seven times as much uranium as No, III. From the present 

 slope of the curve it appears to have little more than one-half as 

 much ionium, relatively to the uranium as No. Ill, whereas the re- 

 lative initial quantity of radium is about twice as great as in No. III. 

 After eight years, that is in 1917, the quantity of radium produced 

 from the uranium should be about equal to that which will have by 

 then been produced from the ionium present. A distinct upward 

 slope should be detectable in the growth curve some time before this. 

 But this is the best, if the estimate of the period of ionium assumed 



* Soddy, Le Radium, 1910, vii. 297. 



t Geiger and NuttaU, PhU. Mag., 1911, xiiii. 613 ; 1912, xxiii. 439. 



