290 Messrs. F. Soddy and T. D. Mackenzie on the 



three months, the extraction was commenced. The aqueous 

 residue was concentrated and re-extracted till nearly all the 

 uranium was removed. It was then precipitated with 

 ammonium carbonate in excess, and the small precipitate con- 

 taining the uranium X, as shown by a test on a photographic 

 plate, was sealed up in a flask on June 26th, 1906. It was 

 tested for radium on July 9th, in an electroscope giving 1*9 

 to the 7-ray test, and gave a leak of 6*5, equivalent to 

 13 x 10~ n gram radium. It is possible, in the numerous opera- 

 lions with the uranium since the last test, that a trace of the 

 radium-containing residues from earlier work contaminated 

 the quantity, which would account for the somewhat high 

 result. 



This preparation was left until June 20th, 1907, but the 

 gas pumped out on that date was lost. On July 1st, 1907, 

 another test was made in an electroscope giving 2*68 to the 

 7-ray test, and gave a leak of 10*5, equivalent to 15 X 10 -11 

 gram of radium. It is possible that this slight increase in 

 the quantity of radium during the 370 days the experiment 

 has been in progress is real, and due to a partial separation of 

 the hypothetical parent of radium with the uranium X, but 

 in view of the comparatively large amount of radium present 

 it is at present very doubtful, and further results must be 

 awaited. 



Expekiment IV. 



This experiment and the next have to do with the possi- 

 bility that actinium is an intermediate member in the uranium- 

 radium series. An attempt has been made to detect the 

 growth of actinium in the uranium solution employed in the 

 last two experiments. Actinium was tested for by passing a 

 rapid stream of air for two hours through the uranium 

 solution boiling under reduced pressure in a flask with upright 

 condenser placed on the water-bath. After the first half-hour 

 the air was passed into a brass cylinder provided with a 

 central insulated and removable electrode charged negatively 

 to 250 volts with respect to the cylinder. The volume of the 

 condenser was 50 c.c., and the rate of the air-stream 85 c.c. 

 per minute measured at atmospheric pressure. The pressure 

 was kept at about 100 mm. of mercury, and it w T as estimated 

 that the time taken for the gas to pass from the uranium solu- 

 tion into the neighbourhood of the negative electrode was not 

 greater than 5 seconds, and probably less. So that a fraction 

 of somewhat less than one-half of any actinium emanation 

 generated would deposit its products upon the electrode. 

 The latter was coiled up and tested in the electroscope before 



