282 



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



after solution must be corrected for the escape of emanation 

 from the powdered mineral. An experiment showed that 

 about 8' 7 per cent, of the emanation generated escaped from 

 the powdered mineral, so the first tests are corrected for this 

 amount. 



It was not known whether the result obtained with a 

 reaccumulation of emanation in the uraninite solution could 

 be relied upon owing to the possibility of the radium pre- 

 cipitating out in a non-emanating form after solution of the 

 mineral, and the experiments included an examination of 

 this point. It was found that the results with the emanation 

 in the solid mineral, corrected for the escape as indicated, 

 agreed satisfactorily with later ones from the same sample 

 of mineral in solution, so that it would be easier in calibrating 

 an electroscope to dispense altogether with the initial tests 

 with the solid and the necessity of determining the rate of 

 escape from the solid, and to start with a known period of 

 reaccumulation after boiling out the emanation from the 

 dissolved mineral and sealing up the flask. The following 

 Table shows the results obtained with three different quantities 

 of mineral : — 





Date. 



Milligrams 

 Uraninite. 



Observed 



leak per 



milligram. 



Time of 

 accumu- 

 lation. 



Corrected 

 leaks per 

 milligram. 



y-ray test 

 of Electro- 

 scope. 



Leak in 

 Electro- 

 scope of 

 mean sensi- 

 tiveness. 



1.. 



14/6/-07 



13-5 



13-3 



Solid 



14-6 







2... 



19/6/-07 



3-5 



10-7 



;j 



11-7 







3... 



26/6/-07 



2-9 



13-5 



„ 



14-3 * 



32 



12-O^j 



4... 



27/6/-07 



3-5 



12-25 



6 days 



181 



3-2 



15-3 [Mean 

 13-0 [14-3 



5... 



10/7/-07 



2-9 



12-8 



13 '75 days 



13-8 



2-85 



6... 



10/7/-07 



3-5 



15-3 



13 days 



168 



2-7 



16-8; 



* In this test the solid was sealed up in the flask for two days before the 

 test, and this is allowed for in the correction. 



The mean of the last four results gives 14"3 per milligram 

 in an electroscope of mean sensitiveness, so that a leak of 1 

 corresponds to 1*4 X 10 -11 gram of radium. The minimum 

 detectable quantity of radium in any individual experiment 

 was never less than one-half and usually about one-fifth of 

 this amount. 



The earlier result obtained by the use of radium directly 

 gave 10 ~ 10 gram of radium as the equivalent of 1 division 

 per minute, and this is therefore probably about seven times 

 too great. It will be recalled that there was an error in the 



