Relation between Uranium and Radium. 217 



Discussion of Results. 



Naturally the relative errors in the determination of these 

 infinitesimal amounts of radium growing in large volumes 

 of solutions are considerable, and the difficulties are greatly 

 increased by the length of time over which tbe measurements 

 have extended. The chief sources of error are in the possible 

 change of the constant of the instrument,, the change in the 

 standards used to calibrate the instrument, and the actual 

 errors of determination which apply equally to the calibration 

 of the instrument as to the tests on the preparations themselves. 

 Naturally, if the measurements were restarted now, it would 

 be possible, with the greater knowledge and experience 

 now available, to improve the former results. The difficulty 

 is always to be sure that the measurements done, say, five 

 years ago are in every way comparable with those now 

 being done. In future it is proposed to- avoid the use of 

 liquid emanation standards, prepared from uranium minerals, 

 altogether. 



The quantity of radium in a pure solid barium chloride 

 preparation containing about 6 X I0~ 10 g. of radium per gram 

 is beiiw determined once for all, and in future fresh weighed 

 quantities of O'l to 0'5 gram of this preparation will lie used 

 to calibrate the instrument as required. In this way, the 

 measurements over long periods of time may be expected to 

 agree better with one another. 



The general character of the results is, however, now fairly 

 clear. As is to be expected, Preparation IV., though the 

 youngest of the four preparations, gives already the most 

 information owing to the very large quantity of uranium — 

 from 8 to 12 times the quantity of any of the other preparations. 

 The growth of radium during the first period of three years 

 from purification was only about one-third the growth in the 

 second period of three years, in agreement with what is to be 

 expected, if the growth in this preparation is entirely due 

 to the uranium and if ionium was initially absent. On this 

 assumption, the period of ionium calculated from the present 

 results agrees fairly well with that calculated for the period 

 on the same assumption three years ago, which indicates that 

 the assumption itself cannot be seriously at fault. 



Sir E. Rutherford has shown* that the initial growth of 

 radium from uranium is represented by 



where 11 is the number of atoms of radium grown from the 

 number of atoms of uranium in equilibrium with R atoms 

 • ' Radioactive Substances and their Radiations,' p. 466. 



