Diffusion of Actinium and Thorium Emanation. 413 



of 235 for the emanation, a number approximating closely to 

 that predicted by theory. The results of these experiments 

 suggested that it might be advisable to determine the co- 

 efficient of diffusion of actinium emanation, itself presumably 

 monatomic, into a monatomic gas. For this purpose a quantity 

 of argon was prepared, and the coefficient obtained by two 

 methods. 



Whatever the nature of the supposed deviations from the 

 ordinary laws of diffusion referred to above, it might reason- 

 ably be expected that they would be exhibited to an equal 

 extent by the different emanations. A direct comparison 

 was therefore carried out over a fairly wide range of pressure 

 between the diffusion of actinium and thorium emanations, a 

 comparison of their coefficients leading directly to the ratio 

 of their molecular weights. 



Method of Experiment. 



The theory of the method here adopted for obtaining the 

 diffusion coefficient has been given by Rutherford*. If a 

 layer of radioactive material be placed at the bottom of a 

 cylindrical vessel a gradient of the emanation is set up, the 

 partial pressure at any layer distant w from the base being 



_ /I 

 given by p=p e v K ' x , where X is the radioactive constant 



of the emanation and K its diffusion coefficient. It is not 

 convenient in practice to measure the partial pressure of the 

 emanation, but if a metal rod be placed along the axis of 

 the cylinder in question and kept at a high negative potential, 

 the active deposit produced by the emanation is directed 

 to the rod soon after formation. We have thus a layer 

 of active matter on the rod, the gradient of which is that of 

 the emanation within the cylinder. The distribution of active 

 matter along the rod can be found by measuring the ioniza- 

 tion produced by the a rays emitted by successive small 

 segments of the rod. This gives us the variation of p with 

 x in the above equation, and knowing the value of X, that of 

 K is determined. 



Debiernef has shown that the above theory is also appli- 

 cable when the emanation is allowed to diffuse up into the 

 space between two parallel plates. In an exactly analogous 

 way a deposit of active matter is obtained on the plates, the 

 gradient of which is that of the emanation itself. 



Both of these experimental arrangements have been used 

 in the course of this work. 



* ' Radioactivity,' p. 27"). 

 t Le Radium, J une 1907. 



Phil. Mag. IS. 6. Vol. 17. No. 99. March 1909. 2 F 



