Valency of the Radio-elements. 595 



Godlewski *, on the other hand, investigating the properties 

 of solutions of the active deposit of radium in water, found 

 that the charge of the particles can be changed by the 

 addition of multi-charged cations and anions, a property 

 which is highly characteristic of colloids. In order to 

 obtain a further proof of the colloidal nature of such solu- 

 tions and to obtain, by experiment, an estimation of the size 

 of these small particles, the diffusion of radium D, E, and F 

 and of thorium B in alkaline, neutral, and dilute acid solution 

 was measured. 



Diffusion in Alkaline Solution. 



Alkaline solutions of the above-mentioned elements are 

 not completely stable, since the particles discharge them- 

 selves as time goes on and are precipitated on the walls of 

 the vessel. It is no longer sufficient, as in the diffusion 

 of stable solutions, to determine the ratios of the concen- 

 trations of the initial and final layers of the diffusing 

 substances, since the initial layer loses a p:irt of its content 

 through coagulation which will play no part at all in the 

 diffusion f. After the termination of the experiment the 

 concentrations of all the layers are determined and added 

 together. The total value so obtained is taken as the 

 content of the initial layer before the experiment began. 

 The possible error that may be made by this assumption 

 has the result that the velocity of diffusion seems somewhat 

 greater than it actually is, and the substance therefore seems 

 less colloidal ; the actual value, however, is not substantially 

 altered. 



Diffusion of polonium into \ N ammonia. 



D160.3. Dl8°« 



n - g^r • °- 13 CV14 



ni. ^ °' 21 °' 22 



iv. §^r o-ie o-n 



Mean value for Di 8 o is 0*19 sq. cm. per day. 



* Godlewski, Bull, de V Acad. d. Science de Cracovie (1913). 



t Frl. Meitiier Las already noticed the precipitation of radium C from 

 solution by addition of ammonia ; it is of interest to mention that the 

 failure of the v. Lercli method in alkaline solution is due to radium C 

 being present in a colloidal state. 



