THE HEAT OF JJELAXATION Ol' THE EKAD LATI rf'K 



119 



the shorter range of the recoil particles the latter method is indeed 

 considerably more sensitive than the method first described. In the 

 study of the self-diffusion of lead ions in lead iodide it has been found 

 possible to determine that the measurement of a-radiation and recoil 

 yield the same result ; in spite of the fact that the recoil measurements 

 register processes in th(> top few hundred molecular layers they were 



28 27 26 2S 2^ 23 22 21 20 19 IS 17 



^T'lO^ 



Fig. 1. Rate of Diffusion of labeUed solid Lead. Schmelztemperatui — 



Melting point. 



Iff 



as highly reproducible as the a -measurements. Recoil measurements at 

 lead surfaces, on the contrary, indicated a high sensitivity of the metallic 

 surface to external effects. For example, contact of the lead sampk 

 with air for a short time was sufficient to lower the diffusion values 

 and even the values obtained in a carefully purified nitrogen atmos- 

 phere were rather lower than those determined when working in a 

 vacuum. In spite of the uncertainty arising for these reasons in the 

 recoil values, the experimental points obtained by this method also 

 lie approximately on the straight lines obtained with thea-measurements 

 as the basis (cf. Fig. 1) ; it must also be borne in mind that the recoil 

 range in lead is not accurately known and that it must be calculated 

 by extrapolation from the values measured in air. 



The temperature coefficient of the velocity of diffusion is in quite 

 good agreement with the results of the already mentioned preliminary 

 experiments, but the difference in behaviour of single crystals and crys- 



