1907.] 



BARUS— NUCLEI IN DUST-FREE WET AIR. 79 



which is an example chosen at random from many similar cases. 

 The mean b (excluding the last) is thus .000,0045, at least three 

 times as large as the electrical coefficient. 



If but a part ii of the nuclei are caught by the exaustion, n' escap- 

 ing, — dn/dt — dn'/dt = bn~ -{- 2bnn' -\- bn'-. Hence if but i/iii 

 of all the ions are captured, the coefficient of decay, b, found should 

 be about iii times too large as compared with the true value, or 

 dn/dt = — uibn-. But this fails to explain the increase of b with 

 i/n, unless the nuclei grow smaller during decay (or virtually by 

 loss of charge) and so pass beyond the scope of exhaustion. But 

 this is improbable ; the experiments show that b increases while the 

 number of nuclei present decreases, no matter whether these re- 

 duced numbers of nuclei are due to w^eak radiation (generating but 

 a few), or to low^ exhaustion (catching but a few), or to the decay 

 of a larger nucleation (where only a few survive in the lapse of 

 time). If — dn/dt = — a ^ en -{- bn-, where a is the number of 

 ions generated per second by the radiation, en the number inde- 

 pendently absorbed per second and bn- the decay per second by 

 mutual destruction, the integrated equation very fully reproduces 

 the observed nucleations n, when b = .000001 and e = .0356. 



Brown University, 

 Providence, R. I. 



