94 Scientific Proceedings, Royal Dublin Society. 



patches. As the activity decayed, the average interval before detonation increased, 

 as shown below, where a typical set of results is recorded : — 



from 

 specimen, I'l cms. in every case. 



An interval of over an hour elapsed between the fourth and fifth tests, so that 

 the concentration of a particles only averages about one-tenth as great in the last 

 four tests as in the first four. The concentration per square centimetre is nearly, 

 but not quite, proportional to n, as small variations occurred in the areas of the 

 patches. The average value of nt for the first four tests is 4-35 x 10', and for 

 the last four 3"35 x 10', so the evidence here favours the view that nt is 

 approximately independent of the concentration. This result was amply confirmed 

 by other series of observations, though the mean value of nt was generally higher, 

 probably owing to the specimens being less diy, so that some of the iodide was 

 insensitive. A set of tests at distances -varying from 1 to <i cms. indicated an 

 increase in the value of nt with distance, i.e. a decrease in the detonating efficiency 

 of a particles with decrease of velocity. The results are rendered somewhat 

 indefinite by unavoidable differences in sensitivity among the test specimens, and 

 by the large variations in number of particles per explosion, inherent in the effect. 

 The detonating efficiency of an a particle at 1 cm. is apparently about twice as 

 great as tliat at 3 cms. 



We see, then, that all the evidence favours the theory that a single a particle 

 is responsible for the detonation. It is natural to ascribe this to a nuclear 

 collision, probably with either a nitrogen or a hydrogen atom. Such a collision 

 would almost inevitably disrupt the molecule to which the atom belonged, and 

 may even, as Sir Ernest Paitherford has shown, disrupt the nitrogen atom itself. 

 The probabilities involved are quite of the order we should expect. The decrease 

 in the detonating efficiency of an a particle with its velocity also accords well with 

 this view. 



As regards the second question, it is evident that any compound containing a 

 light element would occasionally have one of its molecules disrupted in a stream 

 of a particles. This would appear to be certain in the case of the disruption of a 

 nitrogen atom, and almost certain in any case of nuclear collision. In the case of 

 iodide of nitrogen the disruption of one molecule appears frequently to entail the 

 detonation of tbe whole. We would expect that the probability of detonation of 

 any body sufficiently unstable to be exploded at all by a single molecule would be 

 of the same order of magnitude as that for iodide of nitrogen. 



The following is a brief record of the various substances tested. In no case 

 did detonation occur, though some of the bodies are notoriously sensitive. The 



