172 



It is further assumed that the curves for different gasee 

 are of the same form ; in other words, that the function 

 f(v) is the same for all gases. 



A complete test of this hypothesis would require an accu- 

 rate delineation of the ionisation curve in the case of each gas. 

 As has already been said, this would be a difficult task, inclu- 

 sive, indeed, of our present purpose. But a comparison of the 

 curves in different gases, so far as they have been obtained, 

 shows that the principle is at least approximately true. For 

 example, the ratio of the range of RaC to that of RaA is the 

 same in all gases within errors of exjDeriment, and again the 

 ratio of the maximum abscissa of the RaC curve to the abscissa 

 / is also constant, so far as I have measured it. As examples 

 of the constancy of the first of these two ratios, I have at 

 different times found it to be 1*46 in air, 1*47 in pentane, 1*47 

 in ethyl chloride, 1*44 in carbon dioxide, 1*48 in ethyl alco- 

 hol, and 1*49 in ethylene. The differences here are probably 

 experimental only. As regards the second ratio, I have found 

 it to be 1*36 in air, 1*37 in ether, 1*44 in ethylene, 1*35 and 

 again 1*41 in ethyl chloride. This ratio is much more liable 

 to error than the former ; for all ionisations are harder to 

 measure correctly than ranges, and the peak of the ionisation 

 curve is an especially uncertain point. Also there is a special 

 difficulty due to the existence of a peculiar phenomenon, which 

 must now be considered. 



It is to be observed that the ionisation curves in different 

 gases will not correspond unless the potential employed is 

 enough to saturate at all points of the path of the a particle. 

 More electric force is required as the particle slows down. This 

 may be deduced from figures given in the paper "On the 

 Recombination of Ions, etc.," p. 196. It is there stated that 

 the ratio of the saturated ionisation current to that at 25 

 volts per cm. in the case of the ions made by the a particle 

 of RaC at a distance of 6*25 cm. from its origin was found 

 to be 1*29 ; whereas, when the distance was reduced to 5*05 

 cm., it was found to be 1*19. Each of these ratios is the 

 mean of four determinations. (By an arithmetical error, one 

 of the latter is incorrectly given in the paper quoted : 1*23 

 should be 1*20.) 



Again, the effect is clearly shown by the curves of Fig. 3, 

 which represent the results of experiments on a mixture of 

 ethyl chloride and air. It will be seen that the curve does not 

 show the characteristic increase of ionisation with distance 

 when the electric force is small, the reason being that it is so 

 much more difficult to collect the ions made by the a particle 

 at the end of its path. 



