170 



depth of the ioiiisation chamber might become indetermi- 

 nate. 



We must now consider the assumption that the area of 

 the ionisation curve may be represented by the product of the 

 co-ordinates HI as already defined. This is really equivalent 

 to the supposition that the ionisation resulting from the ex- 

 penditure of a quantity of energy 8e by the a particle is 

 equal to /i/fyjSc, where f(vj is a function of the velocity of 

 the particle and k is a constant, depending on the nature of 

 tlie gas molecule. It implies in the first place that the area 

 of the ionisation curve in any gas is not dependent on pres- 

 sure and temperature, and that, if the form of the curve is 

 altered by a variation of these conditions, it is only in so far 

 that all the ordinates are miultiplied by some factor, and all 

 the abscissae divided by the same factor. It implies, in the 

 second place, that the ionisation curve of one gas can be made 

 to coincide with the curve of any other gas, by multiplying 

 all the ordinates by some factor, and all the abscissae by some 

 other factor. Let us examine the evidence in favour of these 

 statements. 



If the hypothesis is true 1\I must be independent of pres- 

 sure and temperature. As regards pressure, some results were 

 quoted in a paper "On the Recombination of Ions in Air and 

 Other Gases,"" which showed this to be correct in the cases of 

 air and ethyl chloride ; and further evidence will be found in 

 the results given at the end of this paper. For, without hav- 

 ing made any exhaustive comparison of the values of BI at 

 different pressures in each gas, I have often used various pres- 

 sures in the determination of the specific ionisation of a gas ; 

 and the general agreement between the results obtained is 

 good evidence that pressure is without effect. 



In the same way, since many determinations in the case 

 of the same gas have been made at different temperatures, the 

 close agreement shows also that temperature has no influence 

 on the ionisation. More direct confirmation can be obtained 

 from the following results. During a number of the deter- 

 minations of UI, the ionisation chamber was connected in 

 parallel with a second chamber containing a uranium layer. 

 The ionisation currents acted against each other. Thus the 

 values of the currents in the radium apparatus could be de- 

 termined by balancing against the uranium ; the latter wa^ 

 always at the temperature of the room and therefore formed 

 a fixed standard. The extent of the surface of the uranium 

 could be varied by means of a semaphore, having^ a graduated 



Trans. Roy. Soc. S.A., 190."), p. 187. 



