]7^ 



Fig. 3. 



It is necessary to refer to one more assumption which is 

 made in calculating the results, viz., that the Rl of a mix- 

 ture can be determined from a knowledge of the lil of each 

 component. For example, it is supposed that, the Rl of air 

 being 100 and of ethyl chloride 132, then the Rl of a mix- 

 ture in such proportions that the a particle spends half its 

 energy in each is 116. 



For I have not been able to prevent the leakage of air 

 into the apparatus when raised above ordinary tempera- 

 tures, and it is necessary to measure and allow for the air 

 present in each experiment. The apparatus holds very well 

 when not heated ; but it is sometimes necessary to raise the 

 temperature to 60° or 70" C. in order to obtain a sufficient 

 density of the vapour under treatment. Fortunately, how- 

 ever, the air present may be a considerable fraction of the 

 gas when measured by pressure, and yet be of little import- 

 ance when measured in terms of the energy spent in it. Thus 

 the correction for air present is usually quite small, as will 

 be seen from a consideration of the numerical results in § 2. 



The assumption is by no means an obvious one. If any 

 part of the ionisation in a gas is secondary, and is due to radi- 

 ation originating in one molecule and acting on a neighbour- 

 ing molecule, it might well be that complications would arise 

 in a mixture of gases. I have made several direct attempts 

 to find whether any such effects existed : the results of some 

 of them are shown in the following tables. Each table refers 

 to a set of experiments carried out consecutively. The per- 

 centage of gas in each mixture is determined from the stop- 

 ping power, and the percentage of energy spent in the gas 

 is then calculated. The value of Rl for tlie gas is calculated 

 by multiplying the observed value for air by the specific ioni- 

 sation of the gas, as taken from the final tables eiven at the 

 end of the paper. For example, in the first set, Rl for air 

 is 198-5, and Rl for ethyl chloride is taken to be 189-5 x 1*32 



