310 Rev. P. J. Kirkby : Chemical Effects of the 



considerable part in these reactions. This would be the case 

 if the rate at which water was formed depended to an 

 appreciable extent upon the amount of water- vapour present, 

 for then the observed reaction would depend to some extent 

 upon the rate at which the apparatus permitted the water- 

 vapour to diffuse away from the region of the discharge. 

 Now on our hypothesis the chemical reaction resulting from 

 the motion to the electrodes of each pair of ions is very 

 likely to be affected by the amount of water-vapour present, 

 for water-vapour condenses on ions of both kinds (J. J. 

 Thomson, ' Conduction, &c./ pp. 153, 154), retarding their 

 motion and increasing their effective mass. The res alt of 

 this condensation may be to increase the chemical reaction : 

 an impinging ion maybe more likely to dissociate a molecule 

 if the mass of the ion is thus increased up to a certain 

 point even though its velocity of impact is diminished. 

 Perhaps this is the reason why the final slope of the curve 

 (j) = 4"5 mm.) of fig. 2 is not so uniform as that of the 

 curves corresponding to the pressures 2*1, 1*05 mms. : the 

 slight increase in the slope may be due to this, that as 

 the positive column is extended the ions which traverse it 

 have a greater chance of attaching water-vapour to them, 

 for all but a small percentage of the ions in the positive 

 column pass through its whole length. This effect, however, 

 would tend to disappear at the lower pressures owing to the 

 greater speed of the ions. 



In view of these considerations and of the discrepancies 

 alluded to above, the observations recorded in this paper 

 cannot claim to be independent of the form of the apparatus 

 except very roughly. Nevertheless, they agree among 

 themselves with great precision, and therefore furnish an 

 accurate record of the amount of chemical action which took 

 place in the present apparatus in the various circumstances 

 of the discharge. 



Influence of the Current. 



It was shown in the preceding paper that Ap/AQ is 

 independent of the current for small values of the distance 

 between the electrodes. This, however, should cease to be 

 true when that distance is large, for then the distribution of 

 electric force may be so modified by a large change in the 

 eurrent that the chemical reaction is appreciably changed. 

 This conclusion was confirmed by observation. Thus, when 

 the distance between the electrodes was 16 cms. and the 

 current *0015 ampere was used, Ap/AQ proved to be 5*5 



