THE CORROSION OF IRON AND OTHER METALS 653 



iron atoms or ions enter into solution in exchange for the 

 hydrogen ions, hence the appearance of iron in the solution 

 in the experiments referred to. 



The answer to these contentions is that no one has yet 

 dealt with pure water and that it is practically inconceivable 

 that any one ever should. 



What has been done is to take ordinary water and purify it, 

 after boiling out dissolved volatile " impurities," by repeated 

 distillation ivi glass vessels : as purification proceeds, the power 

 of the liquid to conduct an electric current diminishes and 

 becomes very, very slight ; it is only logical to assume that 

 if pure water could be obtained, i.e. water free from every 

 other substance, it would have no conducting power. As even 

 the hardest glass is to some slight extent attacked by water, 

 it is inconceivable that pure water should ever be obtained at 

 all events by the use of glass vessels ; and if vessels, say of 

 iridium — the least attackable of metals — were used, there 

 would still be the difficulty, again all but insuperable if not 

 entirely so, of getting rid of surface impurities derived from 

 the environment during working. The conductivity test is 

 of such delicacy that the most minute impurity tells. 



No " electrolytic theory " can apply to water if water be 

 a non-conductor, the essence of the electrolytic theory being 

 that action takes place because an electrolyte is present which 

 is resolved by the passage of the current through it into its 

 ions, one of which at least is active towards the electrode at 

 which it is delivered. 



That two substances, such as iron and water, neither being 

 an electrolyte, should interact (without the intervention of a 

 third) is improbable, to say the least, also for another reason. 

 When two conductors are brought into contact, it is well known 

 that they are at a different electrical potential at the junction : 

 years ago there was much discussion as to the origin of this 

 difference — whether it might be regarded as an actual contact 

 difference or whether it should not be supposed that some 

 minute amount of chemical change took place at the junction 

 in which a film of moisture condensed on the surfaces brought 

 into contact was concerned. The question has never been 

 settled and there are almost insuperable difficulties in the 

 way of settling it experimentally. But assuming that such 

 an electrical difference exists at the junction of two substances, 



