THE RUSTING OF IRON 415 



bubble of air passed the bend, the upper surface of the iron 

 was exposed to the oxygen whilst the lower surface remained 

 immersed in water. In one experiment, after distillation an 

 amount of air was passed during five weeks which thirty times 

 exceeded the quantity required completely to convert the iron 

 into ferric oxide : not even one speck of rust appeared. To 

 show that the treatment with chromic acid had not in any way 

 modified the tendency of iron to rust in presence of carbon 

 dioxide, in a similar experiment in which iron had been exposed 

 to air and water during three weeks, the glass tube was cut at 

 the point Q immediately above the distilling flask. Air con- 

 taining carbonic acid, cleaned by previous passage through a 

 tower containing pumice stone moistened with water, was then 

 passed slowly through the apparatus. After six hours the surface 

 of the iron was distinctly tarnished ; after seventy-two hours 

 the whole of the surface of the metal was corroded and a 

 considerable quantity of red rust had collected in the bend 

 of the tube. 



These experiments show beyond any doubt that purified 

 oxygen and purified water are insufficient to cause the rusting 

 of iron. The view that iron, pure oxygen and water interact 

 directly to form ferrous oxide and hydrogen peroxide is thus 

 incorrect, so that the theory of rusting based on this assumption 

 becomes untenable. The peculiar behaviour of substances which 

 are known to arrest the rusting of iron cannot be attributed to 

 their power of decomposing hydrogen peroxide but must be 

 assigned to another cause. This question has been investigated 

 by Dr. Moody with especial care. The action of alkalis, such 

 as lime or caustic soda, is at once explained by their power of 

 destroying carbon dioxide. The influence of salts of weak acids 

 such as sodium nitrite and potassium ferrocyanide is also due 

 to their power of absorbing carbon dioxide, the volatile acids, 

 nitrous and hydrocyanic acid, being liberated. In presence of 

 sodium nitrite, ferrous oxide is not oxidised to the ferric state. 

 Weak solutions of chromic acid, which rapidly decompose 

 hydrogen peroxide, do not prevent iron from rusting rapidly, 

 whilst potassium iodide, which also decomposes hydrogen 

 peroxide, not only does not inhibit rusting but actually accele- 

 rates it. 



It has also been made clear that during the formation of 

 rust under natural conditions a large proportion of ferrous 



