OXIDATION OF IRON. 113 



sphere ; but, as carbonates and even bicarbonates act in a 

 similar manner, it is evident that this view of the action 

 was erroneous. • 



The results above stated not coinciding with those ob- 

 tained by Payen, and given in Gmelin (vol. v. p. 185), 

 I repeated his experiments, and found that the portions of 

 the blades of iron immersed in dilute alkaline solutions as 

 described by him are preserved from oxidation, but that 

 the portion out of the liquor becomes covered with oxide, 

 which increases in inverse ratio to the strength of the 

 solution, and that when the proportion of 1 part of the 

 alkali to 2000 parts of water is reached, the oxidation of 

 the iron proceeds rapidly on that part of the blade just 

 above the surface of the water. 



This series of experiments confirms the opinion already 

 expressed that the iron exists in a passive condition, and 

 that this condition only attains sufficient power to resist 

 oxidation by atmospheric oxygen when there is 5 per cent, 

 of either hydrates or carbonates of alkalies in solution. 



I have repeated also M. Pay en's experiments on the 

 preservative action of weak solutions of carbonate of soda, 

 and find that my results do not correspond with his. He 

 states that 1 part of a saturated solution of carbonate when 

 diluted with 54 parts of water will preserve iron from rust, 

 while 1 with 59 will not. I find that the iron rusts in both 

 cases, thereby proving that stronger solutions are necessary 

 to prevent oxidation. 



I have made many experiments with the view of dis- 

 covering the cause of this interesting preservative action 

 which alkaline solutions exert on iron, but have failed. 

 Knowing the destructive influence which sea- water has on 

 iron, and the serious injury done to iron ships through the 

 action of bilge- water, a series of experiments were made 

 with sea-water to which was added such a quantity of 

 caustic soda or potash or their carbonates that, after the 



SER. III. vol. v. 1 



