IRON RUST. 317 



that its magnetic properties remained unimpaired. The next 

 question to determine was whether it was still a combina- 

 tion of the two oxides having undergone some molecular 

 change, or whether the protoxide had not become converted 

 into peroxide by absorption of oxygen from the atmosphere. 

 I therefore dissolved some of the red powder in hydro- 

 chloric acid, and to tlie solution added some ferridcyanide 

 of potassium, but not a trace of the protosalt was indi- 

 cated by that reagent. The magnetic peroxide may also 

 be conveniently prepared by fusing together for some minutes 

 the common l)lack oxide with three or four times its weight 

 of nitrate of potash. Prepared in this way it is more readily 

 dissolved by hydrochloric acid than that obtained by simple 

 ignition. 



" I have searched various chemical works, but can find no 

 mention of a magnetic peroxide of iron. I therefore pre- 

 sume that this peculiar compound has escaped observation." 



M. Sarzeau (Chemical Neivs, I., 1860, p. 137, and Jour, 

 de Pharm. et Chemie, 23 Jan. 1860), in describing a method 

 of preparing aerated water from plates of iron placed in a 

 solution of carbonic acid in water at 50° F., says : — " We 

 obtain a liquid with well marked inky flavour, and which has 

 all the properties of a solution of a protosalt of iron. If 

 the liquid be heated it is at first turbid, and then becomes 

 milky, afterwards ochrey, and at last when it boils it suddenly 

 changes to a deep brown. On cooling it deposits a substance 

 of the same colour, which, separated by filtration, dried in 

 the air, and dissolved in hydrochloric acid, gives a mixture of 

 protochloride and perchloride of iron. If a magnet be 

 brought in contact with the brown matter when dry it is found 

 to adhere, showing that by heating a solution of carbonate of 

 iron magnetic oxide is formed." 



I did not come across the accounts of the experiments by 

 Mr. Robbins and M. Sarzeau until my own were practically 

 completed, as they are not quoted in any of the Chemical 

 Dictionaries and works of reference, but I have since 

 repeated them, and can confirm their results. 



Robbins' experiment of heating finely-powdered magnetite 

 was rejjeated on several samples of magnetite, and in each 

 case the resulting sesquioxide of iron was found to be wholly 

 attracted by a magnet, although no trace of ferrous oxide 

 could be detected in the red powder b}^ dissolving in hydro- 

 chloric acid and adding potassium ferricyanide. 



