325 



for the alcohol with the same effect, but in this case some of the 

 sesquioxide will remain unaltered and mixed with the black oxide. 

 The alcohol or ammonia is correspondingly changed by oxidation 

 derived from the oxygen which has been released from combination 

 with the iron. If the hydrated Fe 3 O 4 be heated in contact with air, 

 it immediately (even when it has been kept for many months) be- 

 comes Fe 2 O 3 by oxidation from the atmosphere, but if heated to red- 

 ness in vacuo, it cools unchanged. [Can the black powder of alu- 

 mina be A1 3 O 4 , formed in a similar way ?] The process of catalysa- 

 tion by Fe 2 O 3 is thus evident ; the heated ses.quioxide loses a por- 

 tion of oxygen to the alcohol and becomes Fe 3 O 4 , which is instantly 

 reconverted into Fe 2 O 3 by receiving oxygen from the air, and this 

 alternation is constantly going on in every portion of the glowing 

 mass. It is not a mere action de presence, but alternate reduction 

 and oxidation of the sesquioxide, producing a continuous oxidation 

 of the alcohol. 



This suggests a consideration apparently adverse to the atomic 

 hypothesis of Dr. Dalton. How can a single compound molecule 

 Fe 2 O 3 be changed by deoxidation into another compound molecule 

 Fe 3 O 4 , when, according to theory, there are in it but two combining 

 proportions of iron, whereas the resultant contains three ? and how 

 (by deoxidation) can the resultant molecule contain four combining 

 proportions of oxygen when the primary contained only three ? We 

 can indeed represent to the imagination that three molecules of the 

 sesquioxide, acting as if they were one triple molecule, lose one com- 

 bining proportion of oxygen, and are converted into two molecules 

 of the black oxide ; and conversely, that two molecules of the black 

 oxide, acting as if they were one double molecule, combine with one 

 atom of oxygen, and are converted into three atoms of sesquioxide. 

 The only way to account for this, in accordance with the popular 

 atomic theory, seems to be, to assume that the notation for these 

 oxides is incorrect, and that 



for Fe 2 O 3 we should write Fe 6 O 9 , 

 and for Fe 3 O 4 we should write Fe 6 O 8 . 



If the current notation be retained, and any law be admitted, in 

 virtue of which three molecules of sesquioxide may suffer reduction 

 as if they were only one molecule, and divide into two molecules of 

 the magnetic oxide, we might conceive a peculiar structure in the 



