99 
one of these samples was supplied to me by Sir Charles Fox, 
as taken from the outside of Conway Bridge, the other 
secured by myself at Llangollen, North Wales. These speci- 
mens gave the following results when submitted to analysis: 
Conway Bridge. Llangollen. 
Sesquioxide of Iron 93 - 094 92-900 
Protoxide of Iron.. 5 -8 10 6*177 
Carbonate of Iron 0-900 0-617 
Silica 0*196 0-121 
Ammonia Trace Trace. 
Carbonate of Lime 0-295 
These results clearly show the correctness of Sir Charles 
Fox’s foresight, that the composition of the rust of iron is 
far more complicated than is stated in our text books. 
Therefore the question may be asked, is the oxidation of iron 
due to the direct action of the oxygen of the atmosphere, or 
to the decomposition of its aqueous vapour; or does the very 
small quantity of carbonic acid which it contains determine 
or intensify the oxidation of metallic iron? To reply to 
it I have made a long series of experiments, extending 
over two years, and which I hope will throw some light on 
this very important question. 
Perfectly cleaned blades of steel and iron having a gutta 
percha mass at one end, were introduced in tubes which 
were placed over a mercury trough, and by a current of 
pure oxygen conducted to the top of the experimental tube, 
the atmosphre was displaced, and it was then easy to intro- 
duce in these tubes traces of moisture, carbonic acid, and 
ammonia, 
After a period of 4 months the blades of iron so exposed 
gave the following results: — 
Dry Oxygen 
Damp „ 
Dry Carbonic Acid 
Damp „ 
.No oxidation. 
In three experiments only one 
blade slightly oxidised. 
No oxidation. 
Slight appearance of a white 
precipitate of the iron, found 
to be carbonate of iron. Two 
only out of six experiments 
did not give these results. 
