49 
obtained, and the corresponding quantity of carbon 
monoxide is formed. 
That this is the case was ascertained by analysing the 
oxychloride and by collecting the gas over caustic potash 
and measuring the volume of insoluble gas which came 
over. This gas was shown to burn with a blue flame, and 
the quantity was found in two experiments to correspond 
closely with the theoretical amount. The following is the 
composition of the oxychloride thus prepared : — 
Found. 
Niobium 43 '48 
Chlorine 47 '03 
Oxygen 9 ’49 
100-00 
A decomposition of carbon dioxide by a metallic chloride 
has hitherto not been observed. Phosphorus trichloride 
absorbs free oxygen on distillation in the air, with formation 
of oxychloride, but it does not decompose carbon dioxide, 
“On the Retention of Saline Impurities by Hydrated 
Ferric Oxide,” by Mr. Charles Frederick Cross, Dalton 
Scholar in the Owens College. Communicated by Professor 
H. E. Roscoe, F.R.S. 
It is a fact well-known to all chemists that ferric oxide, 
when precipitated in the form of hydrate from solutions 
which contain either organic matter or the fixed alkalis, 
carries down a certain quantity of these substances in adhe- 
sive union. 
That this adhesive action is exerted upon saline bodies in 
solution, although an obvious inference, I have found 
nowhere stated to be the case. The importance of the 
analytical bearings of such a property, if it do exist, led me 
to undertake a quantitative investigation of the matter, the 
results of which I proceed briefly to detail. 
I. Two grams fine iron wire were dissolved in HC1, the 
solution oxidized with HN0 3 , the excess of acid driven off 
