An Attempt to analyse Cast Irori. S5^ 



irianganese. This makes an addition in the weight of 

 0'2I5 oxysen. Thus 100 parts of oxide of manganese 

 after exposure to a red heat consist of 70'25 of manganese 

 and 29'75 oFoxj'gen, and 100 parts of metallic manunnese 

 take up 42-16 parts of oxygen in forming this oxide. This 

 so nearly agrees with the red oxide of iron, that, without 

 any remarkable error, we may calculate the quantity of 

 metal bv omitting the trifling portions of manganese 

 oxide, and counting the whole for oxide of iron, as J have 

 done in the first and less successful attempt to oxidize the 

 cast iron with nitric acid. A want of the metal of man- 

 ganese ])revented me from repeating this experiment on a 

 greater quantity, to attain greater precision. But the result 

 as stated is sufficiently accurate fur the calculation we here 

 require. I also must observe, that the oxide of manganese 

 obtained in the analysis of the iron was afterwards treated 

 with nitric acid in a crucible of gold, on which its weight 

 was found not to have altered in the least ; so that the 

 manganese in both instances was in an equal degree of 

 oxidation ; which is a circumstance of great moment in the 

 calculation. 



When from 6*925 of oxide from the cast iron we de- 

 duct 0325 oxide of manganese, there remain &•& grammes 

 for oxide of iron (contaminated with a most minute por- 

 tion of earths), which correspond to 4'5764 of pure metallic 

 iron; and as 0-325 of a granime of oxide of manganese indi- 

 cate 0-2283 of metallic maniianese, they make together 

 4'805 gramtnes of metal. If these be deducted from the 

 five grammes of cast iron employed, there remain 0'195 

 of a gramme for carbon, which during the experiment had 

 escaped in the form of carbonic acid ; and this divided into 

 five gives 3-9 per cent, which is 0'06 per cent, less than in 

 the experiment for oxidizn^ the cast iron with nitre. These 

 experiments may therefore be considered as perfectly cor- 

 roborating each other. They also prove that the carbon 

 in cast iron is not present in the state of diamond, but in 

 the state of graphit, or such carbon as is contained in car- 

 bonic acid to the amount of 28 per cent.j otherwise the 

 loss in the last calculated experiment ought to have been 

 Tifiuch less than the quantity of carbon in that with nitre. 

 That it turned out greater in this experiment is, because a 

 loss could not take place in it but from erroneous data for 

 calculating the quantity of metal. 



In order to prove decidedly that this cast iron contained 

 neither su!i)hur nor phosphorus, 1 dissolved five grammes 

 thereof in pure nitric aciil ; the solution was correctly sa- 

 ■ Vol,4(V. No. 174. Oct. 1612. K tu'ratcd 



