All Attempt to analyse Cast Iron. 653 



•f oxide of manganese undissolved, and then the dissolved 

 oxide of manganese precipitated with prussic alkali; after 

 which the magnesia was separated with caustic potash*. 



This analysis gives the following products from the cast 

 iron : 



Red oxide of iron 23' 660 



Oxide of manganese contaminated with magnesia 1-485 



Silica ] • 1 25 



Burnt extractive substance 0-336 



27-606 



If 100 parts of red oxide of iron consist of 30-66 parts 



of oxygen, and 6934 parts of ironf, then these 2566 of 



oxide of iron correspond to 17-85 of pure iron in the jg-i 



grammes of the analysed cast iron. 



If 100 parts of burnt oxide of manganese consist of 29:f 

 oxygen and 70| metaljic manganese, then 1*485 gramme 

 of oxide indicates 1-043 metal of manganese. By adding 

 the manganese to the quantum of iron, then, what is defi- 

 cient in 19-1=0-1 1 of a gramnje will be loss, and carbon. 

 Arranged according to per centage, the component parts 

 consist of Iron contaminated with base of siiex . , 93-8/5 



Manganese , 5*460 



Carbon and loss 0*665 



1000- 00 



The loss exhibited by the foregoing analysis is too small. 



I thereiore thought it probable that a great portion of 



sulphuric acid, notwithstanding the strong ignition, might 



have remained in the oxide of the iron; which determined 



me to ^cp^at the analysis in a different manner. 



1. To ascertain in a pDsiiive manner the quantity of car- 

 bon present in the cart iron, I pulveriztJ a part of the 

 bruised crystals in a cast-iron mortar, and sifted the powder 

 through a t'.nc linen cli.ih. I did not collvci what first went 

 through, and which migiit possibly b(; (kfiicd by dust from 



* Since these experiments were made in the bei^inning of the year 1808, 

 M. Gay-Lu>-sac has proved that ihemost part of ilie metals r.re precipltited 

 from tlieir neutral solutions in acetic ci 1 by incaiif. of a stre.'.in of sul- 

 phuretted liydrogon, which also may be tnppHv empi ved for sepata ing 

 oxide of nianganes" troin mayiiesia and from ilmt. BerihoUct's above- 

 mentioned soluble cotiipoimd of sulphur^.' led h\ dro^i^en ind oxide of man- 

 ganese is not formed when hjdrosulphurel of pi.: -h is mixed with a solu- 

 tion of manc^anese, an. V»uqi!clin*s proposed mt-' idi* of course very ap- 

 plicable to the same p'lip' se — i he Aurnuii 



f Sec the aiihoi » tvxpe:i:neii's on Definite Proportions, Annalex de 

 Cliimu-, ixil, Aijgust; and »\m /ifluindimsar i Ftjiik, Kemi och MineraUigi. 

 ?H. 1810. pag. IJiiO. 



the 



