124 MB. P. C. CALVERT AND MR. R. JOHNSON ON THE 



slowly ; therefore, about two hours are required to burn all 

 the carbon existing in about three grammes of cast iron. 

 By this method, two analyses of the same sample seldom 

 presented a greater difference than 0.005. We also always 

 took the precaution to dissolve the oxide of iron obtained, 

 after combustion, in order to see that no hydrogen was given 

 off and consequently no metallic iron remained. 



Silicium. 



There is considerable difficulty in determining with pre- 

 cision this element in pig iron, and it was only after several 

 fruitless trials of various processes that we adopted the follow- 

 ing, which gave us very satisfactory and concordant results. 



Five grammes of pig iron were dissolved in aqua regia con- 

 taining excess of nitric acid, the whole was then evaporated 

 to dryness and fused in a platinum crucible with three times its 

 weight of a mixture of pure carbonate of potash and carbonate 

 of soda. The mass obtained was dissolved in water and boiled 

 with aqua regia until the whole of the peroxide of iron had 

 entered into solution, and then was evaporated a second time 

 to dryness and heated carefully to about 200° C. The mass 

 was then heated with hydrochloric acid and water, and the 

 silica, on being gathered on a filter, was washed with dilute 

 hydrochloric acid until it was perfectly white ; it was then 

 dried, and calcined, and its weight gave the amount of 

 silicium in the iron analysed. 



Sulphur. 

 In consequence of the small proportion in which this 

 element exists in pig and wrought iron, considerable diffi- 

 culty is experienced in ascertaining with accuracy the various 

 proportions of sulphur existing in iron; and this difficulty 

 was increased by the fact that none of the methods recom- 

 mended gave satisfactory results in our hands. Thus, for 

 example, the method which consists in determining the sul- 



