1028 IRON 



of ammonium, vanadate of ammonia is deposited as a white or yellow crystalline 

 powder. To test for chromium only, the mass after fusion with nitre is extracted 

 with water, and then boiled with carbonate of ammonia ; the solution is neutralised 

 with acetic acid, and then acetate of lead added ; the production of a yellow precipi- 

 tate indicates chromic acid. 



Aluminium. This metal is best separated from iron, by first reducing the latter to 

 the state of protoxide by sulphite of ammonia, then neutralising with carbonate of 

 soda, and afterwards boiling with excess of caustic potash, until the precipitate is 

 black and pulverulent. The solution is then filtered off, slightly acidulated with 

 hydrochloric acid, and the alumina precipitated by sulphide of ammonium. 



Calcium and Magnesium. These metals are found in the solution from which the 

 iron and aluminium have been separated ; they both exist probably (together with 

 the aluminium) in the cast iron in the form of slag, and are best detected in the black 

 residue which is left on dissolving the iron in dilute sulphuric or hydrochloric acid. 

 After digesting this residue with caustic potash, and burning away the graphite, a 

 small quantity of a red powder is left, which is composed of silicic acid, oxide of iron, 

 alumina, lime, and magnesia ; if 500 grains of cast iron are operated upon, a suffi- 

 cient quantity of insoluble residue will be obtained for a quantitative determination 

 of its constituents. 



EggertJs Methods. Eggertz's method of determining combined carbon in iron or 

 steel depends upon the discoloration produced by carbon in solution of pernitrate of 

 iron, which, under ordinary circumstances, is colourless, or at most of a greenish tint. 

 The standard series of colours is made by dissolving quantities weighing 1 diagramme 

 of steel of known composition in nitric acid at low temperature, and diluting with water 

 to a standard volume. The solutions which give different shades of brown are preserved 

 in glass tubes. 



A similar weight of the steel to be examined is dissolved in pure nitric acid under 

 the conditions observed in making the standard series. The solution is decanted from 

 the residue, poured into a burette of the same diameter as the tubes containing the 

 standard series, and diluted with water until it matches one of the tints. The amount 

 of carbon is then found by calculation from the relative volumes of the solutions of 

 steel, with a medium amount of carbon, say 0'8 per cent., gives a yellowish green 

 solution ; a very hard variety, with 1-5 per cent, brownish red; and the softest, with 

 0'40 per cent., only a slight greenish tinge. 



The Swedish classification of Bessemer steel by numbers, based upon the percentage 

 of carbon determined by the above process, is as follows : 



No. 1 contains 2 per cent. ; No. 1'5, 175; No. 2, 1'5 ; and so on, up to No. 4-5, 

 with only 0-25 per cent., below which point the scale is not extended. 



The following determinations of carbon in various kinds of iron and steel made in 

 Sweden are by Eggertz : 



Softest Swedish Bessemer iron contains 0'08 per cent, of carbon 



Soft steel 075 



Best quality of cast steel . . . 1*4 to 1*5 



Natural forge steel .... 0'99 2'44 



Cement steel 0*5 1'90 ,, 



Cast steel 0'36 1-9 



Hardest welding cast steel . T80 



Malleable cast iron .... 0-88 1'52 



Draw plate steel 3*30 ,, 



The following modification of this process is adopted at Seraing, in Belgium, for 

 the determination of carbon in mild Bessemer steel. Two samples, each weighing 

 0*2 gramme, of the ingot to be tested are taken ; one being in the state of filings, 

 and the other of borings. These are treated with 20 cubic centimeters of nitric acid, 

 of specific gravity 1*2; the solution being effected in a water-bath at 80. For the 

 test-standard the same weight of two different samples, of the hardest steel made in 

 the works, whose carbon has been already determined and containing 0'61 and 0'63 

 per cent, respectively, are dissolved up in the same manner. The whole of the four 

 solutions are then brought down to one tint by adding water to the darker ones, care 

 being taken to conduct the operation in tubes of exactly the same diameter. The 

 amount of carbon can then be computed from the volume of the solutions. If tlio 

 difference in the results given by the two samples does not exceed 4 '03 per cent., the 

 arithmetical mean between them is adopted as the true amount; but should it be 

 larger, the operation is repeated. 



In order to obviate the necessity of making standard solutions for every set of 

 determinations, different coloured liquids have been employed with a view of obtaining 

 a permanent scale of colours. Among these may bo mentioned caramel, or burnt 

 sugar, which gives various shades of yellow and brown, but alters very quickly ; and 



