554 TRANSACTIONS OF SECTION B. 



evolution of gas ceased. The insoluble matters, consisting mainly of carbides 

 and phosphides, were filtered off, washed and dried, and were ground down 

 in an a^ate mortar to a still finer powder, so as to liberate any mechanically 

 entangled sulphides. The powder so dealt with was again treated with acid as 

 before, after which the residue was filtered off, thoroughly washed with water, 

 was transferred to a separate vessel, and was boiled with strong caustic-potash to 

 dissolve any decomposition products. 



The residue was again filtered off, was washed and dried, and submitted to 

 analysis. The residue when dried weighed about 45 per cent, of the original 

 metal, and contained as follows : — 



Per cent. 



Iron . 9243 



Carbon 



Silicon 



Sulphur 



Phosphorus 

 Water, &c. 



606 



012 



012 



097 (62 per cent, phosphide of iron) 



030 



10000 



A second trial was made with the same metal ; but, in this case, repounding and 

 acid treatment were repeated three times, so as to eliminate the possibility of 

 mechanical inclusion of sulphide or iron. The sulphur found in the remaining 

 carbides was 0"1 per cent. 



As the manganese in this metal was not sufficient to form manganese sulphide 

 with the sulphur, it seemed desirable to determine whether or not when the 

 manganese is in sufficient quantity sulphur would crystallise with the carbide. 

 For this purpose the white chilled part of a crushing roll was experimented upon. 

 The centre part was open grey iron, and contained 3 - l per cent, of the carbon as 

 graphite. 



The white chilled portion contained : — 



Per cent. 

 Combined carbon ......... 3'75 



Graphitic carbon ......... Trace 



Manganese .......... 0'65 



Silicon 070 



Sulphur 0-10 



Phosphorus 023 



It was crushed to powder and treated exactly in the same way as previously 

 described for the separation of carbide. The residue contained by analysis : — 



Per cent. 



Silicon 0028 



Sulphur 0016 



a result showing that only a minute quantity of sulphur was crystallised with the 

 carbide. Whether a different result would follow if both sulphur and manganese 

 were greatly increased has yet to be determined. 



Having proved that sulphur in some undetermined 6tate of chemical combina- 

 tion does crystallise with carbide of iron, an attempt was made to determine the 

 maximum amount of that element the carbide will retain under the most favour- 

 able conditions. With this object in view a considerable quantity of very pure 

 white iron, containing only traces of silicon, sulphur, and phosphorus, and 3'5per 

 cent, of carbon, wa6 melted in a plumbago crucible; and when in a molten con- 

 dition sticks of roll sulphur were forced under the surface of the metal, and 

 afterwards the mixture was briskly shaken up with the sulphur which had 

 liquefied on the surface. 



Precisely the same result was obtained as described by Karsten, who had made 

 a similar experiment. A metal was produced having a white fracture and large 

 cleavage faces. The micro-structure was similar to that of hypereutectic iron 

 carbon alloys. Large plates of carbide cut the metal in many directions, whilst 

 between the carbide plates was located the triple carbide-sulphide-pearlite eutectic, 

 so accurately described by Mr. Donald Levy. 



The carbide plates themselves were peculiar in having circular prismatic h> 



