?68 



NA TURE 



[August 15, 1895 



ordinary- steel really is. reference must be made to some 

 facts recorded in Xati-RE, vol. xli. 1S89, p. 32. An 

 attempt was therein made to show that notwithstanding 

 the importance of the part played by carbon in the 

 hardening of steel, the phenomena of hardening cannot 

 be explained solely by a change m the relations of 

 carbon to iron. The iron itself appears to change its 

 state, and M. Osmond has shown that it probably assumes 

 at least three distinct allotropic forms, which he designates 

 respectively as a, fJ, and y iron. 



The fact that the iron itself may e.\ist in more than 

 one state, brings into prominence the causes which under- 



lie the difference between an ordinary rock-section and 

 that of a metal or metallic alloy. In granite, for instance, 

 as the fused mass cools the quartz, mica and feldspar 

 fall out of solution in distinct crystalline masses : and 

 although the fusibility of the mass, and consequently its 

 structure, may be greatly influenced by the presence or 

 absence of a small quantityof impurity, say two or three per 

 cent, of sodium, still, so far as we know, complications 

 do not arise from allotropy of the constituent elements 

 of the rock. In the case of a specimen of carburised 

 iron the conditions are widely different. It is certain 

 that jone very vital change in the relations between the 



carbon and the iron does actually take place at 650 C. 

 that is to say, at a temperature far below the fusing point 

 of the mass. The decomposition of the carbide of iron, 

 FcjC, may lake place at various rales. Ccmcniite can, 

 for example, under sufTicient pressure, resist decomposi- 

 tion at a lempcralure well above that al which it would 

 ordinarily decompose, and we are confronted with the 

 complications which ensue when carbon is united, not 

 merely with a iron, but with j3 or y iron, so as to form 

 cither FcjS^C or I-'ey'C. 



A few examples will serve to make the method of in- 

 vestigation clear. The effect of annealing slcel is very 



NO. 1346, VOL. 52] 



remarkable. The operation consists in raising the metal 

 to a high temperature and in allowing it to cool slowly. 

 .•\ granular structure is thus developed in the metal, the 

 size of the polyhedral grains being proportional to the 

 temperature to which the metal is raised. If the tem- 

 perature is over 1000 C. the grains of ferrite iron) will 

 be large, while the pcrlite remains outside the grains and 

 arranges itself in the joints rour.d them. Fig. 1 represents 

 a sample of \ery mild steel containing o'i4 per cent, of 

 carbon which had been forged and etched with dilute 

 nitric acid ; while Fig. 2 represents the same steel which 

 has been cooled from an initial temperature of 1015" C. 



In it the ferrite has arranged itself in larger grains than 

 was the case in the tirst section, which had not 

 been raised to nearly so high a temperature before 

 cooling. Now compare this with Fig. 3, which shows 

 the effect of raising the steel to an initial tempera- 

 ture of 960' C, allowing it to cool down to a tem- 

 perature of 770' C, and then cooling it rapidly liy 

 quenching it in water. Microscopic examination shows 

 that the interstitial matter is martensitc, together with 

 some troostitc, while the principal mass is still ferrite in 

 grains. These three specimens, chosen, it should be re- 

 marked, from the eighty-five beautiful photographs given 



4* 





by M. Osmond, serve to show how much the stnu tuie 

 of the same \ariely of steel will vaiy with the thermal 

 treatment to which the metal has been subjected. Fig. 4 

 shows a sample of more highly carburised steel polished 

 with rouge, which presents ;i vermicular surface of ferrite 

 and perlite. 



There would appear lo be no limit In tin- .ipplic atinns 

 of micrographic analysis, as all metals and all alloys 

 maybe subjected 10 its action, ll is known, for instance, 

 that the t|ualities of the copper .illoys are gre.itly modi- 

 fied by the addition of minute (|uantities of deoxidising 

 agents, such as ])li<isphoriis, alinniiiium, or silicon, ;mn 



