494 Professor H. C. H. Carpenter [Marcb 7, 



that during quenching fugitive and unstable varieties are formed, 

 one of which may be ^ iron. I cannot help thinking that this 

 answer would never have been thought of if the equilibrium diagram 

 had been well established before M. Osmond advanced his theory. 

 The fact is, however, that Avhen the allotropic theory was put forward 

 knowledge of the diagram was very far from being as complete as it 

 is to-day. 



In the second place, this theory can be tested by experimental 

 work published nearly six years ago by Messrs. Rosenhain and 

 Humfrey.'' These investigators carried out a research on the 

 tenacity of mild steel containing O'l per cent carbon between 1080° 

 and 612° C. This material contained therefore a, /? and y iron. 

 Strictly speaking, their research does not test hardness in the 

 ordinarily accepted sense of the term, but it cannot be denied that 

 determinations of tenacity do test the cohesion of a material which 

 may be properly regarded as intimately connected with hardness. 

 The authors found what they describe as two very well marked 

 discontinuities in the temperature — tenacity curve, corresponding 

 to the critical points Ag and A2 ; but whereas the discontinuity at A3 

 was large that at A., was quite small, and their results must be 

 regarded as negativing the view that (3 iron is harder than 7 iron. 

 In fact, if their curve be examined it will be found that the iron has 

 a decidedly greater tenacity just above Ar^ than just below it, even 

 though a decrease in tenacity would be expected owing to the higher 

 temperature. At 950° C, where their iron was in the 7. range, the 

 tenacity Avas markedly greater than at 850° C, where it was in the 

 /? range. It is very interesting that these results of Rosenhain and 

 Humfrey fit in better with the original theory of Osmond, which was 

 really a 7 iron theory, than with the later one. 



It must also be regarded as improbable that Beta and Alpha iron, 

 which, apart from the fact that the former is non-magnetic and the 

 latter magnetic, only exhibit slight diiferences in properties, and, in 

 particular, between whose crystalline forms no diiference has yet been 

 discovered, should differ so strikingly in one particular property — 

 hardness — such as is demanded by the above theory. From the work 

 of Sir George Beilby, Avhich will be subsequently quoted, it would 

 appear that two varieties of a metal possessing the same crystalline 

 form would be likely to have the same hardness. 



The carbonist theory of hardening, as the name implies, throws 

 the chief, in fact the sole, emphasis on the function of carbon. It 

 owes its inception to the pioneering work of Sorby, of Sheffield, who 

 made a very careful study of the micro-structures of quenched and 

 annealed steels. On this view, the hardness of quenched steel is 

 to be ascribed to the conversion of the pearlite, stable below Ar^, 

 into an amorphous constituent called hardenite, which corresponds 



* Journal of the Iron and Steel Institute, No. 1, 1913, pp. 219-268. 



