ZOOLOGY AND BOTANY, MICROSCOPY, ETC. , 247 



•out some years ago by Osmond and others, was apparently undertaken 

 by the authors, J. 0. Arnold and A. McWilliam, with the object of 

 verifying or disproving the conclusions then reached. The three steels 

 employed contained respectively 0*21, 0*89, and 1'7H p.c. carbon, the 

 total percentage of elements present, other than iron and carbon, being 

 very small. The critical ranges of these steels, on heating and cooling, 

 were determined. Samples were heated to different temperatures, 

 quenched, and microscopically examined. By quenching a piece of the 

 O^S!) p.c. C steel when it was passing through the Ac. 1-2-3 change, a 

 section showing both pearlite and hardenite was obtained. Dark etching 

 boundary lines between these two constituents are identified by the 

 authors with Osmond's troostite. The authors give their reasons for 

 declining to accept martensite, troostite, sorbite, and austenite as con- 

 stituents ; they do not object, however, to the terms " troostitic struc- 

 ture," " martensitic structure," etc. In an appendix, definitions of the 

 constituents of steel are given. Fe 3 C of cementite and Fe 3 C of pearlite 

 are stated to be physically different substances. Excellent micrographs, 

 which have apparently been drawn and not photographed, illustrate the 

 paper. 



Important contributions to the discussion were made by J. E. Stead 

 and H. le Chatelier. J. E. Stead expressed the opinion that the 

 evidence in favour of the allotropic states of iron was now overwhelming. 

 The term " eutectoid," as applied to steel of pearlite composition, was to 

 be preferred to " saturated." Martensite, as a homogeneous solid solution 

 •of carbon or carbide of iron in iron, should be recognised as a constituent. 

 H. le Chatelier accepted the authors' experimental results as being in 

 harmony with previous observations, but disagreed with some of the 

 theoretical explanations put forward. The alleged insolubility of 

 •cementite in hardenite up to 900° C. was improbable. The claim of 

 troostite and austenite to be recognised as constituents is defended. The 

 authors' assumptions as to the constitution of these bodies were purely 

 gratuitous, and not supported by any evidence. The explanation of the 

 separation of graphite in high carbon steel was contrary to well known 

 facts of chemistry. An alternative explanation is advanced. The 

 authors of the paper replied to these criticisms at some length. 



The Presence of Greenish-coloured Markings in the Fractured 

 Surfaces of Test-Pieces.* — H. G. Howorth has examined, microscopi- 

 cally and otherwise, a large number of defective test-pieces from gun 

 forgings. The defects were due to the presence of foreign matter, 

 resembling Siemens-Martin slag, usually yellow or green in colour. Two 

 substances, which appear to be sulphide and silicate of manganese, were 

 distinguished in the coloured matter. 



The Nature of Troostite.f — C. Benedicks, regarding this constituent 

 as a product of transformation of martensite, dismisses Boynton's 

 hypothesis that troostite is y8-iron, as untenable. Troostite appears to 

 be a pearlite with ultra-microscopically small particles of cementite, con- 

 taining also more or less hardening carbon, and offers an interesting 

 analogy to the colloid solutions. 



* Journ. Iron and Steel. Inst., lxviii. (1905, 2) pp. 301-19 (13 figs.). 

 t Tom. cit., pp. 352-70 (2 figs.). 



