August 15, 1895] 



NATURE 



367 



thorough knowledge of the inhabitants, whether animal or vege- 

 table, of oceanic islands. The work must be done speedily, or it 

 will be too late ; and it is work that can hardly be undertaken 

 on a sufficiently extensive scale without aid from Government. 

 Haileybury College. F. W. Headley. 



MICROGRAPHIC ANALYSIS. 



MEr.-\LLURGIST.S would have been greatly aston- 

 ished if they had been urged at the beginning of 

 the present century to gather information as to the com- 

 position of samples of iron and steel by merely looking at 

 polished and etched specimens through a microscope. 

 The operation is, nevertheless, rapidly taking its place in 

 the ordinary routine of a works laboratory. 



As regards the history of the development of this new- 

 branch of investigation, it appears that micro-metal- 

 lography has not been developed from petrography. It is 

 the natural extension of the study of meteoric iron, and, 

 as has often happened in the history of science, it 

 seems to have had more than one independent origin. 

 Priority of date rests with our own countryman Dr. 

 Sorby. In 1864 he submitted to the British .Association 

 photographs of opaque sections of various kinds of iron 

 and steel, and he endeaxoured to develop a method for 

 the industrial examination of such sections under high 

 powers, preferring polished sections to fractured surfaces. 

 The abstract of his paper is very brief; but looking back, 

 it seems strangely comprehensive and suggestive. He 

 claimed that the sections showed "various mixtures of iron, 

 two or three well-defined compounds of iron and carbon, 

 of graphite, and of slag ; and these, being present in 

 different proportions, and arranged in various manners, 

 give rise to a large number of varieties of iron and steel 

 differing by well-marked and \exy striking peculiarities 

 of structure.'' 



Later, Prof. Martens, in Berlin, without neglecting the 

 examination of sections, carefully studied, in 1878, the 

 general laws which govern the occurrence and formation 

 of fractures, fissures, blow-holes, and crystalline structure 

 in metals and alloys. His work, therefore, presents all 

 the characteristics of perfect originality. It was not long 

 after the publication of Martens' work that M. Osmond, 

 then engineer at the Creusot Works, began, with his 

 colleague .M. W'erth, investigations on the cellular 

 structure of cast steel. This work was published by the 

 Actidi'mic dcs Sciences in 1885, and in order to trace the 

 progress which has been made in micro-metallography 

 during the past ten years, it would be difficult to do 

 better than consult the beautiful monograph by M. 

 Osmond which has recently been published by the Societe 

 if Encoiirtigcincnl of Paris.' 



As .M. Osmond justly observes, metallography should 

 in its early days be descriptive ; it should enable 

 us to determine the form and nature of the various 

 constituents of alloys, to ascertain their mode of dis- 

 tributi(m, and to measure their dimensions. Later on, 

 when sufficient data have been established, it will be 

 possible to apportion the observed facts to their respec- 

 tive causes 1 1 ) by ascertaining the way in which the 

 structure of a given metal changes under the influence 

 of the three combined factors — temperature, time, and 

 pressure, and (2) it will be possible to trace the relations 

 l)ctween the observed facts and their consequences by 

 defining the mechanical properties which correspond to 

 a particular structure. 



The first step in the complicated procedure is to cut 

 and polish the opaque specimens of steel. The methods 

 do not admit of condensed description, and the original 

 memoir must be consulted, as even the technical manuals 

 of crafts, in which the polishing of metals plays a part, 



* " Mclhode giincrale pour TAnab'sc microgr.'xphiquc des acicrs au 

 * .irlmne," p.ir M. K. Osmond (/>'»//. tie ia Soc. d' Encouragement, vol. x. 



p. 480. 1S95). 



NO. 1346, VOL. 52] 



give but little information that is useful in the preparation 



of metallic sections for the microscope. It must, how- 

 ever, be added that one method of polishing is specially 

 designed with a view to wear away the softer constituents 

 of the specimen, and bring the harder into relief. It is 

 often useful to attack a polished specimen of steel with 

 a reagent which will colour certain constituents only. 

 For this purpose M. Guillemin treats sections of bronze 

 by oxidation, at regulated temperatures, which produces 

 varied colourations on several constituents of the alloy, 

 while M. G. Charpy prefers an electrolytic attack. It is 

 somewhat surprising to find that an infusion of coco (a 

 popular French term for an infusion of liquorice) is very 

 useful for the purpose, which recalls the fact that Japanese 

 artificers have, for centuries, used plum-juice vinegar, 

 decoctions of finely-ground beans {Glycine hispida), or 

 extracts of the roots of certain plants, as valuable agents 

 for colouring the peculiar alloys which they employ in 

 art metal-work. It may be that the micro-metallographer 

 has much to learn from the Japanese. 



The "attack" of polished specimens is made by suit- 

 able reagents, which may be divided into the three classes 

 — acids, halogens, and salts. Of the acids, nitric acid of 

 36" Baume appears to be the most useful. Of the halogens 

 the pharmaceutical tincture of iodine gives excellent 

 results, as it removes carbon from the steel, and colours 

 certain portions of the specimen. Such treatment, the 

 nature of which has been so briefly sketched, will serve 

 to reveal the main constituents of steel. These are five 

 in numljer, and it has been found convenient to give 

 mineralogical names to them, following the suggestion of 

 the distinguished .American metallurigist, Mr. Howe. 

 Thus pure iron is called /jvvv'/cy the carbide of iron, Fe^C, 

 of .Abel, cementiie. This is not coloured by the infusion 

 of coco or tincture of iodine, which latter leaves it of a 

 silver-white brilliancy under vertical illumination. Dilute 

 nitric acid in the cold does not affect ccmentite. The third 

 material is one of the components of the " pearly 

 constituent of Sorby," which may be coloured by coco 

 or by iodine, and M. Osmond proposes the name 

 of sorbite for it, though he is uncertain as to its exact 

 constitution. The fourth constituent, to which he 

 gives the name of inartensite, is that which is ordin- 

 arily obtained by the rapid cooling of a specimen of 

 steel during the familiar operation known as "harden- 

 ing." It is a crystalline, fibrous substance which iodine 

 colours readily either yellow, brown or black, according 

 to the amount of carbon it contains. Now, martensite 

 preserves its characteristic forms equally well in very low 

 carbon-steels which have been hardened, as well as in 

 high carbon-steels which have been subjected to this 

 process. It may be urged, therefore, that martensite is 

 not a carbon-iron compound which has liquated out of the 

 mass, but that it represents the crystalline organisation, 

 formed under the influence of carbon by one of the allo- 

 tropic forms of iron. 



The last of the five constituents of steel, marks the 

 transition of soft iron into hardened steel. The name of 

 troostite is after the eminent chemist, and it resembles 

 sorbite, but its composition is as yet uncertain. This 

 name is not well chosen, as a variety of silicate of zinc 

 has long been known as troostite. 



It will be evident that a micro-section of a mass of 

 steel closely resembles a rock-section which has con 

 stituent minerals distributed through it. It should, how- 

 ever, be pointed out that there are cases in which the 

 existence of these several constituents cannot be sharply 

 defined, as it is frequently necessary to deal with transi- 

 tion forms which defy classification. Sorbite, troostite, 

 and martensite appear to be solidified solutions of 

 various forms of carbon in diverse forms of iron, for it 

 seems clear that metallographic work on steel brings into 

 prominence the existence of allotropic forms of iron.' 



In order to realise how complicated the structure of 



