254 



NATURE 



[January 14, 1892 



Consequently, the physical properties of the metal are 

 modified in accordance with the relative proportions of 

 each present. 



It is thought that pure iron may contain both o and ^3 

 iron ; but it is certain that when alloyed with other bodies 

 —more especially carbon— these allotropic changes of 

 iron become very marked. In this connection the author 

 noted long ago that the carbon-iron alloys were more 

 liable to change, and more sensitive to variations of tem- 

 perature, than iron alloyed with other bodies ; and, as 

 the results of experience given, he goes so far as to advo- 

 cate the substitution of other than carbon-iron alloy, in 

 the following words : — 



" For the production of steel in large masses suitable 

 for ship and boiler plates, rails, &c., the traditional rules 

 of the old school of steel-casters cannot well be applied. 



" What is desired is a strong ductile material, capable 

 of resisting sudden shock or impact, and sudden or ex- 

 treme changes in temperature— a material as insensible 

 as may be to all influence except that due to fair wear. 

 This material — i.e. low carbon, Bessemer, or Siemens 

 steel — has almost superseded wrought-iron, and yet is not 

 always to be relied upon. Under certain ill-understood 

 conditions this metal sometimes behaves in a manner 

 which has not been satisfactorily explained. 



" Must we infer that carbon may be the culprit, and 

 that the carbon-iron alloy is more sensitive to external 

 influences, and more liable to molecular changes, than 

 other iron alloys free from carbon ? ^ 



" It is quite possible that further study and experiment 

 may result in the production of an iron alloy, capable 

 alike of being forged or cast in amorphous masses free 

 from any tendency to hardness or temper." 



Osmond's researches have deservedly attracted great 

 attention, and, in conjunction with the work of Prof Ball 

 and Prof. Roberts-Austen, have on the whole been ap- 

 proved and accepted by men of science. 



^yith the aid of the Le Chatelier's pyrometer, the 

 critical points of temperature, i.e. the points at which 

 molecular changes take place, have been determined. 

 The whole series of these masterly researches cannot 

 here be given in detail ; but after careful consideration 

 the writer thinks they go far to prove that the undoubted 

 molecular changes which take place when iron is heated 

 simply represent marked periods of dissociation. That, 

 broadly speaking, the results are in accord with the 

 periodic law ; with the spectroscopic work of Mr. 

 Lockyer, and the researches of Mr. Crookes, together 

 with the author's research on the behaviour of iron at 

 varying temperatures ; to say nothing of the work of other 

 chemists all pointing in the same direction. As before 

 said, the physical properties of iron are a function of 

 temperature ; indeed, one is fain to say that the absolute 

 elementary body can only exist as such at a given tem- | 

 perature ; at any other, it cannot, strictly speaking, be | 

 that body. ! 



Appreciable dissociation can, however, only be noted \ 

 at comparatively wide intervals of temperature ; minor i 

 changes must, of course, be beyond recognition to our 

 senses. It is possible that the first beginning, so to 

 speak, of dissociation may be detected with the spectro- 

 scope ; but unless carefully worked, the first spectroscopic 

 indications of dissociation are not altogether trustworthy. 

 The temperature of the spark and also the arc is sub- 

 ject to irregularities not easily controlled ; and unless 

 worked by the experienced spectroscopist, the results are 

 apt to be unsatisfactory. 



It is remarkable that many eminent chemists reject the 

 spectroscope as a means of research, and prefer other 

 methods. Others, again, consider that these methods of 



' Practically, we know that it is so; only pure carbon-iron alloys can be 

 tempered and hardened with facility, the presence of other bodies tending 

 to neutralize this effect ; or, more plainly, the steel-smith terms it bad steel, 

 which cannot be tempered or hardened properly. 



NO. I 159, VOL. 45] 



procedure fail to differentiate minute differences, which 

 yet substantially exist. The method of spectrum ana- 

 lysis, however, seems capable of registering reactions 

 beyond the scope of ordinary analysis. In fact, the 

 instrument may be said to afford the same aid to the 

 analyst in enabling him to note infinitesimal chemical 

 changes or reactions, as the microscope affords for the 

 discrimination and classification of microscopic objects. 

 On the whole, the results of a long series of patient 

 investigations show that pure iron is probably a very com- 

 plex body (as regards the commercial metal the micro- 

 scopic researches of Sorby go to prove that it clearly is 

 not, as usually assumed, a homogeneous body, but rather 

 a heterogeneous one), extremely sensitive to external influ- 

 ences modified by the presence of minute proportions of 

 other bodies; these latter, according to Prof Roberts- 

 Austen, governing its physical properties in accordance 

 with their atomic volume as compared with that of iron, 

 and in accord with the periodic law. It is evident that 

 it is not enough to give merely the simple percentage of 

 carbon, sulphur, &c., in iron ; the relative mass (as com- 

 pared with iron) of the foreign ingredient must also be 

 noted, and thus only can the absolute percentage com- 

 position be determined. Practical results appear to 

 confirm those attained in theory. 



The writer, when all these researches were not even 

 thought of, in some notes on the working of steel, 

 gathered from his own practice and confirmed by that of 

 others, considered that the heat at which the steel was 

 manipulated played a very important part, for the 

 following reasons : — 



Many discrepancies have been noted in the behaviour 

 of steel— the results of the tests applied not at all corre- 

 sponding to what might have been expected from its 

 chemical composition. For instance, steel containing 

 traces of silicon, "17 per cent, carbon, -06 of sulphur, 

 •07 phosphorus, -40 per cent, manganese, stood a tensile 

 strain of 27 to 30 tons per square inch ; other samples 

 of the same composition showed an increase of 31 to 35 

 per square inch, and under these latter conditions the 

 carbon had to be reduced '12 per cent, to stand the 

 normal test of 27 to 30 tensile strength. 



The cause of this sudden change was not satisfactorily 

 explamed. I think a plausible explanation of these dif- 

 ferences, which are of constant occurrence, may be traced 

 to dififereiices in temperature, and, strictly speaking, a 

 steel of given chemical composition must be worked at 

 a corresponding temperature to insure good results. 

 Some clue to the temperature required may, however, 

 be obtained. The fusion point of steel varies with the 

 amount of other matter alloyed with the pure iion. 

 All concerned in the manipulation of steel are well 

 aware that hot or cold rolling makes a great difference 

 in quality, and that the critical welding-point of the metal 

 is confined to a narrow limit of heat ; it is often a matter 

 of great practical difficulty to decide upon the most suit- 

 able temperature necessary for good work. It is well 

 known that steel may be seriously damaged either by 

 under or over heatmg. In the first instance the steel 

 may be too hard to bear rolling, or even if passed through 

 it may not be sufficiently plastic to yield readily to 

 pressure. We may in this case assume that minute 

 cracks or flaws occur which cause a corresponding de- 

 terioration in strength. On the other hand, it is well 

 known that if steel be over-heated it falls to pieces on 

 further manipulation. The reason is apparent — it is ap- 

 proaching to a semi-fluid condition, and therefore cannot 

 be rolled out. It may be, therefore, that for the success- 

 ful working of steel it is not enough to rely simply on 

 chemical analysis. 



The foregoing seems to indicate that steel must be 

 worked at a certain fixed, but as yet unknown, tempera- 

 ture below its fusion-point ; at which heat the flow of 



