2>7o 



NATURE 



[Fed. 20, 1890 



(Nature, xH. pp. 11, 42). I desire, in the first place, to 

 point out the bearing of the singular minimum of the 

 viscosity of hot iron (^oc. cit., p. 34) on the interpretation 

 given of Maxwell's theory of viscosity {Phil. Mag. (5), 

 xxvi. pp. 183, 397, 1888 ; xxvii. p. 155,1889). When iron 

 passes through Barrett's temperature of recalescence, its 

 molecular condition is for an instant almost chaotic. This 

 has now been abundantly proved (cf. John Hopkinson, 

 Phil. Trans., London, clxxx. p. 443, 1889, where the 

 literature may be found ; cf. Osmond, below). The 

 number of unstable configurations, or, more clearly, the 

 number of configurations made unstable because they are 

 built up of disintegrating molecules, is therefore at a 

 maximum. It follows that the viscosity of the metal 

 must pass through a minimum. Physically considered, 

 the case is entirely analogous to that of a glass-hard steel 

 rod suddenly exposed to 300°. If all the molecules passed 

 from Osmond's /3 state to his a state together, the iron or 

 steel would necessarily be liquid. This extreme possi- 

 bility is, however, at variance with the well-known prin- 

 ciples of chemical kinetics. The ratio of stable to 

 unstable configurations cannot at any instant be zero. 

 Hence the minimum viscosity in question, however rela- 

 tively low, may yet be large in value as compared with 

 the liquid state. 



(2) My second point has reference to the function of 

 carbon in steel. It is not to be understood that we ignore 

 the importance of the changes of carburation produced 

 by tempering steel. To explain the varied physical phe- 

 nomena which accompany temper, it is sufficient to re- 

 cognize some special instability in the tempered metal. 

 This is given by the carbide configuration, and the phy- 

 sical explanations in question may be made without 

 specifying its nature further. Hence the permissibility 

 of the purely physical considerations. 



On the other hand, it is indeed surprising that, on the 

 part of engineers and chemists, the important subject of 

 temper has been but inadequately dealt with, as Prof. 

 Austen justly remarks. Sir Frederick Bramwell (cf. 

 Nature, xxxviii. p. 440), in his inaugural address at 

 Bath, in 1888, dwelt at some length on the subject of 

 temper. The question is again touched upon by Mr. 

 Anderson at the Newcastle meeting of the British Asso- 

 ciation. Neither of these gentlemen, however, really 

 shows forth the gist of the matter. Indeed, even in 

 Ostwald's massive " Lehrbuch der AUgemeinen Chemie" 

 (Leipzig, W. Engelmann, 1887), full of examples as it is, 

 bearing on all points of chemical physics, the frequent and 

 exceptionallyimportant case of tempered steel is altogether 

 absent. And yet the chemical interpretation to be given to 

 the phenomena of temper seems to be closely at hand. Dr. 

 Strouhal and 1 {\Vied. Ann,->ii. p. 390, 1880; Bulletin 

 U.S. Geol. Survey, No. 14, chap, ii., 1885) showed that, 

 by the process of hardening, the electrical resistance of 

 steel may be increased by more than three times its value 

 for the soft metal. If the hard rod is now softened, the 

 resistance again decreases by an amount depending on 

 the temperature to which the hard metal is exposed and 

 on the time of such exposure, in a way which, throughout 

 the whole research, is beautifully sharp and character- 

 istic. Eventually, the relatively low resistance of soft 

 steel is again reached. Now suppose the carbon mole- 

 cule of steel to be dissolved in the metal, forming an 

 alloy of Matthiessen's Class II. Seeing that the quantity 

 of carbon contained is not large, the electrical resistance 

 of hard steel is at once an expression of its chemical com- 

 position, structurally unknown though it be. Hence in 

 the electrical diagram of the phenomena of temper con- 

 structed by Dr. Strouhal and myself, the time variations 

 of resistance of hard steel at any given temperature may 

 be interpreted as a case of Wilhelmy's {Pogg. Arm., Ixxxi., 

 pp. 413, 499, 1850) rate of chemical reaction {Rcactions- 

 geschwindigkeit), and expressed in accordance with his 

 well-known exponential law. This indeed is the character 



of the observed time curves. Hence also the full diagram 

 of the phenomena of temper, considered both in their 

 variation with time and with temperature, is available for 

 the elucidation of most points relative to the effect of 

 temperature on rate of chemical reaction.^ 



(3) A further remark may be made relative to Osmond's 

 {Antiales des Mines, July- August, 1888, pp. 6-7 ; Mem. de 

 VArtillerie de la Marine, Paris, 1888, p. 4) iron of the a 

 and the ^ type. The assertion that mere strain partly 

 changes a into /3 iron is in conformity with the viscous 

 behaviour of the metal. For it appears that the effect of 

 any mechanical strain as well as of temper, is marked 

 decrease of the viscosity of the metal. Osmond's theory, 

 however, appears to explain too much. Since most metals 

 can be similarly hardened by straining, it would follow 

 that there should be a and ^3 varieties in all these cases, 

 even though a molecular change corresponding to Gore's 

 phenomenon in iron has only in a few instances been 

 observed (iron, nickel, platinum-iridium alloy). I believe, 

 however, that there is reason to be urged even in favour of 

 this extreme view.^ The ion theory of metallic conductivity 

 is fast gaining ground. 



J. J. Thomson states it in his well-known book 

 (" Applications of Dynainics," p. 296). Giese ( Wied. 

 Ann., xxxvii. p. 576, 1889) has outlined an ion theory of 

 electric conduction, uniformly applicable to metals, 

 electrolytes, and gases. It seems to me, if a preliminary 

 hypothesis be made relative to the evolution of a magnetic 

 field out of an electric field ; if advantage be taken of the 

 spiral distribution of points which frequently results from 

 the symmetrical interpenetration of two congruent Bravais 

 systems ; ^ if, finally, in metals, the function performed by 

 a bodily transfer of ions can also be performed by an 

 exchange of the charges of charged atoms (Giese, in- 

 directly Helmholtz), that the possibility of an ion theory 

 of magnetism may be suspected. Quite apart from the 

 influence of a field, the conditions of exceptionally close 

 approach favourable to the transfer of charges from atom 

 to atom, are given by the distribution of the heat agitation 

 in the metal. 



(4) I will close this note by some remarks on the change 

 of the character of diffusion when occurring in solids. 

 Studying the coloured oxide coats on iron, Dr. Strouhal 

 and I (Bull. U.S.G.S., No. 27, p. 51, 1886) pointed out 

 that the outer surface of the film is oxidized as highly as 

 possible in air ; and that the inner surface of the film, 

 continually in contact with iron, is reduced as far as 

 possible. This distribution of the degree of oxidation 

 along the normal to the layer, is equivalent to a force in 

 virtue of which oxide is moved through the layer, from its 

 external surface to its internal surface. The formation of 

 an oxide coat is thus a case of diffusion. Conformably 

 with this view, the film, during its formation, behaves like 

 an electrolyte, as was pointed out by Franz, Gaugain, 

 and Jenkin, and more recently by Bidwell and by S. P. 

 Thompson. 



We then adverted to the crucial difference be- 

 tween diffusion in solids and diffusion in liquids, in- 

 asmuch as in the former case (solids) diftusion de- 

 monstrably ceases after a certain small thickness is per- 

 meated. The limit thickness of the film is reached 

 asymptotically, through infinite time. It has a definite 

 value for each temperature, increasing as temperature 

 increases. In the light of other evidence since gained, 

 this explanation is substantiated. The formation of the 



' An ulterior consideration presents itself here relative to an extension of 

 the thejry of Arrhenius \i^' ied. Ann., iv. p. 391, 1878) to metallic con- 

 ductivity. Arrhenius and Ostwald find in the maximum of electrolytic con- 

 ductivity a measure of rate of reaction. I must pass over this question here, 

 since it is without immediate bearing on the text. 



- I have spent much time in endeavouring to throw light on this question, 

 and will indicate the results later. My methods were (i) to find the effect of 

 mechanical strain on the carburation of steel ; (2) to find the effect of strain 

 on the rate of solution ; (3) to find the hydro-electric effect of stretching. 



i A good account of the relations of the iiravais .ind the Sohncke system is 

 given by H. A. Miers, in Nature, xxxix. p. 277. 



