PREFACE.

Early in 1881 one of us submitted to Prof. G. F. Becker, geologist in charge of the Division of the Pacific, a brief digest of the facts showing the singular adaptability of the electrical properties of the iron-carburets for the classification of these products, with the request that permission be given us for the extension of the work in the laboratory of the Geological Survey. Professor Becker at once indorsed the project enthusiastically, and owing to his advocacy our proposal shortly after received the assent of the Hon. Clarence King, then Director of the Survey,-given with the proviso that the electrical researches be not prosecuted to such an extent as to occupy us exclusively.

Divers special investigations and routine duties, together with the labor involved in providing for the organization of a physical laboratory, prevented us from giving the furtherance of the proposed researches the attention necessary. Nevertheless, data of a varied character continually accumulated, while the scope and the conception of our general problem enlarged at an unexpectedly rapid rate. The publication of our results in some connected form, therefore, urged itself more and more seriously upon us.

About a year ago our plan was effectually encouraged by Prof. F. W. Clarke, chief chemist of the United States Geological Survey. The experiments to be discussed in this memoir have occupied our available time during the last five years. Notices, more or less complete, have appeared abroad from time to time in places not readily accessible to the public. Some of the papers it was deemed necessary to publish in German with considerable fullness. But all English publication has been purposely delayed, not only because we desired to reduce the results originally expressed in terms of the German standards to the more current and now legal denominations of ohm, volt, etc.,1 but principally because it seemed expedient for facility of comparison to refer all our data to the uniform temperature, zero centigrade. This

1In making this reduction the legal equation, 1 ohm=1.06 S. U., was made use of in all chapters with the exception of III and IV, the results of which were reduced at an earlier date and when 1 ohm = 1.05 S. U. appeared to be nearer the truth. This, however, is of no serious significance, because in these chapters the relative values of resistance are alone of interest. The absolute accuracy of the reduced values is of course immediately dependent on the absolute accuracy of the German standards (Siemens) at our disposal.

premised an accurate knowledge of the relation between electrical conductivity and temperature for iron, for steel in different states of temper, and for cast iron, and required excessively tedious labor.

The results in Chapter I on the electrical temperature-coefficient of iron-carburets present an unexpected range of variation, and thus possess intrinsic interest.

In Chapter II we investigate and discuss the conditions of the operation of tempering. This chapter is fundamental. Such facts as essentially sustain the argument underlying the whole of the present work are, therefore, emphasized with a larger number of experimental data than would otherwise be necessary.

In Chapter III we attempt to throw new light on the laws set forth in Chapter II by following them into their ulterior consequences. With the aid of certain allied electrical properties of alloys and of malleable cast iron, the nature of the phenomenon of hardness as presented by steel is discussed from every available physical and chemical point of view, within the scope of the present purposes.

In Chapters IV, V, and VI, the method for the accurate definition of hardness, and the scheme of operations for tempering developed in Chapter II, are consistently applied to analogous magnetic phenomena. The nature of the dependence of magnetization on the three independent variables of cylindrical rods, viz: carburation, ratio of dimensions, hardness, for given conditions of structure, is carefully discussed and in part graphically represented. Rules are finally given for the treatment of magnets, such that exceptionally great retentiveness, both as regards the hurtful effects of temperature and time and of shocks, may be conveniently attained with the least available sacrifice of magnetization.

We may add that a supplementary Bulletin is now in preparation, in which the very remarkable annealing effect of high temperatures (4000..1000°) will be magnetically discussed, and furthermore the degree of approximate coincidence between the physical state (of the necessarily linear rod) characterized by the unique maximum of magnetizability and the physical state of maximum density of steel, will be determined. That these states must be found very nearly coincident, our present results permit us to predict.

In Chapter VII, finally, we endeavor to generalize upon the foregoing results, as a whole; to restate the fundamental laws with greater accuracy and breadth of scope than was possible in the earlier chapters; and, finally, to deduce from all a method for the physical definition of iron-carburets.

Minor discussions, the relevancy of which does not justify their im mediate introduction into the chapters proper, are frequently introduced as addenda.

We desire in this place particularly to emphasize that throughout the present work the terms "thermo-electrically positive" and "thermo-elec trically negative" are to be understood in the way defined by the original

investigators, Seebeck,' Becquerel,3 Hankel; i. e., with reference to the series arranged thus:

- Bi... Cu... Fe.... Sb+,

5

an acceptation which we believe to be general on the Continent. In England the above terms are received in a sense which is precisely the opposite of this; that is, with reference to the thermo-electric series, arranged thus:

Of the two methods of designation, the latter is obviously the more logical and consistent, as will readily be seen, for instance, if the galvanic and the thermo-element be analogously described. And if we refrained from embodying the latter acceptation in this memoir, we have done so merely because of the great liability to error encountered in changing the sense of every thermo-electric expression and diagram, as well as the signs of all of the many thermo-electric data. Isolated constructions are too apt to be overlooked, and this in a way completely to mar the drift of the context. But the English reader will find no difficulty in making this change of sign for himself in any set of thermoelectric data which may interest him. To avoid all misconception, moreover, we give the direction of current in each essential case.

Much of the work was done abroad in Professor F. Kohlrausch's laboratory. It is a pleasant duty which permits us to extend to Professor Kohlrausch, in this place, our grateful acknowledgments, not only for the kindly interest with which he regarded the progress of the experiments throughout their extent, but for much valuable advice by which the papers have materially profited.

We desire to mention, in conclusion, that work done by us conjointly, if published in German, is to be put under S. and B.; if in English, under B. and S., conformably with an original agreement.

PHYSICAL LABORATORY,

UNITED STATES GEOLOGICAL SURVEY,

C. BARUS.

V. STROUHAL.

Washington, December 1, 1884.

2

Seebeck: Gilb. Ann., LXXIII, pp. 115 and 430, 1823.

3 Becquerel: Ann. de Chim. et de Phys., XLI, p. 353, 1829. *Hankel: Pogg. Ann., LXII, p. 197, 1844.

'Cf. Wiedemann: Lebre v. d. Elektrieilät, II, p. 248 et seq., 1883. Mousson: Physik, III, p. 381 et seq., 1875. Jamin: Cours de Physique, 2d ed., T. III, p. 42, 1869. Cf. Jenkin: Electricity and Magnetism, p. 175 et seq., 1880. Maxwell: Electricity and magnetism, 2 ed., Vol. I, p. 338-9, 1881.

(599)