The facts may be briefly stated thus, solid iron is a solvent for carbon which passes into solid solution, mainly in the form of a carbide Fe,C, when the iron is cooled below 1100°C. But iron is itself an allotropic body, and the particular allotropic form in which iron exists, determines the amount of carbide retained in solution by the iron as the mass cools down from 1100°. This carbide falls away from the solid solvent at 890°, because, in the case of this particular steel, the iron passes at that temperature from what is know as the y to the B form which cannot dissolve more than o'9 per cent. of carbon as carbide. When the temperature of the mass falls to 680° the rest of the dissolved carbide separates as laminæ.

In all the cases represented in the plate, analysis would indicate the presence of precisely the same amount of carbon, 1.5 per cent. The series begins with a section revealing the structure of the metal just as it comes from the cementation furnace. This section is the circular disc in the centre of the series. In No. 1, the steel has been heated to 1,000°, worked and slowly cooled. In it the carbide occurs in confluent masses, the ground work being made up of alternate bands of iron and carbide so that the structure resembles Mother of Pearl, the association being called "pearlite." In No. 2, the same steel has been heated to 850°, and quickly cooled in air. The confluent masses have to some extent disappeared, and there is a ground work of carbide in the distinct form to which the name "sorbite" has been given. In No. 3, the same steel has been heated to 850°, and quenched in water and has thereby been "hardened," and the resulting structure is minutely acicular the product being called "martensite." In No. 4, the interest deepens; the same steel has been heated to 1,050°, and quenched not in water but in iced brine. The result is totally different from "martensite ;' there are broad chevron-like bands of "martensite" enclosing lighter and much softer portions to which the name of "austenite" has been given by its distinguished discoverer M. Osmond; but this result hardly prepares us for what happens when the steel is still further cooled by placing it in liquid air, thereby cooling it to-243°. The material, as No. 5 shows, swells up and becomes martensite or something very like it, forming a fern-like structure of singular beauty. In No. 6, the treatment has resulted in partial reversion to the original structure, but the metal has been quenched from a temperature near its melting point, and what is known as a

"burnt" steel is the product. No. 7, shows the result of annealing the original steel at a temperature of 650° for a long time, followed by slow cooling; the result is better developed bands of cementite FC, and of the ground work of pearlite than in the specimen No. 1. Finally, by heating any of the specimens except No. 6 to 850°, working them, and allowing them to cool slowly, the structure with which we started in No. 1, is practically again obtained, as shown by No. 8.

For the sake of ready reference, the several specimens, too, have been arranged as segments of a circle round the section representing the original mass of "cement" steel, and the point I ask you to bear in mind is that these complex and varied molecular changes have occurred, except in the case of No. 6, in the solid, for in no other case has the steel been raised to a higher temperature than 1,050°, which is more than 300° below its melting point. It may be added that profound changes would have been brought about by annealing No. 3 for, say 500 hours, at a temperature of only 140°, but the structure would be too minute to be effec tively shown on the scale of magnification adopted in this series.

Evidently this extraordinary case of molecular unrest has been brought about by the molecular mobility of the carbide of iron and by a remarkable molecular change of the iron itself, on which

Change of Direction,

Budden Evolution of Heat

Brolution

of Heat

I will not now dwell. The connection of the phenomena we have been considering with Graham's work is obvious, for the transformations present a superb case of the influence of osmotic pressure. On it depends the diffusion of the carbide in solid iron, and by it the lowering of the allotropic change points of the iron, to which I cannot further allude, has also been affected. The molecular mechanism producing the lowering of the change points is, moreover, closely connected with the molecular unrest which enabled the 1.5 per cent. of carbon contained in the steel to lower the freezing point of the solvent iron from 1,600° to 1,380°. This latter fact brings us to one other fact which is the direct outcome of Graham's work on the occlusion of hydrogen by metals. A research upon which I have long been engaged,* has placed the relations of iron and hydrogen in a new light. The cooling curve of electro-iron given in fig. 3 shows, in addition to evolutions of heat at 1,1320, 895°, and 766°, the existence of two distinct evolutions of heat which occur respectively at 487° and 261o. These points on the cooling curve were recorded after the electro-iron had been heated in vacuo to 1,300' three successive times, but after repeated heatings of the metal in vacuo, they become so small that it is impossible to identify them with certainty. There would, therefore, seem to be but little doubt that they are due to the presence of hydrogen in the iron. The upper point, that at 487°, moreover, reappears after the metal has been re-charged with hydrogen. It may be pointed out that the ordinates, showing the evolution of heat, are obtained by a differential method which is described in the Report, to which a reference has been given. There are other results, which need not be referred to here, which indicate that the molecular changes in iron are dependent upon, or are influenced by, the presence of hydrogen in the iron, and the whole investigation shows how important Graham's original observations as to the relations of iron and hydrogen are likely to prove in the immediate future.

We are learning to look at the elements present in a metal, not merely from the quantitative point of view, but from the action they exert on the solid, stimulated as that action may be by the thermal treatment to which the metal is subjected. We are beginning to realise that the mechanical properties of a metal

* Fifth Report to the Alloys Research Committee of the Institution of Mechanical Engineers, 1899.

depend on the osmotic pressure exerted by the added elements, and, therefore, that "molecular unrest," and industrial applications are closely connected. This view, it may fairly be claimed, Graham originated, and I doubt whether he would have wished any other recognition than that so universally accorded to him of being the leading atomist of his age.

XIII-History and Art: a Florentine Outline.

BY

JAMES A. MORRIS, I.A., F.R.I.B.A.

[Read before the Architectural Section, 12th March, 1900.]

A KNOWLEDGE of history, like a knowledge of heredity, is one of those silent factors in the world's life, one of those iron keys which opens massive doors into still chambers, wherein strong men may look, and wise men learn.

Real history is the comprehensive and impartial record of the past, and the knowledge of a nation's antecedents is as essential to a right understanding of its present or future, as is a knowledge of heredity to the well-being of the family life; for, possessed of that knowledge, one can the better understand the recurrence or development of hereditary traits in the national or family life, whether of good or ill.

Now the influence of heredity, is, I think, discernible also in the arts; and, in a measure, goes towards the explanation of those recurrent art periods which seem so puzzling to the historian; but, are they not puzzling, just because so many histories are not comprehensive, in the sense that they unfortunately ignore art, or, if it be recognised at all, it is with scant courtesy, and seeming ignorance. The futility of this procedure is made further manifest, when one considers how much historical documents vary in detail, or present only one side of a matter, according as they have been penned by partisans or others; and in seeking justly to apportion these, it is worth while to remember how little we can assure ourselves of the real facts of political or social life, even of to-day; and, if this be so, how much less are we able accurately to