Owing to certain inexplicable failures of tires, The Standard Steel Works authorized us to make a series of experiments to determine, if possible, the cause of the trouble. As a result of numerous analyses and physical tests, both from defective and experimental tires, we are able to present the following facts concerning piping and segregation, as they exist in the original ingot, and their effect upon the finished tire. At the autumn meeting of the Iron and Steel Institute, in a paper on segregation, after giving analyses from various parts of different-sized ingots, makes the following statement: “It may be fairly assumed that ordinary ingots are not seriously affected by this redistribution,” that is, by the segregation of the elements.

We think the following data show conclusively that this is not the case, and that there are conditions under which ordinary ingots are seriously affected by this redistribution of the elements.

Our opinion is that segregation and the existence of piping in the ingots used are the causes of by far the larger percentage of failures of steel tires. If this opinion can be substantiated, then, as the two evils of piping and segregation occur in the same part of the ingot, we are able, by the removal of that part and the utilizing only of the metal which is more nearly perfect both physically and chemically, to reduce the percentage of failures to a minimum.

Before going further we would define segregation as a concentration of a part of the elements in a certain portion of the ingot, forming compounds richer in these elements than the mother-metal. Concentration actually takes place in the part which solidifies last.

The elements in which there seems to be the greatest tendency to segregation are carbon, phosphorus and sulphur. We have in the course of our experiments found only one or two instances where there was a decided variation in silicon and manganese.

It must be borne in mind that the ingots in question are comparatively small; it may be due to this that segregation of silicon and manganese has not been observed.

Segregation occurs around the sides and at the base of the pipe; samples from the extreme top of an ingot usually give lower results than those from the body. We do not wish to convey the idea that the remainder of the ingot is perfectly homogeneous, since this is not the case. We do believe, however, that segregation to a damaging extent occurs in the immediate vicinity of the pipe only.

It is a question in our minds whether some of the hard spots found in tires after first or second turning may not be due to segregated material. There is little doubt that the majority are due to sliding of wheels. Whether or not a hard spot was due to segregation could readily be determined in any individual case by analyses of material from the spot in question.

In our experimental work the etchings have proved almost invaluable. From the etched surface samples for analyses can be taken intelligently, as there is a decided difference in appearance between the segregated and the non-segregated metal; that is, when segregation occurs to a marked degree. The etched surface shows the location of segregation by a porous or granular appearance. This is much more apparent in the etching itself than on a print made from it.

We are in the habit of referring to the segregated metal as being porous and to the non-segregated metal as being solid; wherever these terms are used by us they will be understood to have the above meanings.

The chief difficulty experienced is in getting enough material for a complete analysis from porous spots. The results are lowered by the addition to drillings taken for analyses of solid material surrounding the spots. Another point to be taken into consideration is the liability in taking drillings to get into solid material underneath porous spots.

We make a practice of etching one-half of the broken test-pieces as close to the point of fracture as possible, and we have yet to find a case exhibiting a marked variation in physical tests from the same tire, the reason for which was not shown by the etchings of the broken test-pieces. By following this plan we think a reason might be found for some of the inexplicable results that are met with in the testing of material. While it is a point hard to determine it seems that the line between segregated and nonsegregated metal is pretty sharply drawn.

https://www.youtube.com/watch?v=fZyPZaVEuBY

That a decided segregation may take place in a small ingot is shown by the following experiment:

Fig. 1 is a sketch of a section of an ingot, cut lengthwise through the center. The ingot was 12 inches in diameter and 13 inches long, weighing about 400 pounds. It was one of a group, and was bottom-cast in a mould of cast-iron with a cast-iron cover. The sketch shows piping and also the location of holes from which samples were taken for analyses.

Fig. 2 is from a print of an etching and shows the portion of the ingot in immediate vicinity of the pipe. The segregated material can be located by light spots on the print, the original appearance on the etching being granular. Fig. 2 is full-size.

In Table I. we give analyses from different parts of the ingot. They show, it seems to us, a very homogeneous metal outside the immediate vicinity of the pipe. As will be seen, the sample from hole 13 gives the lowest results. Drillings from hole 2 were taken with a ¼-inch drill in order to keep as far as possible inside of porous spot. There was only enough material in the sample for a carbon- determination. Carbon was found to be 1 per cent., and from analogy there is every reason to believe the sample would have shown high phosphorus and sulphur.

In order to determine what results could be expected from a long ingot, the following experiment was made:

A bottom-cast octagonal ingot, 5½ feet long, 12 inches across the flats at bottom and 11 inches at top, was cut lengthwise through the center. A good print was obtained from an etching of the section of this ingot, but, on account of its size, it cannot well be reproduced.
There were only a few traces of segregation shown by the etching, and analyses showed that segregation had only taken place to a small degree.

We have not attempted to give all the analyses made of samples taken. A sufficient number, however, are given to show the character of metal in different portions of the ingot.

Table II. gives analyses of samples taken from the holes shown on Fig. 3.

We would call attention to several facts:

1. That samples from holes 16, 22 and 25 show higher results than are shown from corresponding holes on either side of the ingot;
2. That samples from holes 28 and 33 show about the same results as corresponding holes on the sides;
3. That samples from holes 36, 40, 44 and 47 show results slightly lower than samples from corresponding holes on the sides.

Silicon and manganese determinations made on samples taken from holes as shown in Fig. 3 gave the following results (Table III.).

The average analyses taken from the series of holes on the left side of the ingot, beginning at 18; from the middle of the ingot, beginning at 16; and from the right side of the ingot, beginning at 17, are remarkably close. We give these averages in Table IV.

The succeeding experiments show the effect of piping and segregation in finished tires and also the results obtained from tires made of billets from long ingots, the top or piped end having been cut off.

Fig. 4 is a print from an etching of a tire made from a short ingot, say 12 inches long. The effect of piping in the ingot is at once apparent; in this case the flange-side of the tire corresponds to the bottom of the ingot.

The analyses given of samples, from different portions of the etched section, show a marked segregation, but not as decided as we sometimes find. The segregation occurs in the immediate vicinity of the pipe, as is shown by Fig. 1A, which gives the location of the

drill-holes from which samples were taken for analyses. The most interesting feature in the case of this tire is the variation in physical results. Test-piece No. 2 was from the piped and segregated portion of the tire; No. 4, from the solid portion. Test-pieces were taken longitudinally, or, we might say, in the direction of the grain.

Fig. 5 is a print from an etching of broken test-piece, No. 2; Fig. 6 is a print from an etching of broken test-piece, No. 4. It is evident why test-piece No. 2 gave such poor results.

Table V, gives the chemical analyses and physical results obtained from this tire.

Fig. 7 is a print from an etching of a tire showing one of the

worst cases of segregation we have found. This tire broke after having been on center a few days, and before it was put in service.

The shrinkage was excessive, but, as other tires had stood the same amount, there is little doubt in our minds that segregation was, to a great extent, at least, the cause of failure.

The fracture was clean. With a small glass no traces of defects were found; the only peculiar feature being a polarized appearance; and etching showed the location of pole to be at point of worst segregation.

Fig. 8 shows location of holes from which samples were taken for analyses, and Table VI. gives results of chemical analyses and physical tests.

With regard to the carbon given for hole 3, we would say, that in order to satisfy ourselves and check our work, we had a determination made by Booth, Garrett & Blair, whose results are given, and correspond almost exactly with those obtained by us.

Mr. H. M. Howe, in his Metallurgy of Steel says: “ Heterogeneous composition usually implies heterogeneous strength and ductility; and the strength of a heterogeneous substance is usually nearer the strength of the weakest component or part than the average of all the parts. The piece tends to break down piecemeal. So with ductility.” This being so, the character of such tires as are shown above is to be judged, not so much by the best results that can be obtained from a section of the tires, as by the poorest.

Fig. 9 is a sketch of an ingot 5½ feet long, from which four billets were cut, marked A, B, C and D respectively. These billets were made into tires, the top portion of the ingot, or the piece not marked, having been scrapped. The billets ranged in weight from 760 to 800 pounds.

Fig. 10 is a print from an etching of a section of one of these tires, and is typical of them all. Figs. 10A, 10B, 10C and 10D, show

the location of the holes from which samples were taken for analyses.

The tires were all cooled in air as they came from the mill, without being subjected to any further treatment.

Table VII. gives results from the four tires.

In every instance, test-piece No. 2 is from the face-side and No. 4 from the flange-side of the tire.

With regard to the chemical analyses we would say, that this steel was made some time ago. Our present specifications to the Otis Steel Company, Limited, from which we buy our steel, call for lower silicon and phosphorus, and in most classes of service, for lower carbon, giving (as shown by Table VIII.) lower ultimate tensile strength and increased elongation.

Table VIII. gives the results obtained from a tire made from one of a number of billets, as in the preceding instance. The etching of a section of this tire showed no traces of segregation. The etchings of broken test-pieces showed no imperfections. The analyses correspond more closely with our present specifications.

Fig. 11 shows the location of the drill-holes from which samples were taken for analyses. As in previous cases, test-piece No. 2 is from the face-side of the tire and No. 4 is from the flange-side.

As a result of these and numerous other experiments the Standard Steel Works have abandoned the use of short ingots in the manufacture of tires, and are now using billets cut from ingots about 5½ feet long, of varying diameters, the top or piped end, which is also that portion of the ingot where segregation is liable to occur, being scrapped.

Before closing we would call the attention of those interested in this subject to an instructive series of etching prints in the Western Railway Club Proceedings of the meeting held in October, 1890. The prints are given by Mr. Rhodes in connection with a paper on steel-tired wheels and their fastenings.

The mixture used for etching is composed of three parts of sulphuric acid, one part of hydrochloric acid and nine parts of water.

Mr. Osmond’s paper on “ Microscopic Metallography” places us under renewed obligations to him for his clear and concise statement of the present status of the new and important science which he has done so much to establish. While, as he says, “ it has not acquired perhaps either the full consciousness of its future, or the full possession of its field,” we must not forget that great advances have already been made in judging steels from the microscopical examination of their fractures and etched specimens showing their structure. At the present time, for steels requiring hardness and toughness combined, or high elastic limits in proportion to the ultimate strength, no one relies wholly upon the chemical composition except as determining the grade of steel. It is necessary to bring that special composition into the highest condition of which it is capable, by both mechanical and thermal treatment. In other words, we seek to render the geometric elements of structure or the mineral aggregates as small and homogeneous as possible.

The constructor of ordnance, observing his tensile bars in the testing-machines, judges the condition of the metal as much by the appearance of the surface under elongation as by the relations of the elastic limits to the ultimate strength. If the bar elongates with a smooth surface, he knows that the mineral aggregates are small and elongating uniformly, and that the metal is in good condition. On the other hand if the bar in elongation presents a reticulated surface, the mineral aggregates are much larger, and elongation is taking place more rapidly through the planes of cleavage between, than through the aggregates. The metal is not in the best condition and must be improved by further thermal treatment.

The fact that chemical homogeneity by no means necessitates mechanical homogeneity should be extensively disseminated among consumers of steel rails, tires and ordinary grades of steel. Even at the present time there are many who suppose that a given chemical composition, irrespective of the form of the section and method of manufacture, means very definite physical properties in the rails and tires. This is largely due to the fact that steel rails, having been once molten, were said, when first introduced, to be homogeneous in contrast to the built-up pile of the iron rails. That crystalline forces, during solidification and subsequent cooling and heating of the metal, form the structure, was not considered. The care and time given to the manufacture of the early rails from small hammered ingots, and the many passes in the rail-trains at low temperature, gave to the small thin heads a solid structure of small mineral aggregates, presenting great resistance to flow from the pressure of the wheels of passing trains. From several of these old rails, etched specimens show, at 50 diameters, only a confused structure, the ferrite being distributed in small grains, and the pearlyte seeming to be granular. Each aggregate is so thoroughly interlocked with its neighbor that it is difficult to break it down and cause flow of any or all of its constituents. The wear of those rails has been so satisfactory as to raise hardly any question with regard to their homogeneity.

Generally speaking, the elastic limits of these rails ranged from 40,000 to 45,000 pounds, and the ultimate strength from 80,000 to 90,000 pounds, but, as I have said, each aggregate was thoroughly supported by those adjacent, and the wheel-pressures were carried and distributed so well that the breaking-down of the aggregates, due to the cold-rolling of the surface of the rails by the wheels, rarely exceeded 0.02 inch in depth. In many of the best-wearing rails it was less than 0.02 inch. The flow of metal to the side of the head was very limited and rarely occurred on the rails, except on curves and heavy grades. With the wheel-pressures of to-day a slight flow would be produced in such rails.

Fig. 1 shows a micro-photograph of the structure of such steel magnified 50 diameters, the upper part being the top of the rail. The pearlyte and ferrite are thoroughly intermixed in a confused state, the former predominating, seeming more granular than the lamellar type. Such rails wore, in track-man’s parlance, “smooth and bright as a silver dollar.” Under the microscope, the lamellae seen on the surface of the rails are very small and appear firmly attached.

In the deep-headed type of rails, similarly shown in Fig. 2, the structure is coarser, and the pearlyte and ferrite are entangled with each other, showing on the etchings like vermicular markings. The paerlyte seems to be rather of the lamellar than of the granular type. The lamellae are feebly united, being formed under high temperature; the mineral aggregates are large and friable, and the surface of the rail breaks down more than 1/32 of an inch in depth, readily flowing under the wheel-pressures. Under the microscope, the rails seem to be formed of scales or lamellae, overlapping each other, of both pearlyte and ferrite. Little fragments are constantly becoming detached from the surface, or aggregates are flaking out, while larger masses are flowing to the side of the head and becoming detached in large fragments.

Such rails do not wear smooth. They have a dull luster, are pitted and present, in the track-man’s parlance, a “ mealy appearance.” Figs. 2, 3 and 4 are micro-photographs, magnified 50 diameters, of the transverse structure in a deep-headed rail containing 0.26 carbon. Fig. 2 is from the head of the rail, just beneath the uninjured surface, and is very coarse in structure compared with Fig. 1. Fig. 3 is from the upper portion of the rail. The structure shows the aggregates broken down and under flow, even below the portion shown at the bottom of the picture. As the figure indicates, the top of the rail soon becomes a series of thin lamellae, portions detaching as soon as the limits of elongation of the metal are reached. The surface-lamellae of the rails are between the upper and nether millstone, and portions are ground to powder upon the passage of each wheel.

This is a practical illustration of axial compression in Mr. Osmond’s cylinder, A B (Fig. 1, Trans, xxii., 244). In this case we have a series of superimposed lamellae and flow caused by axial compression between all of their surfaces of contact, instead of the two similar halves of the cylinder placed upon their common base, xy, which he considers “ an extreme case.” Applied to rails, it is not an extreme case for the upper aggregates to flow, or portions of the metal to break out and flow over an interior gas-bubble or small pipe.

The flow on soft rails is not confined to a few lamellae on the surface, but includes the breaking of several aggregates in depth, which finally flow until they overhang the edge of the rail, as shown in Fig. 4, then accumulate in larger masses, and eventually become detached. The wear and deformation of the rail-section largely depends upon whether the metal has great resistance to surface-flow or flows readily under the wheel-pressures.

Many years of experience with steel rails have clearly demonstrated that those which flow the least under the wheel-pressure give the most service, and those that flow the most wear and abrade with the greatest rapidity. This same remark applies to the flange-abrasion on curves. As I have already observed, the most serviceable rails have a fine structure, and small mineral aggregates, while the less serviceable rails have a coarse structure and large mineral aggregates, though, in many cases, the so-called chemical composition is identical. In the language of the constructor of ordnance, the condition of the metal for the grade of its composition was better for the service in the older than in the later rails.

It is now more than ten years since I made a broad thin head for rails of heavy sections. The roll-pressures and rapid cooling contribute to make the structure finer and more homogeneous in heads of this form than it was possible to secure in the deep-headed type then in general use.

During the past nine years one railroad has put 80,000 tons of my section into trunk-line service, and, notwithstanding over 50 per cent, increase in the wheel-pressures, the rate of wear of these rails has been very small. Many railroads, seeing its good wearing qualities, have copied the design, and it is now a leading type.

In 1890 I introduced a series of rail-sections of different weights, especially designed to produce a fine structure in the heads by the modern process of rolling. I designed also a special 95-pound rail for the Boston and Albany Railroad Company.

It was long ago observed that as the weight of the section increases, the difficulties of producing a fine structure in the head are also increased. In order to secure a fine structure in my large sections, especially in the Boston and Albany 95-pound, I proposed the following composition: carbon, 0.60; manganese, 0.80 to 0.90 ; silicon, 0.10 to 0.15; phosphorus, not to exceed 0.06 ; sulphur, not to exceed 0.07.

While great doubt was expressed as to the possibility of securing toughness with so high a percentage of carbon, we were able to control the condition of the metal in the final product of the rails, so that they were of unexpected toughness. The butts stood a drop-test of 2000 pounds, falling 20 feet, and rarely broke; while hundreds stood 2 to 4 blows of 30 feet before failing (supports 4 feet apart). At the side, the flange gave 16 to 18 per cent, elongation per inch under the point of impact of the drop. It must be said to the credit of President Bliss of the Boston and Albany Railroad Company, that he was willing to meet the manufacturers in a fair spirit and pay an increased price per ton for the high-carbon rails.

Up to the present date more than 150,000 tons of these high-carbon rails, in my series of sections of 60, 65, 70, 75, 80, 95 and 100 pounds, have been rolled; about nine-tenths of the product being in the weights above 70 pounds. The carbon is slightly decreased in the lighter sections and increased in the 100-pound section. The rails are showing good results in trunk-line service. Out of 75,000 tons, none of them failed in the exceptional cold of 1892-93 the temperature in many places being 30° F. below zero.

The structure in all the sections is now under study and comparison. The same section having been produced in many cases at different mills under different methods of manufacture, a vast field is offered for research.

Practically for rails the structure of the ingot cannot be omitted from consideration, as its solidity and continuity are quite as important as a good micro-structure. In brief I may say that my studies have enabled me to improve the texture and increase the toughness of the metal at the same time.

As to microscopic examinations, Mr. Osmond observes, that “ natural illumination will serve only for very low powers, which are usually insufficient for this work.” Dr. Sorby makes the same remark, which was strictly true when he made his research. Our American opticians were the first to produce wide-angled, low-power objectives, which, admitting so many rays of light, are well adapted for the study of opaque objects by natural illumination. I have found a 2-inch of 9 or 10 degrees aperture and a 1-inch of 30 degrees aperture to work well by natural illumination on either fracture or etchings. They cover a large field and stand deep eye-piecing for amplification. A ½-inch objective of 100 degrees aperture is also serviceable. Higher powers must be used for resolving the fine laminae. For photo-micrography the wide-angled objectives are not essential, because the plate will be sensitive to lines in the spectrum far beyond what the eye can see.

A good photo-micrograph often shows what it is impassible to see under the microscope. An important point is, that a skilled operator can bring out structure on his plate, that the inexperienced eye could hardly recognize under the microscope. Mr. Osmond’s photo-micrographs, taken with his simple inexpensive apparatus, are so excellent that they reflect his skill as an operator and his selection of plates, the emulsion of which was suited to the optical combination of his microscope and lamp. These are important matters.

The paper of Professor Martens, which I have not had opportunity to examine fully, is evidently another contribution of the highest value in this important department of research. And our fellow-member, Mr. Sauveur, has confirmed and added to these results from abroad by his excellent work in the same line.

The importance of the subject of segregation, treated in M. Pourcel’s paper, is coming to be generally recognized. The segregation of carbon, sulphur and phosphorus from the setting metal to that which solidifies last, often becomes a serious matter in rail-ingots. Upon examination of a number of rails broken during the severe cold of last winter, it was found that they were the top-rails from the ingots, with so great an excess of carbon and phosphorus as to be brittle.

For the last two years we have done considerable microscopical work at the South Works of the Illinois Steel Company, with gratifying results. Some of those results are embodied in my paper on the “ Micro-Structure of Steel,” presented at this meeting. I should only like to add here a few remarks which have been suggested to me by Mr. Osmond’s paper, and which may be of interest at this meeting.

In regard to what Mr. Osmond writes concerning the polishing of the specimens, I would say that after numerous and repeated trials of all the polishing-agents used, and recommended for such work, we have adopted the following manipulations as giving the best and quickest results. The specimen is first cut to a convenient size(½ inch to 1 inch square and about 1/8 inch thick will answer very well) and then gradually ground with the following polishers, in the order named:

1. Emery-paper, grade No. 2, mounted on a smooth board of wood.
2. Emery-paper, grade No. 1, similarly mounted.
3. Emery-paper, grade No. 00, similarly mounted.
4. Flour-emery, spread wet over a cotton cloth, stretched over a block of wood, with a perfectly plain surface.
5. Tripoli, wet, on cloth in like manner.
6. Best grade of jewellers’ rouge, used wet on a piece of chamois-skin, stretched tightly over a block of marble.

In dealing with cast-iron, the first emery-paper may often be omitted. By following this process it will be found after a little practice, that a piece of steel can be polished and freed from even microscopical scratches in about an hour, while one-half hour is generally sufficient for samples of cast-iron. According to Mr. Osmond, oblique illumination is necessary to bring out certain moire effects, which he describes. These effects we have not studied; but, unless used for that purpose, oblique illumination, which in all events may only be employed with low powers, is not to be recommended ; for it does not give a truly accurate appearance of the real structure of the specimen. As a source of light we use the oxy-hydrogen flame playing on a cylinder of calcium or of zirconium. The latter we have found to give as good results as calcium, but no better.

With such brilliant illumination the time of exposure for photographing is greatly shortened. Using orthochromatic plates (sensitometer 16 to 20) an exposure of from 4 to 10 seconds will be found sufficient for a magnification of from 50 to 100 diameters. With higher powers a longer exposure is necessary. A magnification of 500 diameters requires an exposure of from 30 to 40 seconds. We may say in passing that we have succeeded in obtaining some fair photographs of spiegeleisen magnified 1000 diameters. For that purpose we used a 1/12-inch homogeneous objective and an amplifier.

We have assumed that cementite and the hard constituent of pearlyte are the same carbide of iron (probably Fe3C), but we agree with Mr. Osmond that much more evidence is needed to settle this point. Until, then, however, we shall accept that hypothesis, first, because it accounts remarkably well for the structure of the different grades of steel; secondly, because there is as much, if not more, evidence in its favor than in favor of the opposite view; and, thirdly, because it is a simple one, which supposes the formation of only one carbide of iron in all steels not quenched.

I should like to call attention to the concordance between most of Mr. Osmond’s conclusions and our own. Mr. Osmond says that the temperature, increasing the pearlyte, tends to form more and more large and regular polyhedra. I express the same view in my proposition IV.: “ The higher the initial temperature the larger the grain for a given chemical composition.” Again, talking of steel of medium hardness, Mr. Osmond says that quick cooling prevents the structure from developing, thus causing a finer grain. In my proposition V. I say, “ The slower the cooling the larger the grain for a given composition.” In my last proposition, however, I advance the view that the size of the grain is independent of the amount of work the metal has received. Mr. Osmond agrees to this, but only when the metal is finished at a relatively high temperature. If the forging or rolling has been continued until a lower temperature is reached, then Mr. Osmond claims that work, as such, does influence the structure of the metal. Of course, our proposition applies only to hot working; and, perhaps, the working-temperature at which the structural change, observed by Mr. Osmond, takes place is so low that it actually belongs to cold-working.

I was especially gratified to read the strong plea which Mr. Osmond makes in favor of the use of the microscope as a means of “ reconstructing the thermal treatment to which the steel has been subject,” and consequently of detecting improper heat-treatment. This is, indeed, one of the main points which I desire to emphasize in my paper.

With regard to some suggestions and criticisms contained in Prof. Marten’s brilliant paper, I will say a few words.

While we must all agree in assigning the very greatest importance to the study of the micro-structure of steel, yet I think that we should by no means neglect the study of the fracture. Through it very important light may be thrown on the problem with which we are now confronted. The fracture takes place along planes of weakness. Now, the study of these planes of weakness and of their distribution and size under different conditions and after different conditions of heat-treatment may, it seems to me, be of incalculable benefit to us.

The use of specific names for the microscopic components of iron and steel I think is, on the whole, desirable as a matter of convenience. It is inevitable that in order to convey our ideas to each other we should make use of some kind of name. Before I suggested the specific names “ Cementite,” “ Ferrite,” etc., writers inevitably, even if unconsciously, fell into using names which were equally specific, but simply more cumbrous. There was “ The pearly constituent,” “ The intensely hard compound,” etc. The specific names which I gave were really no more specific than those which preceded them, but simply shorter and more convenient.

We cannot converse or write about such substances without using some sort of name. That name may be brief like “ quartz,” or it may be a long phrase like ” the intensely hard, transparent, vitreous substance which crystallizes in the hexagonal system; ” but such a phrase is simply a long inconvenient name. In the early stages of our knowledge names must be given provisionally, because we cannot fully distinguish the different substances, and we may incorrectly group under a single name some unlike substances.

We must recollect that in the early history of mineralogy names were sometimes incorrectly given, and sometimes two classes of minerals were incorrectly grouped under one name; but the errors and confusion which were occasioned by such error in nomenclature were insignificant, and not to be weighed against the advantage as regards convenience which the use of brief names for minerals gave.

Now, the only question is whether we shall give brief names, recognizable and easily written, to the microscopic constituents of iron, or use long and cumbrous ones. The brief ones have the great advantage of clearness and convenience. They have, perhaps, a slight disadvantage in suggesting a more complete knowledge than we actually possess, and also in that, when we do later come to a more perfect knowledge and have to remodel our nomenclature to fit the new facts, it may be less easy to drop them than it would be to drop long ones which had become odious through their inconvenience. I agree with Prof. Martens that we should all insist that, for the present, the specific names must be regarded as provisional.

I think that, running through Mr. Sauveur’s interesting paper, there is a tacit assumption that there is a constant and known relation between the coarseness of the grain of steel and its quality or physical properties, and that in general (other things being equal) the finer the grain the better the quality. This assumption is natural and quite proper. I do not intend to criticise it in the least. But I do want to point out that we have now reached a stage in the study of these questions which makes it desirable that we should seek direct and accurate testimony as to the relation between the size of the grain and the physical properties, and that we should no longer take this important matter for granted. If the physical properties of every piece of steel which is subjected to microscopic examination were also determined, I think that it would be a great help to us, and that we could interpret the results of our microscopic examination very much more confidently. We must not forget that it is knowledge of the physical properties after all that is our objective point, rather than knowledge of the micro-structure. We do not study micro-structure for itself, but as a key to the physical properties.

The specimens on the table before us show that, in certain cases at least, the size of the grain as revealed by fracture gives no indication of the physical properties of the metal. These two pieces of the runner of a manganese-steel casting were originally one continuous piece. The fracture of one, as you see, is reasonably fine. This is the piece before water-toughening, and, therefore, while relatively brittle. The fracture of the other is extremely coarse and columnar; this is of the water-toughened and very strong and ductile metal. In short, here in one and the same continuous piece, and but a short distance apart, a relatively fine fracture accompanies weakness and brittleness, while a coarse columnar fracture accompanies strength and ductility combined.