Search Documents

By P. A. Larssen, P. R. Gage, R. W. Bartlett

The kinetics of formation of intermediate sili-cide compounds from MoSi2/Mo and WSi2/W couples were studied using a cylindrical sample configuration. Concentration profiles of the constituent metals were deter mined from electron-microprobe analysis of cross-sectioned samples after isotlzernzal diffusion annealing. These data also yielded hand widths of phases. Growth of intermediate phases depends on diffusionin them and obeys a parabolic rate law. The rates of formation of the intermediate phases decrease in the order Mo5Si3 > W5Si3 > Mo3Si. Crowth rates. for the isomorphous phases Mo5Si3 and W5Si3 had the same activation energy. 43 kcal per mole. SILICIDE coatings employed to protect refractory metals from high-temperature oxidation are usually formed by a diffusion process requiring concurrent transfer of silicon to the refractory metal surface. For a simple binary system and a coating thickness of one or two mils, the coatings on molybdenum and tungsten are primarily the disilicides WSi2 and MoSi2, respectively. During use of the coated re- fractory-metal system at high temperatures, in the absence of the silicon source, further diffusion causes growth of lower silicide compounds and depletion of the disilicides. Recent investigations have shown that the lower silicides are less effective than the disilicides in forming a glassy oxide film which impedes oxidation.' Consequently, failure of coatings at high temperatures may coincide with complete conversion of the disilicide phase. The purpose of this part of the study is to investigate the growth kinetics of the intermediate compounds in the W-Si and Mo-Si systems. There are three intermetallic compounds in the Mo-Si system: MoSi2 with a tetragonal structure

Jan 1, 1964

By Foster Bain

MOST of us householders in the East burn anthra-cite, so the problem that interests us most is the distribution and supply of the domestic sizes of anthracite. That, however, is only a small part of the coal problem and has its own peculiarities. Whether anthracite prices are high or low is largely a matter of relativity, depending on the size of the transaction, the available supply, and other things. As an isolated example which I hesitate to quote because it is not entirely representative, you may be interested to know that the Government fuel yard has been delivering to the Government buildings in the District of Columbia, white-ash stove coal at 16 c. a ton less during January, 1923, than during the same month of 1922. The reason is that the Government deals only with large companies who have been wise enough not to boost their prices and the small increase at the mine has been more than offset by the decrease in railroad rates. The total cost is $12.50 per ton, but I want to make it perfectly clear that it does not follow that this would be a just price for a retail dealer to charge the domestic consumer. For example, in the Government business there is no such thing as a bad debt, and certain other fixed or general charges are charged to other departments. The Bureau of Effi-ciency made an effort to determine what would be a. legitimate handling charge for a retail dealer, in Washington, and concluded that if the Government's cost of handling was $1.00, then $1.88 would be the corresponding fair charge to pay a retail dealer carrying the same volume of business, allowing for bad debts, complete depletion and depreciation of his equipment, and 10- per cent. profit. Assuming that this is the correct differential for Washington, it does not' follow that it would be equally just for Pittsburgh, New York, or elsewhere; nor does it by any means follow that it is a fair differential this year, because the dealer, after being unable to use his equipment in the summer months, now has to deliver an unusually large quantity of coal. I merely cite this instance to show that on the basis of primary cost of coal plus railway charges the retail price is not necessarily higher this-year than last.

Jan 4, 1923

By R. N. Stevens

GRAIN boundary sliding plays a part in the high-temperature deformation of polycrystalline solids, and many investigations1-11 have been carried out to determine the deformation contributed by this mechanism. The principal method employed has been to measure one of the components of the grain boundary displacement, see Fig. 1, at a large number of boundaries, and from the mean value so obtained to calculate the elongation contributed by sliding. A formula connecting these mean values and the elongation contributed is evidently required, and many have been derived. They are usually of the form where Egh is the tensile elongation contributed by sliding, is the mean value of the component, x, N is the mean number of grain boundaries intersecting a line of unit length, and K is a dimensionless parameter known as the averaging factor. Table I summarizes the equations used by various investigators. Examination of these shows that their derivations usually involve doubtful assumptions, and are often in serious error. The purpose of this note is to point out that rigorous solutions to the problem of calculating Egb from x have been available in the literature for some years.' Since this is so little known, and also the original derivations do not make some important points clear, it is proposed to rederive the basic equations. These will be extended further, and the conditions in which the extended equations may be applied will be indicated. The mean values of the components of the sliding vector (I, d, and I:, Fig. 1) will depend on the method by which boundaries are sampled in order to make measurements.3,11 Boundaries are usually sampled by marker lines parallel to the stress axis, or lines perpendicular to the stress axis, the measurement being made at the point where the marker line intercepts the boundary. We will denote the mean values determined by these various sampling methods as follows. lT, dT and vT are mean values obtained by making measurements at pointswhere a transverse line intercepts the boundary while lL, dL, and vl, are the mean values from a longitudinal line. The number of boundaries per unit length also may depend on the direction of measurement, and so we define NT and Nl to be the number of boundaries per unit length, after creep, in the transverse and longitudinal directions, respectively.

Jan 1, 1967

By Frank Nicholson

Copper ore was first noticed in Ste. Genevieve County in 1563. The diswvery mas made by a German farmer named Simon Grass, who had occasion to make a road from his farm down the hill into the neighboring valley, and while so engaged noticed pieces of a green-colored mineral, some of which he collected. Not knowing the character of the ore he had found, and supposing it to be valua ble for its brilliant green color only, he carried a piece with him into the town of Ste. Genevieve and showed it around the various bar rooms as a curiosity. No special attention was paid to the matter until nine years later, when Mr. 0. D. Harris sent a specimen of the ore to Dr. Theodore Fay, of St. Louis, for analysis. This analysis, which was the first ever made of Ste. Genevieve copper ore, showed the specimen to contain : Copper,. .........17.75 per cent. Zinc,..........3.21 " Iron........... 24.50 " Sulphur ..........1.09 " Organic matter, silica, carbon, and oxygen,. .. 53.45 " 100.00 " From this analysis it is evident that the sample was taken from near the surface, and was probably recognized by its green color. The almost entire absence of sulphur proves it to be an oxidized form. There is on file, at the office of the Ste. Genevieve Copper Works, a paper* giving three arsays of the ore made in 1875 by Henry B.

Jan 1, 1882

By A. J. Lanza

The prevention of illness among the employees of the mining industry is especially important in view of the importance of the industry, the unsettled conditions of labor, which emphasize the economic necessity for conservation of labor, and for humanitarian reasons. The burden of chronic illness is altogether unwarranted and unnecessary. If the industry is to maintain its labor supply in an efficient manner and keep the standard high, it is essential that it make every effort to keep its labor in good physical condition. The basic principles underlying health conservation are the same for any industry, but they are here applied especially to the mining industry, and more particularly, perhaps, to metal mining, as it is with this phase of the industry that the writer is most familiar, and as it is conceded that metal mining presents more health hazards than coal mining. The first step in the prevention of illness lies in securing employees who are in sound health and free from organic disease. I do not mean that all men who go underground should be physically perfect, but they should be free from organic defects of the heart, lungs, and other organs, or from anatomical defects that would markedly increase their liability to accident. The number of men working underground in mines in this country whose physical condition totally unfits them for such work presents a situation that should not be allowed to continue, and which I believe is not equalled in any other industry in the country. There is but one way to secure men who come up to the required standard, and that is by a thorough physical examination before employment. I am fully aware of the objections felt toward such a procedure by many employers and employees. There is much to be said on both sides, but to any one who is at all familiar with the mining industry as it is today in this country, and who approaches the matter with an unbiased mind, there can be absolutely no doubt as to the necessity for physical examination. As a purely logical proposition this is self-evident, but there remains the further necessity of placing physical examinations on such a basis that they can be administered in a just and equitable manner free from the abuses and suspicions to which they have been subject in the past.

Jan 1, 1919

By Edwin Ludlow

THERE has been so much misinformation sent out through the newspapers, and I find so few people who are really acquainted with the true facts in regard to the coal strike, that I feel it would be advan-tageous to present the facts as I have been able to gather them. One of the great points that the miners are now mak-ing is the fact that the operators of the central competi-tive field, consisting of the states of Illinois, Indiana, Ohio and Pennsylvania, have refused to meet them in a joint conference such as has been customary in previous years. There are two reasons why the operators de-clined to hold such a meeting. In the first place, the last time they held a meeting of this kind they were all indicted under the Sherman Act for making an agree-ment in restraint of trade. In the second place, the miners at their convention did not authorize their dele-gates to negotiate a new contract, but instructed them that nothing less than a scale based on the present wage, with the addition that there should be a six-hour day, a five-clay week, time j and a half for over time, and double time for Sundays, would be accepted. These additions would very much increase the cost of coal over the present wage scale. The operators could not consider renewing a contract on the present scale, as it would mean ruin for them, and as the miners were not authorized to negotiate a new scale on a fair basis that would enable the operators to mine coal in competi¬tion, with the open-shop mines the operators did not feel that there was any use in meeting the delegates. Thirty per cent. of the output of bituminous coal comes from the open-shop mines of West Virginia, Alabama and Pennsylvania. During the strike, in 1919, this 30 per cent. was increased to 50 per cent. and it has affected the operations of the union mines to such an extent this year that they have been able to run only 15 to 30 per cent. of their time, while the open shop mines have run from 60 to 80 per cent. as the wages in these mines had been reduced to the 1917 scale.

Jan 5, 1922

By A. Eilers

(Butte Meeting, August, 1913.) A NUMBER of the copper refineries in this country have lately separated some of the rarer metals from the slimes in the refinery tanks One of these has furnished me the following table of recoveries from given amounts of blister. I publish . this to show the proportions, to gold and silver and to each other, in which these rarer metals may be expected to occur in copper from certain localities.

Jan 6, 1913

By Henry F. Adams, Guy H. Ruggles

IN this paper the authors aim to chronicle the experience and salient points brought out in changing flotation reagents at a concentrator which had probably been using a mininium amount of oil at a minimum cost per ton, to a more complicated system requiring a partial chemical control. The original oils were coal tar and pine oil. The new reagents are lime, xanthate and pine oil. It was found that there were no insurmountable difficulties and eventually the cost of the new reagents per ton of ore proved to be very little more than that of the old. LIME PLANT When the question of handling lime came up there were two propositions open: (1) to handle the lime dry; (2) to handle it wet in the form of an emulsion. Granting that a dry lime feeder would function properly, delivering lime uniformly, the chief objection was the dust that would be raised when these feeders were being charged or operated, as the only possible location was on a gallery beneath which men were working. With this objection the handling of lime as an emulsion seemed to be the better scheme. The lime plant which was first installed was for experimental work on one section. This was gradually changed and developed until it was of sufficient size to accommodate 20 sections. The various changes will not be discussed and the following description deals only with the fully developed plant. Crude as it is, and perhaps somewhat expensive to operate, the lime-handling plant has been quite successful. Much has been learned which will pave the way for a better installation in the future. Burnt lime is received loose in standard gage box cars. This is the product of the kilns and has had no screening or sorting. A grab sample is taken from each car as it is unloaded. An average of the analyses of samples from 44 cars gave the following: [ ]

Jan 1, 1928

By Lo-Ching Chang, Kenneth M. Goldman

INTRODUCTION THE kinetics and mechanism of transfer of a constituent across a slag-metal interface are fundamentally important because many metallurgical processes involve the existence of a slag phase and a metal phase. Industrial processes seldom proceed to equilibrium, and information of the rates and modes of reactions are therefore often of greater practical importance than equilibrium data. A study of the kinetics of sulphur transfer across a slag-metal interface under reducing conditions will contribute a great deal toward a better understanding of desulphurization in the blast furnace process. The general principles existing in the transfer of a constituent, such as sulphur, across a slag-metal interface can be extended to similar transfer phenomena in other metallurgical processes. The present paper will demonstrate how such studies will lead to fruitful results from both the practical and theoretical standpoints. EXPERIMENTAL METHOD The apparatus used was simple. A graphite crucible 1 5/8-in. id, a wall 1/8-in. thick, and a depth of 4 in., was packed in an outer fused silica tube with powdered magnesia. The crucible assembly was rotated at a speed of 235 rpm inside an induction coil. A charge of 250 g of Armco ingot iron was introduced into the graphite crucible and melted by induction heating. The melt was held at a desired temperature [ ] for 3-5 min., after which 7.5 g of granular ferrous sulphide was added and the bath was held for another 3-5 min. A metal sample was then taken by means of a fused silica tube and a rubber suction bulb. Immediately after the metal sample was taken, 30 g of a prefused and crushed slag were introduced. The slag reached the temperature of the furnace in 1 to 1 1/ 2 min. Slag

Jan 1, 1948

By Paul Torrey

IN his recent article, Stanley C. Herold described some of the principles involved in oil-field rejuvenation. The purpose of this paper is to present certain facts, based upon observations of systematic water flooding of oil sands in the Bradford and Allegany fields of northwestern Pennsylvania and southwestern New York, which, it is believed, will partially substantiate Mr. Herold's opinions. The writer gratefully acknowledges his obligation to the Forest Oil Corpn., the Kendall Refining Co., the Petroleum Reclamation Co., W. J. Healey, of Bradford, Pa., and the Producers Oil Corpn., of Olean, N. Y., for information included in this article; and he is especially indebted to Jerry B. Newby, of the Petroleum Reclamation Co., for helpful suggestions and criticism in the preparation of the paper. It is obvious that, in the calculation of mathematical formulas for the prediction of oil recovery, uniform conditions must be assumed; otherwise, the applicability of the results of such work will be, by necessity, restricted to the area or individual property for which they were designed. Geologic and physical conditions are so variable throughout the fields where water flooding is practiced that it is difficult to devise methods that can be universally applicable. However, it is believed that there are certain fundamental principles which, when applied with specific modifications designed to meet the conditions existing in local areas or individual properties, may yield considerably greater recoveries of oil.

Jan 1, 1927

It is claimed that the first mining of copper by Americans in Arizona was done at Ajo, near the Mexican border, in 1854,* a year after this region had been added to the United States, under the terms of the Gadsden purchase. A group of adventurers, organized at Los Angeles under the name of the Arizona Mining & Trading Company, came, to the district. This party of prospectors consisted of 20 men, led by Major B. Allen, J. D. Wilson, and William Blanding. † They went first to Yuma, where the party divided, some of them going southward into Mexico while the remainder took the trail to Tinaja Alta, where they heard of the copper deposits of Ajo, which is 85 miles southeast from Yuma.‡ They found attractive outcrops of copper ore, and set to work energetically, but their operations suffered many interruptions, because that part of the country was still controlled by the Mexicans, who were uncertain about the exact position of the new international boundary. The Mexicans tried to expel them, but they held their ground. The first shipment of ore, consisting of native copper and cuprite, came from what is now the western end of of the New Cornelia workings; it was hauled in ox-carts to San Diego, 400 miles across the desert. § Later shipments were hauled to Yuma and floated on barges down the Colorado river to the Gulf of California, whence the ore was loaded on

Jan 1, 1932

By Rodolfo V. de la Cruz

The accurate assessment of the in situ deformability of rock masses is very important in the de and construction of underground openings for mining and geotechnical engineering applications. This paper presents a method and device for determining the in situ deformability of rock masses by relating the borehole jack load to the diametral displacements at a number of orientations around the borehole wall. The method called mu1tiple measurement method include the steps in which: a) diametral displacement sensors are installed at different orientations around the borehole to measure the changes in length of various diameters of the borehole; b) a uniaxial load is applied on two opposed octants of the borehole wall, and; c) using elastic theory and assuming a reasonable value for the Poisson's Ratio of the rock mass, the in situ deformability of the surrounding rock materials is determined from the measured diametral displacements and applied unidirectional load. The multiple measurement method has a significant advantage over existing methods since it provides several independent data in a single test from which to calculate the in situ deformability of the rock mass. The multiple data provide a check and counter check on the reliability and accuracy of the various measurements. The multiple data would also provide a means of quantifying the degree of fissuring on the borehole wall. Information on the degree of fissuring is very important in numerical modeling of rock masses.

Jan 1, 1984

By J. L. Whitlock

An attached growth aerobic biological treatment process has been developed at Homestake Mining Co.'s Lead operation which not only oxidizes free and complexed cyanides, including the stable iron complexed cyanides, but also thiocyanate, and the oxidation byproduct ammonia. Through the employment of a mutant strain of bacteria which has been gradually and specifically acclimated to the waste, these potential pollutants are mineralized to relatively harmless sulfates, carbonates, and nitrates. The resultant effluent, through toxicological testing, has been shown compatible with the receiving stream, which serves as a cold water marginal trout fishery.

Jan 1, 1985

By N. B. Wittman

Among the objects of interest which the members of the Institute are invited to examine in connection with this meeting is the Brown segmental wire-gun, which will be seen in process of construction at the works of the Pennsylvania Diamond Drill and Manufacturing Company, at Birdsboro. This is a 5-inch breech-loading (highpower) rifle, 19 feet long, constructed upon the system of Mr. J. H. Brown, an inventor of recognized genius in many branches of mechanical industry, and under the engineering direction of Lieutenant G. N. Whistler, Fifth U. S. Artillery. The system consists essentially of the subdivision of the inner core or tube of the gun, which takes up the initial compression, into longitudinal segments of such size that a higher physical condition and degree of special elasticity may be set up therein than it is possible to produce in the larger masses of metal heretofore used for the tube of the modern high-power cannon. The process consists in winding layers of square steel wire, subjected to a constant tension of 130,000 pounds per square inch, around these segments, which, when so held together, form the tube of the gun. A lining-tube may or may not be used. The steel segments used in the gun now under construction are 18 feet 4 inches long, and 3 inches thick at the breech, from which they taper to 0.8 inches at the muzzle. The physical condition of the segments has been determined by test-pieces, cut from each end of each segment, which show the following characteristics: Tensile strength, 165,000 pounds per square inch. Elastic limit, 105,000 pounds per square inch. Elongation at rupture, 14 per cent. in 2 inches. They are of crucible chrome steel, made by the Carpenter Steel Works, at Reading, Pa. The wire which holds them together has 0.005 square inches sectional area and is wound under a constant tension of 650 pounds on each strand, equivalent to 130,000 pounds per square inch. The winding has been done by a machine designed

Jan 1, 1893

By David H. Browne

The use of coal-dust fired reverberatory furnaces, or indeed of rever-beratory furnaces of any description, was for the Canadian Copper Co. a matter of necessity, and not of choice. For 20 years smelting had been done in blast furnaces alone, and with the Herreshoff furnaces used prior to 1904 there was no trouble in treating fine ores. But little flue dust was produced, and this, following the time-honored custom, was wetted down and put back with the charge. Whether the flue dust really smelted, or whether it was worn out by being chased around in a circle, was a problem that troubled no one. With the installation of modern blast furnaces and high-pressure blowing engines, in 1904, flue dust commenced to assert itself. Evidently more dust was made than could be smelted, but so many vital. problems engaged our attention at that time that this minor question was pushed to one side. In 1906, the details of blast-furnace smelting and the conversion of matte had been worked out to a satisfactory conclusion, and the ever-increasing piles of flue dust and fine ore on the stock yard demanded serious consideration. Numerous experiments in sintering, briquetting, mixing with converter slag, forming blocks of flue dust with green-ore fines and cement, and so on, were undertaken. None of these showed much promise. Our problem was still further complicated by the question of treating converter slag. The ore was basic, the slag was not needed as a furnace flux, and it was felt that under these conditions the old method of pouring slag in molds and resmelting in the blast furnace was an unnecessary expense. If the converter slag could be settled in basic-lined reverberatory furnaces, in which at the same time flue dust and green-ore fines could be smelted, two problems might thus be solved. Reverberatory practice with our ores was, however, an unknown factor. As carried on in the West, on siliceous ores and concentrates, at least 25 per cent. of fuel was required, and even this ratio depended greatly

Jan 1, 1915

The 38 companies listed below are those substantially engaged in the mining of metals, non- metals, or coal which appeared among the country's top 500 (on the basis of sales volume) listed in the July 1966 issue of Fortune. General Motors, with $20.7 billion in sales, heads the entire list, and a prominent member of the mining industry, Island Creek Coal, occupies the 500th spot with sales of $109.6 million. The lists of "industry medians" following mining's "Top 38" reveal how well the mining industry fared in 1965 in various salient categories of business performance.

Jan 9, 1966

By Frank Firmstone

It is proposed to consider here only comparisons made between results obtained when the materials employed are precisely the same, two furnaces at the same works for example, or the same furnace under different conditions as respects temperature of blast, method of charging, etc. With this limitation, the most important point in the comparison is the amount of fuel consumed per ton of iron. This, in fact, is the only item in the prime cost of the iron which can properly be said to vary in consequence of changes in the furnace process; for of the other principal items, the ore used per ton of iron, and also the amount of fluxing material, will depend on its richness, which, under the assumed conditions, will be practically invariable when taken over any considerable period; the labor cost also, from its nature, should be very nearly constant when estimated on the ton of iron. Moreover, as it consists principally in handling and moving material, it is greatly affected by circumstances quite independent of

F. C. HENDERSCHOTT,* New York, N. Y.-I am going to take, as the text of what I shall discuss, a portion of the second paragraph of Mr. Stanford's paper. It read as follows: "The most vital need of the management is for more brains. The logical course for an engineer ;in charge of an industrial organization is, therefore, to direct whatever forces may be at his command toward the development, training and guiding of-all the brains in his organization. This is easier said than done, for it involves the carrying on toward two objectives, the maximum of productive work and the development of the individuals, who perform it;" and the last sentence of the paper; "Industry has no greater asset than an intelligent workman." The United States of America, at the outbreak of the present war, had the best equipment of any nation in the world. That statement holds not only for our -shops and our transportation systems and our stores and our farms and our homes, but it holds generally.. There was no other nation in. the world that had so good equipment as the United States. There was no other nation in the world that could be called progressive that has such poorly trained workers as the United States. Even Norway, Denmark, Sweden and New Zealand ranked ahead of us. Now that is the, problem, generally speaking, and there was recognition of that fact, a growing recognition which-took form, and one of its forms was what we call the corporation school. Now obviously a corporation has no particular educational functions, but as is said in the closing sentence of Mr. Stanford's paper, or as is implied, an industrial institution is, after all, the reflection of the aggregate of the efficiency of the individuals who compose that organization. There is no way of getting away from that. If your individuals, from your president down, are efficient, if your individual has developed his brains, if he is a trained man, you have got an efficient organization, and if you haven't that condition, you have got weaknesses and your organization is no stronger than those weaknesses. I maintain that the same thing holds for a nation; .a nation is as efficient as the collective efficiency of the individuals who compose the nation, and no more so.

Jan 4, 1918

By Franklin J. Kay

Introduction The methods used in surface, mining vary throughout the world, depending on many variables such as: the types of over- burden materials, the depth of the overburden, the thickness of the seam, and other operational choices, both financial and functional. However, the stripping machines are typically draglines, stripping shovels, excavators, scrapers, loaders, and bulldozers. In all cases the economic feasibility of mining thin seams depends on the market needs, the market location, the equipment used, the mine management, the quality of the coal, and the labor market. Slight negative combinations of these parameters have closed many mines while positive combinations allow others to exist. The limited mobility or slow movement of large stripping shovels, draglines, and bucket wheel excavators make them more economically fitted to areas where large quantities of overburden are to be moved. This is necessary to keep the transfer and setup time between pits a negligibly small percent of the total operational time. Equipment such as the scrapers, loaders, backhoes, trucks, and bulldozers that are either very mobile or can be trailered is more versatile and can be used in the smaller area mines.* Mobile equipment is also more practical in areas with shallow overburden because of the constant movement required while removing overburden. In thin coal seams the equipment is usually paired such that the operational equipment can be listed as: (1) dragline-loader, (2) shovel-truck- loader, (3) scraper-loader. (4) loader-truck-loader, or (5) loader-loader. In these operations the overburden is removed with the first machine(s) of the pairing and the loader or loaders are used for coal production. Recent articles in Coal Age (Chironis, 1978) and Mining Engineering (Wiebmer, 1978) have discussed varying philosophies in mining similar seams of coal. One operator uses draglines for competitive mining while the other uses scrapers for the same purpose. Both cite strong cases sup- porting their choice of equipment as good sound economic decisions and their success supports their choices. The drag-line operation is adding new larger draglines to decrease the cost of mining, rather than to increase coal production. The scraper operation has also gone to more and larger scrapers for basically the same reasons. These operations and all

Jan 1, 1981

By H. F. Moore

We have studied the simple case of repeated stress, without considering impact, which might bring in other factors. This investigation has shown that steel under repeated applications of stress, reversed from positive to negative, will not fail below some fairly clearly defined limiting stress that, so far we can see, does not bear any definite relation to the ordinary elastic limit, being as large as the elastic limit in some cases and about one-half the elastic limit in others. According to one theory, fatigue is the result of changes in crystalline structure by repeated stress. I do not know of any evidence in favor of this theory. The crystals are broken under repeated stress. According to Bauschinger, under repeated stress some inherent property of the material changes its elastic limit; possibly some property of the amorphous cement between the crystals. A third theory is that all fatigue of metals is the result of the spread of damage from little localized overstresses. While I do not feel justified in accepting the last theory rather than the second, I have yet to run across a case of failure of steel or other metal, either in the laboratory or in service, that could not be explained by the gradual spread of damage from some nucleus. We have found, for example, that it is possible to subject a homogeneous steel, like Armco iron or a 0.90 per cent. carbon steel, that has been thoroughly annealed, to one hundred million reversals of stress as high as the elastic limit of the steel with no sign of failure. On the other hand, a complex structured steel, like a chrome-nickel steel that has been hardened with very little drawing, will fail at a stress about 58 per cent. of its elastic limit. Between these two we get a variety of results. In the steel work of any building, there are undoubtedly many thousands of places where structural damage was done—where riveted members were bent into place cold and where the rivets were fitted with drift pins, etc. Locally the steel in the building was stressed beyond its yield point in many places. These little localized overstresses do no damage at all; the building is entirely safe. But if this building were subjected eight or ten times a day to a mild earthquake, these little localized overstresses might be nuclei from which damage would spread

Jan 1, 1922