Search Documents

By V. C. Perini

The North Central Texas district as herein defined includes the counties of Brown, Callahan, Coleman, Comanche, Coryell, Eastland, Erath, Fisher, Hamilton, Haskell, Hood, Jones, Lampasas, Mills, Nolan, Palo Pinto, Parker, San Saba, Shackelford, Stephens, Stonewall, Summer-ville, Taylor, Throckmorton and Wise. These counties as listed are the counties of District 7B, State Railroad Commission. In the new designation of districts for the Transactions, the North Central Texas district includes most, but not all, of the counties of the previously classified West Central Texas district. This district covers the south central crest and the west flank of the Bend flexure north of the Llano uplift. The west flank of the Bend flexure is on the east boundary of the Permian Basin of West Texas. The greater part of the oil and gas produced in this district has been from Pennsylvanian strata. Some oil and gas has been produced from the lower Permian (by the latest, selection of the Pennsyl-vanian-Permian boundary) and minor amounts from the Mississippian and Ordovician sediments. Increasingly important quantities of oil and gas are to be expected from the older sediments by deeper development in proven fields and by new discoveries. Most of the oil and gas in the district has accumulated in localized, well-defined structural traps in limestone and sandstone reservoirs and in traps due to stratigraphic changes in sedimentation, mainly porosity traps in limestone and sandstone lenses, both usually associated with some structural feature. These features are found by surface and subsurface geological structural mapping, by core drilling and by geophysical surveying. Geochemical methods are also the basis for some wildcatting. The demand for sweet crude oil from this district is firm for local and national consumption, but the market price for the oil remains practically the same as in 1941. The price is not adequate to meet the requirements of the average independent operator with the continued advance in production, tievelopment and wildcatting costs. Stripper wells are being plugged daily and should be jealously saved for the future local and national reserves, especially in view of the fact that new discoveries of commercially important oil and gas reserves are not keeping pace with consumption. An expected economic disruption attendant upon the national war effort has arisen and continues in the future outlook for this district. The lack and delay of securing replacement parts for almost all oil-field equipment has handicapped and delayed operations and has decreased efficiency in wildcatting, development and production. Sufficient experienced drilling, scientific and office personnel is lacking because of inductions in the armed forces and migration to war industries during the period before manpower conservation became effective. The office personnel of the average oil and gas producer continues to be swamped by the increasing demands for reports by state and federal bureaus. An average of IQ days

Jan 1, 1944

By W. W. Norton

The gradually increasing proportion of sulphide ores which lead smelters of to-day are called upon to handle has caused the roasting problem to become one of ever greater importance. Mines have increased in depth, methods of concentration have improved, competition in the purchase of ores has become more keen and it has been clear that new and modern methods of eliminating sulphur must receive a large share of attention. We may look back a score of years or so and recall the old hand roasters, some of which turned out a finely divided roasted product, and others fashioned with a "fuse-box" wherein the roasted ore was slagged. Then came mechanical furnaces of several types, the Bruckner cylinder, the Brown-O'Harra, the Ropp, and others, all designed to cut out the cumbersome hand labor of the old reverberatories. Roasting costs were thereupon reduced and tonnages stimulated to a gratifying extent. However, it seems to have been early recognized that this very feature of a greater amount of roasted sulphide ore as compared with oxide ore brought with it a train of difficulties at the blast furnaces. The mechanically roasted ore was fine physically, the blast furnaces were choked and the troubles of the smelterman were not at an end. Then the Huntington-Heberlein pot system of roasting appeared upon the scene, and later on the Dwight-Lloyd machines were invented, and inasmuch as both of these processes were designed not only to roast the ore but to agglomerate it as well, it was apparent that distinct steps in advance had been achieved. At the Murray plant, modern up-to-date roasting practice is fully exemplified and there are now in successful operation roasting furnaces or devices of several sorts: namely, Godfrey revolving-hearth furnaces, Wedge multiple-hearth mechanical roasters, Dwight-Lloyd sintering machines, and Huntington-Heberlein pots. Godfrey and Wedge furnaces will properly handle material high in sulphur, say ores with 25, 30, and 35 per cent. of that element; D. & L. machines and H. & H. pots will positively not treat efficiently ores or mixtures containing anywhere near the sulphur content mentioned, but are confined to charges containing in the neighborhood of 15 or 18 per cent. In passing, it may also be explained that, so far as the knowledge of the writer goes, Godfrey and Wedge furnaces do not economically eliminate sulphur to an extent

Jan 1, 1915

By Patrick H. O’Neill

FUNDAMENTALLY a dredge designed for operating under arctic conditions and particularly when the temperature is below freezing is not greatly different than one for use in more moderate climates. Because of the accumulation of ice it is necessary to design suspension units to carry an increased weight when ice forms on the digging ladder or stacker. Winch room and house must be made with double walls and insulated. The dredge must be heated either by steam or by circulating warm air electrically heated. The stacker must be covered and heated to avoid having the drive units frozen tight. Actually the worst trouble that might develop in a stacker, because of cold weather, is the freezing of spilled material and ice accumulation on the head pulley and the underside of the belt. For this reason, stackers are completely enclosed and the pan kept open to get rid of any spilled material and to prevent overloading. Since the stacker is the most vulnerable place on a dredge in cold weather, heat is turned on there long before it is necessary any other place on the dredge. The principal points of interest in' connection with dredging, under subfreezing conditions, concern the preparation of the ground being dredged and the frozen ponds. In the interior of Alaska, and also in the Yukon territory of Canada, most of the gold bearing gravel deposits are permanently frozen and they are also covered by variable thicknesses of fine grained sediment, locally called muck. The better pay areas in this frozen ground have generally been worked by the earlier miners who sank shafts to bedrock in the frozen ground, then drifted in four directions and dug out the areas between the drifts as they re- treated toward the shaft. They thawed only the area they were excavating, doing so with steam from wood-fired boilers, and hoisted pay gravel to the surface for subsequent washing.

Jan 11, 1954

By N. E. Maccallum

(Cleveland Meeting, October, 1912.) ALMOST half a century has passed since the Siemens brothers, after tedious and costly experiments, finally began the manufacture of open-hearth steel. The furnace of that time was very small, having a hearth-capacity of from 3 to 4 tons, and we can appreciate the wonderful development that has taken place in those 50 years by the knowledge that to-day furnaces of 50 times the original capacity are in operation. As the new steel process assumed importance, and furnaces began to grow in number and in size, a question arose, to which no satisfactory answer has ever been given, and that was: ?What is the safe and economical limit of the open-hearth furnace?" Various stopping-places have been suggested, and 20, 30, 40, and 50 tons have each been advocated, but all such arbitrary restrictions have been passed, and 60 tons would probably represent the average furnace, although in a few instances from 70 to 80 tons are regularly tapped into one ladle. The furnaces which are the subject of the present paper have gone far beyond even these limits, and since, in my opinion, they mark another step in the progress of open-hearth steel, they will be considered fairly and frankly, in order that those to whom the matter may be of interest may be able to judge and form their own Conclusions. Short accounts of the work of these furnaces have appeared from time to time in various technical journals, of which the following is a brief résumé. These furnaces are the property of the Phoenix Iron Co., of Phoenixville, Pa., and their distinctive and important feature is, that they charge to twice the capacity of one ladle, and tap simultaneously into two ladles, with a device for controlling the flow into either. In this way the capacity of the furnace is not limited, as formerly, but heats of any size within the capacity of two ladles may be charged.

Oct 1, 1912

By R. L. Stephenson

IN plants casting electric-furnace ingots and converting them immediately into blooms, billets, slabs, bars or forgings, the criterion of quality performance is the soundness of the semifinished or finished product. When rejections are necessary, it is not always obvious whether the defects were caused by an unsound ingot or by improper treatment during the heating and forming operations. To evaluate the effects of the numerous factors involved, it is necessary to accumulate a large number of detailed observations and to analyze and interpret these data properly. Even after a thorough investigation has been conducted, it is impossible, in some cases, to isolate the cause of the defect and say that it is definitely the result of an unsound condition in the ingot or mistreatment of the ingot during further processing. Two factors that are extremely difficult to evaluate are the chemical composition of the steel and time from finish pour to charge in the soaking pits or "track time," and it is necessary to evaluate their combined effects before many other factors can be studied. Plain carbon, fine-grained steel in the range of 0.18 to 0.23 pct carbon, is very susceptible to cracking and the longer the elapsed time between the pouring of the ingots and stripping for charging into the soaking pits, the higher the billet rejections for cracks are found to be. Below 0.15 pct carbon and above 0.30 pct the track time

Jan 1, 1947

By D. W. Fuerstenau

A detailed study of the residence time distribution in an industrial balling circuit is presented. Using fired recycle pellets as a tracer impulse in each experiment, material was found to pass through the drum under plug flow conditions with a mean residence time of about 20 seconds. The time for a single pass through the entire circuit is about 60 seconds. Approximately three passes are required for fresh feed to report as product. Segregation of the tracer pellets in the system, according to their diameter, was observed and reasons for this are suggested. These results should be of potential use in the modeling and control of ball balling circuits in the iron ore industry.

Jan 1, 1985

By William B. Plank

NEVER before have so many men chosen the mineral technology field for their college training. In the college year 1936-'37, 7190 such students were enrolled in the 53 schools of the United States. My last study of the mineral technology schools of the United States and Canada was made for the year -1934-'35, and the current study, as the previous one was made for the A.I.M.E. under the auspices of the Division of Mineral Industry Education. Altogether, four studies of enrollment have been made. The first three were for successive college years beginning with 1932-'33. This present one covers a two-year period, for the purpose of these studies may he served satisfactorily if they are made biannually. The data on the schools were collected in the same way as for former studies, 'namely, by questionnaire, dated in this case Nov. 18, 1936. The figures reported are therefore comparable with those of former cars.

Jan 1, 1937

By James M. Stapleton

It has long been recognized by blastfurnace men that correct top distribution of materials is very important in efficient and economical furnace operation. Thousands of experiments on top design, filling, sizing, and other operations have been carried out to improving furnace performance by securing uniform distribution. Many of these experiments have worked out satisfactorily and have been passed along to the industry. Many of them have been discarded and forgotten. It has been the custom to judge the results of such experimentation by the results shown in improved or lowered furnace efficiency. While the effect on furnace practice will have to continue to be the final criterion of experimental changes, it has always been a desire of blast-furnace men to ascertain the reason for failure or success. Efficiency in the blast furnace approaching the maximum can be measured in several ways: (1) by actual operating results as revealed in the coke rate and tonnage figures, and (2), in the degree to which the top gas approaches the limiting ratio of CO/CO2 for the reduction of iron oxide by CO gas. This ratio is shown in the equilibrium curve in Fig. I for the system Fe-O-C and CO2-CO-C. The closer the top gases approach this limiting ratio (set at 1.08 by Martin1 because of the effect of limestone calcination) the more efficiently we may expect a furnace to operate. The fact that low CO/CO2 ratios parallel eficient operating performance caused us to investigate the cause of good or poor distribution and the effect of filling and other changes on furnace operation. Heretofore it has been possible to take a sample of furnace gas (usually from the dust catcher) and from it determine how closely the reducing gases were to the ideal or limiting value as set by the equilibrium diagram. However, to find just where in the top area the distribution was distorted and the top-gas CO/CO2 ratios were distant from equilibrium values has been a difficult and expensive investigational procedure. In an effort to overcome this difficulty, apparatus was designed to take gas samples and temperatures across the top radius of the furnace. It was hoped that with the aid of this apparatus traverse sampling of the ascending gases could be simplified, and the cost of sampling substantially lowered. Routine experimenting could then be maintained and results checked with furnace operation. It must be emphasized that a traverse sampling is necessary for a true picture, as single samples at any one point are unreliable. This statement can be verified by an examination of Tables 3 to 6, showing results obtained at our furnaces in a large number of tests. Apparatus The design of the equipment for taking the samples was governed by the need for providing a simple, routine test that could be performed fairly quickly without adversely affecting normal furnace operation.

Jan 1, 1943

By James M. Stapleton

It has long been recognized by blastfurnace men that correct top distribution of materials is very important in efficient and economical furnace operation. Thousands of experiments on top design, filling, sizing, and other operations have been carried out to improving furnace performance by securing uniform distribution. Many of these experiments have worked out satisfactorily and have been passed along to the industry. Many of them have been discarded and forgotten. It has been the custom to judge the results of such experimentation by the results shown in improved or lowered furnace efficiency. While the effect on furnace practice will have to continue to be the final criterion of experimental changes, it has always been a desire of blast-furnace men to ascertain the reason for failure or success. Efficiency in the blast furnace approaching the maximum can be measured in several ways: (1) by actual operating results as revealed in the coke rate and tonnage figures, and (2), in the degree to which the top gas approaches the limiting ratio of CO/CO2 for the reduction of iron oxide by CO gas. This ratio is shown in the equilibrium curve in Fig. I for the system Fe-O-C and CO2-CO-C. The closer the top gases approach this limiting ratio (set at 1.08 by Martin1 because of the effect of limestone calcination) the more efficiently we may expect a furnace to operate. The fact that low CO/CO2 ratios parallel eficient operating performance caused us to investigate the cause of good or poor distribution and the effect of filling and other changes on furnace operation. Heretofore it has been possible to take a sample of furnace gas (usually from the dust catcher) and from it determine how closely the reducing gases were to the ideal or limiting value as set by the equilibrium diagram. However, to find just where in the top area the distribution was distorted and the top-gas CO/CO2 ratios were distant from equilibrium values has been a difficult and expensive investigational procedure. In an effort to overcome this difficulty, apparatus was designed to take gas samples and temperatures across the top radius of the furnace. It was hoped that with the aid of this apparatus traverse sampling of the ascending gases could be simplified, and the cost of sampling substantially lowered. Routine experimenting could then be maintained and results checked with furnace operation. It must be emphasized that a traverse sampling is necessary for a true picture, as single samples at any one point are unreliable. This statement can be verified by an examination of Tables 3 to 6, showing results obtained at our furnaces in a large number of tests. Apparatus The design of the equipment for taking the samples was governed by the need for providing a simple, routine test that could be performed fairly quickly without adversely affecting normal furnace operation.

Jan 1, 1943

By William E. Ford, Edward Salisbury Dana

A. Anhydrous Carbonates The Anhydrous Carbonates include two distinct isomorphous groups, the CALCITE GROUP and the ARAGONITE GROUP. The metallic elements

Jan 1, 1922

By E. H., Edwerd H. Robie

WILLIAM CHURCH was one of the founders and the first president of the Detroit Copper Mining Co. and was the first man to interest the Phelps Dodge company in the possibilities of the Morenci district, in the '70's. One rather interesting story concerning him is told by James Colquhoun, at one time president of the Arizona Copper Co., in his "History of the Clifton-Morenci Mining District." It seems that a geologist visited the Detroit company's mines and gave Alec McLean, then mine superintendent, quite a lesson on the proper mineralogical terms for what he saw in passing. Later, McLean handed out a few of them to Mr. Church as they walked through the mine on a tour of inspection. The latter had even less regard for a geologist than he did for sulphide ore, which meant little indeed, for in those days the rich ore was all oxidized, and there was no money in the sulphides. So Church was completely unimpressed by the geological and mineralogical terms that McLean had acquired from the visiting geologist. Church turned on him and warned him that so long as he was a member of his staff he was to understand that there were only two kinds of rock in his mines-the one was smelting ore and the other was waste rock, and he was not to forget it.

Jan 1, 1942

So commonplace that they are seldom noticed, mining-machine bits have a defi-nite and important bearing on the cost of coal production. At the average mine many thousands of bits are used during the year. The actual cost represented by the pur-chase of the bits used, as well as that of sharpening and handling them, represents a considerable sum of money. From the point of view of indirect costs they are im-portant, as the lack of good bits at a mining machine can seriously impair production. Good bits tend to ensure an adequate supply of coal.

Jan 1, 1940

By L. Massé, Th. Koulomzine

Note on Haalck's Formula Quantitative interpretation of magnetic anomalies is admittedly a difficult process. Few authors have attempted a general approach to this problem. A number of publications can be found widely scattered through scientific periodicals, which present particular solutions obtained by the introduction of simplifying assumptions, such as the reduction to an equivalent magnet or pole, or the hypothesis of a very thin vein. The practical applications of these solutions are necessarily very restricted though they have given satisfactory results in isolated cases. Probably the most widely known equation of this type was given by Haalck.' His classic work contains curves calculated for typical conditions, and has now become a standard reference. However, a close scrutiny of Haalck's derivation reveals that, aside from the hypothesis of a thin vein, it rests on the

Jan 1, 1949

By M. E. Wadsworth

The chief value of a paper like this consists in its statistical tables, putting on record material useful to future inquirers. The data here given have been compiled from time to time as far back as 1874, and subsequently brought up to date. Much of the data could only be procured by personal inspection of the catalogues, as for example those relating to the Lawrence Scientific School, which were taken at that time from the catalogue of Harvard University, on file in the University Library, by the present writer, then an instructor in that institution. These tables of statistics were first prepared for publication by the present writer in 1888-1890 and published in his "Report to the Board of Control of the Michigan College of Mines " (Michigan Mining School) for 1886 to 1890. Later they were revised, enlarged and brought down to date in his reports to the same Board for 1890-92; and since that time they have been, at his request, revised, greatly enlarged and kept up to date by Mrs. Frances H. Scott, the efficient Librarian and Secretary of the Michigan College of Mines. Great credit is due to her for her devotion to this work and her untiring efforts to have the statistics complete as far as possible. The accompanying running commentary is given simply as a string to hold these statistics together, and in order to call attention to a few points. In compiling any tables like those accompanying this paper, one is confronted by serious difficulties, some of which render these results only approximate. In order to insure greater correctness and fullness in the statistics to be published the writer's reports for 1890-92 were sent to the president of every college, university, and technical school in the country, so far as known. Accompanying this was a personal letter which asked of each one his aid in correcting any errors, and in giving fuller statistics. But relatively few answers were obtained, and subsequently from one to many letters have been sent to each institution not formerly reply-

Jan 1, 1898

By George F. Comstock

IT has been reported recently1 that the hardenability and toughness of forging steels may be improved appreciably by the use of complex deoxidizers containing titanium, aluminum, and vanadium. In the light of these results, an investigation of similar deoxidizers containing various hardening and grain-refining elements was proposed. These deoxidizers included three manganese-silicon-aluminum-titanium al- [ ] loys with and without boron and calcium, three titanium-aluminum ferroalloys with vanadium, molybdenum and zirconium, respectively, an aluminum-zirconium ferro-alloy, and plain ferroboron. Their approximate compositions are given in Table I, and the results obtained so far in the investigation of their effects on steel are presented in the later parts of this paper. METHOD OF USING DEOXIDIZERS The deoxidizers listed in Table I were used in three series of 0.40 per cent carbon steels, which were melted in a small basic-lined induction furnace and cast in the form of I7-lb. ingots about 2.75 in. square. The basic charge for each heat was 10 lb. Armco iron and 6 lb. of clean low-carbon sheet-steel scrap. When melted, this charge was deoxidized with 15 grams of ferrosilicon, then about 500 grams of pig iron was added to give the desired carbon content. After the pig iron was dissolved, nickel [ ] shot and high-carbon ferrochromium were added if required, and additions of 80 per cent high-carbon ferromanganese and 50 per cent ferrosilicon were made to provide the proper amounts of those elements. Aluminum was added to every heat 2 min. after the silicon, for grain-size control and preliminary deoxidation. One minute later the special deoxidizer was added to the clean bare surface of the molten steel, and 2 min. afterward the steel was poured into the ingot mold, using a hot-top and pipe-preventing compound after pouring to decrease the depth of the shrinkage cavity. The special deoxidizers were used mostly in the form of about ½ to ¾ -in. lumps,

Jan 1, 1942

By R. A. Walker, R. G. Bradburn, B. C. Flintoff

The events which led to computer control of the grinding circuit at Brenda Mines are described, beginning with mill operating philosophy. High production rates are encouraged so that, despite severe upstream disturbances, a manual control method to meet this philosophy had developed. Unfortunately, due to the time interval between manual control adjustments, losses in production resulted. An analog control method was devised and, although superior to manual control in some situations, was found to suffer from several practical limitations. The data logging capabilities of a control computer seemed warranted. Circuit testwork combined with an analysis of operating data showed that (1) metallurgical personnel at times could increase production rates by as much as 10% over that achieved by grinding operators; (2) the chosen data base required an increase of 2% or more to be statistically significant; and (3) for a 2% increase there would be a significant rate of return of investment in computer installation. Consequently one grinding circuit was fully instrumented, and a rented computer was installed by Brenda personnel. Prior to installation, Brenda metallurgical and instrumentation personnel were trained in the use of the computer. An experimental program was devised with short and long-term objectives. The short-term objective was to develop a practical control strategy which would yield immediate production improvement. This goal has been successfully met as described in this paper. Statistically significant increases in excess of 2% have been established. Work is being completed for the permanent installation of computer control on all four grinding circuits at the present time. Long-term objectives are (1) off-line analysis of the circuit using mathematical models of process units; (2) on-line experiments to study alternative control tactics; and (3) possible expansion of computer facilities to the crushing plant.

Jan 1, 1978

NEW MEMBERS The following list comprises the names of those persons who became members during the period of Septa 10, 1917, to Oct. 10, 1917. ALDER, ALFRED 1020 Kansas City St., Rapid City, So. Dakota. AUBEL, VERNE WALLACE, Night Supt. of Works, Broken Hill Proprietary Co., Ltd., Newcastle, N. S. W., Aust. BAKER, ROBERT EARL Works Mgr., Keokuk Electro-Metals Co., Keokuk, Iowa. BALLAMY, JOHN H., Min. Engr., Delaware, Lackawanna & Western Railroad Co,Scranton, Pa. BRADLEY, ALLEN PALMER, Met., American Smelting & Refining Co., Murray, Utah. BRIGGS, WALTER M Pres. Regal Mines Co. Dedham, Mass. BRONSON, EDMOND B., Pres., Bagdad Copper Co., 120 Broadway, New York, N. Y. BROOK, NILES WINFIELD, District. Mgr., Cleveland Min. Co., 625 Bank & Insurance Bldg., Dubuque, BUSKETT, EVANS 1V., Assayer, Custom Assay Office, 620 Joplin St., Joplin, Mo. CAMPBELL, JOHN HATES, Chem. Engr., Robert W. Hunt. & Co., 2200 Insurance Exchange, Chicago, Ill.

Jan 11, 1917

By C. R. Austin

Tan investigation described in this paper deals with the development of high-temperature alloys of the Konel series over a considerable period of time at the Research Laboratories of the Westinghouse Electric and Manufacturing Co. The word "Konel" has been coined to designate a class of alloys which are essentially nickel-cobalt-iron with other elements added, usually in small quantities. Most of these alloys contain titanium, usually about 2.5 per cent. The main consideration related to strength at high temperatures but the thermionic characteristics of the alloys was also of considerable importance. It is, however, with the mechanical properties at elevated temperatures that we are herein concerned. Much attention has also been devoted to their age-hardening characteristics. There is little need to stress the utilitarian aspect of this subject, as it has been discussed frequently. Alloys exhibiting superior properties at normal or elevated temperatures rapidly find many suitable applications. There may be some question, however, as to what constitutes a "superior alloy, " or an alloy that may have good service possibilities at high temperatures. The problem is complex and various means have been employed to predict the useful life that may be expected of a material. In recent years the "creep limit" has been used as the best means of evaluation. Such tests necessarily are prolonged and hence collection of data is very slow. In a general survey of many alloys therefore the creep test is not desirable and recourse must be had to much shorter time tests in order more rapidly to select alloys which appear to show promise. For this reason the present survey has been made using the high-temperature tensile test and a bend test developed by Howard Scott.1 In the tensile test the conditions have been standardized for all tests, so that values obtained are at least comparable, one alloy with another, although the absolute values recorded merely indicate the mechanical properties at the given temperature under a definite rate of loading.

Jan 1, 1931

By James Douglas

No one mill attribute it to the partiality of friendship if I claim for Thomas Sterry Hunt the right to a place among that little band of men whom nature has endowed with what seems an almost magical power of insight into her mysteries. But those of us who knew him well, while me never ceased to wonder at the breadth of his genius, the fertility of his imagination, and the wonderful capacity of his memory, learned that these native powers would have availed little to win him his rank in science, had they not been stimulated by ceaseless mental activity, indomitable perseverance, and unflagging industry. Thomas Sterry Hunt was born in Norwich, Connecticut, on September 5, 1826. He came on his mother's side from that good old stock which gave to Puritan England the gentle mystic Peter Sterry and that uncompromising preacher Thomas Sterry, who wrote the notable tract, " The Rot among the Bishops " in 1667, and that gave to New England Consider and John Sterry, the mathematicians. While lie was a child the family moved to Poughkeepsie on the Hudson. There the father died when Thomas, the oldest son, was twelve. The mother and her family of six young children returned to the old home in Connecticut. For a short time Thomas attended the public school; but it was for a short time only, as he, thus early, was required to share the burden of the family support and to seek employment. This he found first in a printing-office, then with an apothecary, afterwards in a book-store. The surroundings may have been congenial, but the occupation probably was not; for he re-

Jan 1, 1893

By Erwin S. Sperry

The formation of cracks in metals is one of the most perplexing obstacles encountered during the process of rolling. When occurring in brass they may be due to several causes: 1. Shrinkage-cracks. These are caused by a rough or cracked mould which prevents the metal from shrinking uniformly and without rupture. It is consequently torn apart. In the majority of instances such cracks are visible to the naked eye, and often penetrate the metal to a considerable depth; in others they are nearly if not quite microscopic, but

Jan 1, 1899