Search Documents

Jan 1, 1946

By William M. Brewer

Until quite recently, in fact within the past two gears, but little attention has been given to the outcrops on the west coast of Vancouver Island, and their copper-contents. During the past few months the writer has been engaged in examining and developing some of these prospects. In many respects he has found characteristics associated with these prospects which, in his experience, are unique. In the first place, nearly all the outcroppings which overlie chalcopyrite-ore along the west coast of the island are composed of a high-grade magnetite. The magnetic qualities of some of these outcrops arc so pronounced that the magnetite possesses polarity. Although the writer hinlsclf had no analyses of this magnetite made, yet from its appearance he can readily believe that analyses made for other parties, which show a yield' of 62 or 63 per cent. of metallic iron, with only traces of phosphorus and silica, are correct. At a very shallow depth, masses of chalcopyrite, yielding in some instances as high as 32.6 per cent. of copper, occur associated with the magnetite; and in one instance, which recently came under the writer's observation, a solid body of high-grade chalcopyrite, fully 4 feet in thickness, occurs Within 6 feet of ; the surface. Usually this solid sulphide-ore carries low values in gold. Probably ail average of' $2 per toil would be fairly representative. But in some instances development has determined the occurrence of narrow stringers of pyritous quartz, associated with the sulphide-ores, which yield by assay more than $20 per ton in gold. In the districts examined by the writer, the country-rock is usually crystalline limestone, with dikes of igneous rock as intrusions. Sometimes the outcrops of magnetic iron-ore are

Jan 1, 1900

By Dewitt T. Ring

Oil development in 1934 has been largely restricted to drilling offset and lease requirement wells. Practically no effort has been made to discover new fields, chiefly because of price structure and proration allowance. A summary of developments during 1934 of necessity is brief. No state agency is maintained for the collection of data. The State Bureau of Mines issues permits for drilling and inspects the plugging of wells in coal-bearing measures only. Therefore data are not in existence with which to compile tables similar to those in other papers of this symposium. Forty-seven counties in the state produce oil, productive depths ranging from 100 to 5300 ft. Approximately 35 different formations in Ohio have produced either oil or gas. These reservoir rocks occur through a stratigraphic interval ranging in age from the shallow Goose Run sand of the Upper Pennsylvanian in Washington County to the St. Peter sand of Lower Ordovician in Seneca County. However, 75 per cent of these formations are unimportant from the standpoint of real production. Most of the oil and gas in Ohio has been produced from the Clinton, Oriskany, Berea and Big Injun sands and the Trenton Lime. Of these, the Clinton is first and Berea second in importance. Fields The oil and gas-producing territory of the state can be grouped into five fields or districts, made up of a series of "pools" usually separated by nonproductive areas, and is in no sense a belt of continuous production. Trenton Field of northwestern Ohio, is scattered through 15 counties. Development began in 1885. This field reached its peak in 1896 with a production of 20,575,138 bbl. for the year. For 1934, production had declined to 990,000 bbl. The oil is classified as Lima grade. Clinton Field extends from the Ohio River in Lawrence County to Lake Erie in Lorain, Cuyahoga and Ashtabula counties. First Clinton-

Jan 1, 1935

By Sebastian Paul Musco

The importance of the recovery of precious metals, specifically platinum group metals, was recently expressed by Robert McNamara, former head of the World Bank, who stated that the United States should stockpile strategic metals. More than 90% of the platinum metals are imported into the United States from South Africa and Russia. Mr. McNamara's concern is based on the possibility of a disruption of obtaining these metals due to any racial unrest in South Africa and a renewal of the Cold War with Russia. The United States has a platinum mine from coast to coast. Since 1974, the automotive industry has consumed 3 million ounces of platinum and palladium, and the recovery of these metals from auto catalyst converters is estimated to be 300,000 ounces of platinum per year. This would make the United States the 4th largest producer in the world.

Jan 1, 1984

By C. H. Curtis

The Esperanza mine of the Copper Division of Duval Sulphur & Potash Co. is located in the Twin Buttes District, 32 miles southwest of Tucson, Ariz. Records of mining activity in this vicinity date back to the 1850's. The New Year's Eve mine, which was last active briefly during the years of World War II, made the first actual penetration into what is now Duval's Esperanza property. Duval's exploration department became interested in the Esperanza in the fall of 1954. The first test hole was started on May 8, 1955. Two years of intensive exploration work together with metallurgical studies developed the basis for the decision to establish an operation.

Jan 11, 1961

By R. C. Croft

Chlorination is frequently employed as a method of chemical attack and separation of the constituents of minerals. Occasionally, bromination has been used similarly. Separation of metal halides resulting from the use of either of these halogens generally entails difficult fractional distillation of the halides. Recovery of the halogens for re-use has often been attempted in the case of chlorine by oxidation of chlorides. Though attractive thermodynamically, this method of recovery fails due to the formation of stable oxychlorides. Although a little less reactive than either chlorine or bromine, iodine offers advantages in the field of mineral treatment. It has been shown, for example, that iron may be removed completely from ilmenite by iodinating the latter in the presence of carbon or hydrogen at temperatures near 800°c where good reaction rates are obtained. Under these conditions titanium does not form an iodide and is not transported in the vapour phase as is the Fe12 formed. Other good separations obtained between zinc and iron in willemite and tin and tungsten in scheelite concentrates confirm that separations of metal iodides by sublimation are more specific than similar separations attempted by distillation of volatile chlorides. Readiness of iodides to form stable vapour-phase complexes appears to contribute to this specificity. The very low stabilities of most metal oxyiodides permit quantitative recovery of iodine by oxidation of metal iodides. In the case of Fe12 derived from iodination of ilmenite, full recovery of iodine can be achieved by air oxidation of this compound at temperatures ranging from 350 to 500° C.

Jan 1, 1973

By W. O. Borcherdt

THE 650-ton underground mill of The Empire Zinc Company of Colorado (a subsidiary of The New Jersey Zinc Co.) serves the Eagle mine at Gilman in the Battle Mountain mining district of Colorado, The topography is rough, the mine and mill being under Battle Mountain and the adit entrance in the north wall of the canyon of the Eagle River, which is here a steep gorge. The town of Gilman stands on a shoulder of the mountain at altitude 8975 ft., above sea level, along the Pike's Peak ocean to ocean highway and contains the company office and the usual facilities of a modern mining town. The railway contact is at Belden, altitude 8370 ft., at the bottom of the gorge directly below Gilman, where the westbound track of the Denver and Rio Grande Western Railroad is cut along the north bank and the eastbound track along the south bank of the river. A general description of the mine, mill and service facilities, with an account of the geology, character of the ore and methods of mining and milling having previously been published in a series of articles1, this paper will discuss with as little repetition as practicable the reasons for building the mill underground rather than as a conventional structure, the methods of construction employed and the experiences gained in what is apparently the pioneer plant of this type and extent. There is a large historical background for underground plant construction. From the dim past caves have been used for dwellings, stables, storage and shops. It was natural that where caves were available their many advantages led to their enlargement to suit convenience and where they were not available but the ground conditions, topography and drainage were favorable, to the construction of artificial caves. In the vicinity of Epernay, France, extensive underground chambers protect the manufacture and storage of champagne-an intricate and ingenious application of chemical engineering-and in recent years these plants have been mechanized by the installation of conveyor systems. In the

Jan 1, 1937

By H. R. Aldrich

This paper presents some of the results obtained during the field season of 1919 while mapping, in detail, the stratigraphy of the Gogebic Range in Wisconsin. The detailed stratigraphic section for the range was first published, in 1919, by W. 0. Hotchkiss,1 State Gcologist of Wisconsin. The idea of the continuity of the definite succession thus established was the result of underground work in the productive part of the range in Michigan and Wisconsin, where most of the required data also were obtained. The bulk of the range in Wisconsin is heavily covered by drift; in fact, there is but one satisfactorily exposed section across the entire formation, few partial sections, and only scanty outcrops of any size. Evidently mapping of the detailed members of the formation could only be established at a few points; and if from these small areas the .unexposed portions were to be mapped by projection, the only apparent method for projection was straight-line interpolation modified or supported by data from the scanty outcrops and from test workings. It was found that by tracing a magnetic line, the position of which in the formation may be determined from test pits and outcrops, the projection of that member through the drift-covered areas can be accomplished. The magnetic line serves as a datum horizon within the formation to which reference may be made in correlating outcrops and ledge matter thrown from test workings. The foot wall is easily located also, and from the two datum horizons thus available and the known proportional thickness of the several members in the type sections, outcrops and artificially exposed ledges can be correlated by lithological characteristics. As a result, the tracing of magnetic lines became the basis of our work on the stratigraphy and we were able to carry the mapping of detailed members throughout the length of the range. It was found that the magnetic horizon at the base of the Tyler slate was most easily followed by the dip needle; that discontinuity caused by cross-faulting could be established; that folding along the strike could be detected in a repetition

Jan 1, 1922

By G. W. Tracy, R. D. Carter

A method based on an approach by Hurst' is developed for calculating water influx behavior. Using this method, superposition calculations may be eliminated. The principal diflerence between this method and that of Hurst is that, over finite intervals of time, constant oil-productton rates are assumed by Hurst whereas constant water influx rates are assumed in the presently described method. By adopting the assumption of constant water influx rate for finite time periods, a method is developed for calculating water influx behavior which lends itself to convenient combination with the material balance equation in the Schilthuis form. Above the bubble point, an equation is given for the explicit step-wise calculation of pressure history for a prescribed oil-production history. Use of this method of calculating water influx behavior should result in a considerable saving of time as compared to superposition calculations when the calculations must be made without using a digital computer. The connection is shown between the method described herein and the method developed by van Ever-dingen and Hursr2 (which requires superposition cal-culations). Also, results obtained using the two methods are compared.

THE morning session on non-metallic minerals was opened briefly by the chairman, Oliver Bowles, giving the usual instructions and then turning immediately to the introduction of the speakers. "Min-ing and Preparation of the St. Peter Sandstone in' Arkansas" by D. D. Dunkin was first presented by W. M. Weigel, mining technologist for the Missouri Pacific R. R. The paper covered a discussion of the geology of the sandstone, correlating it with the sand-stones at Klondike, Mo., and at Ottawa, III. Its physical appearance, structural characteristics and thickness; the mining methods, blasting, loading, haulage and classification were taken up. A series of slides aided in clarifying the details. The discussion consisted of questioning Mr. Weigel on operating details. The second paper, "Geology and Mining of Diaspore and Flint Fire-Clays of Missouri," written by H. S. McQueen, geologist with the Missouri Bureau of Geology and Mines, and C. R. Forbes, professor of mining, Missouri School of Mines and Metallurgy, was presented by Mr. Forbes. It thoroughly covered the geology, stressing the peculiar sink formations in which all diaspore clay deposits of the region occur. The crudeness of the mining methods was mentioned and believed justifiable because of the small size of the indi-vidual deposits. It was stated that the diaspore and flint clay deposits of Missouri were confined to an area 60 mi. square and it was found in a survey that the average deposit had dimensions comparable to the fol-lowing: 100 ft. long, 75 ft. wide and 50 ft. deep. The largest deposit found was 800 ft. long. 500 ft. wide and 20 ft. deep. It is estimated that the known reserve of clays is from 750.000 to 1,500,000 tons. Mr. Forbes stressed the fact that in mining clays the major problem is entirely one of sorting. This is done under the supervision of a pit foreman trained to make this classification by eye. The mine-run clays are sepa-rated into diaspore clay ranging in alumina content from 70 to about 85 per cent, first grade Burly clay 00 to 70 per cent, and Flint clay 52 to 60 per cent. Discussion of the paper consisted in questioning Pro-fessor Forbes on labor. cost, haulage, drilling, and geology.

Jan 3, 1928

By Wallace F. Lovejoy

The conservation of mineral resources as a public policy question has received a great deal of attention in the area of petroleum and practically no attention in other mineral areas, except, of course, soil and water. Some insights into mineral conservation problems and how these problems can be approached in the social decision-making process are significant. THE MEANINGS OF CONSERVATION Conservation means many things to many people .2, 11 Too often it is a label attached to a proposed program or policy to add respectability and to gain public acceptance. It is a term applied to specific resources in some contexts and to resources generally in other contexts. The oft-quoted comment of President William Howard Taft that conservation is something everybody is for, no matter what it means, succinctly illustrates the confusion over both the meaning of the concept and the policies required to implement a conservation program. This confusion goes quite deep and is more than a matter of semantics or emphasis. For example, the efficient exploitation of a mineral deposit which might be applauded as good conservation in terms of that mineral, might well result in destruction of or harm to the environment which might be viewed as bad conservation. Economics provides some tools to help solve this sort of problem, but there may be fundamental differences in value systems among groups in this society, so that ultimately society must collectively make a subjective political choice as to what is good or bad conservation policy.a An Economic Definition of Conservation The economic concepts of conservation have more relevance and are more easily applied to minerals than to many other natural resources. This does not mean that economics goes all the way in understanding mineral conservation, but at least it provides a framework for analysis which later can be altered to include noneconomic variables. Perhaps the greatest confusion in discussing mineral conservation arises in attempting to relate purely physical concepts with economic concepts. An economist is interested in maximizing or conserving values; he is not interested in maximum physical recovery or utilization. He wants to know the costs and benefits of using, or not using, resources. The economic theory of mineral-resource conservation can be labeled variously as the theory of exhaustion, the theory of resource exploitation, the theory of resource utilization, or simply as a special case of the broad theory of capital.4,28 Consideration of these labels illuminates the concerns of the econo-

Jan 1, 1976

By G. S. Hume

In Alberta, an oil well with an estimated capacity of 60 to 100 bbl. was "brought in" at Wainwright, the oil being derived from the Colorado group of rocks. In the Norman fields, Mackenzie River area, two new holes completed showed no oil in commercial quantities whereas the "Discovery" well has now an increased flow amounting to about 100 bbl. In Kent Co., Ontario, a well from low down in the Trenton formation gave an initial flow estimated at 200 bbl. This opens up new possibilities for the Ontario fields as this is the first oil from this depth in the Trenton. ACTIVE drilling operations for oil were carried on during the summer of 1923 in various parts of western Canada, particularly in Alberta. During the early part of the summer, interest centered in southern Alberta, in the Coutts area, just north of the producing Kevin-Sunburst field of Montana; but on Nov. 14 a strike of oil by British Petroleums Ltd. at its well near Wainwright, 120 mi. east of Edmonton, diverted public attention to a new field. According to present estimates, this well has a possible production of from 60 to 100 bbl. a day; the oil is 15° Bé. with an asphaltic base and is obtained from a sand at the depth of 2019-2036 ft. The oil stands at 1700 ft. in the casing and there is sufficient gas pressure to cause it to flow when the well is opened. The oil horizon is in a sandstone in rocks of Colorado age. It is known that there is a terrace structure of which the western edge extends in a northwesterly direction from the vicinity of Hardisty to Viking, where it probably swings more to the north. At Viking a small dome-like structure on this terrace gives rise to an important gas field. Wainwright lies directly northeast of Hardisty. The exact structure in the Wainwright field is unknown. At Fabyan, 6 mi. northwest of Wainwright, the Imperial Oil Co.1 last year brought in a large gas well estimated at 14,000,000 cu. ft.; the escaping gas sprayed a small amount of heavy black oil over the derrick. The gas was obtained at a depth of

Jan 2, 1924

By W. B. Hincke, C. G. Maier

Commercial processes of lead softening directly involve the behavior on fusion of mixtures of the oxides of antimony and lead, and the vapor pressures of these materials. Practically no quantitative data have been available for the discussion of lead softening from the point of view of the fundamental metallurgy of the process. The Pacific Experiment Station of the United States Bureau of Mines, in cooperation with the University of California, in the course of its program of study of the fundamental properties of oxides and sulfides, recently made vapor pressure measurements upon the various forms of pure antimony trioxide.1 When these measurements were considered in connection with the softening and "fuming" furnaces which are often used for lead purification, an unexpectedly anomalous condition appeared. In brief, this anomaly consisted in the fact that in certain softening furnaces, where lead bullion of approximately 1 per cent antimony content was being treated at a temperature of approximately 1300" F., no appreciable volatilization of antimony occurred, there being produced only a "skim" containing from 15 to 20 per cent Sb, 65 to 60 per cent Pb, as mixed oxides. In the "fuming furnaces," where the temperature was only about 1400" F. and where an accumulation of "skims" from the primary softener was being treated with certain other antimony-containing materials, a considerable volatilization of antimony as oxide occurred; an "impure" skim of some 60 to 65 per cent Sb and 10 per cent Pb was also produced as mixed oxides. The latter material cannot be suitably treated in hearth-type or softener furnaces. The essential chemical difference between the first and second step is the customary use of some reducing material (such as petroleum coke) in the later operation. An obvious interpretation of the effect of coke in facilitating the fuming of the antimony is the supposition that the antimony may be reduced to metal, and volatilized as such. This supposition is improbable, however, when it is realized that a slight extrapolation of the vapor-

Jan 1, 1932

By H. M. Chance

Probably no other prosperous mining district is so little known as the Black Hills. The name leads one to assume that the district is a rolling country, consisting of more or less insignificant hills. Nothing could be more untrue, as the Black Hills is an elevated region rising to heights of 7000 and 8000 feet above sea-level, and broken by deep gorges and ravines. It is, in fact, a mountainous country in all that that term implies ; but the cañons and gorges are of such a character that the country is easily penetrated by railroad-lines, and good wdgon-roads exist to-day, affording access to every portion of it. Aside from the stampede to the Black Hills country in 1876, when gold-placers were discovered, and 15,000 or 20,000 people became infected with the mining fever, the region has never had a mining " boom." Quartz-mining commenced in 1877 or 1878, and has been successfully prosecuted ever since; the output of gold steadily growing, until at present the yield is about $7,000,000 annually. This great production is about equally divided between two entirely different classes of deposits. The first class, comprising the so-called quartz or free-milling ores (of which the Homestake mine is the principal producer), are veins of generally steep dip, in the older metamorphosed rocks, and similar to the veins worked in other mining districts. The second class of mines is located in what is known as the " siliceous gold-belt." The Siliceous Gold-Belt. This belt, more particularly to be described in this paper, lies in the immediate vicinity of Deadwood and Lead City, and covers an area 6 or 7 miles in length by 3 or 4 miles in width. The formation in which the siliceous ores are found is a nearly horizontal series of sandstones and shales, generally

Jan 1, 1901

By J. de la Vergne, L. O. Cooper

The hydrostatic steel mine shaft liner (HSL) Provides a permanent wall lining for drilled shafts. It is watertight and suitable for bad ground conditions. After a shaft has been drilled to its full depth, the HSL is lowered into it and secured in place with cement grout. The liner consists of a cylindrical steel shell usually reinforced with exterior stiffening rings. The shell is basically designed for compressive strength and the rings are provided to give the shell added resistance to buckling. There are no published standards for the correct procedures to design an HSL. This paper gives a method that appears logical. A design solution can be complex with increased opportunity for error. This paper presents simple formulas to produce sufficient accuracy for construction design. Two new concepts are presented, referred to as transition pressure and critical pressure. An example demonstrates use of the formulas and concepts presented. To facilitate comprehension and use of the formulas, Greek letters are avoided and most of the English alphabet symbols used refer by association to their definitions. An appendix shows derivation of the formulas.

Jan 1, 1984

The research project on size reduction, sponsored by the mining industry through financial support from individual companies, is now embarking on a second three-year program. The first 3-year program had as its objectives assimilation of the literature on comminution and development of simulation programs encompassing the basic mechanisms of grinding. This involved investigating conditions under which a particle in a mill is captured, the stress to which it is subjected, its break probability, and the resulting particle size distributions. The results were used to obtain the most representative simulation models based on current knowledge of mechanisms. As part of this work a set of computer programs has been developed for the following cases: . Batch mill, by numerical integration for completely general selection and breakage functions. . Continuous open-circuit mill by same method as above. . Continuous open-circuit mill by general numerical integration method, including transient behavior

Jan 1, 1969

By Tkeoclore Marvi

IT is singular that an industry quite the antithesis of flying should record tremendous strides in the utilization of aviation through- out the entire depression period, .while in the same years private interests lagged and airplatie manufacturing plants closed to such an extent that he purchase of new planes of certain types was practically impossible. Though receiving little notice from aviation publications and national associations, this use of the airplane progressed by sheer merit. It saved money, conserved energies, and made minutes in the air the equal of weeks on foot or water. The advent of aviation into mining and petroleum operations dates back many years, yet it \vas not until the depression that this new facility of the mineral industry was put on its proper economic basis. There, by its possibilities for saving effort, time, and money, it takes its place equally with other great technical advances in nietl

Jan 1, 1934

By J. S. Johnson, W. C. McCulloch

All phases of plant operation must be synchronized so that the A number of interruptions in the flow of material may be reduced to a minimum. In the majority of plants operating under a labor agreement, time lost during the operating shift cannot be made up with overtime. With the work week restricted to 35 hr., the ratio of productive time to total time justifies extra effort in keeping the plant operating at capacity during the productive shift. This does not apply when central plants with adequate storage are operated by multiple shifts. In such installations the problems become similar to those of any factory or mill where added units, parallel circuits, or excess capacity serve to maintain the average production during intervals of maintenance shut- down. These plants, by special agreement, may run continuously, barring mechanical interruptions. The control of plant operation is in those procedures that are practiced to obtain from the run-of-mine product a high recovery of clean coal products to meet adopted trade standards and consumer specifications at the lowest possible cost. Correlation of all factors for continuous operation requires more than mechanical perfection in the plant. The run-of-mine coal must be delivered to the preparation plant without delays and the clean coal and refuse products taken away. Power supply and water requirements must be adequate. Signal systems must be established for emergency stops and to facilitate resumption of coal flow.

Jan 1, 1943

By G. C. Wallick, W. R. Foster, J. G. Savins

Inasmuch as the differentiation and integration methods represent different modes of rheological analyses, a dual scheme of analysis using both methods should lead to a generalized method of data analysis. A dual differentiation-integmtion method of analysis is applied here to the Poiseuille flow of a variety of ideal Generalized Newtonian and viscoplastic models. Using machine-processing techniques, the result is a spectrum of response patterns which are expressed in terms of certain derivative functions. It is shown that these characteristic functions form the basis of a highly-sensitive analytic technique for optimizing the selection of the most appropriate functional relationship between shear rate and shearing stress. INTRODUCTION In the first paper of this series, Savins, Wallick and Foster1 presented an historical review of the salient features of the differentiation method of rheological analysis in Poiseuille flow, and also indicated how the method could be applied to problems involving plane Poiseuille flow. It was shown that the differentiation and integration methods, although basically not incompatible, do represent different modes of rheological analysis. This suggests that valuable background information regarding the probable response characteristics of real data obtained with the differentiation method can be obtained from an integration method-differentiation method analysis of the response of a variety of ideal rheological models. The present paper describes how this dual method of analysis has been applied to suites of idealized models representing a wide variety of Generalized Newtonian and viscoplastic behavior which have received attention at various places in the literature.

By J. D. Hunt, R. J. McDonald

It is proposed that extensive fluid motion occurs between the dendrites of a Partially solid casting and it is suggested the A segregates in steel castings are a direct result of this motion. The fluid flow arises because of density differences between liquid ahead of the dendrites and the interdendritic liquid. Experiments are described on a model system investigating this fluid motion. The importance of interdendritic fluid flow is discussed with reference to longitudinal segregation. SERIOUS macro segregation is often found in large killed steel ingots. Among the most common segregates are those known as A and V, Fig. 1. The A segregates are arranged over the surface of cones, the wide parts being at the base of the ingot. The A segregates are pencil like in shape and extend from the columnar into the equiaxed region. The V segregates are distributed over inverted cones but have a less well defined shape. They occur at the center of the ingot. Because of the commercial importance of macro inhomogenities these segregates have attracted wide-scale attention. The segregates contain alloying elements and impurities which depress the freezing temperature and are thus the last liquid to freeze. A number of explanations have been proposed to explain their presence. It has been suggested1'' that the segregates are the result of the equiaxed zone growing towards the columnar region together with the interaction of convection within the bulk liquid. It has also been proposed that the segregates are the result of nucleation and rapid growth ahead of the dendritic front.3 Neither of these explanations is completely satisfactory because A segregates are pencil shaped and are not found as continuous cones. In this paper an alternative mechanism is proposed, and experiments are described on a model system to determine whether the mechanism occurs in practice. It is thought that A and V segregates are not peculiar to steels but should occur in many other large castings. Fluid motion within the bulk liquid region of a casting has been extensively discussed. Little attention has been paid to fluid motion within the partially solid regions. The work415 that has been done has been mainly concerned with the fluid motion which results from the volume contraction on freezing. The volume contraction is used to explain various types of macro segregation. The present authors suggest that in many cases the fluid flow resulting from volume contraction will be small compared with that due to other causes. When the columnar region is growing, the liquid still remaining between the dendrites has a much lower temperature near the ingot walls than at the dendrite tips. The composition of the liquid in contact with solid will vary with temperature. The density of the liquid will thus vary both because of changes in temperature and composition. The layers of different density liquid will give rise to a slow but steady liquid flow through the interdendritic regions. If there is no density change with composition, the temperature gradient within the dendritic region should give rise to a downward flow between the dendrites, shown schematically in Fig. 2. When the liquid density changes appreciably with alloy concentration this motion could either be en-

Jan 1, 1970