Search Documents

By C. D. King

The advances in basic open-hearth practice which have occurred during the past decade have been principally in the direction of the physical development of larger units, decreased fuel consumption and increased length of furnace campaigns. However gratifying the economies which have resulted from these improvements are, it must be granted that like advances in chemical, physical-chemical and metallurgical knowledge pertaining to open-hearth practice would have added greatly to these already attained reductions in costs and it is the purpose of this paper to stimulate the discussion and study of these phases of cooperation and their relation to ingot yields. Definition and Calculation oF Losses In discussing basic open-hearth yields, the writer refers to the ingot yields obtained in the process of converting pig iron, scrap and ore to steel. In this conversion some scrap is produced. The difference between the total weight of ingots and scrap produced and the original metallic charge constitutes the loss. Basic open-hearth yields vary considerably, and while some of the divergencies can be explained metal-lurgically because of differences in slag analysis, proportion of pig charges, and character of steel made, there often remain differences that apparently do not readily disclose themselves. In reconciling yields by the usual accepted methods of calculation, one is often forced to evaluate large losses as ((unaccounted for." Such a system is prone to accept these large losses as inherent to the operat.ion and not, susceptible of appraisement and correction. It is obvious that if the ingot yield plus the scrap recovered subtracted from 100 per cent. determines the loss, a drop in yield may increase the loss or the scrap, and both contribute to a monetary loss. To be sure, one can subdivide these unaccounted for losses into various components, such as metallics in slag, free Fe in slag, spillage, etc., and thereby reduce the amount of losses considered as "unaccounted for." If empirical methods could indicate with some degree of accuracy all losses, however slight, the undetermined losses become appreciably smaller and resultant calculations will check actual

Jan 1, 1929

By Henry S. Rawdon

Any laboratory corrosion test, as judged from the practical point of view, is valuable only to the extent that it foretells what will, in all probability, occur in service. Such a test is most properly to be considered as a "pilot test," that is, a test which indicates the direction along which action may be expected to occur rather than as a truly quantitative test which would be expected to tell just how and to what extent the action would proceed. It is generally recognized that laboratory corrosion tests should be chosen with particular reference to the character of service expected for any particular type of metal, so far as it can be foreseen. Even when this requirement has been fulfilled, however, the question whether the actual service behavior of the material is in accordance and general agreement with predictions based upon such laboratory tests is always a pertinent one. The information available in the technical literature bearing on the correlation between the service behavior of metals from the corrosion standpoint, and the results obtained with the same metals in laboratory corrosion tests is very meager, indeed. The results which have been obtained in weather-exposure tests of sheet duralumin which form the body of this discussion are presented here for their bearing on this particular phase of the problem of the corrosion resistance of duralumin. Although the exposure tests described have not been completed, in the sense that all of the tests, in the series initially laid down, have been accomplished, still the general trend shown by the results is so clear that a number of definite conclusions at this stage (subject, of course, to possible modification in the light of later results) are believed to be warranted. The tests reported upon here form one phase of the much broader study of this subject which has been under way for several years in which the National Advisory Committee for Aeronautics, the Bureau of Aeronautics of the Navy Department, the Army Air Corps and the Bureau of Standards have cooperated. Without the very helpful cooperation of the manufacturers in furnishing the

Jan 1, 1929

By Henry S. Rawdon

Any laboratory corrosion test, as judged from the practical point of view, is valuable only to the extent that it foretells what will, in all probability, occur in service. Such a test is most properly to be considered as a "pilot test," that is, a test which indicates the direction along which action may be expected to occur rather than as a truly quantitative test which would be expected to tell just how and to what extent the action would proceed. It is generally recognized that laboratory corrosion tests should be chosen with particular reference to the character of service expected for any particular type of metal, so far as it can be foreseen. Even when this requirement has been fulfilled, however, the question whether the actual service behavior of the material is in accordance and general agreement with predictions based upon such laboratory tests is always a pertinent one. The information available in the technical literature bearing on the correlation between the service behavior of metals from the corrosion standpoint, and the results obtained with the same metals in laboratory corrosion tests is very meager, indeed. The results which have been obtained in weather-exposure tests of sheet duralumin which form the body of this discussion are presented here for their bearing on this particular phase of the problem of the corrosion resistance of duralumin. Although the exposure tests described have not been completed, in the sense that all of the tests, in the series initially laid down, have been accomplished, still the general trend shown by the results is so clear that a number of definite conclusions at this stage (subject, of course, to possible modification in the light of later results) are believed to be warranted. The tests reported upon here form one phase of the much broader study of this subject which has been under way for several years in which the National Advisory Committee for Aeronautics, the Bureau of Aeronautics of the Navy Department, the Army Air Corps and the Bureau of Standards have cooperated. Without the very helpful cooperation of the manufacturers in furnishing the

Jan 1, 1929

By D. J. McAdam

Results of investigation of corrosion-fatigue of metals at the U. S. Naval Engineering Experiment Station have been presented by the author in four papers.1, 2, 3, 4 In those papers references were given to the work of Haigh in 1917.5 The investigation at Annapolis included a variety of ferrous and non-ferrous alloys. The experiments showed that even slight corrosion simultaneous with fatigue may cause failure at stresses far below the ordinary endurance limit. The experiments also showed that, for most metals, severe stressless corrosion prior to fatigue is much less damaging than even slight corrosion simultaneous with fatigue. The term "corrosion-fatigue" was restricted to mean the simultaneous action of corrosion and fatigue. The combined effect of corrosion followed by fatigue was called "prior-corrosion fatigue." A stress-cycle graph representing corrosion-fatigue is a curve approaching a horizontal asymptote, the ordinate of which may be called a "corrosion-fatigue limit." To avoid confusion in later discussion, the term "endurance limit" will be restricted to mean the fatigue limit obtained by tests in air with specimens as free as possible from stress concentration and from corrosion. The cndursnce limit as thus obtained presumably represents an inherent property of the metal.

Jan 1, 1929

By D. J. McAdam

Results of investigation of corrosion-fatigue of metals at the U. S. Naval Engineering Experiment Station have been presented by the author in four papers.1, 2, 3, 4 In those papers references were given to the work of Haigh in 1917.5 The investigation at Annapolis included a variety of ferrous and non-ferrous alloys. The experiments showed that even slight corrosion simultaneous with fatigue may cause failure at stresses far below the ordinary endurance limit. The experiments also showed that, for most metals, severe stressless corrosion prior to fatigue is much less damaging than even slight corrosion simultaneous with fatigue. The term "corrosion-fatigue" was restricted to mean the simultaneous action of corrosion and fatigue. The combined effect of corrosion followed by fatigue was called "prior-corrosion fatigue." A stress-cycle graph representing corrosion-fatigue is a curve approaching a horizontal asymptote, the ordinate of which may be called a "corrosion-fatigue limit." To avoid confusion in later discussion, the term "endurance limit" will be restricted to mean the fatigue limit obtained by tests in air with specimens as free as possible from stress concentration and from corrosion. The cndursnce limit as thus obtained presumably represents an inherent property of the metal.

Jan 1, 1929

By C. E. Reinhard

A useful accelerated weathering test should be capable of placing any series of metals quantitatively in the same order of endurance as that noted under a particular set of actual exposure conditions. The development of such a test must necessarily depend to a considerable extent upon data obtained from a wide variety of exposures. This paper offers preliminary data from exposure tests on nine metals placed in 11 locations in the Lehigh Valley, within a radius of 20 miles from Palmerton. The exposure conditions vary from a severe industrial atmosphere to a pure rural atmosphere. The data are offered as evidence of the necessity for the identification and evaluation of the variables of corrosion and as illustration of some of the variations to be considered in the development of a useful accelerated test. It is quite clear, from this evidence, that a useful accelerated test must be versatile to the extent of showing the probable variations of corrosion resulting from differences in atmosphere as well as from differences in the metals. The paper is presented solely with the object of pointing out the wide variations in relative merits of several common metals in resisting corrosion and without any thought of initiating controversy regarding the merits of particular metals. For this reason the metals are designated only by symbols in the tables and no key is given to these symbols. Test Procedure Specimens, 4 by 6 in., were cut to accurate dimensions from each of the following metals or alloys: aluminum, ascoloy, copper, copper steel, galvanized iron, and zinc. Additional grades of these metals bring the total to nine types. Since it was impossible to obtain the required number of copper-steel specimens having equal surface conditions of rust and scale, it was thought advisable to pickle these specimens in a hot solution of 10 per cent. sulfuric acid. All other specimens were carefully cleaned in ether and alcohol and weighed just prior to exposure. Four specimens of each type of metal were mounted on a test rack by means of porcelain insulators. The racks were so constructed that

Jan 1, 1929

By C. E. Reinhard

A useful accelerated weathering test should be capable of placing any series of metals quantitatively in the same order of endurance as that noted under a particular set of actual exposure conditions. The development of such a test must necessarily depend to a considerable extent upon data obtained from a wide variety of exposures. This paper offers preliminary data from exposure tests on nine metals placed in 11 locations in the Lehigh Valley, within a radius of 20 miles from Palmerton. The exposure conditions vary from a severe industrial atmosphere to a pure rural atmosphere. The data are offered as evidence of the necessity for the identification and evaluation of the variables of corrosion and as illustration of some of the variations to be considered in the development of a useful accelerated test. It is quite clear, from this evidence, that a useful accelerated test must be versatile to the extent of showing the probable variations of corrosion resulting from differences in atmosphere as well as from differences in the metals. The paper is presented solely with the object of pointing out the wide variations in relative merits of several common metals in resisting corrosion and without any thought of initiating controversy regarding the merits of particular metals. For this reason the metals are designated only by symbols in the tables and no key is given to these symbols. Test Procedure Specimens, 4 by 6 in., were cut to accurate dimensions from each of the following metals or alloys: aluminum, ascoloy, copper, copper steel, galvanized iron, and zinc. Additional grades of these metals bring the total to nine types. Since it was impossible to obtain the required number of copper-steel specimens having equal surface conditions of rust and scale, it was thought advisable to pickle these specimens in a hot solution of 10 per cent. sulfuric acid. All other specimens were carefully cleaned in ether and alcohol and weighed just prior to exposure. Four specimens of each type of metal were mounted on a test rack by means of porcelain insulators. The racks were so constructed that

Jan 1, 1929

By C. L. Mantell

Although so common and well known a metal, tin is really a less abundant element than many of those less familiar and usually ranked with the scarce or rare elements, such as cerium, yttrium, lithium, beryllium, titanium, zirconium and vanadium. The color of tin is white, with a slightly bluish tinge. When it is compared with nickel, the latter metal is brown. When placed alongside of tin, chromium is definitely blue. Tin is whiter than silver or zinc. Exceedingly thin films of tin show various shades of brown by transmitted light. The metal has a brilliant luster, and when mirror-polished, has high light reflectivity. The luster depends to a large extent upon the pouring temperature at which the metal is cast. If the temperature be too high, iridescent colors may show on the surface. These are the result of oxide film formation. If the pouring temperature be too low, the surface is dull. Small amounts of foreign metals, such as lead, arsenic, antimony and iron, decrease the luster of tin and impart a yellowish tinge to the metal. The structure is decidedly crystalline. The cast metal is a mass of crystals. When a bar of tin is bent, it emits a characteristic creaking sound. This is called the "cry" of tin, resulting from the grinding of the crystals one against the other during the bending of the metal. Tin metal may lose its crystalline structure by cold working, but grain growth occurs rapidly under the influence of heat. Metal which has been cast at a low temperature may become amorphous as the result of mechanical work. If such material be heated to temperatures well below the melting point, say at 110° C. for a half hour or for a few minutes at 150" C. recrystallization occurs. If the surface of tinned plate, tin foil, or tin meta1 in cast form be etched with hydrochloric acid containing a little free chlorine, attack occurs at the intererystalline faces. The surface shows patterns resembling the frost flowers on window panes in winter. The etched appearance is called moire metallique. Fine dendritic surface crystals of tin may be obtained by casting the metal on a surface of polished steel. In commercial tin plate, the intercrystalline boundaries are shown by fine grooves or channels, .which result from the method

Jan 1, 1929

By C. L. Mantell

Although so common and well known a metal, tin is really a less abundant element than many of those less familiar and usually ranked with the scarce or rare elements, such as cerium, yttrium, lithium, beryllium, titanium, zirconium and vanadium. The color of tin is white, with a slightly bluish tinge. When it is compared with nickel, the latter metal is brown. When placed alongside of tin, chromium is definitely blue. Tin is whiter than silver or zinc. Exceedingly thin films of tin show various shades of brown by transmitted light. The metal has a brilliant luster, and when mirror-polished, has high light reflectivity. The luster depends to a large extent upon the pouring temperature at which the metal is cast. If the temperature be too high, iridescent colors may show on the surface. These are the result of oxide film formation. If the pouring temperature be too low, the surface is dull. Small amounts of foreign metals, such as lead, arsenic, antimony and iron, decrease the luster of tin and impart a yellowish tinge to the metal. The structure is decidedly crystalline. The cast metal is a mass of crystals. When a bar of tin is bent, it emits a characteristic creaking sound. This is called the "cry" of tin, resulting from the grinding of the crystals one against the other during the bending of the metal. Tin metal may lose its crystalline structure by cold working, but grain growth occurs rapidly under the influence of heat. Metal which has been cast at a low temperature may become amorphous as the result of mechanical work. If such material be heated to temperatures well below the melting point, say at 110° C. for a half hour or for a few minutes at 150" C. recrystallization occurs. If the surface of tinned plate, tin foil, or tin meta1 in cast form be etched with hydrochloric acid containing a little free chlorine, attack occurs at the intererystalline faces. The surface shows patterns resembling the frost flowers on window panes in winter. The etched appearance is called moire metallique. Fine dendritic surface crystals of tin may be obtained by casting the metal on a surface of polished steel. In commercial tin plate, the intercrystalline boundaries are shown by fine grooves or channels, .which result from the method

Jan 1, 1929

By J. L. Gregg

A study of copper-base and silver-base alloys was made with the object of finding, if possible, useful alloys subject to dispersion hardening. These studies led to the discovery of several alloys showing considerable dispersion hardening, the most important of which, the alloys of copper with iron, nickel, cobalt and chromium silicides, have since been very ably reported by M. G. Corson.l This paper is offered as a few notes covering alloys other than those described by Corson and some further data on the silicide alloys. Copper Alloys For the purpose of making a rapid survey of copper alloys that might show useful dispersion-hardening phenomena, various alloys were prepared containing 3 per cent. of additional elements added in such proportion that possible compounds could be formed without an excess of either added element. These alloys in the form of rolled or forged pieces were quenched in water from a temperature of 871" C. (1800" F.) and subsequently aged at various lower temperatures for l-hr. periods. The comparison of the Rockwell hardness of the quenched alloys with the maximum hardness of the alloys after aging was used as a measure of the dispersion hardening. Table 1 shows the Rockwell "E" hardness 1/8-in. ball, 100-kg. load) of the quenched alloys, together with the maximum hardness resulting from aging. The reheating temperature given in this table is the aging temperature at which maximum hardness resulted from a l-hr. sojourn. From this table it may be observed that only the silicide alloys showed significant dispersion hardening. Of these, the only one not described by Corson is the Cu-Be-Si alloy. This alloy contained 1.35 per cent. beryllium and 1.78 per cent. silicon, which is approximately equivalent to 3 per cent. Be2Si. It had a yellow color and scaled much less than the other silicide alloys upon heating. It could be worked easily either hot or cold but hardened very rapidly on cold working. The annealed alloy upon drawing from a diameter

Jan 1, 1929

By J. L. Gregg

A study of copper-base and silver-base alloys was made with the object of finding, if possible, useful alloys subject to dispersion hardening. These studies led to the discovery of several alloys showing considerable dispersion hardening, the most important of which, the alloys of copper with iron, nickel, cobalt and chromium silicides, have since been very ably reported by M. G. Corson.l This paper is offered as a few notes covering alloys other than those described by Corson and some further data on the silicide alloys. Copper Alloys For the purpose of making a rapid survey of copper alloys that might show useful dispersion-hardening phenomena, various alloys were prepared containing 3 per cent. of additional elements added in such proportion that possible compounds could be formed without an excess of either added element. These alloys in the form of rolled or forged pieces were quenched in water from a temperature of 871" C. (1800" F.) and subsequently aged at various lower temperatures for l-hr. periods. The comparison of the Rockwell hardness of the quenched alloys with the maximum hardness of the alloys after aging was used as a measure of the dispersion hardening. Table 1 shows the Rockwell "E" hardness 1/8-in. ball, 100-kg. load) of the quenched alloys, together with the maximum hardness resulting from aging. The reheating temperature given in this table is the aging temperature at which maximum hardness resulted from a l-hr. sojourn. From this table it may be observed that only the silicide alloys showed significant dispersion hardening. Of these, the only one not described by Corson is the Cu-Be-Si alloy. This alloy contained 1.35 per cent. beryllium and 1.78 per cent. silicon, which is approximately equivalent to 3 per cent. Be2Si. It had a yellow color and scaled much less than the other silicide alloys upon heating. It could be worked easily either hot or cold but hardened very rapidly on cold working. The annealed alloy upon drawing from a diameter

Jan 1, 1929

By K. S. Seljesater

Since the development of dispersion-hardenable lead-antimony alloys' in the laboratories of the Western Electric Co., Inc., studies have been made of the effect of various third constituents on these alloys.2 This paper covers a study of the effect of small amounts of arsenic. This element is important since it occurs as an impurity in many commercial antimonial leads and the results obtained are particularly interesting as indicating a class of addition agents which exert a stabilizing effect on dispersion-hardened alloys. Experimental In order to study the effect of increasing amounts of arsenic, a series of alloys were made up containing 1 per cent. antimony, and arsenic varying from 0. to 0.10 per cent., the balance Doe Run lead (99.997 per cent. lead). The alloys were cast into bars 1 1/2 in. wide by 3/4 in. thick and cold-rolled to about 1/8 in- thickness. Three samples of each composition were annealed 1 hr. at 225" C. and quenched in water. One set was aged at room temperature, one at 50" C., and the third at 100" C., the

Jan 1, 1929

By K. S. Seljesater

Since the development of dispersion-hardenable lead-antimony alloys' in the laboratories of the Western Electric Co., Inc., studies have been made of the effect of various third constituents on these alloys.2 This paper covers a study of the effect of small amounts of arsenic. This element is important since it occurs as an impurity in many commercial antimonial leads and the results obtained are particularly interesting as indicating a class of addition agents which exert a stabilizing effect on dispersion-hardened alloys. Experimental In order to study the effect of increasing amounts of arsenic, a series of alloys were made up containing 1 per cent. antimony, and arsenic varying from 0. to 0.10 per cent., the balance Doe Run lead (99.997 per cent. lead). The alloys were cast into bars 1 1/2 in. wide by 3/4 in. thick and cold-rolled to about 1/8 in- thickness. Three samples of each composition were annealed 1 hr. at 225" C. and quenched in water. One set was aged at room temperature, one at 50" C., and the third at 100" C., the

Jan 1, 1929

By C. E. Wood, T. L. Joseph

This paper is a brief report of experimental work undertaken determine whether barium oxide in any quantity increases the desulfur ing action of blast-furnace slags. Industrial furnace operation wi slag containing small quantities of barium oxide, referred to later, intc ested S. A. Sanford of the Commercial Chemical Co., Youngstown, Oh to conduct experiments to furnish comparative results on desulfurizati with normal slag and slag containing small quantities of barium oxic The U. S. Bureau of Mines agreed to assist on the problem because the possible saving in fuel if the desulfurizing action of blast-furna slag can be substantially increased, also because the experiments woi yield new information on sulfur elimination in general, a matter economic importance in furnace practice. A more complete and func mental study of desulfurization of iron by slag is in progress. Distribution of Sulfur iN the Blast Furnace More than 90 per cent. of the total sulfur entering the blast furnz is in the coke. Sponge iron formed by reduction of iron oxides absor part of the sulfur as ferrous sulfide. Kinney's examination of sample taken from various levels of a 3004,011 furnace, shows that the me continues to take up sulfur until it reaches a level near the tuyer Desulfurization must therefore take place as the metal falls through t slag and as the slag and metal exist as layers in the crucible. The relative amounts of sulfur removed while the metal falls throu the slag and while the slag and metal exist as layers in t crucible are not definitely known. Preliminary experiments by t

Jan 1, 1929

By E. H. Dix, F. Keller

The use of magnesium as an alloying element in aluminum alloys has been limited, in general, to comparatively small quantities. In duralumin-type, strong aluminum alloys, magnesium is present to the extent of 35 to 3/4 per cent. and various grades of Magnalite and the "Y" alloy of the National Physical Laboratory of England contain up to 1 1/2 per cent. In this country small quantities of aluminum alloy sheet containing 4 per cent. magnesium are produced commercially. All commercial aluminum contains sufficient silicon to convert a portion of the magnesium added to the compound Mg2Si. It is well known that the solubility of both magnesium and magnesium-silicide decrease with decrease in temperature. It is, therefore, impossible to separate the age-hardening effect of the magnesium-silicide from that of magnesium, except by the use of aluminum of such high purity that the silicon content is negligible. The present paper is the fifth l, 2, 3, 4 of a series from the laboratories of the Aluminum Co. of America reporting the results of investigations of the equilibrium relations in aluminum-rich alloys made from the high-purity aluminum manufactured by this company. 5 Previous Investigations A number of investigations have been made of the constitution of the complete aluminum-magnesium system and several determinations made

Jan 1, 1929

By F. Keller, E. H. Dix

The use of magnesium as an alloying element in aluminum alloys has been limited, in general, to comparatively small quantities. In duralumin-type, strong aluminum alloys, magnesium is present to the extent of 35 to 3/4 per cent. and various grades of Magnalite and the "Y" alloy of the National Physical Laboratory of England contain up to 1 1/2 per cent. In this country small quantities of aluminum alloy sheet containing 4 per cent. magnesium are produced commercially. All commercial aluminum contains sufficient silicon to convert a portion of the magnesium added to the compound Mg2Si. It is well known that the solubility of both magnesium and magnesium-silicide decrease with decrease in temperature. It is, therefore, impossible to separate the age-hardening effect of the magnesium-silicide from that of magnesium, except by the use of aluminum of such high purity that the silicon content is negligible. The present paper is the fifth l, 2, 3, 4 of a series from the laboratories of the Aluminum Co. of America reporting the results of investigations of the equilibrium relations in aluminum-rich alloys made from the high-purity aluminum manufactured by this company. 5 Previous Investigations A number of investigations have been made of the constitution of the complete aluminum-magnesium system and several determinations made

Jan 1, 1929

By C. C. Furnas, S. P. Kinney

The efficient operation of a blast furnace depends primarily upon efficient contact between the descending streail1 of solid materials and the ascending stream of gas. A program of research dealing with the interior of the present-day furnace is being carried by the U. S. Bureau of Mines along two lines: (I) observations on commercial furnaces and (2) experiments in the laboratory. Through the courtesy of the officials of the Illinois Steel Co. it was possible to make a detailed investigation of the interior of the shaft of NO. 4 furnace at the company's South Works, South Chicago, Ill. The experimental work is reported in a IT. S. Bureau of Mines bulletin' which will include the results of several thousand observations on the interior of the shaft of this furnace. Such of the data as are presented later in this paper constitute only a brief summary. The second part of the work has been conducted in the laboratory. Only one phase of the laboratory investigation has been completed at the present time, namely, the flow of gases through beds of broken solids." The object of this papcr is to correlate the results of the two different lines of research and to present a. more or less complete picture of the condition of the interior of a blast furnace by giving approximate, quantitative data of the actual size-distribution of solid particles in a typical furnncc, and of the efficiency of contact between the gas and solid phases. With such data before him the operator or the experimenter not only has a much better idea of the actunl conditions inside a furnace, hut he also knows in what direction to look for possible improvement. The discussion which follows presents only a few of the items studied and reported; those persons who are particularly interested in all the result's of this investigation must refer to the aforementioned publication.

Jan 1, 1929

By N. A. Zeigler

The present publication is a continuation of the work on gas analysis described in a paper presented before the Institute of Metals Division year ago.' While that paper was largely descriptive in character dealing with the various parts of the apparatus and method of operations the present one gives numerical data obtained as a result of a series c experiments with iron-carbon alloys. The apparatus and the method applied are essentially the same as previously described, changes bein only in minor details. Attention should be called to the fact that in the present work we ar dealing with the gaseous mixtures obtained as a result of vacuum meltin of various ferrous alloys. It should be understood that when the term "degasified sample" and "products of degasification" are used, the refer to the sample remaining after the vacuum fusion of the alloy an the gaseous products removed therefrom, respectively. The write does not wish to create the impression that the metallic products of thj process are absolutely free from any gaseous elements, nor that the gase obtained are present as such in the original sample. These two point are not yet definitely known and are subjects for future investigatior The point, however, which has been established beyond doubt, i that all the gases dealt with come from the metallic charge and ar present either in the charge as such, or are produced as a result of chemic: reactions going on within the charge. The suggestion that the gases an due to the reactions between the metallic charge and the refractor materials has no valid basis. General Procedure A series of iron-carbon alloys was prepared in a magnesia-line graphite crucible placed in a high frequency laboratory furnace open t the atmosphere. The materials used were Westinghouse electrolyti iron and Acheson graphite. Each alloy, after it was completely molter was kept in the liquid state for 10 min., cast into a small iron mold, an

Jan 1, 1929

By O. W. Ellis

The writer has dealt with the effect of various methods of melting copper upon the gas content of the metal.' The copper referred to in his earlier paper was melted in the foundry both in the oil furnace and the electric furnace. As the metal did not measure up to conductivity standard, the term "casting copper" was applied to the material in the title of the paper. It was thought that some information in regard to the gases in a sample of overpoled copper might also be of interest, and through the courtesy of C. R. Stevens, superintendent of the copper mill of the West-inghouse Elec. & Mfg. Co., such a sample was obtained. The metal was chill-cast in the form of an ingot about 7 in. long, 2 1/2 in. deep (from the base to the raised surface of the ingot) and from 1 3/8 to 1 7/8 in. wide. A photograph of part of the ingot is shown in Fig. 1. The effect of over-poling on the surface of the metal is here quite clearly shown.

Jan 1, 1929

By O. W. Ellis

The writer has dealt with the effect of various methods of melting copper upon the gas content of the metal.' The copper referred to in his earlier paper was melted in the foundry both in the oil furnace and the electric furnace. As the metal did not measure up to conductivity standard, the term "casting copper" was applied to the material in the title of the paper. It was thought that some information in regard to the gases in a sample of overpoled copper might also be of interest, and through the courtesy of C. R. Stevens, superintendent of the copper mill of the West-inghouse Elec. & Mfg. Co., such a sample was obtained. The metal was chill-cast in the form of an ingot about 7 in. long, 2 1/2 in. deep (from the base to the raised surface of the ingot) and from 1 3/8 to 1 7/8 in. wide. A photograph of part of the ingot is shown in Fig. 1. The effect of over-poling on the surface of the metal is here quite clearly shown.

Jan 1, 1929