Search Documents

Jan 1, 1949

By N. S. Greensfelder

The firing of explosive charges by electricity dates back to 1745 when a Doctor Watson is said to have used an electric spark for igniting gunpowder. His method failed in practical application because, to ignite blasting powder, heat must be applied for a period longer than the duration of an electrostatic spark; Wheatstone found that this does not exceed 1/24000 sec. The usual result, therefore, of an electrical discharge in gunpowder is to scatter the grains, without igniting them. Benjamin Franklin partly overcame this difficulty by confining the powder and crushing grains that lay in the line of discharge. In 1822, after the primary cell had been discovered by Volta, Robert Hare, professor of chemistry at the University of Philadelphia, noticed the heating of a short piece of wire placed in the circuit of a voltaic battery. Later work by Oersted, Faraday, Wheatstone, and Henley resulted in the developmcnt of the magneto blasting machines. Fulminates of gold and silver are thought to have been discovcrcd by Cornelius Drebbel, about 1600. Pepys, in his diary for November 11, 1663, recounts a conversation with Doctor Allen, who told him about aurium fulminans "of which a grain ... put in a silver spoon and fired, will give a blow like a musquett and strike a hole through the silvel spoon downward, without the least force upward.'ll This fallacious belief that explosives act downwards only is still held by many who should know better. Fulminate of mercury was discovered by Howard in 1799. J. Shaw of Philadelphia is said to have invented a fulminate-of-mercury cap in 1814; his idea was later adopted by gunmakers in England and France. The principal advantages of mercury fulminate for use in detonators are the simplicity of exploding it and the great shock developed by its explosion.

Jan 1, 1923

By N. S. Greensfelder

The firing of explosive charges by electricity dates back to 1745 when a Doctor Watson is said to have used an electric spark for igniting gunpowder. His method failed in practical application because, to ignite blasting powder, heat must be applied for a period longer than the duration of an electrostatic spark; Wheatstone found that this does not exceed 1/24000 sec. The usual result, therefore, of an electrical discharge in gunpowder is to scatter the grains, without igniting them. Benjamin Franklin partly overcame this difficulty by confining the powder and crushing grains that lay in the line of discharge. In 1822, after the primary cell had been discovered by Volta, Robert Hare, professor of chemistry at the University of Philadelphia, noticed the heating of a short piece of wire placed in the circuit of a voltaic battery. Later work by Oersted, Faraday, Wheatstone, and Henley resulted in the developmcnt of the magneto blasting machines. Fulminates of gold and silver are thought to have been discovcrcd by Cornelius Drebbel, about 1600. Pepys, in his diary for November 11, 1663, recounts a conversation with Doctor Allen, who told him about aurium fulminans "of which a grain ... put in a silver spoon and fired, will give a blow like a musquett and strike a hole through the silvel spoon downward, without the least force upward.'ll This fallacious belief that explosives act downwards only is still held by many who should know better. Fulminate of mercury was discovered by Howard in 1799. J. Shaw of Philadelphia is said to have invented a fulminate-of-mercury cap in 1814; his idea was later adopted by gunmakers in England and France. The principal advantages of mercury fulminate for use in detonators are the simplicity of exploding it and the great shock developed by its explosion.

Jan 1, 1923

By Daniel Houdouin

Data is frequently collected in mineral processing plants in order to evaluate operating performance or to model process units. Metal concentrations, circulating loads, products flow rates, classifiers efficiencies, flotation units recoveries, grinding mill efficiencies are typical parameters calculated from plant data. The validity of these calculations depends strongly upon the reliability of the data used for that purpose. The objective of this paper is to illustrate that flowrates, efficiencies and rate parameters of mineral processes are more reliable when calculated from balanced data rather than from raw data. Using variance-covariance matrices of balanced data produced by automatic mass balance packages (including flow rates, concentrations, percentage of solids and size analyses), variances of separator efficiencies (flotation or classification) and of mill grinding rates are derived. The variances are compared with those of variables calculated from raw data. The advantage of using balance data is illustrated with real data from flotation, grinding and classification units.

Jan 1, 1984

By M. R. Klepper, H. Landsberg

Many of the sedimentary rocks contain small amounts of radioactive constituents. These vary in quantity in different layers. Some recent deposits show rather high activity as; for example, the deep sea oozes and some lake deposits.lv2 Measurements made in boreholes have also shown considerable changes in activity from horizon to horizon. The first to publish data on this fact was Ambronn,3 and recently Russian geophysicists have shown remarkable variations of activity from one layer to the next in various wells. Spak4 felt even that he was able to line up horizons from data measured in two neighboring wells.‡ The measuring of radioactive radiation is fairly difficult. Large errors are possible by the circulation of gaseous emanations in the well and porous strata. It is, therefore, never certain whether the activity of the layer where the instrumental device or collector is located will be measured, or that of some adjacent layer. For this reason it seemed to be advantageous to take samples of rocks into the laboratory and test them there for their contents of radioactive material. Such a method has been followed in the past to distinguish radioactive ore from other rocks. Radioactivity of rocks, especially of the igneous type, was also determined in the laboratory by numerous authors for absolute age investigations.= Corresponding data from studies of sedimentary formations are scarce. Specimens taken from a core of the Bradford oil sandstone were tested by Landsberg and Ingham,7 with the results illustrated in Fig. 1, which shows that within relatively small vertical distances the differences of relative activity were considerable. If one assumes that the distribution of radioactive material is fairly uniform throughout a layer, measurements of radioactive radiation in various horizons would offer additional information for stratigraphic

Jan 1, 1940

By M. R. Klepper, H. Landsberg

Many of the sedimentary rocks contain small amounts of radioactive constituents. These vary in quantity in different layers. Some recent deposits show rather high activity as; for example, the deep sea oozes and some lake deposits.lv2 Measurements made in boreholes have also shown considerable changes in activity from horizon to horizon. The first to publish data on this fact was Ambronn,3 and recently Russian geophysicists have shown remarkable variations of activity from one layer to the next in various wells. Spak4 felt even that he was able to line up horizons from data measured in two neighboring wells.‡ The measuring of radioactive radiation is fairly difficult. Large errors are possible by the circulation of gaseous emanations in the well and porous strata. It is, therefore, never certain whether the activity of the layer where the instrumental device or collector is located will be measured, or that of some adjacent layer. For this reason it seemed to be advantageous to take samples of rocks into the laboratory and test them there for their contents of radioactive material. Such a method has been followed in the past to distinguish radioactive ore from other rocks. Radioactivity of rocks, especially of the igneous type, was also determined in the laboratory by numerous authors for absolute age investigations.= Corresponding data from studies of sedimentary formations are scarce. Specimens taken from a core of the Bradford oil sandstone were tested by Landsberg and Ingham,7 with the results illustrated in Fig. 1, which shows that within relatively small vertical distances the differences of relative activity were considerable. If one assumes that the distribution of radioactive material is fairly uniform throughout a layer, measurements of radioactive radiation in various horizons would offer additional information for stratigraphic

Jan 1, 1940

By A. W. Fahrenwald

The function of the c1assifier in modern fine-grinding practice is to remove a finished product from the grinding-mill discharge, leaving material that needs further comminution. The classifier, therefore, makes two products: (1) "finished" product, which overflows the rim of the classifier tank, sometimes referred to as "slimes" or as "classifier overflow," and (2) "unfinished" product, which settles to the sloping bottom of the classifier tank and is removed by dragging or raking. This product is generally referred to as "sands" or "oversize." One of the outstanding features of the closed-circuit classifier is its ability automatically to return the unfinished product of the ball-mill discharge to the feed end or scoop box of the ball mill. This is a distinct advantage because it avoids dilution of the ball-mill feed, and, furthermore, the particu1ar method of removing the settled sand, by raking it up the inclined bottom of the classifier, gives a classifier discharge containing a minimum of water. This leads to maximum classifier efficiency, in this type of classifier, because the maximum volume of water is displaced upward by the settling sands. The closed-circuit classifier is required to deliver a finished pulp of closely specified density requirements, therefore added hydraulic water generally is not permitted. Relation of Classifier Efficiency to Grinding Efficiency The energy of crushing, and therefore the cost, according to Rittinger's Law,l is proportional to the new surface exposed in crushing. Since the new surface produced in crushing is nearly proportional to the reciprocals of the diameters crushed to, it is obvious that the cost of crushing in the finer sizes is tremendously greater than in the coarser sizes for equal ratios of reduction. High circulating load of finished sand in the circuit of the classifier and grinding mill is therefore obviously expensive practice. It is the classifier's function to reduce this to a miniMum.

Jan 1, 1930

By H. U. Schutt, J. M. Toguri

REGARDLESS of the degree of purification effected during the electrorefining of copper, a danger of re-introducing impurities exists in the operation of melting and casting of the refined copper. Logically only a combination of the purification and casting operation would remedy the situation. A well-known method of metal purification is the "Gross"-type process, whereby a volatile halide is produced, purified, and decomposed.' In the present case, this concept was used to remove the impurities rather than to form possible volatile copper halides. The object was to produce volatile halides of the impurities present and distill these away, leaving high-purity copper. It is interesting to note that such techniques are used in spectrographic procedures in order to shorten exposure time and enhance the density of impurity lines.' In the present investigation, copper hydroxy fluoride was used, since the temperature of decomposition of copper hydroxy fluoride and the melting point of copper nearly coincide. The fluorine released by the thermal decomposition of copper hydroxy fluoride purges the molten copper. Copper hydroxy fluoride was prepared by dissolving high-purity copper in nitric acid and adding hydrofluoric acid and ammonium hydroxide in the required proportions. The melt was made in a high-purity graphite crucible heated by an induction furnace. Trial experiments showed that impurities can be introduced from the crucible material. Therefore, preceding each cast, the crucible itself was cleaned by decomposing copper hydroxy fluoride and discarding the residual copper. Best experimental conditions were obtained by adding 30 g of copper hydroxy fluoride to 500 g of nor-

Jan 1, 1967

By Amitava Gangulee

A recent investigation has shown that solid solution hardening is almost absent in the quasibinary system SnTe-pbTe.1 The lack of solid solution hardening was attributed to the small difference of only about 3 pct between the atomic sizes of lead and tin, and to the fact that the formation of these solid solutions involved changes only in the sublattice occupied by the lead and tin atoms. The investigation reported here is concerned with the quasibinary system PbSe-PbTe which, similar to the SnTe-PbTe system, forms a continuous series of solid solutions2 and has the same NaC1-type (Bl) structure. The size difference between the selenium and tellurium atoms, however, is slightly larger than 7 pct. No indication of any ordering of the lead and tin atoms at. selected compositions had been found in the system SnTe-PbTe, which had been explained by the weak second-nearest-neighbor interactions between lead and tin atoms.' The possible occurrence of an ordering transition in the system PbSe-PbTe was also investigated in the present work. Polycrystalline specinlens of PbSel-1-xTex (x = 0, 0.20, 0.25, 0.33, 0.50, 0.67, 0.75, 0.80, and 1.0) were prepared by melting appropriate amounts of 99.999 pct pure Pb, Se, and Te j.n evacuated and sealed Vycor capsules. The capsules were quenched in water. The solidified ingots were homogenized at 800°C for 14 days. Chemical analysis of the homogenized ingots showed that the deviations from the stoichiometric compositions were less than 0.1 at. pct for each component. Part of each ingot was ground to pass through a 75 g sieve and pressed into 25.4 mm by 12.7 by 2 mm briquettes. Duplicate sets of specimens, each containing at least one briquette and part of the ingot of each composition, were sealed in evacuated Vycor capsules. Both sets were annealed at 500°C for 7 days; one set was quenced in ice-water and the other was cooled to room temperature over a period of 14 days.

Jan 1, 1970

By H. Herbert Hughes

THE trend in modern building construction is definitely toward the use of weight-reducing materials. The basic advantage of lighter structural weight is obvious; reduction of dead load with retention of equivalent strength affords the possibility of increasing the live load, or if this is not desirable or necessary it makes feasible a reduction in size of structural steel members and corresponding savings in other phases of construction. Among the weight-reducing materials and practices rapidly gaining favor are the use of light-weight patented insulating board to replace back-up tile or units, particularly in home construction; introduction of aluminum alloys to replace steel for many purposes; manufacture of porous brick, the use of which materially reduces the weight of masonry walls and partitions; and the use of self-rising vesicular concrete, diatomite and sawdust tile, porous gypsum, or gypsum tile for nonload-bearing partitions and walls. Even building accessories are being designed to save weight and space; one of the most striking examples of this is improved radiator construction. Concrete technique is keeping pace with the rest of the industry. Burned shale aggregates now available in many sections of the country will make concrete weighing only 100 lb. per cubic foot, saving roughly 35 per cent. in weight and sacrificing none of the strength of a rock-sand mix. It may be more expensive, but actual experience has shown that the saving in dead load will effect a reduction in structural steel which in many instances will more than offset the increased cost of the aggregate. The vesicular nature of most light-weight aggregates gives the concrete especially good insulating, fireproofing and soundproofing properties; yet absorption is comparatively low. Strangely enough, however, the increased use of light-weight aggregates appears to have augmented the demand for ordinary aggregates in concrete construction rather than to have decreased it. Light-weight aggregates have widened the field for concrete structures and, since the producers do not attempt to recommend the use of their material except where weight or other special properties are important factors, impetus- has been given to sales of other aggregates.

Jan 1, 1931

By John Chipman

IT has been long known that the gas evolved during the boil in the open-hearth furnace is mainly carbon monoxide associated with smaller quantities of other gases. A number of attempts have been made to determine the precise effects of carbon and oxygen content of the bath, and of other factors, such as temperature, upon the evolution of carbon monoxide. In his investigation of the carbon-iron oxide product in liquid steel, Herty1 showed that the value of this product at equilibrium could be approximated by plotting the value of the product observed in open-hearth reactions against rate of carbon elimination, and extrapolating to zero rate. The results of this study showed a slight but definite effect of temperature upon the carbon-iron oxide reaction. Experimental studies of the evolution of gases from rimming-steel ingots have not yet attained the degree of precision necessary for an exact understanding of the effects of pouring temperature, carbon content, and iron oxide in the metal upon the composition of these gases. Klinger2 showed that the gas consists mainly of carbon monoxide, and that con-siderable variation in its composition occurs during the solidification of the ingot. In order that these variations may be understood in their proper relationship to the progress of solidification of the ingot and the segrega-tion that accompanies solidification, it is essential to have a knowledge of the condition of equilibrium in the several chemical reactions by which the gases are being formed. It is also necessary to understand the effects of temperature upon the equilibria, and further, to know the manner in which the temperature of the liquid metal within the ingot changes as the skin of the ingot is being built up. The reactions in which we will be chiefly interested are those involving in the one case the oxides of carbon, and in the other case steam and hydrogen, with iron oxide and carbon in the liquid metal. The reactions as they occur in unalloyed liquid iron, and their respective equilibrium constants are written as follows:

Jan 1, 1937

By J. T. Jones, H. V. Winchell

In the history of mining in this country there has been nothing exactly similar to the Mesabi iron-range in Minnesota, and the conditions there prevalent. Soft-ore mines are found in many parts of the country. In some the deposits are vertical, and are worked in open pits, or have been so worked, until it was found cheaper to go under ground. In other localities, the ore is scattered through clay beds in the form of hard lumps. These clay beds are sometimes excavated in open pits, and the ore separated by washing. On the Mesabi, the ore itself occurs in strata, and these strata appear quite flat to one standing upon the ground. The ore is soft in the main, and does not require much sorting to free it from foreign or deleterious substances. It is usually covered by the glacial drift-mantle, but occasionally passes beneath a capping of jasperoid rock called " taconite." The ore is twice or three times the thickness of the layer of drift lying upon it. This drift consists of clay, sand, and gravel, with occasional beds of large boulders. There is great difference of opinion among competent miners as to the cheapest methods of mining this ore. It is admitted that local conditions will govern largely ; but there is such a similarity among many of the mines that it seems possible to arrive at some general conclusions on the subject. A typical example of a Mesabi mine is found in the Biwabik. It is one of the largest, and is the most advanced in development. Not a single ton of ore has yet been shipped from this mine, but the expense of the explorations upon which the following description is based has exceeded $60,000. It is seldom that such accurate and complete exploration of a deposit of this magnitude has been possible in the space of time which has elapsed since the discovery of this ore-body. It lies in a new. region, and was remote from railroads until last August, yet, not only the width and length but the depth and quality of the ore-deposit have been learned, and the probable expense of' mining its 20,000,000 tons of ore has been closely figured. After careful consideration, the method of stripping and

Jan 1, 1893

By Orson Shepard

THE leaching method that was first widely used with the cyanide process consisted of percolation leaching of crushed ore in vats or leaching tanks. It was frequently necessary to separate the sand for cyanide treatment and discard the slimes, because only granular material allows the necessary amount of solution to percolate through it in a reasonable time. Later a slime process consisting of agitation followed by thick-ening and filtration was developed, and is extensively used to the pres-ent day. Even though the agitation-leaching slime process is considered satis-factory, the cost of plant and of operation are both greater than for a percolation-leaching plant of equal capacity. This situation has been recognized in the hydrometallurgical treatment of copper-bearing slimes; and a considerable amount of j work has been done at the Southwest Experiment Station of the U. S. Bureau of Mines in an effort to develop a percolation method of treating slimes. Experimental work on the copper-leaching problem was started by H. E. Keyes1, and later tests were made by J. D. Sullivan and A. P. Towne2. The Bureau of Mines work resulted in the development of a process by which slimes were moisture-agglomerated with sand and crushed ore, to form a permeable mass through which solution could be trickled without breaking clown the glomerules. The agglomeration trickle-leaching method was devel-oped on a small scale and finally tried out in a 12-ton vat. Successful results were obtained as long is the leach solution was added slowly, so that the agglomerated charge was never flooded. If the leach solu-tion flooded the charge, the glomerules broke down and, the charge became impervious. There are two important difficulties in the application of the agglom-eration trickle-leaching-method to the cyanide process for gold and silver ores:

Jan 1, 1937

By R. L. Slobod, H. A. Blum

A semiquantitative method for measuring the wettability of reservoir rocks has been developed. These data are needed for reservoir analysis and for interpretation of laboratory displacement studies. The wettability of a core sample is measured by the contact angle for the system oil-water-solid. These contact angles are calculated from the displacement pressure (threshold pressure) obtained by use of the centrifuge, using first oil-water and second air-oil in the same core sample. This method is based on the assumption that the air-oil and oil-water interfaces occupy similar positions in the porous medium when desaturation of the wetting phase is initiated. The assumption is also made that the Contact angle for the air-oil-solid system which is close to zero in value does not change appreciably even when the contact angle for the oil-water-solid system experiences marked changes. "Apparent contact angles'' for five different solids ranging from 31" to 82" have been determined, and changes irr the "apparent contact angle" of a given sample with laboratory use from 33" to 53' {lave been observed. INTRODUCTION Reservoir rocks vary in their wettability, 1,2,3,4 some being preferentially water wet while other are apparently preferentially oil wet. The degree of wettability, in theory, is measured quantitatively by the contact angle for the system solid-oil-water, hut in 1)practice this quantity is extremely difficult to determine." The need for obtaining some measure of this quantity has become obvious in recent years as a result of both laboratory. and field observations. Several reservoir.: such as the Bradford Sand in Pennsylvania and the Wilcox' at Oklahoma City are reported to be oil wet. If these report:: correctly reflect the true nature of the surface of the rock in the ground, then these reservoirs probably will not perform as predicted on the basis of a water wet rock. In the laboratory many core samples have been Observed to be oil wet. These samples, while perhaps not correctly representing the reservoir, have markedly different properties from water wet rock. The location of the phases is different (oil is in contact with rock instead of water), and other quantities such as the connate water, the capillary pressure curve; residual oil. relative permeabilities. and the recovery of oil by water flooding are markedly affected by the value of the contact -4 A further complication observed in the laboratory is the chang-ing of the wettability of a rock specimen with use. such changes mean that repeat runs on the same core are not the duplicate tests desired, but represent new experiments on a more oil wet material. The changes in surface characteristics are sometimes so rapid that by the time an experiment requiring several days is completed, the results may not be representative of the core as originally described, The, rock material itself (Silica. carbonates. etc.), whell Clean and uncontaminated, is water wet. There seems b( : little evidence of the presence of oil wet minerals such as heavy metal sulfides. The oil wetness of reservoir samples, therefore, is believed to be caused by the accumulation of an adsorbed film in which the polar group of a large organic. molecule is adsorbed on the surface leaving the organic or hydrocarbon part of the molecule projecting out from the solid. Such a surface is much more readily wet 17" oil than by water. These considerations indicate the need for a method which will provide some measure of tile wettalbility (Contact(angle if possible) for reservoir rock ..samples) The method should be capable of (1.1 distinguishing between tile wettalility of dif. ferent formation,.. (2) detecting (change. ill wettalbility of a core with use. and (3) measuring the changes in Wettability which may be accomplished with cleaning operations. such as the use of sodium silicate for increasing water wrettalbility." A preliminary report on such a method is preseted below. not with the idea that this procedure alone will Solve the wettabil-ity problem. but rather that it may provide encouragement to

Jan 1, 1952

By James E. Dunn

For several years, ground support at the Pitch uranium mine has been a serious headache to the management. The "moving" ground raised havoc with standard timber sets as well as steel sets. The author discusses this problem which has led to the adoption of heavy cribbing practices in drifts and raises.

Jan 6, 1961

By E. P. Bucklen

Simulation is a tool useful in the evaluation of the probable effect which various changes may have on the performance of time-dependent systems. This is especially true as far as rail transportation in mining is concerned. The VPI-OCR RAILSIM program is a computer program designed specifically for solving a broad range of problems encountered in the design of haulage systems of new mines as well as in the operation of haulage systems in existing mines. This program is one of several computer applications for evaluating mining systems developed at the Virginia Polytechnic Institute and State University under contract with the Office o f Coal Research of the U.S. Dept. of the Interior. The RAILSIM program uses Monte Carlo techniques to introduce variability in the occurrence of events such as would be observed in real systems. Data are required to describe the haulage network, locomotives, stationary facilities, and other miscellaneous parameters. Output from the program lists measures of performance that are useful in assessing the capabilities and limitations of given system configurations. The output information may cover many shifts of operation to indicate the broad range of variation and interaction which can occur as the system responds to varying levels of demand for service. The RAILSIM program was tested extensively in mining operations. The results of some of these tests are used to illustrate the usefulness of this program.

Jan 1, 1972

On Jan. 1, the office of the. National Service Committee was established at 502 McLachlen Building, Washington, D. C. The first work was to review current legislative information; which is being continued and good results are accomplished. Whenever matters of importance appear in the Congressional Record, notice thereof with such documents as are germane are forwarded to the members of the committee for their information and advice. The first formal report of the chairman was on the so-called Smith-Howard bill to secure Federal endowment for the establishment of engineering research. The chairman of the Committee, at the request of the Executive Committee of the Engineering Council, appeared before the Senate Committee on Education and Labor for the purpose of opposing those features of the so-called Kenyon bill which would create a practically permanent commission dominated by the Engineering Corps of the Army over all Federal engineering projects as well as state and municipal projects receiving Federal aid. The greater part of the time has been expended in correspondence and consultation on the matter of the establishment of a Department of Public Works. The results of this work lead to the belief that there is much promise of the establishing of such a department by the 66th Congress Hall members of engineering societies, as well as the societies themselves, will unanimously agree on and support a measure, and in order to secure such unanimous support, it will be advisable to call a conference of engineering societies. .

Jan 4, 1919

By C. P. Williams

THE lead-bearing area of Missouri has been subdivided, geographically, into the Southeastern, the Central, and the Southwestern

Jan 1, 1877

By Albert J Hoskinson

Progress on the earth depends to a large extent upon the rapid interchange of ideas and commodities between the various nations of the world. The smooth flow of commerce, by which these ideas and commodities move, is accelerated by accurate and complete maps of the various land and water areas of the earth. The production of accurate maps depends upon a complete knowledge of the figure and dimensions of the object to be charted, therefore a knowledge of the shape of the earth is a prime necessity for accurate charting purposes. Gravity observations were first undertaken to aid in determining the shape of the earth so that the geodetic surveys of its land and water areas might be more accurately connected to form an undistorted map. A large part of the earth's surface is water and it is therefore necessary to get gravity values at sea as well as on land before a complete and accurate figure of the earth can be determined. Dr. F. A. Vening Meinesz, of Holland, was the first scientist to perfect a pendulum instrument that would accurately measure gravity at sea. He had previously developed a multiple pendulum apparatus for measuring gravity upon swampy and unstable land areas in Holland and it was therefore quite natural that he should adapt the instrument to give accurate results upon a still more unstable support, the surface of the sea. His instrument as perfected gives satisfactory results when mounted in gimbals on any vessel where the roll and pitch are not more than about 5" and where the vibrations from rotating machinery and wave action are not large enough to cause the pendulums to shift on the knife-edges. This degree of stability can be obtained on surface craft only in protected waters or on very calm days. Therefore, Dr. Meinesz mounted his instrument in a submarine so that he could submerge below the disturbed surface waters and get the desired stability in nearly any kind of weather by simply diving to greater depths. Most of the gravity-at-sea work has been done with a Meinesz instrument and a very large number of the observations have been made by Dr. Meinesz himself. He has occupied nearly 1000 sea, stations and probably has traveled as many miles by

Jan 1, 1940