Search Documents

By Jessica M. Wempen, Michael K. McCarter

"Differential Interferometric Synthetic Aperture Radar (DlnSAR), a satellite-based remote sensing technique, has potential application for measuring mine subsidence on a regional scale with high spatial and temporal resolutions. However, the characteristics of Synthetic Aperture Radar (SAR) data and the effectiveness of DlnSAR for subsidence monitoring depend on the radar band (wavelength). This study evaluates the effectiveness of DinSAR for monitoring subsidence due to longwall mining in central Utah using L-band (24 cm wavelength) SAR data from the Advanced Land Observing Satellite (ALOS) and X-band (3 cm wavelength) SAR data from the TerraSAR-X mission.In the Wasatch Plateau region of central Utah, which is characterized by steep terrain and variable ground cover conditions, areas affected by longwall mine subsidence are identifiable using both L-band and X-band DlnSAR. Generally, using L-band data, subsidence magnitudes are measurable. Compared to X-band, L-band data are less affected by signal saturation due to large deformation gradients and by temporal decorrelation due to changes in the surface conditions over time. The L-band data tend to be stable over relatively long periods (months). Short wavelength X-band data are strongly affected by signal saturation and temporal decorrelation, but regions of subsidence are typically identifiable over short periods (days). Additionally, though subsidence magnitudes are difficult to precisely measure in the central Utah region using X-band data, they can often be reasonably estimated.INTRODUCTIONDifferential Interferometric Synthetic Aperture Radar (DlnSAR) is a satellite-based remote technique that can be used to measure surface displacement over large regions with high spatial resolution. Under good conditions, displacements can be measured with centimeter to subcentimeter accuracy (Buckley, 2000; Massonnet and Feig!, 1998). DlnSAR also has high temporal resolution, and imaging periods typically range from 10 to 50 days (Rosenqvist, Shimada, and Watanabe, 2004; Roth, 2004). In the last two decades, the application of DlnSAR for mine subsidence monitoring has been demonstrated in coal basins in Europe, Australia, China, and the United States (Camec and Delacourt, 2000; Ge, Chang, and Rizos, 2007; Ge et al., 2008; Ismaya and Donovan, 2012; Ng et al., 2010; Perski, 2000; Perski and Jura, 2003; Popiolek and Krawczyk, 2006; Wegmuller et al., 2005; Wright and Stow, 1999; Zhao et al., 2014; Xinglin et al., 2014). Overall, these studies have demonstrated good data resolution, strong relationships between mine development and subsidence, and reasonable agreement between displacements measured by DinSAR and displacements measured by conventional surveys."

Jan 1, 2016

By R. R. Cellan

There have not been many opportunities to date for comparison between laboratory and commercial neutralization rinse data. Homestake Mining Company's Santa Fe project is one such comparison. This paper discusses laboratory procedures and results from a 100' leach pad height simulation column neutralization rinse test and compares principal data with that achieved in actual commercial leach pad neutralization rinsing procedures. Principal comparisons include the rate of WAD cyanide neutralization, decrease in effluent pH, stabilization of dissolved silver and copper, and the quantity of dissolved gold recovered during rinsing.

Jan 1, 1996

By M. Fiset, D. Belley

"In a country with a cold climate like Canada, the cost of replacing worn parts of snowplow equipment such as blades, is enormous. A great number of these blades, with different mechanical properties, are available, but few studies have been made to facilitate their selection according to the optimum ratio of wear resistance to cost. To attain this objective, laboratory tests have been developed to correlate test results with those obtained from various road service conditions. The test materials, 1080steel in different states, 1090steel and D2 tool steel were chosen to cover a large range of hardness as well as different microstructures. Comparison of the wear resistance of the test materials relative to a standard revealed, in general, a good correlation between the results of laboratory cement wheel tests and those from the road tests. Scanning electron microscope wear scars of laboratory and road test specimens were compared to determine if similar removal material mechanisms were active. IntroductionThe correlation between wear tests in service and those performed with the aid of laboratory equipment is of considerable interest. As a matter of fact, when correlation exists, it allows one to evaluate the performance of a material more easily and at a lesser cost.Several authors have been interested in simulating various wear circumstances, especially abrasive wear. In particular, the simulation of wear on implements working the soil, has been tried by several researchers. The test using the rubber wheel seems to be the one giving the best correlation if the soil is rather sandy and its moisture content is low.However, little work has been done concerning wear on the plates, skids and blades attached to trucks used for levelling and snowplowing; such equipment being used to keep roads clear in northern countries. Now each year, companies introduce new steels by claiming that the y are superior in performance to their predecessor s and invariably they demand a higher cost. The study presented here was undertaken with the goal of making a more judicious choice of these steels by simulating their behaviour in the laboratory. Moreover, the examination of abraded surfaces in situ and in the laboratory have also allowed us to compare the respective wear mechanisms."

Jan 1, 1988

By Leong, William. Chong, Kam Weng

Deep Vibro Techniques namely Vibro Compaction (VC) and Vibro Replacement (VR) are widely used in ground improvement to densify sands and strengthen cohesive soils. This paper describes the application of these techniques for the construction of 20 Ha Iron Ore Stockpile Yard with the maximum stockpile height of 14 m in Quang Ngai, Vietnam. The challenges of this project are to limit the stockpile allowable settlement and the surrounding pile’s lateral movement. The soil condition can be summarized as 12m – 18m thick of loose to medium dense SAND with qc around 5-18MPa and followed by 2.0 – 6.0 m thick soft clay with qc around 0.4 – 0.8MPa. Localized very dense sand layer was found. VR using stone columns were installed in soft clay layer while VC was adopted to densify sand layer. Additional shallow stone columns were built in sand strata to minimize pile top lateral movement. The post construction performance was verified by a large 6 m by 12 m plate load test (PLT). The predicted PLT deformation using Priebe’s improved properties for the VR will be presented and compared with the measured plate load test results.

Nov 1, 2022

By Brent W. Madsen

To assist in fulfilling the Bureau of Mines goals of maintaining adequate mineral supplies to meet national economic and strategic needs, and to minimize the environmental impacts associated with mineral-processing operations, a leaching-solvent extraction-electrowinning technique was compared with a leaching-cementation technique for processing low-grade copper ores. Large-scale laboratory column leaching tests (6 to 7 tonnes) were conducted on minus 37-cm chalcocitic ore using ferric sulfate leach solutions at 25° C. Two columns contained 11 percent minus 1.27-cm fines, and two columns contained 32 percent minus 1.27-cm fines. After 416 days of leaching ore samples containing 11 percent minus 1.27-cm fines, copper extraction was 48 and 51 percent for the leach-cementation and leach-solvent extraction-electrowinning procedures, respectively. When leaching ore samples containing 32 percent minus 1.27-cm fines, the leaching-solvent extraction-electrowinning technique extracted 60 percent of the copper after 568 days, whereas the leaching-cementation technique extracted only 48 percent of the copper. In the latter test, leach solution ponded on the ore, which reflected a decrease in ore column permeability. The decreased permeability was attributed to the generation of excessive amounts of iron salts during cementation; the salts precipitated on the fine ore particles and in the interstices, thus plugging the column. This decrease also lowered the available oxygen, resulting in poor aerobic bacterial activity which slowed the oxidation rate of ferrous ions to ferric ions. Sulfuric acid required to maintain a leachant pH of 2 was nearly halved when copper was recovered by solvent extraction and electrowinning rather than by cementation.

Jan 1, 1979

By Jowett A

To compare open circuit lead primary rougher results at Mount Isa Mines Ltd., eleven surveys were made over five weeks of plant operation with varying ore sources, including a specially hauled sample of 100% Black Star ore. Three of the surveys which represented extremes in performance were selected for sizing, cyclosizing and assaying for lead, zinc and iron. A comparison of the flotation kinetics has revealed the similarities (including common problems) and differences for the surveyed ores. The concepts of entrainment and collectorless flotability are used to aid interpretation of the kinetic data.

Jan 1, 1978

To simplify programming, image analyzers commonly measure either linear intercepts or the areas of particles and grains to determine liberation based on polished section mounts. However, the concentration of minerals by flotation is based on reagents that interact with the exposed surfaces of the minerals. Thus, it is often perceived that image analyzers should measure the perimeter of the mineral of interest in the ore particles. A comparative liberation study of thirteen samples collected from a flotation plant processing a complex base metal sulfide ore showed that the liberation of the ore minerals determined by area measurements is very similar to the liberation determined by the exposed perimeters. The liberation determined by exposed perimeters is more appropriate only for those cases where the mineral texture is so complex that it is retained in the small particles generated in conventional grinding operations. Key words: Mineral liberation, Flotation, Image analyzers

Jan 1, 2003

By A. Pavlides

The purpose of this study is to compare estimation methods for lignite reserves at the Amyndeo mine and to provide a map of the estimated thickness of the coal seams. The reserves are first estimated with the method of Voronoi diagrams, using a group of mirror points to close the polygons around the boundaries of the mine. The second estimation method is based on Ordinary Kriging using the Spartan variogram model. Finally, the reserves are estimated by the mean lignite thickness value of the drill-hole data Two data sets are investigated. The first one corresponds to an exhausted part of the mine. The derived reserves estimates are compared with the actual data obtained by the Public Power Corporation (PPC) to validate the analysis. The second data set extends over the entire mine, and it is used to derive an approximate estimate of the total lignite reserves. A geostatistical analysis of the total layer thickness is performed. An interpolated kriging map of the total thickness variation is produced accompanied by a map of the kriging uncertainty.

Jan 1, 2008

By J. W. van der Merwe, A. Palazzo

"Owing to the corrosiveness of the untreated brackish cooling water typically used for steel mills (and other environments), it is important to treat the water and monitor corrosion in such systems. Generally, corrosion rates are monitored with corrosion probes inserted into a pipeline or vessel. This has been standard practice for many years, and is widely accepted in industry. Typically, two kinds of probes are used - electrical resistance and linear polarization resistance (LPR) probes.In this study, the effectiveness and accuracy of the LPR technique was evaluated by comparing the electrochemical measurements with the results of mass loss tests using corrosion coupons. The corrosivity of the environment, a synthetic brackish water, was varied by varying the calcium hardness and alkalinity, and to simulate actual plant conditions temperatures of 35°C and 45°C were used. In addition to the corrosion rate measurements, the iron concentration was measured, as well as the imbalance from the probe.The corrosion rates obtained by LPR were from 57% lower to 385%higher than those from the immersion tests. Most probe measurements were higher than the immersion results, and 50% of the probe results were50% or more higher than the immersion results. The best correspondence between the two methods was obtained at low calcium levels, except for one measurement that was 93% higher than the coupon results. There was no clear correlation between parameters such as temperature and total alkalinity and the difference between the results. It would therefore appear that LPR measurements can differ significantly from immersion results, and LPR results should therefore be used with caution in industrial applications."

Jan 1, 2015

By A. J. Brown

This paper presents an economic evaluation of resin ion exchange and solvent extraction for the purification and concentration of uraniferous solutions derived from sulphuric acid leaching operations. Capital and operating costs are related to the uranium tenors of pregnant solutions. The conclusions drawn herein apply only to cases where resin and liquid ion exchange are technically feasible process alternatives. It is shown that solvent extraction is economically favoured for pregnant solutions at tenors greater than 0.90 gpl U3O8. Pregnant solutions with tenors below 0.35 gpl U3O8 are best treated by resin ion exchange. Between these concentrations, the economics are marginal and the favoured process is indistinguishable within the limits of accuracy of this study.

Jan 1, 1979

By Tracy Brettmann, Shailendra Endley, Jason Leary, Shashank Valluru

"Augered Cast-in-Place (ACIP) piles have long been used for deep foundation support in sands due to their high capacity and speed of installation. Very little information exists on the effects on capacity if the auger needs to be withdrawn from the hole prior to grouting. This condition can occur due to various reasons including grout plug ups, equipment issues, and grout supply delivery problems. It can become a major problem for the project team if this happens during production because the capacity of that pile is then in question, potentially resulting in re-design, installation of replacement piles and/or schedule delays. This paper compares the results of two high capacity test piles installed into a dense sand where one test pile was installed by first completely withdrawing the auger from the hole after drilling to the pile tip depth and the other test pile was installed utilizing conventional industry standard methods. Identical load tests were then performed on both piles to compare the effects of the two different installation procedures. The load test results were surprisingly similar in this case although the grout take required for the test pile installed in the pre-drilled hole was significantly higher to produce a similar grout return depth.IntroductionAugered Cast-in-Place (ACIP) piles are used extensively along the Texas Gulf Coast due to their combination of high capacity and speed of installation resulting in one of the most cost effective deep foundation systems relative to other options. Soil conditions in this region typically consist of alternating layers of sand and clay in the upper several hundred feet. ACIP piles are routinely installed to depths of over 100-ft to achieve the required capacity depending on the site specific soil conditions and therefore typically extend through and tip into sand layers. Occasionally, due to various reasons like equipment issues and plug-ups, the auger is withdrawn from the hole without grouting. Significant sand in the soil profile could lead to questions about the effect of the auger withdrawal on the pile capacity and load deflection relationship.This paper summarizes the results of two identical pile load tests performed on adjacent piles of the same length with the main difference being one was installed using the standard installation procedure and the other was installed by first completely withdrawing the auger from the hole after drilling to the pile tip depth. The load test results were then evaluated using load transfer analyses to estimate the effect of the withdrawal of the auger on the load-deflection characteristics and capacity."

Jan 1, 2017

By The Hollinger Mill Staff

ALTHOUGH cylindrical grinding mills have long been used, there is wide divergence of opinion as to what constitute best operating conditions. Questions such as mill diameter, diameter-length ratio, revolutions per minute, type and size of grinding media, volume of charge, circulating load, pulp level, pulp consistency, liner design, and size of feed, have all been given serious consideration by competent scientific investigators as well as by operators. While many laboratory and plant tests have been made that would seem to justify certain definite conclusions, there is a dearth of definite data giving actual results obtained when a grinding mill working between fixed limits is operated at its maximum capacity under varying conditions as to ball load, circulating load, speed, etc. In tests conducted recently at the Hollinger, extending over a period of thirteen months, to determine the best operating conditions for a ball-mill 6 ft. 6 in. diameter inside the liners, some very interesting facts have been ascertained. This paper is presented in the desire to make public the results of these tests and not with the idea that the last word has been spoken in regard to grinding. We feel, however, that these facts justify certain definite conclusions and hope that, by the fullest and frankest criticism and discussion, the essentials necessary for the correct design and operation of grinding mills will become better known.

Jan 1, 1937

By The Hollinger Mill Staff

MR. E. L. Longmore (l), in presenting this paper at the Annual General Meeting, suggested that it was neither desirable nor necessary to read any but the briefest portion of the paper. He wished to emphasize that the scale of the milling operations at the Hollinger not only provided several units in continuous operation to ensure that the test work was being carried on under normal conditions, but also provided a large and competent staff to supervise such an extensive experiment and to criticize intelligently the recorded results. He wished to acknowledge the enthusiastic co-operation of the group of assistants whose excellent work was responsible for the confidence with which the conclusions could be presented. The following data were given as a supplement to the published paper. Following conclusion of the test work described in the paper, the size of the balls fed to the mill was changed from 3-in. diameter to 2Yz-in. Within a short time, a gradual rise in the tonnage rate took place and, as no other condition had been changed, the increased tonnage was credited to the changed character of the ball load. At present, the average daily tonnage for the mill with low discharge is 1,507 tons, an increase of 77 tons over the 1,430 tons ground when feeding a 3-in. ball. A similar increase in tonnage should be credited to the high-discharge test, if the 2;Y2-in. ball had been used at that time. This would raise the 1,117 tons ground at that time to a new figure of 1,177 tons. From a study made to determine those factors influencing the rate of pulp flow through the low-discharge mill, it became evident that the ball screen used on the high-discharge mill must have been restricting the flow through the mill, although this was not recognized at the time. Accord-ingly, it was decided to repeat the high-discharge run. Additional length was added to the 21-in. ring grizzly originally used, and the length of the reverse spiral was increased. The end of the grizzly was left open so that, if necessary, excess pulp might flow out. A coarse mesh 3 ft. by 3 ft. flat screen was provided to catch stray balls.

Jan 1, 1937

By Yuexin Han

The characteristics of comminuted hematite products using a high pressure grinding roller (HPGR) and conventional jaw crusher (JC) were comparatively studied by screening, optical microscope, scanning electron microscope (SEM) and Brunauer, Emmett and Teller (BET) techniques. Major properties investigated included product size distribution, particle cracks, specific surface areas and pore volumes. Results showed that in a closed-circuit operation with a screen opening of 3.2 mm, the HPGR and JC gave a similar P80, but the HPGR achieved a more uniform size distribution, more fines and a higher crushing efficiency. Three modes of cracks were identified, namely transgranular cracks, innergranular microcracks and intergranular microcracks. Transgranular cracks originated from shearing breakage. Innergranular and intergranular microcracks were generated by HPGR via the two main nonrandom breakage forms, preferential breakage and interfacial breakage, depending on their respective ability to endure stresses when mineral phases (particles) were loaded against each other. The BET values for the same-size fractions showed that HPGR products gave higher specific surface areas and more pore volumes than JC products. This indicates that HPGR-crushed products are easier to grind because of their high porosity and extra cracks.

May 1, 2012

By Seth H. Hamblin, John E. Regan, Dimitrios Palantzas

"Through reliability-based calibration of larger pile load test databases, recent developments of the Load Resistance Factor Design (LRFD) methods, which include geotechnical load and resistance factors for deep foundations design, have advanced the state of practice for geotechnical engineering. These load test databases upon which the current geotechnical load and resistance factors have been developed are limited, however, and do not provide sufficient data to support the development of resistance factors for uplift (tension) pile capacity. As such, some recommended resistance factors for uplift capacity estimation methods are simply derived by reducing the recommended resistance factors for piles in compression by a factor of 0.10.The results of three (3) tension loading tests performed to failure on 508 mm (20- in) square precast, prestressed concrete piles, driven though granular soil in southeastern Massachusetts, are presented and compared to static pile capacity calculations and dynamic analyses performed using the Pile Driving Analyzer (PDA) during installation. These results are presented along with current recommended LRFD resistance factors for driven piles under uplift (tension) loading conditions. These comparisons offer an assessment of the applicability and limitations of current resistance factors methods for piles under uplift loading and may be useful additions to the existing database for future calibrations of LRFD resistance factors for driven prestressed concrete piles in sand.IntroductionIn 2010, two (2) natural draft cooling towers were constructed by Kiewit (Kiewit) Construction Company of Woodcliff Lake, New Jersey at the Brayton Point Power Station in Somerset, Massachusetts. These massive 500 foot tall, 400 foot diameter cooling towers were designed to convert the existing “open loop” system into a “closed loop system” thereby dramatically reducing the thermal impact to Mt. Hope Bay.Based on design guidance from the German VBG Design Code and the Massachusetts State Building Code (MSBC), tension (i.e. uplift) and lateral loads due to wind loads and design seismic events represented the critical loading conditions.Weighing approximately 190,000 kips upon completion, the cooling tower loads are transferred to 44 foundation nodes (i.e. pile caps) evenly spaced along ring beams which run the circular perimeter of each tower base. Each pile cap node was designed to resist a 2,000 ton compression load, 200 tons in tension (i.e. uplift), and 600 tons in the lateral direction."

Jan 1, 2015

By Fuerstenau MC

Zinc cementation and carbon adsorption are both presently used for precious metals recovery from cyanide leach solutions. Each of these techniques exhibits unique processing features which renders them suitable to treat solutions with different properties and in different applications. In this study, the two processes are reviewed and compared. Operating data from a number of Merrill-Crowe precipitation and carbon adsorption plants in North America have been obtained and used in this analysis. The relative capital and operating costs for the two processes have also been considered.

Jan 1, 1993

By J. D. Noll, J. A. Organiscak, A. B. Cecala

Significant strides have been made in optimizing the design of filtration and pressurization systems used on the enclosed cabs of mobile mining equipment to reduce respirable dust and provide the best air quality to the equipment operators. Considering all of the advances made in this area, one aspect that still needed to be evaluated was a comparison of the efficiencies of the different filters used in these systems. As high-efficiency particulate arrestance (HEPA) filters provide the highest filtering efficiency, the general assumption would be that they would also provide the greatest level of protection to workers. Researchers for the U.S. National Institute for Occupational Safety and Health (NIOSH) speculated, based upon a previous laboratory study, that filters with minimum efficiency reporting value, or MERV rating, of 16 may be a more appropriate choice than HEPA filters in most cases for the mining industry. A study was therefore performed comparing HEPA and MERV 16 filters on two kinds of underground limestone mining equipment, a roof bolter and a face drill, to evaluate this theory. Testing showed that, at the 95-percent confidence level, there was no statistical difference between the efficiencies of the two types of filters on the two kinds of mining equipment. As the MERV 16 filters were less restrictive, provided greater airflow and cab pressurization, cost less and required less-frequent replacement than the HEPA filters, the MERV 16 filters were concluded to be the optimal choice for both the roof bolter and the face drill in this comparative analysis case study. Another key finding of this study is the substantial improvement in the effectiveness of filtration and pressurization systems when using a final filter design.

Aug 1, 2016

By E. Topuz, Xing Wu

INTRODUCTION The air flow in the mine ventilation network can be categorized in three ways: natural splitting, controlled splitting, and semi-controlled splitting. In natural splitting, the air flow in branches of a ventilation network apportions itself according to the aerodynamic resistance of the branches. In controlled splitting, a prescribed quantity of air is circulated through each branch. In semi-controlled splitting, flow in some branches is fixed to a desired quantity while in other branches it is split naturally. In practice, the mine ventilation networks present the latter two situations. Usually, the resistance of each branch in the network is known and the quantity of air in the working place is predetermined by safety considerations. If the number of branches whose airflow quantities are known is equal to or larger than the number of the fundamental meshes in the network, then the network becomes a controlled splitting network (Wang, 1982 a). Otherwise, it is a semi-controlled splitting network. If no booster fans are used in the network, the problem can be solved as either controlled or semi-controlled splitting networks in order to determine the main fan pressure and regulator size and location. All techniques are based on Kirchoff's flow and energy conservation laws and Atkinson's Equation. Booster fans can be used in controlled and semi-controlled splitting networks. It has been proven both theoretically and practically that using booster fans in these ventilation systems can improve flow distribution and reduce consumption of power (McPherson et al., 1985). The current techniques used in determination of booster fan location and size include simulation, linear programming, CPM and cutset operation, out-of-kilter, and critical-path crashing. These techniques can be classified into two groups. The first group, including linear programming, CPM and cutset operation, out-of-kilter, and critical-path crashing techniques, deals with controlled flow networks. The second, simulation, can be applied to the handling of semi-controlled flow networks. Mathematically, the problem of controlled splitting networks can be solved more easily than that of semi-controlled splitting networks. This is because the controlled flow network can be formulated as a linear programming problem, for the solution of which there are several sophisticated algorithms. Moreover, the semi-controlled flow network is usually modeled through nonlinear programming, for which there are no universally applicable solutions. The procedure described by Wu and Topuz, 1986, can be used to convert a semi-controlled network to a controlled flow network, and the problem can then be easily solved by the techniques used for solution of a controlled flow network problem. In practice, a mine ventilation network may change from controlled splitting to semicontrolled splitting with advancement of mining. This dynamic nature of mine ventilation makes the problem more complicated and necessitates development of new solution procedures which will consider the overall economics of the system. CONTROLLED SPLITTING NETWORK The problem in controlled splitting networks with booster fans is to determine the operational characteristic of the main fans, and the location and size of booster fans and regulators. The objective is to satisfy the air flow rate requirement for each branch while minimizing consumption of energy. The techniques used for the controlled splitting network problem include linear programming, CPM and cutset operation, out-of-kilter, and critical-path crashing. The objective function and constraints can be expressed as follows (Wang and Pana, 1971): [ ]

Jan 1, 1989

By V. E. Ross

A comparison was made of three methods for determining the contributions of true flotation and entrainment during batch flotation tests. In the first method size-by-size recoveries are measured in both the presence and the absence of a collector, a frother being present in both tests. In the second method the mass of solids and water recovered after a chosen time of flotation is varied by changing the rate of froth removal and the depth of the froth. The third method uses a transfer factor to describe the classification of solids and water, and hence the changes in the composition of the pulp, during a single flotation test. Batch flotation tests were carried out with artificial mixtures of pyrite and gangue, the amount of floated material recovered being calculated according to each of the three methods. The discrepancies betweeen the calculated amounts are analysed in terms of the assumptions that are made for the various methods and the most suitable conditions for the application of each method are identified

Dec 1, 1991

By BARONJUPTNER VON JONSTORFF, Andrew A. Blair, GUNNAR DILLNER

IT is a well-known fact that the results of different analysts, when operating on the same identical sample of steel or iron, are far from concordant, and it not infrequently happens that great annoyance and trouble in the commercial community are occasioned by these discrepancies. Attempts have been made at various times to investigate the causes of such differences and to obtain more reliable methods of analysis. The most important of these was made by the British Association in 1889: The following gentlemen, with Sir W. Roberts-Austen as Chairman, and Professor T. Turner as Secretary, were constituted as a committee to investigate the question, viz.: Sir F. A. Abel, Mr. Edward Riley, Mr. G. J. Snelus, Mr. John Spiller, Professor J. W. Langley, and Professor Tilden. Average .samples of steel were prepared in America under the superintendence of Professor J. W. Langley. The steels, in the form of turnings, were divided and distributed to sub-committees in Great Britain, America, France, Germany, and Sweden. In Sweden Professor Åkerman acted as Secretary to the Swedish committee. The Government Department of Tests was appointed in Germany; Mr. Ferd. Gautier in France; the British Association Committee in England; Messrs. William

Mar 1, 1905