Search Documents

By Vladimir P. Nebera, Kostas A. Matis, Piotr M. Solozhenkin, Anastasios I. Zouboulis

"Considered combination of two processes-selective sorptions of metals on biosorbents, inorganic sorbents and their following flotation by dissolved air(DAF). The process of biosorption, using as sorbent suspended non-living biomass (of Streptomyces genus, S. clavulgerus and S. griseus, fungi-Penicillium chrisogenum and Phizopus arrhizus, active slime) in fine particulate dispersion, combined with the application of flotation, for the consequent solid / liquid separation of metal-loaded biomass, has been studied for the removal of toxic and noble metals from dilute aqueous solutions. Biomass of microorganisms possesses great sorptive capacity, exceeding capacity of synthetic and natural sorbents. For instance, Micrococus luteus sorbed gold for 2h Au 253mg/g.The optimal conditions of effective flotation (more than 90%) of actinomyces biomass, loaded with toxic metal-cadmium were found. The adsorpbing flotation technique for copper, arsenic (III, V), chromium removal of metals on pyrite mineral fines (suitable industrial by- product) is an effective process.Dissolved-air flotation was also apllied in the case of dolomite/lead aggregates removal. A fatty acid was used as a flotation surfactant. The presence of dolomite fines aids the efficiency of the process, in comprarison with the flotation of alone lead hydroxide.A laboratory investigation was undertaken on the gold-thiourea system over the whole pH range, aiming at the separation complex of gold [Au(CS(NH2)2]2+ by flotation from dilute solutions. The reagents were found to improve gold recoveries remarkably: the recovery reached over 95%. A separation technique has been investigated, involving gold adsorptionon powdered activated carbon. Foam flotation was employed for the subsequent solid/liquid separation, using sodium oleate as surfactant. The content of Au were 500 mg/L in froth and recovery gold about 100%."

Jan 1, 2003

By Harri J. Lehto, Marja I. Oja

"The purpose of grinding circuits is to produce a particle size distribution that has as much as possible the wanted particle size class. In practise, however, the production of the wanted size class is not at all as simple as it sounds. In the case of industrial minerals for example the end user may give very tight specifications for the particle size distribution and for the chemical composition of the product. This leads in many cases to a very demanding grinding process, if we want to reach and maintain these specifications. In other words, the required size distribution is not necessary the best possible for the chemical composition.Screen models aim to predict the size distribution and flow of the products when a feed size distribution, feed tonnage, and grade efficiency curve are known. The approach used in this study was to select the most suitable empirical model for classification. Suitability of the existing models has been evaluated and accuracy compared. The tests showed that best grade efficiency model was the Rosin-Rammler model.The performance of a screen was evaluated in order to minimize the amount of fine material in the screen overflow. The experiments were carried out using Derrick single-deck screen, first alone with recycling and then in the closed circuit with a rod mill using two different scales. The separate screen tests showed the expected behaviour of the grade efficiency curves. In the pilot tests the Leonard screen efficiencies stayed almost constant. This suggests that the screen works better in a milling circuit. However the loading of the screen turned the product of the mill coarser that helped screening. The grade efficiencies obtained from the pilot tests had the same general shape. The regression analysis of the pilot test runs did not give an acceptable correlation for the parameters. This suggests that pilot tests cannot be directly used to scale or to optimise industrial milling circuits. However they can be used to predict the trends of behaviour of an industrial circuit."

Jan 1, 2003

By Tuan V. Nguyen, Louis Mittoni, Murray Rudman, CSIRO Minerals, Jean D. Swift, John B. Farrow, Phillip D. Fawell

"Research projects towards developing a sound understanding of the factors controlling the performance of gravity thickeners have been conducted by the AJ Parker Cooperative Research Centre for Hydrometallurgy since 1988 through AMIRA International. The current phase of this work, the three year P266D “Improving Thickener Technology” project, has 27 sponsors comprising mineral processing companies, reagent suppliers and thickener manufacturers.These projects have resulted in the development of a range of tools and techniques to address flocculation, thickening and associated dewatering issues. A number of Computational Fluid Dynamics (CFD) models have been developed to predict likely full-scale thickener performance under various process conditions, based upon a knowledge of thickener design, plant flows and assessment of the flocculation behaviour of the feed slurry. At the same time, various laboratory and plant studies have been used to quantify important thickener processes, and to provide validation of the CFD models. These numerical and experimental capabilities have been applied to various thickener issues across a broad range of mineral processing operations, leading to significant process improvements and substantial financial benefits.This paper presents an overview of the work conducted during these projects, focussing on the research tools, the nature of the information that has been gained and the implications for thickener performance."

Jan 1, 2003

By Hong Nguen Vu, Pavla Ovcarí, Peter Fecko, Barbara Tora

"The goal of the work was verification of bacterial leaching on the samples of black coal. From the results obtained it was clear that the coal is very well leached by Thiobacillus ferrooxidans and this technology can be used for coal desulphurization. A leaching time of one month is required, to remove 65% of total sulphur and 85% of pyritic sulphur.Microscopic evaluation of samples from Jaworzno showed, that it should be classed as a coal with low coalification, probably from black coal hemiphase. Macerals of the liptinite group were well distinguishable; they were of a dark brown colour. Evidence of coal mass oxidation was apparent on the fresh sample; in the test samples (during bacterial leaching) small grains were oxidised."

Jan 1, 2003

By M. Wright, A. H. Mwale

Opportunities to promote enhanced eco-efficiency (i.e. the simultaneous improvement in economic and environmental performance) for the minerals industry exist across the whole value chain. However, given the current operational models of major companies within the industry, and their focus on refining of primary ore bodies, it is important to understand where such opportunities might lie within minerals beneficiation, as it is here that major environmental pressures are felt. This paper examines the combination of unit technologies in a minerals beneficiation train, and explores the inter-relationship between energy input and water consumption for a given mineralogy and product recovery, as these are two critical environmental issues which affect the overall viability of many minerals plants. In particular, we review constitutive relationships for each of the sub-processes of comminution, flotation, and tailings dewatering, and suggest how these might be integrated into a unified hybrid model, with particle size distribution as the key design variable. Such a model can then be used to explore the trade-offs between different performance objectives (be these technical, economic or environmental) as part of a multiple-objectives optimisation exercise. In this way, it may become practical to envision situations in which concentrate grade and recovery are sacrificed, marginally, in order to reduce energy demand in comminution, and improve overall water recovery. In addition, this work provides a useful baseline against which to analyse the potential for dry processing. This becomes more relevant as the industry commits to advancing the sustainable development agenda in the face of the increasing life -cycle costs and impacts of minerals processing operations, coupled to declining ore body grades, and the desire to maximise recovery of minor and trace elements. Some specific observations for copper are offered, supported by operational data from Zambian copper mines.

Jan 1, 2003

By Frank Cappuccitti, Stephen Chryssoulis, Stamen Dimov

Surface organic microanalysis of platinum (Pt) and palladium (Pd) minerals by TOF MS1 is an attractive complement and potential alternative to conventional collector screening methods. Its attractiveness lies on its ability to focus on individual Pt and Pd mineral particulates, therefore providing a more targeted approach. The only requirement is knowledge of the occurrence (mineral composition, grain size and association) of Pt/Pd in the feed sample. This newly developed technique can detect, and identify collectors and quantitate loadings on individual Pt/Pd mineral particulates at plant concentration levels. Using this technology, collector loadings were measured on Pt and Pd metal plates and on PtS2 and PdS particulates which had been exposed together to aqueous solutions of different collectors. Thus, data were obtained on collector selectivity, the effect of sulphur; and of depressants, the latter in combination with TOF-LIMS2 results. Staged rougher tests on North American and UG-2 ores linked Pt/Pd flotation kinetics with collector loadings and ultimately grade/recovery.

Jan 1, 2003

By Jan J. Cilliers, Nicolas Barbian, Esther Ventura-Medina

"Froth structure and stability are known to play an important role in determining flotation recovery and selectivity. An industrial sampling campaign was performed to develop a clearer understanding of the effect of flotation variables on froth stability and the role played by the resulting stability on flotation performance.Froth stability was estimated as the fraction of bubbles bursting on the surface of the froth through image analysis and the solids loading on the bubbles at the top surface. The image processing was performed using the Smartfroth software on the top surface of the froth. The system gives measurements of bubble size distribution, average bubble size and froth velocity. The froth velocity was used to estimate the volume of bubbles overflowing the weir yielding to the fraction of air overflowing the cell. This measurement is the complement of the rate of bubble bursting on the top surface. Also samples of the lamellae were taken by touching the top surface of the bubbles with a glass slide. Then, the amount of solids and liquid per unit area in the bubble lamellae was estimated gravimetrically. Both measurements were used to establish stability criteria.It was generally found that high air flowrates resulted in a drop of froth stability showed by a drop in the fraction of air overflowing and a decrease in the solids loading and simultaneously produced low copper recoveries and low concentrate grades. Deep froths only allowed a small fraction of air to overflow and consequently produced low recoveries. An increase to high frother dosage did not result in an improved fraction of air overflowing, hence producing low recoveries. Likewise, low frother addition only allows low recoveries because the resulting unstable froth cannot support itself. Low aeration was found to be preferential as it allowed high particle loading, therefore high bubble stabilisation, and low gangue entrainment, resulting in the highest flotation performance."

Jan 1, 2003

By Stoyan N. Groudev, Ioannis Paspaliaris, Plamen S. Georgiev, Irena I. Spasova, Kostas Komnitsas

"The uranium deposit Curilo located in Western Bulgaria for a long period of time was a site of intensive mining activities including both the open-pit and underground mining techniques as well as in situ leaching of uranium. The ore was rich in pyrite and, apart from uranium, contained some non-ferrous metals. The mining operations in the deposit were ended in 1990 but since that time the deposit is a permanent source of acid drainage waters. The generation of these waters is connected mainly with the oxidation of pyrite and other sulphide minerals by the indigenous acidophilic chemolithotrophic bacteria.Laboratory and pilot-scale studies revealed that an efficient clean-up of the above-mentioned waters was achieved by means of passive systems of the wetland type. In 2001 a wetland was constructed in the deposit to treat a portion of the acid drainage waters which were discharged to a small ravine. This ravine reaches into the Teina, a small river that is a tributary of the upper reaches of the Iskar river.The wetland was constructed in the ravine and its bottom was located on intrusive rocks with a very low permeability. The rocks were covered by a 30 – 50 cm layer consisting of a mixture of soil, plant compost and crushed limestone. The wetland was characterized by an abundant water and emergent vegetation and a diverse microflora. Typha latifolia, Typha angustifolia and Phragmites australis were the prevalent plant species in the wetland but species of the genera Juncus, Eleocharis, Carex and Poa as well as different algae were also well present.The acid drainage waters treated by the wetland had a pH in the range of about 2.5 – 4.1 and contained uranium, radium, some heavy metals (iron, manganese, copper, zinc, cadmium, lead), arsenic and sulphates in concentrations much higher than the relevant permissible levels for waters intended for use in the agriculture and/or industry. The watercourse through the wetland covered a distance of about 0.2 – 1.0 l/s."

Jan 1, 2003

By Genrich A. Denisov

Principles of ecologically friendly geo-technologies construction are formulated on the basis of the works of the leading scientists in the field of rational processing of mineral resources of the Russian Academy of Sciences. One of these concepts is a new direction in rotational disintegration of mineral and technogenic raw material with the purpose of effective subsequent enrichment. A large amount of practical experience has been accumulated in the industry on selective destruction by application of vibrational method on initial material with consideration of its physical-mechanical properties and mineral structure. This method is successfully realised in a conic inertial pulverisation devices constructed by Mekhanobr Institute (St.-Petersburg). A new construction of conical-inertial pulverisation with a modification of disintegration force has been developed recently at Mekhanobr Institute. It allows, in some cases, to increase selectivity of subdivision and to avoid over-disintegration. Advantages of “wet” subdivision methods in conic inertial disintegration devices are proved. The efficiency of such subdivision can be increased due to rational use of the phenomenon of a dislocation-dynamic diffusion. Its advantage is proved by the effect on disintegrated material by frequency-modulated variable electrical fields of low power. It was shown in industrial conditions, that the productivity of the spherical mill can be increased by 10 %, and that it also prevents ecologically harmful surface active substances (SAS) in milling zone and it reduces the dust content in the air. The techniques of tentative estimation of efficiency of various proceedings technogenic materials and their classification are created. The scientific fundaments of rational, ecologically friendly technology of processing and regeneration of various kinds of technogenic raw material are developed. A number of effective technologies on processing of wastes is developed and inculcated in concentration factories and metallurgical factories.

Jan 1, 2003

By P. Huang, D. W. Fuerstenau

Various factors that control the natural flotability of talc, including differences due to crystal face and edge effects, were delineated. For study of talc depression phenomena, dextrin was used as a model depressant. For the enhancement of talc flotation, dodecylammonium acetate (DDAA) was used as the collector. The interaction of DDAA with talc was assessed by flotation experiments and electrokinetic (zeta potential) measurements, from which the role of hydrophobic bonding can be demonstrated. The results presented in this paper provide useful information on how the crystal chemistry and surface chemistry control the electrical double layer, adsorption phenomena, wettability, and flotation behavior of talc in aqueous media.

Jan 1, 2003

By F Battaglia-Brunet, D. H. Morin, Foucher S. ’

In the mineral-processing field, bioleaching of metal sulphide concentrates is currently operated in very large mechanically agitated tanks. Such bioreactors have considerable requirements in terms of power consumption for mixing, aeration and heat transfer. The use of slurry bubble column, where the particles are suspended by gas motion as well as by liquid upward flow, could be an alternative to achieve bioleaching in more economical conditions. In order to compare these two techniques, a unit of stirred reactors in cascade and a bubble column were comparatively operated at laboratory scale in the bioleaching of a pyrite concentrate. For both pieces of equipment, experiments were performed not only in batch mode but also in continuous conditions.

Jan 1, 2003

By Jeremy Bosman, Niel Lourens

"Cyclones are used in four main applications in primary wet heavy mineral processing plants i.e. upfront desliming, densifying, and concentrate stacking and tails deposition.UPFRONT DESLIMINGMany heavy mineral ore bodies are associated with high slimes varying from 15 to 60%. The increased viscosities arising from the slimes impact negatively on heavy mineral recovery in the spiral plant. Recoveries can be reduced by up to 10%. Cyclones have proven to be able to remove slime efficiently with acceptable loss of valuable heavy mineral in the overflow. For low slime applications, typically below 25%, it is possible to achieve the desired results using a single stage of cyclones, whilst for higher slimes, two stages of cyclones are required. The overflow of the primary stage is retreated through a second stage to minimize heavy mineral loss."

Jan 1, 2003

By Victor Avdokhin, lvan Kovlekov

"Studies of the process of concentration of gold-containing sands and technogenic resources show that the main reason of gold losses is misbalance between size of gold particles and degree of roughness of mineral bed. This results in free circulation of small gold particles in pores with following washing off the particles from the bed by turbulent flows.We elaborated design of magnetic sluice, intended for improving efficiency of recovering fine gold by optimizing magnetic field of a magnetic system on the base of forming catching coat from magnetic mineralsThe magnetic sluice includes trough-shaped body from magnetic-conducting material. On the bottom of the trough, magnets are placed, and opposite poles of the magnets are oriented outwards. Thickness of the bottom is determined to be sufficient for locking magnetic lines of force of the magnets on each other. The magnets of rectangular plate shape are placed on the rough bottom, covering the whole surface or with spare areas; the magnets are placed across the trough with alteration of rows with the same pole orientation."

Jan 1, 2003

The introduction and subsequent ascendancy of autogenous and semi-autogenous milling for comminution circuits has undoubtedly lead to many economic advantages. From a process viewpoint, however, at least one downside has been the sensitivity of these machines to process input variations. Foremost among these is ore competence, though feed size comes a close second. This sensitivity is due to the reliance of ag/sag mills on the feed ore for their grinding media competence and size. Autogenous (ag) mills are the most sensitive in this respect with semi-autogenous (sag) mills being increasingly less so as the ball charge is increased. Increasingly designers and operators are recognising the importance of feed size and the benefits of manipulating it to improve performance. This has been achieved by installing additional crushing/screening capacity ahead of the ag/sag, better operation of the primary crusher, better control of segregation cased by stockpiles and bins and by changing blast design/practices to modify the ROM size distribution. This paper describes the influence of feed size distribution on ag/sag mill performance with reference to a range of plant data. It particularly concentrates on how changing blasting practice to modify ag/sag feed size can have a considerable overall benefit.

Jan 1, 2003

By M. T. Spyridopolous, S. J. Neethling

"There has been a significant research interest in the stability of single thin films, including prediction of their failure based on disjoining pressure and film thickness experiments and DLVO theory. It is, however, notoriously difficult to correlate information on the failure of a single thin film with measurements of foam stability. This is particularly true for the prediction of dynamic foam stability indicators, such as equilibrium froth height and surface bubble size.An alternative experimental approach to studying single thin film failure for predicting foam stability is to measure directly the coalescence probability between two approaching bubbles in a surfactant solution. These experiments can be performed under different chemical and physical conditions by using a micromanipulator that allows variation of bubble size, surfactant concentration, approach velocity and stationary position. The interface radius of curvature, the time to coalescence and the film area between the bubbles can be measured directly using image analysis.In a foam, the force exerted on the thin film separating two bubbles depends on the local radius of curvature of the Plateau borders and vertices, which depends in turn on the local liquid content. Results from different bubble sizes in the micromanipulator therefore translate into regions of different liquid content in the foam. Small bubbles, having a high radius of curvature correspond to the low-liquid content (Polyhederschaum) region higher in the foam, and large bubbles to the wetter (Kugelschaum) region closer to the liquid-foam interface. Experiments with different holding distances between the bubbles correspond to differences in bubble size in the foam, as the size of the film between the bubbles varies."

Jan 1, 2003

By Kal Sastry, Anshuman Prasad

"Agglomeration of fine iron ore concentrates is carried out in two steps:Balling, which refers to the production of green pellets by tumbling moist fines powders in drums, discs, or cones, and Induration, which refers to the subsequent thermal treatment of the green pellets in straight grate or grate-kiln systems.This paper addresses computer simulation of balling or green pelletization step. The process of agglomeration from a mechanistic point of view can be described as the transformation of moist feed to well-formed species or pellets. The moist feed is composed of fines of infinitesimal size while well-formed species are of finite size and number. These well-formed species undergo changes in their size and number through different mechanisms of agglomerate formation and growth. Nucleation, coalescence and layering are the three main mechanisms that describe the formation and growth of granules during an agglomeration process. A population balance equation enables quantitative description of the evolution of agglomerate size distribution due to the occurrence of each of the mechanisms.A numerical solution for the population balance model is necessary and is obtained by the use of a sizediscretization approach. The numerical solutions provide a suitable basis for the analysis of the agglomeration process. Also a coalescence kernel has been introduced that is capable of describing the simultaneous occurrence of nucleation, coalescence and layering mechanisms. The effect of feed moisture has been described using a moisture factor with an exponential functional dependence on the moisture content. The effect of the feed to seed ratio is modelled by the introduction of a crowding factor. It was assumed that only the feed particles which are much more in number experience a crowding or restricted in space interactions effect. A computer simulation program has been developed with MATLAB as the simulation language."

Jan 1, 2003

By S. J. Neethling

"Flotation is a complex process with many factors that influence the performance of a cell. This paper examines the impact of several key variables on the performance of flotation cells. The study makes use of a simulator for the froth phase of the flotation vessel, as the main purpose of this study is to separate the impact of different variables from one another.In experiments it is very difficult to separate the effect of variables from one another as a change in a singe input parameter, such as air rate, invariably also affects many of the other key variables, such as bubble size and bubble bursting rates. This interaction between variables makes it very hard to experimentally separate the impact of, for instance, the air rate from that of bubble size or bursting rate. In the simulator, though, it easy to make independent changes to these variables.In order to achieve this a base case simulation that is typical of many of the rougher cells found in sulphide flotation circuits was carried out. Four key variables, air rate, incoming bubble size, surface bubble size and the fraction of the bubbles flowing out of the cell unburst, were varied in order to obtain the system’s response.The most significant results were obtained when the bubble size was varied. An increase in the incoming bubble size has the effect of reducing the pulp zone recovery, but increases the froth zone recovery. If it were possible to vary the incoming bubble size without influencing the overflowing bubble size, the impact of the change would be unclear because of these two competing effects. The effect of a change in the overflowing bubble size has an opposite effect on the froth recovery, with larger bubbles resulting in lower froth recoveries."

Jan 1, 2003

By Sun Chuanyao, Jia Muxin

Generally, if the bond valence (BV) between metal ion and oxygen of the mineral is greater than 0.8, only surface species –Me-O-H or –Me-O exposures on the mineral interface, if the BV is between 0.2-0.8, the metal ion exposures on the interface, and if the BV is less than 0.2 the metal ion dissolves into water. The PZCs of oxide minerals and island silicate minerals are determined by the bond valence of the bond in the mineral crystals, the higher the bond valence the lower the PZC of the mineral. So the PZC sequence of Al2SiO5 minerals is kyanite> andalusite> sillimanite. The general rule for framework silicate minerals is as: The higher the Al content, the lower the PZC of the mineral, when the ratio of Al to Si in the tetrahedron skeleton is below 1:3. The higher the Al content, the higher the PZC of the mineral, when the ratio of Al to Si in the skeleton is between 1:3 and 1:1. The general rule for layer silicate minerals is summarized as: the three-layer silicate minerals have much lower PZCs than that of the two-layer silicate minerals. The threelayer silicate mineral with the octahedron layer occupied by lower charged cations has lower PZC than that of the three-layer silicate mineral with the octahedron layer occupied by higher charged cations. The twolayer silicate mineral with the octahedron layer occupied by lower charged cations has higher PZC than that of the two-layer silicate mineral with the octahedron layer occupied by higher charged cations. The layer silicate mineral with Al substitution in the tetrahedron layer has lower PZC than that of the layer silicate mineral without Al substitution in the tetrahedron layer. Such pyroxene minerals whose M1 is occupied by a trivalence cation and the M2 is occupied by a monovalence cation has lower PZCs than those both M1 and M2 are all occupied by bivalence cations. Such hornblende minerals whose M2 is occupied by a trivalence cation and the M4 is occupied by a monovalence cation has lower PZCs than those both M2 and M4 are all occupied by bivalence cations. There is no Si4+ or 4 coordinated Al3+ exposures on the surface of framework silicate minerals, the metal ions of three layer or chain silicate minerals are compact between tetrahedron layersor chains, they hinder the metal ions to interact with oleate acid radical, the metal ion octahedron layer exposures on the surface of a two layer silicate mineral, but the the oleate acid radical can not attach to a metal ion by replacing a –OH- that connects three metal cations in the same time, So the floatability of such silicate mineral is bad. The metal octahedron chain style of an island silicate mineral affects it floatability, the mineral exposures octahedron chain metals with two dangling bonds in two directions floats better than the mineral exposures octahedron chain metals with two dangling bonds in one direction.

Jan 1, 2003

By K. A. Natarajan

Interaction with microorganisms results in significant surface chemical changes on minerals. For example, chemolithotrophs such as Thiobacillus ferrooxidans and Thiobacillus thiooxidans significantly alter the surface chemistry of several sulfide minerals such as pyrite, chalcopyrite, sphalerite and galena. Heterotrophic bacteria such as Bacillus sp. could alter the electrokinetic properties of oxide minerals such as hematite, alumina, silica, kaolinite and calcite. Microbe-mineral interactions could thus be beneficially utilised in mineral beneficiation as different from bioleaching. Microbially induced mineral beneficiation unlike bioleaching deals with interfacial phenomena at mineral-solution-bacteria interfaces and surface chemical changes are brought about at very rapid rates amounting to a few minutes resulting in mineral separations in an aqueous medium.

Jan 1, 2003

By Briony C. Ruse, David V. Boger, Fiona Sofra1

"The sustainability of the mining industry must be addressed through improving tailings management. The use of thickened tailings and pastes stems from a paradigm shift towards designing tailings disposal sites to suit their surrounding environment, rather than manipulating the environment to accommodate the tailings. Thickened tailings and pastes usually exhibit non-Newtonian behaviour, so a thorough understanding of the rheological characteristics of the material is required in order to determine the disposal plant operating conditions.As environmental factors translate into economic issues, minimising waste production using dry disposal methods is gaining popularity. The use of rheological information is of high importance in evaluating the spreading characteristics, the transportation and the dewatering of thickened mineral tailings. The principles outlined may be applied to many industries encompassing a wide range of waste materials. The main obstacle to the widespread implementation of dry disposal methods has been the high initial capital cost of the methods due to the dewatering step, and the high cost of pumping thick suspensions across any great distance. Proper and thorough costing of tailings disposal activities combined with an accounting approach that shows closure and post-closure care costs on a per tonne basis of tailings disposed throughout the life of the tailings facility will show that dry disposal methods are more favourable than they might seem on initial appraisal."

Jan 1, 2003