Opaque and ore minerals associated with sedimentary rocks

Opaque phases are minor constituents of most sediments, either as part of the elastic component and hence related to a source area, or as authigenic phases and so indicative of the depositional environment and diagenetic conditions. Many sediments contain both detrital and authigenic opaque minerals. To overcome the problem of the paucity of opaques in most sediments, it is advisable either to select atypical specimens (for example, heavy mineral bands from elastic sequences) or to upgrade the opaque content by making grain mounts. For all sediments, grain mounts are prepared after heavy mineral separation, but, in addition, carbonate-rich rocks can be digested in dilute acid solutions in order to upgrade their opaque mineral content. Sections of the typical (and unprocessed) sediments also are prepared. Upgraded material is used to investigate the full range of mineral species present and to determine their relative proportions. However, textural information is more readily obtained from representative material; this is especially true for the authigenic opaque mineral component. Reflected light studies are potentially useful in helping to provenance clastic grains and, because many of the opaques are iron or iron-titanium oxides, the studies also play a role in explaining geomagnetic results. Polished sections of very fine-grained rocks, like shales and mudstones, have better surfaces than those of thin sections and this allows better fabric analysis and more accurate modal analyses to be made. The use of polished thin sections rather than thin sections in the archaeological study of pottery has illustrated the advantages of this technique. (Amstutz and Bubenicek, 1967; Berner, 1984).

The opaque and accessory minerals generally belong to the heavy mineral suite of oxides and silicates. Detrital sulphides are uncommon except in young, immature sediments or those formed under unusual reducing conditions. In maturer sediments, primary sulphides occur only within sealed rock fragments where they have been protected from sedimentary oxidation. Original iron-titanium oxide minerals and mafic silicates (especially biotite) are extensively altered to haematite and TiO₂. minerals, but from textural evidence it is possible to determine their original mineralogy. Authigenic, often euhedral, haematite and TiO₂. minerals form overgrowths on detrital grains of the same mineralogy but have a different minor element chemistry. Carbonaccous material also forms part of the detrital fraction and measurements of the reflectance of vitrinite are used to determine the thermal history of sediments.

In addition to the quartz, feldspars, phyllosilicates and rock fragments that comprise the bulk of these rocks, they contain opaque detrital phases. These include magnetite, ilmenite, (often oxidized to haematite and TiO₂, minerals to give 'martite' and 'leucoxene', respectively), rutile, anatase, brookite, chromite, sphene and heavy mineral silicates, including zircon, tourmaline, epidote, pyroxenes and amphiboles, along with carbonaceous material. Less commonly, cassiterite, mixed oxides of the rare metals (Nb, Ta, U and Ti), monazite, xenotime and native gold are present. Sulphides are restricted to rock clasts.

Authigenic components include quartz, calcite, dolomite, ankerite, chalcedony, zeolites and sulphate cements, together with lesser amounts of haematite, rutile, anatase, pyrite, marcasite, pyrrhotite and base metal sulphides. The replacement of magnetite along (111) crystallographic directions by haematite forms the very distinctive pseudomorph known as martite.

The replacement of ilmenite oxidation-exsolution lamellae in magnetite by TiO₂, minerals, coupled with the dissolution of magnetite to give an open boxwork texture, is often called leucoxene. (Golding, 1961; van Houten, 1968; Dimanche and Bartholome, 1976; Turner and 'Archer, 1977; Holmes et al., 1983)

The mineralogy of the opaque detrital fraction is similar to that of sandstones but has a much finer grain size and only the TiO₂ minerals, carbonaceous material and sulphides are readily apparent without using oil immersion lenses at high magnifications. Authigenic sulphides are common and include framboidal pyrite formed from precursor iron sulphides (greigite and mackinawite), pyrite and marcasite associated with carbonaceous matter or authigenic carbonates and discrete grains of pyrrhotite and base metal sulphides.

Siderite nodules, which are essentially pyrite and siderite, often have base metal sulphides, especially sphalerite, within them. (Love and Zimmerman, 1961; Raiswell, 1976; Hudson, 1982; Love et al., 1983, 1984)

These, too, have the same opaque mineralogy as sandstones and shales, especially if the limestones are argillaceous or arenaceous. Haematite, TiO₂ minerals and carbonaceous material are the most common detrital opaque phases, and pyrite and marcasite are the most abundant diagenetic sulphides. Other common base metal sulphides, including sphalerite, galena and chalcopyrite, are present in minor amounts. Framboidal and recrystallized framboidal pyrite and the selective replacement of coarse-grained carbonates by sulphides, especially within fossils, are characteristic textures. (Love, 1970)

All coals have a sulphur content and, in addition to pyrite and marcasite, a range of sulphides have been reported including sphalerite, galena, chalcopyrite, millerite and bornite. The sulphides (mainly pyrite and marcasite) form framboids, recrystallized framboids or irregular aggregates associated with nodular siderite. They replace plant remains or form veinlets along the cleat. (Bocter et al., 1976; Laznicka, 1981; Swaine, 1984)

Placers are Tertiary to Recent in age, and include a number of well recognized subclasses which reflect both the mechanism of concentration and depositional environment. Amongst the most important are alluvial and beach (littoral) placers, although screes (deluvial) and aeolian placers are mined. Many placers can be related to their proto-ores, although for reworked placers this is more difficult. Sulphides are unusual as modem placer minerals, although they have been recorded. Placer minerals have a number of properties. They need to be chemically stable in the weathering cycle, be mechanically strong, lack good cleavages, be hard (Moh's hardness > 6) and have a high density. Historically, placers have produced much of the world's noble metals and gemstones and still remain important sources of them.

Placers from ultramafic or mafic parents are mined for platinum group metals, gold, diamond, ilmenite, magnetite, chromite, rutile, and leucoxene; those from alkalic ultramafic parents produce pyrochlore group minerals, apatite and Nb-Ta-bearing minerals. Placers from acidic volcanic and plutonic rocks produce ilmenite, rutile, cassiterite, wolframite, scheelite, columbite, thorite, zircon, monazite and gold. Unconsolidated placer ores are usually prepared as grain mounts. Consolidated fossil placers can be regarded as normal ores and made into polished blocks or polished thin sections. (Hails, 1976)

The Witwatersrand banker gold-uranium placer of the Kaapvaal Craton is perhaps the most famous example of a fossil placer. There has been much speculation on the origin of the deposit and two main arguments have been proposed: either that it is a fossil placer, with detrital native gold, uraninite and pyrite grains mechanically introduced into a basin followed by redistribution of elements (notably uranium) by metamorphism, or that the gold, uranium-bearing minerals and pyrite are epigenetic and hydrothermal in origin. The host rock is an oligomictic conglomerate of vein quartz, chert, jasper, banded iron formation and quartzite pebbles in a matrix of secondary quartz and phyllosilicates. It contains detrital, abraded pyrite, chromite, zircon, uraninite, gold and other heavy minerals. Locally, gold and uranium are found within narrow carbonaceous-rich horizons. Over seventy minerals have been recorded.

These are common and widespread and are often associated with sedimentary iron ores. There are a number of classes of deposit, including: deep sea manganese nodules, precipitates from stagnant water, non-volcanogenic quartz-clay-glauconite and manganese carbonate associations, and volcanogenic sedimentary manganese ores.

Sedimentary manganese ores mainly comprise manganese oxides and hydroxides or manganese-bearing carbonates. Manganese sulphides are uncommon ore minerals (unlike iron sulphides) and manganese silicates are characteristic of metamorphic or metamorphosed deposits. Individual manganese minerals, especially oxides and hydroxides, are difficult to distinguish in reflected light, hence the terms 'wad' or 'psilomelane' are often used to denote the presence of mixtures of unidentified soft and hard manganese minerals, respectively. X-ray powder diffraction is a more successful method for the identification of these species. Similarly, the characterization of manganese-bearing carbonates requires a knowledge of their chemistry. (Roy, 1968,1976, 1981; Bonatti et al., 1972; Burns and Bums, 1979). Small (up to 30 cm in diameter) manganese nodules occur in vast numbers over large areas of the deep ocean floor. They are authigenic in origin and often nucleate on detrital grains, organic material or earlier nodules.

They comprise mixtures of oxides and hydroxides of iron and manganese, together with clays, zeolites, quartz, opal, baryte and feldspar. The nodules carry significant amounts of cobalt, nickel and other minor metals sorbed onto the major phases and are regarded as having potential economic interest.

In reflected light, individual minerals cannot be identified by their optical properties. However, within the collomorphic banding that is characteristic of the nodules, separate bands can be distinguished by reflectance differences and by the presence or absence of bireflection and anisotropy. Amorphous iron and manganese hydroxides have low reflectances (less than 9%), polish poorly and are isotropic, whereas todorokite and birnessite-rich bands polish better, have reflectances of approximately 15% and show marked bireflectance and anisotropy. X-ray powder diffraction of the different bands is the usual method employed to identify their mineralogy.

Elsewhere, manganese crusts are associated with modem deep sea volcanogenic sulphide deposits and other volcanic activity close to seamounts; these crusts are often cobalt enriched. (Bums and Bums, 1977a,b; Sorem and Fewkes, 1977, 1979; Bischoff and Piper (eds), 1979; Muller, 1979; Heath, 1981)

These are the manganese equivalent of bog iron ores, and comprise hydrated manganese oxides.

These occur within shallow marine deposits in stable cratonic areas. The ores are conformable and form lenticular or nodular horizons showing mineralogical zoning, from oxide-rich to mixed oxide-carbonate to carbonate-rich facies. These facies types primarily are related to palaeogeography but also to subsequent supergene enrichment processes. In the shallow water oxide zone the ore mineral assemblage comprises manganite, pyrolusite-manganite or psilomelane associated with limonite. This zone passes into a transitional zone of manganese-bearing carbonates and manganese oxides and hydroxides which, in turn, passes into the deeper water carbonate zone. This zone consists of nodules and lenses of cryptocrystalline manganese-bearing carbonates, including calcic rhodochrosite, manganocalcite and manganoan calcite.

These are divided into the Moroccan-type, found in stable platform environments, and the Appalachian-type, found in geosynclinal sequences. Moroccan-type deposits comprise manganese oxides, pyrolusite and psilomelane. They occur in limestones and dolomites, often in association with red bed sequences. Appalachian-type deposits are found in limestones, dolomites and volcanic sequences where they form beds and lenses of manganocalcite and calcic -rhodochrosite within carbonate host rocks.

These are widespread but individual deposits are small and of little economic importance. The ores are stratiform or lensoidal, and comprise manganese oxides interbanded with chert and jasper on a centimetre scale. Many are associated with volcanogenic sulphide deposits. Metamorphism of these deposits produces complex silicate ores that vary in mineralogy with the degree of contact and regional metamorphism. Braunite, hausmannite, rhodonite, bustamite, tephroite, pyroxmangite and piemontite associated with haematite, magnetite, limonite, and iron and base metal sulphides are typical minerals of this association. (Zantop, 1981)

Iron is a common constituent of sediments, and, hence, sedimentary iron ores are widespread and all classes of deposit have been exploited from the Iron Age to the present. In order of increasing economic importance, the main classes are bog iron ores, clay ironstones, minette-type ores, Clinton-type ores and iron formations. Only the last two are mined. (James, 1954; Murray, 1979)

These are of very localized occurrence and are found in present-day lake and marsh sediments where iron has been fixed by humic complexes. The iron is subsequently released by bacterial action and is precipitated as ferric oxides and hydroxides associated with minor amounts of siderite and vivianite.

Limonite is the generalized term for poorly characterized iron hydroxides which include mixtures of goethite, lepidocrocite, akaganeite and feroxyhyte.

These deposits are widespread but low grade. They occur within marine sequences and represent the fossil equivalents of bog iron ores. Dolomite and siderite are the main iron-bearing minerals and, historically, the ores were roasted to convert iron carbonate to iron oxide. (Curtis et al., 1975)

This is the commonest type of unmetamorphosed ironstone, but, with an iron content of 30-35% by weight, minette-type ores are generally uneconomic. The ironstones are found in shallow marine carbonaceous shale, mudstone, marl, limestone and ironstone sequences. The main iron-bearing minerals are limonite, siderite and chamosite with lesser amounts of magnetite, haematite, greenalite and pyrite. Typically, the iron ores comprise ooids of chamosite, haematite and limonite which have nucleated about detrital grains or fossil fragments. Quartz, calcite, cellophane and silica are the main gangue minerals. Ores of Jurassic age were mined in Europe in the nineteenth and twentieth centuries. (Kimberley, 1979, 1981; Maynard, 1986)

The class is named for the Silurian Clinton Formation of the eastern United States of America where oolitic ironstones, together with argillaceous and carbonaceous shales, limestones and dolomites, occur within shallow marine sediments. Haematite, chamosite and siderite are the main iron-bearing minerals and are associated with calcite and silica.

Iron formations are defined as thinly-bedded or finely laminated chemical/biochemical sediments comprising chert (jasper), plus iron-bearing oxides, carbonates, silicates or sulphides. These are characteristically enormous deposits and range from Archaean to Lower Palaeozoic in age, although most are Proterozoic. They are subdivided, partly on their tectonic environment, into Algoma-type, found in volcano-sedimentary sequences in greenstones belts, and into Lake Superior-type, found in continental-shelf sequences of quartzites, cherts, black shales and dolomites. The mineralogy of the ores is variable and this reflects primary sedimentary facies variations, later diagenetic effects, and metamorphism.

Unmetamorphosed iron formations can contain four facies types, which are believed to represent original variations in the sedimentary conditions. These are an oxide facies formed in well oxygenated waters and comprising magnetite-chert or haematite-chert; a carbonate facies dominated by siderite; a silicate facies containing chamosite, glauconite, greenalite and locally minnesotaite and stilpnomelane; and a sulphide facies of pyrite and minor pyrrhotite formed in an oxygen-poor environment often associated with black shales. Silica is the main gangue of the oxygen-rich facies, whereas carbonates including siderite, ankerite and dolomite are important in the more reduced facies. Many iron formations have been metamorphosed, their original sedimentary and diagenetic features destroyed and their original mineralogies altered to complex iron silicate associations. Supergene oxidation of magnetite to martite and primary iron minerals to limonite is widespread and economically important. (Eichler, 1976; Kimberley, 1978; Trendall and Morris (eds), 1983)

These are the major ores of aluminium, but with increasing amounts of hydrated iron oxides, bauxites grade into laterites. Bauxites, which are Cretaceous to Recent in age, can be divided into a laterite crust-type formed by the weathering of aluminous non-calcareous rocks and a karstic-type formed from aluminous limestones or dolomites.

In bauxites the aluminium oxide mineralogy changes with increasing metamorphic grade from gibbsite to boehmite to diaspore and finally to corundum. These minerals are accompanied by quartz, kaolinite, halloysite, limonite and haematite, together with a very large number of relict phases from the parent rock, including anatase, rutile, sphene, zircon, pyrite and hogbomite. Pisolitic, oolitic or conglomeratic fabrics are common in bauxites and many ores retain relict fabrics of their parent rocks. Individual oxide and hydroxide minerals are very difficult to distinguish in reflected light and X-ray diffraction methods are usually employed to characterize the mineralogy of bauxites. (Bárdossy, 1982)

In unmineralized basic and ultrabasic rocks, nickel is held in olivine and to a lesser extent in other mafic silicates, magnetite and pyrrhotite. It also forms discrete sulphides and arsenides, including pentlandite, millerite, heazlewoodite and niccolite. The nickel contribution from olivine is the most significant, and, hence, the most productive parent rocks are peridotites and similar ultramafics although, norites are also important. Nickel, together with magnesium and iron, is released on weathering and either reprecipitates with iron (and cobalt) and silica close to the surface to form nickel-rich limonitic ores (called nickeliferous iron ores), or passes down through the upper weathering zones to precipitate with magnesium and silica as nickel-bearing hydrous silicates (which are collectively known as garnierite) to form nickel silicate ores. Beneath these ores lie altered and then fresh parent rock. (de Waal, 1971; Golightly, 1981)

Many primary sulphide deposits have been subjected to secondary alteration and oxidation. Oxidation is the major process above the water table and sulphides are oxidized to sulphates, oxides and hydrated oxides, especially limonite or, more rarely, haematite. Pocks which mainly comprise limonite and silica and which overlie sulphide ores are called gossans. Gossans typically show a cellular structure of goethite and silica that have precipitated along grain boundaries, cleavages or twin planes of the primary sulphides. Brightly coloured secondary base metal carbonates and sulphates, relict sulphides and native metals are also present. The textures and mineralogy of gossans reflect the textures and mineralogy of the original sulphide ores, especially if the primary ores were coarse grained. Some of the metals go into solution and migrate downwards to meet the water table. Here, conditions change from oxidizing to reducing and the metals reprecipitate as secondary sulphides. Commonly, this dramatically increases the grade of the ore, especially where the primary ore comprises disseminated sulphide. This zone is known as the supergene enrichment zone. Supergene enrichment occurs as a series of oxidation and hydration reactions acting upon the original sulphide species. The degree of oxidation of the more susceptible minerals often controls the subsequent oxidation and alteration of the less susceptible ones. The generalized sequence of susceptibility is: pyrrhotite > chalcopyrite > fine-grained pyrite > sphalerite > galena > coarse-grained pyrite. (Sillitoe and Clark, 1969; Blain and Andrew, 1977; Andrew 1980; Sangameshwar and Barnes, 1983; Nickel and Daniels, 1985)

These deposits range from Proterozoic to Tertiary in age and are hosted by shales, siliciclastics and carbonates. They are important ores of copper, lead, zinc, uranium and silver, and can be syngenetic, syndiagenetic or epigenetic in origin. Older deposits are often metamorphosed. (Sangster (ed.), 1983)

These stratiform deposits account for a significant proportion of the world's copper reserves. Two main copper provinces: the Upper Proterozoic Zambian Copper Belt and the Lower Permian Kupferschiefer of central and northwest Europe have been studied intensively. Both provinces have huge, essentially syngenetic/diagenetic ores found in shallow marine sediments associated with major transgressions. Anoxic conditions and bacterial reduction of sea water sulphate were important controls on mineralization. The ores have both lateral and vertical mineralogical zonation related to palaeogeographical conditions. In the Kupferschiefer, this zoning is copper- and silver-rich passing upward into lead- and zinc-rich ores, whereas in the Zambian Copper Belt, it is chalcocite passing basinwards into bornite, then chalcopyrite and finally pyrite. In unmetamorphosed examples the sulphides are fine-grained and their distribution reflects primary sedimentary features of the host sediments. Associated volcanicity is minor in extent or absent.

The main sulphides of the unmetamorphosed Kupferschiefer are fine-grained pyrite, chalcocite, galena, sphalerite, digenite, djurleite, bornite, chalcopyrite and covelline. Minor minerals include anilite, tennantite, luzonite, mooihoekite, haycockite together with trace amounts of cobaltite-gersdorffite, smaltite-chloanthite, clausthalite, molybdenite, native gold, native silver and platinum group minerals.

Some shallow marine clastics contain minor green-grey siltstone, light-coloured sandstone or carbonate-rich horizons within a dominantly purple-red sequence. The subordinate horizons are hosts to low grade copper-silver ores which are believed to be diagenetic in origin. Copper, iron and silver stripped from the main oxidized reddened sequence are reprecipitated as pyrite, followed by copper-iron and copper sulphides. Although discrete silver minerals are present in trace amounts most of the silver occurs within copper sulphides ('chalcocite'), including digenite, anilite, djurleite and chalcocite. Typically, the sulphides are finely disseminated or form veinlets parallel with the bedding of the host rocks. (Hayes and Einaudi, 1986; Lange et al., 1986, 1987)

There are two main classes of deposit: uranium deposits and red bed copper (lead and zinc) deposits. The ores are mined for uranium, copper, vanadium and minor silver, lead and zinc. Both classes are hosted by continental to very shallow marine sediments laid down under and conditions. The sequences are characteristically reddened by fine-grained haematite coatings (called 'pigment') around the clastic grains. The mineralization is conformable, often associated with primary sedimentary features and is concentrated in bleached or grey-green organic and pyrite-rich beds. The ores are diagenetic to epigenetic and, because of the porosity of the host sediments, are often characterized by a large number of brightly coloured secondary oxidation zone minerals.

Large scale deposits are found in the western states of the USA, mainly within Mesozoic fluviatile sandstones and siltstones. The deposits form flats, pods or the characteristic roll-fronts at the junctions between oxidized and reduced clastics. Uranium, in association with pyrite, is initially fixed in the clastics by organic material; on oxidation of these sediments (reddening), the uranium is oxidized and moves in solution until it encounters reducing conditions when it reprecipitates (often replacing plant material). Uraninite (as pitchblende), coffinite, uranium-rich organic material, carnotite and tyuyamunite, together with many secondary uranium and vanadium phases, are the main ore minerals. These are associated with lesser amounts of pyrite, marcasite, galena, sphalerite, chalcopyrite, bornite and chalcocite. (Fischer, 1968; Harshman, 1968; Kelley et al., 1968; Young, 1984)

Although some are lead-zinc-rich and economic, the majority are copper-rich, small and sub-economic. The ores are concentrated into bleached horizons within reddened sequences, where they form syndiagenetic cements around elastic silicate grains. The main sulphides include the blue copper sulphides (chalcocite), bornite, pyrite, galena, sphalerite and covlline, together with many secondary copper, lead and zinc carbonates and sulphates. The primary (and secondary) mineralization is controlled by the local porosity and permeability of the host rocks. (Holmes et al., 1983; Zielinski et al., 1983)

This class has similarities with the older exhalative shale-hosted lead-zinc deposits (Mount Isa type) and those of the Mississippi Valley type. The type examples are found in Lower Carboniferous shelf carbonate sequences of the Central Irish Plain. The carbonates lie transgressively over Devonian siliciclastics and mineralization is localized close to Caledonian-trending faults. The deposits occur in argillaceous to dolomitic carbonates and most have simple mineralogies but complex synsedimentary textures. The main sulphides are pyrite, sphalerite, galena and marcasite, and locally chalcocite and bornite accompanied by minor amounts of tetrahedrite-tennantite, boulangerite, bournonite and other sulphosalts. Mineralization is believed to be the result of fluid exhalations onto the seafloor, accompanied by replacement and void infilling of the underlying sediments. (Hitzman and Large, 1986; Russell, 1986)

This class of deposit is mined for lead, zinc, fluorite and baryte. These epigenetic deposits usually occur within Phanerozoic limestones and dolomites in sequences that are often associated with evaporates. Individual deposits may be small, but, collectively, they occur within large ore fields. Mineralization is concentrated along erosional surfaces, karstic surfaces or in faulted and brecciated carbonate, and comprises void infilling and metasomatic replacement ores. Associated wallrock alteration is restricted to dolomitization, silicification and the introduction of pyrite. Mineralization is polyphase. Three mineralogical subclasses are recognized, the most common is the zinc-rich, but lead-rich and fluorite-rich subclasses are also important. (Heyl, 1983; Kisvarsanyi et al. (eds), 1983; Hagni, 1986; Kyle and Price, 1986)

Both void infilling and metasomatic replacement deposits occur within Lower Carboniferous limestones, or more rarely dolomites, and have many of the characteristics of the fluorite-rich subclass of the Mississippi Valley deposits. Galena and sphalerite are accompanied by bravoite, pyrite, marcasite and chalcopyrite within a fluorite, baryte and calcite 'gangue'. (Dunham, 1983, Ixer, 1986).

Lower Carboniferous limestones and sandstones, that have been mineralized, lie unconformably upon granitic basement. The mineralization within the orefield is zoned: an inner higher temperature zone is characterized by a pyrrhotite, chalcopyrite and quartz association with minor lead and zinc, and is surrounded by galena-rich and sphalerite-rich outer zones. The minor ore mineralogy within all of the zones is complex, with nickel-cobalt-iron arsenides and sulpharsenides, bismuth, silver and antimony minerals and rare earth phosphates and fluorocarbonates. This mineralogy suggests a geochemical contribution from the underlying granite. (Dunham, 1948, 1983; Brown et al., 1987)

There are few examples of this class of deposit, but they are large and produce significant by-product silver. The deposits are Middle Proterozoic to Palaeozoic in age and are found in shallow water black shales, siltstones, cherts and evaporates or their metamorphosed equivalents. Basalts and tuffs form part of these sequences but do not appear to be related genetically to mineralization. In unmetamorphosed deposits the sulphides are very fine-grained and often contain much framboidal pyrite; with metamorphism the sulphides coarsen in grain size. Pyrite, sphalerite, galena, pyrrhotite and freibergite are the main ore minerals and, amongst the silver-carriers, freibergite, stephanite, pyrargyrite, argentopyrite and native silver are important. Basinal compaction with the expulsion of interstitial saline brines along faults is the most widely accepted model for the genesis of these deposits. (Love and Zimmerman, 1961; Stanton, 1963; Mathias and Clark, 1975; Finlow-Bates and Stumpfl, 1979; Perkins, 1984)