Abstract The Central African Copperbelt is the world’s premier sediment-hosted Cu province. It is contained in the Katangan basin, an intracratonic rift that records onset of growth at ~840 Ma and inversion at ~535 Ma. In the Copperbelt region, the basin has a crudely symmetrical form, with a central depocenter maximum containing ~11 km of strata positioned on the northern side of the border of the Democratic Republic of Congo and Zambia, and marginal condensed sequences <2 km in thickness. This fundamental extensional geometry was preserved through orogenesis, although complex configurations related to halokinesis are prevalent in central and northern parts of the basin, whereas to the south, relatively high-grade metamorphism occurred as a result of basement-involved thrusting and burial. The largest Cu ± Co ores, both stratiform and vein-controlled, are known from the periphery of the basin and transition to U-Ni-Co and Pb-Zn-Cu ores toward the depocenter maximum. Most ore types are positioned within a ~500-m halo to former near-basin-wide salt sheets or associated halokinetic structures, the exception being that located in extreme basin marginal positions, where primary salt was not deposited. Stratiform Cu ± Co ores occur at intrasalt (Congolese-type), subsalt (Zambian-type), and salt-marginal (Kamoa-type) positions. Bulk crush-leach fluid inclusion data from the first two of these deposit types reveal a principal association with residual evaporitic brines. A likely signature of the ore fluids, the brines were generated during deposition of the basin-wide salt-sheets and occupied voluminous sub and intrasalt aquifers from ~800 Ma. Associated intense Mg ± K metasomatism was restricted to these levels, indicating that capping and enclosing salt remained impermeable for prolonged periods of the basin’s history, isolating the deep-seated aquifers from the upper part of the basin fill. From ~765 to 740 Ma, the salt sheets in the Congolese part of the basin were halokinetically modified. Salt was withdrawn laterally to feed diapirs, ultimately leading to localized welding or breaching of the former hydrological seal. At these points, deeper-level residual brines were drawn into the intrasalt stratigraphy to interact with reducing elements and form the stratiform ores. It is probable that salt welding occurred diachronously across the northern and central parts of the basin, depending upon the interplay of original salt thickness, rates and volumes of sediment supply during accumulation of salt overburden, and tectonism. The variable timing of this fundamental change in hydrologic architecture is poorly constrained to the period of halokinetic onset to the earliest stages of orogenesis; however, the geometry of the ores and associated alteration patterns demands that mineralization preceded the characteristically complex fragmentation of the host strata. Thus, while an early orogenic timing is permissible, mineralization during the later stages of extensional basin development was more likely. In situ reducing elements that host Zambian-type stratiform Cu ± Co ores were in continuous hydrological communication with subsalt aquifers, such that ore formation could have commenced from the ~800 Ma brine introduction event. The nonhalokinetic character of the salt in this region allowed the intact seal to have maintained suprahydrostatic pore pressures, facilitating fluid circulation until late stages of basin growth and possibly early stage orogenesis. Leachate data from ores positioned in the depocenter maximum and southern parts of the basin that underwent relatively high grade metamorphism record mixing of residual and halite dissolution-related brines. Salt dissolution was likely triggered by emergence of diapirs or thermally and/or mechanically induced increased permeability of halite. While it is certain that halite dissolution occurred during and after orogenesis, conditions favorable for salt dissolution may have existed locally during extension in the depocenter maximum. The permeability of salt increased to a point where it became the principal aquifer. The salt’s properties as an aquiclude lost, originally deep-seated residual brine mixed with new phases of evaporite dissolution-related brine to produce ores at middle levels of the basin fill. During the final stages of ore formation, recorded by postorogenic Pb-Zn-Cu mineralization in the depocenter maximum, the salinity of fluids was dominantly derived from the dissolution of remnant bodies of salt.

Abstract The Kalahari Copperbelt in northwestern Botswana is characterized by structurally controlled, stratabound, mineralogically zoned copper-silver deposits hosted along a major redox boundary within a late Mesoproterozoic rift succession. Copper-silver mineralized rocks occur on the limbs and in the hinge positions of regional-scale folds that characterize the Pan-African Ghanzi-Chobe zone fold-and-thrust belt. Regional facies changes along the base of the transgressive marine D’Kar Formation, the host to the majority of mineralized rocks, delineate a series of synsedimentary basin highs and lows. The facies changes were identified through both lithostratigraphic analysis of drill holes and along-strike variations in magnetic lithostratigraphy, a technique that correlates the magnetic fabrics of second vertical derivative aeromagnetic maps with changes in lithostratigraphy. Basin highs controlled the development and distribution of favorable lithostratigraphic and lithogeochemical trap sites for later sulfide precipitation. Major facies changes across the Ghanzi Ridge area straddle a significant crustal structure identified in gravity datasets that appears to have influenced extensional activity during basin development. During basin inversion, the basin highs, cored by rheologically stronger bimodal volcanic rocks, localized strain within mechanically weaker rock types of the Ghanzi Group metasedimentary rocks, leading to the development of locally significant permeability and the formation of structural trap sites for mineralization by hot (250°–300°C), oxidizing, metalliferous Na-Ca-Cl brines. Structural permeability was maintained within trap sites due to silicification and/or feldspar alteration during progressive deformation and associated hydrothermal mineralizing events.

Abstract This special volume provides a comprehensive review of the current state of knowledge for rare earth and critical elements in ore deposits. The first six chapters are devoted to rare earth elements (REEs) because of the unprecedented interest in these elements during the past several years. The following eight chapters describe critical elements in a number of important ore deposit types. These chapters include a description of the deposit type, major deposits, critical element mineralogy and geochemistry, processes controlling ore-grade enrichment, and exploration guides. This volume represents an important contribution to our understanding of where, how, and why individual critical elements occur and should be of use to both geoscientists and public policy analysts. The term “critical minerals” was coined in a 2008 National Research Council report (National Research Council, 2008). Although the NRC report used the term “critical minerals,” its focus was primarily on individual chemical elements. The two factors used in the NRC report to rank criticality were (1) the degree to which a commodity is essential, and (2) the risk of supply disruption for the commodity. Technological advancements and changes in lifestyles have changed the criticality of elements; many that had few historic uses are now essential for our current lifestyles, green technologies, and military applications. The concept of element criticality is useful for evaluation of the fragility of commodity markets. This fragility is commonly due to a potential risk of supply disruption, which may be difficult to quantify because it can be affected by political, economic, geologic, geographic, and environmental variables. Identifying potential sources for some of the elements deemed critical can be challenging. Because many of these elements have had minor historic usage, exploration for them has been limited. Thus, as this volume highlights, the understanding of the occurrence and genesis of critical elements in various ore deposit models is much less well defined than for base and precious metals. A better understanding of the geologic and geochemical processes that lead to ore-grade enrichment of critical elements will aid in determining supply risk and was a driving factor for preparation of this volume. Understanding the gaps in our knowledge of the geology and geochemistry of critical elements should help focus future research priorities. Critical elements may be recovered either as primary commodities or as by-products from mining of other commodities. For example, nearly 90% of world production of niobium (Nb) is from the Araxá niobium mine (Brazil), whereas gallium (Ga) is recovered primarily as a by-product commodity of bauxite mining or as a by-product of zinc processing from a number of sources worldwide.

Abstract For nearly 50 years, carbonatites have been the primary source of niobium and rare earth elements (REEs), in particular the light REEs, including La, Ce, Pr, and Nd. Carbonatites are a relatively rare type of igneous rock composed of greater than 50 vol % primary carbonate minerals, primarily calcite and/or dolomite, and contain the highest concentrations of REEs of any igneous rocks. Although there are more than 500 known carbonatites in the world, currently only four are being mined for REEs: the Bayan Obo, Maoniuping, and Dalucao deposits in China, and the Mountain Pass deposit in California, United States. The carbonatite-derived laterite deposit at Mount Weld in Western Australia is also a REE producer. In addition to REEs, carbonatite-related deposits are the primary source of Nb, with the Araxá deposit, a carbonatite-derived laterite in Minas Gerais state, Brazil, being the dominant producer. Other commodities produced from carbonatite-related deposits include phosphates, iron, fluorite, copper, vanadium, titanium, uranium, and calcite. Types of ores include those formed as primary magmatic minerals, from late magmatic hydrothermal fluids, and by supergene enrichment in weathered horizons. Although the principal REE-bearing mineral phases include fluorocarbonates (bastnäsite, parisite, and synchysite), hydrated carbonates (ancylite), and phosphates (monazite and apatite), the dominant mineral exploited at most mines is bastnäsite. Bastnäsite typically is coarse grained and contains approximately 75 wt % RE 2 O 3 (rare earth oxides; REOs). Processes responsible for REE enrichment include fractional crystallization of the carbonatitic magma, enrichment of REEs in orthomagmatic or hydrothermal fluids and subsequent precipitation or subsolidus metasomatic redistribution of REEs, and breakdown of primary carbonatitic mineral phases by chemical weathering and sequestration of REEs in secondary minerals or in association with clays. Carbonatites are primarily associated with continental rifting, but some carbonatites are associated with orogenic activity. Although there is debate on how carbonatite magmas are generated, the parental magma and REEs are clearly derived from mantle sources.

Abstract Highly evolved alkaline/peralkaline igneous rocks host deposits of rare earth elements (REE), including Y as well as Zr, Hf, Nb, Ta, U, and Th. The host rocks spanning from silica-undersaturated (nepheline syenites) to silica-oversaturated (granites) occur in intraplate tectonic environments, mainly in continental settings and are typically associated with rifting, faulting, and/or crustal extension. They range in age from Neoarchean/Paleoproterozoic to Mesozoic, but several significant deposits are of Mesoproterozoic age. The deposits/prospects can be subdivided into three types. The first is hosted by nepheline syenitic rocks of large, layered alkaline intrusions where the mineralization commonly occurs in layers rich in REE-bearing minerals, which mostly show cumulate textures (e.g., Thor Lake/Nechalacho, Canada; Ilimaussaq, Greenland; Lovozero, Russia; Kipawa, Canada; Norra Kärr, Sweden; Pilanesberg, South Africa). The second type includes mineralization in peralkaline granitic rocks where REE-bearing minerals are usually disseminated. The mineralization is typically hosted by pegmatites (including the Nb-Y-F type), felsic dikes, and minor granitic intrusions (e.g., Strange Lake, Canada; Khaldzan-Buregtey, Mongolia; Ghurayyah, Saudi Arabia; Bokan, Alaska, United States). The third type is disseminated, very fine grained, and hosted by peralkaline felsic volcanic/volcaniclastic rocks, mostly of trachytic composition (e.g., Dubbo Zirconia and Brockman/Hastings, Australia). The bulk of the REE is present in ore/accessory minerals which in some mineralized zones, particularly in cumulate rocks from alkaline complexes, can reach >10 vol %. Mineralization is composed of a variety of REE-bearing minerals, which frequently show complex replacement textures. They include fluorocarbonates, phosphates, silicates, and oxides. Economically most important are bastnäsite, monazite, xenotime, loparite, eudialyte, synchysite, and parasite. Many other minerals are either sparse or it is difficult with present technology to profitably extract REE from them on a commercial scale. Compared to carbonatite-hosted REE deposits, the REE mineralization in alkaline/peralkaline complexes has lower light REE concentrations but has commonly higher contents of heavy REE and Y and shows a relative depletion of Eu. Elevated concentrations of U and Th in the ore assemblages make gamma-ray (radiometric) surveys an important exploration tool. The host peralkaline (granitic, trachytic, and nepheline syenitic) magmas undergo extensive fractional crystallization, which is protracted in part due to high contents of halogens and alkalis. The REE mineralization in these rocks is related to late stages of magma evolution and typically records two mineralization periods. The first mineralization period produces the primary magmatic ore assemblages, which are associated with the crystallization of fractionated peralkaline magma rich in rare metals. This assemblage is commonly overprinted during the second mineralization period by the late magmatic to hydrothermal fluids, which remobilize and enrich the original ore. The parent magmas are derived from a metasomatically enriched mantle-related lithospheric source by very low degrees of partial melting triggered probably by uplift (adiabatic) or mantle plume activity. The rare metal deposits/mineralization related to peralkaline igneous rocks represent one of the most economically important resources of heavy REE including Y. In addition to REE, some of these deposits contain economically valuable concentrations of other rare metals including Zr, Nb, Ta, Hf, Be, U, and Th, as well as phosphates.

Abstract Ion adsorption-type rare earth element (REE) deposits are the predominant source of heavy REE (HREEs) and yttrium in the world. Economic examples of the deposits are confined almost exclusively to areas underlain by granitic rocks in southern China. These deposits are termed “ion adsorption-type” because the weathered granites contain more than ~50% ion-exchangeable REY (REE + Y), relative to whole-rock REY. The ore grades range from 140 to 6,500 ppm (typically ~800 ppm) REY, and some of the deposits are remarkably enriched in HREEs. The Yanshanian (Jurassic-Cretaceous) granites that weather to form the deposits are products of subduction-related or extensional intraplate magmatism. These parent granites for the REE deposits are biotite- and/or muscovite-bearing granites and are characterized by >70% SiO 2 , <0.08% P 2 O 5 , and metaluminous to weakly peraluminous (ASI < 1.1) compositions. The highly differentiated (SiO2 >~75%) muscovite granites are HREE enriched relative to the biotite granites and are notably characterized by occurrences of fluorite and hydrothermal REE-bearing minerals, particularly REE fluorocarbonates that formed in a deuteric alteration event. Magmatic allanite and titanite are either altered to form hydrothermal REE-bearing minerals or almost completely broken down during weathering. The weatherable REE-bearing minerals, including fluorocarbonates, allanite, and titanite, are the source minerals for the ion adsorption ores. The HREE grades of the ion adsorption ores are strongly influenced by the relative abundances and weathering susceptibilities of these REE-bearing minerals in the parent granites. The presence of easily weathered HREE minerals in the underlying granites appears to be the primary control of the HREE-rich deposits, although solution and solid phase chemistry during development of the weathering profile may influence REE fractionation. Monazite, zircon, and xenotime are also present in the granites, but because they are more resistant to chemical weathering, they are typically not a source of REEs in the weathered materials. The REE-bearing minerals are decomposed by acidic soil water at shallow levels in the weathering profile, and the REE 3 + ions move downward in the profile. The REEs are complexed with humic substances, with carbonate and bicarbonate ions, or carried as REE 3+ ions in soil and ground water at a near-neutral pH of 5 to 9. The REE 3 + ions are removed from solution by adsorption onto or incorporated into secondary minerals. The removal from the aqueous phase is due to a pH increase, which results from either water-rock interaction or mixing with a higher pH ground water. The REEs commonly adsorb on the surfaces of kaolinite and halloysite, to form the ion adsorption ores, due to their abundances and points of zero charge. In addition, some REEs are immobilized in secondary minerals consisting mainly of REE-bearing phosphates (e.g., rhabdophane and florencite). In contrast to the other REEs that move downward in the weathering profile, Ce is less mobile and is incorporated into the Mn oxides and cerianite (CeO 2 ) as Ce 4 + under near-surface, oxidizing conditions. As a result, the weathering profile of the deposits can be divided into a REE-leached zone in the upper part of the profile, with a positive Ce anomaly, and a REE accumulation zone with the ion adsorption ores in the lower part of the profile that is characterized by a negative Ce anomaly. The thickness of the weathering profiles generally ranges from 6 to 10 m but can be as much as 30 m and rarely up to 60 m. The negative Ce anomaly in weathered granite terrane is thus a good exploration indicator for ion adsorption ores. A temperate or tropical climate, with moderate to high temperatures and precipitation rates, is essential for chemical weathering and ion adsorption REE ore formation. Low to moderate denudation, characteristic of such a climate in areas of low relief, are favorable for the preservation of thick weathering profiles with the REE orebodies.

Abstract Ancient and modern types of sedimentary placer deposits formed in both alluvial and coastal environments have been signficant sources of the rare earth elements (REEs). The REE-bearing minerals in placer-type deposits are primarily monazite [(Ce,La,Nd,Th)PO4] and sometimes xenotime (YPO4), which are high-density (heavy) minerals that accumulate with the suite of heavy minerals. Monazite has been extracted from many heavy mineral placers as a coproduct of the economic recovery of associated industrial minerals, such as titanium oxide minerals (ilmenite, rutile), zircon, sillimanite, garnet, staurolite, and others. Xenotime has been produced from some alluvial deposits as a coproduct of tin (cassiterite) placer mining. Placers are mineral deposits formed by the mechanical concentration of minerals from weathered debris. Placers can be classified as eluvial, alluvial, eolian, beach, and fossil (paleo) deposit types. Monazite-bearing placer-type deposits can occur in residual weathering zones, beaches, rivers and streams, dunes, and offshore areas. The detrital mixture of sand, silt, clays, and heavy (dense) minerals deposited in placers are derived primarily from the erosion of crystalline rocks, mainly igneous rocks and moderate- to high-grade metamorphic rocks (amphibolite facies and higher). In fluvial settings, slope is an important factor for the concentration of heavy minerals from detritus. In coastal settings, the actions of waves, currents, tides, and wind are forces that concentrate and sort mineral particles based on size and density. Placer deposits containing monazite are known on all continents. In the past, by-product monazite has been recovered from placers in Australia, Brazil, India, Malaysia, Thailand, China, New Zealand, Sri Lanka, Indonesia, Zaire, Korea, and the United States. More recently, monazite has been recovered from coastal and alluvial placers in India, Malaysia, Sri Lanka, Thailand, and Brazil. In particular, along the southwestern and southeastern coasts of India, beach deposits rich in heavy minerals have experienced renewed exploration and development, partly to recover monazite for its REEs as well as its Th, to be used as a nuclear fuel source. Exploration designed to locate heavy mineral placers in coastal environments should identify bedrock terranes containing abundant high-grade metamorphic rocks or igneous rocks and identify ancient or modern coastal plains sourced by streams and rivers that drain these terranes. Trace elements associated with heavy mineral placers, useful as pathfinder elements, primarily include Ti, Hf, the REEs, Th, and U. Radiometric methods of geophysical exploration are useful in discovering and delineating deposits of heavy mineral sands. Several minerals in these deposits can produce a radiometric anomaly, but especially monazite, due to its high thorium content. Some beach districts in India and Brazil have been demonstrated as areas of high background radiation with potential dose exposure to humans and others, primarily due to the Th and U in detrital grains of monazite and zircon. Monazite- or xenotime-bearing placers offer several advantages as sources of REEs. Ancient and modern deposits of heavy mineral sands that formed in coastal settings can be voluminous with individual deposits as much as about 1 km wide and more than 5 km long. Grains of monazite or xenotime in placer deposits are mingled with other heavy minerals of industrial value. Monazite and xenotime are durable and often the heaviest minerals within the sand-silt deposit, which makes them relatively easy to mechanically separate. Thus, the REE ore minerals, monazite or xenotime, can be recovered from heavy mineral placers as a low-cost coproduct along with the economic production of the associated industrial minerals.

Abstract Each year an estimated 56,000 metric tons (t) of rare earth elements (REEs), including 23,000 t of heavy REEs (HREEs), are mined, beneficiated, and put into solution, but not recovered, by operations associated with the global phosphate fertilizer industry. Importantly, the REEs in sedimentary phosphorites are nearly 100% extractable, using technologies currently employed to meet global phosphate fertilizer needs. Our evaluation suggests that by-product REE production from these phosphate mines could meet global REE requirements. For example, the calculated REE flux accompanying phosphate production in the United States is approximately 40% of the world’s total and, alone, could supply 65% of global HREEs needs. Moreover, recognition that the tonnages and HREE concentrations of some unmined phosphorite deposits dwarf the world’s richest REE deposits suggests that these deposits might constitute stand-alone REE deposits. The hypothesized genesis of these REE-rich occurrences strongly supports the long-debated suggestion that oceanic REE contents vary in a secular fashion and that associated high-grade REE abundances reflect oceanic redox state transitions during specific time periods. Here, we use this new process-based model, based on observed variations in global-secular REE abundances, to identify phosphorite horizons deposited during periods favorable for highgrade REE accumulation.

Abstract China is the world’s leading rare earth element (REE) producer and hosts a variety of deposit types. Carbonatite-related REE deposits, the most significant deposit type, include two giant deposits presently being mined in China, Bayan Obo and Maoniuping, the first and third largest deposits of this type in the world, respectively. The carbonatite-related deposits host the majority of China’s REE resource and are the primary supplier of the world’s light REE. The REE-bearing clay deposits, or ion adsorption-type deposits, are second in importance and are the main source in China for heavy REE resources. Other REE resources include those within monazite or xenotime placers, beach placers, alkaline granites, pegmatites, and hydrothermal veins, as well as some additional deposit types in which REE are recovered as by-products. Carbonatite-related REE deposits in China occur along craton margins, both in rifts (e.g., Bayan Obo) and in reactivated transpressional margins (e.g., Maoniuping). They comprise those along the northern, eastern, and southern margins of the North China block, and along the western margin of the Yangtze block. Major structural features along the craton margins provide first-order controls for REE-related Proterozoic to Cenozoic carbonatite alkaline complexes; these are emplaced in continental margin rifts or strike-slip faults. The ion adsorption-type REE deposits, mainly situated in the South China block, are genetically linked to the weathering of granite and, less commonly, volcanic rocks and lamprophyres. Indosinian (early Mesozoic) and Yanshanian (late Mesozoic) granites are the most important parent rocks for these REE deposits, although Caledonian (early Paleozoic) granites are also of local importance. The primary REE enrichment is hosted in various mineral phases in the igneous rocks and, during the weathering process, the REE are released and adsorbed by clay minerals in the weathering profile. Currently, these REE-rich clays are primarily mined from open-pit operations in southern China. The complex geologic evolution of China’s Precambrian blocks, particularly the long-term subduction of ocean crust below the North and South China blocks, enabled recycling of REE-rich pelagic sediments into mantle lithosphere. This resulted in the REE-enriched nature of the mantle below the Precambrian cratons, which were reactivated and thus essentially decratonized during various tectonic episodes throughout the Proterozoic and Phanerozoic. Deep fault zones within and along the edges of the blocks, including continental rifts and strike-slip faults, provided pathways for upwelling of mantle material.

Abstract Porphyry Cu and porphyry Mo deposits are large to giant deposits ranging up to >20 and 1.6 Gt of ore, respectively, that supply about 60 and 95% of the world’s copper and molybdenum, as well as significant amounts of gold and silver. These deposits form from hydrothermal systems that affect 10s to >100 km 3 of the upper crust and result in enormous mass redistribution and potential concentration of many elements. Several critical elements, including Re, Se, and Te, which lack primary ores, are concentrated locally in some porphyry Cu deposits, and despite their low average concentrations in Cu-Mo-Au ores (100s of ppb to a few ppm), about 80% of the Re and nearly all of the Se and Te produced by mining is from porphyry Cu deposits. Rhenium is concentrated in molybdenite, whose Re content varies from about 100 to 3,000 ppm in porphyry Cu deposits, ≤150 ppm in arc-related porphyry Mo deposits, and ≤35 ppm in alkali-feldspar rhyolite-granite (Climax-type) porphyry Mo deposits. Because of the relatively small size of porphyry Mo deposits compared to porphyry Cu deposits and the generally low Re contents of molybdenites in them, rhenium is not recovered from porphyry Mo deposits. The potential causes of the variation in Re content of molybdenites in porphyry deposits are numerous and complex, and this variation is likely the result of a combination of processes that may change between and within deposits. These processes range from variations in source and composition of parental magmas to physiochemical changes in the shallow hydrothermal environment. Because of the immense size of known and potential porphyry Cu resources, especially continental margin arc deposits, these deposits likely will provide most of the global supply of Re, Te, and Se for the foreseeable future. Although Pd and lesser Pt are recovered from some deposits, platinum group metals are not strongly enriched in porphyry Cu deposits and PGM resources contained in known porphyry deposits are small. Because there are much larger known PGM resources in deposits in which PGMs are the primary commodities, it is unlikely that porphyry deposits will become a major source of PGMs. Other critical commodities, such as In and Nb, may eventually be recovered from porphyry Cu and Mo deposits, but available data do not clearly define significant resources of these commodities in porphyry deposits. Although alkali-feldspar rhyolite-granite porphyry Mo deposits and their cogenetic intrusions are locally enriched in many rare metals (such as Li, Nb, Rb, Sn, Ta, and REEs) and minor amounts of REEs and Sn have been recovered from the Climax mine, these elements are generally found in uneconomic concentrations. As global demand increases for critical elements that are essential for the modern world, porphyry deposits will play an increasingly important role as suppliers of some of these metals. The affinity of these metals and the larger size and greater number of porphyry Cu deposits suggest that they will remain more significant than porphyry Mo deposits in supplying many of these critical metals.

Abstract Rare-element granitic pegmatites represent highly concentrated sources of rare metals, including Li, Rb, Cs, Be, Sn, Nb, Ta, Zr, Y, REE, and U. In today’s markets, pegmatites are the principal sources of Ta, and one pegmatite (Tanco, Canada) is the sole commercial producer of Cs for use as deep drilling fluid in the form of Cs formate solution. Growth in the demand for Li-based batteries has prompted exploration for spodumene-and petalite-bearing pegmatites, and several new Canadian prospects are slated for mining. Pegmatite bodies that contain minerals in which these elements are essential structural constituents constitute less than ~1 to 2% of all pegmatites in a given pegmatite-bearing terrane, and the economic production from many such bodies is limited by their small size (i.e., they may be economic in grade, but not for mechanized mining). Because pegmatites are found in cratons and orogenic belts, however, pegmatite-hosted resources are widespread and likely to be significant secondary, if not primary sources of rare metals for local economies or in times of disruption of global supplies from other types of deposits. Pegmatites are primarily igneous in origin, and the most likely processes that enrich them in rare elements include crystal-melt fractionation together with the creation of locally flux-and rare-element-enriched domains of melt in otherwise rather ordinary granitic melt. The mechanism of constitutional zone refining, in which fluxes and incompatible components are enriched in a boundary layer of melt adjacent to crystal growth fronts, represents the most effective means of concentrating rare elements. Whereas Rayleigh fractionation produces an exponential increase in the abundances of incompatible rare elements, constitutional zone refining leads to a sharp, “L”-shaped inflection in the concentration of incompatible elements with the progress of crystallization. The absolute concentrations of trace elements at the end of constitutional zone refining can be orders of magnitude greater that those that are attainable by Rayleigh fractionation (between mineral and bulk melt). In rare-element pegmatites, some trace-element enrichment patterns show the gradual increase in abundance that is expected of Rayleigh fractionation, whereas pegmatites in which the transition from ordinary mineral assemblages to those enriched in rare elements is sharply defined, more closely match the elemental fractionation that is derived from constitutional zone refining. Although pegmatites are igneous, pegmatite-forming melts crystallize well below their liquidus, and perhaps even below solidus temperatures. The textures and zonation that are hallmarks of pegmatites arise in response to the inception of crystallization from highly undercooled, viscous melt. Graphic granite, the one texture that is unique to pegmatites, constitutes prima facie evidence of such conditions. The crystallization of rare-element minerals, such as beryl, spodumene, tantalite, and pollucite can also be reconciled to the low-temperature crystallization of melts. Pegmatite-hosted ores are entirely endogenic, and many pegmatites exhibit little or no exogenic wall-rock alteration. Narrow and sporadic zones of alteration envelopes, unpredictable size in relationship to degree of fractionation, and sharply defined zonation of rare-element ores to inner units combine to make granitic pegmatites difficult targets for exploration.