Abstract Skarn consists of coarse-grained Ca-Fe-Mg-Mn silicates formed by replacement of carbonate- bearing rocks accompanying regional or contact metamorphism and metasomatism. The major processes which result in skarn include metamorphic recrystallization of impure carbonate rocks, bimetasomatic reaction between unlike lithologies, and infiltrational metasomatism involving hydrothermal fluids of magmatic origin. Metal deposits that contain skarn as gangue, termed skarn deposits, may be formed by any combination of the above processes. However, the majority of the world’s major skarn deposits are thought to be related to magmatic-hydro-thermal systems; these are the skarn deposits treated in this paper. Skarn deposits are among the most abundant and variable of all types of mineral deposits; yet, in spite of their diversity, they exhibit systematic geologic, petrologic, and mineralogic features which permit classification and detailed investigation. The most useful classification of skarn is based on the dominant calc-silicate mineral assemblages. Thus, skarn that replaces dolomite largely consists of magnesian silicates such as forsterite and serpentine and is termed magnesian skarn. Skarn that replaces limestone largely consists of Fe-Ca silicates such as andradite and hedenbergite and is termed calcic skarn. Skarn deposits, on the other hand, are best classified on the basis of the dominant economic metal; six major subclasses, consisting of Fe, W, Cu, Zn-Pb, Mo, and Sn, are discussed in this paper. Although all subclasses can occur in either magnesian or calcic skarn, magnesian skarn deposits of W, Cu, and Zn-Pb are notably sparse. Variations within the subclasses of skarn deposits are recognized as a function of magma type, depth of emplacement, reducing capacity and composition of host rocks, distance of carbonate horizon from the magmatic source, and degree of meteoric water involvement. These factors can combine to yield continuous transitions between some classes. The majority of skarn deposits are of Mesozoic or younger age. The few important Paleozoic examples are W and Sn skarns which as a group may represent relatively deep environments of formation. All skarn types are abundant in Mesozoic time, but Cu and Zn-Pb skarn deposits, which in most cases represent a relatively shallow environment, are dominantly of Tertiary age. Most likely, this age distribution reflects differences in level of erosion rather than an evolution of ore-forming processes. Calcic magnetite skarn deposits virtually are the only skarn type found in oceanic island-arc terrains. They are widespread in the Urals, Philippines, and coastal British Columbia. Characteristic features include: (1) their association with epizonal diorite stocks emplaced in cogenetic basalt-andesite; (2) an Fe-rich calc-silicate gangue consisting of epidote-grandite-ferrosalite with retrograde chlorite-actinolite, which reflects intermediate oxidation states; (3) extensive epidote-pyroxene or albite-scapolite alteration of plutonic and volcanic rocks; (4) a low sulfide content; and (5) a minor metal suite of Cu, Zn, Co, and Au. Minor sulfides reflect low sulfidation states and include pyrrhotite, arsenopyrite, chalcopyrite, and sphalerite; cobaltite and bornite commonly occur in trace amounts. In contrast, magnesian magnetite skarn deposits are common in continental margin orogenic belts associated wih mesozonal to epizonal felsic plutons; they are found in most base metal sulfide skarn districts where dolomite is present. In this case, the high magnetite content is a function not of the igneous rock association but, rather, of the dolomitic wall rocks, in which Fe-rich calc-silicates are not stable. Inner diopside-spinel and outer forsterite-calcite zones of the early, high-temperature stage are overprinted at lower temperatures by humite, borates, magnetite, phlogopite, and serpentine. Minor sulfides refleet intermediate sulfidation states and include pyrrhotite, pyrite, calcopyrite, and sphalerite. Deposits of this type commonly display transitions to Cu skarns. Tungsten skarns and base metal (Cu, Zn-Pb, Mo) sulfide skarns are most characteristic of continental margin orogenic belts and are commonly thought to be related to subduction-related I-type magmas. Tungsten skarn deposits typically are associated with coarse-grained granodiorite to quartz monzonite stocks and batholiths emplaced in eugeoclinal limestone-shale ± volcanic sequences. Contact metamorphic and reaction skarn assemblages are overprinted by stratiform metasomatic calcic skarn, consisting of garnet-pyroxene (± scheelite) and outer wollastonite-idocrase zones in marble, and pyroxene-plagioclase-epidote in plutons and pelitic hornfelses. Zones of hydrous silicates, particularly biotite and hornblende with accessory quartz, feldspar and calcite, crosscut early skarn patterns and commonly contain abundant scheelite and sulfides. Mineral compositions in calcic W skarns, a general function of depth of skarn formation and host-rock composition, can be expressed as a continuum between two end members: reduced types formed in carbonaceous host rocks and/or at greater depths, and oxidized types formed in noncarbonaceous or hematitic host rocks and/or at lesser depths. Diagnostic minerals of the reduced types reflect low oxidation states (e.g., hedenbergitic pyroxene, almandine-rich garnet, Fe-rich biotite and hornblende, magnetite) and low sulfidation states (e.g., pyrrhotite, rare pyrite, traces of native bismuth). Diagnostic minerals of the oxidized types reflect intermediate oxidation states (e.g., salitic pyroxene, andraditic garnet, epidote) and intermediate sulfidation states (pyrite, minor pyrrhotite, traces of bismuthinite). In general, the geology and mineralogy of W skarns is radically different from that of Cu and Zn-Pb skarns; these differences point to W skarn formation at a relatively higher temperature and deeper environment than the base metal sulfide types. The majority of Cu skarn deposits are associated with epizonal granodiorite and quartz monzonite stocks in continental crust; relatively few occurrences are known from oceanic island-arc settings associated with quartz diorite and granodiorite plutons. As a group, calcic Cu skarn deposits are characterized by an association with felsic porphyry-textured stocks of hypabyssal character, proximity to stock contacts, high garnet to pyroxene ratios, relatively oxidized assemblages (e.g., andraditic garnet with diopsidic pyroxene, magnetite with hematite), and moderate to high contents of sulfide minerals of intermediate sulfidation state (e.g., pyrite-chal-copyrite, minor tennantite, sphalerite). Within this class, a continuum may exist between: (1) smaller Cu skarns associated with unaltered barren stocks and displaying relatively minor retrograde alteration and (2) larger Cu skarns associated with altered and mineralized porphyry copper stocks and commonly displaying more ferric-rich garnet, extensive retrograde alteration, and greater sulfide content. Some skarns mined for Cu display a more complex metal suite, including W, Mo, Bi, Zn, and Au and less-oxidized mineral assemblages (e.g., grandite with ferrosalite) than those discussed above; these deposits may represent transitions either to oxidized W skarns or to some Mo-bearing skarns. Calcic Zn-Pb skarn deposits form in the middle to late orogenic stages of continental margin belts and are associated with granodioritic to granitic magmatism. These skarns are characterized by their occurrence along structural or lithologic contacts at some distance from plutonic contacts, high pyroxene to garnet ratios, distinctive Mn- and Fe-rich minerals (e.g., early johannsenitic pyroxene, minor andraditic garnet, and late bustamite, rhodonite, dannemorite, and ilvaite), and the association of significant amounts of sulfides (e.g., sphalerite, galena, pyrite, pyrrhotite) with pyroxene rather than with garnet or other silicate minerals. Variations within this class may be related to distance from causative plutons; proximal Zn-Pb skarns are less Mn-rich, contain more sulfides in skarn than in limestone replacement ore, and display higher garnet to pyroxene ratios and lower Pb to Cu ratios than do distal skarns. Distal Pb-Zn skarn deposits commonly contain the bulk of ore in carbonate gangue beyond the skarn zone and may be linked with certain manto and vein deposits of Pb-Zn-Ag. An important factor in the formation of Zn-Pb skarn deposits is the travel distance of hydrothermal fluids between source and reactive limestone, which results in depletion of fluids in Mg, Al, and Cu and relative enrichment in Mn, Fe, Zn, and Pb. Tin skarns are associated with ilmenite-series granites of both I- and S-type emplaced late in the orogenic cycle of continental magmatic arcs or in relatively stable or incipiently rifted cratonic environments. The granites commonly contain greisen alteration associated with lithophile element deposits. Magnesian Sn skarns display an evolutionary sequence involving: (1) an early skarn stage of spinel, pyroxene, and forsterite; (2) an intermediate tin-borate stage of phlogopite, magnetite, and tin-bearing Fe-Mg borates; and (3) a late cassiterite stage of cassiterite, fluoborite, magnetite, and micas. The late stage is commonly accompanied by deposition of minor amounts of sulfides of low sulfidation state, including arsenopyrite, pyrrhotite, galena, and sphalerite. In some localities, the intermediate stage displays a calcic skarn overprint of magnetite, idocrase, and tin-bearing andradite. Calcic Sn skarns show evolutionary trends similar to those described above, in that Sn is not deposited as cassiterite until the system evolves to relatively low temperature and acidic conditions. The early stage in calcic skarns forms Sn-bearing andradite, wollastonite, and malayaite (e.g., Japan, West Malaysia), or idocrase-magnetite-fluorite, hedenbergitic pyroxene, and spessartine-bearing grandite (e.g., Tasmania, Alaska). During later stages, Sn is released by alteration of andradite-malayaite to cassiterite, calcite, quartz, and fluorite, and by alteration of idocrase-pyroxene-grandite to fluorite, amphibole, phlogopite, tourmaline, and magnetite. In both types of Sn skarns, the amount of cassiterite, and hence the recoverable grade of Sn, is directly related to the degree of retrograde alteration. High-grade Sn deposits in massive sulfide replacement bodies in dolomite (e.g., Renison Bell, Tasmania) may represent the low-temperature distal analogue of magnesian Sn skarns. The descriptive base outlined above indicates that broad correlations exist between the metal content of skarns and their igneous rock association and tectonic setting: the more mafic igneous rock types of oceanic island-arc settings produce Fe-rich (magnetite) skarns with significant Cu, Co, and Au contents; the intermediate to silicic calc-alkaline magmas of continental margins produce W skarns and minor Zn skarns in the mesabyssal environment and Fe, Cu, Mo, Pb, and Zn skarns in the hypabyssal environment; the more evolved granitic magmas of late- or postorogenic continental environments produce Sn, W, Mo, Zn, Be, and F skarns. These associations are compelling evidence for the dominantly igneous source of metals in skarn deposits. Further links exist between bulk compositions of skarn, compositions of garnets and pyroxenes, and amount of sulfides and their sulfidation state. In a general sense, skarn deposits can be classed on the basis of calc-silicate and iron oxide associations on a scale toward increasing oxidation state, with some W and Sn skarns at the reduced end and some Fe and Cu skarns at the oxidized end. Correlation between this scale and the sulfidation state of associated sulfides is suggested by the trend from pyrrhotite, native bismuth, and arsenopyrite in the more reduced skarns to large amounts of pyrite in the more oxidized skarns. Less direct are correlations between the oxidation-sulfidation state of skarn deoposits and the combination of factors involving depth, reducing capacity of host rocks, and intrinsic oxidation state of magmas. In general, the reduced and low-sulfur end of the scale correlates with the more reduced S-type or ilmenite-series magmas and with I-type or magnetite series magmas of mesa-byssal environments, whereas the oxidized and high-sulfur deposits correlate with the more oxidized I-type magmas of hypabyssal environments. Such correlations remain an important area for future research and must be tied to an increased understanding of the behavior of fluorine, chlorine, sulfur, and metals during late-stage magmatic processes. The major unifying feature of skarn deposits is their evolutionary style. Underlying the variation in metal content, magma association, tectonic setting, and mineralogy described above is a common pattern consisting of (1) essentially isochemical contact metamorphism accompanying emplacement of magma; (2) metasomatic skarn formation and initial ore deposition accompanying crystallization of the magma, initial cooling of the pluton, and evolution of an ore fluid; and (3) retrograde alteration and continued ore deposition accompanying the final cooling of the system. Mineral zoning patterns of each successive stage commonly crosscut earlier patterns as a consequence of shifting hydrothermal conduits during structural evolution. Metasomatic minerals commonly occur as overgrowths on, or veinlets in, metamorphic minerals, and these in turn may break down to polymineralic mixtures during retrograde alteration. The degree of development of any given stage varies widely between classes. Thus, the metamorphic stage is more intense in mesozonal skarns located at pluton contacts (e.g., W skarns) than in epizonal skarns located at some distance from plutons (e.g., distal Zn-Pb skarns). On the other hand, the retrograde stage is more intense in epizonal skarns located at stock contacts (e.g., porphyry Cu skarns) than in epizonal distal skarns (e.g., Zn-Pb skarn) or in mesozonal skarns (e.g., W skarns). Detailed field and petrographic-analytic studies, combined with fluid inclusion and stable isotope studies, yield estimates of P-T-X conditions during skarn evolution. Initial skarn formation occurs between 650° and 400°; higher temperatures are more characterisitc of deeper occurrences (1 to 3 kb) than of shallower occurrences (0.3 to 1 kb). The metasomatic fluid is characterized by low CO 2 content (X CO2 less than 0.1) and moderate salinities (10 to 45 percent NaCl equivalent). Boiling appears to be more characteristic of shallower environments, but the number of studies is limited. The source of sulfur is generally ascribed to magmatic or deep-seated rather than local sources, and the origin of H 2 O varies from magmatic during the early stages to magmatic + meteoric in the late retrograde stages of some deposits. Prograde skarn zoning patterns are interpreted as the results of infiltration metasomatism, with diffusion in intergranular fluids playing a minor role. Diffusion models are commonly used to explain zonal patterns in terms of component activity gradients. However, such models are strictly applicable only to zoned envelopes on single-stage veins and lack general applicability to the complex mineral patterns of large skarn deposits. Accurate estimates of gradients in solution temperature and composition must await a more complete experimental and theoretical data base on thermodynamic properties of complex solid solution minerals. Sulfide deposition generally takes place after the main period of skarn growth, as a consequence of declining temperature, local oxidation-reduction reactions implied by the preferential association of sulfides with specific calc-silicate zones, or neutralization of the fluid at the marble contact. A limiting factor in the quantitative interpretation of sulfide deposition in skarns is the lack of experimental studies of sulfide solubility in systems buffered by common skarn calc-silicates.

Abstract Textural and mineralogical studies of the Bald Mountain Cu-Zn-Au-Ag massive sulfide deposit, northern Maine, document a well-preserved premetamorphic hydrothermal evolution involving both exhalative and subsea-floor replacement processes. The 30-million metric tonne (Mt) Bald Mountain deposit forms a thick bowl-shaped accumulation of sulfides up to 215 m thick within a synvolcanic sea-floor graben of Early Ordovician age. Five principal stages (facies) of mineralization are recognized. Stage I mainly developed Fe sulfide mounds composed of fine-grained pyrite (As- and Sb-rich) and probably marcasite, with locally abundant sphalerite, sparse galena, and silica. Framboidal pyrite and colloform pyrite ± sphalerite ± galena are present locally, near the stratigraphic top of the deposit. During late stage I mineralization partial collapse of the sulfide mounds took place, probably due to dissolution of matrix anhydrite, producing thin to very thick (up to 20 m) accumulations of pyrite-quartz breccias. Following this mound collapse, stage I resumed with exhalative mineralization that filled the graben to its rim. Related mineralization formed volumetrically minor replacements of rhyolite ignimbrite. Stage I massive sulfides have a geochemical signature marked by generally high contents of Zn, Pb, As, Sb, Ag, Au, Hg, and Tl. This mineralization was succeeded, mostly within the graben structure, by the precipitation of low-temperature deposits of silica and Fe oxyhydroxides (now hematitic chert) that cover stage I deposits to depths of up to 28 m, as a hydrothermal cap to the massive sulfides below. Stage II formed mainly pyrrhotite-chalcopyrite replacements of stage I sulfides in the deep subsurface at temperatures of ca. 340° to 400°C based on arsenopyrite geothermometry. Stage II developed mainly after precipitation of the exhalative ferruginous silica cap, which may have sealed in the system thermally and chemically against shallow seawater entrainment. In addition to high Cu, stage II deposits contain abundant Co and Se. During this and subsequent stages of mineralization, older stage I sulfides underwent extensive recrystallization and zone refining at ∼250° to 325°C, accompanied by the replacement of pyrite by sphalerite ± galena, formation of euhedral quartz, arsenopyrite, and pyrite ± electrum, and remobilization of galena into small veins. Geometric relationships involving mineral assemblages and whole-rock (massive sulfide) geochemical data, together with textural information, suggest that Zn, Pb, As, Sb, Hg, and Tl in stage I deposits were dissolved and transported both upward and laterally by the zone refining for at least 150 m, resulting in very low contents of these elements within underlying stage II deposits. Gold was also remobilized and locally concentrated by the zone refining. Stage III deposits consist of wavy quartz ± chalcopyrite veins and replacements, reflecting their emplacement into unlithified massive sulfide mounds. Coeval to younger stage IV mineralization produced a complex assemblage of coarse pyrite with major amounts of chalcopyrite, magnetite, and greenalite; siderite, quartz, minnesotaite, ferropyrosmalite, and sphalerite are generally minor. Like stage II deposits, those of stage IV are significantly enriched in Co and Se, relative to stage I deposits. Stage IV also developed by subsea-floor zone refining, superimposed on older stages as veins and replacements, preferentially in the lower part of the deposit. Epigenetic hematization and silicification of fine-grained hanging-wall sediments (now argillites) between andesite flows, and of overlying fine-grained rhyolite ignimbrites, may have occurred when stage IV fluids breeched the ferruginous silica cap. Later, stage V mineralization formed siderite-rich veins with variable amounts of quartz, pyrite, marcasite, pyrrhotite, sphalerite, greenalite, magnetite, hematite, and calcite, both in the massive sulfide body and the stringer zone. Occurrences of very Fe rich silicates in stages II, IV, and V, and of siderite in stages IV and V, contrast with the absence of Mg-bearing silicates and carbonates throughout the deposit (excluding the footwall stringer zone). These compositions record the involvement of end-member Fe-rich hydrothermal fluids during formation of late sulfide veins and replacements, without appreciable amounts of shallow entrained (unreacted) seawater. Stage IV and V deposits precipitated from CO 2 -rich fluids based on their abundant siderite gangue. Stage IV formed mainly by the oxidation of stage II pyrrhotite, producing assemblages of pyrite ± magnetite (and rare magnetite without pyrite) together with locally prominent Fe 3+ -bearing greenalite; this relatively high f O2 environment continued during stage V mineralization, forming greenalite and hematite. Oxidation may have resulted from fluid boiling due to breaching of the ferruginous silica cap and consequent lowering of pressure in the hydrothermal system.

Abstract The Kidd Creek mine is an Archean volcanogenic Cu-Zn deposit with total past production and current reserves of more than 138.5 Mt at 2.4 percent Cu, 6.5 percent Zn, 0.23 percent Pb, 90 g/t Ag, and up to 0.15 percent Sn. The massive sulfides occur at the top of a locally thickened felsic volcanic pile, within and overlying a succession of massive rhyolite flows, volcaniclastic rocks, and coarse epiclastic units. The felsic volcanics occupy the core of an anomalous, S-shaped fold structure and attain a maximum thickness of approximately 300 m beneath the deposit. Massive autobrecciated rhyolite occurs at the base of the mine sequence and is interpreted to be a proximal vent facies. The local volcanic basement comprises mainly ultramafic flows, intercalated with minor rhyolite. The ultramafic rocks are interpreted to be early extrusive lavas associated with the development of an extensional rift. Basaltic pillow lavas and breccias occur in the hanging wall of the mine and are extensively intruded by gabbroic sills. South of the mine, this stratigraphy is truncated along the contact with younger, regional metasedi-mentary rocks. Kidd Creek is typical of a class of large volcanogenic massive sulfide deposits that occur within thick successions of permeable felsic volcaniclastic rocks and are dominated by large, stratiform, Zn-rich lenses with laterally extensive zones of ore-grade Cu mineralization. The deposit consists of three main ore-bodies (the North, Central, and South orebodies) that are distributed along an inferred boundary fault of a linear, grabenlike depression. The present deposits have a restored strike length of at least 2 km, indicating remarkable continuity of the hydrothermal system along the length of the graben. The main ore lenses formed by infilling and strata-bound replacement of volcaniclastic rocks, coarse volcanic breccias, and finer grained tuffs that filled the graben. Abundant relics of silicified rhyolite within the massive sulfides, gradational contacts between the massive sulfides and unmineralized fragmental rocks at the margins of the ore zones, and extensive replacement within the hanging-wall breccias confirm that a large part of the deposit formed below the sea floor. Burial of the deposits by mass flows was coincident with mineralization, and subsea-floor deposition of sulfides progressed laterally into the volcaniclastic rocks adjacent to the ore lenses. Metalliferous sediments or exhalative horizons are notably absent, and there is little evidence that widespread venting of high-temperature fluids occurred at the sea floor. Deposition of sulfides within the thick sequence of basin fill ensured that ore-forming fluids were confined to the graben and relatively little metal was lost to high-temperature discharge. The development of the three main orebodies is best explained by a long-lived, low-temperature hydrothermal system punctuated by several higher temperature pulses of Cu-rich fluid. Focusing of the fluids was caused by intense silicification of the rhyolite above and adjacent to the main upflow zone. Extensive lateral flow occurred within the bedded volcaniclastic rocks, and the highest temperature fluids appear to have occupied a number of high-level aquifers beneath the deposits. These are marked by conformable lenses of chlorite alteration, semimassive chalcopyrite, and strata-bound chalcopyrite stringer mineralization. The larger alteration envelope is broadly conformable to the ore lenses and consists of quartz and sericite, together with chlorite, Fe-rich carbonate, and minor tourmaline. Two main ore suites occur at Kidd Creek: a low-temperature, polymetallic suite enriched in Zn, Ag, Pb, Cd, Sn, Sb, As, Hg, ±Tl, ± W, and a higher temperature suite of Cu, Co, Bi, Se, In, ± Ni. The massive ores consist mainly of pyrite, pyrrhotite, sphalerite, and chalcopyrite, together with galena, tetrahedrite, ar-senopyrite, and cassiterite, in a quartz and siderite gangue. However, more than 60 different ore minerals and ore-related gangue minerals are present, including complex assemblages of Co-As sulfides, Cu-Sn sulfides, Ag minerals, and selenides. Tin is present as cassiterite in the upper part of the massive sphalerite lenses and as stannite in the underlying chalcopyrite-rich ores. Despite the high Ag content of the deposit, Kidd Creek is remarkably Au poor. The ores exhibit a close chemical affinity with their immediate felsic host rocks, including strong coenrichments of Ag, Pb, As, Sn, W, and F However, the complex metal assemblage suggests that a more primitive mafic suite may also have played a role in metal supply. The extensive metagraywackes to the south of the mine are younger than Kidd Creek and therefore could not have been a source for metals. An abundance of pyrrhotite, arsenopyrite, high Fe sphalerite, and Fe-rich chlorite indicates predominantly low fO 2 –fS2 conditions, and the abundant siderite in the ore indicates that the hydrothermal fluids were highly enriched in CO 2 . Sulfur isotope compositions range from -2.4 to +3.3 per mil, with the bulk of the massive sulfides having S 34 S values close to 0 per mil. The mineralogy and bulk composition of the Kidd Creek ores bear a closer resemblance to those of many Phanerozoic Zn-Cu-Pb deposits (e.g., Bathurst, Neves Corvo) than to other Archean Cu-Zn deposits. The predominance of Zn-rich ores (ca. 70–80 Mt) implies that most of the deposit formed at low temperatures (ca. 250°C). Solubility modeling indicates that a large hydrothermal system at relatively low temperatures would have been sufficient to account for about 75 percent of the metals. The significant enrichments in Ag, Pb, and Sn reflect not only the abundance of felsic volcanic rocks in the mine sequence but also the sustained, low-temperature venting history of the deposit. In contrast, the Cu-rich ores appear to have been introduced during relatively short-lived, hydrothermal pulses at much higher temperatures. The higher temperatures most likely coincided with discrete felsic magmatic events that occurred at several intervals during the ∼3.5 m.y. history of the volcanic complex. The late-stage introduction of Cu may indicate that the Cu-rich fluids evolved separately from the lower temperature, con-vective part of the hydrothermal system. This model is supported by the presence of a high-grade bornite zone in the South orebody, which represents a massive influx of Cu metal at peak hydrothermal temperatures late in the development of the Cu stringer zone. Kidd Creek resembles sulfide deposits that are currently forming in young, intraoceanic back-arc rifts, such as the Lau basin, and this may be an appropriate modern analogue for the Kidd Creek setting. The combination of voluminous mafic-ultramafic flows in the footwall of the deposit, punctuated by anomalous felsic volcanism, and the extensive deposits of coarse epiclastic rocks and volcaniclastic sediments suggest that Kidd Creek formed within a subsiding rift basin. The importance of a plumelike source for the ultramafic melts and the longevity of the hydrothermal system may indicate that rifting occurred above a stationary hot spot.