©2005 Society of Economic Geologists, Inc. , v. 100, pp. 773–779

SULFIDE MELT INCLUSIONS AS EVIDENCE FOR THE EXISTENCE OF A PARTIAL MELT AT BROKEN HILL, AUSTRALIA

HEATHER A. SPARKS † AND JOHN A. MAVROGENES Research School of Earth Sciences, Australian National University, Canberra, ACT, 0200, Australia and Department of Earth and Marine Sciences, Australian National University, Canberra, ACT, 0200, Australia

Abstract Polyphase sulfide melt inclusions are hosted within garnetite rocks and quartz veins in garnetite surround- ing droppers and large masses of the orebody at Broken Hill, Australia, and record the presence of a former sulfide melt. Sulfide melt inclusions are either primary or occur along healed fractures in both garnets and quartz veins. Common daughter minerals in the inclusions are , , arsenopyrite, , tetrahedrite-tennantite, and minor amounts of argentite, , dyscrasite (Ag3Sb), and gudmundite (FeSbS). The inclusions exhibit a strong enrichment in low-melting-point chalcophile elements compared to the main orebody. Experimental reequilibration of sulfide melt inclusions shows homogenous melt at temperatures as low as 720° ± 10°C and 5 kbars, well below that of peak metamorphism at Broken Hill (800° ± 10°C and 5 kbars). Thus, these inclusions are interpreted to represent a trapped sulfide melt formed during peak meta- morphism at Broken Hill, Australia.

Introduction enriched in low-melting-point chalcophile elements to the Broken Hill, New South Wales, Australia, is the world’s point where remobilized may form discrete high-grade largest Pb-Zn-Ag deposits. Despite over a century of study, pockets. Based on these experimental and empirical observa- aspects of remain elusive (for review see Stevens, tions, Mavrogenes et al. (2001) and Frost et al. (2002) con- 1975). White et al. (1995) postulated that emplacement of the cluded that at least some of the Pb-Zn-Ag ore at Broken Hill orebody occurred after peak metamorphism. Others must have melted. (Gustafson and Williams, 1981; Phillips et al., 1985; Stevens Hofmann (1994) and Hofmann and Knill (1996) unam- et al., 1988) suggested that the orebody is premetamorphic biguously established that the of Lengenbach, Switzer- and syngenetic in origin (i.e., exhalative). Broken Hill is land, partially melted during metamorphism. They described hosted within a suite of complexly folded and metamor- polyphase sulfide melt inclusions trapped in quartz. These phosed Proterozoic metasediments and metavolcanic rocks sulfide melt inclusions exhibit strong enrichment in low-melt- (Stevens et al., 1988) and reached peak metamorphic condi- ing-point chalcophile elements, including Pb, Tl, As, Sb and tions of at least 800°C and 5 kbars at 1600 ± 5 Ma (Phillips Bi, and fully homogenize at reasonable temperatures and Wall, 1981; Page and Laing, 1992; Cartwright, 1999). (<500°C). They proposed that sulfide melts formed during The effects of metamorphism on the Broken Hill orebody metamorphism were trapped as sulfide melt inclusions. The are not entirely clear, although some studies (Brett and Challenger Au mine in South Australia was shown to be a Kullerud, 1966, 1967; Lawrence, 1967) have suggested that metamorphosed Au deposit (Tomkins and Mavrogenes, 2002) syngenetic Pb-Zn-Ag ores may have melted during peak by the recognition of polyphase melt inclusions in peak meta- metamorphism. Partial melting of silicates is well docu- morphic mineral assemblages. Frost et al. (2002) established mented from textural and chemical criteria (Phillips, 1980; that the massive sulfide ores of Snow Lake, Manitoba. melted Phillips and Wall, 1981). In contrast to silicate melts, sulfide during metamorphism. More recently, partial melting has melts quench to complex intergrowths of sulfide minerals been used to explain the distribution of sulfide and sulfosalt that tend to reequilibrate at very low temperatures (Frost et mineral at Hemlo, Ontario (Tomkins et al., 2004). Until now, al., 2002). As a result, textures of sulfide melts are rarely however, no direct evidence for the existence of a sulfide melt preserved. at Broken Hill has been documented. Recent experimental work (Mavrogenes et al., 2001) has Materials Studied demonstrated that eutectic melting in the system PbS- Sulfide melts can migrate and concentrate into pockets, Fe0.96S-ZnS-(1% Ag2S) begins at 795°C at 5 kbars. This tem- perature is well within independently derived estimates for known as “droppers,” and these were the focus of the present peak metamorphic conditions at Broken Hill. Frost et al. study. Droppers were first identified and described as sulfide (2002) has shown that the addition of low-melting-point chal- dikes by King and O’Driscoll (1953), Mackenzie (1968) and cophile elements (Ag, As, Au, Bi, Hg, Sb, Se, Sn, Tl, Te, Cu, Maiden (1975, 1976). They are interpreted as piercement Pb, Fe, Mn) depresses the onset of sulfide partial melting. structures of remobilized ore extending out from the main orebodies (1–50 m) into the country rocks. They typically Wykes and Mavrogenes (2005) show that the addition of H2O lowers sulfide eutectics. Frost et al. (2002) also suggest that crosscut foliations and igneous rocks (Fig. 1). Droppers are with progressive melting a polymetallic melt will become surrounded by an alteration package of rocks primarily com- prised of garnetite. Garnetite is composed of ~95 to 98 per- † Corresponding author: e-mail, [email protected] cent (by volume) equidimensional (~100 µm), orange-brown

0361-0128/05/3521/773-7 $6.00 773 774 SCIENTIFIC COMMUNICATIONS

wherever possible. As with previous experimental studies (Mavrogenes et al., 2001) the presence of quenched molten sulfide (in experimentally reequilibrated samples) was indi- cated by characteristic myrmekitic intergrowth of (Fig. 2A). Where these textures were present, numerous area scans were performed over the entire exposed sulfide melt in- clusion to acquire an average composition. High-pressure homogenization experiments were per- formed using a 12.7 mm end-loaded piston cylinder appara- tus located at the Research School of Earth Sciences, Aus- tralian National University (RSES, ANU). A series of experiments at 620°, 720°, and 840°C, and 5 kbars was per- formed. For the 620°C run, a volume of garnet separates was loaded into a drilled-out MgO rod, and heated for 4 hours. For the 720° and 840°C runs, approximately 0.16 g of garnet grains and one quartz sample (two chips, each ~2 mm diam) were loaded into -palladium capsules (3 garnets, one quartz) and welded shut. All four capsules were loaded into machined MgO rod and run simultaneously. The experiment durations were four hours for the 720 °C run and one hour for 840 °C. Runs were quenched by cutting the power to the ap- paratus. For further details regarding sample assembly and run procedures see Hermann and Green (2001). Inclusions were analyzed using laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) at the RSES. Si, S, Ca, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Sr, Mo, Ag, Cd, In, Sn, Sb, Ba, Re, Os, Pt, Te, Au, Hg, Tl, Pb and Bi were investigated. Laser output energy was set at 100 mJ with a repetition rate of 10 Hz. Relatively slow ablation rates smoothed out the signal, allowing qualitative estimates of the compositions of the analyzed phases. To minimize overlap be- FIG. 1. Cross section through the Broken Hill orebody (coordinate 1101S) tween isotopes, two isotopes of the same element were mea- showing 1 lens (Zn lode) overlying 2 lens (Pb lode) from which a dropper ex- sured simultaneously. Analyses for major and trace element tends into the country rock. This clearly illustrates the remobilized nature of the dropper (after Mackenzie, 1968). Drill hole N4689 passed through this concentrations of entire unexposed sulfide melt inclusions dropper at 26–51.2 m. were acquired in real time, ensuring maximum control of the ablation procedure (Fig. 3). Each acquisition started with a collection of carrier gas for approximately 30 to 40 s (carrier gas in Fig. 3). The laser was then turned on (On in Fig. 3) ini- tially ablating pure host (garnet host in Fig. 3). Upon inter- spessartine garnets. Interstitial material in the garnetite com- section of a sulfide melt inclusion the signals include material prises minor galena and rare quartz veins. We sampled gar- ablated from both host and inclusion, with the contribution netite and one quartz vein in four drill holes (3028, N4689, from each component evolving as the ablation pit deepens Z3031, and C144) from Perilya Ltd.’s Broken Hill mine. (sulfide melt inclusion = SMINC in Fig. 3). Once the entire sulfide melt inclusion was ablated, pure host was again inter- Methodology sected and the analysis was stopped (Off in Fig. 3). Spot sizes Garnet grains were separated from the garnetite and were selected based on the diameter of individual SMINCs. mounted in epoxy approximately one garnet layer deep, and Data was acquired in blocks of 10 with external standards an- polished to expose sulfide melt inclusions. Quartz chips were alyzed at the beginning and end of each block. mounted separately and carefully polished until sulfide melt A similar method to that described by Halter et al. (2004) inclusions were exposed along healed fractures. All samples was employed to calculate sulfide melt inclusion composi- were then inspected using reflected light microscopy and tions obtained by LA-ICP-MS. First, the background count imaging with a scanning electron microprobe (SEM), using rate determined from ablation of the host (garnet) was sub- backscattered electrons. The compositions of the sulfide in- tracted from the sulfide melt inclusion to give background clusions were estimated using a JEOL 6400 SEM equipped corrected count rates. If inclusions were close to the surface, with an Oxford Link ISIS energy dispersive (EDS) detector representative host signal from the same sample was used. located at the Research School of Biological Sciences, Aus- Second, element ratios were calculated by referencing to tralian National University (ANU). A 15-kV accelerating volt- NIST 610, which was analyzed as the external standard. Hal- age, a 1-nA beam current, and 120-s counting times were ter et al. (2004) assumed their sulfide melt inclusions to be employed. Bulk compositions were obtained from a ~10-µm stoichiometric (Fe, Cu)S, and element concentrations were area scan. Individual phases were analyzed by spot analysis normalized accordingly by assuming 50 mol percent S. This

0361-0128/98/000/000-00 $6.00 774 SCIENTIFIC COMMUNICATIONS 775

FIG. 2. Broken Hill sulfide melt inclusions. A. Backscattered electron image of a fully homogenized melt inclusion quenched from 840°C and 5 kbars. Note the myrmekitic texture typical of rapidly quenched sulfide melts. B. Reflected light photomicrograph of polished garnetite, showing the distribution of inclusions. Note the abundance of randomly distributed melt inclusions within garnet grains, and the monomineralic character of interstitial sulfides. C. Transmitted light photomi- crograph of a sulfide inclusion trail in a healed fracture in quartz. Within one inclusion array, all inclusions are composition- ally similar. D. Backscattered electron image of a negative crystal-shaped polyphase melt inclusion containing 8 daughter phases (as labelled). This inclusion graphically illustrates the high levels of Ag (tetrahedrite and dyscrasite), As (arsenopyrite) and Sb (tetrahedrite and gudmundite) present in Broken Hill melt inclusions. E. Reflected light photomicrograph of a rounded polyphase melt inclusion containing four daughter phases (as labelled). F. Backscattered electron image of a par- tially homogenized melt inclusion quenched from 620°C and 5 kbars. Note that roughly 10 percent of the inclusion shows quenched melt textures.

0361-0128/98/000/000-00 $6.00 775 776 SCIENTIFIC COMMUNICATIONS

FIG. 3. Ablation depth (time) vs. Mn, Cu, Zn, As, Ag, Sn, Sb, and Pb counts for an individual garnet-hosted sulfide melt inclusion (M337, sample Z3331-33.8). This profile first ablates through garnet, then a complete (unhomogenized) sulfide melt inclusion, followed by garnet. Note that individual daughter phases can be recognized in the ablation profile. Initially, galena (Pb) is ablated, followed by tetrahedrite (Sb, As, and Ag), and finally chalcopyrite (Cu). In fully homogenized inclu- sion profiles, no individual phases are seen. This inclusion contains approximately 50 wt percent Pb, 4 wt percent As, 1 wt percent Zn, 1 wt percent Cu, and 800 ppm Ag.

method is inappropriate for the current study as the bulk within one single sulfide melt inclusion (e.g., Fig. 2D). compositions of the inclusions are not stoichiometric metal- Many of the phases have euhedral to subhedral morphology, sulfide compositions, as demonstrated by EDS analyses of suggesting slow cooling and crystallization from a homoge- quenched melt from homogenized inclusions. Therefore, el- nous melt during post-entrapment cooling. Common ement concentrations were normalized to 100 percent (S + daughter minerals are galena, sphalerite, arsenopyrite, chal- Mn + Fe), where S, Fe and Mn values were collected by copyrite, tetrahedrite-tennantite, and minor amounts of ar- EDS, an approach that also accounts for the host element gentite, bornite, dyscrasite (Ag3Sb), and gudmundite contribution. (FeSbS). Other inclusions are monomineralic (e.g., ar- A total of 134 individual ICP-MS analyses were processed senopyrite and lollingite) and exhibit their own crystal habit. and plotted. To ensure analysis of entire inclusions, only sub- In all garnetite samples studied, all interstitial sulfides were merged (unexposed) inclusions were analyzed. Analyses are monomineralic galena or pyrrhotite (Fig. 2B). None of the considered representative of the melt only if their composi- low-melting-point chalcophile element-rich phases recog- tion was polyphase in nature and not monomineralic. Mineral nized as daughter minerals in sulfide melt inclusions (e.g., inclusions were easily identified as such and rejected. tetrahedrite, gudmundite, argentite, or dyscrasite) have been found in garnet interstices. Results Partial homogenization of sulfide melt inclusions was rec- The presence of sulfide melt inclusions in all garnetite ognized at 620ºC and 5 kbars (Fig. 2F), and total homoge- samples studied and in one quartz vein sampled within the nization was observed at 720ºC and 5 kbars (Fig. 2A). EDS garnetite supports the existence of a former sulfide melt. analyses of quenched melt from 30 garnet-hosted and 7 Sulfide melt inclusions are randomly distributed inside gar- quartz-hosted inclusions from both 1 atm (not reported; see net grains (Fig. 2B) or along planes in healed fractures in Sparks, 2003) and high-pressure heating experiments in garnet or quartz (Fig. 2C). Extensive petrographic study re- which complete homogenization occurred are compiled in vealed that sulfide melt inclusions generally show the nega- Table 1. Individual sulfide melt inclusions within a single pop- tive crystal shape of their host (Fig. 2D) or are spherical ulation (e.g., a single healed fracture) homogenized at the (Fig. 2E), with several daughter minerals present inside. At same temperature, further suggesting that these inclusions least eight discrete daughter minerals have been observed represent trapped melt.

0361-0128/98/000/000-00 $6.00 776 SCIENTIFIC COMMUNICATIONS 777

TABLE 1. Concentrations of S, Fe and Mn (wt %) in Homogenized Sulfide PbS-FeS-ZnS-Ag2S would result in an increase in Ag concen- Melt Inclusions (used in data reduction from both garnet and quartz hosts) tration in the melt, whereas in the present study Ag increased Drill hole S Mn Fe Total (Fig. 4B) with increasing Cu/Pb. Previous work suggests that Ag increased in the melt until tetrahedrite saturation oc- Garnet host curred. This is supported by our observation that tetrahedrite 3028-60.8 18.30 1.56 10.45 30.31 is the major Ag host in sulfide melt inclusions. Thus, Ag be- 3028-60.8 19.76 1.19 11.35 32.30 haved incompatibly during galena fractionation. 3028-60.8 21.04 1.01 12.73 34.78 Garnet-hosted sulfide melt inclusions reveal the same cor- 3028-60.8 21.02 1.87 11.64 34.53 relation between Cu and Pb as those hosted by quartz (Fig. 3028-60.8 17.71 1.62 7.74 27.07 3028-60.8 20.66 1.49 10.91 33.06 4A), and different samples record a similar phenomenon. For 3028-60.8 19.05 1.05 10.94 31.04 example, sample N4689-51.2 (solid circles; Fig. 4C) and sam- 3028-60.8 17.23 2.92 10.19 30.34 ple Z3031-107.8 (open triangle; Fig. 4C) plot at the opposite 3028-60.8 21.04 1.37 10.12 32.53 ends of the fractionation trend shown by Cu/Pb. In contrast 3028-60.8 19.99 2.19 11.82 34.00 3028-60.8 20.62 0.94 11.09 32.65 to Ag, Co behaves compatibly during fractionation. This is il- 3028-60.8 19.06 1.86 13.09 34.01 lustrated by sample Z3031-107.8 (open triangle; Fig. 4D), 3028-60.8 17.30 1.59 10.85 29.74 which has a much higher Co content than sample N4689-51.2 3028-60.8 20.58 1.41 12.20 34.19 (solid circles; Fig. 4D) and may have been trapped earlier. 3028-60.8 22.30 1.31 14.49 38.10 The similar incompatible behavior of other elements may 3028-60.8 18.90 1.35 11.15 31.40 3028-60.8 20.18 2.58 10.52 33.28 explain why droppers are enriched in Ag, Sn and Sb. Extreme 3028-60.8 24.28 1.33 18.29 43.90 sulfide melt fractionation may eventually to melts that 3028-60.8 21.43 1.71 12.40 35.54 are very rich in these elements. This might also explain the N4689-51.2 21.79 1.89 11.81 35.49 anecdotal correlation at Broken Hill between high Au grades N4689-51.2 21.54 1.89 12.33 35.76 N4689-51.2 21.26 1.67 11.21 34.14 and garnetite. N4689-51.2 21.18 1.06 10.78 33.02 Average Ag grades at Broken Hill are hard to determine N4689-51.2 20.64 0.74 13.25 34.63 owing to the variable nature of the lodes, but the average N4689-51.2 20.95 1.00 13.11 35.06 composition of the entire deposit has been estimated to be N4689-51.2 18.53 1.93 12.37 32.83 148 ppm Ag (Parr and Plimer, 1993), and grades higher than N4689-51.2 19.72 2.29 10.34 32.35 Z3031-107.8 22.16 0.59 16.97 39.72 400 ppm are not reported. Our estimates of sulfide melt in- Z3031-107.8 22.69 0.23 16.45 39.37 clusion bulk compositions yield extreme enrichment of Ag, as Z3031-107.8 19.02 1.14 13.02 33.18 well Pb and Cu, compared to the Broken Hill main lodes Z3031-107.8 19.00 0.72 9.82 29.54 (Table 2 and Table 3). The Ag/Pb ratio of sulfide melt inclu- Z3031-107.8 21.50 0.55 12.91 34.96 Z3031-107.8 21.03 0.72 13.86 35.61 sions is one order of a magnitude higher than the average ore grades of the main lodes (Table 2). It is likely that a sulfide Quartz host melt formed during the waning stages of metamorphism at C144 22.27 0.00 11.18 33.45 Broken Hill, based on the textural evidence in the form of sul- C144 22.78 0.28 13.26 36.32 fide melt inclusions, and melt likely persisted to temperatures C144 22.78 0.53 13.94 37.25 as low as 720°C, and potentially even lower. C144 23.25 0.05 15.52 38.82 C144 22.41 0.27 12.02 34.70 Discussion and Conclusions C144 22.49 0.00 11.97 34.46 C144 21.71 0.00 9.12 30.83 The presence of sulfide melt inclusions within garnets as- C144 22.28 0.00 13.42 35.70 sociated with the ores of Broken Hill conclusively establishes Average garnet 20.35 1.42 12.13 33.89 that the ore partially melted during metamorphism. That σ 1.60 0.60 2.13 4.33 Average quartz 22.50 0.14 12.55 35.19 these inclusions homogenize at temperatures reasonable for σ 6.28 0.43 3.53 10.24

Determined by EDS TABLE 2. Comparision of Average Compositions of Sulfide Melt Inclusions and Ore Grade

Pb (wt %) Zn (wt %) Ag (ppm) Ag (ppm)/Pb (wt%) The compositions of 44 quartz-hosted and 45 garnet-hosted SMINC 51.1 0.8 6,005 118 sulfide melt inclusions are plotted in Figure 4. Only S, Ca, Mn, Fe, Co, Ni, Cu, Zn, As, Ag, Sn, Sb, Au, Hg, Pb and Bi Orebody1 were present above background. The negative correlation be- No. 3 lens 7.8 11.9 169 21.7 tween Cu and Pb shown in Figure 4A (closed symbols) is No. 2 lens 16.4 12.4 118 7.2 likely due to fractionation of chalcopyrite and galena which No. 1 lens 9.6 22.4 53 5.5 are two major phases revealed by reflected light microscopy A lode 4.3 10.4 31 7.2 B lode 4.3 12.4 33 7.7 and SEM analyses of exposed sulfide melt inclusions in these C lode 3.2 6.4 34 10.6 samples. However, we had no prior knowledge of the behav- ior of these phases during sulfide crystal fractionation. Mavro- SMINC = sulfide melt inclusion genes et al. (2001) suggested that fractionation in the system 1 From Haydon and McConachy, 1987

0361-0128/98/000/000-00 $6.00 777 778 SCIENTIFIC COMMUNICATIONS

FIG. 4. Chemical trends of LA-ICP-MS analyses of individual sulfide melt inclusions. Each point represents an individual inclusion. A. Pb vs. Cu in sulfide inclusions in garnet (open squares) and quartz (solid squares) form a linear array that must have resulted from a chemical process such as fractionation. B. Ag vs. Cu/Pb in inclusions in quartz showing Ag increasing with Cu/Pb due to the incompatibility of Ag during galena fractionation. C. Pb vs. Cu in inclusions in garnet from four sep- arate garnetite samples. Note that these four separate samples define a linear trend toward the high Pb, low Cu end of the spectrum. Inclusions in sample N4689-51.2 (filled circles) plot at the extreme other end of the trend. Thus, the different samples appear to have trapped melt at different stages of the melting history. D. Co vs. Cu/Pb in inclusions in garnet from four separate garnetite samples, again define a chemical trend with Co behaving compatibly during fractionation. Note that the most evolved sample from C (N4689-51.2; filled circles) plots at the lowest Co (highest Cu/Pb) end, whereas the least fractionated sample from C (Z3031-107.8; open triangles) plots at the highest Co, (lowest Cu/Pb) end of the trend.

TABLE 3. Average Sulfide Melt Inclusion Compositions Determined by LA-ICP-MS

Pb (wt%) Cu (wt%) Zn (wt%) As (ppm) Ag (ppm) Sb (ppm) Sn (ppm) Co (ppm) Ni (ppm) Au (ppb)

Garnet host 54.6 7.2 1.3 9,559 7,078 6,685 217.5 45.5 18.1 4 Quartz host 48.2 14.7 0.3 2,250 5,125 7,378 188.6 36.1 256.8 4 All inclusions 51.1 11.3 0.8 5,542 6,005 7,066 201.6 40.3 149.3 4

0361-0128/98/000/000-00 $6.00 778 SCIENTIFIC COMMUNICATIONS 779 peak metamorphic conditions at Broken Hill and display sys- Johnson, I.R., and Klingner, G.D., 1975, Broken Hill ore deposit and its en- tematic chemical trends, establishes that they are trapped vironment, in Knight, C.L., ed., Economic Geology of Australia and Papua New Guinea: Melbourne, Australasian Institute of Mining and Metallurgy, melts. Furthermore, their enrichment in metals well estab- v. 1, p. 476–492. lished as constituents of the Broken Hill lodes suggests that King, H.F., and O’Driscoll, E.S., 1953, The Broken Hill Lode, in Edwards, sulfide melt inclusions formed by the melting of ores. A.B., ed., Geology of Australian Ore Deposits: 5th Empire Mining and Met- The abundance of sulfide melt inclusions in garnets asso- allurgical Congress, p. 578–600. Lawrence, L.J., 1967, Sulphide neomagmas and highly metamorphosed sul- ciated with droppers confirms the previous suggestions that phide deposits: Mineralium Deposita, v. 2, p. 5–10. droppers represent solidified sulfide dikes. Dropper ores as Mackenzie, D., 1968, Lead lode at New Broken Hill Consolidated Limited: well as associated sulfide melt inclusions are enriched in Cu, Australasian Institute of Mining and Metallurgy Monograph, Series 3, p. Sb, As, Ag, Ni and Au compared to the main lodes, consistent 161–169. with their derivation from the main lodes. Maiden, K.J., 1975, High grade metamorphic structures in the Broken Hill orebody: Australasian Institute of Mining and Metallurgy, Proceedings, v. Acknowledgments 254, p. 19–27. ——1976, Piercement structures formed by metamorphic mobilization in Heather Sparks received financial support through a stu- the Broken Hill orebody: Australasian Institute of Mining and Metallurgy dent research grant from the SEG Foundation. John Mavro- Proceedings, v. 257, p. 1–9. Mavrogenes, J.A., MacIntosh, I.W., and Ellis, D.J., 2001, Partial melting of genes received support from the Australian Research Coun- the Broken Hill galena-sphalerite ore: Experimental studies in the system cil. John Ridley in particular is thanked for his constructive PbS-FeS-ZnS-(Ag2S): ECONOMIC GEOLOGY, v. 96, p. 205–210. criticism of an earlier version of this manuscript. We also Page, R.W., and Laing, W.P., 1992, Felsic metavolcanic rocks related to the thank Perilya Broken Hill Ltd for access to drill core, in par- Broken Hill Pb-Zn-Ag orebody, Australia: Geology, depositional age, and ticular, Ian Groves, Jane Murray, and Noel Carol. Mike Shelly timing of high-grade metamorphism: ECONOMIC GEOLOGY, v. 87, p. 2138–2168. and Charlotte Allan assisted with LA-ICP-MS analyses. Dis- Parr, J.M., and Plimer, I.R., 1993, Models for Broken Hill-type lead--sil- cussions with Joerg Hermann, Richard Arculus, Ron Frost, ver deposits, in Kirkham, R.V., Sinclair, W.D., Thorpe, R.I. and Duke, J.M., Carl Spandler, Jeremy Wykes, and Chris McFarlane were eds., Mineral deposits modeling: Geological Association of Canada: Special helpful. Finally, John Vickers’ help with sample preparation Paper v. 40, p. 245–288. Phillips, G.N., 1980, Water activity changes across an amphibolites granulite was invaluable. facies transition, Broken Hill, Australia: Contributions to and Petrology, v. 75, p. 377–386. REFERENCES Phillips, G.N., and Wall, V.J., 1981, Evaluations of prograde regional meta- Brett, R., and Kullerud, G., 1966, Melting relations of galena-- morphic conditions: Their implications for the heat sources and water ac- pyrrhotite assemblages: A homogenous sulphide melt at 718°C: ECONOMIC tivity during metamorphism in the Willyama Complex, Broken Hill, Aus- GEOLOGY, v. 61, p. 1302. tralia: Bulletin Mineralogic, v. 104, p. 801–810. ——1967, The Fe-Pb-S system: ECONOMIC GEOLOGY, v. 62, p. 354–369. Phillips, G.N., Archibald, N.L. and Wall, V.J., 1985, Metamorphosed high-Fe Cartwright, I., 1999, Regional oxygen isotope zonation at Broken Hill, New tholeiites: Their alteration and relationship to sulphide mineralization, Bro- South Wales, Australia: Large scale fluid flow and implications for Pb-Zn- ken Hill, Australia: Geological Society of South Africa, Transactions, v. 88, Ag mineralization: ECONOMIC GEOLOGY, v. 94, p. 357–374. p. 49–59. Frost, B.R., Mavrogenes, J.A., and Tomkins, A.G., 2002, Partial melting of Sparks, H.A., 2003, Partial melting at Broken Hill: Evidence from sulfide sulfide ore deposits during medium- and high-grade metamorphism: melt inclusion: Unpublished B.Sc. Honours thesis, Canberra, Australian Canadian Mineralogist, v. 40, p. 1–18. National University, Department of Geology, 81 p. Frost, B.R., Swapp, S.M., and Heine, T.H., 2002, Partial melting of massive Stevens, B.P.J., 1975, Broken Hill type lead-zinc-silver deposits, in Markham, sulfide ore bodies in the Snow Lake district, Manitoba [abs.]: Geological N.L. and Basden, H., eds., The mineral deposits of New South Wales: Association of Canada; Mineralogical Association of Canada, Joint Annual N.S.W. Geologic Survey, Sydney, p. 45–60. Meeting Program with Abstracts, v. 27, p. 37. Stevens, B.P.J., Barnes, R.G., Brown, R.E., Stroud, W.J. and Willis, I.L., Gustafson, L.B., and Williams, N., 1981, Sediment-hosted stratiform de- 1988, The Willyama supergroup in the Broken Hill and Euriowie blocks, posits of , lead and zinc: ECONOMIC GEOLOGY 75TH ANNIVERSARY New South Wales: Precambrian Research, v. 40, p. 297–327. VOLUME, p. 139–178. Tomkins, A.G., and Mavrogenes, J.A., 2002, Mobilization of gold as a poly- Halter, W.E., Pettke, T. and Heinrich, C.A., 2004, Laser-ablation ICP-MS metallic melt during pelite anatexis at the Challenger deposit, South Aus- analysis of silicate and sulfide melt inclusions in an andesitic complex I: An- tralia: A metamorphosed Archean gold deposit: ECONOMIC GEOLOGY, v. 97, alytical approach and data evaluation: Contributions to Mineral Petrology, p. 1249–1271. v. 147, p. 385–396. Tomkins, A.G., Pattison, D.R.M., and Zaleski, E., 2004, The Hemlo gold de- Haydon, R.C., and McConachy, G.W., 1987, The stratigraphic setting of Pb- posit, Ontario: An example of melting and mobilization of a precious metal- Zn-Ag mineralization at Broken Hill: ECONOMIC GEOLOGY, v. 82, p. sulfosalt assemblage during amphibolite facies metamorphism and defor- 826–856. mation: ECONOMIC GEOLOGY, v. 99, p. 1063–1084. Hermann, J., and Green, D.H., 2001, Experimental constraints on high pres- White, S.H., Rothery, E., Lips, A.L.W., and Barklay, T.J.R., 1995, Broken Hill sure melting in subducted crust: Earth and Planetary Science Letters, v. area Australia, as a Proterozoic fold and thrust belt: Implications for the 188, p. 149–168. Broken Hill base metal deposit: Transactions of the Institute of Mining and Hofmann, B.A., 1994, Formation of a sulfide melt during Alpine metamor- Metallurgy, Section B, Applied Earth Sciences, v. 194, p. 1–17. phism of the Lengenbach polymetallic sulfide mineralisation, Binntal, Wykes, J.L. and Mavrogenes, J.A., 2005, Hydrous sulfide melting: Experi- Switzerland: Mineralium Deposita, v. 29, p. 439–442. mental evidence for the solubility of H2O in sulfide melts: ECONOMIC GE- Hofmann, B.A. and Knill, M.D., 1996, Geochemistry and genesis of the OLOGY, v. 100, p. 157–164. Lengenbach Pb-Zn-As-Tl-Ba-mineralisation, Binn Valley, Switzerland: Mineralium Deposita, v. 31, p. 319–339.

0361-0128/98/000/000-00 $6.00 779 0361-0128/98/000/000-00 $6.00 780