Lake-Sediment Geochemistry Reveals 1400 Years of Evolving Extractive Metallurgy at Cerro De Pasco, Peruvian Andes

Lake-sediment geochemistry reveals 1400 years of evolving extractive metallurgy at Cerro de Pasco, Peruvian Andes

Colin A. Cooke, Alexander P. Wolfe, and William O. Hobbs* Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, Alberta T6G 2E3, Canada

ABSTRACT The geochemical record preserved in lake sediments is a potentially powerful tool in A 80°W 75°W 70°W archaeometallurgy. Here, sediments from Llamacocha, a small lake in the central Peruvian Andes, are used to reconstruct an ~1400 year legacy of metal extraction from Cerro de Pasco, once the largest silver mine in the world. The earliest evidence for anthropogenic lead (Pb) enrichment occurs ca. A.D. 600 and is confi rmed by Pb stable isotope ratios that 5°S PERU match those of Cerro de Pasco ores. Early Pb pollution is attributed to precolonial smelting km for silver production, which relied on galena-based fl uxes. Following colonial control of the 0 200 400 mine ca. A.D. 1600, the switch to mercury (Hg) amalgamation for winning silver resulted in Cerro de atmospheric Hg emissions, as registered in Llamacocha sediments. Both Pb and Hg deposi- Pasco 10°S tion increased through the twentieth century, attaining peak values in A.D. 1968 and 1942, Study region respectively. Principal components analysis (PCA) identifi es a gradient that differentiates Lima anthropogenic from natural metals within the record, confi rming that early smelting led to Wari the volatilization of trace metals associated with local ore mineralogy. These results represent the fi rst evidence for a major precolonial mining industry at Cerro de Pasco, provide a Cuzco 15°S chronological framework for evolving extractive technologies, and are the fi rst to document widespread Hg pollution associated with colonial Hg amalgamation.

INTRODUCTION losses associated with amalgamation, suggest 76°W Lake sediments offer the potential to pre- that New World colonial Hg emissions totaled B Cerro de Pasco serve interpretable records of preindustrial 196,000 t, averaging ~612 t a−1 (Nriagu, 1993, Hg amalgamation remains metallurgical activities. In Europe, lake-sed- 1994). However, the very existence of colonial N (Chipian Valley) iment geochemistry has been used to address Hg pollution remains equivocal because natu- archaeological questions of metallurgical evo- ral archives collected outside the Andes, but 20 km lution (Bränvall et al., 2001; Renberg et al., still in the Southern Hemisphere, show no evi- 1994), while similar efforts are just beginning dence for preindustrial Hg pollution (Biester et 11°S Laguna Junín for the Andes (Abbott and Wolfe, 2003; Cooke al., 2002; Lamborg et al., 2002). The Andean et al., 2008). Our knowledge of the timing region itself has not been investigated system- Junín Llamacocha and evolution of Andean metallurgy remains atically. Thus, lake-sediment archives recov- incomplete because of extensive looting, and ered proximal to major amalgamation centers C 1 km N 4200 over 400 years of mine-site degradation asso- afford a unique opportunity not only to shed Stream Lake ciated with colonial and industrial mineral light on the evolution of New World metal-

extraction. Lake-sediment archives may poten- lurgy, but also to address long-standing ques- 4300 tially contribute not only to our understanding tions regarding the preindustrial global cycle of the environmental legacy of archaeometal- of Hg. Here, we utilize a lake sediment core

4400 lurgical activities, but also toward identifying to reconstruct the history of mining and metal- 11°4’S patterns associated with the technological evo- lurgy around Cerro de Pasco, once the world’s lution of Andean metallurgy. largest silver mine. Cerro de Pasco in the central Peruvian Andes contains one the most intensively exploited sil- STUDY SITE Llamacocha 75°48’W 4200 ver deposits in the world. Although colonial Cerro de Pasco is located on the Junín plain silver mining in the region is well known from in central Peru (Fig. 1). Following colonial Figure 1. A: Location map of study region within Peru. Shaded area indicates zone historical accounts (e.g., Fisher, 1977), preco- discovery ca. A.D. 1630, it quickly became of Cu-Pb-Zn-Ag polymetallic ores, which lonial exploitation patterns remain cryptic at one of the world’s foremost producers of sil- broadly corresponds to crest of Andean Cor- best. Large-scale mercury (Hg) mining began ver (Fisher, 1977). Cerro de Pasco, and the dillera. B: Location of study lake in relation in the late sixteenth century with the develop- numerous nearby smaller mines surrounding to Cerro de Pasco. C: Detail of study lake ment of Hg amalgamation, a process that facil- it, exploit rich polymetallic (Pb-Zn-Cu-Ag) and surrounding area. itated the extraction of silver from low-grade ore deposits, the primary minerals of which ores and stimulated large-scale silver produc- are enargite (Cu3AsS4), arsenopyrite (FeAsS), tion in the New World. Historical records of aramayoite (Ag[Sb, Bi]S2), argentiferous 1977; Ward, 1961). Orebodies at Cerro de Hg production and sale, coupled to known galena ([Ag, Pb]S), argentiferous tennan- Pasco are estimated to have contained over 2 4 6 6 tite ([Ag, Cu, Fe]12[Sb, As]4S13), grantonite × 10 t Ag, 2 × 10 t Pb, and 4 × 10 t Zn,

*Current address: Department of Geosciences, (Pb9As4S15), sphalerite ([Zn, Fe]S), bismuthi- with lesser amounts of gold (Au) and bismuth University of Nebraska–Lincoln, Lincoln, Nebraska nite (Bi2S3), and native silver (Ag) (Einaudi, (Peterson, 1965). 68588-0340, USA.

© 2009 Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or [email protected]. GEOLOGYGeology, No, vNoembervember 2009; 2009 v. 37; no. 11; p. 1019–1022; doi: 10.1130/G30276A.1; 2 ﬁ gures; Data Repository item 2009257. 1019 This study is based on a 100-cm-long sedi- RESULTS ment core recovered from Llamacocha (cocha Calibrated year A.D./B.C. 0 2000 1500 1000 500 AD/BC 500 meaning lake in the native Quechua). Llama- Core Chronology 210 210 20 Pb cocha (11°4′S, 75°48′ W; 4190 m elevation) is Unsupported Pb activity values decline in 14 )

2 C m small (0.08 km ), 14 m deep, and lies within near exponential fashion to a depth of ~35 cm c

40 (

2 h a pristine, 1.31 km granitic catchment. At a (see Fig. DR1A in the GSA Data Reposi- t

60 p distance of ~60 km southeast from Cerro de tory1). The constant rate-of-supply (CRS) e 80 D Pasco—distal to any known past and present model was applied to calculate sediment ages A mining activities—Llamacocha is strategi- and sedimentation rates spanning the past 100 cally located to record regional-scale atmo- ~100 a (Fig. DR1B) (Appleby and Oldfi eld, spheric deposition of metals associated with 1978). Calculated CRS sedimentation rates B 1.24 smelting and amalgamation in the Cerro de were relatively constant, averaging 74 (±8) g 206Pb/207Pb Pasco region (Fig. 1B). m–2 a–1 (Table DR1; Fig. DR1C). CRS ages 1.22 were then combined with the calibrated 14C METHODS dates (Table DR2 [see footnote 1]) using the local ores 1.2 mixed-effects model of Heegaard et al. (2005), 10 Core Collection and Chronology which incorporates the age-depth relationship Pb A sediment core containing an intact sedi- of the dates while integrating both within- and 1 ment-water interface was recovered using a between-sample uncertainties (Fig. 2A). The percussion corer fi tted with a 7-cm-diameter resulting age-depth model provides a conserva- 0.1 polycarbonate tube (Blomqvist, 1991). The tive interpretation of available dates incorporat- core was extruded into continuous 1 cm inter- ing the two sources of random variability (i.e., vals in the fi eld, and the upper sediments dating uncertainty and variable time elapsed Hg 10 were dated using 210Pb activities measured by between dated horizons). Accordingly, the age α-spectroscopy. The lower portion of the core model was used to derive metal fl uxes from the was dated using accelerator mass spectrometry concentration data, which compensates for vari- 14 (AMS) C age determinations obtained from able rates of sedimentation and facilitates com- 3 1 macroscopic charcoal (from grasses) and cali- parisons to other studies (Fig. 2B). brated using SHCal (McCormac et al., 2004) 2 within Calib 5.0 (Stuiver and Reimer, 1993). Geochemistry Of the elements measured, Pb and Hg are 1 Sediment Geochemistry cornerstones of our interpretations (though Ti To assess the history of regional mining and down-core stratigraphies for all elements mea- 0 2

smelting activity, we analyzed the concentra- sured are shown in Fig. DR2). This is because s

tions of 16 elements using a Perkin Elmer both are relatively immobile in lake sediments o 1

c PCA1 Elan 6000 quadrupole inductively coupled (Gallon et al., 2004; Rydberg et al., 2008), and, s

l 0

plasma–mass spectrometer (ICP-MS). To furthermore, they are uniquely representative p ascertain changes in Pb provenance, stable of the two predominant ore-processing tech- m a -1 Pb isotopic ratios were determined using a niques used in the Andes: Pb-based smelting S Nu Plasma multicollector ICP-MS coupled and Hg amalgamation. Titanium (Ti) is used to 2000 1500 1000 500 AD/BC 500 Amalgamation to DSN-100 introduction system; Tl was used represent catchment-derived lithogenic inputs Smelting to correct for mass bias. The relative standard (Fig. 2B). For nearly a millennium prior to 1.0 deviation of all Pb isotopic ratios was <0.1%. ca. A.D. 600, the accumulation of Pb is sta- C Cr Mg Fe –2 –1 )

% Mn Elements were extracted using 1 M HNO3 ble and low, averaging 0.07 (±1) mg m a . Cu 9 Al

1 overnight, a standard extractive method for This represents the accumulation of natural, ( Ca

2 Sc Zn

lake sediments (Graney et al., 1995). Dupli- nonpollution Pb accumulation in Llamacocha s i As x Na

cates were run every tenth sample and were sediment. Indeed, prior to A.D. 600, the depo- A Ti within 5% of each other. Concentrations of sition of Pb and Ti is signifi cantly correlated A Sb C Sr Pb 2 P total Hg were determined on a Mileston Inc. (r = 0.76; n = 35; P < 0.001), indicating that Hg DMA80 direct mercury analyzer. Blanks, weathering of local bedrock within the Llama- -0.4 duplicates, and standard reference materials cocha watershed largely controlled Pb delivery -1.0 PCA Axis 1 (61%) 1.0 were run every tenth sample. Blanks aver- to the lake. After ca. A.D. 600, the deposition aged 0.16 ng g–1; relative percent difference of Pb increases, becomes completely decou- Figure 2. A: Composite age-depth model and between duplicates was 11%; and standard 95% confi dence interval (CI) integrating con- reference materials were within 1% of certi- stant rate-of-supply dates (open squares) 1GSA Data Repository item 2009257, tables of and calibrated 14C median ages (shaded cir- fi ed values. To explore any underlying rela- 210Pb, 14C, and stable Pb isotopic data; plots of 210Pb, cles). B: Stratigraphic profi les of 206Pb/207Pb tionships between the analyzed elements, a CRS dates, and sedimentation rates; down-core pro- ratios, Pb, Hg, Ti, and PCA axis 1 sample correlation matrix of the centered and stan- fi les of each element measured; plots of stable Pb scores. Vertical dashed lines indicate onset dardized sediment geochemical data (21 ele- isotopic ratios; and the down-core fl ux ratio profi le of smelting and amalgamation with 95% CI for both Pb and Hg, is available online at www.geo- (shaded regions). Note Pb and Hg are plot- ments) was subjected to principal components society.org/pubs/ft2009.htm, or on request from edit- ted on log scales. C: Biplot of PCA results analysis (PCA) using CANOCO version 4.5 [email protected] or Documents Secretary, GSA, for leading two axes, showing loadings for software (ter Braak and Šmilauer, 1998). P.O. Box 9140, Boulder, CO 80301, USA. each analyzed metal.

1020 GEOLOGY, November 2009 pled from Ti (r2 = 0.02; n = 63), and 206Pb/207Pb the expansion of the Wari Empire out of the unexcavated colonial smelters and grinding ratios markedly decrease (Fig. 2B). Ayacucho Valley (Fig. 1A). The Wari were wheels—presumably used during Hg amalga- By ca. A.D. 1600, Pb accumulation rates had the largest pre-Inca empire in the Andes, and mation and the smelting of Ag-bearing ores— risen to 0.6 mg m–2 a–1. However, in contrast they maintained both direct (i.e., occupation) is located in the Chipian Valley, <10 km east to Pb and the other trace metals, Hg fails to and indirect (i.e., trade) control of the central of Cerro de Pasco and ~60 km northwest of increase over this 1000 a period, remaining Andes until ca. A.D. 1000, when they abruptly Llamacocha (Fig. 1B; Cooke and Abbott, constant at ~1.5 (±0.2) µg m–2 a–1 (Fig. 2B). collapsed (Isbell, 2008; Moseley et al., 2005). 2008). The process of Hg amalgamation domi- This likely refl ects the absence of Hg within While there have been few Wari silver artifacts nated metallurgy in the Andes for the next ~400 local ores, and the fact that Hg was not directly excavated archaeologically, there seems little years. However, the environmental legacy of utilized during precolonial smelting activities. doubt from the Llamacocha sediment record this Hg use has remained, until now, elusive. The fi rst clear increase of Hg within Llama- that smelting activity increased steadily during Nriagu (1993, 1994) estimated that Hg amal- cocha sediment occurs at A.D. 1600. Hg fl ux the ~400 years of Wari infl uence. However, the gamation resulted in annual atmospheric Hg more than doubles (Fig. DR2), rising to 3.9 µg collapse of the Wari appears to have carried lit- emissions on the order of 400–1200 t. If dis- m–2 a–1 by 1650 (Fig. 2B). The deposition of tle consequence for smelting at Cerro de Pasco, seminated globally, these emissions would Pb also increases across this period, but to a since rates of Pb deposition increase across the have resulted in an average increase in Hg lesser degree. We therefore suggest that this cultural transition (Fig. 2B). This observa- deposition rates between 0.8 and 2.4 µg m–2 a–1. fi rst appearance of Hg pollution represents tion suggests that while metallurgy at Cerro Remarkably, the Llamacocha record registers the adoption of Hg amalgamation as the pre- de Pasco was perhaps initiated by the initial an ~2.3 µg m–2 a–1 increase in Hg accumula- dominant silver-extractive technology at Cerro expansion of Wari, it was decoupled from the tion at the onset of amalgamation, very near to de Pasco, closely following colonial takeover empire’s demise. the upper end of Nriagu’s (1994) estimate. The from indigenous metalsmiths (see following This subsequent interval covers the expan- majority of these emissions likely occurred dur- discussion). sion and apogee of the Inca civilization ing the heating of the Ag-Hg amalgam and thus Further increases in Pb and Hg are noted in (ca. A.D. 1450–1532), the largest precolonial were most likely in the form of gaseous ele- Llamacocha sediment during the nineteenth cen- empire of the New World. Due to the low sedi- mental Hg (Hg0). Hg0 has an atmospheric resi- tury, and peak values for each metal are reached mentation rates within Llamacocha at this time dence time of 0.5–2.0 years, and it circulates in A.D. 1968 and 1942, respectively (Fig. 2B). (Fig. DR1C), the Inca period is poorly resolved hemispherically (and potentially also globally) Twentieth-century Pb deposition peaks two in the record. However, it is during this interval (Lin and Pehkonen, 1999). However, lake- orders of magnitude above background levels that 206Pb/207Pb isotopic ratios fi rst fall within sediment archives collected outside the Andes (Fig. DR2), representing massive enrichment the range of Cerro de Pasco ores (Fig. 2B). A have shown little evidence for preindustrial Hg associated with extensive local mining activity. simple two-component mixing model between pollution (Biester et al., 2002; Lamborg et al., This Pb enrichment is accompanied by sediment end-member 206Pb/207Pb ratios (1.245 for local 2002). While resolving this discrepancy will isotopic ratios that match closely those of Cerro bedrock and 1.200 for Cerro de Pasco ore) sug- require additional records—especially from de Pasco ores (Fig. 2B; Fig. DR3) (Mukasa et gests that ~95% of the Pb within Llamacocha the Southern Hemisphere—it is evident from al., 1990; Sangster et al., 2000). sediment was derived from atmospheric pollu- this study that the Hg amalgamation generated Our interpretation of the geochemical data tion from Inca time onward. Indeed, archaeo- regional-scale Hg pollution. is supported by the PCA results (Fig. 2C). The logical research suggests that regional silver Around A.D. 1800, atmospheric emissions fi rst PCA axis (hereafter PCA1) explains 61% production increased under Inca authority, increased dramatically, as indicated by both of the variance observed within the geochemi- likely driven by demand from Inca nobility for Hg and Pb fl ux data and the trend of PCA1 cal data set. Those trace metals that are associ- precious metals as a tribute tax (Costin et al., (Fig. 2B). During the nineteenth century, Cerro ated directly with metallurgical pollution (e.g., 1989). The geochemical record therefore sug- de Pasco was producing in excess of 60 t a–1 of Hg, Pb, Sb, As, and Zn) all produce strong gests that smelting at Cerro de Pasco not only pure silver (Brading and Cross, 1972; Purser, positive loadings on PCA1. In contrast, ele- occurred continuously, but it likely increased 1971), and during the early twentieth cen- ments that are attributed primarily to lithogenic progressively over the entire precolonial inter- tury, construction was completed on the cen- sources (e.g., Ti, Sr, Al, Sc, and Na) produce val captured in sediment record. Although tral Peruvian railway. The railway led to the negative PCA1 loadings. These relationships knowledge of early extractive technologies rapid expansion of mining activities within the imply that PCA1 defi nes a synthetic geochemi- remains fragmentary, the Inca are known to have Cerro de Pasco region, and the central Andes in cal gradient that separates metals delivered to exploited argentiferous galena (locally known general. The Cerro de Pasco Corporation was Llamacocha sediments by catchment processes as soroche) as a fl ux during smelting, which founded in A.D. 1901, initiating a shift from from those associated with atmospheric depo- was conducted in clay-lined, wind-drafted fur- silver to copper production. By 1905, two cop- sition. Thus, when arranged stratigraphically naces called huayras (Bakewell, 1984). The use per smelters were in operation (Purser, 1971). (Fig. 2B), increasing PCA1 sample scores of soroche is unquestionably associated with Accumulation rates associated with twenti- record the relative importance of anthropo- increased atmospheric Pb emission and recruit- eth-century mining operations greatly exceed genic contributions from metallurgical sources ment to down-wind lakes (Abbott and Wolfe, those achieved during colonial or precolonial to the overall geochemical record. 2003), and it is likely manifested in Pb stable intervals, and Pb stable isotopes indicate that isotopic signatures (Cooke et al., 2007). The virtually all of the sediment Pb burden at pres- DISCUSSION AND CONCLUSIONS results from Llamacocha therefore suggest that ent can be attributed to atmospheric pollution Given Llamacocha’s proximity to Cerro de similar technologies may have been known to (Fig. 2B). Despite recent declines, modern Pasco, we attribute the initial increase of sedi- metalsmiths from considerably earlier stages of accumulation rates of Pb remain 94 times mentary Pb fl ux and the synchronous decline Andean cultural development. greater than background Pb accumulation in 206Pb/207Pb ratios to the rise of smelting European discovery of the Cerro de Pasco rates. In contrast, recent rates of Hg accumula- in the region. Our chronology suggests that silver deposit occurred ca. A.D. 1600 (Fisher, tion are six times background Hg accumulation this occurred ca. A.D. 600, coinciding with 1977; Purser, 1971). A large collection of rates (Fig. DR2), which fall in line with remote

GEOLOGY, November 2009 1021 regions from around the Northern Hemisphere Brading, D.A., and Cross, H.E., 1972, Colonial silver Lin, C.J., and Pehkonen, S.O., 1999, The chemis- (Biester et al., 2007). mining: Mexico and Peru: The Hispanic Amer- try of atmospheric mercury: A review: Atmo- While the silver deposits at Cerro de Pasco ican Historical Review, v. 52, p. 545–579, doi: spheric Environment, v. 33, p. 2067–2079, doi: 10.2307/2512781. 10.1016/S1352-2310(98)00387-2. are generally believed to have been discov- Bränvall, M.-L., Bindler, R., Emteryd, O., and Ren- McCormac, F.G., Hogg, A.G., Blackwell, P.G., ered by colonial metallurgists ca. A.D. 1600, berg, I., 2001, Four thousand years of atmo- Buck, C.E., Higham, T.F.G., and Reimer, P.J., the record presented here documents over a spheric lead pollution in northern Europe: A 2004, ShCal04 Southern Hemisphere calibra- millennium of prior mining activity at levels summary from Swedish lake sediments: Jour- tion, 0–11.0 cal kyr BP: Radiocarbon, v. 46, nal of Paleolimnology, v. 25, p. 421–435, doi: p. 1087–1092. suffi cient to deliver airborne metal pollution 10.1023/A:1011186100081. Moseley, M.E., Nash, D.J., Williams, P.R., deFrance, ~60 km from the mine site. Given the paucity Cooke, C.A., and Abbott, M.B., 2008, A paleolimno- S.D., Miranda, A., and Ruales, M., 2005, of intact precolonial smelting facilities in the logical perspective on industrial-era metal pol- Burning down the brewery: Establishing and archaeological record, lake-sediment records lution in the central Andes, Peru: The Science evacuating an ancient imperial colony at Cerro associated with smelting are of considerable of the Total Environment, v. 393, p. 262–272, Baul, Peru: Proceedings of the National Acad- doi: 10.1016/j.scitotenv.2007.12.034. emy of Sciences of the United States of Amer- value in reconstructing ancient metallurgical Cooke, C.A., Abbott, M.B., Wolfe, A.P., and Kittleson, ica, v. 102, p. 17,264–17,271, doi: 10.1073/ histories. Furthermore, the sediment record J.L., 2007, A millennium of metallurgy recorded pnas.0508673102. is sensitive to evolving ore-processing tech- by lake sediments from Morococha, Peruvian Mukasa, S.B., Vidal, C., C.E., and Injoque-Espinoza, nologies, recording the switch from reductive Andes: Environmental Science & Technology, J., 1990, Pb isotope bearing on the metallogene- v. 41, p. 3469–3474, doi: 10.1021/es062930+. sis of sulfi de ore deposits in central and southern smelting in traditional huayras to colonial Hg Cooke, C.A., Abbott, M.B., and Wolfe, A.P., 2008, Peru: Economic Geology, v. 85, p. 1438–1446. amalgamation. Finally, our data confi rm the Late-Holocene atmospheric lead deposi- Nriagu, J.O., 1993, Legacy of mercury pollution: hypothesis, fi rst proposed by Nriagu (1993, tion in the Peruvian and Bolivian Andes: The Nature, v. 363, p. 589, doi: 10.1038/363589a0. 1994), that colonial amalgamation resulted in Holocene, v. 18, p. 353–359, doi: 10.1177/ Nriagu, J.O., 1994, Mercury pollution from the past widespread Hg pollution. 0959683607085134. mining of gold and silver in the Americas: The Costin, C., Earle, T., Owen, B., and Russell, G., 1989, Science of the Total Environment, v. 149, p. 167– The impact of Inca conquest on local technol- 181, doi: 10.1016/0048-9697(94)90177-5. ACKNOWLEDGMENTS ogy in the Upper Mantaro Valley, Perú, in van Peterson, U., 1965, Regional geology and major ore Funding for this project was provided by grants der Leeuw, S.E., and Torrence, R., eds., What’s deposits of central Peru: Economic Geology from the National Geographic Society and the Natu- New? A Closer Look at the Process of Innova- and the Bulletin of the Society of Economic ral Sciences and Engineering Research Council of tion: London, Unwin Hyman, p. 107–139. Geologists, v. 85, p. 1287–1295. Canada. We are grateful to Pedro Tapia for assistance Einaudi, M.T., 1977, Environment of ore deposition Purser, W.F.C., 1971, Metal-Mining in Peru, Past and in the fi eld; Jack Cornett, Janice Lardner, and John at Cerro de Pasco, Peru: Economic Geology Present: New York, Praeger Publishers, 339 p. Southon for sediment dating; Guangcheng Chen, and the Bulletin of the Society of Economic Renberg, I., Persson, M.W., and Emteryd, O., 1994, Robert Creaser, Antonio Simonetti, and Prentiss Bal- Geologists, v. 72, p. 893–924. Pre-industrial atmospheric lead contamination com for assistance in the laboratory; and two anony- Fisher, J.R., 1977, Silver Mines and Silver Miners detected in Swedish lake sediments: Nature, mous reviewers. in Colonial Peru, 1776–1824: Liverpool, Cen- v. 368, p. 323–326, doi: 10.1038/368323a0. tre for Latin American Studies, University of Rydberg, J., Gälman, V., Renberg, I., Bindler, R., REFERENCES CITED Liverpool, 150 p. Lambertsson, L., and Martínez-Cortizas, A., Abbott, M.B., and Wolfe, A.P., 2003, Intensive pre- Gallon, C., Tessier, A., Gobeil, C., and Alfaro-De La 2008, Assessing the stability of mercury and Incan metallurgy recorded by lake sediments Torre, M.C., 2004, Modeling diagenesis of lead methylmercury in a varved lake sediment de- from the Bolivian Andes: Science, v. 301, in sediments of a Canadian Shield lake: Geo- posit: Environmental Science & Technology, p. 1893–1895, doi: 10.1126/science.1087806. chimica et Cosmochimica Acta, v. 68, p. 3531– v. 42, p. 4391–4396, doi: 10.1021/es7031955. Appleby, P.G., and Oldfi eld, F., 1978, The calcula- 3545, doi: 10.1016/j.gca.2004.02.013. Sangster, D.F., Outridge, P.M., and Davis, W.J., tion of lead-210 dates assuming a constant rate Graney, J.R., Halliday, A.N., Keeler, G.J., Nriagu, 2000, Stable lead isotope characteristics of lead of supply of unsupported 210Pb to the sediment: J.O., Robbins, J.A., and Norton, S.A., 1995, Iso- ore deposits of environmental signifi cance: En- Catena, v. 5, p. 1–18, doi: 10.1016/S0341-8162 topic record of lead pollution in lake sediments vironmental Reviews, v. 8, p. 115–147, doi: (78)80002-2. from the northeastern United States: Geochim- 10.1139/er-8-2-115. Bakewell, P., 1984, Miners of the Red Mountain: In- ica et Cosmochimica Acta, v. 59, p. 1715–1728, Stuiver, M., and Reimer, P.J., 1993, Extended 14C da- dian Labor at Potosí, 1545–1650: Albuquerque, doi: 10.1016/0016-7037(95)00077-D. tabase and revised CALIB radiocarbon calibra- University of New Mexico Press, 213 p. Heegaard, E., Birks, H.J.B., and Telford, R.J., 2005, tion program: Radiocarbon, v. 35, p. 215–230. Biester, H., Kilian, R., Franzen, C., Woda, C., Relationships between calibrated ages and ter Braak, C.J.F., and Šmilauer, P., 1998, CANOCO Mangini, A., and Schöler, H.F., 2002, El- depth in stratigraphical sequences: An esti- Reference Manual and User’s Guide to Canoco evated mercury accumulation in a peat bog mation procedure by mixed-effect regression: for Windows: Software for Canonical Commu- of the Magellanic Moorlands, Chile (53°S)— The Holocene, v. 15, p. 612, doi: 10.1191/ nity Ordination (version 4): Ithaca, New York, An anthropogenic signal from the Southern 0959683605hl836rr. Microcomputer Power, 352 p. Hemisphere: Earth and Planetary Science Let- Isbell, W., 2008, Wari and Tiwanaku: International Ward, H.J., 1961, The pyrite body and copper ore- ters, v. 201, p. 609–620, doi: 10.1016/S0012- identities in the Central Andean Middle Ho- bodies, Cerro de Pasco Mine, central Peru: Eco- 821X(02)00734-3. rizon, in Silverman, H., and Isbell, W., eds., nomic Geology and the Bulletin of the Society Biester, H., Bindler, R., Martinez-Cortizas, A., and Handbook of South American Archaeology: of Economic Geologists, v. 56, p. 402–422. Engstrom, D.R., 2007, Modeling the past atmo- New York, Springer, p. 761–782. spheric deposition of mercury using natural ar- Lamborg, C.H., Fitzgerald, W.F., Damman, A.W.H., chives: Environmental Science & Technology, Benoit, J.M., Balcom, P.H., and Engstrom, v. 41, p. 4851–4860, doi: 10.1021/es0704232. D.R., 2002, Modern and historic atmospheric Manuscript received 30 March 2009 Blomqvist, S., 1991, Quantitative sampling of soft- mercury fl uxes in both hemispheres: Global Revised manuscript received 3 June 2009 bottom sediments: Problems and solutions: and regional mercury cycling implications: Manuscript accepted 22 June 2009 Marine Ecology Progress Series, v. 72, p. 295– Global Biogeochemical Cycles, v. 16, p. 1104, 304, doi: 10.3354/meps072295. doi: 10.1029/2001GB001847. Printed in USA

1022 GEOLOGY, November 2009