Report

Intensive Pre-Incan Metallurgy Recorded by Lake Sediments from the Bolivian Andes

See allHide authors and affiliations

Science  26 Sep 2003:
Vol. 301, Issue 5641, pp. 1893-1895
DOI: 10.1126/science.1087806

Abstract

The history of pre-Columbian metallurgy in South America is incomplete because looting of metal artifacts has been pervasive. Here, we reconstruct a millennium of metallurgical activity in southern Bolivia using the stratigraphy of metals associated with smelting (Pb, Sb, Bi, Ag, Sn) from lake sediments deposited near the major silver deposit of Cerro Rico de Potosí. Pronounced metal enrichment events coincide with the terminal stages of Tiwanaku culture (1000 to 1200 A.D.) and Inca through early Colonial times (1400 to 1650 A.D.). The earliest of these events suggests that Cerro Rico ores were actively smelted at a large scale in the Late Intermediate Period, providing evidence for a major pre-Incan silver industry.

New World metallurgy emerged in the Andean region of South America between the Initial Period (1800 to 900 B.C.) and the Early Horizon (900 to 200 B.C.) (1). The oldest well-dated archaeological site containing metal artifacts is Mina Perdida (Lurín Valley) in coastal Peru, where hammered foils and gilded copper are preserved in contexts dating to 1400 to 1100 B.C. (2). The tradition of sheet-metal working (hammering, gilding, annealing, and repoussé) remained pervasive in the Andes throughout the Early Intermediate Period (200 B.C. to 600 A.D.) and the Middle Horizon (600 to 1100 A.D.) (3). By 1000 A.D., large-scale copper smelting and bronze production is evident at sites such as Batán Grande on the northern Peruvian coast (4). Beginning in the Late Intermediate Period (1100 to 1450 A.D.), intensive copper working became widespread on the Bolivian altiplano, with the production of materials of copper-tin alloy (i.e., bronze), in contrast to the copper-arsenic artifacts found in Peru (5). By this time, silver and gold were well-established as precious metals among Andean cultures. Although silver was highly sought by royalty for symbolic and ritual purposes (6), the geographic distribution, intensity, and timing of Late Intermediate Period silver mining in the Andes remains unclear. Here, we infer a regional history of metallurgy from lake sediments retrieved adjacent to the largest silver deposit of the Bolivian tin belt.

Laguna Lobato (hereafter LL) (7) is small (0.2 km2), relatively deep (11 m), and occupies a nonglacial catchment of 3.9 km2 (Fig. 1). The lake overflows only during the wet season (December to March), and it has no hydrological connection with surface waters draining Cerro Rico, 6 km west of the lake. Because westerly winds prevail for 8 months of the year (April to November) (8), LL is strategically located to record atmospheric deposition of metals volatilized during smelting or transported as fine-grain particulates. This study is based on a 74.5-cm core recovered from the deepest portion of the lake and dated by 210Pb, 137Cs, (table S1), and 14C analyses (9) (table S2).

Fig. 1.

Location map of the study site in relation to the Tiwanaku capital, Lake Titicaca, Potosí, and the Quelccaya ice core. The shaded area indicates the central Bolivian tin belt (27), which broadly corresponds to the crest of the Andean cordillera.

Cerro Rico lies within a zone of xenothermal mineralization related to Middle Tertiary intrusions (10). In addition to native silver, the richest ores contain combinations of acanthite (Ag2S), andorite (PbAgSb3S6), chlorargyrite (AgCl), matildite (AgBiS2), miargyrite (AgSbS2), pyrargyrite (Ag3SbS3), and tetrahedrite [(Ag, Cu, Fe, Zn)12Sb4S13] (11). Tin is associated primarily with cassiterite (SnO2). To assess the history of smelting, we measured the concentrations in lake sediments of five metals (Ag, Bi, Pb, Sb, and Sn) associated with ore composition (9). Of these metals, Pb serves as the cornerstone of our interpretations for two reasons. First, the Incan smelting technology relied on argentiferous galena [soroche (Pb, Ag)S] as a flux during smelting, which was conducted in charcoal-fired, wind-drafted furnaces lined with clay (huayras) (12, 13). The use of soroche led to excessive Pb volatilization, resulting in lake sediment concentrations that are orders of magnitude higher than those of the other analyzed metals. Second, Pb is largely immobile once deposited in lake sediments (14, 15). Although molecular diffusion rates for Ag are higher than those for Pb (16, 17), in both cases they are insignificant in comparison to average sediment accumulation rates in LL (∼1 mm year–1). Thus, postdepositional mobility is not a confounding factor in the interpretation of the record, a conclusion supported by the largely parallel trends observed for each of the metals.

Before 1000 A.D., concentrations of all five metals in the sediments of LL were low and stable, representing natural background levels of metal accumulation (Fig. 2). An additional 15 samples spanning the earlier period from 2000 B.C. to 600 A.D. have similarly low Pb concentrations (∼20 μg g–1), despite stable isotopic evidence for pronounced hydrological changes in this time interval (18) (fig. S1). Thus, Holocene climate variability exerted little influence on nonpollution metal fluxes to the lake's sediments. Metal concentrations initially rose well above background shortly after 1000 A.D., reaching a first peak around 1130 to 1150 A.D. (Fig. 2). Concentrations of Pb exceed 100 μgg–1 in this interval, approximately one-third of the peak Pb burden reached subsequently in early Colonial times. Such enrichment trends are directly comparable to those reported from lakes proximal to major Medieval mining sites in Europe (19). Because we know of no natural processes capable of inducing this degree of Pb enrichment, and given the proximity of LL to a major source of metal pollution, we associate the magnitude of metal enrichment in the lake's sediments with the intensity of ore smelting.

Fig. 2.

Chronology of the core from LL, Andean archaeology and paleoclimate, and sediment metal concentrations. (A) Constant Rate of Supply (CRS) age model based on 210Pb analyses of the upper core section (39 measurements) and macrofossil Accelerator Mass Spectrometer (AMS) 14C dates (9). (B) Generalized archaeology [Peruvian chronology from (1) and (24), historical events from (12), and regional paleoclimate from (23, 25)]. (C) Concentrations of Pb, Sb, Bi, Ag, and Sn from sediments of LL over the past 1300 years. Shaded zones identify three distinct metallurgical zones: Tiwanaku (1000 to 1250 A.D.), Inca–early Colonial (1400 to 1650A.D.), and the rise and crash of tin mining (1850to 1950A.D.). The vertical dashed line (1545 A.D.) separates Incan and Colonial periods. Background concentrations (pre-1000 A.D.) are 20.2 ± 2.2, 0.19 ± 0.15, 0.02 ± 0.01, 0.03 ± 0.01, and <0.01 μg g–1 for Pb, Sb, Bi, Ag, and Sn, respectively (n = 15).

This initial rise in metal concentration coincides with the late stages of the Tiwanaku Empire that controlled the Lake Titicaca basin and extended beyond Oruro in the southern altiplano (Fig. 2). Very few Tiwanaku silver artifacts appear to have survived, except for early discoveries at Chucaripupata, on Lake Titicaca's Island of the Sun (20). Therefore, it is difficult to ascertain which cultural group was smelting Cerro Rico ores before the Tiwanaku collapse, or where artifacts were being fabricated and traded. However, the synchrony between declining LL metal concentrations and the collapse of the Tiwanaku state suggests that metallurgy at Cerro Rico was directly linked to the larger Tiwanaku polity. Although the demise of the Tiwanaku is a subject of debate (21), it has been linked in part to pervasive droughts that forced the abandonment of raised-field agriculture (22). The level of Lake Titicaca dropped by as much as 6 m between 1100 and 1250 A.D. (23). The corresponding reduction of smelting intensity at Cerro Rico suggests some degree of state level administration over the extraction of silver from the mountain.

During the subsequent Altiplano Period (∼1100 to 1400 A.D.), which separated Tiwanaku and Inca imperial regimes, decentralized polities arose across the Titicaca basin (24). Metal concentrations from LL sediments suggest that smelting at Cerro Rico continued during this interval, although at greatly reduced levels of activity. Decreased social organization after the fall of Tiwanaku likely curtailed the demand for ceremonial metals. Furthermore, regional population may have declined as a consequence of crop failures, given the strong evidence from the Quelccaya ice core for persistent drought in the 1250 to 1310 A.D. interval (25).

By 1400 A.D., a renewed intensification of smelting at Cerro Rico is indicated by increases in the concentrations of Pb and Sb in LL sediments. We attribute this to the rise of Incan metallurgy, which lasted until the Spanish conquest. At the height of this interval, Pb concentrations reached 150 μg g–1 (Fig. 2). During the Inca period however, Ag, Bi, and Sn concentrations did not return to levels recorded during the earlier smelting peak, suggesting technological advances by Incan metallurgists aimed at minimizing volatile losses of silver. This notion is supported by distinctive spikes of these three metals immediately after the Spanish arrival at Potosí (1545 A.D.). Early Colonial smelting used bellowed Castilian stone furnaces that had proven successful elsewhere in the Andes, but repeatedly failed at Cerro Rico (12). These furnaces overheated the ore, thus volatilizing metals, including the silver targeted for extraction. Accordingly, the maximum sediment Ag enrichment of the last millennium occurs in conjunction with colonial experimentation. Due to the failure of Spanish extractive techniques, the indigenous huayra technology was retained at Cerro Rico, leaving the smelting process largely in the hands of Incan metallurgists under colonial rule. In the 27 years that followed the Spanish arrival, thousands of active huayras adorned the mountain at any given time (11, 12). Their collective atmospheric Pb emission, largely associated with soroche use, raised sediment Pb concentrations in LL to nearly 300 μg g–1 (Fig. 2).

With the depletion of silver-rich surface ores in 1572 A.D., smelting was replaced by mercury amalgamation as the primary extraction process, a technology brought to Bolivia from Mexico (26). Amalgamation facilitated silver extraction from the lower grade ores recovered in subterranean adits. Atmospheric fluxes of metals progressively declined, as registered by their decreasing concentrations in LL sediments (Fig. 2). Silver production from Cerro Rico declined until its eventual abandonment in 1930 A.D. (10, 11). As the silver supply dwindled, large-scale tin production became important at Potosí in the late 19th and early 20th centuries A.D., ultimately fueled by high demands during World War I. Peak sediment Sn concentrations faithfully track both the transition to tin production and the industry's subsequent crash (1950 A.D.).

The results from LL extend the record of precious metal smelting in the southern Bolivian Andes by several centuries. Although legend attributes the discovery of silver at Cerro Rico to the 11th Inca ruler, Huayna Capac, in the mid-15th century A.D. (11), our data suggests that the deposit was known and exploited as a source of silver in the Late Intermediate Period, as early as the 11th century A.D. The scarcity of coeval pre-Incan silver artifacts is, therefore, due neither to the absence of adequate extractive technologies nor to a limited knowledge of argentiferous ores. Rather, we propose that the earliest silver artifacts to originate from Cerro Rico were looted, likely both before and after European contact. Although it remains impossible to determine precisely the amount of preColonial silver that was extracted from Cerro Rico, total production since 1545 A.D. is estimated between 20,000 and 40,000 metric tons (10, 12). Assuming that pre-Incan technologies operated at efficiencies comparable to huayras and recognizing the initial presence of grades as rich as 25% Ag (11), our data imply that several thousand tons of silver were produced in pre-Incan times. Although major new archaeological discoveries in the Andes remain a distinct possibility, the likelihood seems equally probable that most of this silver was recycled and transported elsewhere in the Americas before conquest, or eventually exported overseas by the Spanish.

Supporting Online Material

www.sciencemag.org/cgi/content/full/301/5641/1893/DC1

Materials and Methods

Fig. S1

Tables S1 and S2

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article