A few years ago, one of my colleagues was looking for historical documents about forced labor in Tena, former capital of the Ecuadorian Amazon. During her lunch break, a middle-aged Kichwa woman asked what she was doing there. My colleague explained and without hesitation her questioner exclaimed, Yo fuí ven- dida! (“I was sold!”). When she was young, she said, her family lived on a nearby hacienda. In the mid-1960s, the landowner sold their debt to a neighboring hacendado, who became their boss. By then, of course, slavery in Ecuador had been outlawed for more than a century; debt servitude and Indian tribute had been prohibited for years; minimum wage laws had been enacted and communal lands were protected by law. Yet concertaje (peonage) in Ecuador and other parts of the far western Amazon persisted―albeit without the whip and stocks―through the 1960s, perhaps even later (Beghin, 1964; Muratorio, 1991) 1.
In this article, I analyze the Rubber Boom (1885-1930) and its impact on indigenous people in western Amazonia. Academic discussions of rubber there have been largely shaped by Julio César Arana’s Peruvian Amazon Company (PAC), which controlled the lower Putumayo River until around 1920 (Map 1). At first, Arana depended heavily on brute force―not just his private goons, but also Peruvian soldiers and gunboats. Nonetheless, his reliance on brutality was short-lived and he soon fell back on more ordinary kinds of debt peonage (Santos Granero & Barclay, 2000: pp. 49-50) . In effect, Arana’s violence represented the most notorious point on a spectrum of Amazonian servitude running from outright enslavement to near autonomy. Unfree labor systems, usually based on debt, emerged around the world as international demand for industrial commodities rose during the 19th century2.
In constructing this analysis, I have tried to draw connections between the massive disruption of indigenous societies―equivalent to the fur trade in North America―and their subsequent “retribalization” later on. After three generations of forced labor, former ethnic boundaries had been largely worn down or
Map 1. The Western Amazon.
erased; new patterns emerged to take their place. As the post-rubber agricultural economy took shape, backward-looking affiliations counted less than common residence, shared languages and recent family ties. As Whitten (1976) and others have written, affiliation remained flexible and shifted with circumstances. Out of such looser ties, modern native communities emerged. A clear-eyed view of the recent past remains essential if we are to understand the origins and trajectory of indigenous groups in Amazonia today3.
2. Labor Systems in the Western Amazon
Wild rubber collection in the Brazilian Amazon began in the early 19th century. In 1855, 2,000 tons of cured latex were exported from Brazil; by 1885, this figure had risen six-fold to 12,000 tons (Weinstein, 1983: p. 9) 4. Plantations there were unsuccessful because Hevea brasiliensis―the main source of high-quality rubber― suffered from a leaf rust when planted close together (Dean, 1987; Grandin, 2009) . Instead, wild Hevea was harvested by seringueiros (tappers or collectors), including impoverished peasants from Northeast Brazil or hopeful chancers from Britain, the United States, Europe and elsewhere (Pearson, 1911; Yungjohann, 1989 [ca. 1920]; Woodroffe, 1914) . Seringueiros received tools and supplies from their patrãoes―usually small merchants or landowners (seringalistas) who controlled Hevea stands in distant forests (Weinstein, 1983: pp. 13-24) . At least once a year, seringueiros delivered their harvest to these patrãoes and were paid in cash or goods worth perhaps a quarter of the product’s market price. Payments were usually insufficient to cover their debt, so the cycle began again5.
Remote traders and seringalistas operated on credit from larger merchants (aviadores) in Pará, Manaus or Iquitos. Rubber collected in the forest was sent downriver to these aviadores; larger export houses then bought it for shipment abroad. “Politicians and men of influence stake as much as they can on [virgin forest] and file a claim with the government,” wrote an American seringueiro, John Yungjohann, around 1920 (1989: p. 11). “As soon as they have secured themselves that way they put themselves at the disposal of the exporter who in turn will give them an almost unlimited credit. Having gone that far, their first move is to establish an agency in the next best hotel to induce and welcome any young man who is willing to go and risk his life…”
Compulsion played little role in recruiting seringueiros. “For the most part,” explains Weinstein (1983: p. 23) , “the tapper was much too mobile and too far removed from the patrão’s control for his debts to operate as an effective means of restraint.” Instead, poverty and a lack of alternatives drew seringueiros into the forest. Isolated on remote rivers, dependent on his patrão for supplies and transportation, “the tapper entered a relationship that he could not normally afford to break unless he was prepared to leave the area…” For the next 30 years, endless variations on this relationship played out in the Brazilian Amazon.
In Bolivia, where Hevea brasiliensis also grew, rubber production took a very different form. In the 1870s, quinine collectors descended the Madre de Dios and Beni River basins and sold their chichona bark to a German exporter, the Casa Richter (Gamarra, 2007: p. 55) . By 1880, however, Richter had been eclipsed by Nicolás Suárez, who became the country’s leading trader. As the market for Amazonian quinine dried up, Suárez acquired large rubber concessions, built up an extensive infrastructure of warehouses, roads, boats and related facilities, raised capital in Europe, bought out most of his rivals and replaced independent seringueiros with indebted peons (Fifer, 1970: pp. 113-146) . By 1912, his companies claimed to own 18,760 estradas (rubber groves with 100 - 150 trees apiece) scattered over 8 million hectares (Gamarra, 2007: p. 190) . Like Arana, Suárez and his few competitors ran integrated operations extending from London to isolated outposts 2000 km upriver from Manaus.
Labor shortages remained a constant worry. In the mid-19th century, only around 18,000 Indians lived in Bolivia’s Amazon region, along with several thousand whites. Between 1849 and 1874, Indian populations, struck by disease and other hardships, fell by half (Van Valen, 2013: p. 97) . Indigenous groups that violently resisted encroachment were quickly annihilated (Gamarra, 2007: p. 72; Van Valen, 2013: p. 66) . With few workers available locally, rubber barons like Suárez cast a wider net. Between 1881 and 1920, the Bolivian government broke up highland communities and dispossessed their residents (Klein, 2003: pp. 146-147; Larson, 2004: p. 222) . Despite reports of whipping and ill-treat- ment in Amazonia, displaced peasants flooded into the forests. By 1900, according to the US commercial attaché, around 200,000 people―mostly recent arrivals―lived in Bolivia’s tropical lowlands (Schurz, 1925: pp. 206-225) . Nearly 8000 indebted tappers labored in Suárez’s rubber camps (Schurz, 1921: pp. 169- 172) .
Rubber collection took a very different course in Peru. Beginning in the 1860s, it spread along the Marañón and upper Amazon Rivers and their tributaries. Here, Castilloa trees―the source of caucho―were cut and drained into pits. By the 1880s, however, Castilloa forests along the Marañón were played out, so tappers moved into the lower Napo and Putumayo valleys and also onto tributaries of the Ucayali (Santos Granero & Barclay, 2000: pp. 29-31; Bellier, 1994: p. 35) . In 1884, 1500 tappers were harvesting caucho along the middle Ucayali itself; a decade later, 5000 peones had moved into the lower Ucayali valley and adjacent waterways. Most workers were Indians who had become “dependent on manufactured goods for their livelihood but lacked the means to obtain them” (Santos Granero & Barclay, 2000: p. 37) . Debt peonage offered virtually the only way for Indians to acquire tools and other goods. For patrones, it ensured the steady supply of willing hands.
North of the Marañón, Castilloa forests began to produce caucho again around 1890. Peruvian caucheros paddled up the Tigre, Pastaza and Napo Rivers, where they competed with Ecuadorian and Colombian rivals6. A network of small steamers―owned by Iquitos trading companies or their representatives―bought rubber and peddled their goods throughout the Napo, Curaray and Pastaza valleys. In 1905, Peru closed these rivers (or at least the parts it controlled) to foreigners, who then moved to the Napo. From rubber stations called fundos, they organized expeditions into distant forests that lasted up to a year. Rubber was hauled over trails known as varaderos and shipped to Puerto Leguízamo (Colombia) on the Putumayo or Rocafuerte (Ecuador) on the Napo. There it was delivered to agents of Iquitos commercial firms and sent downriver.
South of the Marañón, Peruvian caucheros also expanded their territory as older Castilloa forests were depleted or new areas were discovered7. By 1888, rubber collection there was dominated by Carlos Fermín Fitzcarrald, who bridged the Urubamba and the Madre de Dios river systems, opening vast new production areas. According to Brown & Fernández (1991: p. 63; see also Santos Granero & Barclay, 2000: pp. 67-68) , Fitzcarrald’s army “combed the jungle for Asháninka and Piros who could be induced―at gun point, if necessary―to work rubber concessions along the Urubamba, Manú and Mishagua rivers.” Eventually, Fitzcarrald controlled four major Piro groups on the Urubamba and most Asháninka on the Ucayali and Tambo. As in Bolivia, resisters were eliminated (Reyna, 1942: p. 86) .
Unlike Hevea, rubber from Castilloa trees was usually collected by large work gangs under the direct command of their patrón or his overseer. In the early years, these workers included mostly poor mestizos and catechized Indians from the nearby Andean foothills. But as demand for rubber grew, patrones drew in more remote and “uncivilized” people from beyond the frontier (Santos Granero & Barclay, 2000: p. 39) . Santos Granero & Barclay (2000: p. 43) describe how this system worked:
A local patron would outfit an important headman, providing him with goods in exchange for a certain amount of rubber. The chief distributed part of these goods among his followers, retaining the rest as a sort of commission. He then led his men in a rubber-gathering expedition that could last up to six months…This was achieved thanks to the generalized desire for manufactured goods, as well as through the use of a certain dose of implied or actual violence…8
By the 1890s, write Brown & Fernández (1991: p. 68) , “Native headmen had become virtual warlords, armed by rubber traders and thoroughly integrated into a system of debt-slavery.”9
Perhaps the most infamous headman was Venancio Amaringo Campa, an Asháninka chieftain who got his start in the 1880s working for Fitzcarrald10. Around 1900, Venancio commanded a force of 100 Ashininka families who collected rubber on the Purús River. Within a few years, he was also raiding other groups to capture women and children for sale as domestic servants (Santos Granero & Barclay, 2000: p. 42; see also Gow, 1991: p. 41) . Passing through in the early 1920s, the Swedish geographer Otto Nordenskjold commented on “the scant Asháninka population along the large rivers and… the presence of observation towers… to observe the Ene and Tambo Rivers, from where the rubber dealers’ raids would be launched…” (quoted in Varese, 2002: p. 137 ).
For the most part, such events escaped public notice. But not “the scandals of the Putumayo,” as Julio César Arana’s depredations were called11. In 1907, Arana reorganized his PAC as a London-based enterprise with British directors. Two years later, an American named Walter Hardenburg turned up with alarming accounts of Arana’s atrocities and documentary evidence to back them up. The British Parliament demanded an investigation. In 1910, the Foreign Office dispatched its consul general in Brazil, Roger Casement, to the Putumayo and was appalled by his report.
Why did Arana abandon debt servitude and instead turn to mutilation and beatings? At its height, PAC controlled 31,000 square kilometers of forest in the lower Putumayo valley with more than 13,000 Indian workers (Stanfield, 1998: pp. 104-105) . Even so, Arana struggled to survive against better financed European and British competitors (Santos Granero & Barclay, 2000: p. 86) . To expand his operations, he decided to raise more capital on the London Stock Exchange. Here he faced a major obstacle: he owned no land or trees, nothing that would significantly improve the price of PAC’s stock. His only option was to boost his output, albeit temporarily. Between 1902 and 1906, then, his field managers increased their production six-fold fold by using what Baptist (2014) in a similar context has called “the whipping machine.”12
From the beginning, of course, Arana recognized that brutality provided only a stop-gap solution to his labor shortage. After his stock offering, he returned to more conventional forms of debt peonage and allowed rubber production to decline again. But Parliament was not so easily mollified. To placate it, Peruvian authorities brought criminal charges against Arana and a few henchmen, although with little effect. Judges were bought or intimidated, witnesses and evidence disappeared, cases dragged on for decades. To forestall further attacks, Arana had himself elected senator (with immunity from prosecution) and eventually died in his sleep. The “culture of terror,” as Taussig (1984, 1987) later called it, turned out to be part of a short-term market scam intended to deceive London investors.
Indians were not the only victims of forced labor: even foreigners could become indebted servants. In 1914, for example, the New York Times published lurid accounts of Joseph Woodroffe, a young Englishman who went to Peru in 1905. Woodroffe first settled in Nauta, a short distance upriver from Iquitos, where he became a small-time rubber trader. Eventually, he bought the debts of 58 Indian men and 14 women and set out to collect caucho along the upper Tigre River (Woodroffe, 1914: p. 91) . He and his Indians worked there for nearly a year before returning to Nauta. But by then international prices had dropped and Woodroffe was unable to recover his expenses. To make matters worse, his manager in Nauta had run off with the rest of his money and goods; now his workers began to desert him. Facing debts of £1000 [$5000, worth around $100,000 today], he sold their outstanding loans to a Peruvian cauchero who was setting out for a four-year excursion to the Purús River in Brazil (Woodroffe, 1914: p. 112) .
Yet his odyssey wasn’t over. Just as Arana’s terror was reaching its peak, Woodroffe was hired as a bookkeeper by the Peruvian Amazon Company. And there the failure of his earlier Tigre venture finally caught up with him. The caucho that he had sent on commission to Europe nearly two years earlier “brought exceedingly poor prices,” he explains (1914: p. 141).
This had the effect of placing me heavily in debt, to the extent of several hundreds of pounds sterling, and my creditors, knowing of my presence in the Putumayo with the firm of Arana, applied to the Iquitos branch for payment of my debt, a demand which was acceded to without any reference to myself, though no money changed hands, as the Iquitos traders were debtors to Arana, and the accounts were simply placed to my debit and their credit… I was now a victim of peonage; from this day on my life was a living hell, and it must not be wondered at that for a while, I lost all hope of ever becoming a decent member of society again.
It took him four years to repay Arana. In 1913, he departed for England with little to show for his years in Amazonia.
3. Lords of the Riverbanks and Their Subjects
Caucho forests extended far up the Napo and Putumayo rivers into the Andean foothills of Ecuador and southern Colombia. There, too, rubber collection trans- formed labor systems that had taken shape half a century earlier (Pennano, 1988; Barclay, 1998; Esvertit, 2008) . By the 1850s, indigenous people in Ecuador’s Amazonian region (known as the Oriente) fell into three groups: 1) indios mansos (“tame Indians”), Kichwa-speaking people also called runa; 2) “peaceful but uncivilized” Zápara bands scattered between the Napo and the Pastaza; and 3) infieles (“heathens”) primarily Jívaro families (Shuar and Achuar) whose territory extended to the Marañón in Peru.
Although indios mansos lived in isolated households, they were not free. Every household owed large debts to government officials and their merchant partners under a system of repartos: forced sales of cheap trade goods (cotton cloth, beads and metal tools) on credit at inflated prices. Tame Indians were required to repay their loans with gold, smoked fish, pita fiber and tobacco. They also made long expeditions down the Napo by canoe or walked over hazardous trails to trade for the things their masters wanted: curare, salt and slaves from Jívaro and other groups living farther down the Marañón Basin13.
Rubber collection reoriented these networks and the ethnic niches within them. “Peaceful but uncivilized” Zápara intermediaries quickly disappeared―or rather, they joined the mass of Kichwa-speaking laborers brought by Ecuadorian and foreign bosses to a dozen fundos on the middle Curaray, Napo and adjacent rivers (Barclay, 1998: p. 143; see also Reeve, 1985: pp. 78-80; Hill, 1999: p. 750) . By the early 1920s, 4000 - 5000 indebted peons worked on 55 fundos extending downriver into Peru (Gamarra, 1996) . As Kichwa families moved to the fundos, older settlements in the Andean foothills (known as the montaña or piedemonte) virtually disappeared.
By 1915, Ecuadorian caucheros (known as señores ribereños, “lords of the riverbanks”) had consolidated their hold on the upper Napo and its tributaries14. In the Putumayo watershed, however, a different pattern took shape. Beginning in the 1880s, Colombian traders occupied strategic junctures along the upper Aguarico, San Miguel and Putumayo Rivers. Many of these patrones had come into Amazonia from Pasto and Popayán to harvest quinine ten years earlier and stayed on to collect rubber (Domínguez & Gómez, 1990: p. 139; Pineda, 2000: p. 41) . As early as 1890, they controlled virtually every Indian household there: Cofán, Siona, Oyo, Coreguaje, Macaguaje and even Kichwa people from the Napo whose debts they bought (Wasserstrom, 2014: pp. 525-548) .
In general, Colombian patrones traveled with their workers, who harvested caucho along the Andean foothills. Supervision was limited to what one patrón or his overseer could manage by himself with 50 or 75 widely scattered workers. Waiting just offstage, other patrones frequently tried to “seduce” his laborers by paying off their debts. As rubber trees near the foothills became scarce, caucheros moved eastward across the Putumayo to the upper Caquetá. By 1910, few Indians remained in Ecuador north of the Napo. Like señores ribereños, Colombian bosses also took their workers to the Marañón or even further afield if prospects there seemed brighter. And periodically, when caucho prices collapsed, as they did in 1901, 1907 and 1911, Colombian and Ecuadorian rubber men sold their peones in Peru or Brazil to repay their own creditors. As late as the 1970s, native communities along the lower Napo and Marañón in Peru recalled their origins in Ecuador’s distant foothills (Mercier, 1979: pp. 243-247; Hudelson, 1981: p. 216) .
As we know, the epicenter of rubber production in Colombia lay in Arana’s territory several hundred kilometers downriver along the lower Putumayo. Although this region produced mainly jebe débil, it had one major advantage: relatively large native populations. According to Pineda (2008: p. 48), by 1900 nearly 70,000 Indians, divided among eight major ethnic groups, lived nearby. What brought these people―who until then had largely escaped peonage―into the rubber camps? Isolation exacted a high cost: steel tools, although greatly desired, were almost impossible to come by. Around 1880, Colombian traders began arriving and established 18 major collection posts, where they exchanged cheap tools on credit for rubber. By 1904, Arana had bought all of these outposts or forced their owners out of business. He then launched his reign of terror, which subsided once his stock offering in London was complete15.
Throughout Amazonia, the common thread among these diverse operations was debt. Despite their early liberal convictions, 19th century elites soon recognized that “the institutional and normative apparatus of coerced labor would not soon be dismantled” (Larson, 2004: p. 8) . Every reduction in colonial-era head taxes or personal service requirements made Indians less willing to build roads and bridges, clear forests, drain swamps or construct public buildings (Hassaurek, 1868; see also Van Valen, 2013: p. 37) . In contrast to the highlands, as Gow (1991: p. 46) points out, lowland Indians could not usually be forced to work by expropriating their land or incorporating it into haciendas. In the Amazon, debt provided a ready solution: it commoditized labor and enabled workers to be bought, sold or mortgaged, while preserving the fig leaf of personal liberty16.
Debt also provided a critical tool to minimize risk for the commercial elites that controlled rubber production. Three kinds of risk were especially dangerous: unpredictable price swings in the international market; fluctuating exchange rates; and losses due to cheating or disaster17. Yungjohann, a young tile-setter from Upstate New York, described how he was ruined by loss. In 1906, he signed on as a rubber tapper in Acre, 2,500 km up the Amazon. It took a month to reach his patrão’s trading post and another 45 days to paddle up the Xapurí River with six other seringueiros (Yungjohann, 1989 [ca. 1920]: p. 23) . Within a few weeks, Yungjohann was alone. First, their guide cut their canoe loose and abandoned them. Then Yungjohann’s companions died of malaria. Seven months later, he tied his bales of raw rubber together with vines and rafted them downriver to his patrão―a journey of four and a half months. Still in debt, he picked up two new companions―an American named Joe and an unnamed Portuguese―and headed back to work (1989 [ca. 1920]: p. 47). Three months later, Joe succumbed to beriberi and the Portuguese man died even before reaching Yungjohann’s camp.
As the second year ended, he took his jebe fino (and Joe’s) back downriver. “When I arrived at the station,” he wrote (1989 [ca. 1920]: p. 49), “I had my rubber weighed and I learned in despair that I was still in debt.” Now he tried his hand at caucho. He and a Swiss partner bought 20 Indians from a Peruvian trader and after repaying their own debts collected enough caucho to make a small profit. Yungjohann decided to roll the dice again: he used his share to outfit other seringueiros in return for half of their future Hevea. When the time came, though, his seringueiros had disappeared. Most likely, they had sold their rubber to someone else and moved to another river. In 1916, after 20 years in Brazil, he gave up and returned to New York, almost as poor as when he had left. Multiplied by thousands, such incidents brought ruin to many medium and even larger traders in Iquitos and Manaus (Stanfield, 1998: pp. 47-48) .
4. Ethnic Reorganization after the Rubber Boom
In 1914, the market for Amazonian Hevea collapsed under a tsunami of plantation rubber from Asia18. By 1915, Manaus was becoming a ghost town as great fortunes disappeared overnight. In 1920, wild rubber supplied only 10% of the world market (Lawrence, 1931: p. 67) . Seringueiros flooded out of Brazil. Fifteen hundred kilometers upriver, caucho hunters hung on until 1930, when the Great Depression swept them away, too. Nearly 70 years after the quinine boom first began, Amazonia entered another period of major transformation.
In Peru, many caucheros abandoned the forests and settled along major rivers, where they produced cotton, coffee, cattle, rice, corn, sugarcane, ivory nuts, timber and eventually, barbasco (for fish poison) and rosewood oil. “Closely associated with the merchant houses of Iquitos, which financed their operations, fundo owners replaced rubber patrones as the new rural elite (Santos Granero & Barclay, 2000: pp. 163-164) . In outlying areas, Indians workers occasionally opted for isolation or “revert[ed] to a tribal way of life.”19 But elsewhere Indians joined an emergent class of hacienda laborers. “Some maintained an indigenous identity but no longer led a communal tribal life. Others had mixed indigenous identities but did not feel part of any specific Amerindian people, whereas still others became part of the increasingly larger mixed-blood stratum.”
Peter Gow (1999: p. 32) has provided a remarkable account of how Piro, Campa, Machiguenga, Cocama and Amahuaca on the Urubamba and Ucayali Rivers in Peru were transformed from isolated “tribal” groups in the early 19th century to “people of mixed blood” a century later. Before the Rubber Boom began, Piro lived as semi-independent bands trading between mission communities and infieles. “This trade…involved the exchange of forest produce such as medicines for steel tools…They also stole Machiguenga women and children for sale in the Andes.” Working as caucheros, however, “all of these people…came to live together, and to intermarry.” They settled around emerging haciendas and redrew ethnic affiliations along lines that anthropologists later found in the 1970s and 1980s20.
Another striking example of ethnogenesis or ethnic reorganization involves Kichwa-speaking people (Canelos runa) in Ecuador. After 1920, former rubber workers on the lower Curaray moved far upriver and settled near surviving estates in the Andean foothills21. There they often formed groups of intermarried families known as llacta, organized around a powerful shamans or headmen (Reeve, 2014) . But such communities weren’t free. Llacta formed specific alliances with one or another hacendado and also with local traders. According to Whitten (1976: pp. 235-236; see also Macdonald, 1999) , they supplied the hacendados with “free labor, particularly to clear forest areas… They brought pelts, gold, cinnamon, broom fiber in to the stores, and took out salt, candles, gunpowder, shot caps, dynamite, fishhooks, pots and pans… Credit was also extended [and] bound llacta groupings to particular [merchant] families.” In return, llacta members appealed to their patrons for help when problems arose under Ecuador’s Jim Crow-like racial laws. This was not an idle threat: from 1932 until around 1950, Indians in the Oriente were required to carry pasaportes orientales [“Eastern passports”] identifying their master and the reason for their travel. Otherwise, they could be forced to work on public projects or rented out to landowners who paid their fine.
Around the same time, 5000 indigenous Quijos runa workers abandoned the middle Napo and settled in upriver areas that had been deserted for a generation or two: Coca, Loreto, Ávila, Tena, Archidona and Baeza. In the past, travelers had noted the hostility and distrust that prevailed among runa groups in the foothills22. But a half-century of “forced labor, fractured clans, ruptured families and communities, [and] dispersion to unknown parts of the forest” had disrupted pre-existing forms of territorial organization (Cabodevilla, 1994: p. 143) . Like Canelos runa, Quijos people settled in extended families organized around a powerful shaman or leader. Dispersed along the upper Napo, they, too, faced constant labor demands. To assemble work crews for field labor, road construction or other tasks, for example, white officials appointed headmen with traditional-sounding titles: justicias, capitanes and guaynaros23. This system lasted into the 1960s, when landowners turned to raising cattle and no longer needed field hands. By then, many runa families had received land under Ecuador’s agrarian reform laws and also produced cattle or other cash crops. Faced with growing land shortages, their descendants migrated to other parts of the Oriente or to Lago Agrio, Riobamba, Quito and Guayaquil.
Did indigenous groups in Ecuador or Colombia evade the Rubber Boom (or minimize its impact) by hiding in the forest, as several recent researchers have suggested? It is certainly possible, but they must have remained the rare exception, not the rule. One often-cited case involves the Waorani, who play a large role in anthropological discourse (e.g., Rival, 2002 ; additional examples are cited in High, 2015, and Wasserstrom, 2016 ). Even after caucheros departed, Rival argues (2002: p. 38) , “the Huaorani, fierce isolationists, strictly confined themselves to the hinterlands” avoiding everyone else. Wilson & Yost (2001: p. 118) give another reason for their concealment. During the Boom, they write, “Waorani and others were captured as slaves and forced to work in unspeakable conditions.” As a result, all non-Waorani were viewed as cannibals. “What is unusual,” Wilson & Yost conclude, “is not that they chose to reject interaction with people who had been aggressive and untrustworthy; it is that they included everyone, including indigenous neighbors, in their definition of those who were not human.”
Yet historical records suggests a different interpretation. Waorani bands collected and traded with white caucheros into the 1910s, when they were attacked by rubber workers eager to occupy their homelands (Wasserstrom, 2016) . Later, these (mostly runa) workers abandoned the lower Curaray and Napo rivers, leaving Waorani without a ready source of tools and other supplies. Waorani bands then relocated to the forest margins and set up raiding camps about two days’ walk from the nearest runa settlement. Raiding along the frontier continued into the 1950s and in some places even later. This was “war in the tribal zone,” as Ferguson & Whitehead (1992) have described it, not isolation. Eventually, violence subsided when American missionaries offered tools and shotguns in return for peace.
One other instance of tribal warfare occurred in north-eastern Ecuador, this time in the upper Putumayo valley. During the Rubber Boom, most Cofán and Siona people moved to Capuchin missions across the Colombian border or joined large multi-ethnic work gangs harvesting caucho (Friede, 1952: p. 206) . In 1923, however, they began to filter back into Ecuadorian territory after a measles outbreak spread panic through the borderlands. There they bumped up against a small band of Tetete who had hidden from rubber bosses. By living in the swamps, Tetete had preserved their independence, but they were also critically short of tools. At first, they tried to barter with returning Cofán and Siona families―string bags and hammocks for machetes, knives and fishhooks. When that failed, they ambushed Cofán and Siona fishermen and stole their equipment. Sometime in the late 1930s, Siona raiders put a stop to the harassment (Wasserstrom, Reider, & Lara, 2011) . Eventually, three elderly survivors were found in 1966, although they were not seen after 1973.
As far as we know, the Tetete’s (ultimately unsuccessful) effort to evade caucheros remained unique in Ecuador and southern Colombia, even though it later held pride of place in the Amazonian imaginary. In contrast, Cofán and Siona people―who just barely survived the Rubber Boom―did not cut themselves off from the outside world. For the next 30 years, Cofán families panned for gold, which they traded with local landowners or merchants for shotguns, tools, beads and cloth (Holloway, 1932: p. 411; Robinson, 1979: p. 33) . Around 1940, Shell geologist Karl Goldschmid found them eager to work as guides and boatmen. Many of them spoke Spanish (Goldschmid, 2005: p. 174) . As Hill (1999: p. 751) has noted, “The Rubber Boom marked the final and definitive incorporation of indigenous peoples of lowland Ecuador into the political and economic life of the emerging nation-state.” Across the Colombian border, diverse groups―Cofán, Siona, Oyo, Coreguaje, Macaguaje and others―were increasingly absorbed into a mestizo population that flooded down the newly completed road from Mocoa to Puerto Asís.
5. Conclusion: “I Was Sold.”
In 1963, anthropologist Francisco Javier Beghin (1964) surveyed Kichwa-spea- king people in the Ecuadorian Oriente. Estimating their population at 25,000, he divided them into three categories: “progressive,” “stagnant” and “enumerated.” Progressives (around 15,000) grew crops for sale near the region’s spreading road system, where they increasingly contended with highland migrants eager to settle on their land. “Stagnant” families were scattered along the middle Napo between Ahuano and Coca, where they traded animal skins, chickens and plantains for knives, machetes, cloth and dynamite (which replaced barbasco for fishing). Beghin calculated their number at about 1000 “or maybe fewer.”24 Finally, “enumerated” families―Indians listed as debtors in a landowner’s account books―clustered around two haciendas near Tena, where they grew sugarcane and cotton or panned for gold. They included the woman who told my colleague that she had been sold in the 1960s. Another 5000 or so enumerated peons inhabited a dozen semi-abandoned fundos on the lower Napo.
Similar conditions existed in Peru. According to Bellier (1991: pp. 74-79) , “the 1920s brought an end to rubber collection,” but new patrones quickly replaced departing caucheros. These events fell hardest on groups like the Mai huna, who lived near major rivers. Along the lower Napo and Putumayo, Mai huna recalled the years between 1920 and 1940 as worse than the Rubber Boom itself (Bellier, 1994: pp. 38-45) . Well into the 1950s, Mai huna and their neighbors collected rosewood and vegetable ivory (tagua) or tended cattle as indebted workers. Other groups drifted into a marginal life of subsistence farming, periodic labor and unending debt―the lowest stratum of rural society. Not far away, Yagua people were also drawn into the expanding agricultural economy. “As soon as rubber fever subsided,” writes Chaumeil (1994: p. 211; see also Chaumeil, 1981) , “the Yagua successively became loggers, barbasco collectors and fur hunters for local bosses―mostly former caucheros―through a system of habilitación that still exists informally to this day.”
Such accounts should make us cautious about imputing historical isolation to groups that seem marginalized today. “All identities, without exception,” writes Scott (2009: pp. ix-xii) “have been socially constructed. [Tribal] peoples are best understood as runaway, fugitive, maroon communities who have, over the course of two millennia, been fleeing the oppressions of state making projects―slavery, conscription, taxes, corvée labor, epidemics and warfare.” He also warns us against taking up the dominant narrative about people who live in tribal zones, even if they themselves embrace it. “Civilizational discourses never entertain the possibility of people voluntarily going over to the barbarians, hence such statuses are stigmatized and ethnicized. Ethnicity and ‘tribe’ begin exactly where taxes and sovereignty end…” In these “shatter zones,” he argues (2009: p. 28), “the human shards of state formation” seek unoccupied land and safety. Almost identical patterns occurred in western Amazonia, where state formation ebbed and flowed with successive commodity cycles. By the 1870s, it would be difficult to say that any native group was truly isolated.
“Had it not been for the rapid decline of rubber gathering in lowland South America,” notes Hill, (1999: p. 753; see also Taylor, 1994) , “it is doubtful that any indigenous peoples of the southern lowlands would have survived into the twentieth century.” This was certainly true of rubber zones on the western frontiers, where once marginal areas suddenly became a source of great wealth. After the Boom moved on, different forms of ethnic organization took shape, including “a kind of ‘self-barbarization’ adopted by peoples whose location, subsistence, and social structure are adapted to state evasion” (Scott, 2009: p. x) . In Amazonia today, gold miners, illegal loggers and drug producers have mounted another round of resource extraction, displacing or eliminating the survivors of earlier cycles. Yet some things haven’t changed. In dimly lit corners of the region’s eco- nomy―for example, among Indian children working as household servants―you can still glimpse older forms of servitude. “The past is never dead,” wrote William Faulkner (1951) . “It’s not even past.”
This article is dedicated to the memory of Roger D. Casement and Bill Vickers. I would like to thank Brian Ferguson, Brooke Larson and David Stoll for commenting on the manuscript.
1Throughout this article, “western Amazon” refers to the tropical lowland regions of Bolivia, Colombia, Ecuador and Peru.
2According to Brass (1999: p. 230) , unfree labor systems seldom relied primarily on overt violence, because “continuous physical oppression is counter-productive and therefore uneconomic.” A similar point has been made by Malik (2016) . In the US, pre-Civil War chattel slavery remains a glaring exception, although it, too, gave way to sharecropping and other forms of compulsion (Daniel, 1972; Foner, 1988; Blackmon, 2008; Beckert, 2014; Smith, 2015) .
3In contrast to this view, many recent authors have argued for a resurgent essentialism that emphasizes long-term continuities of ethnic identity in spite of external pressures. For examples, see Rival (2002), Yashar (2005), Cepek (2012) and Erazo (2013) , among others.
4In this article, I use rubber terms that were common in the Spanish-speaking countries of western Amazonia, not Brazil. Jebe fino was extracted from H. brasiliensis; jebe débil from H. benthamiana and H. guaianensis; caucho and balata from Castilloa, Manilkara and related varieties. In most contemporary documents, caucho (sold primarily for industrial hoses, valves and other machine parts) and balata (for insulating undersea cables) are indiscriminately called caucho. With good management, H. brasiliensis might be tapped continuously for up to 50 years. H. benthamiana (Arana’s main crop) produced good harvests for five or six. Meanwhile, Castilloa and Manilkara trees were cut and drained into pits; it took as long as 20 years for them to regrow. Upriver from Iquitos, H. brasiliensis becomes scarce, but other varieties extend as far as the Andean foothills. For a discussion of rubber species and their distribution, see Santos Granero & Barclay (2000: p. 23) and Domínguez & Gómez (1990, pp. 82-85) .
5In Portuguese, loans (in cash or merchandise) from larger merchants to smaller ones, or from merchants to rubber collectors, were generally called aviamento. In Spanish, aviamiento often described loans among merchants or traders, while habilitación referred to debts (real or fictitious) incurred by rubber tappers. But usage varied from one place to another.
6 Rice (1903: pp. 401-418) described the beginning of a year-long trip to remote tributaries of the Napo.
7Caucho terminology can be confusing. Rubber merchants and patrones were often called caucheros, as were their workers. In context, however, the meaning is always clear: no document confuses white patrones with their laborers.
8After 1888, laborers in Peru were required to carry work books listing their contracts and patrones, like passes in South Africa. Anyone caught without a pass or who had not worked within six months was guilty of vagrancy and assigned to a cauchero. A few years later, Ecuador also required Indians in the eastern provinces to carry passes. As in Guatemala and other parts of Latin America, these measures persisted until World War II.
9 Santos Graner & Barclay (2000: p. 44) provide similar examples: “Contemporary sources are full of accounts of Ashininka leaders from the Ucayali River river raiding the interior Ashaninka and Asheninka; of Conibo headmen enslaving the Amahuaca and Ashaninka; and of Piro chiefs raiding the Ashaninka, Conibo, and Machiguenga… This increased inter- and intratribal violence to previously unknown levels.”
11Essential accounts include Casement (1913), Eberhardt (1913) and Hardenburg (1921) . Casement’s diaries can be found in Mitchell (1997, 2003) . For more recent historical accounts, see Collier (1968), Goodman (2009) and Stanfield (1998) .
12One additional case of widespread brutality against rubber workers became notorious: the Congo Free State, owned by King Leopold of Belgium. According to Hochschild (1998: p. 233; see also Morel, 1906) , nearly 10 million people―half the colony’s population―died from abuse. Why did Leopold’s resort to genocide? Like other rubber barons in Europe, he probably foresaw the impending end of wild rubber collection: it was widely discussed in newspapers and trade journals. But he was a spendthrift, littering Brussels with monuments and palaces while he looked for other colonies to buy. Unable to shift gears, he apparently chose the whipping machine to maximize short-term production while he could.
13A more detailed account of trade networks is provided in Wasserstrom & Bustamante (2015: pp. 1-18) .
14Designations like “upper,” “middle” and “lower” often reset at international borders. For example, Ecuadorians spoke of an upper, middle and lower Napo ending at Rocafuerte. For Peruvians, however, however, the entire Ecuadorian river was called “upper Napo.” I have followed the same practice here: thus there is an upper Napo in Ecuador and another one in Peru, depending on which country is under discussion.
15Arana also created a buffer zone to block upriver competition. In 1905-1906, he acquired a string of rubber posts extending from the middle Napo to the Caquetá and moved their workers to his other outposts (Domínguez & Gómez, 1990: p. 125) . This no-man’s-land (stretching five or six days’ travel below Rocafuerte) effectively defined the northern boundary of rubber collection on the Napo and Putumayo. According to Barclay (1998: p. 96) , most of the area had originally been depopulated during the 1880s as indigenous people on the middle Napo fled from caucheros who raided them for laborers.
16In Amazonian Ecuador, debt servitude, indentured labor and compulsory service to the provincial government were all imposed on native communities at least by the 1850s. See Jameson (1958: pp. 337-349); Orton (1871); Wiener (2011 : pp. 162-264); Hassaurek (1868) and Muratorio (1991) . As more settlers arrived in the 1870s and 1880s, they also spread highland-style concertaje (peonage) and racial ideologies.
17According to Barham & Coombs (1996: p. 45) , merchants who financed rubber collection faced astonishingly high losses: out of every 100 workers recruited in Pará, “75 would die, desert, or leave because of illness.” In addition, up to 25% of rubber shipments never reached them because of accidents or bad luck. Poorly cured latex might reduce deliveries by another 15%. And labor was expensive―eight times higher than in Asia―while supplies were largely inelastic (Barham & Coombs, 1996: p. 51; see also Campbell, 1923; Tully, 2011) . By extending credit on unfavorable terms, aviadores forced rubber tappers to take on a large share of these risks.
19 Johnson (2003: pp. 34-35) describes Matsiguenga families in the upper Urubamba watershed who withdrew into the forest. According to Santos Granero & Barclay (2000: p. 178) , “Indigenous peoples could certainly have retreated to the interfluvial areas and some did. The Cashinhua, Amahuaca, Matses, Uni, as well as segments of the Ashaninka, Ticuna, Yagua, and Urarina, opted to seclude themselves in out-of-the-way headwater or terra firme areas. But most of the indigenous peoples that had become voluntarily or forcibly engaged in commercial relations during the rubber era were dependent on market goods and opted to remain close to the region’s fluvial trading routes. To do so, they had to attach themselves to a fundoowner, given the absence of suitable free riverine lands.” For another excellent discussion, consult Gow (1991: pp. 44-56) .
20Writing of the Asháninka, Varese (2002: p. 17) emphasizes the “high degree of elasticity and flexibility in their social organization” that allowed them “to adapt to modern economic boundaries”― unlike the vast majority of groups that chose flight or isolation. “The Asháninka were able to survive,” he concludes, “thanks to favorable (external) historical and (internal) cultural circumstances.”
21Carta del Teniente Político de la Coca al Jefe Político del Cantón Napo, 2 de abril de 1923, Archivo de la Gobernación de Napo, Tena, Ecuador (hereafter AGN; documents are unidentified except by their title); R. Serrano, Jefe Político del Cantón Napo, al Director General de Oriente, 27 de marzo de 1926, AGN; Walter Palacios, Jefe Político del Cantón Napo, al Director General de Oriente, 25 de diciembre de 1926, AGN. See also Reeve (1988: pp. 18-34) . The largest of these estates was Hacienda Ila, owned by Carlos Sevilla. In the mid-1930s, according to his former manager (Peña Flores, 1992: pp. 39-40) , “we had more than 100 tambos [settlements] of indebted laborers with three or four families living in each one… All of the laborers on the left bank of the Anzu worked for D. Carlos Sevilla.”
22Population movements after 1930 are documented in AGN, Libros de Cuentas, 1938; also Samaniego & Toro (1939) . See Jiménez de la Espada (1998 [1927-1928]) for hostility among run a communities before the Rubber Boom.
23Oficio No. 2 del Jefe Político del Cantón Napo al Teniente Político de la Parroquia de Napo, 5 de junio de 1950, AGN. In 1946, the deputy commissioner in Puerto Napo reported that 4,000 free Indians lived in his district near Tena. Carta del Teniente Político de Puerto Napo, Rodolfo Rodríguez F., al Sr. Jefe Político del Cantón, 2 de agosto de 1946, AGN. On labor control, see Cuadrilla No. 24, 18 de enero de 1923, AGN; Carta de A. Ramírez al Jefe Político del Cantón Napo-Aguarico, 2 de marzo de 1923. Dozens of similar documents are located in AGN.
24Setting a pattern for later researchers, Beghin thought that “stagnant” Indians had probably lived there for centuries. In reality, they were descended from rubber workers on the lower Curaray who had moved up to the foothills after 1920. In the montaña, they survived mainly as debt peons working for two prominent hacendados, Carlos Sevilla and Sam Souder.
 Barclay, F. (1998). Sociedad y economía en el espacio cauchero ecuatoriano en la cuenca del Río Napo, 1870-1930. In P. García Jordán (Ed.), Fronteras, colonización y mano de obra indígena: Amazonía Andina (pp. 127-238). Lima: Pontífica Universidad Católicadel Perú.
 Beghin, F. (1964). Putumayo. Napo. Pastaza. Informe general de las investigaciones y estudios realizados sobre la Región Oriental por el etnólogo Francisco Javier Beghin. Quito: Instituto Nacional de Colonización.
 Bellier, I. (1994). Los Mai huna. Tucano occidental. In F. Santos Granero, & F. Barclay (Eds.), Guía etnográfica de la Alta Amazonía (Vol. 1, pp. 2-179). Quito: Instituto Francés de Estudios Andinos.
 Brown, M., & Fernández, E. (1991). War of Shadows. The Struggle for Utopia in the Peruvian Amazon. Berkeley, CA: University of California Press.
 Casement, R. (2013). Correspondence Respecting the Subjects and Native Indians Employed in the Collection of Rubber in the Putumayo Districts [the “Blue Book”]. House of Commons Seasonal Papers, 68 (14 February 1912-March 1913). London: Printed for His Majesty’s Stationary Office by Harrison and Sons.
 Chaumeil, J.-P. (1994). Los Yagua. In F. Santos Granero, & F. Barclay (Eds.), Guía etnográfica de la Alta Amazonía (Vol. 1, pp. 181-308). Quito: Instituto Francés de Estudios Andinos.
 Eberhardt, C. (1913). Slavery in Peru. Message from the President of the United States Transmitting a Report of the Secretary of State, and Accompanying Papers, Concerning the Alleged Existence of Slavery in Peru. Washington DC: Government Printing Office.
 Fifer, J. (1970). The Empire Builders: A History of the Bolivian Rubber Boom and the Rise of the House of Suárez. Journal of Latin American Studies, 2, 113-146.
 Friede, J. (1952). Los Kofán: Una tribu de la Alta Amazonía Colombiana. In Proceedings of the Thirtieth International Congress of Americanists (pp. 202-219). London: Royal Anthropological Institute.
 Hill, J. (1999). Indigenous Peoples and the Rise of Independent Nation-States in Lowland South America. In F. Salomon, & S. B. Schwartz (Eds.), The Cambridge History of the Native Peoples of the Americas. Vol. III. South America. Part 2 (pp. 704-764). Cambridge: Cambridge University Press.
 Jiménez de la Espada, M. (1998). Diario de la expedición al Pacifico llevada a cabo por una comisión de naturalistas españoles durante los años 1862 a 1865. Boletín de la Sociedad Geográfica (Madrid), 67-68, 1-246.
 Reeve, M. E. (1988). Cauchu Uras: Lowland Quichua Histories of the Amazon Rubber Boom. In J. Hill (Ed.), Rethinking History and Myth: Indigenous South American Perspectives on the Past (pp. 19-34). Urbana, IL: University of Illinois Press.
 Smith, S. (2015). Emancipating Peons, Excluding Coolies. Restructuring Coercion in the American West. In G. Downs, & K. Masur (Eds.), The World the Civil War Made (pp. 46-74). Chapel Hill, NC: University of North Carolina Press.
 Taussig, M. (1984). Culture of Terror, Space of Death: Roger Casement’s Putumayo Report and the Explanation of Torture. Comparative Studies in Society and History, 26, 467-497.
 Taylor, A.-C. (1994). El oriente ecuatoriano en el siglo XIX: “El otro litoral”. In J. Maiguashca (Ed.), Historia y Región en el Ecuador, 1830-1930 (pp. 17-68). Quito: Corporación Editora Nacional.
 Wasserstrom, R., & Bustamante, T. (2015). Ethnicity, Labor and Indigenous Populations in the Ecuadorian Amazon, 1822-2010. Advances in Anthropology, 5, 1-18.
 Wilson, R., & Yost, J. (2001). The Creation of Social Hierarchy. In R. B. Morrison, & C. R. Wilson (Eds.), Ethnographic Essays in Cultural Anthropology (pp. 107-137). Belmont, CA: Wadsworth Cenage Learning.
 Yashar, D. (2005). Contesting Citizenship in Latin America: The Rise of the Indigenous Movements and the Post Liberal Challenge. Cambridge: Cambridge University Press.