Tropical Agriculture Introducer: Christine Padoch

Abstract

This essay introduces two of Denevan’s articles that contributed profoundly to a reevaluation of Amazonian agriculture both in the present and in the past. Denevan changed the way swidden farming is perceived, especially the ways in which this common form of agriculture relates to forests. He showed swidden to be extremely complex and highly flexible, multistaged, largely invisible, and “hiding” within what appear to be natural forests. Perhaps even more significantly, Denevan directed Amazonian researchers to look beyond swidden and in doing so changed fundamental ideas about early Amazon agriculture, Amazon forests and forest management, and the nature of traditional and smallholder resource management throughout the tropics.

Introduction

Swidden agriculture seems among the most visible of farming methods as its more common name, slash-and-burn, graphically describes. Tree felling is easy to detect; falling trees can be seen, even heard, from far away. The forest gaps that swiddening creates stand out dramatically and remain distinct in satellite images for many years. Moreover, the fires that characterize swidden in most places are not only conspicuous, but especially in recent years, dramatic images of fires have served as the “proof” that swiddening is “primitive,” “wasteful,” and essentially criminal.

In the course of some groundbreaking research and with several provocative articles, William Denevan, however, convinced those who observe and study shifting cultivation that what we thought we saw, i.e., the “visible swidden,” is only a small part of what swiddening actually is and that we were overlooking most of what swidden systems contribute to Amazonian livelihoods and landscapes. Denevan’s Bora project profoundly changed the way swiddening is perceived and understood and how it relates to forests, especially in Amazonia (Denevan and Padoch 1988). He showed shifting cultivation to be far more complex than previously assumed and often “hidden” within what appears to be natural forests. And then, once we had learned all of that, and thought we understood the importance of managed swidden fallows, Denevan went on to convince us that even these “invisible” phases of the swidden cycle were historically both more and less important than we had believed. He revealed how the visibility and drama of slash and burn continued to blind us to the larger story of Amazonian forest management and of the management of Amazonia’s rivers and soils, animals, and plants.

This essay introduces two of Denevan’s articles (1984, 1992) that contributed profoundly to these reevaluations of how Amazonians produce food, fiber, medicines, and many other goods both in the present and in the past. Based on these and similar writings, new discoveries about present-day tropical forests and peoples, waters and soils, and new insights into the processes of change that produced them continue to be made in both Amazonia and elsewhere in the tropics. Denevan’s research truly transformed how we look at and think about – or should look at and think about – tropical resource management around the globe and throughout history.

Saving Swidden

In 1957 FAO Staff, in its forestry journal Unasylva, published an “appeal…to governments, research centers, associations and private persons who are in a position to help” to join a coordinated research effort on shifting cultivation or rather to help with “analyses of the effects of shifting cultivation on soils and forests, as well as with proposals for improving the situation. This entails the study of all facets of the problem. FAO has, therefore, the intention of mobilizing the contributions of as many scientists as possible to help solve this problem in all interested countries” (FAO Staff 1957, 9).

The Unasylva article did not mince words; the “problem” it asserted was that “shifting cultivation, in the humid tropical countries, is the greatest obstacle not only to the immediate increase of agricultural production but also to the conservation of the production potential for the future, in the form of soils and forests” (FAO Staff 1957).

Furthermore, according to the article, shifting cultivation or swidden was “not only a backward type of agricultural practice…[but] also a backward stage of culture in general. In all respects it corresponds to the Neolithic period through which humanity passed between the years 13,000 and 3,000 BC, considering that the substitution of iron tools for polished stone has made no substantial difference in the way of life” (FAO Staff 1957). Moreover, FAO proposed a number of “points of attack.”

Uninformed and wholesale condemnation of swidden agriculture and of its cultural concomitants was of course not surprising in 1957, nor is it now a thing of the past. What is remarkable, however, is that in the very same year as its call to eliminate shifting cultivation, i.e., 1957, the FAO published Hanunoo agriculture , anthropologist Harold Conklin’s path-breaking study of shifting cultivation in the Philippines (Conklin 1957). Conklin’s work was not the first piece of research to take a careful look at what swidden farmers actually did and what values they produced not only for themselves but for broader landscapes and societies. Hanunoo Agriculture did, however, profoundly affect the discussion of swiddening. It stands to this day as a pioneering work. To those willing to read it and learn, Hanunoo Agriculture demonstrated that shifting cultivation is in many instances complex, diverse, and developed and an environmentally and economically rational way to make a living in the humid tropical uplands. Conklin’s work spurred much further research on shifting cultivation in Southeast Asia and elsewhere.

As in Southeast Asia, studies of swiddening in Amazonia have a long history. Much of what has been said about Amazon shifting cultivators is equally marred by misunderstanding, misjudgment, and condemnation. Amazonian swidden studies , however, often have had a different focus and emphasis than their Southeast Asian counterparts. They have emphasized issues such as carrying capacity, labor input, and energy efficiency and human dietary issues. Indeed, all shifting cultivation systems may have more in common with each other than not. However, the chacras, roças, or swidden fields of the lowland Amazon, dominated by tall, vegetatively reproduced cassava (manioc) plants that grow in the shade of even loftier banana plants and papayas and are underlain by a great diversity of herbs and tubers, indeed look very different from Hanunoo upland rice fields that climb the steep hills of Mindoro Island in the Philippines. Furthermore, these different visible structural features also signal different labor demands, management strategies, crop patterns and yields, and often very different long-term trajectories.

Perhaps, the most important issue that Conklin and many later studies on shifting cultivation addressed is the supposed wastefulness and destructiveness of shifting cultivation. Long before “deforestation” was swiddeners’ principal purported crime, accusations that shifting cultivators were “nomadic” and that their farming wasted the soil resources that were needed to avert future famines (FAO Staff 1957, 2) were made. The “shifts” of annual crops from one wooded site to another that characterize swiddening were simply attributed to heedless overuse of soils that made shifting imperative. The most important and memorable change in these views that followed Hanunoo Agriculture and other important works was probably the argument that the shifting of plots was an orderly and planned process, based upon a cycle of “natural fallow” and recovery, punctuated with periods of agricultural production, rather than the misplaced assumption that shifting sites entailed a sequence of soil exhaustion, forced abandonment, and endless migration to new forests. Conklin suggested that the Hanunoo were not unlike other more “advanced” farmers who rotated crops or fallowed fields, but swiddeners rotated fields rather than crops, and their fallows were woody and their cycles were protracted.

Thus, swidden cultivation was saved – at least among those willing to take a careful look – from the ignominy of condemnation as wasteful and its practitioners as ignorant nomads. Shifting cultivators indeed kept cutting forests, but when they shifted, those very same forests would regenerate and the soils be restored in a potentially endless cycle of clearing and regrowth. Many of us, who studied shifting cultivation, represented those shifts in neat diagrams of rectangles and arrows that recreated swidden cycles. The arrows brought the farmed and fallowed plots back to a point zero where the cycle of swidden and fallow would begin yet again. That drawing was later repeated, graphically illustrated, and interpreted by others – including many who had never actually been in a shifting cultivation field. But perhaps, that picture was just a little too neat, too regular, and too rotational. The impressive complexity, flexibility, and diversity of swidden systems was lost; many of us, even against our better judgment, represented swiddening as altogether too predictable and equilibrated, too certain and unchanging, and too simple.

Revealing the Invisible

Twenty-five years after Conklin’s and other studies helped us to understand why swiddeners abandoned their fields after a year or two of cropping, Denevan and the Bora project, represented here by “Indigenous Agroforestry in the Peruvian Amazon” (Denevan et al. 1984; also see Denevan and Padoch 1988), convinced us that swidden fallows were not really abandoned at all, or at least that “abandonment,” if it actually occurred, was not an event but a complex process. The paper showed that what observers of shifting cultivation had long called “abandonment” was rather a change of management: extended, complex, contingent upon a plethora of observed conditions, and largely invisible to the eyes of outsiders. It added yet another layer of intricacy to a system that we had just recently learned was not at all primitive. Around the time that the Bora project took place, agroforestry had become an area of scientific focus and promise since it was deemed to be, among other things, an “alternative to slash and burn.” The characterization of Bora swidden-fallow management as a form of agroforestry raised it to the level of an advanced, sophisticated, and up-to-date, improved way to farm and manage forests rather than an anachronism.

The Bora project grew out of previous research done by Peruvian agronomist Salvador Flores Paitán in the Bora community of Brillo Nuevo along a tributary of the Amazon River in Peru. Flores Paitán had been investigating Bora patterns of managing young secondary forests by enriching them with important economic plants such as coca and a broad array of fruits. The project headed by Denevan, however, helped put the forest management into a swidden context and then placed the swiddens into a broader forest landscape. No longer was a swidden field seen as abandoned when annual crop harvests ceased; nor did the fields then rotate through a sequence of natural succession that restored them to their initial conditions. Swiddens were now places where farmers made a complex series of decisions through the years and throughout the plot. Just as there was no distinct break between the farm phase and the forest phase, there was also a complex blurring of where management began and ended spatially, with edges gradually being left to something close to “natural” regrowth, while other parts of the plots were manipulated for even longer periods. The individual swidden plot therefore did not have a history distinct from the broader landscape matrix within which it was found. Swiddening became the management of a complex, anthropogenic, uneven-aged landscape with, at best, indistinct edges. It was a place where a great variety of plant and animal species were managed, none of them quite detached from the forest, and none of them quite “natural.” There was an identifiable sequence from an “original” vegetation with some economic plants present, to a swidden with many food and other annual and perennial plants, and on to an orchard fallow agroforestry phase combining managed economic plants and natural vegetation. Finally, what developed was a forest fallow in which economic plants were fewer than in the swidden, but still present in greater numbers than in an original forest.

The Bora project did not emphasize the “visible swidden” of annual crops and burning slash and the essential regularity of the swidden cycle. Bora swiddens and swidden-fallows were messy, visually, ecologically, and conceptually. Some areas were managed far longer than were others. All areas had somewhat distinct trajectories, and the effects of active and more passive management, and of what was “natural” in the forest, were blurred, confusing, and essentially invisible.

The Bora of Brillo Nuevo, while doubtless distinct in some of their resource management practices, is not the only Amazonians – be they officially regarded as indigenous or not – to manage forests for food and other goods in such complex ways. Ribereño and caboclo peasant farmers up and down Amazonia’s rivers have adapted swidden-fallow management systems to a variety of particular spatial, ecological, and economic conditions. Beyond Amazonia and going back to the classics of shifting cultivation and related land use systems in Central America, Southeast Asia, and Africa, references have long been made to managed swidden fallows (including in Hanunoo Agriculture itself), but these were little emphasized and under-appreciated until quite recently. “Indigenous Agroforestry in the Peruvian Amazon,” published in the international journal Interciencia, helped to launch a great many studies around the tropics that took a closer look at what really happens in swidden fallows. Although the exact influences may be difficult to trace and intellectual debts often are unacknowledged in writing, those intercontinental and intergenerational transfers of insights, perspective, and focus have been important for many scientists researching both Asian and Amazonian systems.

Discovering the Vanished

Swiddens and swidden-fallows not only still provide the products with which smallholders around Amazonia feed their families and supply urban markets, but they also hold important clues to the contours of far earlier resource use systems in the South American tropical lowlands. These systems are now concealed by the upheavals of a particularly violent history of invasion, destruction, and erasure and the transformations wrought by time and rivers. They are also obscured by layers of false assumptions, conceptual blindspots, and prejudices. Today’s practices of plant, animal, land, and water management and manipulation doubtless are legacies of the past, albeit transformed, and the configurations of present-day forests are archives of the resource management technologies used by earlier Amazonians.

Combining the ethnographic and ecological clues and insights with a broad knowledge of what historical data exists, as well as some detective work and deductive reasoning, in “Stone vs metal axes,” Denevan arrived at an even more transformative idea than in the Bora studies. In the paper, he made a shocking statement: “shifting cultivation, as an ancient practice in Amazonia, seems to be a myth” (Denevan 1992, 161).

This sentence delivered a jolt to all of us who had long accepted and frequently repeated that shifting cultivation is the basic way in which tropical forests around the globe have been managed for food and other human needs by agriculturists. Despite evidence to the contrary, we have believed and even insisted that all other production systems we observed – the house garden, the managed fruit grove, and the intensively farmed plot on the riverbank – were important to village economies and ecologies but were fundamentally just add-ons, late stages, or variants of the swidden. Making gaps by cutting trees, and then (usually) using fire to clear the spot for planting crops, seemed primal, reasonable, and unquestioned – until Denevan called it a myth. The essential production system, he argued in the article that follows, was the carefully selected site, and the managed, transformed, transplanted, weeded, and manipulated garden and agroforest. The gap that was made using slash-and-burn techniques may have been the supplement, the “early stage” of the manipulated secondary forest, and the rare add-on.

Several recent archeological discoveries, including Heckenberger’s “garden cities” on the upper Xingu (Heckenberger et al. 2008), appear to corroborate Denevan’s provocative idea of extensively managed woodlands and some more intensively managed permanent plots, rather than distinct shifting farms and forests. Amazon forest vegetation gives up its secrets only with a good deal of work and even more good luck. However, the vision of Amazonia’s past suggested in “Stone vs metal axes” seems more plausible with each new discovery. Every added piece of evidence returns some lost history to Amazonia’s forests, lands, rivers, and the people and societies that have seen their histories destroyed and denied. In other regions where local communities and histories have fared somewhat better, there are other examples that suggest that technologies other than swiddening prevailed in times and places where metal tools were scarce. For instance, in interior Borneo, many complex resource management technologies of a variety of cultural groups are also commonly reduced to and decried as “slash and burn.” However, histories of earlier times and agriculture done with virtually no metal tools and largely without resorting to swiddening are also found. The Lun Dayeh of far interior Kalimantan and Sarawak, for instance, recounts that just a few generations ago, stone and wood implements were widely used instead of metal, and trees were rarely cut to make fields (Padoch 1985). In that case, farming was largely done not on the wooded hills but on broad, flat, and well-watered sites that made farming of rice in permanent pond fields the principal mode of food production. These specifically selected production areas were managed by digging, flooding, and weeding and other technologies that required little use of metal tools.

We may never find conclusive proof that would convince everyone that the Amazonians once supported substantial populations without cutting large swathes of forests, but rather by a broad repertoire of skills, technologies, and organization that created not only a wooded but a “humanized landscape” that also produced a great volume of necessary and desired foods, fibers, and other goods. The proofs lie in little researched local resource use technologies that still exist but are largely disappearing and in forests that are being replaced by large monocultures before they have been well understood. However, we do have some powerful clues and insights that have pointed us in the right direction, and new discoveries are doubtless still to be made.

Those studying resource management in Amazonia as in other tropical forest areas owe a debt of gratitude to the scholars who were and continue to be careful, perceptive, and patient observers of production systems – such as shifting cultivation – that are challenging and confusing at best and oftentimes the objects of condemnation. Denevan’s writings including the two articles featured here show unambiguously that he is a masterful observer of what is before him, a critical reader of what has been written earlier, and a perceptive listener to what there is to hear. However, these two articles also show that Bill Denevan’s contributions go well beyond these extraordinary gifts of observation and analysis. Denevan has seen well beyond what is visible, what is immediately perceptible, and what is clearly written on the page.

His major contribution, as the two articles that follow clearly show, is that he has perceived what was hidden, revealed what was invisible, and retrieved what was lost. And then through his research, teaching, and these articles, he has taught all those of us who take the time to read them and to learn from them to also look for and at last to see what is hidden and invisible and lost.

6.1 Stone Versus Metal Axes: The Ambiguity of Shifting Cultivation in Prehistoric Amazonia

Original:

Denevan, W.M. 1992. Stone vs metal axes: The ambiguity of shifting cultivation in prehistoric Amazonia. Journal of the Steward Anthropological Society 20: 153–165. Reprinted by permission of the Department of Anthropology at the University of Illinois, Champaign-Urbana.

Abstract Clearing forest with stone axes is a very laborious procedure, even with the assistance of girdling, tree falls, and fire. Experimental research in Amazonia and elsewhere confirms this. Iron or steel axes are many more times as efficient. Consequently, Denevan argues, shifting cultivation, which requires frequent clearing of forest, was uncommon in pre-metal axe times. Instead, cultivation was permanent or semi-permanent, maintained on even poor soils by the incorporation of ash, charcoal, and organic material. When metal axes were introduced by Europeans, forest clearing became much faster and less labor demanding, and shifting cultivation became more feasible than intensive cultivation as long as mature forest was available. Thus, the use of metal axes resulted in significant reduction of cropping frequency and greater deforestation.Keywords Amazon · Ax use · Forest clearing · Pre-European agriculture Shifting cultivation

Abstract Clearing forest with stone axes is a very laborious procedure, even with the assistance of girdling, tree falls, and fire. Experimental research in Amazonia and elsewhere confirms this. Iron or steel axes are many more times as efficient. Consequently, I argue, shifting cultivation, which requires frequent clearing of forest, was uncommon in pre-metal axe times. Instead, cultivation was permanent or semi-permanent and maintained on even poor soils by the incorporation of ash, charcoal, and organic material. When metal axes were introduced by Europeans, forest clearing became much faster and less labor demanding, and shifting cultivation became more feasible than intensive cultivation as long as mature forest was available. Thus, the use of metal axes resulted in significant reduction of cropping frequency and greater deforestation.

Keywords Amazon · Ax use · Forest clearing · Pre-European agriculture · Shifting cultivation

Introduction

The mastery of the forest by man requires no axe. –Sauer 1958, 108

Donald Lathrap had remarkable insights into Amazonian prehistory, the interaction between subsistence and settlement, and the position of the Amazon as center stage rather than backwater of South America. One of his major interests was the distinction between the cultures and economies of the densely populated floodplain (várzea) and those of the sparsely populated uplands (terra firme) (Lathrap 1970, 1977). A key component of this story is the stone axe and its efficiency for clearing forest. The presence of stone axes in archaeological sites is probably indicative of agriculture (Lathrap 1970, 62–63), but what kind of agriculture? Don would have appreciated what I have to say here, but I am sure he would have had some caustic but constructive comments.

Ethnologists and ethnohistorians have generally portrayed surviving indigenous hunters and gatherers (foragers), shifting cultivators, and other traditional economies as representative of prehistoric food production systems. Even where such groups have clearly undergone considerable acculturation, it has been suggested that their food-getting ecologies are essentially intact and of long standing, despite changes in crops and tools.

This perspective, however, is coming under increasing attack, as witness the debate over the authenticity of the Tasaday in the Philippines (Headland 1992), and the revisionist view of the !Kung of the Kalahari by Wilmsen (1989), who argues that the !Kung are far from living relics of stone age hunters and gatherers. Few groups anywhere have been isolated from the world economy, directly or indirectly.

In lowland South America, we find tribes such as the Nambiquari, Ache, Héta, Sirionó, Yuquí, and Yora which until recently subsisted from hunting and gathering with little or no cultivation. However, rather than being remnants of pre-agricultural societies, it now seems that most are refugee agriculturalists who have fled from more powerful tribes, from Europeans, or from demographic pressure in the floodplains, an argument made in 1968 by Lathrap and even earlier by Claude Lévi-Strauss (1963); and also see Bailey et al. (1989, 65–66).

More recently, archaeologist Anna Roosevelt stated that in Amazonia “theories about pre-Conquest subsistence cannot be tested with ethnographic data” and that “present-day Indians’ resource management modes may not be representative of prehistoric ones” (Roosevelt 1989, 31). Colchester (1984, 311) provides further emphasis: “it is time that we started examining Amazonian societies in terms of the recent radical transformations that have occurred and that are occurring in their technological, demographic, and economic bases…” Furthermore, most surviving Indians are located in the terra firme high forests of the interfluves, where resource conditions (soils, game, and fish) are relatively poor, whereas most prehistoric Indians were located in the resource-rich floodplains on adjacent bluffs.

Our knowledge of indigenous adaptations to the terra firme, adaptations that are now being heralded as instructive for successful rainforest utilization, is based primarily on present-day observations. This model is characterized by short cropping/long fallow shifting cultivation; low-protein crop staples (manioc, sweet potato, plantain); small, temporary settlements (ca. 10–100 people); and very low population density (below 0.5 per km²). Productive agroecological techniques are common, such as soil improvement, fallow management, and polycultural plantings. Considerable debate exists over the reasons for this pattern which applies to most tribes,¹ particularly over whether game (protein) scarcity is the key limiting factor (Denevan 1971, 1992, 208–209; Hames and Vickers 1983).

I believe that the short cropping, long-fallow shifting cultivation pattern is primarily a post-conquest development, reflecting the shift from forest clearing with stone axes to the much more efficient iron or steel axes, and that in aboriginal times forest clearing was too labor intensive to be a common or frequent agricultural strategy. I will first compare tree cutting with stone and metal axes. Then, I will discuss the significance of this for prehistoric Indian agriculture in Amazonia.

Tree Cutting

There is some data available on the technology and the differential labor involved in clearing forest with stone and metal axes in Amazonia, in particular, from Robert Carneiro (1974, 1979a, b), who conducted experimental research with the Amahuaca in eastern Peru, the Kuikuru in the Brazilian Amazon, and the Yanomami in southern Venezuela. Felling trees with stone axes was clearly time-consuming, difficult work.

Carneiro (1979b, 69–70), for example, calculates that a 24-inch (61 cm) diameter tree of moderate hardness could take from 11.7 to 14.4 hours to fell with a stone axe versus 0.52 hours with a steel axe. Felling times of course vary with the type of axe, arm strength, cutting technique,² trunk thickness, and hardness of wood. The ratio of felling time, stone to steel, increases progressively with trunk size. The ratio is only 10 to 1 for a 6-inch (15 cm) diameter tree; it is 23 to 1 for a 24-inch (61 cm) diameter tree; and it is 32 to 1 for a 48-inch (122 cm) diameter tree, or 115 hours vs 2.4 hours. Differences in felling times between stone and metal axes are much greater for hardwoods than for softwoods: “holding diameter constant, a tree twice as hard as another [density or specific gravity] will take twice as long to fell.” This is for a steel axe. The difference for a stone axe would be even greater (Carneiro 1979b, 62).

Hill and Kaplan (1989; Kaplan 1985), working with Ache Indian in Paraguay and the Yora in Peru, confirmed Carneiro’s hypothesis that the rates of clearing times increase disproportionately with increasing tree diameter and tree hardness, with a significantly greater rate of increase for stone compared with steel axes. The hardness, however, had a much greater effect on stone axe felling times than did tree size: “For hardwoods, the time cost for stone-axe clearance can be 60 times greater than for metal tools” (Hill and Kaplan 1989, 331). Overall, the average efficiency ratio was about 10 to 1, stone to steel.

For a family plot of 1.7 acres (0.7 ha) of trees of mixed size and hardness, Carneiro (1979b, 71) calculated a total of 1,229 hours for clearing with a stone axe versus only 64 hours with a steel axe, a ratio of 19 to 1. The former equals 246 five-hour work days, which is simply not tolerable. He asks how swidden clearing could be done then? His answer is that labor time was reduced with the assistance of trunk burning, girdling or cutting a ring through the cambium layer, tree falls to knockdown additional trees, and leaving the largest trees standing. He calculates an average efficiency ratio of only 7 or 8 to 1, stone to steel, using auxiliary techniques, and 10 to 1 if all trees over 2 ft. (61 cm) in diameter are left standing.³ Kaplan (1985), however, found that the multiple tree-fall technique did not reduce clearance time significantly. Killing and deleafing trees by girdling and burning the base of the trunk leave the trees standing, but will bring in sunlight to some of the adjacent ground surface.

Another consideration is the availability of stone for axe heads. It can take several days to make a stone axe and hours to sharpen one (Kozák et al. 1979); proper stone sources may be far away requiring long treks or trade (Denevan 1966, 47–48). Axe heads dull or break and shafts break.⁴ Axes are lost or stolen. The rapidity by which tribes shifted to metal axes when available, their struggle to obtain them, and the major role of metal axes in trade are well known, reflecting their great saving of labor, as is related in “The revolution of the ax” by Alfred Métraux (1959).⁵

Agricultural Implications

The inefficiency of the stone axe has dramatic implications for prehistoric agriculture in Amazonia. Several anthropologists have suggested this without pursuing it (Colchester 1984; Hill and Kaplan 1989). Kaplan (1985), in an unpublished paper, presented the hypothesis that “aboriginal farmers, particularly in interfluvial regions, were highly selective regarding their choice of potential gardening sites and that as a result the distribution of forest types placed important constraints on settlement pattern and subsistence practices throughout the Amazon basin.”

Sites for fields would have been sought where the vegetation lacked large hardwood trees and was dominated by small softwood trees, essentially secondary vegetation such as fallow-field regrowth, or along streams, or sites disturbed by tree falls and landslides. The Machiguenga (Peru) say that when they had stone tools, settlement was concentrated along small streams⁶ where clearing for gardens was easiest (Hill and Kaplan 1989, 332). Even today, they often clear fields from thickets of giant bamboo (Baksh and Johnson 1990, 205). The Yora’s (Peru) reliance on foraging apparently reflected limited availability of metal axes (Hill and Kaplan 1989). Allan Holmberg (1969, 272) reported that the Sirionó gave much more attention to gardening and became more sedentary as soon as he provided them with steel hatchets.

The Yanomami use of stone axes for clearing and the impact of the introduction of steel axes in the twentieth century are described by Colchester (1984). Secondary vegetation and stands of soft-stemmed musaceous species were sought for swiddens because of greater ease of clearing with stone axes than was mature forest, even though more labor is required for weeding in secondary vegetation. Plots were small, the larger trees were not felled, and treking for game and wild plant foods was of major importance for subsistence: “The Yanomami of the seventeenth century were interfluve foragers, who supplemented their subsistence with the cultivation of small plots, widely dispersed about their foraging territory” (Colchester 1984, 308). Early explorer accounts support this pattern. With the introduction of metal axes, the Yanoama (Yanomami) changed from a foraging economy supplemented by agriculture to an agricultural economy supplemented by foraging, with larger fields and villages and less mobility (Colchester 1984, 310). Likewise, Machiguenga Indians said that in the past when it was difficult to obtain steel axes their gardens were much smaller, and they relied more on forest products for food (Johnson 1977, 164).

I do not argue that clearing interfluve forest was rare in pre-metal axe times, but it was probably more restricted and much less frequent than with tropical forest tribes today, most of whom clear new fields every two or three years. Sites were undoubtedly more selective, based on ease of clearing, and once a field was established, it was probably maintained in cultivation as long as possible. It was likely that less labor was required to combat weeds (probably suppressed by controlled shade) and other pests and to use soil maintenance techniques than was required to establish new clearings. However, data is needed to confirm this.

The argument here is hypothetical, as there is little archaeological or early historical evidence on the nature of prehistoric terra firme agriculture, and there is no physical evidence for any form of prehistoric shifting cultivation. There were probably pockets of fairly intensive farmers, mainly along small streams. Overall populations were probably low, but possibly larger than scholars have believed, including myself (Denevan 1992, xxv–xxvii; Meggers 1992). In contrast, on the floodplain and adjacent levees, there was no vegetation to clear or only easily cleared vegetation, and the stone axe was less a liability. Soils were fertile and wildlife resources rich. Fields did not “shift,” and populations were dense. Also, those savannas where the soil fertility and drainage were either not severe or could be managed could have been attractive to permanent farmers given that there were few or no trees to clear (e.g., Denevan 1966, 94–95; Posey 1985, 140–144).

We do know that hardwood forests were cleared for agriculture elsewhere by relatively dense populations using stone tools, as in Yucatán, in western Central America, and in Europe. However, these were areas with soils much superior to those of Amazonia, so that fields could be cropped for numerous years and did not have to be cleared frequently. The considerable labor involved in clearing with stone tools thus could be tolerated. A recent study by Doolittle (1992) argues that prehistoric shifting cultivation in eastern North America was less common than permanent fields.

Conclusions

I am suggesting that shifting cultivation in prehistoric Amazonia was uncommon because of the inefficiency of the stone axe, especially in the mature, high, hardwood forests of the terra firme. Indian shifting cultivation today has a short cropping period, reflecting poor soil, weed and pest invasion, game depletion, and social friction, but it is made possible by the steel axe which makes clearing new plots a relatively easy process – a matter of a few weeks to create a field large enough (0.5–2.0 ha) to feed a family.

Indian shifting cultivation as we know it is the product of the steel axe and also the machete. What then was the nature of prehistoric high forest agriculture? We do not know and may never know. However, there are several possibilities:

• 1. House Gardens : permanent plots of mixed annuals and perennials around the house, with careful weed control and soil management using household refuse for fertilizer. Lathrap (1977), in his classic article “Our father the cayman, our mother the gourd,” maintained that the earliest agriculture in Amazonia was carried out in such house gardens.

• 2. Intensive Swiddens : located on sites where tree clearing was relatively easy, such as naturally disturbed or old field plots with young secondary growth of softwoods. A present-day example of such fields would be the highly diverse or polycultural swiddens described by Harris (1971) for the Waika (Yanomami) of the Upper Orinoco, which are cultivated for up to six years. Such fields contrast with the monocultural swidden dominated by a single species, usually manioc, which is the common form of tropical-forest Indian field today (Beckerman 1983), even for the Yanomami (Hames 1983, 18–19). Most current monocultural fields are only used for 1–3 years. Beckerman (1983, 4–6) gives several reasons for the monocultural field, but he does not consider the role of the steel axe in making short-lived swiddens feasible.

• 3. Agroforestry : forest manipulation via intentional and unintentional planting and management of crops along trails, campsites, fallow swiddens, and other activity areas (“forest fields”) (Denevan and Padoch 1988; Posey 1985).

These three models of terra firme agriculture with a stone axe technology in reality were likely manifested by numerous transitional forms, varying with habitat, mobility, time, and demography. These activities, combined with foraging, contributed to the creation of anthropogenic forests, or semi-managed forests, with a larger than natural number of useful plants present – wild, semi-domesticates, and domesticates. The Amazon forest was not pristine in 1492, nor is it today. Probably, all of these forms of agriculture and agroforestry were present in the terra firme, in a mosaic of variable population densities that may have included sectors of sparse hunters and gatherers in the more difficult forests and larger semi-permanent populations where vegetation was easily cleared.

We do know from archaeological and other evidence that there were some substantial populations in the terra firme forest. Large sites have been seen but not studied. Reports of the Jívaro uprising in 1599 mention mobilization of over 20,000 warriors (Harner 1972, 21). Ethnohistorical accounts indicate Kayapó settlements periodically numbering over 1,000 (Posey 1987, 139, 147). The subsistence base for such numbers is unknown.

The difficulty of clearing mature tropical hardwood forest with stone axes at least partially explains the dramatic differentiation between usually scattered terra firme settlement and dense riverine settlement with large, permanent villages. This is not to say, however, that poor soils and limited animal protein are not contributing factors. Involved here is a major debate in Amazonian cultural ecology, and it has yet to be resolved.

The adoption of metal axes and machetes in the New World was generally very rapid where Europeans were present. According to Hans Staden (1928, 74, 90), who lived with the Tupinambá, iron tools were an important trade item on the Brazilian coast as early as 1554.⁷ More remote regions obtained metal tools indirectly through trade and raiding, probably on an irregular basis. Isolated tribes continued the use of stone axes well into the twentieth century, although few still do so. This raises a question: how was agriculture and associated foraging and settlement affected when people at times had metal axes and at other times still had to depend on stone axes, or when some farmers in a village had metal axes while others did not? There seems to have been little reporting on this. An “overnight” conversion from stone to steel, as reported by Holmberg (1969, 268) and others, was probably unusual.

There are other questions that need to be pursued. How effective were the iron axes that were introduced in Colonial times?⁸ Iron axes must have been less effective than steel axes, but most of the experimental data from Carneiro and others is for steel axes. Also, to what extent were metal axes present in Upper Amazonia in prehistoric times? Lathrap (1970, 178) found bronze axes on the Río Pachitea and Río Pisqui in eastern Peru, clearly traded from the Andes. These were small, however, and probably were not used to cut down large trees. Finally, what was the significance of the stone axe for agriculture elsewhere in tropical America? Gordon (1982, 57–61) discusses the use of stone and metal clearing tools in Central America, including the impact of the machete on forest species manipulation. He notes that: “clearing wet evergreen forest without metal cutting tools would clearly have been a slow and laborious process.” He also believes that polycultural milpas were an integrated component of anthropogenic forests.

To conclude, shifting cultivation, as an ancient practice in Amazonia, seems to be a myth. There is no evidence for it. At best, it was rare, at least in short-cropping cycle form. It is not logical, given the stone axe. It is therefore a relatively modern adaptation resulting from the introduction of the metal axe. What is the significance of this, if valid? Certainly, it tells us something about pre- and post-Columbian adaptation and the impact of European technology: “Production and social organization were altered as each settlement chose a particular method for acquiring manufactured goods, particularly iron tools” (Golob 1982, 269).⁹ However, stone axe technology also tells us that the Amazon forest can be farmed successfully and sustainably with minimal destruction by means other than shifting cultivation, which is one of the main instruments of forest destruction today.

6.2 Indigenous Agroforestry in the Peruvian Amazon: Bora Indian Management of Swidden Fallows.

Original:

Denevan, W.M., J. Treacy, J. Alcorn, C. Padoch, J. Denslow, and S. Flores Paitán. 1984. Indigenous agroforestry in the Peruvian Amazon: Bora Indian management of swidden fallows. Interciencia 9(6): 346–357. Reprinted by permission of Interciencia.

Abstract A distinction has long been made between swiddens (shifting or slash and burn cultivation) and fallows in which crops are abandoned to forest and soil recovery. However, our interdisciplinary team found that for the Bora in Peruvian Amazonia there is a continuity from field to fallow to managed forest to mature forest. The early fallows contain a large number of useful species mixed with forest species. The “enrichment” can be both intentional and accidental results of human activity as well as spontaneous. Other research has indicated that such fallow agroforestry is not unique but occurs elsewhere in Amazonia as well as in tropical Africa and Asia. For a more detailed study of Bora agroforestry, see Denevan and Padoch 1988.Keywords Agroforestry · Amazon · Bora · Fallow management

Abstract A distinction has long been made between swiddens (shifting or slash and burn cultivation) and fallows in which crops are abandoned to forest and soil recovery. However, our interdisciplinary team found that for the Bora in Peruvian Amazonia there is a continuity from field to fallow to managed forest to mature forest. The early fallows contain a large number of useful species mixed with forest species. The “enrichment” can be both intentional and accidental results of human activity as well as spontaneous. Other research has indicated that such fallow agroforestry is not unique but occurs elsewhere in Amazonia as well as in tropical Africa and Asia.

Keywords Agroforestry · Amazon · Bora · Fallow management

Introduction

In recent years, students of Amazonia have emphasized that some of the most successful food-producing adaptations to the rainforest habitat have been those of the indigenous tribes and that consequently we have much to learn from these “ecosystem” people. “Refined over millennia, Amazon Indian agriculture preserves the soils and the ecosystem … If the knowledge of indigenous peoples can be integrated with modern technological know-how, then a new path for ecologically sound development of the Amazon will have been found” (Posey 1982, 18; 1983, 225). For similar statements for other tropical regions, see, for example, Nigh and Nations (1980), Clarke (1977), Eckholm (1982, 34–35), and Klee (1980). In particular, Indian cultivation is characterized by multiple cropping [intraspecific and interspecific] and interaction with natural vegetation.

Attention has been directed to several forms of traditional management of tropical forest resources: (1) the diverse, multi-storeyed swidden (shifting cultivation field) which protects the soil and allows for habitat recovery under long fallow (e.g., Conklin 1957; Harris 1971); (2) the house garden, or dooryard garden, was also diverse and multi-storeyed, but with a large complement of tree crops and with soil additives from household garbage, ash, and manure (e.g., Covich and Nickerson 1966); and (3) the planting, protection, and harvesting of trail side and campsite vegetation (“nomadic agriculture” or “forest fields”), involving wild, semi-domesticated, and domesticated plants (e.g., Posey 1982, 1983, 241–243). A related type of plant management is the manipulation and utilization of swidden fallows, a form of agroforestry involving a combination of annual crops, perennial tree crops, and natural forest regrowth.

Swidden-fallow management is apparently widespread among Amazon tribes and some mestizo farmers and rarely among colono (colonist) farmers. However, it has received little attention; brief mentions include Denevan (1971, 508–509) for the Campa in eastern Peru, Posey (1982, 1983, 244–245) for the Kayapó in central Brazil, Basso (1973, 34–35) for the Kalapalo in central Brazil, Eden (1980) for the Andoke and Witoto in the Colombian Amazon, Smole (1976, 152–56) for the Yanoama (Yanomami) of southern Venezuela, Harris (1971, 487, 489) for the Waika in southern Venezuela, and Torres Espinoza (1980) for the Shuar in eastern Ecuador. Some observers have assumed that all that is involved is a return to abandoned swiddens to search for residual crops left from the former cultivation, but indications are that actual management occurs, including planting and protection as well as utilization of certain useful wild plants that appear at various stages of fallow succession.

The purpose of this paper is to examine the swidden fallows of an Amazon native group, the Bora of eastern Peru, with the objective of demonstrating how fields are gradually abandoned. This contrasts with most studies of shifting cultivation which focus on why fields are abandoned and which present a sharp distinction between the field (swidden) and the abandoned field (fallow). For the Bora, there is no clear transition between swidden and fallow, but rather a continuum from a swidden dominated by cultivated plants to an old fallow composed entirely of natural vegetation Fig. 6.1. Thirty-five years or more may be required before the latter condition prevails. Abandonment is not a moment in time but rather a process over time.

Fig. 6.1

Bora man carrying peach palm fruits and chambira fronds harvested from a 10–15-year-old fallow (Photo by John Treacy)

Agroforestry is currently receiving considerable attention as a potentially stable and ecologically viable form of tropical forest land use (King and Chandler 1978; Hecht 1982; Budowski 1981; Salas 1979; Hart 1980; Spurgeon 1980). One of the major recommendations of the recent US National Research Council (1982, 4, 5, 146) report on tropical development is that the agroforestry systems of indigenous people should be studied and recorded before such knowledge is lost. We believe that certain features of Bora swidden-fallow management can be incorporated into systematic models of tropical agroforestry systems. Indeed, an examination of Bora land use indicates that “agroforestry” is new in name only to native groups in the Amazon. Under denser populations in the past (Denevan 1976), large areas of Amazon forest may actually have been stages of productive swidden fallows. Whole biotic components were largely selected and managed, a condition Nigh and Nations (1980) call “intermediate disturbance” and which Gordon (1969, 69; 1982, 73–78) calls an “orchard-garden-thicket” or “tree garden.”

The Research Area

Field work was undertaken from July to December 1981 in the Bora settlement of Brillo Nuevo on the Yaguasyacu river, an affluent of the Ampiyacu river (between the Napo and Putumayo) which joins the Amazon at Pebas 120 km northeast of Iquitos. The climax vegetation of the area is humid tropical forest. The closest meteorological station to Pebas is Francisco de Orellana, 75 km distant, where an annual average of 2,757 mm of precipitation was recorded (1964–72). There is a distinct seasonal distribution, with rains peaking from December to May and abating from June through November, but with the driest month (August) still having 133 mm. Temperatures average around 26 °C throughout the year (ONERN 1976, 37). Brillo Nuevo is situated beside an oxbow lake formed by the Yaguasyacu. The area is a hilly, dissected fluvial terrace interlaced with numerous seasonal streams. The soils are primarily deep Ultisols. They include red and yellow clay soils, red and brown sandy soils, and a gray soil in depressions. The Bora prefers to farm the clay soils and red sandy soils (Gasché 1979).

There are 43 families living in the settlement. All are descendants of tribal groups brought to the Ampiyacu from the Igaraparaná-Caquetá region of Colombia following Peru’s loss of a border war with that country in 1934. They were resettled on land eventually granted to them by the Peruvian government and to which they retain community title. (The study was undertaken at Brillo Nuevo, rather than with a community long established in its habitat, because of previous unpublished agroforestry research there by project member Salvador Flores Paitán.) The Bora are gradually being assimilated into Peruvian society through missionaries, commerce, and access to Pebas, Iquitos, and Pucallpa. Bora villagers speak Spanish, wear manufactured clothing, and market handicraft items and lumber. Bora subsistence, however, retains many of its traditional elements, with a reliance on swidden agriculture, house gardens, fallow management, high forest collecting, hunting, and fishing. Previous accounts of the Peruvian and Colombian Bora include Whiffen (1915), Jiménez Seminario (1933), Forde (1934), Girard (1958), Gasché (1980), Guyot (1971, 1974, 1975a, b), and Paredes (1979).

Background: Bora Shifting Cultivation

A brief survey of Bora agriculture was conducted to grasp the fundamental dynamics of the system and to understand how cultivation techniques might influence fallow-field character and management. Various aspects of cropping, spacing, and zonation within fields (chacras in Peru) and the schedules of planting, harvesting, and weeding are examined below. Together, these affect the eventual structure and composition of the fallows (purmas in Peru). Almost the entire area of village land is in some stage of secondary forest due to shifting cultivation since the Bora arrived here 50 years ago. However, stands of old, mature forest are within 20 minutes walking distance from the village and extend northward across the Colombian border.

Family fields are dispersed throughout the forest surrounding the communal maloca (residence of the village curaca or ceremonial head). Fields are often closely clustered because farmers find it convenient to visit several on one trip. Most plots are accessible within 15 minutes on foot from the maloca; others are across the Yaguasyacu and are reached by dugout canoe. Both primary and secondary forest are cleared for gardens. Primary forest sites are recognized as more fertile, while secondary forest (fallow) is closer at hand and more easily felled. The oldest clearly identified fallow is about 35 years of age. There is botanical evidence, however, of secondary forest over 40 years of age. (The presence of buried and surface potshards indicates previous occupation of the area at unknown times, by unknown Indian farmers.)

The Bora say that a minimum of ten years of fallowing is needed before a plot can be cut and planted anew. Most swiddens, however, appear to be prepared from fallows 20 years of age or older. For the Bora, one indicator of a fallow ready to be felled and cropped is a lack of shrubby growth near ground level.

Most fields are cut and burned during the months of least rain; however, a field can be prepared any time the weather permits. Field sizes range from a fourth of a hectare to one hectare. Axes and machetes are the only tools used for felling the forest. Cutting is often accomplished within hours by community work teams, but individual families can cut a field over a period of several days. Often small hills are chosen as field sites, the highest part of the hill becoming the center of the field. Fallen vegetation is allowed to dry for two or three weeks before burning. Selective cutting, a common management technique of swidden farmers, is practiced by the Bora. Valuable timber species, such as tropical cedar, are routinely spared during clearing, and various palms and other useful trees are commonly left in or at the edges of newly cleared fields; others may coppice and be protected.

The Bora plant a wide variety of crops (Table 6.1); however, the main staple is manioc. Some 22 varieties of sweet and bitter manioc are known by the Bora, and a newly planted field bristles with manioc cuttings spaced 50–80 cm apart. The Bora intersperse pineapples, fruit trees, and minor annual crops among the manioc. Both seeds and seedlings of trees are planted. Minimum spacing for fruit trees is said to be between 1 and 2 m. However, as the planting period may extend over several weeks, farmers forget from day to day where tree seeds are already planted and often place seeds closer together inadvertently. Consequently, a few planted trees end up growing virtually side by side.

Table 6.1 Common Bora Cultivated and Protected Economic Plants

Some crops are aggregated within the field. Fruit trees are commonly clustered on high land, topography permitting. Areas away from field boundaries, or near trails, also appear to be preferred sites for these trees. Patches from 1 to 2 m² are made into planting beds for tubers on sites selected according to ash distribution or local soil variations. The Bora recognize various soil types, based upon texture and color. Coca is almost always planted in well-tended rows near trails and field entryways.

Peanuts, grown in the second or third-year fields, are planted using a special management technique. In a small area from which manioc has recently been harvested, soil (previously loosened by manioc growth and root decay) is gathered and packed into several dozen mounds measuring from about half to 1 m.² Ashes brought to the fields from home cooking fires are mixed in with the soil as fertilizer. Between 6 and 12 shelled peanuts, previously soaked overnight in a solution of crushed basil leaves to prevent ant predation, are planted in the mounds. From two to four cuttings of sweet manioc are placed laterally into the sides of the mounds.

Bora names for swidden stages are based upon a field’s capacity to produce manioc. A field containing the first, most productive planting of manioc is called an úmihe . As an úmihe is gradually harvested and replanted, it becomes a kapúuwa , the term for a field yielding less productive secondary replantings of manioc. The Bora consider two replantings of manioc the maximum possible. When manioc is no longer replanted, the field is termed a jia, which is roughly equivalent to a fallow field or purma .

Initial crop zonation influences subsequent management options and the pattern of forest regeneration. First, clustering fruit trees in the field center or in areas of access allows them to be easily harvested and weeded as the field matures. Second, heavily disturbed or weeded areas, particularly the coca and peanut zones, will frequently only support sparse, grassy secondary growth. This may be due to local soil exhaustion or compaction, plant allelopathic effects, removal of seedlings of secondary species during intense cultivation, or some combination of these. (See Uhl et al. 1981, for a discussion of microhabitat preferences of secondary seedlings in Amazonia.)

The crop composition of Bora fields can vary widely. Some fields have an apparent low diversity index, planted only with manioc, pineapple, and maize (mainly for poultry) and perhaps a few scattered plantains and bananas. Others are rich in species and numbers and feature complex zonation. While a range of options is to be expected in any swidden system (Denevan 1971), the two extremes seem to be common in Bora swiddens. A similar duality is noted by Harris (1971) for tribes in the Orinoco region of Venezuela, where fields seem to be either primarily monocropped with staples or polycropped with abundant subsidiary plants. In many of these cases, the crop composition in any one field may in part be determined by what a farmer has available from other fields in various stages of development. Since a Bora family may have six or more fields of different ages and crop mixtures, diversity between fields fulfills the same function of assuring a supply of varied crops as does diversity within a single field. Another significant point regarding crops is that simplified fields receive few visits after two or three years of harvests, while diversified fields have longer-lasting utility in the fallow stage.

Bora Swidden Fallows

A series of fields was selected to examine vegetation structures and the process of abandonment. This paper examines plots of three, five, six, nine, and 19 years of age from date of cutting. Each field was measured to determine its approximate size, and the owners were interviewed to record cropping histories and to help inventory plants found within the fields. The vegetation was sampled using the line-intercept method. In each field, zones of vegetation were identified. These included plant communities in areas occasionally weeded and areas of unweeded secondary vegetation. Each zone was sampled by extending two ten m long intercepts into the zone from randomly determined points. Plants along the lines were collected and identified by their Bora names. In addition, Bora informants identified useful plants.

The plots are not strictly comparable in terms of relief or soil type, or planting histories. However, finding a series of fields with identical histories and characteristics is impossible in practice. Nonetheless, a dynamic model of abandonment is revealed by comparing vegetation patterns in the different aged plots.

The swidden fallows described below reflect a strategy of managed succession designed to solve a shifting cultivator’s dilemma of how to maintain field production in the twilight of the cropping cycle, while at the same time permitting forest regeneration. Abandonment is similar to Manners’s (1981, 360) evaluation of the swidden cropping cycle, which he describes as a “successional series partly regulated by human populations on the one hand, and ecological processes on the other.” A kapúuwa is chosen to head the sequence here because it represents a stage when human management is still relatively intense and forest regeneration is only just beginning.

Transitional Field: Three Years Old

Figure 6.2 is a representation of an enriched Bora swidden, cut from 30-year-old fallow, located not far from the settlement center. The field has developed multiple canopies, features complex zonation, and contains at least 20 cultigens. Guava, uvilla, macambo, and peach palm are the dominant tree species, all measuring between three and four m in height. The trees provide a 30% field cover but have not reached their peak yielding periods. Fruit-tree density in general is greatest near the southern end along the trail. The understory of manioc is sparse because tree roots and shade prevent replanted manioc from fully developing. A small peanut patch, also containing chili pepper and other minor crop plants, is located near the northwest corner. Bananas are more or less clustered on the southwest downslope corner. The field is surrounded on three sides by 30-year-old forest and on the south by newer fields less than one year old.

Fig. 6.2

Map of three-year-old transitional field (kapúuwa)

The kapúuwa, or transitional field, is a rapidly changing mosaic of vegetation reflecting Bora management techniques . Weeding, harvesting, and replanting manioc are performed in one small area at a time, producing a pattern of different aged stands of both manioc and associated secondary growth within the field. Weeds are often pulled out by the roots. In Fig. 6.2, the pineapple zone on the left are weeded, and the one on the right is unweeded. Selective weeding, another widespread swidden technique, is practiced by the Bora. Seedlings of useful tree species are often spared; however, it is not axiomatic that all are left untouched.

In fields at this stage, tree coppicing is readily observed. Some of the fruit trees in the field may be coppicing trunks of trees planted in the field when it was a chacra 30 years before. Inga species, useful as soil nitrogen fixers, are persistent coppicers, so abundant that they are nevertheless dispatched with machetes as annoyances. Others, such as copal, resprout and are protected. This tree grows slowly, reaching harvestable age (for edible fruits) within some 20 years. A bonus of useful coppicers appears to be one advantage in clearing purma for new fields.

Because the field is periodically weeded, secondary growth makes little headway except for invasion at the edges where fallen trees were not burned during field preparation. Some two or three m of unweeded field perimeter has been ceded to the encroaching forest. The growth primarily consists of fast-spreading vines and thin saplings.

Transitional/Orchard Fallow: Five Years Old

Some of the processes outlined above were noted in this field, also cut from a 30-year-old purma (Fig. 6.3), but at a later stage of development. The field contains a manioc kapúuwa zone; however, the unharvested manioc plants were small. As cuttings are routinely thrust into the earth after harvesting the roots, manioc can continue to grow without forming much below-ground material. Manioc is also a persistent plant; cuttings merely thrown aside will occasionally take root. Twelve other cultigens were originally planted, of which six were still clearly harvestable: coca, star apple, peach palm, uvilla, avocado, and barbasco.

Fig. 6.3

Map of five-year-old transitional/fruit orchard

Zonation resulting from management is evident. The large coca patch is well weeded and maintained. A small coca patch is abandoned and empty, as is a peanut patch. Secondary growth in both these abandoned areas is limited to short grasses, low herbs, and occasional seedlings of pioneer forest trees. A fruit zone extends the length of the field along the trail. The understory consists of a viney thicket mixed with low herbs growing among old pineapples and stray spindly manioc stems. This thicket forms an intermittent subcanopy 1.5 m in height. The overstorey is primarily comprised of equal numbers of well-spaced, productive star apples (three to five m in height) and peach palm (8–10 m in height).

Secondary vegetation has swallowed about a third of the original plot. The regrowth zone contains trees 10–15 m in height and measuring from 8 to 15 cm in diameter. Cecropia, Jacaranda, and Inga are common. The trees and abundant upper-storey vines form a 100% canopy. The forest floor is a dense tangle of herbs, including abundant Melastomataceae, Piperceae, and Araceae; palms are few.

In this transitional field, pineapples, fruit trees, and minor plants deemed useful are maintained. The pineapples may be harvested for up to five years; thereafter, the fruits produced are small and bitter tasting. Visits to the field follow the ripening schedules of the fruits, although visits for hunting also occur periodically. The main activity besides harvesting fruit is weeding. Coca is weeded every three months; the fruit trees and pineapples receive a slash weeding by machete every three to four months.

The farmer identified many useful plants, both in the weedy orchard and kapúuwa zones and in the reforested perimeter. The most immediately harvestable species are vines and low herbs. These include utilitarian vines and ceremonial plants not now used by the Bora, including reeds, once used to make decorative noseplugs and flutes, and plants yielding body paints. Other useful, but not yet harvestable species were construction and other woods in the seedling stage.

The reforested zone contained a great number of species. Thirty-four plants appeared on two 10 m transects, 13 of which were considered useful. Six were construction woods; four provided materials for weaving and dyeing baskets; and three were firewoods. Most of the useful forest species in this fallow will not be harvestable for 10–30 years. Rapidly growing construction woods are harvestable, but they are so plentiful around Brillo Nuevo that they receive no special care. The Bora casually harvest useful herbaceous plants as needed.

Orchard Fallow: Six Years Old

This orchard fallow is mapped in Fig. 6.4. Cut from primary forest, it is astride a sloping hill surrounded by newer fields on three sides.

Fig. 6.4

Map of six-year-old orchard fallow

The plot consists of two general vegetation communities: a residual fruit orchard occupying about one sixth of the original cleared area and abundant secondary growth surrounding the orchard. The original field was planted with over 26 crops, some of which are tree species now surviving within the orchard. Star apple is the most numerous planted species, and these trees measure 3–5 m in height. A canopy is formed by uvilla (5–8 m in height) and peach palm (10–13 m in height). Several 18 m tall Cecropia trees dominate the orchard area. The orchard has a 70% canopy and is well lit by sun splash. Regular weeding has resulted in an open floor of grassy vegetation covered with slashed mulch. Harvesting fruit in such purma orchards is a casual pastime. The Bora use poles equipped with vine loops on the ends to ensnare and pluck fruit-laden racemes from high branches. Coca, however, has suffered from shading, and harvesting is reduced. Cuttings are removed for replanting in nearby new fields. There is little evidence of manioc besides occasional stubble debris.

Growth surrounding the orchard is topped by 25 m tall Cecropia and Rubiaceae trees towering over dense stands of 10–15 m high trees and old plantain saplings. A thick, shrub understory mixed with abundant short saplings and palm sprouts occupies the entire subcanopy regrowth zone. The forest floor has accumulated a thin layer of leaf litter, and no grasses are present. An array of useful spontaneously appearing species similar to those in the five-year-old field were present in the regrowth zone.

Orchard Fallow: Nine Years Old

This fallow (Fig. 6.5), cut from high forest, demonstrates how long a managed orchard-fallow succession can be maintained. The orchard zone is small, and cultivated trees are few; however, a vigorously growing unshaded coca patch remains. The patch contains 82 evenly spaced, well-tended bushes. Coca is clearly the most valuable crop available here. The owner visits the field on a regular basis to harvest the leaves and on those occasions may refresh himself with uvilla, guava, and star apple foraged from the residual orchard.

Fig. 6.5

Map of nine-year-old orchard fallow

The secondary regrowth is a woody thicket, 10–15 m in height, with many vines and sub-storey shrubs. Several useful trees including cedar were on the field’s perimeter. Because this field is a downriver site, soils and topography differ from the upland sites nearer the settlement zone. The downriver sites are less well drained, and thus the secondary communities differ from the other fallows studied.

Forest Fallow: 19 Years Old

This older forest fallow (Fig. 6.6) was surveyed for useful tree species. The original chacra was cut from mature forest and, according to the owner, planted with at least 11 species, including several varieties of fruit trees.

Fig. 6.6

Map of 19-year-old forest fallow

The forest displayed clear stratification. Low vegetation consisted of herbaceous plants, including ferns, measuring 30 cm to 1 m in height. Above is a second stratum of thin, straight saplings, 5–6 m high, including many palms. Seventy-five percent of the canopy was provided by trees 15–18 m in height, while emergent Cecropia and Jacaranda, both 25 m tall, filled out the canopy. The forest floor was 40% covered in leaf litter, and walking was unhindered, except in small thicket-filled gaps caused by falling trees. All individual trees measuring 15 cm in circumference within a transect 10 m wide and 102 m long (length of the field) were tallied. Some 233 trees belonging to 82 species were counted. Over half the trees were single occurrences. Our informants identified 22 useful trees in the transect, fitting into the following categories:

• (a) Construction materials: 11 species, 25 individuals, including two varieties (three individuals) of highly valued cumala; 13 huicungo palms, used for general thatching, were also present.

• (b) Medicinals: four species, four individuals.

• (c) Food: two species, 11 individuals, consisting of eight macambos and three assai palms bearing edible fruits.

• (d) Artisan material: one individual, a dye-bearing tree.

• (e) Utilitarian: four individuals, four species. These included three palms from which salt is distilled, and one tree from which pitch is extracted and used to seal canoe hulls.

In addition, there were at least two other species of trees from which edible grubs are harvested. The only apparent survivors of the prior swidden were the macambos, which were clustered within the transect 60 m downslope. These are harvested sporadically.

None of the above trees, all of which are common, appeared to receive individual attention. The cumalas are not yet harvestable, nor will they be for about a decade. This old fallow apparently receives few visits for collection purposes, but hunting trips and grub foraging are frequent.

The Process of Abandonment: Analysis

The Bora recognize that two ecological processes, soil depletion and secondary succession, must be confronted. They acknowledge that manioc is not sufficiently productive to merit harvest after 3 or 4 years, mainly because of soil depletion but also because of weed invasion. Abandonment of fields planted almost entirely in manioc occurs within the space of a year. However, if fields are polycropped with trees, weeds may be the major obstacle to extended field use. Management shifts from replanting manioc to dealing with encroaching secondary vegetation threatening tree crops. With periodic weeding, trees can remain productive for several years before disappearing into the secondary forest, often succumbing to the effects of shading and competition for nutrients.

Our observations indicate that the most productive fallow stage is between about four and 12 years. Before 4 years, fruit trees are not yet producing or have limited production. After 12 years, management is minimal, and many of the smaller useful plants are shaded out. Harvesting of some species continues, however, for up to 20–30 or more years. Another important characteristic is seasonality. The various Bora fruit tree species yield sequentially allowing a spread of produce throughout the year.

A number of tree species planted in Bora fields are, however, adapted to growing in dense secondary forests. Umarí and macambo are common cultivated trees found in old fallows, either growing alone or in groups. These survivors of swidden orchards are valued components of Bora fallows. At 20 or 30 years of age, most fruit trees cannot be easily harvested; however, the Bora occasionally gather the fallen fruits. A valuable function of fallen fruit is that they attract game animals. It is common to find an umarí fruit on the forest floor with tooth marks of a majás (Cuniculus paca) or other browser. For this reason, older purmas are good hunting grounds.

The process of abandonment and forest regeneration clearly has a spatial aspect. While successionary processes are complex (Uhl et al. 1981), there is a tendency toward a pattern of centripetal forest regrowth which might be explained largely as a result of the history of weeding. Harvesting and weeding of manioc holds regrowth at bay. Once a manioc zone is abandoned, terrain is gradually surrendered to the forest, and the field shrinks in size.

Abandonment is also related to how harvesting proceeds sequentially from grain-producing annuals (rice and maize) to root crops and pineapples to fruit trees and spontaneously appearing utilitarian trees and vines.

Table 6.2 shows the succession of harvestable plants in Bora fields and fallows. While the Bora recognize many useful fallow plants, many go unharvested and are essentially neglected. The main reason for this is that high forest, from where sturdy construction woods and vines are harvested, is still a short walk away. At present, most plants used for handicrafts, for example, are taken from high forest. Nevertheless, as the high forest frontier becomes more depleted and distant, secondary growth species become more important. There is evidence that this is occurring. The Bora have recently become interested in planting hardwoods and useful palms in swiddens and in fallowed fields.

Table 6.2 Succession of harvestable plants in Bora fields and fallows

Phased Abandonment: Implications for Agroforestry

There are similarities between complex swidden systems and agroforestry systems (Hecht 1982). Agroforestry combines the production of trees and other crops on the same unit of land (King and Chandler 1978), a strategy essentially identical to swidden-fallow management. Both systems rely on the succession of tree crops following the harvests of short-term cultigens.

Viewed in this fashion, Bora agriculture converts to an agroforestry system during the early stages of forest fallow. The enriched swidden to fallow sequence closely resembles the natural succession analogue approach to tropical agroforestry outlined by Hart (1980; also Uhl 1983, 78–79). Hart suggests that select cultigens be placed in the niches normally occupied by common early successional species. The analogue plants would have growth structures and resource requirements similar to those of their weedy counterparts. Thus, rice or maize replaces early annual species, bananas replace wide-leafed Heliconia , and late-appearing tree crops mimic early successional tree species. Whether by accident or design, the Bora seem to follow this approach. Bananas do well in low shady areas, where Heliconia plants are also common. The most obvious example is uvilla which matches its ubiquitous cousin, the Cecropia. Guava is also in the same genus as its semi-domesticated analogue, the shimbillo (Inga sp.) . Further research may reveal other similarities between naturally appearing species and cultigens which could be incorporated into swidden agroforestry-type models.

Another feature of Bora swiddens that could be useful in agroforestry design is the use of space. Bora tree clustering according to local topographical conditions suggests that slope and terrain should be considered when planning agroforestry plots. More important, slowly abandoning ground to secondary forest may be a sound strategy for tropical farming. There is no reason to think that agroforestry plots should have 100% planted standing biomass. Managed forest regrowth could provide useful products, as well as canopy cover for the soil and a source of stored nutrients for when the forest is cleared to begin the swidden and agroforestry cycle anew.

Swidden-fallow agroforestry, enriched with tree crops planted in areas of forest regrowth, could approximate the “tree-garden” model of silviculture which may have been a pre-European agricultural adaptation in the Caribbean lowlands of Colombia, Central America, and the Maya region (Gordon 1982). This “thicket” model involves a combination of over-storey fruit trees and subcanopy woody shrubs, interspersed with areas of maize, bananas, manioc, and other crops. Systematic swidden-fallow agroforestry would have a fruit orchard core, or series of cores, but these would be embraced by areas of regenerated forest. The forest, in turn, could be enriched by a variety of useful analogue species able to compete in the viny subcanopy, or later on as canopy species (fruit, timber) in high-forest fallow. Timber species would be appropriate late-fallow enrichment trees. Over a large area, swidden-fallow agroforestry would resemble Gordon’s image. It would be more a thicket and less a field. Furthermore, the growth rate of managed successions may be as fast or faster than natural successions (Uhl 1983, 79).

Swidden-Fallow Products

The cumulative dietary contribution of fruits and nuts, even when harvested casually, may be significant. Certainly, they provide a continuing (seasonal) variety of minerals, fats, and vitamins to tropical diets dominated by roots and tubers rich in carbohydrates. Some trees, moreover, can provide major staples. The peach palm, very important to the Bora for its fruit and heart, can compete with maize as a nutritious food (Hunter 1969; Johannessen 1966). In addition, plant products useful for beverages, condiments, construction, tools, drugs, and medicines are of more than minor importance to village societies and economies.

As with major and minor natural forest products, those in swidden fallows frequently reach markets beyond the village, at regional, national, and international levels. Even remote traditional cultivators are willing and able to respond to market opportunities for forest products and manage those products accordingly. Pelzer (1978, 286) argues that a large percentage of the rubber, black pepper, copra, coffee, and benzoin harvested for cash in southeast Asia comes from smallholder swiddens through intercropping “in what is ordinarily thought of as the ‘fallow’ period of the swiddens.” The ultimate success of agroforestry systems will depend on such cash cropping.

For isolated communities such as Brillo Nuevo, cash cropping of forest products is problematic. Tropical cedar and other timber trees can be floated downriver to market. One can only be impressed by the Bora planting or protecting tropical cedar seedlings in their swiddens and fallows, anticipating a substantial cash return for their children 30 years later. The use of swidden-fallow products, such as palm and liana fibers, tree bark, and dyes for the manufacture of handicraft items, can bring an income to Bora households. The considerable tourist and export trade in the Iquitos area provides an outlet for traditional items such as hammocks, bags, baskets, bowls, and ornaments. On the other hand, the marketing of perishable food items does constitute a difficult problem for remote villages such as Brillo Nuevo, especially in view of the poorly developed processing and marketing facilities in the region. Toasted macambo nuts, a Bora delicacy, could have market potential. Palms such as Jessenia and Mauritia, potential sources of edible oils (Balick 1982), are common in Amazon forests and could be integrated into agroforestry models.

The history of the Amazon has been one of commercial harvesting of forest products (quinine, copal, sarsaparilla, barbasco, palm heart, Brazil nuts, rubber, timber). Much of that history involved the destruction of important resources by unwise harvesting practices and the economic and social exploitation of indigenous peoples. Sustainable and equitable procedures are possible, and trade in forest products can be enhanced by incorporating forest species of commercial value into agroforestry systems. Such commercial orientation would, of course, necessitate not only the development of specific agroforestry designs and techniques but also appropriate processing, transportation, credit, and marketing facilities. The economic possibilities for Amazonian plants are vast (Myers 1983). An argument might well be made that the potential value of marketable production from sustained-yield agroforestry plots, including swidden fallows, can be significantly greater per year per hectare than that from cattle ranching or shifting cultivation.

Conclusions

The Bora process of swidden abandonment is in reality a conversion of a short-term cropping system into a longer-term agroforestry system. The main conclusions regarding abandonment and fallow management are summarized as follows:

• 1. Fallowing is multipurpose. The secondary forest is not only nutrient storage for future cropping but an important niche for secondary crops and useful spontaneously appearing plants. We identified 131 different useful species in Bora fallows. We propose that an appropriate designation be established to account for enriched fallows, a characteristic which may be common in tropical swidden systems. The term “orchard fallow” could be used to describe the structural and functional aspects of traditional agroforestry. In a subsequent “forest-fallow” stage, economic plants are still present but are more dispersed, fewer in number, and less managed.

• 2. Viewed properly, a swidden site is never completely abandoned as a resource zone. Secondary harvests of fruits, spontaneously appearing species, and even animals continue until the forest is removed for further cropping.

• 3. There exists an identifiable sequence from original forest with some economic plants present, to a swidden with numerous individual economic plants present, to an orchard fallow or agroforestry phase combining managed economic plants and natural vegetation, to a forest fallow in which economic plants are fewer but still present in greater numbers than in the original forest. Likewise, there is a corresponding sequence in the proportions of biomass which are cultivated or managed, spontaneous economic, and spontaneous non-economic.

• 4. Research is needed on analogue species with growth architectures and nutrient requirements adapted to secondary forest environments.

• 5. Swidden-fallow management is not unique to the Bora, (but also appears with other Amazonian groups). It is widespread in Africa (De Schlippe 1956, 215–216; Dubois 1979) and in the Pacific, including the Philippines (Conklin 1957, 125–126; Oración 1963), New Guinea (Clarke 1971, 82–84, 138–139; Hyndman 1982), and Micronesia (Yen 1974). It may once have been common in Middle America (Gordon 1982). These systems need to be studied; it is still practiced by the Huastec in Mexico (Alcorn 1984).

• 6. Agroforestry drawing on traditional management methods and combining planted species and natural secondary vegetation could be an ecologically appropriate and economically viable alternative to destructive short-fallow shifting cultivation in tropical areas. The ideal model would provide food crops during the swidden stage and cash crops and other crops during the fallow stage. The cash crop perennials should be relatively fast maturing species which can be harvested by around ten years so that the cycle can be renewed as soon as possible. Such a model would help fulfill the need for sustained production of food and other needed products and simultaneously do minimal damage to a fragile environment.