Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Human use of amazonian aquatic and wetland ecosystems
Native Amazonians
Millennia before a band of Europeans was swept downstream by eastwardflowing Andean waters (Carvajal, 1942) and others had paddled west against the current to remote Cordilleran headwaters (Acura, 1942), Amerindians had introduced a human dimension to the Amazonian landscape. Indeed, early reports by soldiers and missionaries suggest very large numbers of indigenes at the time of conquest. Yet high estimates for the pre-contact population generally have been viewed with some reserve. Perhaps the reason lies in the failure to allow for the precipitous demographic decline during the early years of contact, whether from deliberate extermination, lack of immunity to Old World diseases, or the general disruption of indigenous ways of life.
Several scholars admit the possibility that Indian nations in Amazonia might have reached several million individuals before this collapse. In considering such estimates, it should be remembered that the pre-Conquest inhabitants might not have had to face some of the hardships presently seen as inherent to the region. Thus, areas now malarial, and shunned as such, may have been free of the disease. Several parasitologists suggest that the aetiological agents of malaria were introduced into the New World by Europeans and their African slaves (Dunn, 1965, 1993; Coatney et al., 1971; Wood, 1975).8 The proposition has not gone unchallenged (e.g. Noble et al., 1989), and much research remains to be done with respect to the initial manifestation in the Americas of malaria and other diseases.
The fact is that substantial archaeological evidence exists for large precontact populations in Amazonia: e.g. pictographs, petroglyphs, and the widespread occurrence of charcoal and kitchen middens in the soil of what might appear to be virgin forests. Salvage work carried out in connection with large-scale hydropower projects has revealed a number of prehistoric sites. In the case of the Balbina reservoir, for example, a map of the lower Uatumã and tributaries shows no less than 121 such sites (SAUHEB, n.d). Similarly, a survey in the area of the projected Porteira hydroelectric station, on the Trombetas River, Pará, identified 43 prehistoric localities (Costa et al., 1986).
The archaeological significance of terra preta (black earth) was recognized more than a century ago. This anthrosol is derived from middens, and contains abundant potsherds and ceramics, some of considerable sophistication, such as the striking Santarém pottery (Palmatary, 1960). The terras pretas are found both on floodplains (Sternberg, 1960) and uplands, where they have long been "highly prized as agricultural grounds" (Brown and Lidstone, 1878).9 Indeed, indigenous settlements were so large and persistent that soil scientists have come to recognize the black earth as a "taxonomic unit" (Falesi, 1967).
In sum, there are grounds to believe that, when Europeans arrived upon the scene, Indians numbered in the millions, and that the Amazon had been, in the words of Palmatary (1939), "a busy ... part of the New World."
As to the time depth involved, indications are that it is greater than formerly thought. Thus, caves and rock shelters under the plinthite that caps the vast reserves of iron ore in the Serra dos Carajás of southern Pará have begun yielding artifacts whose oldest radiocarbon ages are reported to be 8,140 + 130 years (Lopes and Silveira, 1990), or 8,065 + 360 (Magalhães, 1993). They are interpreted as evidence of the presence of pre-ceramic hunter-gatherers and, later, of potters. Among other indications of human antiquity in Amazonia is the eighth-millennium pottery from a prehistoric shell midden at Taperinha, near Santarém, taken to be the earliest yet found in the western hemisphere (Roosevelt et al., 1991). One may expect that continued research will further push back the record of human presence in the region.'° Obviously, this presence relates to the more general and highly controversial issue of timing the entry of Amerindians into the New World. And here, some students (e.g. Torroni et al., 1993; Torroni et al., 1994; Wallace and Torroni, 1992), having applied a new high-tech tool, the analysis of mitochondrial DNA, to the problem, are interpreting the results as consistent with an "early" arrival, possibly some 30,000 years before present or even earlier.
Thanks to the mobility provided by an extensive fluvial network, comprising tens of thousands of kilometres of navigable channels, the wateroriented tribes ranged far and wide. Alexandre Rodrigues Ferreira (1972), whose Viagem Philosophica from 1783 to 1792 has been seen as the forerunner of nineteenth-century scientific reports on Brazil, remarks on the intimate knowledge of Native Brazilians concerning Amazonia's hydrography. He recounts how a Macuxi, questioned about the Rio Branco, took a cord to represent this affluent of the Rio Negro, and attached to it, left and right, as many strands as are the subeffluents, spacing them correctly and endowing them with appropriate windings. Additionally, knots indicated the number and position of Indian villages.
True, not all Indian groups were riparian peoples, a fact evidenced by numerous archaeological sites on the upland interfluves (Smith, 1980). But the "personality" of Amazonia, to a considerable degree, rests on, and is defined by, its waters and wetlands. To them are attached some of the region's most fascinating myths, such as that of the iara or mãe d'agua, siren of rivers and lakes, that of the lecherous bôto, or Amazon River dolphin (Inia geoffrensis), seducer of streamside maidens, or yet that of the supernatural cobra grande, big snake.
Even the cosmologies of contemporary Amazonian fishermen occasionally translate an ethos that, at least theoretically, favours sustained use of aquatic resources (Furtado, 1992; Smith 1981). One belief, for instance, enjoins fishermen to abstain from overfishing and to diversify their catch, in order to avoid angering the "mother-of-fish." Also to be shunned is the wrath of another spirit that punishes those who disturb spawning female turtles.
The Amerindians' economy did not demand the stockpiling of vast food resources. An abundant storehouse was at hand, and they knew how to preserve the product of their fishing, hunting, and gathering, for instance by smoke-drying, moquém, or potting meat in its own rendered grease, mixira. Some carry-over of seasonally abundant fare was thus assured, and they had little incentive to overexploit the waters and wetlands.
Referring to the numerous alluvial islands on which the natives laid out their crops, Acuña observed in 1639:
The river washes [over] these islands every year, fertilizing them with its muds, so they [are ... never] sterile, even though, year in and year out, one demands of them ... maize [Zea mays] and yuca or manioc [Manihot utilissima].
In addition to the staples mentioned by Acuña (1942), Amerindians assembled in their floodplain gardens a variety of food plants. Some of these are believed to have been domesticated thousands of years ago, such as a number of cucurbits (Cucurbita spp.), beans (Phaseolus vulgaris), peanuts (Arachis hypogaea), papaya (Carica papaya), and several species of Capsicum (Ducke, 1946; Kerr, Campos, and Barros, 1986).
Things would change with the arrival of Europeans and Neo-Brazilians. Instead of a moderate, diversified use, mainly for consumption, of aquatic and wetland bioresources, a mercantilist world pushed many species to the brink of extinction.
Conquest and colonization: The waters
Incorporation of Amazonia into the Portuguese empire occurred in the seventeenth and eighteenth centuries. As the Amazon Basin was conquered, river after tributary river, it yielded up all sorts of drugs and spices - aromatic and valuable cargo for Europe-bound ships. Plants and animals enriched merchants, and suffered great havoc in the process. The aquatic fauna early became the object of veritable carnages, as in the case of the Amazon turtle (Podocnemis expansa) and the previously mentioned manatee.
Coutinho (1868), writing about turtles, once present in vast numbers, decried their capture before oviposition, and recalled guidelines no longer observed:
Nobody was allowed to approach before the eggs had been laid. Only when this was completed, did one proceed to the viração [the "turnover," by which the chelonians were immobilized], each person receiving half a turtle, the remainder being given back to the river.
While two-thirds of the clutch of eggs had been used in the manufacture of manteiga (literally, "butter," an oily substance derived from the yolks), onethird had been reserved for reproduction. With commercialization, the destruction of eggs reached mind-boggling proportions. Fat obtained from the eggs was used in the preparation of food and as lamp oil, as was that extracted from turtle meat (fig. 6.6).
Manteiga was also melted down from fatty tissues of the manatee and put to the same use. The massive slaughter of the sirenian got under way in the seventeenth century (Vieira, 1659; Heriarte, c. 1662). Part of the resultant products, notably rendered fat, was shipped to Europe. Much of the flesh apparently was consumed in the Guianas and the Caribbean, where English and French colonies were supplied by Dutch traders. 11
The preceding illustrates how overexploitation directly endangers bioresources.12 These may also be threatened indirectly by habitat modification. A wide range exists of potential anthropic impacts of this kind on the waters of Amazonia, which hold sway over the alluvial wetlands and are complex ecosystems in themselves.
Human activities may subvert hydrologic characteristics, such as volume and seasonal variations of runoff. They may also affect the particulate and dissolved load of river waters, altering their chemical and other properties, and modifying their interaction with soils, plants, and animals. Further, they may trigger obscure biological chain reactions by eliminating keystone species or, conversely, by introducing disruptive exotic organisms.
Any of these, and a number of other anthropogenic changes, would probably have a noticeable impact on the human population. The life of floodplain dwellers (varzeanos, varzeiros, ribeirinhos) is closely linked to the yearly rise and fall of water levels, which conduct the cadence of current economic activities on, and in the proximity of, the várzea. One example is the gathering of aquatic macrophytes as fodder cattle huddled on platforms or rafts during high water. Another is the felling and removal of timber in várzea forests, the former at low, the latter at high, stage. To the degree in which waters saturate, submerge, or leave the land dry, they influence not only the extent of the area available for cropping but the span of time during which it can be used, and even the manner of its use.
Conjectured human-induced trends in river stage
An upswing in the severity and frequency of major floods in the Amazon lowlands was predicated at least as early as 1903 by Paul Le Cointe, who returned to the subject several times (Le Cointe, 1922, 1935, 1948). Leaving aside the possibility of water slope adjustment to neotectonic events (Sternberg, 1950, 1955, 1987a), shown to be significant in westernmost Amazonia (e.g. Dumont, 1993; Räsänen, 1991, 1993),(13) elevation of flood levels such as suggested by Le Cointe could result from increased highwater flow, siltation of channels and bottomlands, or a combination of both processes. Although these might be induced or exacerbated by societal behavior, Le Cointe did not invoke the possible role of human activities.
However, given the contemporary rate of deforestation, it is now reasonable to anticipate that, in time, and starting with small or medium-sized, intensely ravaged, tributary watersheds, streamflow will be affected. Reduced transpiration, as moisture extraction by plant roots decreases, compounded with curtailment of water storage, due to soil degradation and erosion, may be expected to raise flood levels. An increased delivery of debris may contribute to the silting up of waterways and thus to a general elevation of gauge readings. In the case of low-water discharge, impairment and removal of soil should lead eventually to a diminished carry-over into dry periods.
Prominent among activities that contribute to deforestation and thus might affect the flow patterns of Amazonian waters is the construction of new roads opening the interfluves to "economic development." In Brazil, this was translated essentially in terms of giant cattle ranches, established at the expense of pristine selva. In addition, following the new roads came wave upon wave of farm folk, either landless or desiring to exchange minuscule, worn-out plots in the south or north-east of Brazil for larger tracts carved out of the forest.
New inducements for the destruction of Amazonia's rain forest have arisen. Mining is one of them. Take, for instance, the Grande Carajás region,14 named after the massif noted for its immense mineral reserves, especially high-grade iron ore. Far more extensive than the forest areas directly affected by opencut mining in the district are those where wood is extracted for the charcoal used in smelting (Penna, 1989). Charcoal kilns that use forest trees are found in a wide radius around the blast furnaces. One estimate is that an area of 26,000 Km² for tree plantations eventually will be needed to satisfy the demand (Kohlhepp, 1987).15
Other incitements to forest clearing, less readily discernible at this time, are certain to lie ahead. One might be the commercial production of lowland coca (Erythroxylum coca, var. ipadu), a close relative of the Andean E. coca, var. coca (Plowman, 1981). Grown initially by Indians, the powdered leaf had been observed in use in 1774 by a Portuguese administrator (Sampaio, 1825). Later, planted under conditions of shifting cultivation, it was noted by Martius (Spix and Martius, 1822-1831) on "a hill, cleared of forest south of the village of Ega," now Tefé. In the Andean realm, for example in the Huallaga Valley of Peru, coca has become the single most important cause of deforestation and a significant factor in water pollution by fertilizer, herbicide, and industrial wastes (Dourojeanni, 1992). To the degree that eradication programmes are effective in the mountains, production may shift to the Amazon lowlands.16
Given actual and potential motivations to divest cordilleran and lowland watersheds of their cover, it is defensible to argue that hydrologically disruptive effects of clearing will eventually spread to the main stem. It is another thing to posit present-day changes in the lowland Amazon induced by deforestation in the Andes. Yet it has been suggested that higher and more persistent flood crests, as well as decreased low-water stages, were perceived as far downstream as Iquitos (Gentry and López-Parodi, 1980, 1982). Consistent with this suggestion is the claim by varzeanos in the Itacoatiara area that floods had become increasingly severe. Confirmation of this observation was sought in the stage levels at Manaus (Smith, 1981), but the suspected tendency of the maxima at this station, after a spike in 1976, did not hold. Indeed, trend analyses relative to heights, duration, and timing of yearly floods, applied to data for the period 1902-1985, showed no statistical significance, and did not support the above propositions (Sternberg, 1987a). When the series was extended to 1992 (fig. 6.7), both annual maxima and mean daily water levels continued to reveal no significant tendency. However, the slope of the linear trend relative to the low-water stage exhibited a slight, but distinct, upward bent, and requires further investigation.
Figure 6.7 Annual maximum, mean, and minimum stages the Rio Negro at Manaus.
Series for the period I January 1903 to 31 December 1992. Given the negligible gradient of the water surface, the Manaus gauge, about 20 km up the Negro estuary, provides a reasonable surrogate for the stage of the Amazon River at the confluence. Upper graph: annual maximum for each year; middle graph: mean of all daily values for each year; lower graph: annual minimum for each year. The horizontal lines give the mean level of the respective series for the entire 90-year period. In the search for a possible change in the level of the water surface, trend analyses were carried out for the three series in the manner of Brillinger (1988). The results were as follows. For the maximum: slope, 0.0068440m/yr, SE, 0.0055143; ratio (tstatistic, with 17 degrees of freedom), 1.241); two-sided p-value, 23.14%. For the mean: slope, 0.011471m/yr; SE, 0.0062402; redo, 1.834; two sided p-value, 8.36%. For the minimum: slope, 0.019263m/yr; SE, 0.0081425; redo, 2.366; two-sided p-value, 3.01%. Whereas the trends for the annual maximum and for the mean of daily values lack significance at the commonly employed 5% level, the upward trend of the annual minimum is significant at this level. (Source of data: Manaos Harbour Ltd. and Portobris).
Deterioration of aquatic ecosystems
The hydrograph can be significantly altered by manipulation of the natural conduit through which the river flows (e.g. by impoundment
or construction of embankments). When, in the 1960s, the Hudson Institute of New York proposed the building of a gigantic earth dam across the Amazon at the Óbidos narrows (Panero, 1967, 1969), the scheme met with strong opposition - largely on emotional issues of sovereignty, rather than because of its inherent silliness - and was abandoned. By contrast, a programme to harness effluents of the trunk stream or of the Rio Pará has been actively pursued. As a result, the Tucuruí hydroelectric power plant, one of the world's largest, was inaugurated in 1984 on the Tocantins River, which has been envisaged as an unbroken chain of huge reservoirs (Sternberg, 1981). Plans for similar development on other tributaries were drawn up in rapid succession, and some were implemented before the country's financial difficulties interrupted the programme.
Engineers argue that the area of forest drowned by even the largest hydroplant is but a small fraction of what goes up in smoke every year. They are probably right. However, it is not the area lost to reservoirs that is so important, but the effects of the impoundment on entire aquatic ecosystems. Recall that the life cycles of many species of freshwater fauna are, in great measure, adjusted to yearly inundations of the forest and to an open communication between river channels and floodplain lakes (Goulding, 1980; Marlier, 1967; Smith, 1981; Sternberg, 1975). It seems quite possible, therefore, that an important supply of protein, of special value to underprivileged rural populations, will be much reduced by impounding Amazonian rivers. Dams do not merely affect characteristics of flow; they can also alter physical and chemical properties of the water.
Noteworthy among other activities that constitute a direct menace to the quality of Amazonia's waters are the extraction, processing, and transportation of mineral resources. In some tributary catchments, the purity of water has already been seriously vitiated by mining and metal refining. Illustrations are found typically in Andean headwaters, but, increasingly, in the lowlands. An example comes from the Madeira River basin, in the state of Rondônia. The region has been the scene of a gold rush that mobilized an adventitious labour force, estimated, at times, to exceed half a million garimpeiros (folkminers, prospectors) in Brazil's Amazonia.17 On the Madeira, gold is extracted from the channel bed by means of locally-built suction-dredges (fig. 6.8). In 1989, the Rondônia Cooperative of Garimpeiros (COPEGRO, 1989) denounced the frenzied increase in the use of heavy equipment on the river, alerting against an exhaustion of the placer. That COPEGRO suggested a minimum distance of 50 m between dredges, in order to reduce congestion and the frequent collision of outboards, gives some idea of the crowded conditions at the time on bountiful sections of the river. Prospecting for gold along the 180-km stretch known as the reserve garimpeira fell off considerably since COPEGRO's pronouncement, after a high point of some five to ten thousand dredges was reached - a fact that appears to reflect the depletion of the deposit (Veiga, 1993).18
Reworking of the channel bed, roiling of the waters, redeposition of sediments downstream, and dumping of spent diesel oil in the rivers are not the most serious impacts of gold mining in Amazonia. Metallic mercury is used on the dredges, as it is in land-based operations, to amalgamate fine gold particles. When the precious metal is recovered by ustulation, the vaporized quicksilver eventually falls back to earth, contaminating soils and waters. No form of pollution can be more insidious. It is hard to believe, for instance, that the limpid, apparently pure, waters of the Tapajós River, downstream from Itaituba, an important gold-mining centre, can pose serious health risks. Yet, there is evidence that this, precisely, is the case. Mercury (particularly the organic compound, methylmercury) identified in Amazonian river sediments, plants, and fish (see, for instance, Castro, Albert, and Pfeiffer, 1991; Malm et al., 1993; Pfeiffer et al., 1991; and the review by Pfeiffer et al., 1993), has a profound and generally irreversible impact on the human central nervous system. The levels of this metal in several fish samples collected in Amazonia have exceeded the World Health Organization's safety limit for human consumption; high values found in species widely consumed by humans, such as the omnivorous tambaqui (Colossoma macropomum), are of the greatest concern. Some plants, such as the water hyacinth (Eichornia crassipes), used to feed domestic animals, seem to concentrate mercury (Martinelli et al., 1988).
In contrast to pollution by mercury, which is not perceptible to the senses, the presence of even a thin film of petroleum, another threat to Amazonian aquatic and wetland environments, is betrayed by iridescence. However, Brazil, an unequally developed country, aspires for "first world" status. It is committed to an urban-industrial model of development, having chosen to rely on highways and airways for its long-distance transportation - notably in its attempts to end the isolation of Amazonia. Oil spills, while potentially catastrophic to waters and wetlands, are not immediately life-threatening to human communities. The national society is likely to see economic benefits and pay little, if any, attention to the possible environmental costs of a petroleum-driven development of Amazonia.
What appear to be important oil and gas reserves were found in 1986 more than 600 km west of Manaus (fig. 6.9), in the Urucu river drainage (Pimentel, 1986). The prospect of drilling for, pumping, transporting, and refining oil in Amazonia is disquieting, given the effects that a major spill could have on waters and wetlands. A relatively minor mishap occurred in 1988 but received little or no attention outside the region. An estimated 20,000 litres of crude befouled the narrow Urucu River (A Crítica, 1988a, 1988b, 1988c, 1988d, 1988e), whose waters eventually reach the Solimões, about 450 km upstream from this river's confluence with the Rio Negro.
Despite the billions of dollars spent on cleanup efforts in the Gulf of Alaska following the 1989 Exxon Valdez spill of more than 40 million litres of oil, the results have been judged unsatisfactory. Destructive as were the environmental effects of the Alaskan disaster,(19) they pale in comparison with what could happen if a large discharge of oil were to take place in Amazonian waters. In the Prince William Sound, the land area directly affected, mainly rocky headlands and pebbled beaches, was limited to the intertidal zone. In the mosaic of levees, backswamps, ridge-and-swale features, and shallow lakes of the Amazon wetlands (see figs. 6.2 and 6.5), the very concept of shoreline has little meaning. At floodtime, waters spread over an area that may comprise as much as 90 to 95 per cent of the várzea (Junk, 1985b). In a low-energy environment lacking the wave action that helps break up oil spills at sea, the slick, depending on the river stage, could be carried dozens of kilometres through open water, flooded forest, and floating meadows. One can imagine the immediate effects of oil settling on, and penetrating, the fine sands, clays, and silts of the floodplain, impacting on soil organisms, plant life, and aquatic as well as terrestrial fauna. In the long term, however, the intense microbial activity of the tropical milieu may expedite the weathering of the pollutant. As to the chances of a significant oil discharge into waters and wetlands of Brazil's Amazonia, the question is not whether, but when and how much.
One final instance of the kinds of human activities that have the potential for causing grievous harm to the region's aquatic ecosystems: consider the plan for a waterway from the Caribbean to the Rio de la Plata, to be achieved by linking the Orinoco with the Amazon system, and this with the Plata basin. Given Brazil's current lack of interest, there is no immediate prospect of its implementation.(20) Nevertheless, the point needs to be made that interbasin linkages, involving the Amazon drainage - an idea which has been under consideration for at least two centuries (Anal, 1773; Humboldt and Bonpland, 1820-1822) - could produce environmental changes of unforeseeable ramifications. A major concern among many students stems from the role of waterways in the diffusion of organisms. Although not all exchanges need prove harmful, the gamut of possible biological consequences from the intermingling of biotas pertaining to previously discrete ecosystems is considerable. Hazards include competition for resources, direct predation, perhaps even the spread of parasitic diseases among geographic isolates. An estimation of this risk must be based on an understanding of the degree of biological separateness of the catchments involved, and the specific co-evolutionary history of hosts and pathogens, in the case of both endemic fauna and potential invaders.
Conquest and colonization: The fertile interface
Because waters create, refashion, and exert a dominant influence on the wetland ecosystem, human actions that denature the former can injure the latter. Thus, given the importance of ichthyocory, modification of the waters qua habitat for dispersal agents may indirectly affect the makeup of the wetlands vegetation. Naturally, the most demonstrable anthropic pressures on an ecosystem are those applied to it directly - and the Native Amazonian farmers stepped lightly on these lands. Numerous tracts of terra preta testify to sustained use of the floodplains. The same is true, dowovalley, for the estuarine and tidal lands, whose mudflats and mangrove forests, in the words of the historian Varnhagen (1927), "offered ... inexhaustible mines of crabs." They also offered an abundant harvest of molluscs, easily harvested from the roots of mangroves (Rhizophora mangle, Avicennia, nitida, and Laguncularia racemosa). That the utilisation of such resources by humans goes far back in time is evidenced by the number and size of coastal shell mounds, long quarried by colonists as a source of lime (Le Cointe, 1945).
Among the earliest Europeans to cultivate the Amazonian wetlands were the religious orders: first, in the manner of subsistence farming, to feed the missions; then in the production of cash crops, such as bananas, sugar cane, cacao, cotton, rice, maize, and manioc, with which to finance their evangelisation activities. Plantations and cattle ranches were established in the estuarine section of the river, where ruins of sugar mills and vestiges of drainage and irrigation canals have been identified (Lima, 1956). One mideighteenth-century inventory of Jesuit holdings on Marajó Island (fig. 6.10) lists seven ranches, running more than a hundred thousand head of cattle.
The Portuguese and neo-Brazilian upriver settlements were essentially punctate, dispersed, allowing ample room for new arrivals to settle out interstitially, along the riverine avenues of penetration. Gradually, farming developed alongside the basic gathering activity. Although some plants were introduced, indigenous crops took pride of place. Native manioc, or cassava, was then - as it is now - the staple food crop, and the transformation of the tuber into farinha, or meal, was among the first rustic industries.
After the middle of the nineteenth century, penetration and settlement of Amazonia followed a pattern that mirrored the geographical distribution of the seringueira, or rubber tree (Hevea brasiliensis). At a time when the demand for rubber was increasing, disastrous droughts uprooted a considerable part of the rural population in Brazil's semi-arid north-east, the Nordeste. Arriving in Amazonia, they formed the workforce of the new extractive economy. Other nordestinos, eschewing the dream of making a fortune in the distant rubberrich headwaters ("altos rios"), settled on the floodplain with their families, to be joined later by those who beat a retreat from the seringais when the boom collapsed. To the fertile river-banks that emerge during falling stages they applied the traditional lavoura de vasante ("ebb farming") they had used in the seasonally dry river beds of their homeland. Well-drained natural levees were cultivated with a diversity of native and exotic crops: annuals, replanted after each flood, and, starting with endemics, arboreal perennials capable of surviving the yearly inundation. Home gardens contained native fruit trees, such as cupuaçu (Theobroma grandiflorum), graviola (Annona muricata), papaya (Carica papaya), açaí (Euterpe oleracea), and cacao (Theobroma cacao) - for a time an important cash crop. Much later came jute (Corchorus capsularis), introduced in the early 1930s by Japanese settlers and once a significant element of the varzeano's economy (Pinto, 1966).
The end result is a landscape incorporating a rich repertory of native and introduced crops, distributed in conformity with the topography of the várzea. Crops intolerant of humidity or requiring more time for maturation, such as banana, cacao, manioc, or beans, are located on the levee crests, and those more tolerant to humidity, notably jute, make use of the bottoms (fig. 6.11), In addition to current agricultural pursuits and animal husbandry, there has been some small-scale rubber tapping, logging, and extraction of non-wood products of the selva. With the river and the forests supplying a rich bounty of animal protein, this land-use system has been treated as an example of what now goes by the designation "agroforestry" (Bahri, 1992).
The distribution of various crop combinations (in 1956), adjusted to the subdued relief of the floodplain, is indicated in a generalized manner by segments of streambank, not by individual establishments. The depiction of some crops, such as beans, sweet potato, rice, and vegetables, which are significant on certain farms, is not compatible with the degree of generalization of the map. Fruit trees are disseminated throughout most properties, and some stretches of the paranás are lined with mangoes. (Source: Sternberg, 1975.)
Nowhere is it better represented than on the floodplain island of Careiro,(21) especially on the levees of the Paraná do Careiro, which bounds it on the south, and the subsidiary Paraná do Cambixe (figs. 6.4, 6.11 and 6.12). Here, refugees from the 1888 drought in the Nordeste received small long-lot tracts of a size proportionate to the number of active family members. The stable, close-knit community that resulted favoured the retention, alongside truck gardening, of the north-east's time-honoured livestock farming tradition. The landscape came to be increasingly influenced by dairying, with the island becoming a key component of the Manaus milkshed (Sternberg, 1956,1966). The spatial organization of the Careiro wetlands permits the periodic utilization, as supplementary pastures, of lowlying areas, where native fodder plants become available as the waters recede.
During severe floods, however, forage must be cut and transported by boat to cows bunched on patches of unsubmerged levee crest, or confined to barns upraised on piles or waterborne on rafts (fig. 6.13). The main native forage, canarana (Echinochloa polystachya), which thrives in white water, is indispensable for getting through critical inundations. When not consumed by cattle, it and other grasses are set on fire after the flood subsides, nutrients carried in the turbid water then being incorporated as ash into the soil.
Although the humanized landscape of the Careiro represents an alteration of the wetlands, it is far from the radical make-over some might find desirable. Swamps, marshes, and forest-covered tropical floodplains in their natural state tend to have few champions. The most extreme transformation of the aquatic-terrestrial ecosystems is, of course, to take the "wet" out of "wetlands." The word saneamento (from L. sanare, to restore to health), as used in Brazil, is most aptly translated by the English "reclamation." In Western thinking, improvement and proper use of wetlands often is equated with exsiccation. Perhaps this mindset derives, in part, from the widely acclaimed achievements of Dutch hydraulic engineers. Generations of dykemasters from the Netherlands have been called to work all over the world (Veen, 1962) and it is not surprising that, early in this century, a Dutch engineer was hired to study the manner in which Marajó Island might be drained.
The optic of certain ancient civilizations, such as those that flourished on the east coast of Mexico, appears to have been very different. The density and achievements of these pre-Columbian populations do not seem compatible with what, to modern eyes, appears as an inhospitable, waterlogged region. But, instead of removing water, the natives chose to discipline it and use it to supply protein in the form of aquatic fauna, while raising ridges, upon which they presumably produced the necessary complements of carbohydrates (e.g. Flannery, 1982; Harrison and Turner, 1978; Turner and Denevan, 1985). Water, properly organized, represented a resource, not an obstacle.22
A major hydrogeomorphic transformation of wetland and lacustrine environments along the Amazonas was attempted by the federal Agronomic Institute of the North in 1949, when it set out to effect, on a gigantic scale, what várzea farmers sometimes accomplish on their property: the diversion of silt-laden floodwaters into lowlying tracts. The target was Lake Maicuru, also known as Lago Grande de Monte Alegre, or, simply, Lago Grande (see fig. 6.2), between Santarém and Monte Alegre, Pará State (Camargo, 1958; Sioli, 1951). Eventually, the lacustrine depression was to have been brought under cultivation, but it was thought that, even during the initial stages of siltation, water buffalo could be raised there. A number of canals from the Amazonas were cut. The largest, completed in 1953, was 30 m across and 6 m deep at bank full stage, and in 1954 was discharging 275 m³/sec. (Camargo, 1958). Recently, the President of the Society for the Preservation of Amazonia's Natural Resources issued a warning, based on his perception that, as water pours in, Lake Maicuru is expanding, swallowing up small neighbouring lakes and streams, and threatening the natural levees that separate it from the Amazon (O Liberal, 1989).
The most ambitious operation hitherto attempted to remake Amazonian wetlands was begun in the mid-1970s by São Raimundo Agro-industrial Ltda. (SRAL). This was a division of the empire established by US billionaire Daniel K. Ludwig on the banks of the Amazon and its left-bank tributary, the Jari. With Amazonia envisaged as Brazil's most promising rice bowl, the project to cultivate the grain on 12,700 hectares of wetlands used powerful earthmoving equipment to build 460 km of dikes and 250 km of canals. Agricultural operations were mechanised, employing heavy farm machinery; seeding, and the application of fertilizer, herbicides, and pesticides were carried out by aeroplanes (SRAL, n.d.). However, weed control is very difficult in the Amazonian wetlands. An exploratory survey had signalled the presence of several potentially dangerous weeds in floodplain rice fields and the difficulty of targeting them selectively (Barrett, 1975). Also, elevated temperatures, coupled with high humidity, favour pests and diseases. Indeed, biocides did prove to be a significant part of production costs on the 3,500 ha eventually brought under cultivation (SRAL, 1984).
In 1982, a Brazilian consortium incorporated the Companhia do Jari, with the objective of taking over Ludwig's tree farm, pulp plant, and mineral deposits. Two years later, it also acquired the rice operations. The buyers have stated that, although aware of the serious problems faced by SRAL, they were motivated by the social merit of an enterprise dedicated to the production of a food staple, and by the desire to complete the "nationalization of 'Projeto Jari'." Four years and US$ 13 million later, officers of the Companhia became convinced that "the degree of technological sophistication" built into the gigantic project was just "not compatible with rice farming in Amazonia," stating that this should be undertaken "with simple and inexpensive processes that rely more on people, less on machines" (Companhia do Jari, 1988).
