Cover Image
close this bookAmazonia: Resiliency and Dynamism of the Land and Its People (UNU, 1995, 253 pages)
close this folder5. Silviculture and plantation crops
View the document(introductory text...)
View the documentSilviculture for pulp
View the documentRubber plantations
View the documentOil-palm Oil palm
View the documentCoconut

(introductory text...)

Silviculture for pulp
Rubber plantations
Oil-palm
Coconut

Agriculture, livestock raising, and silviculture are often pinpointed as major factors in the destruction of Amazonian forests. But human occupation of the region does not necessarily provoke a downward ecological spiral. Many indigenous groups have prospered with relatively sophisticated cultures for protracted periods of time in various parts of Amazonia. Answers to the difficult question of promoting sustainable agriculture in Amazonia will come from a blend of indigenous knowledge and modern, scientific input. Also, a blend of native and exotic plants will provide the mainstay for farmers, ranchers, and plantation operators. Dynamic and successful agricultural and silvicultural systems are always open to new opportunities for incorporating promising plant and animal species, wherever they may come from.

A major focus on agriculture is justified because approximately half the people in Amazonia still live in rural areas and depend on farming, ranching, plantations, and forest resources for a living. Also, a growing number of forestand agriculture-based industries are being established in Amazonia, thereby creating employment opportunities in urban areas.

Sustainable agriculture requires flexible farming systems. No particular mix of crops or agronomic practices in a given area is likely to endure for long. New crops and varieties must be constantly introduced to raise and sustain productivity, and novel agronomic practices adopted in response to changing market and environmental conditions. Such practices are well developed in Amazonia among small-, medium-, and large-scale operators.

Sustainable agriculture is a multifaceted concept, involving both ecological and socio-economic and political dimensions (Smith 1990). A critical defect in much of the discussion about sustainable development in Amazonia, particularly as it applies to agriculture, is the lack of attention to economic realities. Ecologists provide valuable insights with regard to the role of forests in environmental stability and as food sources of pollinators, but ecological perspectives alone will not lead to sustainable development. Also, social scientists tend to stress policy and sociocultural dimensions to development; as important as these perspectives are, they need to be tempered by economic realities.

Our analysis does not suggest that technological fixes will solve sustainability issues in Amazonia. Rather, we stress the need to conserve and manage natural resources while designing agricultural systems that provide some hope of generating income for farmers, ranchers, and plantation owners.

Amazonia can be a graveyard for over-ambitious agricultural development schemes. Derelict rubber (Hevea brasiliensis) plantations, weed-choked pastures, and the early failure of some small farms along pioneer highways are testament to the folly of attempting to establish farming systems that are not attuned to ecological or market realities. Although ecological constraints were factors in the demise of some agricultural development schemes, mismanagement and a distorted policy of fiscal incentives are also to blame. The preponderance of ecological and socio-economic factors in farming difficulties varies with time and location. A major point here is that, with proper management and an openness to incorporating new technologies as well as learning from traditional systems, many constraints to raising and sustaining yields can be overcome. The relative success of many farms and plantations on the uplands attests to the resilience of some agricultural systems that have been deployed.

Some positive developments are under way in Amazonian agriculture and silviculture, indicating that environmental constraints are being surmounted. For example, Jari's plantations of exotic trees for pulp production appear to have turned the corner with respect to increased productivity and profitability. Cash-cropping with several perennial crops is breathing new life into the Bragantina zone east of Belém, an area once thought to be at the terminal stages of slash-andburn farming. An emphasis on what appears to be going right for at least some Amazonian farmers and plantation operators should provide some insights into the development process and could generate helpful information for policy makers.

Our analysis of sustainable agriculture and silviculture thus focuses mainly on perennial cropping systems, mostly geared to generating cash income. An emphasis on income generation is justified by the evolutionary trend of many subsistence systems towards cash-cropping (duo 1989). Jari's experience with pulp production, and the efforts of various companies to establish oil-palm and coconut plantations are discussed here. In subsequent chapters we explore agroforestry systems, ranching, and the potential of the flood plains for increased food production.

Silviculture for pulp

The Jari project, originally owned by the shipping magnate Daniel Ludwig, has been controversial since the ambitious project began in 1967. The immense size of the concession, 1.6 million ha, understandably fanned nationalistic feelings within Brazil. Plans to establish monocultures of fast-growing exotic trees for pulp production stirred concerns about creating ecological deserts in Amazonia and other regions.

Ludwig invested close to a US$1 billion in the Jari project before selling out to a consortium of 23 Brazilian companies and banks in 1982 (Kelly and London 1983: 282). Jari is often portrayed as a failure and as an undesirable development model for Amazonia (Browder 1989b; Parker et al. 1983; Schmink 1988b; Szulc 1986; Uhl, Jordan, and Herrera 1982), but much valuable experience has been gained at Jari that could be useful for environmentally sound development in the humid tropics.

Many of the early failures of the Jari project can be attributed to inept management and to a disregard of the scientific literature on mechanical clearing of tropical soils and silvicultural experiments (Palmer 1986). In the 1960s and the 1970s, management seemed to be operating on a revolving door principle (Kinkead 1981). Ludwig fired as many as 30 directors at Jari (Hecht and Cockburn 1989). Lack of continuity and a slowness to learn from early errors retarded Jari's progress. Under Ludwig's tenure, heavy machinery was used to clear forest for plantations of Gmelina arborea as well as to harvest the broad-leafed trees introduced from tropical Asia. Bulldozers and other machinery scraped away thin and valuable topsoil and compacted the subsoil (Greaves1979).

Now plantations are harvested by hand-held power saws. Whereas in the past some steep slopes were cleared for plantations, many of these erosionprone ravines are being allowed to revert to forest. Soil erosion no longer appears to be much of a problem in plantations at Jari. In Jamaica, the dense mat of needles in commercial stands of Caribbean pine has actually reduced soil erosion rates compared with native forests (Richardson 1982).

At first, Gmelina was envisaged as the main plantation tree for pulp production at Jari. But yield projections for this sun-loving Asian tree were not met at Jari. Gmelina produces reasonably well only on the better soils, mostly alfisols (terra roxa). In Thailand, where G. arborea is native, the species occurs in mixed deciduous forest growing on loamy soil, either calcareous or granitic (Smitin-and 1989). Gmelina fares poorly on the sandier ultisols and heavily leached oxisols found in much of Jari and other parts of Amazonia, and is susceptible to insect and pest attack, such as ceratocystis wilt. Caribbean pine (Pinus caribaea var. hondurensis; fig. 5.1) was brought to Jari in 1973 and two species of eucalyptus (Eucalyptus deglupta and E. urophylla; fig. 5.2) were introduced in 1979 because they are fast-growing, hardier, and better adapted to poorer soils (Fearnside and Rankin 1982). By the late 1980s more eucalyptus species were being tried at Jari, including hybrids.

In 1989, Jari had some 85,000 ha in plantations. Pine and eucalyptus dominate the planted area, a reflection of the generally poor soils (JARI 1989). In December 1989, pine occupied 42,832 ha, eucalyptus covered 22,749 ha, while Gmelina was planted on 19,700 ha, the last down from a high of 64,000 ha (Palmer 1986). The proportion of new plantings to eucalyptus will continue to grow, while the relative contribution of Gmelina will decline further. Eucalyptus accounted for only 3 per cent of the plantation area in 1984 (Hornick, Zerbe, and Whitmore 1984), but by 1989 the two eucalyptus species occupied 27 per cent of the plantations. Gmelina will still be planted for the foreseeable future, however, because some customers prefer its smooth paper for specialized uses, such as airline tickets (Refer 1990).

In 1989, the average yields of Jari's plantations were modest, about 13 m3/ha/year for pine, 15-17 m3/ha/year for Gmelina, and 19 m3/ ha/year for eucalyptus. Nevertheless, productivity is increasing with the deployment of improved genotypes and the adoption of new agronomic practices. Eucalyptus yields of 42 m3/ha/year were not unusual by the late 1980s, and in 1990 several plots of E. grandis were producing 53 m3/ha/year. The latter species is so precocious at Jari that it attains 30 metres within six years. Eucalyptus yields of 60 m3/ ha/year are no longer unusual in the humid tropics and are far greater than the potential of plantations in temperate climates (Evans 1986). In the mid-1980s, for example, Aracruz was regularly achieving eucalyptus yields of 70 m3/ha/year in south-eastern Brazil (WRI 1985a: 39), and, by selecting a mix of superior clones, yields reached as high as 100 m3/ha/year (Zobel, Van Wyk, and Stahl 1987: 142).

Deployment of superior genotypes is also the main strategy for increasing the productivity of plantations at Jari. Seed of Eucalyptus urophylla was originally obtained from Flares in Indonesia, but Jari has recently been purchasing seed from other companies in Brazil. Jari anticipates producing its own E. urophylla seed in the near future. The eucalyptus species employed at Jari do not coppice after cutting, thus enabling foresters to introduce fresh germ plasm after each five-sevenyear cutting cycle.

Jari is drawing on the experience of Aracruz in Espirito Santo and Bahia with vegetative propagation of especially fast-growing eucalyptus material. Aracruz has planted over 60,000 ha of eucalyptus in south-eastern Brazil, much of it in clones. In 1979, for example, Aracruz had rooted 1 million eucalyptus cuttings, but by 1984 some 15 million eucalyptus clones were established. To reduce the danger of disease or pest epidemics, no more than 100 ha are planted to any one clone in the same area (WRI 1985b: 50). From 10 to 20 ha per clone appears to be the best compromise between reducing pest and disease attack and easing the task of planting and harvesting (Zobel, Van Wyk, and Stahl 1987: 213).

Similarly, Jari is conducting trials of over three dozen promising Gmelina selections in order eventually to plant various high-yielding clones in different blocks. In this manner, ecological heterogeneity will be maintained while raising yields. Gmelina cuttings root readily, thereby eliminating the need to graft on rootstock. In 1988, Jari began producing its own pine seed from controlled crosses on a 100 ha orchard near Morada Nova de Minas in Minas Gerais, an area far removed from other pine plantings and therefore free of pollen contamination.

A similar strategy of broadening the genetic base of plantations is under way on the savannas of Amapá, where AMCEL has planted 90,000 ha of tropical American pines. About 90 per cent of the plantings at AMCEL are Pinus caribaea var. hondurensis, with P. oocarpa, P. caribaea var. Caribaea, and P. tecunumanii comprising the remainder. The AMCEL plantings are cut on an eight-year cycle and the logs are sent to the nearby Jari pulp mill. In the past, germ plasm was obtained from scattered populations in Honduras, Guatemala, and first-generation plantings in Fiji and Brazil (McDonald and Fernandes 1984). AMCEL currently obtains its planting material from one of the Jari companies that maintains a nursery in Minas Gerais and from a forestry research institute (IPEF- Instituto de Pesquisas Florestais) in the state of São Paulo. The savanna soils under AMCEL management are poor in nutrients, but have suitable physical structure for silviculture and agriculture. To maintain yields, nitrogen and phosphate fertilizers are applied periodically to the soils in AMCEL's pine plantations.

On the agronomic front, weeding, control of ubiquitous leaf cutter ants (species of Atta and Acromyrmex), and fertilizers are necessary to achieve high yields. Weeds are controlled by herbicides (3 litres Roundup and 1 litre DMA per ha) and tractor mowing. In some cases, helicopters spray preemergent herbicides before planting. Herbicides are cheaper to use than mechanical weeding; also, unwanted plants return more quickly when they are cut. While plantations are becoming established, two herbicide applications a year are usually needed.

Leaf cutter ants are the most serious pests of recently planted Gmelina and eucalyptus. Leaf cutter ants can be serious agricultural pests in other parts of Amazonia and contiguous regions, sometimes forcing farmers to abandon fields (Butt 1970; Goncalves 1957, 1967; Lathrap 1968; Nepstad, Uhl, and Serrão 1990). If fields are closer than 1 km apart, leaf cutter ants are more likely to find cleared areas in the forest during nuptial flights (Fautereau 1955). Jari's large, open spaces are ideal grounds for the build-up of leaf cutter ant populations.

Leaf cutter ants probably became a scourge to farmers soon after the dawn of agriculture in Amazonia. The ants have even been incorporatedin mythology; the Desana of the Uaupes River, for example, believe that the "saliva" of the moon helps keep leaf cutter ants away from manioc fields (Reichel-Dolmatoff 1974: 73). Saúva, as the ants are called in Brazil, were reported as a common agricultural pest in the Amazon in the nineteenth century (Bates 1863: 23) and in Bahia during the sixteenth century (Sousa 1971: 173).

Traditional control measures against leaf cutter ants have been non-existent or largely ineffective. In most cases, fields seriously affected by the ants are simply abandoned. Various control measures were attempted against leaf cutter ants in Bahia during the 1500s: nests were sometimes dug up and burned; the nests were compacted to make it difficult for the ants to gain access to their subterranean chambers; or manioc leaves were spread along the ants' trails to distract them (Sousa 1971: 173). As leaf cutter ants became an increasingly serious problem in coffee groves along the Paraíba valley between Rio de Janeiro and São Paulo after 1830, fires were built by the ants' nests. Bellows were also used to force smoke into the nests, but such practices sometimes damaged the coffee bushes (Stein 1985: 217). In northwestern Amazonia, leaf cutter ants are savoured as a culinary item, but human predation of Atta ants has not impaired their reproductive potential - a leaf cutter ant nest can contain more than 1 million individuals.

Jari staff use three pesticides to control leaf cutter ants. Mirex granules are placed by nests, while white Termicidol powder, which contains Aldrin, is doused into nests. Biabine gas is pumped into the nest galleys. The principal active ingredient in Biabine is methyl-bromide. Mirex has not been sprayed by air in the United States to control fire ants since 1978, a year after it was found to be a carcinogen; Amdro is now used for fire ant control in the United States, but it costs four to five times as much as Mirex (Jerry Stimac, pers. comm.). In 1990, 160 people, split up into several teams, were employed to destroy leaf cutter ant nests in Jari's plantations. Leaf cutter ant control thus requires about 1 person per 530 ha at Jari. In the late 1970s, Jari was spending US$1 million a year on the control of leaf cutter ants (Palmer 1986).

In view of the potential environmental hazards of pesticides, more research on biocontrol or less toxic control methods is warranted. The Kayapó, for example, scrape a forest vine (Tanaercium nocturnum, Bignoniaceae) to kill African bees five minutes prior to extracting honey from the hive (Kerr 1986). This vine can kill leaf cutter ants within a minute and, when placed in the nest, eliminates leaf cutter ants, or reduces them to such low numbers that they are no longer significant pests.

Some 100-150 kg of triple superphosphate are applied per ha to establish eucalyptus and pine plantings. In the case of Gmelina, 70 g of NPK fertilizer are applied at planting time. Pine orchards are fertilized again two to three years after planting. Experiments are under way to see if the addition of lime would reduce the need for relatively expensive phosphate fertilizer. Unfortunately, some of the lime has come from shell mounds near Santarém, thereby partially destroying some valuable archaeological sites (Anna Roosevelt, pers. comm.). Large quantities of Carboniferous limestone occur at Monte Alegre and along the middle Tapajós; such deposits could eventually be exploited for lime fertilizer.

To help raise and sustain plantation yields, Jari has created a superintendency for forestry research with a staff of 33 people. The research division is divided among several departments, including genetic improvement, silvicultural research, plant pathology, soils, plant nutrition, and native forest resources. Consultants are also employed to tackle specific problems, but an in-house scientific capacity provides continuity and greater familiarity with the overall project.

The investment in science is paying off at Jari. The pulp mill, towed from Japan in 1978, functioned mostly at a loss until the main boiler blew up in July 1988. Since the mill came back on line in March 1989, operations have been in the black. Some of the wood for pulp production is being met by bringing in Caribbean pine logs from the AMCEL plantation in Amapá, but Jari expects to be self-sufficient in wood production within a few years. From March to December 1989, the Jari mill produced 230,000 tons of pulp, mostly destined for export. By 1993, annual production of pulp at Jari had reached 330,000 tons per annum (Blount 1993).

Enormous investments were necessary at Jari to turn the corner. A 9,000 km road network was laid out, a 68 km railroad was built to transport harvested logs to the pulp mill, several residential communities were created with support services, and experiments were carried out over two decades. Jari now appears to be a sustainable operation, welcome news for the 5,500 employees and for the 60,000 people who depend indirectly on the enterprise. Plantations can clearly make a dent in rural poverty by providing employment opportunities.

Jari started becoming profitable in 1990 largely because of improved planning, research, and streamlining operations. The current management team took advantage of lessons learned during the learning curve when Ludwig held the helm at Jari. The current consortium operating Jari recently made a profit of US$5 million on pulp sales of US$160 million (Alvarenga 1991). Jari has dropped its cattle operation, including the intercropping of pasture and pine, to concentrate on its core business: pulp production. Jari terminated a highly mechanized rice operation on the Amazon flood plain, owing mainly to the high cost of controlling water levels in rice paddies using diesel pumps, and to low rice prices in the 1980s. The former rice area around São Raimundo has been sold to Grupo Uliana, and this consortium maintains a herd of 13,000 water buffalo on the flood plain. Grupo Uliana also plans to raise pirarucu (Arapaima gigas) in ponds and to operate a hotel emphasizing eco-tourism. The kaolin mine, in operation for over a decade, has been profitable since its inception in the mid-1970s; in 1991, the mine turned a US$12 million profit (Alvarenga 1991).

Some are still concerned that Jari is not an appropriate model for Amazonian development or that profitability does not take into account resource depletion (Vosti 1993). Jari was never intended to be the model for development in Amazonia or elsewhere. Rather it contributes to a mosaic of land uses in the region. The criticism that profits do not consider such factors as soil erosion and loss of biodiversity could be applied to most development schemes throughout the world. Overall, Jari is helping to conserve the environment by maximizing production from a relatively small proportion of the property and by taking stringent safeguards to avoid water pollution.

The idea that Jari is causing major ecological impacts, and therefore cannot be considered a model for agro-silvicultural development, is without merit. Jari anticipates clearing only 145,000 ha, less than 10 per cent of the property. Effluent from the paper mill passes through a series of large settling ponds before it is discharged into the Jari River. When the second pulp mill comes on line in the mid-199Os, the settling ponds will occupy some 450,000 m2, roughly the area covered by 63 soccer fields.

In new plantations established inland from the left margin of the Jari River, extensive corridors of forest, at least 400 metres wide, have been left to provide faunal bridges (fig. 5.3). A similar approach is being studied in Florida to connect otherwise isolated patches of forest (Harris 1984). Only flat areas have been cleared for plantations. Forest patches provide windbreaks for young trees and fire-breaks; in 1987, 600 ha of pine were lost in a lightning-caused fire.



Fig. 5.3 Corridors connecting forest in a portion Jari's pulpwood plantations.. Morro do Felipe, Amapá. Blank areas are forest

The proposed 100 MW hydroelectric plant at the Santo Antonio falls of the Jari River was expected to be ready in the mid-1990s and will supply energy for the pulp mills. The energy needs of the mill are currently met by pulp wastes, diesel oil, and to a lesser extent fuel-wood. Instead of a dam, the hydroelectric station will divert some of the Jari River through a turbine. Fisheries will not be disrupted and the scenic beauty of the basalt falls will remain intact.

Although few would argue that the vocation of Amazonia is converting all of the forest to plantations, the Jari experience could prove useful as world demand for pulp continues to rise and methanol becomes an economically viable fuel by the late 1990s. The cost of pulp production in industrial countries is increasing owing to environmental restrictions. In Scandinavia, for example, ever more stringent emission controls and pollution taxes are driving up the cost of operating pulp mills. In the north-western United States, access to some old-growth timber is being denied to protect mature ecosystems that contain endemic species. The construction costs for a modern pulp mill with a capacity of 1,500 tons of pulp a day are approaching US$1 billion, so more pulp operations are likely to shift to the humid tropics where operating costs are generally lower (Benoit 1990).

Methanol can be made from fossil fuels, such as natural gas, or from biomass, particularly wood. Brazil's current alcohol fuel programme is based on ethanol production from sugar cane. Ethanol from sugar cane, maize, or manioc requires setting aside some of the better farmlands for fuel production. An advantage of methanol production from wood, such as eucalyptus, is that more marginal areas can be used, thereby reducing competition for space to grow food crops (Smith 1981b,c). When petroleum prices rise again, methanol may become an economically attractive fuel for automobiles and trucks in some countries, such as Brazil.

Rubber plantations

In the 1950s, two sizeable plantations of rubber were established in the Bragantina zone. Pirelli developed a 7,000 ha property, while Goodyear obtained a 4,500 ha holding. Both plantations were soon attacked by South American leaf blight (SALB), which is caused by an endemic fungus, Microcyclus ulei. Most commercial rubber plantations in the Amazon are double-grafted to reduce disease problems and to improve yields. A relative of the rubber tree, Hevea pauci-flora, is often employed for the crown because it resists SALB. Still, yields do not compare with those prevailing in SouthEast Asia. At the Goodyear plantation near São Francisco in the Bragantina zone, for example, latex yields average only 300 kg/ha/yr, less than onethird of those achieved on Goodyear plantations in Malaysia and the Philippines where groves are free of SALB (Goodyear 1989: 28). Furthermore, operating costs are higher in the Amazon, particularly with respect to labour and fertilizer.

Rubber fares better as an intercrop in Amazonia, because disease pressure is less severe when the trees are dispersed among other plants. But the immediate outlook for rubber in Amazonia is dim. As mentioned in the discussion of extractive reserves, the Brazilian government is essentially eliminating all subsidies for natural rubber. At best, rubber may have a future only as an intercrop with small farmers, or on well-managed plantations in southern Brazil where disease problems are less severe.

Fiscal incentives promulgated since 1950 have spurred much of the rubber planting in Brazil. For example, SUDHEVEA (Superintendencia da Borracha) was established in 1967 to promote rubber planting, to support rubber tappers, and to improve product quality. In 1989, SUDHEVEA's functions were taken over by the environment agency IBAMA (Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis) before SUDHEVEA had accomplished its goals. Besides, IBAMA is concerned more with nature preservation and wildlife regulations than with stimulating rubber production.

PROBOR (Programa de Incentivo a Produçao da Borracha) was created in 1972 to help Brazil become self-sufficient in natural rubber within a decade. Yet in 1974 Brazil imported 41 per cent of its natural rubber needs, and in 1989 the proportion had climbed to 74 per cent (Goodyear 1989: 26). PROBOR was phased out in 1985, when SUDHEVEA was charged with providing technical assistance only. Even though a suite of fiscal incentives, taxes, and subsidies did little to redress the balance of imports of natural rubber to Brazil, investments were made in rubber plantations in various parts of Brazil. Today, planted rubber covers at least 40,000 ha, mostly in São Paulo, Espírito Santo, Bahia, Mato Grosso, and Pará. If rubber is to remain a viable commercial crop in Amazonia, more research will be needed on resistance to SALB, among other diseases and pests.

Oil-palm Oil palm

In contrast to the rather gloomy prospects for rubber in Brazil, an exotic plantation species shows greater promise. African oil-palm is well adapted to most soils and climates in Amazonia. Pará has the largest plantings of African oil-palm in Brazil (table 5.1), but at least 50 million ha in the Brazilian Amazon are deemed suitable for the quick-growing palm (Alvim 1989). Brazil imports palm oil and the market for vegetable oils is growing. Furthermore, palm oil can serve as a substitute for diesel fuel, so both the short- and long-term prospects for oil-palm are bright.

Table 5.1 Area planted to African oil-palm in Brazil, 1991

State Area planted (ha)
Pará 32,000
Amapá 4,300
Amazonas 2,000
Bahia 15,000
TOTAL 53,300

Source: Falesi and Osaqui (1992)

The Bragantina zone has the oldest African oil-palm groves in Amazonia, on both small- and medium-sized holdings. The largest plantations are operated by DENPASA (Dende do Pará, S.A.), which inaugurated an oilprocessing factory along the Belém-Mosqueiro road in 1976 with fiscal incentives provided by SUDAM (Superintendência do Desenvolvimento da Amazonia). DENPASA currently operates two oil-processing factories with a capacity of 20 tons fruit/hour. DENPASA imports planting material from a British company that uses micropropagation techniques to multiply highyielding varieties. DENPASA's 6,000 ha are mostly sandy, so fertilizers are applied according to foliar analyses. Soils are enriched with nitrogen by a thick mat of planted leguminous creepers (Pueraria javanica and P. phaseoloides; fig. 5.4). DENPASA processes oil-palm year-round and provides steady employment for about six persons per hectare.

Following the lead of DENPASA, several entrepreneurs have established three small-scale African oil-palm processing factories in the Bragantina zone near Belém and Castanhal. These oil-palm processing plants, capable of processing 6-7 tons fruit/hour, service 10-20 ha plantations, often in a cooperative arrangement. At least 3,000 ha of African oil-palm are planted on smallholdings outside of DENPASA's property. The small-scale oil-palm processing facilities use boilers fuelled by fruit stalks, cracked kernels, and wood. Fruit stalks are sometimes mulched for later distribution around oilpalms, while the pressed kernels are sold for livestock feed. Oil-palm production is thus linked to the livestock sector, and provides employment opportunities in rural areas.

A cloud nevertheless looms on the horizon for oil-palm plantations in the Bragantina zone and other parts of Latin America. Spear rot, possibly caused by a viroid or mycoplasma-like organism, appeared in the Bragantina zone in 1985. Spear rot stymied an attempt by SOCFIN to establish an oil-palm plantation near Tefé in Amazonas in the late 1970s. Only 2,000 ha of oil-palm were planted by the Belgian company near Tefé, a fraction of the original plans. In the Bragantina zone, only the DENPASA plantations have been struck by the disease as of 1990, when some 400 ha were affected. By 1992, 1,500 ha of DENPASA's plantation near Mosqueiro in the Bragantina zone had been destroyed in an effort to control spear rot. Some 7,500 ha of oil-palm in the Bragantina zone are thus in imminent danger of spear rot.

Some 200 km south of Belém 13,900 ha of recently planted oil-palm are also at risk to spear rot. In 1989, a subsidiary of Banco Real (CRAI Companhia Real Agroindustrial) purchased 3,340 ha of oil-palm established by Agromendes; the latter was a subsidiary of Mendes Junior, a giant construction firm, which diversified into agriculture with fiscal incentives. In 1986, Agromendes and CRAI owned some 6,000 ha of oil-palm in the vicinity of km 74 of the PA 150 highway (Barcelos et al. 1987). As of 1992, CRAI and the absorbed Agromendes, now renamed Agropalma, cultivate 9,210 ha of oil-palm on adjacent properties. Nearby, DENPASA maintains a 2,500 ha plantation of oil-palm, which is processed by Agropalma. Approximately 100 km north-east, along the PA 252 road, REASA operates a 2,500 ha oilpalm plantation. The Moju-Acará area, with 14,210 ha of oil-palm, thus had the greatest concentration of the crop in Pará, in 1992.

CRAI and Agromendes plan to plant a further 10,000 ha on their 33,000 ha property. Roughly half of CRAI and Agropalma's property is slated to remain in forest in order to comply with Brazil's environmental regulations. Some 20 hybrids are planted, with an average yield of 22 tons fruit/ha/yr from the 5,000 ha of plantations in production. Even this diversity of planting material may not save the plantations from spear rot. Banco Real has invested approximately US$70 million in oil-palm along the PA 150 highway, and failure of the oil-palm plantations at CRAI and Agropalma would put 1,000 people out of work.

Inputs at the oil plantations along the PA 150 highway are kept to a minimum, and environmental impacts appear to be negligible. Agropalma maintains a 100 metre strip of forest on either side of streams on the property and plants only on relatively flat surfaces. No pesticides are used on the plantations. A company agronomist suggests that the surrounding forest and second-growth communities provide biocontrol agents (W. Padilha, pers. comm.). A lepidopteran larva, Anteotricha sp., started attacking oil-palm leaves in 1991, but the caterpillar is apparently controlled by a fungus. Native a,ca' palm (Euterpe oleracea) appears to be one of the pest's hosts.

Pressed fruit cake from the processing plant is recycled to the plantations, which have a ground cover of nitrogen-fixing Pueraria. Fertilizer applications are carefully calibrated according to foliar analyses. Provided that spear rot does not appear along the PA 150 highway, Banco Real's oil-palm plantations are likely to succeed. As in the case of the Bragantina zone, the presence of large oil-palm plantations and processing plants could eventually provide a stimulus for small-scale producers to diversify their crop base with oil-palm. Already, Pará, accounts for almost two-thirds of Brazil's oil-palm production (Falesi and Osaqui 1992).

More than 20,000 ha of oil-palm established in the Huallaga valley in the Peruvian Amazon since 1967 are also threatened by spear rot (Pulgar 1987: 134). If spear rot strikes the Huallaga valley, it will deal a severe blow to the Peruvian government's efforts to find viable alternatives to coca production.

The disease, known as amarelecimento fatal in Brazil and marchitez sorpresiva in Peru, is currently managed with varying degrees of success by early identification of symptoms and prompt elimination of diseased trees. Spear rot is thought to be transmitted by insects, but fogging with insecticides would destroy beetle pollinators. The separation of oil-palm plantations by stretches of second growth and other crops may retard the spread of the disease. Black vultures (Coragyps atratus) relish ripe oil-palm fruits, so plantations are best kept away from large towns and cities where large populations of the vultures congregate.

Genetic resistance to spear rot is the best way to combat the disease, and a near relative of African oil-palm that grows wild along the Amazon flood plain promises to help in this regard. American oil-palm (Elaeis oleifera) resists spear rot and has already been used in crossing work with E. guineensis to reduce the latter's height and thereby facilitate harvesting. American oil-palm has also been employed by breeders to improve the oil quality of African oilpalm. Several hybrids between prostrate caiaue, as the palm is known in Brazil, and African oil-palm are under observation at a sizeable oilpalm breeding and evaluation site operated by EMBRAPA (Empresa Brasileira de Pesquisa Agropecuaria) along the Urubu River north of Manaus. Caiaué's resistance to spear rot underscores the importance of conserving wild habitats in Amazonia.

Coconut

Another exotic palm, coconut, has emerged as a viable cash crop in eastern Amazonia, for both small- and large-scale producers. Fresh coconuts from smallholdings sell briskly at local markets, whereas coconuts from commercial plantations are sent to Belém and northeastern Brazil to be processed for the confectionery trade.

A Brazilian firm based in Maceio, Sócôco, is establishing a 5,000 ha coconut plantation near Belém and has planted 3,600 ha of coconut south of Belém at Acará. The plantation at Acará began in 1981 with support from the International Finance Corporation and SUDAM. Sócôco's plantations at Acará are relatively homogeneous since they contain only a few, high-yielding selections, dominated by PB 121 and PB 111, with 70 and 15 per cent of the planted area, respectively. Genetic homogeneity at Acará has not led to any serious pest or disease problems thus far. Selections planted by Sócôco contain Malayan Dwarf germ plasm, so lethal yellowing - which has ravaged coconut plantations in parts of Central America and the Caribbean - is not a threat. Marchitez, caused by a protozoan, is the main disease problem. After marchitez, the second most important cause of coconut tree mortality is lightening. An unidentified fungal disease of the leaves, called queima de folha, can also be a problem.

Yields at Acará on eight-year-old plantings with a density of 160 trees/ha are in the order of 10,000 coconuts/ha/yr. Fertilizer is needed because the yellow oxisols are infertile. Leaf analysis is performed in Montpellier once a year to calibrate fertilizer doses; seedlings are given 200 g urea, 300 g K, 15 g Borax, and 800 g super-phosphate when planted. Salt is added to the fertilizer mix after the third year to facilitate the uptake of nutrients. Nitrogen fertilizers are no longer applied after the third year as the Pueraria javanica ground cover, which was introduced from Cote d'Ivoire, becomes well established. Pueraria also suppresses weeds, an important consideration in an area receiving around 3,000 mm of rain a year.

Only a small fraction of Sócôco's 17,000 ha property at Acará has been cleared. The vicinity of the plantation is not densely settled, but neighbouring communities benefit from the 600 jobs generated by the Acará coconut operation. A pasture of Brachiaria humidicola is maintained to supply milk for some of the employees; this pasture, already 12 years old in 1991, also supports some 50 mules, which are used to take coconuts to waiting trucks (fig. 5.5).

A 500 ha coconut plantation, operated by TABA, a regional airline, was established near Mosqueiro on the Bragantina coast without fiscal incentives. The TABA plantation contains approximately 50,000 trees and produces some 30,000 coconuts a month. The plantation is genetically heterogeneous, since most of the trees originated from seedlings collected from villages in widely scattered locations in the Bragantina zone and on Marajó. The TABA plantation is gradually becoming more homogeneous, however, as dead trees are replaced by selections made at the plantation.

The only significant pests thus far are beetles, particularly Rinco-pharum palmarum. A well-established ground cover of pueraria protects the soil and provides nitrogen to the palms (fig. 5.6). TABA is diversifying its coconut operation by experimenting with intercrops, such as lime, soursop, and guava (fig. 5.7). Also, several small reservoirs have been created to raise pirarucu, a highly valued fish in local markets.