Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Tree crop farming in the humid tropics: Some current developments

G.A. Watson
IDRC Consultant, Sussex, England


Tree crop farming systems in the humid tropics are in an active state of development In Sri Lanka, Nigeria, and Latin America attention is being devoted to improvement of traditional tree food crops. In larger-scale rubber and oil palm cultivation monoculture is still the norm, but in the smallfarmer sector there is increased interest in intercropping with food crops In coconut, the introduction of high-yielding hybrids, complemented by the development of multicropping systems, is leading to a major upward revision of this crop's potential. Studies of taungya farming have confirmed the value of food cropping in young forest tree crops

So far research has concentrated on the commodity crops and many lessons have been learned which could be applied in the future development of agro-forestry.


Spurred by increasing social, population, and economic pressures, tree crop farming systems in the humid tropics are in a state of very active development. In the commodity sector, rubber, oil palm, coconut, and cocoa plantings are being extended; there is greater use of high-yielding material, more productive methods of management, and shorter replanting cycles, together with increased interest in intercropping. In the traditional multicrop farming sector, there are attempts to establish more productive crop mixes, and the improvement and cultivation of food tree crops are being encouraged. So far, the bulk of detailed research on tree crops has been concentrated on commodities, and many lessons have been learned of possible application in the production of forest and food crops.

Traditional Tree Crop Farming Systems

The traditional farming system of the humid tropics typically involves, in small plots of not more than a few hectares, a mix of trees in association with a variety of perennial and annual crops. Together, they provide food and income for the smallholder farmer. In Indonesia, there is the homestead farm, with multistoreyed crops: the top storey of coconut, middle storey of cloves, citrus, cinnamon, jackfruit, bananas, and pepper, and ground storey of ginger, groundnuts, maize, patchouli, cucurbits, and others. In the West African compound farm, the tree component includes oil palm, cononut, cola nut, citrus, fruit, and cocoa, grown with an understorey of cassava, plantains, yams, groundnuts, and vegetables (Lagemann et al. 1975).

Perhaps the most intensive tree cropping system is found in the Sri Lankan Kandy gardens. These are small farms based on a close association of coconut, kitul, and betel palms with cloves, cinammon, nutmeg, citrus, mango, durian, jackfruit, rambutan, and breadfruit, with a lower storey of bananas and pepper vines, and a peripheral ground storey of maize, cassava, beans, pineapples, and others, often supplemented by an outside field of paddy rice (McConnell and Dharmapala 1978).

Such gardens provide the farmer with a mix of food and cash crops that can offer a large degree of self-sufficiency. Harvests and consequent income are dispersed throughout the year and provide greater stability than is possible with only one or two annual crops. Furthermore, the mix of trees offers almost complete protection to the soil, sustains the nutrient cycle, and eliminates the need for cultivation implicit in the production of annual food crops.

In Sri Lanka, a major programme is under way to develop such gardens on neglected and abandoned tea areas on erosion-susceptible land (The World Bank Tree Crop Diversification Project, Tea, 1). The old tea bushes are slashed heavily, and then planting holes are dug to accommodate a mixture of coconut, cloves, nutmeg, durian, rambutan, avocados, breadfruit, mango, jackfruit, coffee, and pepper. House plots are established at the same time to provide an area for vegetable production and accommodation for a milk cow. Where possible, improved varieties of the different crops are used, and, eventually, the productivity of these plantings should be high compared with that of traditional gardens where only a proportion of the crops yields at any appreciable level.

Food Trees

The humid tropics is the home for many hundreds of potentially important food trees that have traditionally provided starches, sugars, proteins, oils, salts, and vitamins to indigenous peoples. In some territories, a limited number of fruit species have been subject to selection and propagation for organized production, but the potential of other species as suppliers of basic food stuffs has been largely ignored. With increasing population pressures, however, and a greater realization of the limitations in the cultivation of crops such as maize and rice, more attention is now being paid to these crops.

Particularly noteworthy is the work in Nigeria, where Okafor (1977) has been developing vegetative propagation techniques for Irvingia gabonensis (African mango), Treculia africana (African breadfruit), Pentaclethra macrophylla (African oil bean), and Chrysophyllum albidum (star apple). Buddings of the first two species have produced viable fruits in 3.5 and 4 years, respectively. Acceptance by the local population has been good. (See the paper by Okafor in this volume, pp. 103- 107.)

A similar process is beginning in Brazil, with Brazil nuts (Bertholletia excelsa), guarana (Paullinia cupana), and the pupunha palm (Bactris gasipaes) all being brought into improvement programmes and tested in mixed cropping regimes (Watson 1980). In Latin America the pupunha or pejibaye palm may be particularly important. This palm is indigenous to the entire humid tropical zone of Central and Latin America and traditionally has provided a staple food for Amerindians. A great range of genotypes exists, producing fruit with various levels of starch, protein, and oil; palmitos (palm hearts) are obtained from young basal shoots, and the wood is of value for parquet flooring and production of canes, fishing rods, etc. Some 300 ha of small plantations exist in Costa Rica, and there are plans to expand this to 1,000 2,000 ha producing palmitos for export. Work on propagation and selection is under way in both Brazil (Arkcoll 1979) and Costa Rica (Urpi 1979), and precocious fruiting at age two years has been recorded. Some fruit yields have been as high as 35 t/ha a year, and, if they can be maintained at this level in large-scale plantings, this palm may become a major staple food for the tropics. As such, and in common with other food trees, it would have major advantages over annual food crops. It can be sown directly into uncultivated land and be kept in continuous production for many years without any soil cultivation, thus eliminating one of the major limitations facing the small farmer of the tropics.

Commodity Plantations

With commodity crops, operators in both the public and private sectors face great pressures; in many areas there is a shortage of labour for plantation operations. Costs for labour, land clearing, and crop establishment are rising, yet prices must remain in line with competitive materials. Increasing financial pressures and market fluctuations must be offset by great flexibility and sophistication in cropping practices. To improve productivity, farmers must increase inputs and accept the consequent additional technical and administrative problems. A similar situation is likely to be encountered in the development of agro-forestry schemes.


In rubber, the main drive toward increased productivity is for a shortening of the time before the tree reaches tapping size. Measures include use of advanced planting materials, leguminous groundcovers, and fertilizer applications based on a diagnosis of the local soil requirements. Also used are highyielding materials, chemicals to stimulate latex flow, and quicker replanting schedules that allow advantage to be taken of improved planting material as it becomes available (Lim et al. 1973; Abraham 1978; Mohamed Nor 1980).

Intercropping with food crops is possible in the early years of establishment of a rubber plantation but is not favoured in the large-scale operations because of management problems. In such cases, local food supplies are obtained from areas set aside for the purpose: house plots, ravine and "waste" areas, or paddy fields running through the rubber areas but unsuitable for the tree crop.

In the smallholding sector, however, intercropping has been a traditional practice, with pineapples, bananas, cucurbits, and others as cash crops, and with maize and upland rice for subsistence. Data are available from West Africa (Mellis 1978), Malaysia (Wan and Chee 1976;anon.1973) and Thailand (Templeton 1974; de Vries 1974; Speirs 1974; anon. 1974) but possibly the most comprehensive are from Indonesia by Reed and Sumana (1976). They quote the results of a survey of 100 farms in an organized smallholder development scheme (NSSDP), and a further 20 independent farmers from a local village, Babussalam. All farms in the NSSDP had 1 ha of rubber planted at a density of 500 trees/ha intercropped in the wet season but only a portion intercropped in the dry season.

Yields of wet-season rice and rice interplanted with maize were satisfactory, but there was a great deal of variation between farms. The average yield of maize in 45 NSSDP blocks was only 462 kg/ha of rubber, whereas yields of 1.52.0 tonnes should have been possible. The average yield of green gram (Phaseolus aurea) was 261 kg/ha in the NSSDP blocks and only 114 kg/ha at Babussalam. Yields of 400 - 600 kg should have been possible. Reed and Sumana calculated that, to break even, the farmers would need maize yields to rise to 942 kg/ha, rubber yields to 1,077 kg/ha of intercrop, and, for green gram, in Babussalam and NSSDP, respectively, to 449 and 342 kg/ha of rubber. These levels should be achievable if the farmers use good-quality seed and adequate quantities of fertilizer.

At the Aek Nebara smallholder development scheme in Sumatra, it has been shown that, with central assistance involving mechanical cultivation of the interrow areas, largescale intercropping with rice can be successful. In this scheme, over the years 1974-1977, the area of rubber intercropped annually with rice rose from 447 to 8,133 ha for a total production of 25,555 tonnes, with an average yield of about 1,500 kg/ha in the later years (Matondang and Rangtuki 1978). Even this scheme was not without its problems, however; at the end of rice cropping, the smallholders left the area to cultivate food crops elsewhere, and a massive lalang (Imperata cylindrica) invasion developed that threatened the rubber crop and required extensive use of herbicides.

This incident highlights a problem that is likely to occur with the intercropping of rubber and other long-term tree crops: the main crop becomes neglected in favour of the food crops. In development schemes in the Ivory Coast, for example, there is a policy of rubber monoculture with leguminous creepers used as groundcovers. However, farmers unfamiliar with rubber have insisted on intercropping with tomatoes, aubergines, maize, and even sugar cane.

The soil is sandy and infertile, and the results will be that rubber growth will be poorer than expected, the time until tapping will be extended, and the viability of the project as a whole will be affected.

In Brazil a major new planting programme is under way with rubber, but few of the farmers are familiar with plantation rubber and they now are testing systems that stem from alternative cropping traditions. At Belem, for instance, where vine pepper has been traditionally grown as a monoculture, the crop became badly affected by Fusarium, Phytophthora, and nematodes. In an effort to maintain production, trials have been carried out on the intercropping of pepper with rubber: disease incidence is said to be lower but the economic potential of the association has yet to be assessed (Viegas et se. 1980).

In jungle colonizing schemes in north-west Brazil, in an attempt to gain early income and supplement official subsidy payments, farmers are intercropping rubber with upland rice, maize, beans, and pineapples, as well as with coffee. Formal experiments are under way involving coffee grown in different spacing combinations with rubber, and these seem reasonably promising. But, in the farmers' fields, competition between the annual crop, coffee, and rubber, compounded by the absence of fertilizer application and general know-how, indicates that any short-term advantages are unlikely to offset the long-term loss in productivity of the rubber.

TABLE 1. Projected Total Income per Hectare of a Typical Crop Association with Rubber (Ouro Preto, Brazil) (in cruzeiros*). Crop association: rubber with coffee and rice-beans, maize-beans rotations in the 1st, 2nd, 3rd, and 4th year respectively after planting.


Year after planting

1 2 3 4
Sacks/ha Coffee - 15 35 25
(Cr/ha)     (30,000) (75,000) (50,000)
Sacks/ha Rice 30 - 15  
(Cr/ha)   (15,000) (7.500)    
Sacks/ha Beans 6 4,5 3 2
(Cr/ha)   (24,000) (18,000) (12,000) (8,000)
Sacks/ha Maize - 25 - 10
(Cr/ha)     (17,500) (7,000)  
Total Cr/ha   39,000 65,500 94,500 65,000

* UN rata of exchange in July 1979 was: US$1 = Cr 26: in July 1981: US$1 = Cr 87.

A common practice in the colonizing schemes is to rotate rice with maize and beans, between the coffee and rubber. The income from a typical mix of these crops, over the first four years after the rubber was planted, has recently been estimated at 264,000 cruzeiros per hectare (approximately US$7,000) (table 1). Such an income is likely to be dependent, however, on the co-operative efforts in several holdings, because there is a labour shortage in the area. Moreover, visual observation of typical rubber in the area suggests that at least eight to nine years are likely to be required before tapping can begin, so there must be serious doubts over the viability of mixed cropping under the local conditions.

Oil Palm

The great bulk of plantation oil palm is grown in monoculture, and maximum productivity is ensured by use of high-yielding materials, assisted pollination, and heavy fertilizer programmes. In West Africa the oil palm is an important constituent of the compound farm and would need to be incorporated in any mixed cropping development in the area. It has been shown that on good soils, intercropping of oil palm with food crops is quite possible and may even be advantageous. Sparnaaij (1957) has described how, in one major experiment, intercropping for as long as possible with a mixture of yams, maize, and cassava, followed by cocoyams as the palm shade developed, gave a net increase in yield of palm fruit as well as significant production of the food crops. Intercropping stimulated palm growth in the early years after planting and thereafter gave increased fruit production for 12 years of harvesting.

Soil fertility levels, however, impose some limitation on the potential for intercropping in oil palm. In one experiment at Nkwele in eastern Nigeria, on a degraded soil, continuous intercropping with yams and cocoyams was included as one treatment, with household waste and other fertilizer materials applied in the second, fifth, and seventh years after planting. In the first three years of production the continuously intercropped plots yielded twice as much as the control plots, but the effect was not lasting. Whereas the yield in the control plots continued to rise slowly, that for the intercropped treatment declined.

It would seem that the small nutrient reserves in the soil and the original vegetation, which were released by the clearing and cultivation for arable cropping, were quickly exhausted. As a result, soil fertility dropped to such a low level that the intercropped palms, although much taller and better developed than the control palms, could not maintain their yields.

The conclusion drawn from these experiments is that intercropping of oil palm is perfectly feasible on the good soils (and so is compatible with local traditions), but that on poor soils there is a clear risk of nutrient exhaustion and eventual decline in productivity of the palms.


Traditionally a subsistence crop with low input, coconut is now moving into a phase of active development. The productivity of old stands is being raised by the introduction of intercropping, and new hybrids, the progeny of crosses between Asian dwarf palms and certain West African and Asian tall varieties, show promise of bringing the profitability of coconut up to that of oil palm.

Coconuts are particularly suitable for intercropping by the small farmer. In the early years after planting, and at about 812 years when shade lightens with increase in height, the palms do not fully utilize the soil, water, and light resources available. Accordingly, intercropping has been traditionally practiced in some parts of Indonesia, Malaysia, Sri Lanka, and India, and formal studies on cropping combinations have been carried out in India (Nair et al. 1975).

Crops tested included tuber and rhizome species, upland rice, cucurbits, bananas, and pineapples. Most promising were the rhizome species, but performance of pulse crops was not good, probably because of inadequate light penetration. Among perennial crops, cocoa, cinnamon, nutmeg, cloves, and peppers grew well, with cocoa showing the greatest promise.

The beneficial association of cocoa with coconut has been recorded elsewhere and is probably due to weed control and improved nutritional conditions following introduction of the cocoa. In Malaysia the cocoa-coconut combination is considered to be the most profitable of all crop combinations (Yaacob Tunku Mansur and St. Clair-George 1979) and is being adopted rapidly by both the estate and smallholder sectors.

Intercropping of coconuts will present some of the same limitations as those with oil palm, but the coconut is more adaptable than the latter and has a much wider geographic application. For instance, coconuts are often grown on sandy coastal soils where little else that is profitable will grow. Under these conditions, cashew, passion fruit, jackfruit, and guava may provide suitable middle storey crops (Denamany et al. 1978), and cattle and sheep may be grazed successfully on the thin pasture that develops. With improving soil and moisture conditions, cropping intensity can be increased until peak productivity is reached, as with the cocoa coconut combinations of Malaysia, which are generally sited on fertile, marine clay alluvium.

Of particular interest to agro-forestry is that coconut traditionally has furnished fibre, leaf, and wood for a range of local cottage industries. In the Philippines a large-scale replanting programme in coconut has provided the stimulus for the design and production of saws capable of handling large quantities of the tough coconut wood, which hitherto was wasted material. Similarly, intensification of the Malaysian rubber industry has led to the use of rubber timber for charcoal, construction, and household and ornamental articles.

Cocoa and Coffee

Coffee production has stabilized in general, because of fear of over-production, and the only technical development worthy of mention is the production of "arabusta" hybrids and of rust-resistant varieties in Latin America. In cocoa, high yields are sought by use of selected hybrid seed, and in preliminary fashion by the development of high-yielding clonal material. Both crops are often grown together with shade crops, with yield being very dependent on incident Iight.

Of particular interest is work in Costa Rica studying the local traditional practice of growing cocoa, and particularly coffee, in association with the leguminous tree "poro" (Erythrina poeppigiana) as a shade and mulch-providing tree, together with the naturally regenerating timber laurel (Cordia al/iodora). Caturra and other low-growing varieties of coffee are normally planted at about 7,000-8,000 plants/ha, with taller-growing varieties at 3,000-4,000 plants/ha. The poro shade trees are planted at about 200 trees/ha, with laurel occurring as selfsown trees at densities of up to 300 trees/ha. This combination proves highly compatible; the poro is pruned twice a year, with slashed branches and leaves returning nitrogen of up to 80 kg/ha at each pruning (Molleapaza 1979). The laurel is a fast-growing tree and is self-pruning; it has a straight trunk with a narrow, open crown. Although it does not produce heavy shading, its roots will compete with coffee. It is a valuable timber tree, and, with the natural forest now distant from the cocoa-coffee areas, timber costs are high, so that the value of annual production by typical stands of laurel is estimated to be US$644/ha (Combe and Gewald 1979). The trees are generally cut in small numbers as the farmer requires, either for sale or for construction use.

Other tree species often found within the coffee plantings of Central America, at times growing as living fences, along steep banks or in hedges, and used for construction or firewood, are Alnus acuminate (a nitrogen-fixing tree which is also planted with pastures in the highlands), Inga spp. (a legume which produces shade and fuelwood), and Erythrina berteroana, E. costaricensis, and Gliricidia septum (all of which are used for live fence posts and fix nitrogen). Laurel (Cordia alliodora) and Spanish cedar (Cedrela odorata) are often found and tended in coffee and cocoa plantations or in pastures, and they are both very valuable timber trees. The pejibaye or peach palm (Bactris gasipaes or Guilielma gasipaes) is also included for food and shade. Work is under way at CATIE in Costa Rica on all these combinations, seeking to determine the interactions of species and so help to develop optimum combinations.

Taungya Farming

The taungya system of forest establishment is practiced in many areas of the humid tropics, particularly in the Far East, with teak plantings, and in West Africa, with Gmelina arborea Except that the system operates on a more extensive basis, it has many similarities to intercropping systems with rubber, oil palm, or coconut, with labour cultivating annual food crops in the interrow and helping to keep the plantation free of weeds.

Ball (1977) has detailed the Nigerian system and estimates that whereas direct planting with no food crops may cost the Forestry Department N525/ha (1 naira = US$1.8), with traditional taungya, when food crops are grown and belong to the farmer, costs may be reduced to N212.5/ha. If the Department retains the food crops and merely pays the farmers for their labour, costs are further reduced to N117/ha (see the more recent paper by Ball and Umeh in this volume, pp. 72 - 78).

In Costa Rica, CATIE staff have been studying taungya systems since 1962 and conclude that production of a food crop during the first one to two years of tree establishment can offset substantially the costs of. reafforestation, particularly if fertilizers are not used and if high value annual crops (maize and beans) are grown (Combe and Gewald 1979).

In applied work from Surinam, Vega (1979) also determined that the growing of food crops in the early years of establishment is of real value but distinguishes between schemes run with farmers owning the food crops and those run under the control of a state service with the farmers acting purely as paid workers. He supports Ball in concluding that the latter is the more economically satisfactory alternative for the state.

There is an interesting analogy to this from the commodity sector in Malaysia, where the Federal Land Consolidation and Rehabilitation Authority has been running land development schemes. They have found that in peripheral areas with relatively low infrastructure, an allotment of 4 ha of rubber per settler is necessary to achieve a satisfactory level of income. This amount, however, is more than can be handled properly by family labour, at least in the establishment phase, and accordingly it is recommended that the rubber be managed on a shared basis, with the family effectively receiving a wage for their labour rather than a land entitlement (Ti 1977, Mustapha Juman and Gan Teng 1980). Other peripheral areas of tree crop development, such as occur in a number of territories, could conceivably be candidates for the same treatment.

Agro-forestry and Its Development

In the whole spectrum of farming systems of the humid tropics, ranging from primitive shifting cultivation to the intensive commodity crop plantation, traditional taungya farming perhaps represents the first step towards formal cropping, and agro-forestry its updated and more intensive development. With intensification, inputs and production will both rise, and agronomic and management problems must be expected to increase in proportion. Some lessons that the commodity sector has learned will be of interest in this respect.

In the first place, management can only deal competently with a small number of main crops. The provision of highyielding material, development of nursery techniques, planting maintenance, and harvesting schedules are all simplified by concentration on a few planting materials.

Care in the establishment phase, particularly in the degree of shade permissible, the use of fertilizers, and maintenance of weed control, can exert a lasting beneficial effect on the main crop. The need for fertilizer applications sufficient to supply both production and immobilization requirements by tree crops is of major importance, particularly during the years of establishment and on poor soils (Watson 1964; Baule and Fricker 1970; Ng 1977; Sanchez 1979).

The use of leguminous groundcovers during the early years of establishment of a tree crop can have major beneficial effects on soil conditions and tree growth, while at the same time permitting significant savings in nitrogen fertilizer use (Soong and Yap 1976; Ng 1977). However, the cultivation of short-cycle leguminous food crops is unlikely to confer such benefits because they require complete fertilizer applications, including nitrogen, to produce satisfactory yields (Oelsligle et al. 1976).

To obtain satisfactory production of food crops in the interrow areas, without damage to the tree crop, the secondary crops must be kept well away from the young trees; improved varieties should be used; fertilizers applied for both tree and intecrop; and some system of pest control will also be required (anon. 1974; Templeton 1974; Wan and Chee 1976; Reed and Sumana 1976). In addition, if large areas are to be intercropped, mechanized cultivation is essential for adequate planting and harvesting schedules and control of weed problems.

As shade from the developing tree crop canopy causes a cessation of food cropping, provision must be made to prevent any major weed invasion. The best and cheapest way of protecting against weeds is to encourage vigorous tree growth by good husbandry and use of fertilizers.

In the replanting of tree crops, productive cropping can be maintained by staggered felling and introduction of a following crop under the decreased shade (Black and Hubbard 1977).

Just as some commodity crop development schemes have set land aside for food crop production, so Ball (1977) foresees the allocation of areas within any agro-forestry project so that raising of trees, animals, and food crops may separately be intensified. Wetter soils in valley bottoms might be cultivated permanently with rice and vegetables. Steep slopes and catchment areas need to be retained under natural forest that serves as a source of traditional products, including bush meat. Shallow soils on moderate slopes may support short-rotation forest crops, with slightly sloping sites with deep soils being cultivated by mechanical means with short fallow periods. Soil derived from sedimentary deposits on level sites is suitable for forest production with short agricultural cropping periods.

One can see that such developments in time could lead to the establsihment of relatively intense cropping systems for each soil unit in a given project. Egger (1978) foresees the same type of longer-term development but with the emphasis on ecofarming techniques, largely based on tree crops, and with only the minimum necessary inputs, such as fertilizers, from external sources (see the closely related paper by Behmel and Neumann in this volume, pp. 92 - 98).

In Costa Rica and Brazil, among other places, attempts are being made to compare different cropping treatments on small-scale experimental sites. The trials involve a variety of tree crops mixed with lower storey annual and perennial crops or pastures. Although these are likely to give useful indications of the effect of different cropping systems on soil conditions, their application may otherwise be very limited. It is extremely difficult to reproduce field conditions in small plots; edge effects are large, and micro-climatic effects caused by wind and insolation differ appreciably. It is virtually impossible to extrapolate the cost of weeding, pruning, harvesting, and other operations in small plots to field conditions; pasture plots have no animal interaction and so on.

These experiments need to be complemented by production-oriented, large-scale field studies of the more promising systems, on a number of locations, if they are to be meaningful to the farmer and the planner. Such studies do not require that one relearn all the basic facts of tree- annual crop-soil interrelationships. Much is already known, or can be inferred, from related commodity crop studies. Field work should concentrate instead on the selection of the most promising associations for any given location and then on the development of methods for maximizing productivity within the resources available.

Contents - Previous - Next