J.E.M. Arnold is a Senior Research Fellow at the Oxford Forestry Institute, Oxford, United Kingdom.
Factors influencing farmer decisions for or against tree growing, and the impact of tree c ash crops on household food security
In most parts of the developing world, rural people recognize the role of trees in providing a number of locally important goods and services. As nearby natural forests recede or are degraded, farmers have historically tried to protect, plant and manage trees on their land in order to maintain such sought-after outputs. For example, in a recent study of hill areas in Nepal where there has been progressive reduction in the forest resource, Campbell et al. (1987) found that the households surveyed had increased the number of trees on their own land, particularly fodder and fruit-trees.
HOME GARDENS IN INDONESIA - intensive intercropping of trees and annual plants
However, increasing pressures on the land and on tree resources may result in the elimination of trees rather than in their retention or establishment. Pressures for more land for crop and livestock production widely militate against the continued presence of trees. Shifts toward monocrops and mechanization tend to lead to the removal of trees since they impede tractor operation. Trees in the vicinity of crops are removed to eliminate competition for light, water and nutrients. Land is left fallow for increasingly shorter periods, so that a smaller proportion of the area is under tree cover at any given time and the trees do not have time to develop to productive size. Burning associated with pasture management also impedes the growing of trees, as do serial tenure patterns where animals are allowed to roam freely on farmers' lands after the harvest.
Changes in the control of land as pressures on it mount also impinge upon the place of trees in land use. Privatization and nationalization of common lands both heighten pressures on remaining communal resources and increase the number of people excluded from access to trees. Insecurity of land tenure and tree tenure practices that imply that the presence of trees establishes an effective claim on land tend to discourage or prevent tree management.
Nevertheless, the expanding presence of trees in many land use situations indicates that in certain circumstances growing trees can also form an appropriate part of the response to increasing pressures on the farmer's resource base. The purpose of this article is to examine a number of such situations in an attempt to identify the contributions that trees make to the farm household and the considerations that have caused the farmer to adopt them, To this end, case-studies representing three types of situation are examined-farming systems based on low-intensity, extensive use of land; the intensive and highly developed system of home gardens; and the growing of trees as cash crops in the form of farm woodlots.
In situations where it has traditionally been practiced, shifting cultivation, the most basic form of agriculture, is the most efficient use of farmer resources, predominantly family labour. Where there is sufficient land to support a fallow period, no other farming practice will produce a higher return to labour without inputs of capital. The fallow vegetation maintains soil productivity, and the process of clearing and burning permits crop cultivation with minimal inputs for soil preparation and weeding. Though cultivation periods could be extended by increased weeding, it is easier, in terms of requiring less labour, to clear and bum a new area. Similarly, yields per hectare could be increased by more intensive cultivation, but at the expense of lower output per unit of labour. As long as farmers can satisfy their production objectives through less labour-intensive methods, they will logically stick to them (Rambo, 1984; Raintree and Warner, 1986).
A broad sequential process can be identified where, as reduced access to land prevents continuation of sustainable shifting cultivation, farmers eventually start to intensify agricultural practices (Olofson, 1983; Raintree and Warner, 1986). This usually takes the form of small incremental changes involving increased labour, and sometimes capital in the form of fertilizers, herbicides, etc. This evolutionary process away from shifting cultivation may move away from tree cultivation altogether, but it can include tree components.
A common practice at early stages in this process is to enrich the fallow by encouraging or planting tree species that either accelerate or enhance the regeneration of soil fertility or produce outputs of subsistence or commercial value, or both. The cultivation in the Sudan and elsewhere in the semi-arid areas of Africa of Acacia senegal as a fallow crop is an example of a species that does both. A leguminous species, A. senegal improves soil fertility, as well as producing gum arabic for sale, and fuelwood, medicine, fibres and other products for use in the households. Other examples include the management of the babassu palm for both commercial and subsistence products in conjunction with shifting cultivation over large areas of the mid-north of Brazil (May et al., 1985); the planting of rattan as a commercial crop in the swidden cycle in Borneo (Weinstock 1983); the multiple-product-managed swiddens of the Ifugao in the Philippines (Conklin, in Olofson, 1983); and then market-oriented cyclic agroforestry systems practised in parts of the Peruvian Amazon (Padoch et al., 1985). Although quantitative information on inputs and outputs is limited for these systems, it can be expected that, as they generate additional income with only minor increases in labour inputs and minimal changes to the basic shifting cultivation system, they will provide fairly high returns to labour (Raintree and Warner, 1 986).
At the next step in intensification, as pressures on land force the transition toward continuous cultivation, various forms of intercropping are encountered. By incorporating soil-enriching species with food crops, these practices introduce the functions of fallow on a continuous basis. Numerous examples of such continuous fallow strategies are to be found, such as the maintenance of Faidherbia albida (formerly termed Acacia albida) in cultivated areas of much of Africa.
An example of plant mixtures of this nature is the intercropping of Sesbania sesban, a leguminous tree, with maize in areas of western Kenya. When the maize is shaded out after about three years the Sesbania sp. is left as a fallow crop for one to two years, and then cleared and used for fuelwood. The cycle is then repeated. Practised in an area of labour shortages, over a ten-year cycle it has been estimated that it produces less than half the maize per hectare when compared with monocropping but requires less than half as much labour and gives higher maize yields per unit of labour input-in addition to the fuelwood and soil protection benefits (World Bank, 1986).
Considerable research has been directed recently toward developing a more intensively managed continuous fallow system known as alley cropping. This involves the growing of field crops between hedgerows of nutrient-cycling trees or shrubs, which are periodically pruned during the cropping season to reduce shading and provide green mulch for food crops. An economic analysis of the results of alley cropping research conducted at the International Institute of Tropical Agriculture (IITA) in Nigeria, involving intercropping of maize with Leucaena leucocephala and treatments with nitrogen fertilizer and herbicide, revealed that Leucaena intercropping gave the highest economic returns and yields of maize of all the alternatives, but required higher labour inputs than cultivation of maize alone or with fertilizer/herbicide applications. Returns to labour from intercropping, however, were higher than in maize monocropping (Ngambeki, 1985). However, considerable further work in testing the research results under on-farm conditions is needed to ensure that the labour availability and farmer skills assumed are realizable in practice; that the costs and values are consistent with those actually experienced or perceived by farmers; and that the physical input-output relationships and performance are replicable on farms (Balasubramanian, 1983; Raintree, 1989).
IMPROVING LAND-USE PRACTICES intercropping of maize, beans and trees In Kenya
Home gardens (also known as compound farms and as homestead and mixed gardens) are usually located close to the household and form part of the more intensively cultivated areas in the farm system. They are characterized by a mixture of annual or perennial species grown in association. They commonly exhibit a layered vertical structure of trees, shrubs and ground-cover plants that recreate some of the properties of nutrient recycling, soil protection and effective use of space above and below the soil surface found in forests. Home gardens are widely used to supplement outputs from other parts of the farm. They are also used to spread farm work, output and income more evenly throughout the year (Ninez, 1984).
In the densely populated lowland areas of central Java, for example, home gardens form the principal farming system on dry lands while irrigated rice cultivation forms the other main component of the farm system. The gardens are traditionally dominated by perennials rather than annuals, and by woody rather than herbaceous plants.
On farms with sufficient rice land to enable households to meet their basic food needs, priorities in labour and capital allocation are directed to rice production. The garden areas in these farms are essentially forest gardens of trees with commercial value (Wiersum, 1982; Hunink and Stoffers, 1984; Stoler, 1978). However, as a result of rapid increases in population, the area of rice land per caput has decreased and a large proportion of farmers now have no rice land, or at least not enough to meet their basic rice requirements. As the proportion of land devoted to rice production decreases, home garden areas are cultivated more intensively, becoming mixed rather than forest gardens as annual plants are progressively intercropped to provide food and income.
Management is intensified by increasing labour inputs. The scope for increasing productivity is such that labour inputs in small gardens are reported to be on average three times as high as those for larger gardens. Returns to both land and labour are high (Soemarwoto and Soemarwoto, 1984). Intensive garden management has been reported to produce up to 20 percent of household income and 40 percent of caloric intake (Stoler, 1978).
Another approach to intensification of garden area use has been to increase the value added from home garden produce. Penny and Singarimbun (1973) describe how some of the poorest farmers have shifted from the production of coconuts to the production of coconut sugar, which is a highly labour-intensive process but leads to increased returns to coconut-bearing land despite low returns to labour. Other farm-related income-earning activities include pit-sawing and fuelwood gathering (Hunink and Stoffers, 1984).
As landholding size continues to decline, the farmers increasingly have to try to gain income from off-farm employment. At this stage, cultivation of annuals is reduced in order to allow more time for income-earning activities. Trees and other perennials requiring only low labour inputs become the main component of the garden (Stoler, 1978).
In a study of farming practices in southeastern Nigeria, Lagemann (1977) found a similar relationship between population pressure on the land and intensity of tree cultivation. Farms comprise a mixture of fallow, outer and inner fields, and permanently cultivated compounds around the household. The latter contain a variety of tree species, including oil-palm, raphia palms, coconut, banana and plantains intercropped with cassava, yams and other arable crops.
Tree crops in Kerala, India: a low-capital option
A study of tree cropping in the homesteads in Kerala, India, examined a system that combines intensive management of perennial and annual crops with more extensively managed areas. As rising population pressure on the land leads to decreasing landholding size, uncultivated land is first brought into use, resulting in removal of natural tree cover. This is followed by more active management of the home gardens, with the range of cultivated trees being narrowed down to those with multiple uses. Priority is given to those species valued for fruit, fodder mulch and suitable as supporting structures for cultivation of pepper, betel and various climbers. In the process, the density of trees and intensity of their cultivation increases (Nair and Krishnankutty, 1984). As pressure on the land increases further, landholding size falls to the point where farming ceases to provide the main source of income. As labour on the farm declines and moves to off-farm employment, the tree component increases even further as the vegetation reverts toward a forest condition.
However, unlike the cases in Java and southeastern Nigeria, in this area of Kerala capital has become increasingly available, enabling some farmers to intensify land use further through purchased inputs such as fertilizer and herbicides. This reduces the importance of the multipurpose trees for soil nutrient maintenance and weed suppression. Consequently, they tend to have been removed as they have become impediments rather than complements to agriculture. Removal of trees has been accelerated by rapid rises in the prices of timber and land; the latter leading to the shift of land use for cash crops. Trees are then cultivated only where they are competitive as a cash crop. For example Ailanthus triphysa is grown to supply wood stock to the match industry.
Thus, the Kerala experience broadly parallels that of Java and Nigeria up to the point where farmers were able to invest substantial capital in their farms. The subsequent displacement of agroforestry practices seems to confirm that, in the absence of capital, farmers employ trees primarily to provide substitutes for purchased inputs and as crops requiring lower inputs than other crops.
In Nigeria, as in Java, growing population pressure is being accompanied by decreasing farm size and declining soil fertility. As pressures on the land heighten, the proportion of land under compound systems is increased, as is the density of both tree and arable crop cultivation within the compound areas. Lagemann argues that this shift reflects farmer's perceptions that such land use, combined with increased mulching and manuring, offers the most effective way of using their resources to slow down the process of declining soil fertility and to maintain production. Though labour inputs per hectare are no higher than in the fields, yields in monetary terms are five to ten times as much per hectare and returns to labour four to eight times as much. Lagemann identifies the phasing of planting and harvesting in the compound areas to reduce peak workloads, and better physical working conditions under the shade of trees as factors contributing to this higher labour productivity.
With increasing population density, compound areas account for up to 59 percent of crop output and a growing proportion of total farm income, with the proportion of income generated from tree crops rising to a share nearly equal to that from arable crops. Livestock becomes an increasingly important part of the farm system, both as a source of income and of manure. However, as population density continues to increase, yields and returns to labour eventually decline to a point at which farmers have to turn increasingly to non-farm sources of income. As people shift to working off-farm, a lower labour input to the compound areas is adopted, leaving the areas dominated by trees and other perennials.
The overall picture, therefore, as in Java, is one of farmers responding to decreasing land availability by moving to greater dependence on systems that incorporate trees with crops, and livestock. Initially this is because they permit more sustainable intensification of land use and higher returns from available labour inputs, than alternative land uses. When pressures on the land increase further, to the point where individual holdings are too small to be economically viable, income has to be generated mainly from off-farm employment. This results in a further shift toward tree crops in the overall mix, as these systems have the added advantage that they can be maintained, in a modified form, as low-input, low-management forms of land use.
Similar patterns have been recorded for other parts of the tropics in which home gardens are an important part of the overall farming system. As the experience from Kerala underlines (see Box on p. 38), it is a system that exhibits considerable diversity and dynamic change in response to changing market opportunities and availability of capital, labour and other inputs.
TREE CASH CROPS Eucalyptus deglupta for fuel and polewood in Costa Rica
The farm woodlots examined in this section essentially involve the growing of trees as field cash crops. Farmer decisions could therefore be expected to be governed by the question of whether or not the tree crop is more profitable than alternative crops or other uses of the land. Though there are many instances where trees are grown intensively, with inputs of fertilizer, irrigation, etc., the majority of situations reviewed suggest once more that tree crops are primarily selected when a low-input, minimum-management system is required by the farmer.
For example, in the Philippines, farmers now grow Paraserianthes falcataria on a six- to eight-year rotation for sale to a nearby pulp company, Paper Industries Corporation of the Philippines (PICOP), on an area of land previously utilized for low-density extensive agriculture. The average size of landholding is 11 hectares, some of which is devoted to cultivation of food crops. At least part of the land on 45 percent of the farms had been previously used for growing food crops, and other non-food crops had been raised on some parts of 31 percent of the farms. Credit for tree-farming was available but was utilized by only 30 percent of the farmers eligible; generally these were farmers with above average woodlot sizes (Hymen, 1983).
Ex post analysis showed that farmers were able to earn (in their estimation) an acceptable return from their woodlots in most prevailing circumstances, with an internal rate of return of 22-31 percent for successful tree-farmers on costs other than land costs. There are no data to permit the comparison of these returns with returns from alternative land uses, but farmers cite low labour inputs as the principal reason for preferring tree growing (Hymen, 1983). In an area where average landholding considerably exceeds the size that can be cultivated under food or other crops with family labour, pulpwood production enables farmers to expand the area they can put to productive use.
Another instance where growing trees as a cash crop has recently been successfully introduced to farmers is in Haiti. Tree growing was introduced to hill farmers as a means of earning additional income. In this instance, there was already an existing market for charcoal and poles, a strong orientation toward income generation and cultivation of cash crops. In addition, most farmers owned their land. It was hoped that incorporating trees into their land use systems would also help them control the serious erosion problems they were experiencing.
By 1986 approximately 110000 farmers had planted more than 25 million seedlings. Patterns of planting vary considerably from farmer to farmer, but increasingly they have moved from species suited only to the output of fuelwood and poles to multipurpose species, and to intercropping trees with agricultural crops such as maize, sorghum and beans. A recent cost-benefit analysis indicates that most such tree/crop mixtures should have a higher net present value than continued cropping without trees, and that adoption of tree/crop mixing appears to be an economically viable decision for most people (Grosenick, 1986).
Surveys of participating farmers confirm that they perceive increased income as the principal benefit, but are influenced by other motives as well. Many plan to use their trees as a form of savings and value the fact that they can draw down such savings, by harvesting the trees, at a time of their choosing. As was the case in the Philippines, growing trees may enable poor farmers to increase the amount of land they are able to work (Conway, 1986; Balzano, 1986 in Conway, 1987).
In India, there has been an upsurge in tree growing in response to expanding markets for poles and other wood products (e.g. pulpwood) and energetic farm forestry support programmes. The principal factors motivating farmers to move to cultivation of trees rather than other cash crops appear to be rising labour costs and increasing shortages of labour, and declining returns from agricultural crops grown on rainfed areas. The advantages of trees in these circumstances are usually perceived to be: low labour inputs, minimal annual operating costs in most years, greater resistance to drought, and hence reduced risk and uncertainty.
Planting by both small and larger farmers is narrowly focused on monocrops of a few predominantly wood-producing species, mainly eucalyptus. Even among smaller farmers there has been very little development of agroforestry tree cultivation practices.
In each of the situations reviewed above, the decision to grow trees as a cash crop therefore appears to have been a response to one or more of at least three factors. First, income generation, as distinct from food production, played a prominent role in the farmers' production objectives. Second, tree cultivation was perceived to offer advantages in situations of scarce labour and usually scarce capital. Finally! tree crops provided a means of risk reduction.
Tree cash crops in Kenya: low expenses + good markets = rapid growth
Available information about the rapid recent expansion of cash crop growing of trees in parts of Kenya has been assembled by Dewees in a study for the World Bank (World Bank, 1986). Two popularly cultivated species are eucalyptus, grown for poles, and black wattle, grown for poles, charcoal, fuelwood and sticks for "mud-and-wattle" house construction. Markets for these wood products-and for pulpwood and sawn timber in some places-are growing strongly, with farm level production accounting for a large part of the supply.
Tree growing tends to be practiced by poor farmers unable to meet their basic food needs and for whom it is a principal source of farm income. In Vihiga, in Kakamega District, for example, the average farm size is about 0.6 ha, of which approximately 25 percent is under eucalypt woodlots (van Gelder and Kerkhof, 1984).
Gross income per hectare in this area is considerably lower from tree growing than from other agricultural crops. Dewees suggests that farmer preference for tree crops in these circumstances is conditioned by availability of capital and labour, and by attitudes to risk management. Alternative crops often require substantial investments, at levels beyond small farmers' access to capital. Trees, by contrast, require very little expenditure. Tree growing is also attractive to farmers because of the low labour inputs it requires. There is a shortage of labour in this area because of widespread outmigration of the farm households' men, who seek off-farm employment. Where markets for tree products are good, returns to labour from pole production have been estimated to be some 50 percent greater than from maize production (World Bank, 1986). Tree growing is consequently a rational use of resources for poor farmers needing to devote a substantial part of their labour to non-farm employment.
It would be unwise to generalize too widely from such a limited number of studies, and from analyses that focus simply on the economic factors from within the broader set of decision criteria that most farmers are likely to employ. Nevertheless, a number of general points emerge.
It has often been argued that the cultivation of trees is only feasible for wealthy farmers. This assumption is based on the premise that poor farmers have the production of staple food as their main objective. However, the evidence deployed above suggests that in many cases poor farmers' resources are too limited for them to be able to meet their basic needs from their own production, so that income generation becomes increasingly important.
In these circumstances trees may prove to be appropriate cash crops or intercrops with cash crops. Where land rather than labour is the limiting factor, joint tree/crop/livestock systems may give better returns than monocrops. Where availability of labour has become the constraint because of the need to find work away from the farm, low-input tree crops may provide the best way of keeping land in productive use. Although overall returns from the latter would be greater under agricultural crops than trees, it makes sense for poor farmers to raise trees because they cannot afford the costs of producing alternative crops, or the costs of inputs such as fertilizer, herbicide or irrigation that would be required to maintain the productivity of their land in the absence of trees.
In situations where farmers are being increasingly forced to concentrate on income generation, and hence on cash crops, some tree species and forms of incorporation of trees into the farming system can help mitigate the dangers to which the farm household is likely to be subjected as it has to shift from food self-sufficiency to purchased foods. Virtually all trees will produce fuelwood as a by-product for household use. Species that provide fruit and fodder, shade, protection, green manure and soil amelioration are more likely to provide for both subsistence and income needs than single-product species. Multispecies systems such as home gardens help even out seasonal flows of products and income and the demands on labour, as well as contributing to reduction in risk.
It has been argued that some forms of tree growing, for example farm woodlots, can have negative impacts on household food security in that they divert land from food crop production, reduce employment, are vulnerable (as single-product crops) to sharp market fluctuations, and provide income in "lumps" leading to periods with little income. On the first two of these points, the available evidence suggests that the transfer of land to tree crops is usually a response to changing growing conditions that have made food production no longer viable, or to increasing labour scarcity or rising labour costs. They are therefore more likely to be a consequence than an agent of these changes.
However, the issues over income underline the fact that such tree monocropping is an appropriate option only if the household has access to other sources of food and/or income, and if there are stable markets for the tree products. There is a danger that some programmes to encourage "farm forestry" could induce farmers who are not equipped to do so to shift to tree monocrops. Cash incentives and concentration on species familiar to foresters but not necessarily suited to farmers' needs or expectations could bring about negative results. These dangers could be exacerbated by pressures to achieve ambitious targets, and the lack of attention to market prospects, that many of the larger farm forestry programmes exhibit.
The role of tree components in farming systems
Tree-growing practices contribute to a wide variety of existing farming systems. They tend to be prominent parts of systems where capital and physical resources are limited. In such conditions a tree component within the farming system can play one or more of the five following overlapping roles:
1) maintain productivity of land in situations of scarce capital through the substitution of trees for purchased inputs of fertilizer, herbicide, and investments in soil, crop protection and/or in irrigated water;
2) make productive use of land in situations of scarce capital and labour where trees, as low-input, low-management crops, constitute the most effective use of these resources;
3) maximize returns to land in situations where land and capital are limited, and tree/crop/livestock combinations permit fuller use of available labour than alternative uses of the land;
4) increase income-earning opportunities from use of farm resources as size of landholding and/or site productivity fall below the level at which the household's basic food needs can be met from on-farm production of food;
5) strengthen risk management through diversification of outputs, wider seasonal spread of inputs and outputs, and reduction in the likelihood of crop failure through drought, and build-up of tree stocks to provide capital.
Balasubramanian, V. 1983. Alley cropping: can it be an alternative to chemical fertilizers in Ghana? Paper prepared for the Third National Maize Workshop, 1-3 February 1983, Kumasi, Ghana.
Campbell, J.G. et al. 1987. Socio-economic factors in traditional forest use and management: preliminary results from a study of community forest management in Nepal. Banko Janakari, 1(4): 45-54.
Conway, F.J. 1986. The decision-making framework for tree planting the Haiti Agroforestry Outreach Project. Port-au-Prince, Univ. Maine Agroforestry Outreach Research Project.
Conway, F.J. 1987. Case study: the Agroforestry Outreach Project in Haiti. Conference on Sustainable Development, April 1987. London, Int. Inst. for Environment Development.
Grosenick, G. 1986. Economic evaluation of the Agroforestry Outreach Project. Port-au-Prince, Univ. Maine Agroforestry Outreach Research Project.
Hunink, R.B.M. & Stoffers, J.W. 1984. Mixed and forest gardens on Central Java: an analysis of socio-economic factors influencing the choice between different types of land use. Dept Developing Countries, Univ. Utrecht, Utrecht.
Hyman, E.L. 1983. Pulpwood tree farming in the Philippines from the viewpoint of the smallholder: an ex post evaluation of the PICOP Project. Agric. Admin., 14: 23-49.
Lagemann, J. 1977. Traditional African farming systems in eastern Nigeria: an analysis of reaction to increasing population pressure. Weltforum Verlag, Munich, Africa Studien.
May, P.H. et al. 1985. Babassu palm in the agroforestry systems in Brazil's Mid-north region. Agroforestry Systems, 3(39): 275295.
Nair, C.T.S. & Krishnankutty, C.N. 1984. Socio-economic factors influencing farm forestry: a case study of tree cropping in the homesteads in Kerala, India. In Community forestry: socio-economic aspects. Bangkok, FAO/East-West Center.
Ngambeki, D.S. 1985. Economic evaluation of alley cropping Leucaena with maize-maize and maize-cowpea in southern Nigeria. Agric. Systems, 17: 243-258.
Ninez, V.K. 1984. Household gardens: theoretical considerations on an old survival strategy. Potatoes in Food Systems Research Series, Report No. 1. Lima, Peru, International Potato Centre.
Olofson, H. 1983. Indigenous agroforestry systems. Philipp. Q. of Culture and Society, 11: 149-174.
Padoch, C. et al. 1985. Amazonian agroforestry: a market-oriented system in Peru. Agroforestry Systems, 3: 47-58.
Penny, D.H. & Singarimbun, M. 1973. Population and poverty in rural Java: some economic arithmetic from Sriharjo. Cornell International Agricultural Development Monograph 41, Dept Agric. Economics, Cornell University, Ithaca, New York.
Raintree, J.B. & Warner, K. 1986. Agroforestry pathways for the intensification of shifting cultivation. Agroforestry Systems, 4(1): 39-54.
Raintree, J.B. 1989. Social, economic and political aspects of agroforestry. Paper presented at the Int. Conf. on Agroforestry: Principle and Practice, Edinburgh, 23-28 July 1989. Edinburgh, UK, Univ. Edinburgh, IUFRO, IFSC.
Rambo, A.T. 1984. Why shifting cultivators keep shifting: understanding farmer decision-making in traditional agroforestry systems. In Community forestry: some aspects. UNDP/East-West Center/FAO. Bangkok.
Soemarwoto, O. & Soemarwoto, I. 1984. The Javanese rural ecosystem. In A.T. Rambo & P.E. Sajise, eds. An introduction to human ecology research on agricultural systems in Southeast Asia. Los Baños, Univ. Philippines.
Stoler, A. 1978. Garden use and household economy in rural Java. J. Indonesian Studies, 14(2): 85-101.
van Gelder, B. & Kerkhof, P. 1984. The agroforestry survey in Kakamega District: final report. Working Paper No. 6. Kenya Woodfuel Development Programme, Beijer Institute, Nairobi.
Weinstock, J.A. 1983. Rattan: ecological balance in a Borneo rainforest swidden. Econ. Bot., 37(1): 58-68.
Wiersum, K.F. 1982. Tree gardening and taungya on Java: examples of agroforestry techniques in the humid tropics. Agroforestry Systems, 1: 53-70.
World Bank, 1986. Economics issues and farm forestry. Working Paper prepared for the Kenya Forestry Sector Study. Washington, D.C., World Bank. (mimeo)