Cover Image
close this bookAgroforestry in the Pacific Islands: Systems for Sustainability (UNU, 1993, 297 pages)
close this folder9 Institutional agroforestry in the Pacific Islands
View the document(introduction...)
View the documentIntercropping of tree crops/woody perennials with commercial or subsistence ground or tree crops
View the documentPlanting of timber, fuel wood, and general-purpose trees in relation to agroforestry and agriculture
View the documentGrazing with commercial tree cropping and silviculture
View the documentThe future of institutional agroforestry in the Pacific


Intercropping of tree crops/woody perennials with commercial or subsistence ground or tree crops
Planting of timber, fuel wood, and general-purpose trees in relation to agroforestry and agriculture
Grazing with commercial tree cropping and silviculture
The future of institutional agroforestry in the Pacific

The agroforestry systems so far described in this volume have been informal, often traditional systems that were developed on the basis of empirical, non-quantitative experimentation by local practitioners. We turn now to institutional agroforestry in the Pacific Islands, by which we mean those agroforestry activities that are promoted by governments, quasi-government organizations, private agencies, companies, and aid donors and that involve external funding, formal training, agronomic research, and extension services. Because the activities that fall under the rubric "institutional agroforestry" are becoming so many and varied, we provide here only a summary review rather than attempt a comprehensive survey. Other reviews of the topic are the paper by Vergara and Nair (1985) and the reports given by country representatives at a meeting on agroforestry held in 1987 at the University of the South Pacific's School of Agriculture in Western Samoa (Clements 1988).

Inasmuch as plantation agriculture is the core of much institutional agroforestry in the modern Pacific, it can be said that institutionalized activities were first promoted by colonial governments when they encouraged the plantation production of products for export in order to finance government expenditure and to provide cash income to local populations. Agroforestry projects are encouraged today for similar reasons as well as to meet local demand for forest products, to bring grass-fern "wastelands" to greater productive use, and to reduce the adverse environmental effects of rapid deforestation. A major emphasis of recent agroforestry programmes has been the promotion of increased productivity through multiple use, although multiple use is often defined solely with regard to the commodities derived from some mix of commercial crops, livestock, or timber, rather than in terms of the traditional context wherein an agroforested landscape showed increased multi-purpose utility and stability.

Table 10 Categories and examples of institutional agroforestry in the Pacific Islands

Intercropping of tree crops/woody perennials with commercial or subsistence ground or tree crops - Coconuts intercropped with ground and tree crops

  • Cocoa intercropped with ground crops under shade trees and/or coconuts
  • Coffee with shade/nurse trees or shelter-belts
  • Citrus with shelter-belts and multi-purpose species
  • Bananas intercropped with ground crops under coconuts or other trees
  • Vanilla with host plants under coconuts
  • Oil-palm with subsistence and commercial crops
  • Kava with ground crops and/or other trees
  • Other commercial agroforestry intercrops

Planting timber, fuel wood, and general-purpose trees in relation to agroforestry and agriculture

  • Pine plantations in relation to agroforestry
  • Non-pine forestry in relation to agroforestry

Grazing with commercial tree cropping and silviculture (silvipastoral or agrosilvipastoral systems)

  • Livestock under coconuts
  • Cattle under timber species
  • Other silvipastoral activities

Most modern agroforestry projects fall into one of three categories:

  1. intercropping of tree crops or woody perennials with commercial or subsistence ground or tree crops;
  2. planting timber and/or fuel-wood species (mostly exotic) within existing agricultural systems, as intercrops, rotational crops, or as small-scale monocultures, or woodlots; and
  3. 3. grazing under commercial tree cropping.

Table 10 provides examples of the kinds of activities that are practised under each category.

Intercropping of tree crops/woody perennials with commercial or subsistence ground or tree crops

In the first category, the focus has been on monocultural production, with limited intercropping, of major export crops such as coconuts (for copra and oil), cocoa, bananas, coffee, rubber, and oil from oil palms, and, to a lesser extent, Citrus fruit, papaya, mango, kava (Piper methysticum), and vanilla (which requires support, or mother, trees). Although most of the species in this category have been promoted on the basis of large-scale, often foreign-owned, plantations or estates, they have also been actively promoted for smallholder production by colonial and independent governments and aid agencies.

Since the early nineteenth century and until recently, most of the islands in the Pacific basin have been characterized as plantation economies dependent on the production of coconuts, cocoa, and to a lesser extent coffee or bananas. On the atolls and smaller isolated islands, the coconut, which was an integral component of subsistence agroforestry systems, became the main source of terrestrial-based cash income. For some of the atolls of Micronesia, the introduction of the "tin can copra economy" led to abandonment of Cyrtosperma chamissonis taro pits and the erosion of self-sufficiency (Farrell 1972). The cultivation of subsistence crops was replaced by the cultivation of coconuts for copra and coconut oil to earn a cash income. On the larger islands, the establishment of coconut plantations was soon followed by the development of intercropping systems that included cattle and later cocoa under the coconut palms.

Because of the economic significance of plantation crops, especially coconuts, to many of the islands, systems of raising cocoa, cattle, or coffee under coconuts and other trees have been the subject of concern by island departments of agriculture and regional and global development organizations. Most of the plantation crops have been characterized by fluctuating demand and prices, determined largely by metropolitan countries and larger producers, an uncertainty that has motivated many countries in the Pacific to investigate alternative export crops such as ginger and vanilla.

Coconut intercropping

Coconut agroforestry is the most widespread and important form of commercial and government-promoted agroforestry in the Pacific Islands. Although most smallholders have intercropped coconuts with ground crops and other tree crops or protected trees as part of the traditional polycultural agroforestry systems, many governments in the region continue to encourage the intercropping of ground crops and some tree crops with coconuts. In many cases, this effort has taken the form of the systematic planting or replanting in existing garden areas of regularly-spaced palms, which can then be under cropped. Alternatively, undercropping in existing commercial groves of coconut palms has been systematically encouraged.

Very early on in Papua New Guinea, "it was advocated that plantations could be virtually self-supporting with regard to the provision of food for labour" by intercropping subsistence foods between coconuts; "if grown in correct rotation, the fertility of the soil will not only be maintained, but improved" (Gallasch 1976). However, apart from the widespread intercropping of coconuts with cocoa and pasture, and to a lesser extent, with Robusta coffee, interplanting with food crops was actively discouraged by the Department of Agriculture, Stock and Fisheries. In the early 1970s, however, experiments conducted on the intercropping of coconuts with selected forage grasses and legumes, taro, sweet potatoes, corn, and peanuts indicated that a variety of food crops could be grown successfully between coconuts, and that the effect of intercropping tended to be beneficial in terms of reduced weeding and maintenance of all crops, and that such intensive use of scarce land yielded substantially greater returns than coconut monoculture (Gallasch 1976).

Similar experiments have been conducted in Vanuatu, Fiji, Tonga, and Western Samoa. For instance, in Tonga, a succession of coconut replanting schemes over the past two decades has encouraged the replanting of coconuts on smallholder agricultural allotments, and has actively encouraged intercropping with ground crops, a practice that would probably continue regardless of official support. Maude (1965) found that on some of the smaller islands of Ha'apai, where land is scarce, there were few allotments without full coconut coverage, under which ground-cropping was practiced. Thaman (1976) reported that the median coconut coverage on the main island of Tongatapu was 50 per cent, with over 50 per cent of all allotments having recently planted coconuts, most of which were undercropped at least partially with staple ground crops and supplementary crops. A wide range of cultivated and protected tree species also grow throughout the coconut groves. The normal spacing was 30 feet between rows and 20-30 feet within rows. Because of the increase in plough cultivation and the interest in intercropping, there have been moves to increase the between-row distance and to decrease the within-row distance to as little as 10 feet in order to maintain copra yields, while at the same time making cultivation easier and maximizing both subsistence and commercial intercropping between rows.

As Watt (1980, 303) argues, for most island groups and many small islands within groups, institutionalized coconut intercropping and re habilitation is of utmost importance since coconuts constitute the only sizeable wood resource: "only in the case of Papua New Guinea is the size of the coconut resource dwarfed by the commercial timber resource." The development of a successful coconut-milling industry for overmature stems in Tonga and other areas has not only made use of a resource that had been wasted or destroyed, but also has facilitated the improvement of coconut production by removing overmature trees so that effective replanting could take place. This process is facilitated in Tonga by coupling the well developed and ongoing coconut replanting scheme with a chain-saw service to remove overmature palms for milling (MAFF 1985, 41).

Cocoa intercropping

There are numerous examples of the intercropping of cocoa, or cacao, with food crops or with other tree species, many of which serve as shade plants. In the Tolai areas of the Gazelle Peninsula of New Britain in Papua New Guinea, large-scale planting of cocoa, which began in the 1950s, has either displaced food gardens or has resulted in significant changes to the traditional food systems, including the development of cocoa-intercropping systems. The system that operated in the mid-1970s consisted of clearing forest fallow and then planting root crops and bananas. After the root crops were harvested, cocoa was planted under the shade of the bananas, together with Leucaena leucocephala and/or coconuts, so that the food garden became a regularlyspaced cocoa plantation under Leucaena or a cocoa-coconut-Leucaena mixture (Bourke 1976).

In other places in the Pacific, several other intercropping combinations with cocoa have been developed. For example, in the Solomon Islands, cocoa has been intercropped with chillies, ginger, and tumeric; and Leucaena leucocephala and Gliricidia septum have been used as shade trees (Walton 1988, personal communication). In Vanuatu, the joint-venture Metenesel Estates Cocoa Project, covering 1,700 ha, has officially encouraged intercropping and the preservation of buffer zones of alternate 10-ha blocks of uncleared indigenous forest between cocoa planting, with Gliricidia septum as temporary shade. Most of the cocoa in Vanuatu is cultivated under the shade of coconuts or thinned natural bush or forest, although some areas have been formally planted under Gliricidia septum, Leucaena leucocephala, and Erythrina variegate (Jacovelli and Neil 1984, 6). In Fiji, cocoa is commonly intercropped with tannia (Xanthosoma sagitti folium)under Erythrina variegate as a shade tree. The Ministry of Primary Industries has recommended the planting of cocoa under coconuts, taro, bananas, cassava, and kava (Piper methysticum) (MAF 1974).

Coffee intercropping and agroforestry

Like cocoa, commercial coffee plantations are planted under shade trees and/or with wind-breaks. The major species used in Papua New Guinea for such purposes are Leucaena leucocephala, Casuarina spp., Albizia stipulate, and Erythrina variegate. Coffee has also been interplanted under coconuts in coastal Papua New Guinea, Vanuatu, and New Caledonia (Brookfield with Hart 1971, 157). In the Sepik district of lowland Papua New Guinea, Allen (1985, 227) reports that Robusta coffee, grown in permanent blocks, has been integrated into the traditional diversified shifting agricultural system, with food crops being planted during the establishment stage and bananas and Xanthosoma taro being retained in mature gardens, with Leucaena leucocephala serving as the shade tree.

On Tanna, in Vanuatu, intercropping and agroforestry strategies have been actively promoted on a 450-ha Arabica coffee nucleus estate and on a projected 250 ha of associated smallholder plots (of which 50 ha had been established at the end of 1987). The intercrops included peanuts and maize (root cropping is discouraged). Various models have been tried to protect the coffee from wind damage; these included leaving strips of indigenous forest between coffee blocks and planting casuarina shelter-belts. There is also a major French-sponsored research project on Santo, where experiments are being conducted on the intercropping of cocoa and coffee with coconuts and other species such as kava, black pepper (Piper nigrum), and vanilla, using Erythrina and Cyathea spp. (tree fern) trunks as nurse plants/climbing poles for pepper and Gliricidia for vanilla (Barrance 1988).

Apart from the almost universal planting of shade or wind-break trees, very few countries have actively and systematically promoted the intercropping of coffee with food crops. Many have actually discouraged it, even though, as noted above, studies by Carrad (1982) in the Papua New Guinea highlands showed that intercropping of root crops, bananas, and other crops with coffee, in a 15-year cycle, gave smallholder farmers almost twice the real income obtained from coffee monocropping.

Citrus with shelter-belts and intercropping

The Fiji Sugar Corporation (FSC) and its predecessor, the Colonial Sugar Refining Company (CSR), in conjunction the Agricultural Department, have, at various times, officially promoted the establishment of small mixed citrus orchards on smallholder sugarcane farms throughout Fiji. Such orchards are now common, and do provide some cane-farm families with an additional source of income through local sales, often at fresh produce markets in the main towns. In most cases, however, such efforts have met with only limited success beyond satisfying on-farm subsistence needs.

Perhaps most notable was the proposal for the eventual development of a system of 90 ha of smallholder citrus holdings, each of 2.5 ha, to supplement production from 162 ha of monocultural production at the Batiri Citrus Scheme, established on the island of Vanua Levu in 1978, funded by foreign and local investors and the Fiji Development Bank (FDB). Additionally, 200-500 ha of West Indian limes were to be established in nearby Seaqaqa and Bua to generate additional income and employment on Vanua Levu and to extend the processing season at the factory (C.P.O. 1980, 110). As early as 1984, it was suggested that the Batiri Citrus Scheme would not be feasible as a "stand-alone" orange-processing project, given the high costs of operations, inadequate financing, low long-term production levels, delayed investments in other aspects of the project, and the underutilization of land (McGregor 1984).

To achieve fuller utilization of the project, which has been hampered by the very short processing season of the oranges, diversification into pineapples, mangoes, guavas, and passion-fruit was recommended. Various intercropping trials were carried out, but, unfortunately, this highly export-oriented and capital-intensive project ran into major financial difficulties in 1988 and has subsequently been taken over by the National Marketing Authority (NMA). The smallholder "outfarm" Stage II component was never developed, although a few smallholder farmers did plant improved citrus cultivars, and surrounding Vanua Levu out-growers supplied some 10 tonnes of fruit, with another 109 tonnes coming from Rotuma in 1984 (McGregor 1984, 10). A major problem of this monocultural project has been the destruction of 25 per cent of the annual production by a fruit-sucking moth (Fiji Times 26 July 1988, 3).

In 1945 the Cook Islands began a large-scale Citrus Replanting Scheme, with the initial establishment by 1950 of 150 acres on Rarotonga and 50 acres on Aitutaki. which did include deliberate inter cropping with cover crops and multi-purpose trees. Crotalaria and tarapi grass were planted as green manure between rows to be ploughed under thrice yearly during inter-row cultivation, with leguminous dadap trees (Erythrina variegate) planted between alternate citrus trees to be trimmed to provide green manure. Shelterbelts of jambolan (Syzygium cumin)), Albizia falcataria, hibiscus (Hibiscus rosa-sinensis or H. tiliaceus - it was not made clear what species was planted), and hedge panax (Polyscias guilfoylei) were planted, with A. falcataria being grown to provide timber for citrus shipping-cases (Johnson 1951; Johnston 1967). Unfortunately, the scheme did not encourage self-reliance and has been less than successful. Many landowners have fallen into debt and have suffered widely fluctuating yields and returns.

In Nine, which has actively promoted the planting of extensive areas of limes (Citrus aurantiifolia), often as intercrops with coconuts and other crops, the export of limes and lime juice has been the second most important agricultural export after passion-fruit (Carter 1984, 298). Other countries that have tried to encourage limited intercropping of citrus include Tonga, Western Samoa, and French Polynesia.

Banana intercropping

Intercropping of bananas for export under coconuts, with cocoa, and to a lesser extent with subsistence ground crops and other trees, has been promoted from time to time in Fiji, Tonga, Western Samoa, and the Cook Islands, although recently there has been a trend towards more monocultural production.

In Fiji, the export banana industry was based on smallholder production, often under coconuts and other trees with root crops and supplementary crops cultivated in between, at least in the early stages of plantations. In the early 1960s, however, greater official emphasis was placed on monocultural production for expanding export markets in New Zealand and Japan. The major effort in this direction was the Lomaivuna Resettlement Scheme in south-east-central Viti Levu, which was designed to resettle some 200 landless families on 4ha (10-acre) allotments. The objective of the scheme was the monocultural production of bananas on 2 ha, with the balance being used for homestead and subsistence crops. As a banana scheme it failed because of serious outbreaks of black-leaf-streak fungal (Mycosphaerella musicola) and bunchy-top viral diseases, heavy nematode (Radolus similis) infestations, loss of trees rule to a series of tropical cyclones, the unsuitability of the soils and topography for banana cultivation, and the resultant poor quality of the fruit (Eaton 1988b). Because of these problems associated with monocultural banana production, export banana production in Fiji ceased in the early 1970s.

Intercropping of bananas for export with coconuts and staple root crops has also long been promoted in Tonga, Western Samoa, and the Cook Islands, and supported by considerable financial and technical support from the New Zealand marketing agent, Fruit Distributors Limited, and the New Zealand Government. As was the case in Fiji, banana production increased dramatically in the mid- and late 1960s in response to increasing market demands in New Zealand (partly due to the removal of an earlier quota system, after Fiji ceased export production), but then declined just as dramatically due to the same diseases as experienced in Fiji. Banana production in Tonga, for example, dropped 60 per cent in 1969, with further decreases in 1970 and 1971 (Thaman 1976).

In an attempt to combat declining production and improve export quality, the Departments of Agriculture in Tonga, Western Samoa, and the Cook Islands, with subsidization by Fruit Distributors Ltd and New Zealand AID, have all implemented planting or rehabilitation schemes. In Tonga, a Banana Rehabilitation Scheme, which required growers to plant one or more acres of at least 500 plants, with no intercropping, to control weeds to a reasonable standard, to practice desuckering and leaf-trimming, and to destroy bunchytopaffected plants, was instituted in the early 1970s (Thaman 1976). Despite continued efforts to rehabilitate banana production in Tonga, Western Samoa, and the Cook Islands, yields and fruit quality continued to deteriorate in all island groups during the 1960s and 1970s, with the major share of the New Zealand market being increasingly supplied by Ecuadorean bananas.

Efforts in all three countries during the 1980s to rehabilitate production of bananas for export have continued to face problems of disease, cyclone damage, and marketing. Many of these often heavily capitalized efforts have been strongly monocultural; several have failed entirely; none has met with real success; and most banana schemes have contributed to agrodeforestation.

Vanilla intercropping

Smallholder vanilla production, with appropriate support plants, and often under coconuts or other trees, has been increasingly promoted by governments throughout the Pacific. The only areas where vanilla has become a significant export, however, are French Polynesia and Tonga. In French Polynesia, where it has been intercropped under coconuts and other trees using Tecoma stans as a support plant, exports dropped from 28 tonnes in 1970 to a low of 0.6 tonnes in 1981, but renewed interest in the crop had increased production to 14.7 tonnes by 1984 (Douglas and Douglas 1989, 149).

In Tonga, where vanilla is becoming an increasingly important export crop, there is widespread official advocacy for its planting under coconuts and other trees, using physic nut (Jatropha curcas) as the support plant. Vanilla, given its high returns per acre and negligible demands on soil fertility as an epiphytic member of the orchid family, seems to be the perfect plant for Tonga, where land scarcity is increasing. Vanilla has been a major export crop in the northern Vava'u group since the 1960s, and the Department of Agriculture has been seriously promoting its cultivation in all island groups within the Kingdom for the past 10 years. By 1984, there were reportedly 105 vanilla growers with 21.2 ha of vanilla in production, 37.2 ha planted but not yet in production, and another 3.7 ha planted with support plants in preparation for vanilla planting.

Two different models have been developed to determine how best to incorporate vanilla into the traditional mixed cropping system in the belief that Tongan farmers will manage their plantations better if vanilla is introduced into a system that is already familiar (Fa'anunu n.d.). Other countries, such as Fiji, Vanuatu, and Papua New Guinea, have small vanilla development programmes, and a small number of growers in Fiji are producing marketable yields.

Oil-palm intercropping

Papua New Guinea and the Solomon Islands are the only two countries growing oil-palm (Elaeis guineensis) on a commercial scale. Papua New Guinea has three schemes, the first and largest being the Hoskins Oil Palm Development Scheme started in West New Britain in 1967, on the basis of a balance between a company plantation and adjacent smallholders; the Sangara Scheme started in the Northern Province in 1977; and the Bialla Scheme started in West New Britain in 1980. On the Hoskins oil-palm resettlement scheme, smallholders were encouraged to intercrop subsistence crops between oil-palms in the early stages of plantation development. Consequently, in the late 1970s, 31 per cent of the settlers were able to practice market garden ing of traditional foods, and, given access to land, the farming systems of almost all settler groups had remained virtually intact with people growing their own food (Benjamin and Wapi 1982). In the Solomon Islands, however, the large-scale Guadalcanal Plains joint venture oilpalm development scheme is entirely monocultural, with little encouragement of intercropping or agroforestry practices (Carter 1984).

Kava intercropping

Kava (Piper methysticum) is an important crop in Vanuatu, Fiji, Tonga, Pohnpei, and to a lesser extent in Western Samoa. Given its shade tolerance and that it takes 3-6 years to mature, it is particularly amenable to intercropping both with short-term food crops and with coconuts and other trees.

In Vanuatu, where kava cultivation and consumption is, in many rural areas, highly ceremonial and very significant in garden magic, the plant is almost always intercropped amongst trees in polycultural gardens. Kava is also one of the main cash crops promoted as an intercrop with coconuts and in timber plantations of Cordia alliodora.

In Fiji, kava (known locally as yaqona) is a major cash crop for local sale and the major cash crop for villagers in some of the smaller outer islands and the interiors of the larger islands. Programmes to rehabilitate the copra industry have encouraged the establishment of other economic crops such as cocoa and kava underneath coconut palms. Some Fijian kava farmers are beginning to intercrop kava with Calliandra calothyrsus, a nitrogen-fixing species being promoted for agroforestry in Fiji by the Fiji-German Forestry Project. Throughout Fiji, kava is commonly intercropped with staple root crops, particularly taro (Colocasia esculenta). The kava is then allowed to mature amongst existing cultivated and protected trees after the root crops have been harvested.

Whereas copra prices are dictated by outside market forces beyond the control of Fiji and its farmers, most of the kava produced is sold internally and has shown much greater price stability than copra. Of further significance is that almost all kava farms are Fijian-run. Sofer (1985) argues that the success of kava as a commercial crop in Fiji results from the familiarity of Fijian farmers with its husbandry. Also, kava provides higher returns to the grower, per unit of land, labour, and capital investment, than most alternative cash crops; it requires few outside inputs; its planting material is readily available and consists of parts of the plant not normally sold; it can be harvested or processed at any time; it is non-perishable and can be easily stored; and it is well-adapted to the environmental conditions and to the semisubsistence polycultural shifting agroforestry system in Fiji.

In Tonga, kavais also seen as an important local cash crop and is almost always planted under coconut palms and other useful trees.

Other commercial agroforestry intercrops

In Fiji, a number of successful smallholder papaya plantations, sometimes under coconuts or as part of diversified agricultural developments, supply the growing demand for the fruit in the country's tourist resorts. Unfortunately, a major 30-ha capital-intensive papaya plantation developed by the Native Land Development Corporation to air freight fruit to Japan has been a failure. Most recently, however, the Southern Development Company Ltd. the company responsible for directed smallholder tobacco production for Fiji's local cigarette manufacture, has attempted to diversify the production of its directed smallholder farmers into tomatoes, maize, chillies, mangoes, and papayas. Papayas seem to offer the greatest potential, with 18 farmers already growing the crop under the auspices of the company (Eaton 1988, 194-195).

Areas of improved mango varieties for export have also been planted by the Native Land Development Corporation and on the Government-run Yaqara Beef Cattle holding on Viti Levu as part of a diversification programme. Papua New Guinea, Vanuatu, Tonga, Western Samoa, and the Cook Islands have also experimented with improved mango cultivars, as well as avocados in efforts to diversify tree crop production.

Smallholder passion-fruit (Passiflora edulis) production, often as an intercrop with coconuts or other trees or on small plots within existing plantation areas, has been strongly promoted in Papua New Guinea, Fiji, Tonga, Western Samoa, and Nine, as part of agricultural diversification programmes. Results have been mixed. Hurricanes have damaged vines, market availability has been inconsistent, as has ability to supply, given low production because of poor management and insufficient labour to carry out daily the hand pollination of the flowers needed to obtain high yields. The industry has showed promise only in Fiji, where the production is run as a directed smallholder operation by a fruit processing and exporting company that provides most of the inputs, planting material, and research, extension, and managerial services, and in Nine, where passion-fruit is the most important agricultural export (Carter 1984).

In Fiji, the aid-funded Fiji-German Forestry Project, working together with Fiji's Forestry Department and the Extension Service of the Ministry of Primary Industries, is promoting a kind of alleycropping system through the introduction of hedgerows of Calliandra calothyrsus along the contours of sloping land used for various crops, particularly ginger. In less than two decades, ginger has grown to be Fiji's third-largest agricultural export earner, surpassed only by sugar and coconut products. A combination of the environmental requirements of ginger (high rainfall, fertile soils, good drainage, freedom from nematodes) and the low capital resources of the smallholders planting the crop has led to a "ginger frontier," where ginger is grown on plots newly cleared of forest for a few seasons and then moved on to still more recently cleared land, leaving behind a cleared, less fertile area (Overtop 1989, 83). The Calliandra hedgerows, which are recurrently pruned, are intended to slow erosion, which has been serious at times on the steep ginger-growing lands, to provide nutrients and firewood, and to serve as wind-breaks. Attempts to analyse the economic benefits of Calliandra hedgerows suggest that the labour costs of establishing the hedgerows are repaid in two years by savings on fertilizer (Künzel 1992, personal communication).

Planting of timber, fuel wood, and general-purpose trees in relation to agroforestry and agriculture

Because the imminent demise or depletion of commercially usable natural forests can be so readily foreseen in many Pacific Island countries (Watt 1980, 297), governments and development agencies have in several places promoted either some form of restocking or enrichment of commercially logged areas or the establishment of forest plantations on degraded grassland sites. Not all these efforts can be classified as agroforestry, strictly speaking; but in the Pacific context, as in most of the tropical world, the traditional, if transient, shift of land use back and forth between forest and agriculture on any particular site makes it relevant to consider what at first glance appear to be purely forestry projects.

Many of the timber species institutionally promoted have been exotics such as Caribbean pine (Pinus caribaea), West Indian mahogany (Swietenia macrophylla), cordia (Cordia alliodora), and Eucalytus spp., although some indigenous Pacific species such as Albizia falcataria, Agathis spp., Araucaria spp., and Endospermum spp. have been successfully established, often as exotics in areas beyond their natural range. Many other species - including West Indian cedar (Cedrela odorata), the silky oak (Grevillea robusta), teak (Tectona grandis), mahogany (Swietenia mahogoni), toon tree ( Toona australis), cadamba (Anthocephalus chinensis), and Albizia lebbeck along with several indigenous trees - have also been the subject of trials, and planted to various degrees throughout the islands.

Firewood and multi-purpose species that have been successfully introduced include Leucaena leucocephala, Erythrina spp., Casuarina spp., and Gliricidia septum, and, to a lesser extent, Securinega samoana and Adenanthera pavonina. Other species, all of which have been planted experimentally and which seem to grow successfully, but which have not yet become so well established, include Cassia, Acacia, and Calliandra spp. Apart from timber and fuel wood, the major multi-purpose objectives of such plantings are site reclamation and amelioration, erosion control, wind protection, shade, multipurpose construction and handicrafts, nurse cropping, fodder, green manure, and food.

The indigenous casuarinas, particularly Casuarina equisetifolia, have also shown considerable promise for reforestation programmes, and have been planted in Tonga in land reclamation projects, in the Cook Islands for the rehabilitation of degraded lands, and on atolls as sources of fuel wood and to protect coconut plantations from saltwater damage. C. oligodon and C. papuana are traditionally used for reforestation and to enrich fallow land in Papua New Guinea, and are now promoted in some areas for land rehabilitation and as shade plants for coffee.

Pine planting in relation to agroforestry

Of the total area of timber plantations in the Pacific, well over 50 per cent is accounted for by Caribbean pine (Pinus caribaea). The largest area of pine planting is in Fiji, where that country's Pine Commission together with the Forestry Department has established over 50,000 ha of plantation since 1960, mostly on degraded anthropogenic grasslands (Drysdale 1988a, 110; Watt 1980, 301). Some pine timber is used locally, but the wood was intended mainly for export, and a wood-chipping mill is now in operation. In the mid1960s, under a programme now discontinued, woodlots of Pinus caribaea on smallholder sugar-cane farms were promoted by the colonial government.

Sized from 0.4 to 2 ha, these woodlots were planted on steeper non-cane areas of farms to control erosion, provide on-farm supplies of timber and fuel wood, and for undergrazing by farm animals (Eaton 1988b, personal communication). Apart from this woodlot grazing and grazing of cattle in association with larger pine plantations (described below), there has been no institutional support for any form of intercropping or other agroforestry activities in pine plantations (Drysdale 1988b).

Similarly, in the limited areas of pine planting in New Caledonia, Western Samoa, Tonga, and the Cook Islands, there has been little or no link to agroforestry in such programmes, with the main focus being on creating a timber resource, land improvement, erosion control, and employment creation in rural areas.

In highland Papua New Guinea large areas of degraded grassland have been planted with pines (Pinus spp.) and Araucaria spp. Intercropping activities are few and consist of the intercropping of coffee and cardamon on a trial and demonstration basis (Howcroft 1983).

In Vanuatu, P. caribaea var. hondurensis is the main species planted in forest plantations in seasonally dry and highly degraded sites on the southern islands of Aneityum and Erromango, where some 550 ha had been established up to April 1985. The commercial viability of such plantings is still uncertain, however, due to poor access to markets and high transport costs. On Erromango, high costs of clearing land of the indigenous pioneering species Acacia spirobis has stopped the development of pine plantations. Benefits in the form of erosion control and aiding the local economy through wages were the main motives behind these programmes (Neil 1986a).

Non-pine forestry in relation to agroforestry

To judge from programmes in Papua New Guinea, the Solomon Islands, Vanuatu, Tonga, and Western Samoa, there seems to be greater promise and greater institutionalized promotion of intercropping with other, primarily broadleaved evergreen, species than has been the case with pines.

In Papua New Guinea, where extensive areas of Eucalyptus deglupta have been planted, cocoa and coffee have been successfully grown at 4 m x 4 m and 3 m x 3 m spacing, respectively, in conjunction with E. deglupta planted at 10 m x 10 m (Jacovelli and Neil 1984, 10).

Also in Papua New Guinea, severe environmental degradation resulting from rapid urban expansion and associated subsistence gardening and "fuel-wood mining" prompted the cities of Lae and Port Moresby to institute fuel-wood-planting programmes. In Lae, in 1978, it was decided to plant 200 ha of sloping land (20°-30°) in Leucaena leucocephala for firewood and to intercrop fuel-wood species with annual food crops in zones designated for subsistence food gardening. The project, which was allocated K250,000 (US$275,000) over six years, had a management component coupled with a public education programme and a team of local government rangers to control gardening and to police the area (King 1987). Follow-on projects were planned but not carried out because of lack of funding. By 1988 the project had ceased to operate, and the original plantings of some 100 ha of L. Ieucocephala, Acacia auriculformis, and Eucalyptus spp. and 5 ha of "agroforestry plantings" of fuel-wood species with food crops had been cut down or removed completely (King 1987).

In Vanuatu, Cordia alliodora, a hardwood native to Central America, has been the main commercial silvicultural species since the mid-1970s, with over 1,000 ha planted on 12 islands as of 1984 (Neil 1984). Cordia was first planted on various islands in 5-10-ha blocks called Local Supply Plantations (LSP). As the potential contribution of forestry to rural and national development became evident, larger, export-oriented Industrial Forest Plantations (IFP) were established on the islands of Pentecost, Erromango, and Aneityum (Jacovelli and Neil 1984). The rapid expansion of IFPs, sometimes with plantings of up to 200 ha per year on single sites, led to unprecedented demands for land and aroused fears among landowners, especially on Pentecost, that these silvicultural activities would make land unavailable for planting subsistence and commercial crops. This prompted the Vanuatu Forest Service to establish, on Pentecost in 1984, demonstration plots growing a wider range of subsistence and cash crops within forestry plantations of Cordia alliodora (Jacovelli and Neil 1984).

Crops established between line plantings of Cordia alliodora included 8 sweet potato cultivars, 6 cassava cultivars, 13 aroid cultivars from Colocasia esculenta, Xanthosoma sagittifolium, and Alocasia macrorrhiza, 12 yam cultivars, kava (Piper methysticum), and trials with coffee (Arabica and Robusta), cocoa, and cardamon. In addition to these trials, subsistence gardens have also been established under Cordia alliodora by both local landowners and forest workers alike (Jacovelli and Neil 1984, 8).

Because C. alliodora may be severely attacked by root rot (Phelli nus noxius) in some conditions, and does not perform well on some sites, other species currently being tried in Vanuatu include Terminalia brassii, T. calamansanai, Eucalyptus deglupta, Swietenia macrophylla, Toona australis, and Cedrela odorata. However, the barks of both T. brassii and E. deglupta are palatable to cattle (Jacovelli and Neil 1984, 10; MacFarlane 1980). The species showing greatest potential as an alternative species to C. alliodora may be S. macrophylla, and if grown with nurse species to reduce pest problems, intercropping should be possible during the early years of rotation (Neil 1986b).

Several other systematic experiments on tree species, both exotic and indigenous, have been carried out in Vanuatu in a search for species especially suitable for fuel wood, timber, or pulpwood, but none of this research was connected with agroforestry. Research on agroforestry has focused almost exclusively on "cash crops which appear to have great potential, particularly coffee and cocoa, and possibly kava and cocoa" (Jacovelli and Neil 1984, 11).

In Fiji, some 22,953 ha of tropical hardwood forests have been planted as of mid-1986. Of these, 14,987 ha are West Indian mahogany (Swietenia macrophylla), 3,058 ha are Cordia alliodora, 2,963 ha are cadamba (Anthocephalus chinensis), 928 ha are Maesopsis eminii, 438 ha are Eucalyptus deglupta, and 202 ha are the indigenous species Endospermum macrophyllum (ADAB 1986). Despite such considerable silvicultural activity, in terms of both hardwood and pines, it is essentially monocultural, and, as the General Manager of the Fiji Pine Commission has stated: "Institutionalized agrosilviculture is non-existent in Fiji at present" (Drysdale 1988b, personal communication).

Tonga's silvicultural activities are more diverse, some being significantly agrosilvicultural. More purely silvicultural activities include a major reforestation programme begun on the island of Eua in the mid-1960s. Over 40 ha of mixed exotic species including Toona australis, Cedrela odorata, Cordia alliodora, Grevillea robusta, Agathis robusta, Pinus caribaea, and Eucalyptus spp., as well as suitable indigenous species, such as Casuarina equisetifolia, Terminalia catappa, and Dysoxylum tongense, were planted on the Eua Forest Farm. Tests of seed stock from throughout the world were also carried out on the farm. Larger areas were subsequently planted, with 104 ha alone being planted in 1979 (Thaman 1984e, 3).

The species most commonly planted in 1984 were Eucalyptus saligna, E. tereticornis, Toona australis, and Pinus caribaea. Seedling pro auction for these species and other timber species, such as Cupressus lusitanica, amounted to 77,491 seedlings (42,427 of which were planted) in 1979 (MAFF 1985, 100-102). Reforestation continues, as the small areas of remaining indigenous forest on Eua are exploited, with the local mill "approaching the end of its productive life as the local hardwood timber supply is cut out and cannot be replaced from the Forest Farm for at least another 10 years" (MAFF 1985, 99). The only truly agroforestry aspect of the Eua silvicultural activities, a taungya system of combined tree-planting and temporary gardens, was phased out because "it has greatly increased pressures for settlement of unsuitable land, and is thus clearly not in the national interest" (MAFF 1985, 100).

A second and continuing agroforestry activity has been the Forestry Extension Programme, which began in the 1960s to produce seedlings for distribution to smallholder farmers for planting in small woodlots or as windbreaks around their agricultural allotments (see chapter 5 on Tongan agroforestry). The major species distributed included Casuarina equisetifolia, Grevillea robusta, Cedrela odorata, Eucalyptus spp., Agathis spp., and Gmelina arborea (Thaman 1984e, 3).

With the establishment of the Extension Nursery at Mataliku on the main island of Tongatapu in 1978, the programme was expanded to include the propagation and distribution of a wide range of timber trees, "cultural" species, and species providing food, medicine, and ornamentation. The considerable interest shown by the people for planting on both rural and town allotments led to a "blossoming of forest extension work" to the point that, in 1978, the nursery could not cope with the demand, which exceeded 8,000 trees per month (MAFF 1979, 99).

According to programme records, as of 1984, at least 155 species had been tested and/or propagated for distribution on Eua and Tongatapu. Of these, 66 were timber species, 45 ornamentals, 32 "cultural" plants of particular importance to the Tongan society, 11 food plants, 6 plants used for coastal protection or land reclamation, 4 for living fences or hedgerows, 3 medicinal plants, and 2 each for windbreaks and firewood. Among the most popular nontimber species were Casuarina equisetifolia (planted as an ornamental, living fence, or wind-break); culturally important sacred or fragrant plants, known locally as akau kakala, such as heilala (Garcinia sessilis), langakali (Aglaia saltatorum), sandalwood, or ahi (Santalum yasi), pua (Fagraea berteriana), pipi (Parinari glaberrima), huni (Phalaria disperma), perfume tree, or mohokoi (Cananga odorata), allspice (Pimenta doica), and Pandanus cultivars; fruit-trees, such as mango, Malay apple (Syzygium malaccense), and macadamia nut (Macadamia integrifolia); and ornamental or shade plants, such as flamboyant, or poinciana (Delonix regia), hibiscus (Hibiscus rosa-sinensis), Cordyline fruticosa, copperleaf, or beefsteak, plant (Acalypha amentacea), bougainvillea (Bougainvillea spp.), poinsettia (Euphorbia pulcherrima), gardenia (Gardenia spp.), and the hedge panaxes (Polyscias spp.) (Thaman 1984e).

The final major area of activity has been the testing and establishment of trees for land reclamation, such as the project to rehabilitate low-lying areas at Sopu to the west of the capital of Nuku'alofa on Tongatapu. Reclamation work at Sopu began in the 1960s, with the planting of Casuarina equisetifolia to stabilize the area, and has continued to the present with extensive plantings of Lumnitzera littorea, Rhizophora mangle, Bruguiera gymnorhiza, Xylocarpus granatum, and other selected species. As recently as 1980, 6 acres of Lumnitzera littorea, 4 acres of Terminalia catappa, and 3 acres of Queensland kauri (Agathis robusta) were planted. The vegetation has reportedly been well-established, with the operation becoming more maintenance than reclamation.

Grazing with commercial tree cropping and silviculture

Grazing, usually of cattle, with commercial tree cropping and silviculture consists mainly of the widespread practice of grazing cattle under coconuts or commercial timber species, and the limited grazing of cattle under Leucaena leucocephala or other fuel-wood or multipurpose species.

Livestock under coconuts

The grazing of cattle (primarily beef, but also dairy cattle) under coconuts (in some cases with pasture improvement) is by far the most widespread practice. It has been encouraged throughout the Islands since colonial times, particularly on large coconut estates. In addition to providing meat and dairy products, cattle are seen as effective weed control and fertilization agents, thus facilitating plantation management and the collection of fallen nuts.

Although primarily promoted on large, often foreign or state controlled estates or plantations, some governments, such as those in the Solomon Islands, Tonga, and Niue, have encouraged smallholder grazing of cattle under coconuts and other trees. In the case of Tonga, smallholder agriculturalists have been encouraged to fence limited portions of their 3.3 ha bush allotments to graze cattle, and sometimes horses, under coconuts and other tree crops and protected trees, or, alternatively, to tether animals to trees and graze on a rotational basis.

The practice has been particularly important in Vanuatu (both before and after independence in 1980) and New Caledonia, where beef cattle production is a major activity. Beef cattle production became so important in Vanuatu, prior to independence, that some plantations were turned into cattle properties. The importance of cattle grew in the 1950s, when steeply rising labour costs made planters increasingly dependent on cattle to keep their plantations clean. At one period in the 1950s, herds became larger than the plantations could support, especially during dry spells, and by the end of the decade, town butcheries had opened in both Port Vila and Luganville, the two main towns. By the end of the 1960s, copra production had become no more than a sideline on a number of plantations (Brookfield with Hart 1971, 164165).

In Fiji, in 1973, 10.5 per cent of the local beef requirements were supplied by the 9.9 per cent of the cattle population grazed under coconuts (MAF 1973; Manner 1983). This is particularly significant given the large proportion of range-fed cattle raised on extensive large-scale developments in the dry zones of Fiji. Papua New Guinea, the Solomon Islands, Vanuatu, and New Caledonia in Melanesia, and Western Samoa and French Polynesia have also actively encouraged cattle under coconuts with trials having been conducted on optimum stocking rates and pasture improvement. Much of the Western Samoa Trust Estates (WSTEC) Mulifanua Copra Plantation, reportedly one of the largest copra plantations in the world (Carter 1984), is undergrazed by cattle.

The potential for the formal promotion of large-scale grazing of cattle under coconuts is greatest on the larger islands of Melanesia and Polynesia. On smaller islands, such as those in Tonga and the Cook Islands, where high population densities and land scarcity make more extensive agrosilvipastoral developments less relevant, small-scale rotational undergrazing of tethered animals is more appropriate. In Nine, where population density is low because of emigration to New Zealand, there have been problems of overgrazing and lack of fodder during times of drought- for example, during the severe drought of 1977-1978, when hay had to be imported from New Zealand.

Richardson (1983, 59) cautions that grazing under coconuts can create problems of soil compaction and, especially in the case of free grazing, preclude intercropping, which should take precedence in areas with limited land resources. As shown by studies in Papua New Guinea and elsewhere, smallholder beef cattle production can have harmful impacts on subsistence cropping (Grossman 1981). Where cash cropping or subsistence production is feasible, Richardson (1983, 59) argues that intercropping should take precedence over grazing under coconuts.

Cattle under timber species

The grazing of cattle under commercial timber species has been actively promoted in Papua New Guinea, the Solomon Islands, Vanuatu, and Fiji. In Papua New Guinea, reforestation projects in both the highlands and lowlands offer opportunities for beef production, and cattle have been actively promoted to control weeds and reduce fire danger by consuming the fuel. Pinus caribaea planting has also been encouraged in order to provide shade for cattle in open grasslands (Watt 1980, 308). The introduction of pasture legumes into timber plantations and surrounding areas has also been actively encouraged, and the development of pastures, followed by grazing, has been more or less standard practice in a number of forest plantations in Papua New Guinea, where klinki and hoop pine (Araucaria spp.), Pinus caribaea, and Eucalyptus spp. are grown. Government forest plantations are made available to local Braziers who establish adequate fencing and pastures and follow acceptable range management and stocking practices (Howcroft 1974; 1983).

In the Solomon Islands, where there is a "Cattle Under Trees" (CUT) project, cattle have been grazed under Eucalyptus deglupta in forest plantations established by the government in logged forest (Macfarlane and Whiteman 1983; Schirmer 1983, 101; Watt 1980, 308) and in Vanuatu under both "Local Supply Plantations" and "Industrial Supply Plantations" of Cordia alliodora, as well as under Pinus caribaea on Aneityum, Erromango, Pentecost, and Santo (Jacovelli and Neil 1984, 8). Grazing under pines in Vanuatu is seen as a means of reducing the significant fire threat in plantations (Neil 1986a).

It is in Fiji that the practice has probably been tried most exten sively, owing to research undertaken by the Fiji Pine Commission (FPC), a statutory body with the objective of facilitating and developing "an industry based on the growing, harvesting, preserving and marketing of pine and other species of trees grown in Fiji" (CPO 1980, 141). The FPC is responsible for managing over 45,000 ha of Pinus caribaea out of an envisioned gross estate of 80,000 ha on the highly degraded talasiga (sunburnt) soils of the drier leeward grasslands of the two largest islands of Fiji. The relatively infertile and eroded areas are vegetated with a grassland sub-climax of presumed anthropogenic origin, including species such as Pennisetum polystachyon, Pteridium esculentum, Gleichenia liners, Psidium guajava, Dodonaea viscose, and Casuarina equisetifolia. On moister slopes, Miscanthus floridulus forms almost impenetrable thickets. These grasslands are subject to frequent and unauthorized burning.

The FPC undertook research into cattle grazing for two reasons: to examine the effects of cattle grazing on reducing fuel in high fire-risk zones; and to test the use of cattle as a site-preparation tool for clearing the land of Miscanthus floridulus, which proved difficult to eradicate by more conventional means such as slashing and burning (Drysdale 1982). Research has yielded variable results. Vincent (1971) concluded that grazing of cattle under 5- and 6-year-old pine plantations in poor soils had a detrimental effect on the incremental growth of pines, whereas grazing trials in the Nausori Highlands to determine the effect on fire hazard reduction resulted in a reduction in fuel from 2,500 kg per hectare to 800 kg per hectare, an average cattle weight gain of 0.24 kg per day, and no pasture deterioration despite heavy stocking rates (Gregor 1972). At Nawaicoba, Partridge (1977) reported weight gains twice this, when trees were planted at 2 m x 3 m spacing, with two rows in every five missing. In variable spacing trials, Bell (1981) found slight bark damage to trees less than one year old because of trampling, when the trees were spaced 3 m apart within rows and 2.5, 3, 3.5, and 4 m apart between rows, the cattle being introduced into the plantation when the pines were 54 cm high.

In 1982, the FPC reviewed various research projects on cattle under pines and concluded that given "the high overhead and general costs of FPC operations, commercial cattle grazing of unimproved pasture under pines, is an unlikely prospect" (Drysdale 1982, 4). Although fuel loadings were considerably reduced, the cost of using cattle for fuel reduction was "considered unacceptably high compared with alternatives such as burning" (Drysdale 1982, 3). In contrast, the use of cattle as a site-preparation tool where Miscanthus predominates was termed an "outstanding success" (Drysdale 1982, 8) because other methods of clearing the giant grass gave incomplete results, were impractical, or cost too much.

Because of the high cost of fencing, the long-term and extensive grazing of cattle under pines has been found to be an uneconomic proposition for the Fiji Pine Commission, although some 480 cattle are allowed to graze under pines free of charge at Drasa and Tavaka-bo, and some cattle owners unofficially graze their cattle in Fiji Pine Commission forests. Native landowners are also allowed to graze cattle under their own pine plantings, subject to certain restrictions. But cattle owners also are unlikely to find fencing a profitable venture. Open-range grazing with night-time penning may be a possibility. In addition, the economics of cattle grazing on improved pastures under trees in Fiji still needs to be ascertained.

Other silvipastoral activities

Trees such as Leucaena leucocephala are used as fodder in Tonga and Papua New Guinea, where they are browsed by cattle as a dietary supplement (Watt 1980, 308). There is perhaps some scope for the grazing of other animals such as pigs, goats, and chickens on improved legume pastures or fallows under coconuts, commercial timber species, or other trees (Quartermain 1980; Richardson 1983).

The future of institutional agroforestry in the Pacific

In the Pacific, as elsewhere, interest in agroforestry has recently grown rapidly among scientists, land-use experts, conservationists, and the development professionals of national governments and international agencies. As already noted, systems of commercial production that would now be classified as agroforestry were initiated early in the Pacific's colonial past, particularly in the form of multistorey arrangements of coconut palms with other crops or with cattle. With regard to agroforestry systems in the subsistence sphere, this book has sought to demonstrate their prevalence and antiquity in the Pacific Islands. As Yen (1980b, 91) comprehensively expressed it in his discussion of "Pacific Production Systems," there is nothing new about multi-storey cropping even though it has often been suggested to smallholders as an innovative technique they might adopt.

In fact native systems have always involved such techniques in village gardens with descending storeys of palms, trees, productive vines, shrubs, herbaceous root crops, and vegetable plants and ornamentals. Similarly, in swiddens, mixed species and variety plantings are themselves multi-storey. In this case such plantings also take on a successional aspect, for following the root crops, some cultigens such as banana and longer-term plants such as breadfruit and other fruit and nut trees, industrial shrubs, and vines, prolong the production of these gardens.

Geographers and anthropologists who have studied these sorts of indigenous systems find ironic some of the attempts made to introduce institutional agroforestry into the Pacific context. On the other hand, in a time of deforestation and agrodeforestation, it is apt to encourage both of the approaches to agroforestry described in chapter 1- the institutional approach, which generally seeks to introduce commodity-focused systems devised on the basis of modern forms of analysis, and the cultural-ecological approach, which is concerned more with long-standing indigenous systems, empirically devised and deeply embedded in the cultural landscape. Whether or not the two approaches can be usefully meshed remains open to question, although some forms of "progressing with the past" do seem possible (Clarke 1978).

When attention is turned to the future of institutional agroforestry in the Pacific, it can be clearly forecast that if individual smallholders are to benefit over the long term from the introduction of an unfamiliar institutionalized agroforestry system, they will need to receive an ongoing package of inputs and information, which suggests the need for some sort of extension service. Unfortunately, it is acknowledged that extension work in many Pacific countries is generally poor, and extension services often have only secondary ranking within ministries or departments (Hau'ofa et al. 1980, 188-189). How to remedy this deficiency raises several complex but pervasive issues, which have been dealt with at length in a large literature and which can only be superficially treated here.

With regard to the initial introduction of a new agroforestry system, it is easy - given the current popularity of agroforestry in the development world to find funding for workshops and projects, but these by their nature lack continuity, and they are often administered by staff unfamiliar with local agroforestry traditions. The Pacific is littered with projects advanced in support of all sorts of good causes their collapsed remnants remain, like the military paraphernalia rust ing on beaches after World War II. One way to incorporate continuity into projects and to move beyond reliance on inadequate extension services is to form a centralized management system for smallholders (sometimes referred to as a plantation mode of management). Such a system has been successful in several instances, notably the efficient smallholder production of sugar so important in Fiji's economy and also in tobacco production in that same country (Eaton 1988a). Some other attempts have been less successful. The pros and cons of the approach have been cogently summed up by Hardaker et al. (1984a; 1984b) and Ward (1984).

Aside from problems common to any project-based introduction, a specific constraint to the full realization of the potential of agroforestry by institutional means relates to the disciplinary compart-mentalization that characterizes institutions concerned with land use, whereby - as the Director of ICRAF commented - "agriculture and forestry normally fall under different ministries or, if they are under the same ministry, under separate departments,' (Lundgren 1987, 44). Writing specifically of the forestry sector in the South Pacific, Watt (1980, 302-303) noted that "the separation of agricultural and forestry extension services encourages the impression that agriculture and forestry are mutually exclusive alternatives rather than complementary land uses." Following on from and related to this sectoral compartmentalization is each institution's imperative to maximize the individual component that is the focus of that institution. In contrast, as has often been observed:

The subsistence land user's strategy and aims are to use his labour and land resources to optimize, with minimum risk, the production of various products and services required to satisfy all his basic needs. The fundamental inadequacy of conventional-discipline-oriented institutions lies in the failure to acknowledge and understand these basic facts, strategies and aims, and in the inability to adapt to them. The aims, infrastructure, rationale and philosophy of these institutions, as well as the training of their experts, are geared to the maximization of individual components, be they food crops, cash crops, animals or trees. There is little understanding that the land user needs to share out his resources for the production of other commodities or services (Lundgren 1987, 46).

When maximization is aimed at commercial products, as it most frequently is in the Pacific, a set of sometimes contradictory processes comes into play. For example, attempts to produce cash crops while continuing to meet subsistence needs may bring agricultural involution if land is limited, or it may result in an extension of cropping onto marginal sloping lands as cash crops or cattle take over better lands. A specialization in commercial products may not be accompanied by any concomitant increase in labour availability or extension advice (often restricted to larger producers) on how to increase subsistence production (Ward 1986; Yen 1980b).

Even the Fiji-German Forestry Project, which commenced in the mid1980s, appears mainly focused toward facilitating export cash cropping, although its terms of reference suggest a broader approach that includes "providing ecologically sound advisory assistance in the fields of forestry and agroforestry in line with the social, cultural and economic requirements of target groups" (Tuyll 1988, 3). Consultants to the Fiji-German Forestry Project have also made holistic and wide-ranging recommendations, but the Project's current activities, as described earlier in this chapter, are concentrated on improving the production of ginger as a cash crop by introducing exotic trees to prevent erosion and replace artificial fertilizer.

This accomplishment is not to be decried, but the approach, distinguished by its introduction of and experimentation with exotic trees alley-cropped with a cash crop, does little to preserve existing agroforestry systems or to maintain a balance between commercial agroforestry activities and activities that could protect the existing subsistence base. One consultant recommended to the Project that "agroforestry and forestry extension should not attempt to remain with or return to pure forms of subsistence economy but focus on including profitable cash crops at low risks" (von Maydell 1987, 35). This recommendation does indicate an appreciation of the need to minimize risk, but both it and all the other consultants' recommendations to the Project fail to support strongly the maintenance of a viable subsistence base. Another consultant, who had been selected to identify suitable sites for demonstration plots for the Project, was asked to comment on the idea of putting greater emphasis on the subsistence aspects of agroforestry and of analysing existing local agroforestry systems as demonstration plots into which selected improvements could be introduced. He responded that it was quite unrealistic to expect either the Fiji Government or the German funding agency to support such an emphasis in place of an emphasis on using agroforestry as a way to improve monocultural cash cropping.

In summary, export crops, timber trees, and grazing under coconuts have been the continuing focus of almost all official agroforestry activities for the past century. Regardless of whether it has been the colonial or post-colonial agricultural and forestry departments or, re cently, international aid agencies, the focus has been almost exclusively on monocultural, often large-scale production for export or, in the case of timber and fuel-wood production, for import substitution. Even the intercrops are usually cash crops for export or local sale. Consequently, most indigenous wild species and the wide range of traditional cultivars have received little official promotion and have been the focus of only limited research. Few technical experts or development entrepreneurs know enough about traditional mixed agricultural systems and their component plants to be willing or able to promote their expansion or maintenance. It is not only projects intended to develop commercial agriculture and forestry that may displace or degrade traditional agroforestry systems; modern institutional agroforestry projects may themselves play the same role.

Agencies and educational institutions promoting agroforestry

However, there are also movements in support of traditional systems. The growing popularization and recognition worldwide of the value of the "wisdom of the elders" (Knudtson and Suzuki 1992) may motivate increased institutional attention to indigenous polycultural systems of agroforestry in the Pacific. This section provides information on several examples of such attention and on the institutions involved; mention has been made earlier of some of these, but they will be referred to here briefly again to provide a coherent single account.

All the major universities within the Pacific region (University of Guam, both of Papua New Guinea's universities, the University of the South Pacific in Fiji and its School of Agriculture in Western Samoa, University of Hawaii, and the developing francophone institutions in New Caledonia and Tahiti) support staff with interests in traditional matters, including agriculture, agroforestry, and the management of soil and vegetation. Rather than attempt a full listing of course offerings relevant to agroforestry to at least some degree, we note here only that, on the basis of current information at hand, the courses most directly focused on agroforestry are found within the Geography Department at the University of the South Pacific in Suva, Fiji, and the Department of Agronomy and Soil Science at the University of Hawaii in Honolulu. To the best of our knowledge, the University of Hawaii is distinguished by being the only university in the region to have a named Professor of Agroforestry, who is located in the Department of Agronomy and Soil Science. The Col lege of Micronesia in Pohnpei also has staff with active and direct interests in indigenous agroforestry.

Agroforestry promotion by the Fiji-German Forestry Project, a bilateral agency, has been described in the previous section. A different approach is followed by the South Pacific Forestry Development Programme, which is a multilateral 5-year project funded by UNDP, executed by FAO, and now based in Suva, Fiji. The Programme is concerned with forests and trees in 15 countries, so far particularly with forests in the larger countries, but atoll countries are making enquiries about coconuts and other multi-purpose trees. The role of the Programme is to stimulate activities and provide technical advice, not to operate activities itself. For instance, it facilitated the import of seeds of superior rattan from Malaysia for planting in Pacific forests in order to increase their non-timber production capability. Aside from technical advice, the Programme acts as a focal point for information about forests and trees and publishes the quarterly South Pacific Forestry Newsletter. It is also trying to organize the documentation of local knowledge on indigenous agroforestry, with studies planned or under way in Pohnpei, Fiji, Kiribati, Tuvalu, Tonga, and other island countries.

The Programme has worked cooperatively with the international NGO The Foundation of the Peoples of the South Pacific (FSP) on a project intended to develop sustainable forestry in local areas while slowing down or stopping rapid conversion of forests by large-scale industrial logging. This objective is based in part on selling small mobile sawmills to rural entrepreneurs and community groups so that they may develop small-scale but profitable and locally utilitarian logging, carried out in ways that avoid major environmental damage and that maintain the essential structure of the forest for traditional uses and ecological services.

A US Government project based in Hawaii is carrying out work related to several aspects of agroforestry in Hawaii, American Micronesia, and American Samoa. Called Agricultural Development in the American Pacific (ADAP), the project has provided agroforestry educational materials to all the public (land grant) colleges and universities in the American-affiliated Pacific. In association with the US Department of Agriculture and the US Forest Service, ADAP is also developing training programmes in agroforestry.

The Environment and Policy Institute of the East-West Center in Hawaii maintained a strong programme of research, seminars, and publication on agroforestry for several years during the 1980s (e.g., Djogo 1992; Nair 1984). Although agroforestry is no longer a principal focus of its work, the Institute remains a repository of a large volume of published and unpublished material on the topic.

Mentioned at the beginning of this chapter was the report (Clements 1988) of a technical meeting on agroforestry in tropical islands held at the Institute for Research, Extension and Training in Agriculture (IRETA), which is part of the University of the South Pacific's School of Agriculture in Western Samoa. IRETA is also involved in research projects to improve or strengthen atoll agroforestry in Kiribati.

In the Melanesian countries, with their comparatively larger natural forests, forest-resource inventories are under way or planned, generally as a cooperative, aid-funded project between the local Forestry Department and overseas technical personnel. The inventories are intended to provide the information base necessary for effective land-use planning and management, but now, unlike some past forest assessments, the inventory process includes collection of data on watershed vulnerability and on the indigenous ethnobotanical value of forest plants, as in the forest-resource inventory now being completed by the Vanuatu Forestry Department with technical assistance from the Queensland (Australia) Forest Service and the Division of Tropical Crops and Pastures of the (Australian) Commonwealth Scientific and Industrial Research Organization (CSIRO).

Finally, mention should be made of the work of ORSTOM, the French organization that promotes French scientific research in the third world, mainly in the tropics. With centres in the Pacific in Nouméa and Tahiti, ORSTOM has sponsored work not only related to many aspects of modern development but also to traditional cultural-ecological matters, for example, with specific relevance to agroforestry, the work on the cultivars of kava (Piper methysticum) in Vanuatu (Lebot and Cabalion 1986).