Steve Reynolds
Senior Officer,Grassland and Pasture Crops Group, AGPC, FAO, Rome

Presented at the sixth meeting of Regional Working Group on Grazing and Feed Resources of Southeast Asia, October 1998, Legaspi, Philippines.

Slide Show


Worldwide there are between 10 and 11 million hectares of coconuts. With the marked fluctuations and long term decline in copra and coconut oil prices the integration of livestock and coconuts is economically increasingly attractive. Traditionally used for weed control in plantations so that coconuts could be located, cattle (and sheep and goats) are increasingly seen as an equal part of the system. Although there are constraints particularly related to the level of shade under closely spaced coconuts, a number of grass and legume species have been identified which have varying degrees of shade tolerance. Where light transmission is > 50% sustainable grazing of pastures is possible.

The paper reviews some of the main production systems and details animal production levels in grazing and cut-and-carry systems. Key areas for future work are:

- the screening of new forage species for shade tolerance and persistence;

- the focus on systems of coconut spacing which emphasize wide inter-row areas for increased forage production under high light conditions;

- the development of coconut multicropping systems where various management options are modelled to maximize returns for the grower; 

- the increased use of by-products and alternative feed resources by smallholder farmers; and

- continued efforts to identify alternative tree legumes to supplement leucaena where infestation of the leucaena psyillid has devastated production and severely affected smallholder cattle feeding systems. 


The substantial potential for animal production from a number of agroforestry systems has been reviewed by Gutteridge and Shelton (1994). The plantation crop system with perhaps the greatest potential for further development is the integration of livestock, especially cattle (but also sheep and goats) with coconuts (Shelton, 1991).

Worldwide there are probably between 10 and 11 million hectares of coconuts, with more than 90% located in the Asia and Pacific region, where the Philippines and Indonesia are the largest producers of copra and coconut oil.

Integration of cattle production with coconut enterprises is based on the premise that cattle are beneficial to the management of coconuts and that the combined income of the two enterprises is greater than that of coconuts alone. In the past, coconut was the main agricultural activity and cattle management was directed towards reducing plantation weeding costs and increasing copra production (largely from a higher recovery of fallen nuts). In recent years the marked fluctuation in copra prices, both monthly and from year to year, and the structural decline in copra prices since 1950, has encouraged farmers to diversify and to find a reliable secondary source of income.

Reynolds (1988) demonstrated the importance of a secondary source of income in a case study in Western Samoa. The local copra price dropped sharply from US$0.30/kg in early 1975 to US$0.09/kg later in 1975. Based on data for liveweight gain and copra production, the contribution of beef to gross farm income increased from 21 to 46% for a farm with cattle on natural pasture, and from 42 to 71% for a farm with cattle on improved pastures. The farm without cattle suffered a reduction in gross farm income of 70%. Clearly cattle provided a stabilizing influence on farm income. Iniguez and Sanchez (1991) estimated the percentage contribution of the cattle component in a coconut system in Bali, Indonesia to be 75%.

Cattle production is one avenue for diversification. It is beneficial to coconut production and is increasingly economically attractive both through consistent price increases and price stability. In the Philippines retail prices for beef nearly tripled between 1985 and 1992. Although increases in actual farmgate prices may have been lower, cattle production compares favourably with other intercropping options. Similarly the demand for meat is increasing in Indonesia and this has lead to considerable price increases.


Any attempt to grow two or more crops together, and particularly to grow one (forages) beneath the shading canopy of another(coconuts), necessitates some understanding of the environmental factors involved and the degree of competition likely. Important factors affecting the growth of forage species under coconuts are the available soil moisture and nutrients, the amount of light and the degree of competition between the forage species and the coconuts. The yield of plantation crops may be positively or negatively affected by the pasture system, depending on the nature of the interference which develops and the net effect on the crop environment. The influence of the plantation tree canopy on the quantity and quality of light reaching the ground surface, on temperature and humidity and soil moisture levels has been reviewed by Wilson and Ludlow (1991).

On the positive side, cattle are important for weed control and this has been the traditional use of cattle in coconut plantations. light transmission in the commonly used tall coconut varieties decreases from >90% in recently planted coconuts to a minimum of around 40% at an age of 5-15 years, and then increases again with time until the coconuts are due for replanting at age 50-60 years. Light transmission obviously varies depending on variety (with dwarf or hybrid varieties intercepting more light than the tall varieties), tree spacing and management. Much of the area of existing coconut plantations is of tall varieties and often more than 30 years old, therefore light levels are high enough to support an understorey vegetation. Unless it is controlled this understorey vegetation competes with the trees for water and nutrients.

Grazing can reduce competition from the understorey vegetation by recycling nutrients "locked up" in the standing biomass. A near doubling of coconut yield was reported by several researchers when previously ungrazed coconuts were grazed. This was probably only partly related to increased nutrient cycling; the main effect of grazing being related to a higher recovery rate of nuts in short grazed vegetation. Negative effects of any understorey vegetation on coconut yield must be expected if rainfall or soil fertility is marginal for coconut growth, although the latter can be ameliorated by sufficient fertilization. Competition for moisture is likely to occur where annual rainfall is below 1750 mm, particularly if rainfall is not evenly distributed.

As far as animal production is concerned the provision of shade and thus lower heat loads on animals is likely to have a positive effect on animal productivity. The nutritive quality of forages grown in partially shaded environments such as old coconuts is comparable to those grown in full sun (Norton et al. 1991). Incompatability of cattle and coconuts is likely to be caused by unacceptable damage to young trees or interference in the management of coconuts. Damage to fronds of young coconuts could be caused by grazing animals and it is usual practice to keep cattle away from young coconuts until fronds are out of reach of the grazing animals. The time required for coconuts to grow beyond the reach of cattle varies, but periods of 3-8 years have been mentioned in the literature. Small ruminants such as sheep have been successfully grazed in 2-year old coconuts (Simonnet, 1990). Damage to stems of coconuts is minimal although there are concerns over possible soil compaction and increased erosion that may occur when the understorey vegetation is overgrazed.


Shelton et al. (1987) stress that the level of shade is THE MOST SIGNIFICANT factor determining the output of pastures grown in coconut plantations. As indicated above the amount of light penetrating the coconut canopy will especially be influenced by the age and spacing of coconut trees. 

The capacity of plants to accumulate soluble carbohydrate reserves is greatly diminished under shade, hence those species with a large reserve of biomass in roots and/or rhizomes and stolons which escape grazing may be more persistant under heavy shade than erect species which maximize leaf production (Wilson, 1991). Though they may have a more conservative growth performance, stoloniferous species such as Axonopus compressus, Brachiaria miliiformis (Brachiaria subquadripara), Paspalum conjugatum and Stenotaphrum secundatum are reported to perform well in grazed pastures under strong shade. 

Some grasses and legumes are more shade tolerant than others (see Table 1). When light transmission values fall below 40 or 50% then both production values and the range of species are severely reduced. In general herbage production (and therefore carrying capacity) is inversely proportional to tree density (and light transmission values). Wong (1991) defined shade tolerance (agronomically) as "the relative growth performance of plants in shade compared to that in full sunlight as influenced by regular defoliation. It embodies the attributes of both dry matter productivity and persistence". The term persistence includes both the survival of individual plants and seedling replacement. 

Table 1 Shade tolerance of some tropical forages (after Wong 1991, and Shelton et al. 1987)
Shade tolerance Grasses Legumes
High Axonopus compressus
Brachiaria miliiformis
Ischaemum aristatum
Ottochloa nodosa
Paspalum conjugatum
Stenotaphrum secundatum
Calopogonium caeruleum
Desmodium heterophyllum
Desmodium ovalifolium
Flemingia congesta
Mimosa pudica
Medium Brachiaria brizantha
Brachiaria decumbens
Brachiaria humidicola
Digitaria setivalva
Imperata cylindrica
Panicum maximum
Pennisetum purpureum
Setaria sphacelata
Urochloa mosambicensis
Arachis pintoi
Calopogonium mucunoides
Centrosema pubescens
Desmodium triflorum
Pueraria phaseoloides
Desmodium intortum
Leucaena leucocephala
Desmodium canum
Neonotonia wightii
Vigna luteola
Low Brachiaria mutica
Cynodon plectostachyus
Digitaria decumbens
Digitaria pentzii
Stylosanthes hamata
Stylosanthes guianensis
Zornia diphylla
Macroptilium atropurpureum

Indigenous species - native vegetation under coconut varies according to the location and intensity of grazing. Unless there is control of the stocking pressure there may be changes in pasture composition over time with undesirable weed species gradually dominating the sward. Using cattle as "sweepers" or "weeders" without additional selective weed control measures may control the weeds in the short term but allow tough unpalatable species to become dominant. The more promising of the native species include: carpet or mat grass (Axonopuscompressus), buffalo couch grass (Stenotaphrum secundatum), Pemba grass (Stenotaphrum dimidiatum), cogon (Imperatacylindrica), T-grass (Paspalum conjugatum), as well as various legumes such as alyce clover (Alysicarpus vaginalis), Desmodiumovalifolium, Desmodiumtriflorum, hetero (Desmodiumheterophyllum) and sensitive plant (Mimosapudica).

Productivity may vary from low to moderate depending on the relative percentage of productive grass, legume species and weeds, particularly bush weeds. For example, in Western Samoa local pastures dominated by Mimosa pudica and hetero were considered to be particularly productive while in the Solomon Islands there was no significant difference in liveweight gains between improved pastures and naturalized pastures with a high legume content and consisting of Axonopuscompressus, Mimosapudica, Centrosemapubescens and Calopogoniummucunoides

Exotic species - where the aim is to do more than merely keep weeds under control, so that fallen nuts can be located, then various exotic grass and legume species are available. Grass species most suited to the reduced light conditions under coconut palms are sod forming stoloniferous grasses that form short to moderate height swards. They provide moderate carrying capacity, allow fallen nuts to be quickly located, are inexpensive and easy to establish from cuttings, compete well with aggressive weed species, maintain a reasonable balance with companion legumes under grazing and do not compete excessively with coconut production. Such grasses include Angleton grass or Alabang X (D. aristatum), Batiki (I. aristatum), Cori (B. miliiformis), Koronivia (B. humidicola), Palisade (B. brizantha) and Signal (B. decumbens). Although Para grass (B. mutica) is popular in the Philippines, elsewhere it has been shown to be not very shade tolerant and requires good management under the high light conditions (light transmission >75%) of old coconut plantations or where trees are widely spaced (9 x 9 or 10m). Buffalo couch (S. secundatum) and Pemba grass (S. dimidiatum) are well adapted to heavy shade conditions in Vanuatu and Zanzibar, respectively.

In establishing pastures the degree of shade will determine which of the recommended grass species is most suitable. Where light transmission is < 30%, dry matter yields of all species are low so that grazing of existing species (such as A. compressus) may be most appropriate. In open plantations (light transmission >75%) the choice of species is wide but B. brizantha, B. decumbens and B. humidicola are particularly recommended. In more shady conditions (light transmission 50-75%) I. aristatum and B. humidicola should be used, while in heavier shade (light transmission 30-50%) I. aristatum may be suitable, but species such as S. dimidiatum and S. secundatum are probably most appropriate. For cut-and-carry P. maximum and P. purpureum are widely used.

The legumes most suited to coconut plantations include centro (C. pubescens) and Siratro (M.atropurpureum), with puero (P. phaseoloides) and sometimes Calopo (C. mucunoides) used as pioneers (and as cover crops). However, in some humid tropical environments Siratro is subject to Rhizoctonia leaf blight. Legumes that combine particularly well with B. brizantha and B. decumbens include hetero (D. heterophyllum), D. triflorum and A. vaginalis. Sensitive plant (M. pudica) should be utilized where it is indigenous, but needs to be carefully controlled. In Zanzibar, T. labialis was found to combine well with Pemba grass. Leucaena (L.leucocephala), or (on acid soils) gliricidia (G. sepium), can be grown as a double-row hedge (rows 1m apart) between every two rows of coconuts. 

In Malaysia for low light levels (< 50% sunlight) the shade tolerant species P. conjugatum, A.compressus, C. pubescens and D. ovalifolium appear to be best. In moderate shade P. maximum and B. decumbens are suitable and under old coconut plantations with high light transmission to these can be added P. purpureum, S. sphacelata cv. Kazungula, MARDI digit (D. setivalva), D. pentzii, B. humidicola, Zorniadiphylla and S. guianensis

In Vanuatu, B. decumbens, sabi grass (Urochloa mosambicensis) and B. humidicola perform well under coconuts provided at least 70% light reaches the ground - an average situation for a good stand of 60 year old coconuts. Given very careful management involving undergrazing these grasses will probably persist down to 50% sunlight conditions. For replanted coconuts and particularly the hybrids at the recommended spacing of 9m triangular, light conditions will be below 50% from 5-40 years of age. Under such conditions buffalo grass (S. secundatum) is the best available option at present, which combines high shade tolerance and a growth habit which is resistant to overgrazing.

At a workshop in Medan, North Sumatra, Indonesia in September 1990 (Iniguez and Sanchez, 1991) a Working Group, after reviewing past evaluations of germplasm, recommended various forages as a starting point for future evaluations (see Table 2).

Although there have been a number of studies on the shade tolerance of herbaceous legumes less information is available on tree legumes. Leucaena leucocephala has been shown to have limited shade tolerance. In a more recent study of the response of six fodder tree legumes to a range of light intensities (ranging from 100 to 20%) the relative order of shade tolerance was Gliricidia sepium > Calliandra calothyrsus > Leucaena leucocephala > Sesbania grandiflora > Acacia villosa > Albizia chinensis. With the psyllid insect causing serious damage to Leucaena leucocephala in Bali, Indonesia psyllid-resistant tree legumes are required. 

Table 2 Potential candidate forage material for integrated tree cropping and small ruminant production systems (after Iniguez and Sanchez, 1991)
Crop and Age
Young rubber/oil palm
old coconut
3-6 yr rubber/oil palm young coconut Mature rubber/
oil palm
Light Transmission (%)
100-70 60-30 30-10
Brachiaria decumbens
Brachiaria humidicola
Brachiaria mutica
Digitaria setivalva (MARDI digit)
Pueraria phaseoloides
Centrosema pubescens
Stylosanthes guianensis
Arachis sp.
Desmodium ovalifolium
Paspalum notatum
Paspalum wettsteinii
Axonopus compressus
Arachis sp.
Stenotaphrum secundatum

A trial under coconuts (58% light transmission) to identify suitably adapted species concluded that Calliandracalothyrsus, Codariocalyx gyroides, Desmodium rensonii and Gliricidia sepium warranted further study as forage species for use in the coconut plantations in Bali. A similar study was carried out in North Sulawesi where Gliricidia sepium and Erythrina sp. are commonly used as fences and live stakes under coconuts. Calliandra sp. CPI 108458 produced by far the highest leaf yields and other potentially useful species included Flemingia macrophylla, Calliandra calothyrsus (local), Gliricidia sepium (local), Desmodium rensonii and Codariocalyx gyroides. In the drier environment of South Sulawesi, there was little difference between the leaf yield of C. calothyrsus, L. leucocephala and G. sepium. This was before the effect of the psyllid on leucaena.


Cattle have been used traditionally as "sweepers" for brush/weed control to assist in the collection of coconuts on larger coconut plantations. However, over the last few decades, with rising demand for animal protein (and rising prices for meat) and falling copra prices, commercial interest in improving ruminant productivity has increased in both Asia and the Pacific.

In Asia, smallholder farmers often have one or two cattle which are grazed on whatever feed resources are available in their area. This varies considerably, depending on the available resources and farming system. In many situations cattle are grazed on fallow cropping areas before and after rice or other food crops, and are shifted to plantation areas during the cropping period when there is little available land for cattle. Also smallholders have to maximize use of their limited land resources, and coconuts are usually intercropped with food and other perennial crops such as banana, cloves, pepper and vanilla, particularly in areas with high population density. Despite this intensive land use there are often small areas under coconuts available for grazing or the growing of forage crops. Cattle are generally tethered in such intensive farming systems and shortfalls in feed are overcome by cutting naturally occurring grasses from communal areas such as roadsides. In these circumstances tree legumes can play a significant role in increasing protein content of the feed material, and thereby animal production. The use of tree legumes grown along field boundaries is particularly widely used in Bali.

In the Pacific, a large proportion of cattle are grazed under coconuts. In Fiji, Papua New Guinea (PNG), Western Samoa and Vanuatu cattle have been used traditionally to control weeds and thus reduce upkeep costs, and to provide an additional income from extra copra and meat. In PNG a 70% reduction in upkeep costs has been mentioned and substantially reduced labour costs on plantations in the Solomon Islands have also been indicated.

Both cattle breeding and fattening operations are feasible under older coconuts and these may be based on grazing of pastures or cut-and-carry feeding systems.

Cut-and-carry systems extract a considerable amount of nutrients from the forage production area and this is moved to where the animals are fed; particular care is required to return nutrients to the forage area. Neglect to do so may result in loss of coconut yield and cause a sharp decline in forage yield.

Grazing systems are generally found in more extensive coconut production areas such as in North Sulawesi, Indonesia, parts of the Philippines and also in many South Pacific countries. Some tethering is used to control animals but the majority of cattle are herded or animals are allowed to graze freely. A key factor hampering the development of more commercially oriented cattle production systems under coconuts is the lack of marketing facilities in the more remote coconut plantation areas. The importance of market access for the successful development of a viable cattle industry in the South Pacific was clearly demonstrated by Shelton (1991).


Grazing Systems - the level of animal production reported in grazing trials varies greatly (Table 3). Average daily gains (ADG) vary from 0.12 kg/hd/day to 0.51 kg/hd/day and liveweight gains per hectare varied from 44 kg/ha/year to 744 kg/ha/year. Stocking rates (SR) also varied widely from 1 to 4 cattle/ha (varying sizes) and stocking rate was related negatively to ADG.

The variation in animal production was clearly related to the feed resource available. Liveweight gains were lower on natural vegetation than on improved pastures except in the Solomon Islands where the natural pasture consisted of a very high proportion of legumes. In other cases substantial improvement in LWG was obtained by planting improved pasture. The importance of legumes was clearly indicated in many experiments. Other factors affecting forage growth and therefore animal production were soil fertility and/or fertilizer strategy, and light transmission. In general terms, as indicated above, yield of forages is linearly related to the amount of light available, provided that other factors affecting growth are not limiting. Thus in a coconut plantation with 50% light transmission, the yield of a highly productive grass like Panicum maximum

will be approximately 50% of the yield achieved in full sunlight. Animal production is likely to be affected similarly by light transmission. 

Shelton (1991b), analyzing liveweight data for four long-term stocking rate grazing trials under coconuts related animal production to pastures. In three of the four grazing trials persistence of the sown grasses was poor, and the sown grasses tended to be replaced by more grazing-tolerant but less productive grasses such as Axonopus compressus. Legumes such as Centrosema pubescens were initially more persistent than the grasses but eventually also declined while the proportion of weeds such as Mimosa pudica and unpalatable species increased. He concluded that long-term sustainability of improved pastures under coconuts will depend on the use of grasses tolerant to regular grazing and sufficiently aggressive to keep pastures relatively free of weeds.

In Vanuatu, the stoloniferous grass Stenotaphrum secundatum has proven its ability to suppress weeds better than Axonopus compressus or Paspalum conjugatum at the same stocking rate (Macfarlane, 1993) and S. secundatum pastures are able to produce high liveweight gains if grown in association with legumes. On a S. secundatum pasture containing 20% Desmodium heterophyllum and Vigna hosei under coconuts (65-70% light transmission) Macfarlane et al. (1994) reported animal production of steers (300 kg liveweight) - average daily gain was 0.52 kg/hd/day when grazed at 2 steers/ha (380 kg/ha/year) over a 2-year period. This compared with daily gains of 0.30-0.35 kg/hd/day at stocking rates of 1.5-2.0 animals/ha where young steers (200-300 kg liveweight) were grazed on native pasture consisting largely of Axonopus compressus, Mimosa pudica and Desmodium canum.

Cut-and-Carry Systems - small backyard dairy and beef units are common in Bali, Indonesia, Philippines and Thailand, with the grasses Panicum maximum and Pennisetum purpureum being supplemented with leucaena, gliricidia and various by-products. These are widely used in the tropics because of the small size of holdings and the limited grazing area, the fragmentation of land holdings, a lack of fencing in cropping areas and the low cost of labour. These grasses are particularly suitable for plantation crops when the trees are young and vulnerable to damage from grazing animals. Animal production in smallholder cut-and-carry systems is difficult to assess. Rika et al. (1981) compared the growth rates of 12 Bali cattle leased individually to local farmers and Table 3 Cattle production from grazing experiments under coconut

Country Pasture Light transmission 
Liveweight gain 
Stocking rate
Average daily gain 
Solomon Islands
(2900 mm/yr)
Western Samoa
(2900 mm/yr)
(1700 mm/yr)
improved 79 288-505 2.7-6.3 0.22-0.29
(>2000 mm/yr)
(1600 mm/yr)
(>1500 mm/yr)

n.a. - not available.
Source: Adapted from Shelton (1991) and also see Stur et al. (1994)

Table 4 Average daily gain (kg/hd/day) of Bali cattle grazed at various stocking rates (cattle/ha) on improved pasture or fed according to local feeding systems (Rika et al., 1981)
Feeding System Stocking Rate
Liveweight gains
    Period 1
Kg/head/day kg/ha
Period 2
kg/head/day kg/ha
Sown pasture  2.7 0.39a 386a 0.25a 541a
  3.6 0.38a 497b 0.25a 714b
  4.8 0.37a 647c 0.22a 834c
  6.3 0.32b 733c 0.18b 904c
Local feeding
- 0.30 n.a 0.17 n.a

Note: Values in the same column followed by different letters differ at P<0.05. 
n.a. = not available.

fed natural vegetation, banana stem and coconut leaf (a local feeding system) with the growth rates of cattle grazing improved pasture in Bali (Table 4). Average daily gain of cattle in the local feeding system was similar to that at the highest stocking rate in the grazing trial but considerably lower than those obtained at lower stocking rates where animals were able to choose their own diets.

However, a comparison of a cut-and-carry feedlot system, a semi-feedlot system, and free grazing for beef cattle in Johore, Malaysia revealed higher daily gains for stall-fed animals (Sukri and Dahlan, 1986). Trials were carried out with smallholders in West Johore, where coffee was grown as an intercrop under coconuts. Feed rations consisted of coffee by-products (30%), palm kernal cake (37%), urea (2%) and mineral-vitamin premix (1%) and various native forage species (Paspalum, Axonopus, Ottochloa, Ischaemum and Brachiaria) for grazing. The animals under the feedlot system were confined and fed the feed ration ad lib.; the semi-feedlot treatment involved tethering and grazing on the native grasses for 5 hours daily before the animals received the same feed ration ad lib.; the free-grazing animals were tethered to graze the native grasses. Average daily gains of the animals in the feedlot, semi-feedlot and free-grazing systems were 0.48, 0.37 and 0.15 kg (see Table 5) respectively (over period of 178 days). The feedlot and semi-feedlot groups were extended for a further 116 days (trial 2) with average daily gains of 0.60 and 0.38 kg/animal respectively An economic evaluation demonstrated that gross profit was higher for the feedlot animals than the semi-feedlot or grazing groups. It was concluded that feedlot and semi-feedlot systems had great potential for increasing beef production among smallholder farmers and should avoid the major problem of low feed availability (and quality) in dry spells.

Table 5 Economic evaluation of different feeding systems (Sukri and Dahlan, 1986) 
  Feedlot Semi-feedlot  Grazing
Avg feed intake (kg/day)
Cost of ration/kg ($)
Cost of ration/day ($)
Avg daily gain (kg/day)
Revenue from gain ($/day)
Gross Profit
Gross profit ($/day)
Margin over semi-feedlot (%)

In Timor tethered bulls fatten at an excellent rate of over 1 kg/day on an ad lib. diet of leucaena leaves plus a metre of banana stem for moisture each day. The arrival of psyllids has reduced leucaena growth in this system and leucaena has been replaced by other tree legumes such as Sesbania grandiflora, Acacia villosa and Gliricidia sepium. However, in all cut-and-carry systems animal performance depends on the skill and experience of the farmer in ensuring that forages and feeds are provided according to the animal requirements.


Traditionally, cattle have been used as "sweepers" or "brushers", keeping the grass and weeds short, preventing excessive nutrient and moisture competition with the coconut palms and ensuring easy location and collection of fallen nuts. Animals have acted as weeders or biological lawn mowers in the plantations, saving on part of the herbicide cost. The present emphasis in coconut areas is on planting high-yielding hybrids (mainly in large commercial plantations) and/or on coconut based farming systems where complementary enterprises such as livestock are integrated with coconuts to increase productivity per unit area, increase employment opportunities and to provide a buffer against low and fluctuating copra prices. Increasingly, new management techniques have been adopted, improved grasses and legumes have been planted to increase the animal carrying capacity and in smallholder systems increased use is being made of by-products and forage production is being integrated with food crops.

What is likely to happen in the future and can we learn from the experience of livestock integration with other tree crops?

i) For the immediate future the large majority of coconut areas will remain planted at traditional spacings, so there is a continuing need to identify grass and legume species for reduced light situations (and especially < 50% light transmission).

ii) Where high yielding hybrids are planted at even closer spacings than those traditionally used it remains to be seen if intergrazing is feasible and catch cropping prior to canopy closure may be the main intercropping activity. With the positive results from grazing sheep under coconuts in Vanuatu the integration of sheep at low stocking rates may be feasible, with the same need for low light species as in (i).

iii) As long as high prices were obtained for rubber, palm oil and copra and coconut oil then any use of ruminants was as an aid to the management of the key enterprise, the plantation crop. With the fall in prices for rubber, palm oil, copra and coconut oil in recent years there has been more interest in integrating tree crops and livestock and in developing systems where the combined income from the two enterprises is significantly greater than that obtained from the plantation crop alone. It is interesting that in the literature on rubber, oil palm and coconut there is increasing reference to the need for new tree spacings. While trials are underway with wider inter-row spacings, thus for example Tajuddin et al. (1991) suggest that it may be possible to move from the conventional systems of planting rubber at 6 x 3.7 m (450 trees/ha) or 9 x 2.5 m (444 trees/ha) to a hedge-planting system of rubber at 22 x 2 x 3 m (450 trees/ha) in which the light penetration between the hedges is greatly increased and shade tolerant forage species may be of lesser importance, perhaps we can learn from past experience with other tree crops such as work with pastures under Pinus radiata in New Zealand (Knowles, 1991) where twin rows 3.5 m apart and 24.8 m to the next two rows have been investigated, with alder (Alnus accuminata syn. A. jorunillensis) in the dairy region of Costa Rica (Budowski, 1983), with pines in southern USA (Lewis and Pearson, 1987) where the best silvopastoral arrangement in terms of both forage and wood production was a double-row configuration of 1.2 x 2.4 x 12.2 m and black walnut (Juglans nigra L.) in central USA (Garret and Kurtz, 1983) where wider spacings have been used to accommodate the requirements of both or all components of the integrated system. The ILCA Annual Report for 1988 suggests that "in the long term, the economics of spacing tree crops more widely to allow more light to reach the understorey and facilitate production of forage for livestock should be investigated. In the medium term, a search should be made for shade-tolerant grasses and legumes for use under existing tree densities, and ways to integrate them into smallholder farming systems should be investigated."

iv) In many areras seasonality of forage production is a problem. There are large quantities of alternative feed sources which can be used as supplements including banana, cassava, cocoa pod husk, copra cake, oil palm products, rice by-products, sugar cane residues and by-products etc.

Future work is thus likely to focus on four key areas:

A. Screening of new forage species for shade tolerance and persistence

Stur and Shelton (1991) identified two key problems:

- the lack of species available for reduced light situations;

- the need for a greater range of grasses and legumes which will persist and contribute to animal production in low management and input situations.

Recent ACIAR funded work identified the following species for grazing under coconuts:

Grasses - Brachiaria brizantha, Brachiaria decumbens, Brachiaria dictyoneura/humidicola, Ischaemum aristatum, Stenotaphrum secundatum with Paspalum notatum, Paspalum wettsteinii and Paspalum malacophyllum still under study.

Legumes - Arachis glabrata, Arachis pintoi, Arachis repens, Desmodium heterophyllum, Desmodium intortum cv. Greenleaf, Aeschynomene americana cv Glenn and cv lee and possibly the tree legume Calliandra calothyrsus.

B. Systems of coconut tree spacing with emphasis on wide inter-row areas

Under what tree spacing should pastures be established? Presently, the coconut grower wishing to maximize pasture growth as well as yields of copra should use the least number of palms per hectare required to realize this goal. For tall varieties recommended spacings range from 7 x 7 m to 9.1 x 9.1 m and palm densities from 120 to 200 trees/ha. Where pastures are established under palms at closer spacings and higher densities then light transmission values and pasture/animal production will be lower!

However, earlier reports (Plucknett, 1979) had described "group", "bouquet" and "hedge" planting systems to allow more space for intercrops.

Significantly, a report by Manthriratna and Abeywardena (1979) demonstrated not only the influence of spacing on yield in terms of nuts per palm, but also that the same tree density can be achieved by various planting systems with little effect on the mean coconut yield per palm and therefore per ha.

For intercropping (in this case with pastures) a rectangular system with a wide between-row spacing has many advantages over the square system. Manthriratna and Abeywardene (1979) indicated that the planting of coconuts in avenues or hedges, with rows oriented east to west, is much more suitable for intercropping (Hartley, 1977, concluded that a rectangular spacing of oil palm at 6.4 x 12.8 m resulted in a fruit bunch yield only slightly lower than the normal planting density and allowed elephant grass to be grown in the wide inter-row areas). The choice would appear to be between 9.14 m x 4.57 m (30 feet x 15 feet) and 7.62 m x 5.48 m (25 feet x 18 feet), with the former being favoured in terms of intercropping needs. It is also possible that when the returns from the intercrop are considered a lower planting density of 175-200 palms/ha and 35 or 40 feet x 15 feet might be the preferred spacing.

What other evidence is there that traditional spacings designed for monocrop coconuts are being modified to give more emphasis to intercrops?

In Tanzania, the coconut extension service has in recent years recommended a move away from the traditional 9 or 10 m spacing to a 9 x 15 m spacing because surveys of farming households showed that the expansion of coconut planting was not limited by the availability of land, but by shortages of labour. Food crops have priority for the smallholder and in order to get more farmers to plant coconuts the solution was to plant them at wider spacings to benefit food crops, with permanent intercropping benefitting the coconut palms through better maintenance of the intercrop (and therefore the coconuts).

Jayasooriya (1990) suggests that the concept of Coconut Based Farming Systems has initiated a new era in coconut development, with a move away from the traditional monocrop coconut system (and traditional coconut spacings). In Sri Lanka the Coconut Research Institute has undertaken trials with a hedge planting system of 30 x 18 feet (9.14 x 5.48 m) and an avenue system of 10 x 5 m or 200 trees/ha. Liyanage and Dassanayake (1993) suggest that in future replanting or new planting programmes avenue planting system should be adopted (e.g., 10 x 5 m, 12 x 5 m, 15 x 5 m) with wider rows in east-west direction to facilitate light penetration.

In Indonesia on new plantings it is recommended that young coconuts be planted at 5 x 12m, giving a tree population of 160 palms/ha and wide inter-row areas for cultivating other crops.

In Tonga, Opio (1990a) refers to the practice of "hedge planting" (15 x 5 m) which has been introduced in recent years to minimize the problem of poor light penetration so that continuous intercropping can be undertaken at any stage of development of the coconut trees. Hedge planting is based on wider spacing between the rows and closer spacing between the plants. He mentions that while there is no significant difference in plant population per unit area compared to the conventional spacing of 9 x 9 m, yields from hedge planting are generally higher, sometimes by as much as 25%. This is mainly because inter-cultivation encourages adoption of appropriate technology and improved husbandry practices. Hedge planting also allows for continuous inter-cultivation, which minimizes the problem of land shortages, encourages intensive land utilization and increases the productivity of land under coconut.

Shelton (1993) suggests that establishment of pasture with plantation tree crops may not be sustainable unless there is a radical alteration of tree-planting configurations to improve the long-term light environment.

The effect of hedgerow planting systems on species selection

If hedgerow planting systems are adopted so that light transmission conditions in the inter-row areas are in excess of 80 or 90% throughout the life of the coconut trees then the need to identify shade tolerant species will be less of a priority. As such, species which have already been identified and are in widespread use in open areas, can be recommended and selected according to the particular environmental and socio-economic conditions and likely management levels. Although for the immediate future, in the large majority of coconut areas planted at traditional spacings, the need for shade tolerant species remains, many coconut plantations comprise ageing stands of lower-yielding trees which have thinned over the years and which will need replacing (e.g. in the Philippines it is estimated that 55% of the more than 300-400 million coconut palms are older than 40 years and 27% are beyond 50 years). With generally low productivity and profitability decisions will have to be made about whether to diversify and intercrop or replant with pure coconut stands. In Vanuatu, according to Evans et al. (1992), increasingly under conditions of depressed copra prices some commercial smallholders are showing interest in planting coconuts at lower densities so that the more productive improved pastures can persist under grazing.

If new coconut layouts are adopted this will have implications not only for pasture species but also for stocking rates, grazing systems and the management and economics of the whole integrated system.

C. Development of (coconut) multicropping systems where various management options are modelled to maximize returns for the grower

Opio (1990b) described an intercropping/rotational grazing model which on the basis of cashflow analysis gave much better returns in terms of gross revenues and returns to labour than monocrop coconuts. While Local Tall coconuts and cattle had a better return to labour figure than Local Tall alone, the returns to labour from rotational grazing and intercropping with taro and kava were higher than the minimum wage of WS$ 0.60/hour and this system would provide a reasonable alternative use of land during what traditionally would have been a fallow period with a significant increase in income. 

Comparative results from Fiji indicate not only that returns from monocrop coconuts can be increased substantially by intercropping but that intercropping/ rotational grazing of Local Talls (LLT) with taro provided the highest net accumulative return under a coconut based farming system (Opio, 1993).

This approach is not new as shown by the work of Garrett and Kurtz (1983) on black walnut (Juglans nigra L.) in USA, which lends itself especially well to multicropping due to the high value of its wood and the production of a nut crop with intercrops providing early financial returns which can serve to offset establishment costs and provide a sustained income throughout the rotation period. With black walnut planted at 12.2 x 12.2 m Garrett and Kurtz (1983) compared four multicropping management regimes with nine different management options and illustrated the flow of costs and revenues for a specific management option. 

This is an area which urgently requires attention and models with a range of management options should be developed and tested to build on coconut based farming systems work already underway and incorporate spacing changes as mentioned above.

Although a considerable amount of information is available on the coconut system, the long-term systematic investigation of the pasture-livestock-coconut system is lacking in terms of the studies undertaken for example by the Forest Research Institute at Rotorua in New Zealand for Pinus radiata and livestock/pasture integration or at MARDI in Malaysia for rubber and oil palm. The information available on coconut systems tends to come from a series of unconnected studies by individual research workers; perhaps there is need for a university or research institute in South East Asia or the South Pacific to focus on integrated crop-livestock-coconut systems and to carry out a long-term programme of research along the lines of that undertaken for crop-livestock-tree crop systems in other areas.

D. Use of alternative feed resources

While smallholder farmers traditionally make use of a range of alternative feed resources there is considerable scope for expanding their use. An example of the use of rice straw and rice straw plus supplements in Sri Lanka illustrates this.

Results after one year are shown in Table 6. Straw intake increased over time. Mean annual straw dry matter intake was 574 g/head/day on GS treatment compared to 1,171 g/head/day on GSS treatment with annual mean intakes of urea, molasses, rice polishings and mineral mixture at 29, 146, 125 and 15 g/head/day respectively. Supplementation not only increased the intake of straw but decreased the grazing pressure and resulted in heifers reaching a breedable stage much earlier. It is in the reproductive phase through early heat, early conception, earlier heavier and healthy calves, increased fertility, more calves per lifespan and more milk per lactation that the benefits begin to show. 

Table 6. - Liveweight and liveweight gain data for heifers grazing natural herbage under coconuts1 and those fed rice straw and rice straw plus supplements (after Pathirana et al., 1992)
Parameter G GS GSS2
LW (kg/head): Initial
LW gain: kg/head/yr

1 30 year old coconut plantation with palms spaced at 8.4 x 8.4 m (137 palms/ha);

2 G = heifers grazed (at 6/ha) continuously on natural herbage.

GS = as per G + unprocessed, unsupplemented rice straw ad. lib.

GSS = as per GS plus supplements of urea, molasses, rice polishings and a vitamin mineral mixture

abc values within each row bearing different letters are significantly different (P = 0.01).

Annual rainfall was 1,750-2,000 mm.

Liyanage and Pathirana (1992) noted that cattle grazing only natural herbage produced 0.8 l milk/head/day compared with 3.0 l milk/head/day produced by those grazing natural herbage and fed supplemented rice straw. 

Pathirana et al. (1993) reported on experimental results after five years.

Noting that attempts to introduce improved, well managed pastures and cattle (high input systems) under coconuts had made little progress, it was stressed that this experiment was an attempt to study the long-term effects of low inputs, simulating field conditions, involving indigenous cattle, natural herbage and straw as it is; the resources commonly available to small dairy farmers, as a first step towards increased productivity. Reproduction and lactation data are presented in Table 7. What is of considerable economic consequence for the farmer is the amount of milk collected up to a given point in time. Thus the effect of nutrition on the overall total number of days in milk and the total milk yield in each treatment group up to the end of May 1993 is impressive (total milk yield 803 compared to 7,164 kg!).

E. Alternative Tree Legumes

In many areas of S.E. Asia the leucaena psyllid (Heteropsyllid cubana Crawford) has caused widespread damage to existing leucaena stands. Particularly in the initial stages the devastation was so severe that smallholder cattle production in particular was severely affected as leucaena was a major component in the feeding system. Moog (1992) noted in the Philippines that the infestation resulted in stunted growth of leucaena, death of plants and feed shortages so that farmers resorted to other plant materials such as banana leaves and stems, corn stover, coconut fronds etc.

Moog (1991) suggested that there has been over-reliance on leucaena and that the potential of other fodder tree species such as the Sesbanias, the Erythrinas, and the Gliricidias should be tapped. Work is also underway on psyllid resistant species and the Leucnet activities have already demonstrated a number of promising species. Leucaena KX2 from Hawaii (a cross between L. pallida 728 and L. leucocephala 636 has demonstrated both resistance to the psyllid and excellent productivity.


Coconut plantations offer an excellent opportunity for the integration of cattle and a tree crop, particularly in the less populated areas where the land under coconuts is not fully utilized and is weed covered.

Given the appropriate tree spacing there are few major constraints and provided that adapted forages are planted to ensure a high quality sustainable feed resource, cattle production under coconuts can be a profitable and sustainable form of land use.

Unfortunately in many areas tree spacing is such that reduced light availability restricts the range of forage species and their productivity. Also, there has been little work on developing 

Table 7. - Lactation data of animals without straw (G), with straw (GS), with straw and supplements (GSS)1
No. days in milk Milk yield/lactation (kg/head)
  1st lactation 2nd lactation 1st lactation 2nd lactation
  Mean Range Mean Range Mean Range Mean Range
  Average milk yield (kg/head/day) Overall total lactation data+
  1st lactation 2nd lactation Total no. days in milk Total milk yield (kg) Overall avg milk yield 
  Mean Range Mean  Range      

1 Not analyzed statistically due to insufficient data, particularly from G group. Animals were milked as long as possible.

* Data from only two completed lactations; two more lactations continuing on the date of computation (31 May 1993).

** Data from only 5 completed lactations; 4 more lactations continuing.

*** Data from all 9 completed first lactations. data from only two completed second lactations included; five more second lactations continuing.

+ Data from all completed as well as from all continuing lactations for all animals in each treatment group.

farming systems which allow farmers to choose from various management options. While research work is ongoing to identify alternative feed sources there is need to develop and apply low input systems in many coconut areas where poor farmers are faced with feed shortages especially in the dry season.

Areas where future work needs to be focused include: 

i) The identification of forage species better adapted to the low light environment of coconut plantations (<50%) which are capable of persisting under heavy grazing pressure.

ii) The adoption of coconut planting (rectangular) configurations with wide between-row spacing which allow for maximum light penetration, encourage cultivation, improve forage yields and to which to a large extent forage species already available would be well adapted.

iii) More detailed and systematic studies of the pasture-livestock-crop-coconut system and to develop management options for the farmer.

iv) Better utilization of existing by-products and alternative feed resources for livestock in the smallholder coconut based farming systems.

v) Continued efforts to identify alternative tree legumes to supplement leucaena where infestation of the leucaena psyllid has devastated production and severely affected smallholder cattle feeding systems.


More detailed information on the pasture-cattle-coconut system can be found in the following:

Reynolds, S.G.(1995) Pasture-Cattle-Coconut Systems. FAO RAPA Publication 1995/7, 668p.

- Stur, W.W., Reynolds, S.G. and Macfarlane, D.C. (1994) Cattle Production under Coconuts. In Copland, J.W., Djajanegra, A. and Sabrani, M. (ed) Agroforestry and Animal Production for Human Welfare. ACIAR Proceedings No. 55, 106-114.

- Reynolds, S.G. (1993) Pastures and livestock under tree crops: present situation and possible future developments. In Evans, T.R., Macfarlane, D.C. and Mullen, B.F. (ed) Sustainable beef production from smallholder and plantation farming systems in the South Pacific: proceedings of a workshop. AIDAB, 77-117.


Budowski, G. (1983). An attempt to quantify some current agroforestry practices in Costa Rica. In: Plant Research and Agroforestry (ed. Huxley, P.A.). Proc. of Workshop, Nairobi 8-15 April 1981, ICRAF, 43-62.

Evans, T.R., MacFarlane, D. and Mullen, B. (1992). Sustainable Commercial Beef Production in Vanuatu. Vanuatu Pasture Improvement Project Technical Bulletin No. 4 Dept. of Agriculture, Livestock and Horticulture, Port Vila, Vanuatu. 68p.

Garrett, H.E. and Kurtz, W.B. (1983). Silvicultural and economic relationships of integrated forestry-farming with black walnut. Agroforestry Systems 1, 245-256.

Gutteridge, R.C. and Shelton, H.M. (1994(. Animal Production Potential of Agroforestry Systems. In Copland, J.W., Djajanegra, A. and Sabrani, M. (ed) Agroforestry and Animal Production for Human Welfare. ACIAR Proceedings No. 55, 7-16.

Hartley, C.W.S. (1977). The oil palm (Elaeis guineensis Jacq.). Longmans, London.

Iniquez, LC. and Sanchez, M.D. (1991). Integrated Tree Cropping and Small Ruminant Production Systems. Proc. Workshop on Research Methodologies, Medan, North Sumatra, Indonesia, September, 9-14, 1990, 329p., USAID.

Jayasooriya, B. (1990). Extension strategy and methodologies for transfer of technology for populaziration of coconut based farming systems. In: Coconut Based Farming Systems (ed. Silva, de S.), Proc. of the XXVII COCOTECH Meeting 25-29 June, 1990, Manila, Philippines, 420-447.

Knowles, R.L. (1991). New Zealand experience with silvopastoral systems: A Review. In: Proc. of an International Conference on Agroforestry - Principles and Practice. J. of Forest Ecology and Management. 45(1-4), 251-267.

Lewis, C.E. and Pearson, H.A. (1987). Agroforestry using tame pastures under planted pine. Chapter 13. In: Agroforestry: Realities, Possibilities and Potentials (ed. Gholz, H.L.). Martinus Nijhoff Publishers.

Liyanage, M. de S. and Dassanayake, K.B. (1993). Experiences in coconut based farming systems in Sri Lanka. In: Advances in Coconut Research and Development (eds. Nair, M.K., et al.). Proc. of Intern. Sympos., Kasaragod 26-29 Nov. 1991, 357-367. Indian Soc. for Plantation Crops.

Liyanage, M. de S. and Pathirana, K.K. (1992). Productivity and sustainability of the coconut/ cattle integrated system in the Wet Zone of Sri Lanka. Paper presented at the 2nd Asian Farming Systems Symposium, Colombo, Sri Lanka 2-5 November 1992.

MacFarlane, D.C. (1993). Vanuatu Pasture Improvement Project: structure, programmes and achievements. In: Sustainable beef production from smallholder and plantation farming systems in the South Pacific. Proc. of a Workshop (ed. Evans, T.R. et al.), Port Vila and Luganville, Vanuatu, 2-12 August 1993, 126-142.

MacFarlane, D.C., Mullen, B.F., Evans, T.R., McDonald, C.K., and Eberhard, R. (1994). Vanuatu Pasture Improvement Project - Final Technical Report. Port Vila, Vanuatu, AIDAB.

Manthriratna, M.A. and Abeywardena, V. (1979). Planting densities and planting systems for coconut, Cocosnucifera L. 2. Study of yield characters and the economics of planting at different densities. Ceylon Cocon. Q. 30, 107-115.

Moog, F.A. (1991) . Forage and legumes as protein supplements for pasture based systems. In: Feeding Dairy Cows in the Tropics (ed. Speedy, A. and Sansoucy, R.). Proc. of the FAO Expert Consultation held in Bangkok, Thailand 7-11 July 1989. FAO Animal Production and Health Paper 86, 142-148.

Moog, F.A. (1992). Heteropsylla cubana: Impact on feeding systems in South-west Asia and the Pacific. In: Proc. FAO Expert Consultation on Legume Trees and Other Fodder Trees as Protein Sources for Livestock. 14-18 October 1991, MARDI, Kuala Lumpur, Malaysia, 233-243.

Norton, B.W., Wilson, J.R., Shelton, H.M. and Hill, K.D. (1991). The effect of shade on forage quality. In: Proc. of Workshop Forages for Plantation Crops, Shelton, H.M. and Stur, W.W. (ed.). Sanur Beach, Bali, Indonesia 27-29 June 1990. ACIAR Proc. No. 32, 83-88.

Opio, F.A. (1990a). The need for coconut based systems. In: Coconut Based Farming Systems (ed. Silva de, S.), Proc. of the XXVII COCOTECH Meeting, 25-29 June, Manila, Philippines, 1-15.

Opio, F.A. (1990b). Coconut intercropping/rotational grazing model in the South Pacific. In: Coconut Based Farming Systems (ed. Silva, de S.), Proc. of the XXVII COCOTECH Meeting, 25-29 June, 1990, Manila, Philippines, 342-359.

Opio, F.A. (1993). Experiences in coconut based farming/cropping systems in the South Pacific. In: Advances in Coconut Research and Development (eds. Nair, N.K., et al.). Proc. of Intern. sympos., Kasaragod 26-29 Nov. 1991, 369-381. Indian Soc. for Plantation Crops.

Pathirana, K.K., Mangalika, U.L.P. and Gunaratne, S.S.N (1992). Straw based supplementation in a low input feeding system for Zebu heifers. In: Feeding strategies for improving ruminant productivity in areas of fluctuating nutrient supply. Proc. of a final research co-ordination meeting of an FAO/IAEA Co-ordinated Research Programme, Vienna, 30 March-3 April 1992. FAO/IAEA. IAEA TECDOC-691. 23-31.

Plucknett, D.L. (1979). Managing pastures and cattle under coconuts. Westview Trop. Agric. Series no. 2, Westview Press, Colorado, USA.

Reynolds, S.G. (1988). Pastures and cattle under coconuts. FAO Plant Production and Protection Paper, 91. First edition, 321 pp.

Rika, I.K., Nitis, I.M. and Humphreys, L.R. (1981). Effects of stocking rate on cattle growth, pasture production and coconut yield in Bali. Trop. Grasslands 15 (3), 149-157.

Shelton, H.M. (1991). Productivity of cattle under coconuts. In: Proc. of workshop Forages for Plantation Crops, Shelton, H.M. and Stur, W.W. (ed.). Sanur Beach, Bali, Indonesia 27-29 June 1990. ACIAR Proc. No. 32, 92-96.

Shelton, H.M. (1993). Choosing pasture species for shaded environments. In: Sustainable beef production from smallholder and plantation farming systems in the South Pacific. Proc. of a Workshop (ed. Evans, T.R. et al.), Port Vila and Luganville, Vanuatu, 2-12 August 1993, 68-76.

Shelton, H.M., Humphreys, L.R. and Batello, C. (1987a). Pastures in the plantations of Asia and the Pacific: performance and prospect. Tropical Grasslands 21(4), 159-168.

Simonnet, P. (1990). Sheep flock management in a tropical environment under coconut. Oléagineux, 45(10), 451-455.

Stur, W.W. and Shelton, H.M. (1991). Compatibility of forages and livestock with plantation crops. In: Forages for Plantation Crops (ed. Shelton, H.M. and Stur, W.W.). Proceedings of a workshop, Sanur Beach, Bali, Indonesia, 27-29 June, 1990. ACIAR Proceedings No. 32, 112-116.

Sukri, M.I. and Dahlan, I. (1986). Feedlot and semi-feedlot systems for beef cattle fattening among smallholders. In: Proc. 8th Ann. Conf. MSAP (ed. Hutagalung, R.I. et al.) held at Genting Highlands 13-14 March 1984. 74-78.

Tajuddin, I., Ng, K.F. and Chong, D.T. (1991). The potential and prospects for improving forages under rubber in Malaysia. In: Forages for Plantation Crops. Proceedings of a Workshop (ed. Shelton, H.M. and Stur, W.W.), Sanur Beach, Bali, Indonesia 27-29 June 1990. 130-133.

Wilson, J.R. (1991). Ecophysiological constraints to production and nutritive quality of pastures under tree crops. In: Proceedings of the International Livestock-Tree Cropping Workshop (ed. Z.A. Tajuddin). 5-9 December, 1988, Serdang, Malaysia, 39-54.

Wilson, J.R. and Ludlow, M.M. (1991). The environment and potential growth of herbage under plantations. In: Forages for Plantation Crops (ed. Shelton, H.M. and Stur, W.W.) ACIAR Proceedings No. 32, 10-24.

Wong, C.C. (1991). Shade tolerance of tropical forages: a review. In: Proceedings of workshop Forages for Plantation Crops, Shelton, H.M. and Stur, W.W. (ed.). Sanur Beach, Bali, Indonesia 27-29 June 1990. ACIAR Proceedings No. 32, 64-69.