Croton megalocarpus, the poultry-feed tree: how local knowledge could help to feed the world
The high cost of protein sources for poultry has led researchers to direct increased attention to non-conventional feeds. Although most plant leaves are good sources of protein, their use by non-ruminants is limited by the high fibre content and, in some cases, the presence of toxic factors or metabolic inhibitors.
Laboratory analysis and results of experiments involving animals show that ground seeds of Croton megalocarpus, a species indigenous to eastern African montane forests, could be a major ingredient of commercial poultry feeds. Trees of this species already form an important component of agroforestry systems in the region, providing timber and charcoal. The tree provides good shade and has been used as a boundary marker. Its potential as a feed for poultry has been recognized by only a very few local people. The typical local management system for poultry and cultural beliefs could have attributed to this.
First observations on C. megalocarpus revealed that the scarce literature on this species was incorrect in many details. It was found that trees generally start flowering during the fourth year and potential seed yields of mature trees are assessed at 25 kg per year. Enormous variance has recently been observed in different components of seed production (e.g., frequency of flowering, number of flower spikes, number of female flowers per spike, number of seeds per fruit, weight of seeds), and this indicates that, through selection and breeding, yields could be considerably increased.
As the world population grows, increasing the food supply becomes an ever-more-urgent priority. In the year 2040 the world community will have to produce 500% more food than it is doing at the moment. According to the participants (Consultative Group on International Agricultural Research and national research organizations) of a conference, `Ecoregional approaches for sustainable landuse and food production', there should be enough carrying capacity to fulfil the demand for more food. One requirement identified during this meeting was the need for good analytical research of regional production systems. Increased integrated efforts and attention on the local farming systems should result in improved concepts of food production. Indigenous technical knowledge of animals and plants not conventionally regarded as domesticated is expected to be a major component of the successful accomplishment of this difficult task.
Agricultural research has boosted crop and animal yields worldwide, yet the skewed distribution of scientific and technological capabilities robs the poorest areas of the world most in need of scientific and technical knowhow, of the ability to develop their agriculture to the fullest. Experience has shown that no single body of knowledge alone is sufficient to provide the will and direction to overcome hunger and malnutrition. No firm formulas developed in the industrialized nations can satisfy the ecological, cultural and social diversities that characterize the world's regions. On the other hand, new combinations of cross-disciplinary and multisectoral knowledge have the potential to design strategies for feeding and providing nourishment to the peoples of the world. This is the concept behind the new efforts to promote the domestication of indigenous tree crops (Leakey & Newton 1994).
The ongoing and future work proposed here for the development of new poultry feeds attempts to contribute to this challenge. It shows that through a participatory approach of farming systems development, localized technical knowledge has enormous potential to provide food in a much larger part of the world.
In most areas of the world, there are fewer religious or social taboos associated with poultry than there are with pigs and cattle. Thus poultry products provide an acceptable form of animal protein for most people, with the exception of strict vegetarian vegans.
Table 1. World population (millions) of farm animals, birds and humans in selected regions (1987)
|Cattle||Sheep & goats||Pigs||Chickens||Humans|
More than 90% of the world's poultry flocks consist of chickens (FAO 1982, cited in Smith 1993). The actual number of poultry in the world is very large (table 1) and the number far exceeds in quantity (although not of course in total weight) the combined total of sheep, goats, pigs and cattle. The contribution that poultry makes to the supply of animal protein varies from area to area.
According to an FAO survey in 1970 of 90 countries, 69% of the population had available an average of less than 30 g per day of animal protein, and 19% had 10 g or less. Even in those countries that apparently have more than enough animal protein to meet everyone's needs, these figures almost certainly give an unrealistic picture of the actual situation. It is irrefutable that in the poorest countries the consumption of animal protein by the richer people is considerably greater than that of the poorest, who eat few, if any, animal products. Furthermore, lactating women and young children, who in East Africa form nearly 60% of the population (Balldin et al. 1993) need a greater proportion of animal protein than adult men who, in fact, eat the most.
Inequality of animal protein distribution among the people of poor countries could radically be changed through supply of poultry products. This is technically possible because poultry are able to adapt to most areas of the world; their cost is low, their generation interval short, and rate of productivity high. A poultry enterprise can produce meat within 8 to 10 weeks and eggs within 4 months.
Over the last 40 to 50 years rapid changes have occurred within the poultry industry, especially in developed countries. Selection and cross-breeding techniques have enabled the production of a laying fowl that will provide up to 280 eggs a year and a meat-type chicken that will weigh over 2 kg at 7 weeks of age, having been fed on only 4 kg of a balanced diet.
Before 1930, the method of keeping poultry in the developed countries was similar to the way free-range, scavenger poultry flocks are kept in developing countries today. Therefore the question arises, can the developments in intensive poultry production be profitably adopted by developing countries?
Many governments in developing countries have encouraged the development of large-scale modern poultry enterprises. Beyond the capital development, the constraints are those associated with their operation. For example, a major problem arises when scarce foreign exchange has to be used to import poultry food, and to a lesser extent, vaccines and drugs. To produce 1 kg of broiler carcass requires 2-3 kg of a balanced feed, while a laying bird requires up to 50 kg of a similar diet over a year.
The foodstuffs used for poultry feeds are often of a quality that could be fed directly to hungry humans. Intensive poultry production is thus appropriate when a country has surplus vegetable foodstuffs or an export-based economy (Smith 1993). Consequently, in the developing world, governments would be better advised to pursue a policy encouraging free-range poultry production. Although this system of production appears primitive, it can be very efficient because the inputs are very low. With good management, free-range hybrid poultry under subsistence farming conditions in Zimbabwe have produced 170 eggs per bird per year when the diet of the birds was supplemented with maize alone (Huchzermeyer 1973). The question we must therefore ask now is: are there other sources of food suitable for poultry that would not normally be used for human consumption?
Because poultry are monogastric, they are unable to manufacture essential amino acids or the B vitamins, and they cannot exist on high-fibre diets. The diet of birds must contain the materials essential for the processes of maintenance, production and reproduction. The essential nutrients can be grouped as water, carbohydrate, fats and oils, protein (amino acids), vitamins and minerals.
The energy in the diet, necessary for production (meat, eggs) and for the maintenance of vital functions and body temperature, is largely in the form of carbohydrates; fat and amino acids are also required. The synthesis of protein in the body tissues requires an adequate supply of about 20 different amino acids, 10 of which cannot be synthesized by the bird and must therefore be provided in the diet.
Foodstuffs commonly used in poultry diets in the tropics can be classified into five broad classes:
· cereals and cereal by-products (maize, sorghum, millets, wheat, rice, maize bran, wheat bran, rice bran)
· other energy foods (cassava, sugar, molasses)
· animal proteins (fish meal, meat and bone meal, feather meal, blood meal)
· plant proteins (soya bean, groundnut, cottonseed, sunflower seed, linseed)
· mineral supplements
Foods containing high levels of protein (e.g., soya beans, groundnuts, sunflower seeds) are expensive to purchase, while energy foods (e.g., cereal by-products) are often plentiful and relatively cheap.
The most commonly used animal protein for poultry diets is fish meal. It is a high-quality protein food rich in all the essential amino acids, but fish meal is in limited supply in most developing countries.
The high and increasing prices for animal feed have compelled researchers in developing countries to direct their attention to non-conventional feeds, with particular emphasis on protein substitutes. Plant leaves are commonly processed into leaf meals for non-ruminant animals. Among the leaf meals, leucaena (Leucaena leucocephala) and cassava (Manihot esculenta) leaf meals are most popular (Limcango-Lopez 1990). Other species reported on are Trema orientalis, Morus indica, Moringa oleifera, and Sesbania rostrata. In China, pine needles are one of the main leaf fodders. Theleaf meal is produced industrially and used widely in animal feed, especially for pigs and poultry, mainly to supplement vitamins and trace minerals (Zaichun 1990).
The use of leaf meal as feed is limited by its high fibre content and, in some cases, the presence of toxic factors or metabolic inhibitors. Consequently, levels higher than about 5-10% have detrimental results on survival and production (Yoshida 1944, Springhall & Ross 1965, Labadan et al. 1969, Ross & Enriquez 1969, Lopez et al. 1978, Sazon 1988, Cariaso 1988, Ash et al. 1992).
In Kenya, however, a real breakthrough has been made in identifying a local poultry feed supplement. The Kenya Woodfuel and Agroforestry Programme (KWAP), implemented by ETC Kenya Consultants, has in close collaboration with farmers, identified some indigenous knowledge on a species that can be named the `poultry feed tree'.
The seeds of the poultry feed tree, Croton megalocarpus Hutch. (Euphorbiaceae), are reportedly eaten by birds and squirrels (Noad & Birnie 1989) and can be used for poultry (Nicholson 1992). These seeds contain oil and protein, believed to be 30 and 50% respectively (Teel 1994). Analysis at the EATEC laboratories in Eldoret, however, showed an average of 32% oil and 18% protein. Croton oil is also reported to be a powerful cathartic, occasionally used in medicine (anon. 1958).
C. megalocarpus is indigenous to East Africa. Its range is the semi-arid and subhumid highlands, at altitudes between 1200 and 2450 m, with an annual rainfall of about 800 to 1600 mm and average annual temperatures varying between 11 and 26oC. Trees of this species are found in forests and often on farms, where they play a mayor role as boundary markers, windbreaks, shade trees and fuelwood producers.
The few reports on the use of croton seeds as poultry feed are restricted to the Nyeri and Kakamega areas of central and western Kenya respectively. In all cases the seeds are fed as seasonally available supplements to the diet of scavenging poultry flocks. Little is known about feeding ratios and productivity of these birds.
Preliminary results of experiments by KWAP with ground croton seeds partially replacing commercial chick and layers mash show that up to 50% of commercial feed in the diet of highly productive hybrid layers can be substituted by croton seed meal, with no adverse effects on production or hatchability of eggs. Additionally, a 10 to 15% saving in food consumption was observed when croton seed meal was included in the diet.
In the case of chicks, feeds were formulated using commercial feeds and croton seed meal. One-day-old layer chicks were fed on a 10% croton seed diet, and the level of croton in the feed was gradually increased to 25% for one-week-old chicks and maintained at that level. Feed intake, feed efficiency, body weight gain and growth rate of the chicks over a 12-week period were very satisfactory (table 2). Birds slaughtered after this experimental period did not show any internal abnormalities.
Table 2. Growth of layer chicks fed with croton seed meal in the diet
|Week||Croton seed in diet (%)||Food conversion ratio (%)||Weekly growth rate (g)|
Agroforestry refers to land-use systems in which trees or shrubs are grown in association with agricultural crops, pastures or livestock and in which there are both ecological and economic interactions between the trees and other components of the system. Generally, economic factors of agroforestry systems are more important to farmers, but possible negative ecological interactions could be offset by getting an important product or service from the tree (Thijssen et al. 1993a).
At the present time, agricultural scientists and policy-makers, researchers, development workers and extension staff in developing countries are giving increased attention to the use of locally available resources, and economic and ecological reasons have led to a reassessment of the technologies used in agriculture. The practice of `low external input and sustainable agriculture' (LEISA) has led to remarkable improvement of agricultural systems (ILEIA 1989). Practices such as the use of unconventional animals and plants in food production receive much more attention than in the past. Experts expect a great impact from these practices, especially in the arid and semi-arid rainfed areas and other areas that so far are considered to have a low production potential, because of their alleged `low resource base'.
The rapidly expanding interest in agroforestry in recent years, witnessed by the myriad of research and development activities, leaves no doubt that agroforestry as an approach to land development is now accepted by most scientists and development specialists. Increased concern at the highest international policy levels about the sustainability of agricultural development, in the light of the apparent rapid depletion of the natural resource base, has brought agroforestry even further into the limelight.
The tree component can play its beneficial role through the production of useful commodities (e.g., fruits, fodder, firewood, wood products) and by providing certain environmental services such as shade, erosion control and windbreak. The main assumptions are that trees can diversify the range of products obtained from a piece of land and can influence the microclimate favourably for crop growth and animal production while exploring better the three-dimensional character of the soil through the tree's extended root system, its vertical growth, and its perennial existence.
Production figures for seeds of the poultry feed tree are not available. Egli & Kalinganire (1988) report that in Rwanda it flowers and produces fruits every two years, while evidence from Kenya suggests that this species produces fruits every year. Potential seed production for a tree is still unknown but it is generally described as `abundant'. Preliminary observations, however, indicate that 25 kg of seeds per tree per year is feasible.
There seems to be large variability in the major components of seed production (frequency of flowering, number of flower spikes, number of female flowers per flower spike (1-5), fruit diameter (1.7 and 3.8 cm), seeds per fruit (1-4), seed weight (0.67 and 1.18 g). This indicates that significant improvements in seed yield could be achieved through selection and breeding of new varieties. The effects of management on seed production are nut known, but different pruning regimes could also benefit seed production, as flower buds are borne at the end of branchlets.
Propagation of this species is by seed which germinate readily within about 10 days. Contrary to reports that seeds lose viability fast because of the high oil content (Teel 1994), germination percentages can be as high as 80% after one year storage in a plastic container. No reports are available on vegetative propagation of C. megalocarpus.
C. megalocarpus (i.e., `croton with the big seeds') is a hardy and fast-growing tree (table 3) that can reach 35 m or more in height. It usually forms a flattish crown and has horizontal layers of branches. Young plants thrive well even in harsh climatic conditions, and it is not browsed by livestock nor attacked by termites.
Table 3. Growth characteristics of Croton megalocarpus in East Africa
|Region||Altitude (m)||Age (years)||Height (m)|
|Rwanda, Rwerere area||2300||2||3.1|
|Rwanda, Rangiro area||1900||5||11.7|
|Rwanda, Ruhande area||1750||43||20.2|
|Kenya, Maseno area||1500||1||1.7|
|Burundi, Karuzi area||1620||0.5||0.7|
The poultry feed tree is characterized by its multipurpose functions on the farm. Because seeds are available and propagation simple, it is often planted to demarcate boundaries where it plays roles as a live fence, a windbreak, and reduces soil and water erosion. Younger trees coppice well after pruning, so this species is often used as a hedge around home compounds or fields. In the Kenyan coffee-based land-use system of Embu District, for instance, more than 10% of the farms have croton hedges, with an average length of 70 metres (Thijssen et al. 1993b). However, when managed as a hedge this species is unlikely to produce fruits.
Older trees have been reported to be competitive with crops (Noad & Birnie 1989, Thijssen et al. 1993a). The species is therefore often confined to certain niches on the farm such as roadside boundaries, woodlots and paddocks. In grazing areas crotons are valued as shade trees for animals. As a shade tree croton can even be found on home compounds, in markets and at bus stops along the road. Where trees are grown within crop areas the annual extensive leaf fall combined with high levels of nitrogen and phosphorus in the leaves (Niang 1993) give it potential to improve soil fertility and serve as a source of mulch in, for instance, coffee plantations.
Products obtained from C. megalocarpus are timber, building poles and fuelwood, while first-class charcoal can be made out of the wood of this species (Thijssen et al. 1993a). The well- dried nuts are reported to be used in some areas together with charcoal in cooking stoves. Nut shells could be used as mulch in vegetable gardens (one 30-litre bucket can cover an area of 2.25 m2) and as a component of potting mixtures for plants. The croton flowers are bee forage while the seeds, bark and leaves of the tree are reported to have medicinal value (anon. 1958, Kokwaro 1976, Noad & Birnie 1989) including medication for poultry (Charles 1990).
In the Embu area, on the slopes of Mount Kenya, C. megalocarpus has been found on 40% of the farms (Thijssen et al. 1993a) at an average density of about 15 trees per farm (excluding trees managed as a hedge). The majority of farms (84%) had between 1 and 10 trees, while 11% contained between 50 and 100 trees. An estimate of the total number of croton trees on farms in the coffee-based land-use system of this area (ca 400 km2) gives a figure of more than 160 000 trees.
The change in thinking about agricultural technology has also led to a rediscovery and reassessment of indigenous technical knowledge. An increasing number of publications support and document the argument that farmers have a wealth of knowledge about their own environment. They have developed specific skills to use this environment and are very active and creative in adapting the way they use the environment to achieve their objectives. In line with this, the roles and functions of agricultural researchers and extension workers vis à vis work with the farmers is due for reassessment.
Recent literature and meetings on research and extension linkages show the present interest in the development of participatory technology. There are a number of examples where extension and development workers or researchers have successfully worked together with farmers to develop their technology. These examples are increasingly being recognized as pioneers in a powerful approach to development, which needs to be expanded further.
The Kenya Woodfuel and Agroforestry Programme (KWAP) implemented by ETC Kenya Consultants involves participatory technology development for sustainable agriculture. Its first step was building up a relationship of confidence with rural communities and analysing existing land-use systems. Stock was taken through small experiments and literature research, and items were selected for further development. One of these was the poultry feed tree, from which the last three steps of participatory technology development deserve some special attention. The magnitude of the potential that this species has for many areas in the world has implications for the type of future research and development activities to be undertaken.
Croton seed meal has enormous potential on both small and larger scales. In the example given of the coffee-based land-use system of Embu District in Kenya, and at a yield level of only 20 kg of seeds per tree by the estimated 60 000 croton trees, every child in that area could be assured of eating at least 200 eggs per year if the croton seeds could form 50% of the diet of chickens. Alternatively, at the yield of 25 kg of seeds per tree, for every kilogram of seeds produced and fed to broilers, every child in the area could eat one whole chicken every fortnight.
For the production of animal feeds, western countries are importing large quantities of protein material from countries where this material could be used as quality food for the exporting nation. Because of the international demand, however, the price of these materials is too high for common people in the producing countries, and exportation generates higher profits for the producers. If the apparent potential is confirmed, croton seed meal could become an alternative source of protein for poultry feeds all over the world. Since this product is not used for human consumption, it could become a new cash crop for farmers and an export product for tropical countries with appropriate growing conditions. At the same time, introduction of croton seed meal on the international market could lower prices of, for instance, soya bean and fish meal.
Currently, distribution of this species is restricted to some areas in eastern and central Africa. To ensure a rapid development of this croton meal technology, information has to be collected through surveys in these areas. Mature trees should be studied to obtain important phenological information and better describe the ecological niche of this species. At the same time components of production and management regimes should be observed. The presence of croton trees on farms will make it possible to assess the interaction of this tree with crops, livestock and soils in an agroforestry context.
The first step to domesticate this species should examine the qualitative aspects of the seeds for poultry feed. This will involve laboratory analysis and should include a check on antinutritive components, while formal animal experiments should prove the potential of croton seed meal for smallholders. Then, in collaboration with national research organizations dietary details can be worked out to assist in the process of development and adaptation of improved rural poultry husbandry and the use of croton seed meal in particular.
The second step involves the economics and adoption of this new poultry feed. Labour implications and financial returns, especially from a gender perspective, will be important factors for study. Recognizing the importance of good relations with local people and the scale of the task to get communities to collect primary information and to spread this approach over eastern and central Africa, locally active organizations have to work in a participatory way. Some local people have a lifetime of experience with this species and their input will enable a thorough characterization of the poultry feed tree. Potentials for local markets as well as international trade also have to be examined.
The third step in the domestication process, which could go in parallel with step 2, involves the selection of high quality and highly productive trees. As stated earlier, extensive variation exists in the different components of production. Possibilities of vegetative propagation by cuttings or grafting has to be explored to enable multiplication of quality genetic material. The breeding of superior varieties is an alternative option if selection does not produce tangible results, but it would be a slower process and thus probably more expensive. This fundamental type of research should be left to specialized institutions and most probably deserve an ecoregional approach.
The fourth phase in the development of this new crop, that of dissemination, will require careful attention. Experimental results and relevant dietary information have to be made available to the people and organizations concerned in the most appropriate way. This information has to find its way to commercial organizations through the media and formal publications. If the crop is to be adopted by smallholders, the messages must be technically sound, economically feasible, socially desirable and environmentally safe. Continuous interaction with the target group, as spelled out in the participatory technology development concepts, should ensure that the messages to farmers are appropriate and that they lead to a high adoption rate of the new technology.
For countries outside the eastern and central African region, dissemination of quality planting material will be a prerequisite for introduction of the poultry feed tree. Requisition of seeds of different provenances, seed handling and storage, and distribution of the seeds to other regions will be a major task for any project with the mandate to develop the poultry feed tree to its fullest potential.
Finally, the potential of croton seed meal as protein feed is not necessarily restricted to poultry, since monogastric animals are more affected by fibre content, the type of amino acids, and feed inhibitors. The first experimental results with croton seed meal are very promising. Experiments with other animals, for example, pigs, could be carried out first by informal surveys, because of the huge financial implications of formal animal research. Again, collaboration with farmers will be of uttermost importance.
Anon. 1958. A guide to the Nairobi Arboretum. Government Printer, Nairobi. 12 p.
Ash A.J., Petaia L., & Ako H. 1992. Nutritional value of Sesbania grandiflora leaves for monogastrics and ruminants. Tropical Agriculture (Trinidad) 69(3).
Balldin B., Hart R., Huenges R. & Versluys Z. 1993. Child health: a manual for medical assistants and other rural health workers. AMREF, Nairobi. 418 p.
Cariaso E.R. 1988. Feeding value of Mallunggay (Moringa oleifera) leaf meal in broiler diet. College of Agriculture, University of the Philippines, Los Banos. Thesis.
Charles R.A. 1990. Multipurpose trees and shrubs of the Kathama area. Kenya Polytechnic, Nairobi. 84 p.
Egli A. & Kalinganire A. 1988. Les arbres et arbustes agroforestiers au Rwanda. ISAR, Rwanda. 184 p.
Huchzermeyer S.P. 1973. Potential of free range chicken in Zimbabwe. ZARI, Harare. 46 p.
ILEIA. 1989. Participatory technology development in sustainable agriculture. ILEIA, Leusden, The Netherlands. 38 p.
Kokwaro J.O. 1976. Medicinal plants of East Africa. East African Literature Bureau, Nairobi. 384 p.
Labadan M.M., Abilay T.A., Alejar A.S. & Pungtilan U.S. 1969. The effects of feeding high levels of ipil-ipil (Leucaena leucocephala) leaf meal on comb and testes growth of SCWL cockerels. Philippine Agriculturalist 53:65-94.
Leakey R.R.B. & Newton A.C. 1994. Domestication of tropical trees for timber and non-timber products. MAB Digest 17. 94 p.
Limcango-Lopez P.D. 1990. The use of shrubs and tree fodders by nonruminants. p 61-75. In: Shrubs and tree fodders for farm animals, ed. C. Devendra. Proceedings of a workshop in Denpasar, Indonesia, 24-29 July 1989. IDRC-276e. IDRC, Canada.
Lopez P.L., Sayaboc V.S. & Deanon A.S. 1978. The effect of ferrous sulfate on high ipil-ipil (Leucaena leucocephala (Lam) de Wit) leaf meal fed layers. Philippine Agriculturist 62:116-129.
Niang A.I. 1993. Selection generale des espèces. AFRENA annual planning meeting held in Kabale, Uganda. ICRAF, Nairobi. 64 p.
Nicholson M. 1992. Fodder trees for smallholders. Miti Mingi Mashambani, FINNIDA, Nakuru, Kenya. 55 p.
Noad T. & Birnie A. 1989. Trees of Kenya. Nairobi. 308 p.
Ross E. & Enriquez F.Q. 1969. The nutritive value of cassava leaf meal. Poultry Science 48:846-853.
Sazon M.A. 1988. Feeding value of Sesbania rostrata in layer diets. College of Agriculture, University of the Philippines, Los Banos. Thesis.
Smith A.J. 1993. Poultry. Tropical Agriculturalist Series. CTA, Wageningen, The Netherlands. 218 p.
Springhall J.A. & Ross E. 1965. Preliminary studies with poultry rations for the territory of Papua New Guinea. 1. Grower rations with copra, sago and Leucaena leucocephala. Papua Guinea Agricultural Journal 17:117-121.
Teel W. 1984. A pocket dictionary of trees and shrubs in Kenya. KENGO, Nairobi. 151 p.
Thijssen H.J.C., Murithi F.M., Nyaata O.Z., Mwangi J.N., Aiyelaagbe I.O.O. & Mugendi D.N. 1993a. Report on an ethnobotanical survey of woody perennials in the coffee zone of Embu District, Kenya. AFRENA Report no. 62. ICRAF, Nairobi. 32 p.
Thijssen H.J.C., Murithi F.M. & Nyaata O.Z. 1993b. Existing hedges on farms in the coffee based land use system of Embu District, Kenya. AFRENA Report no. 65. ICRAF, Nairobi. 14 p.
Yoshida R.K. 1944. A chemical and physiological study on the nature and properties of the toxic principles in Leucaena glauca (Linn.). Hawaii Academy of Science, Honolulu, USA. 2038 p.
Zaichun X. 1990. Availability and use of shrubs and tree fodders in China. p 295-302. In: Shrubs and tree fodders for farm animals, ed. C. Devendra. Proceedings of a workshop in Denpasar, Indonesia, 24-29 July 1989. IDRC-276e. IDRC, Canada.