An e-publication by the World Agroforestry Centre

AGROFORESTRY A DECADE OF DEVELOPMENT  Printprint Preview

section 3
Prominence and importance of agroforestry in selected regions

Chapter 10
Homegardens: a traditional agroforestry system with a promising future

Otto Soemarwoto
Director
Institute of Ecology, Padjadjaran University
Bandung, Indonesia

Introduction

Homegardens may have originated in prehistoric times when hunters and gatherers deliberately or accidentally dispersed seeds of highly valued fruit trees in the vicinity of their camp sites (Hutterer, 1984). Brownrigg (1985) in his literature review, mentioned that homegardens in the Near Eastern region were documented in paintings, papyrus illustrations and texts dating to the third millenium BC. These ancient gardens may have originated as early as the seventh millenium BC. They were attached to temples, palaces, elite residences and the homes of the common people. The homegarden was mentioned in an old Javanese charter of AD 860 (Terra, 1954). Perhaps the first published report in modern times was that by Raffles in 1818.

From this very brief historical sketch there is evidence that homegardening is a very old tradition which may have evolved over a long time from the practices of the hunters/ gatherers and continued in the ancient civilizations up to modern times.

This chapter is not intended to present a literature review of homegardens, but rather to discuss their salient features, based on the author's experience in Indonesia, as related to their potential and opportunities for future development, and the associated constraints and pitfalls.


The homegarden as an agroforestry system

A common interpretation of homegardens is that it is a system for the production of subsistence crops for the gardener and his family. It may or may not have the additional role of production of cash crops. It can be immediately surrounding the home or slightly further away, but still near the residential area. The Indonesian term pekarangan is derived from the .word karang, meaning a place of residence (Poerwadarminta, 1976) and, hence, pekarangan specifically refers to a garden on the residential site.

The term agroforestry denotes land-use systems consisting of a mixture of perennials and annuals, and often also animals. A major concern in agroforestry research is sustainability (Lundgren and Raintree, 1983). It is determined by the structure of the system, its ecological functions and its continued ability to fulfil the socio-economic needs of the people. Thus, as is implied by the term, homegarden as an agroforestry system should ideally combine the ecological functions of forests with those of providing the socio-economic > needs of the people (Soemarwoto and Soemarwoto, 1984). The ecological functions of forests include hydrologic benefits, microclimatic modification and soil-erosion control, and genetic-resource conservation.


The structure and composition of homegardens

Structure is intimately linked to function. Although this relationship is obvious, too often we ignore it. Not surprisingly, manipulations of structure have often led to the loss of valuable functions, or vice versa, which in turn has resulted in conditions contrary to our expectations or even collapse of the system.

A prominent structural characteristic of the homegarden is the great diversity of species with many life forms varying from those creeping on the ground, such as the sweet potato, to tall trees often metres and more, e.g., the coconut palm, and vines climbing on bamboo poles and trees. These create the forest-like multistorey canopy structure of many homegardens. Well-known examples are those from Java, but they are also found in other parts of Asia such as Malaysia, the Philippines, Thailand, Sri Lanka and India, as well as in other continents. For example, at the First International Workshop on Tropical Homegardens held at the Institute of Ecology in Bandung in December 1985, such systems were reported from the Pacific, from Africa, and from Latin America by various authors. Such land-use systems are also abundantly reported in the literature (e.g., Anderson, 1954; Kimber, 1973; Fernandes et al., 1984). Fernandes and Nair (1986) gave schematic presentations of the structure of different homegardens from various geographical regions and reported that the canopies of most homegardens consisted of two to five layers. Christanty et al. (1980) demonstrated the remarkably close resemblance of the light-interception curve of a homegarden in West Java with that of the Pasoh forest in Malaysia, as measured by Yoda(1974). These measurements provide an objective and precise method for the identification of the canopy layers. The forest-like structure is also derived from the lack of a discernible planting pattern: usually there are no rows, blocks or definite planting distances among components. To the casual observer, homegardens look haphazard.

A popular assumption is that the forest-like structure has been the result of deliberate planning to mimic the forest. This is doubtful, since in Java, where such homegarden structure is very well developed, forests do not have a high cultural value. They are considered dangerous places where wild animals roam and evil spirits dwell. On the other hand babad alas (forest clearing), e.g., for the establishment of a settlement, is considered a noble deed. The term babad alas still remains in the Javanese vocabulary to indicate pioneering activities of a praiseworthy nature, e.g., the foundation of a university. It is therefore inconceivable that people would deliberately create a "forest" in the surroundings of their home. Indeed a person feels offended when his homegarden is said to resemble a forest. It is more likely that the forest-like structure has been an accidental result of people's efforts to fulfil their many kinds of subsistence needs in the absence of a market economy. Karyono (1981) found associations of plants in rural homegardens in which the plants played complementary roles. For example, Table 1 lists the various economic plants with which pineapple is associated in the homegardens in rural areas of Citarum watershed, West Java, Indonesia.


Table 1 Association of pineapple (Ananas comosus) with various other plants in the homegardens in rural areas of Citarum watershed, West Java, Indonesia.

1_Homegardens: a traditional agroforestry system with a promising future

With the development of a market economy the complexity of the homegarden diminishes, and its resemblance to a forest tends to disappear, as described in a later section.

Species diversity and plant density vary from place to place, influenced by ecological and socio-economic factors. Kimber (1973) found 31, 30 and 67 species in three gardens in Martinique. The smallest garden was about 1,500 sq m. In Grenada there were 18 vegetable varieties and 13 distinct types of food trees in a sample garden of less than 2,000 sq m (Brierley, 1985). Thaman (1985) reported from random surveys of homegardens in Port Moresby, Papua New Guinea; Suva, Fiji; Nuku'alva, Tonga; South Tarawa, Kiribati; Nauru Island; and the "Location" contract settlement on Nauru at least 85,114,79,61,33 and 65 different species and distinct varieties of food plants, respectively, in the homegardens in those areas. In addition, a very wide range of non-food plants was found in mixed homegardens which were of considerable importance for handicraft, fuel, medicine, fibre, dyes, ornamental purposes, perfumes and deodorants, livestock feed, and shade and construction materials. A total of 100 species were found in kitchen gardens and 77 species in hut gardens (a hut being a temporary shelter built in the field during the growing season of the field crop) in the Knon Kaen province in north-east Thailand (Kamtuo et al, 1985).

The dynamics of species succession and plant density and composition of homegardens have been intensively studied by Indonesian researchers. In a West Javanese village of 41 households the average number of plant species per homegarden was 56. The total number of species in the village was 219 in the dry season and 272 in the wet season, i.e., an increase of almost 25 percent in the wet season. The highest increase was in the number of vegetable species, which was almost 75 percent in the wet season. The number of spice plants also increased considerably during the wet season, although the number of species remained constant. Another steep increase was in the number of subsidiary staple food plants. Obviously the villagers were taking as much advantage as possible of the rains.

In an extensive survey from the lowlands to the highlands of West Java, Karyono (1981) recorded that the average size of 351 homegardens sampled was 229.1 m2. The size decreased with altitude. The total number of species found in the survey was 501 in the dry season and 560 in the wet season, with a cumulative number of 602 in the two seasons. The average number of species in the dry season was 19.0 per homegarden and 24 in the wet season. Species density was 8/100 m2 in the wet season. The highest number of species per homegarden was in the altitude between 500 and 1,000 m, species density increasing with increasing altitudes.

Both the number of species and the number of plants were highest in the lowest canopy layer, being 62 percent and 35 percent, respectively, of their total numbers and gradually decreasing to 1 percent and 6 percent, respectively, in the highest layer. However, the highest canopy coverage was found in the fourth layer from the ground.

Poor people tend to grow more staple crops, vegetables and fruit trees, whereas well-off people grow more ornamentals and high-economic-value cash crops (Ahmad et al., 1980). Generally, in the more remote areas more subsistence crops are grown and nearer cities the number of cash crops increases. When labour is scarce, people grow more perennials and less annuals since perennials require less labour than annuals (Stoler, 1975). Penny and Singarimbun (1973) observed that when there were not many off-farm jobs, people spent more time on their homegardens and crop diversity increased. A similar situation was found in a widow's garden in Nicoya (Wagner, 1958).

Culture and tradition are the other factors which influence homegarden composition. In a village on the border between West and Central Java, where the traditions of the Sundanese and the Javanese meet, more vegetables were found in the Sundanese homegardens than in the Javanese ones, but more medicinal plants in the Javanese than in the Sundanese (Abdoellah, 1980). It is known that the Sundanese have a strong preference for vegetables in their diet, while the Javanese customarily consume jamu, i.e., traditional herbal medicine for curative and other purposes. Terra (1954) concluded from his extensive studies that intensive homegardening was found in societies with matrilinear traits, e.g., in Central Java, West Sumatra, Aceh, southern Burma, in the Kasi region of Assam in India, in northern Thailand, Kampuchea and in some parts of the mountain ranges between Vietnam and Kampuchea. In the Muslim districts of southern Ethiopia, tobacco, coffee and drugs were intensively cultivated, but they were absent from the homegardens in Christian northern Ethiopia (Simoons, 1965).

Homegardens commonly have animals as components. Brownrigg's literature review (1985) indicated that animals were found in virtually all types of gardens, e.g., poultry, including doves, and fish in the Near-East homegardens, and poultry and livestock in the Luso-Latin and Caribbean homegardens. Other examples of homegardens with animals are the Chagga gardens in Tanzania (Fernandes et al, 1985), the homegardens in Ghana (Asare et al, 1985), Grenada (Brierley, 1985), Indonesia (Soemarwoto et al, 1975; Soemarwoto and Soemarwoto, 1984), Kerala, India (Nair and Sreedharan, 1986) and Bangladesh (Leuschner and Khalique, 1987).

Culture, religion, and economic and ecological factors influence the animal species kept in the homegardens. In Muslim regions, pigs are an absolute taboo, e.g., in West Java and Aceh, but they are very common in the non-Muslim regions, e.g., in Christian North Sumatra and in Bali, Indonesia, where the dominant religion is Hinduism. In West Java with its high rainfall and long wet season, people almost always make a fishpond in their homegardens.


The function of homegardens

The interest in homegardens has primarily been focused on their function of producing subsistence items and generating additional income. They are known for their stable yields, very varied products, continuous or repeated harvests during the year and low inputs. Although light intensities decline in the successive layers because of the stratified structure of the canopies, Christanty et al. (1980) showed that there was no appreciable decrease in photosynthetic rates. Field experiments demonstrated that in these lower layers the yields were still satisfactory (Omta and Fortuin, 1978; Noor, 1981). Obviously the lower-storey plants were selected for their shade tolerance. This shows that the people are capable of effectively utilizing the spatial niches of the homegardens. However, their yields are generally low, although for poor people they can significantly contribute income and/or nutrients to their minimal income and poor diet, particularly during lean periods. For these reasons homegardens are being promoted in many countries, e.g., in Lima, Peru (Ninez, 1985), Ghana (Asare et al, 1985), the Pacific Islands (Falanruw, 1985; Sommers, 1985; Thaman, 1984), Sri Lanka (Jacob and Alles, 1987) and Indonesia.

The Lima gardens, with an average size of 200 sq m, produced an income of US $ 28.33 in five months. Not impressive, but still it added almost 10 percent to the family income during that period (Ninez, 1985). Higher productivity has also been reported. In Indonesia, homegardens can contribute 7-56 percent of the owner's total income (Ochse and Terra, 1937; Ramsay and Wiersum, 1974; Ahmad et al, 1980; Danoesastro, 1980). Homegardens may also produce higher incomes than other land uses. Gonzales-Jacomes (1981) found that in Central Mexico the average income per square metre of homegarden was almost 13 times that of irrigated plots. From our studies it was estimated that in West Java the annual income from homegarden fishponds per unit area was 2-2.5 times that of rice fields, and in tourist areas, where homegardeners sold ornamental plants in pots or plastic bags, the average income could reach almost 20 times that of rice fields. In a study by Stoler (1975) in a village in Central Java, the cash value of production from homegardens was found to be influenced by the size of the homegarden units, the smallest gardens being relatively more productive. She also found that there was an inverse relationship between labour, in terms of hours worked per hectare, and garden size, i.e., there was more labour input in smaller homegardens than in larger ones.

Indonesian data are illustrative of the relationship between productivity and cost. Ochse and Terra (1937) calculated that in Kutowingangun, a village in Central Java, 44 percent of the total calories and 14 percent of the protein consumed came from the homegardens, but only 8 percent of the total costs and 7 percent of the total labour were spent on them. Terra and Satiadiredja (1941) reported that in the same area the gross income as well as net income from homegardens was more than that from dry fields though less than that from rice fields. Similarly, Danoesastro (1980) reported that in three sub-districts in East Java the percentage values of average gross income from homegardens was 39 for rice fields and 72 for dry fields, but the average net income (percentage values) was 84 for the rice fields and 184 for the dry fields. The figures of net income from homegardens varied from 6.6 to 55.7 percent of total income with an average of 21.1, while the cost of production ranged from 4.7 to 28.5 percent of gross income from homegardens (with an average of 15.1). The cost of production from rice fields varied between 47.4 and 66.6 percent, with an average of 55.9 percent of gross income from this form of land use. Thus, the cost of production of homegardens was much lower compared with that of rice fields.

Ahmad et al. (1980) observed that income from homegardens before the rice harvest, the so-called paceklik or lean period, was 25.5 percent of total income, and declined during and immediately after the rice harvest to only 6.4 percent of total income. Terra and Satisdiredja (1941) reported that income from homegardens increased sharply before Idhul Fitri, the major Muslim festival, when people needed more cash. These fluctuations show the versatility of homegardens for meeting varying needs at different times of the year. Because of this, the homegarden is popularly known in Indonesia as lumbung hidup (living store).

It has also been recognized that homegardens are an important source of nutrients. Covich and Nickerson (1966) reported that consumption of 100g each of Manilkara sapotilla, Persea americana, Manihot esculenta and Guilielnuf utilis grown in Choco Indian dwelling clearings urDarien, Panama, was equivalent to approximately 10,12,150, 15.5, 31, 25 and 113 percent of the minimum daily requirement of protein, calcium, carotene, thiamine, riboflavin, niacin and ascorbic acid, respectively. Homegardens in villages in Lawang, East Java, produced a daily average of 398.4 calories, 22.8 g protein, 16.4 g fat, 185 g carbohydrate, 818.4 mg calcium, 555 mg phosphorus, 14 mg iron, 8,362 IU vitamin A, 1,181.2 mg vitamin B and 305 mg vitamin C (Haryadi, 1975). Similar results were reported from the Philippines and Nigeria (Fernandes and Nair, 1986). In Indonesia, the author and co-workers have found that rice fields gave higher yields of protein and calories than homegardens, but more calcium, vitamin A and vitamin C were obtained from homegardens than from rice fields. As mentioned earlier, Ochse and Terra (1937) reported that homegardens could supply up to 18 percent of the calories and 14 percent of the protein consumed in the whole village.

In general, a large proportion of the products of homegardens is consumed by the gardeners themselves. This is especially true in remote areas where the market economy has not yet developed. Danoesastro (1980) reported that in his study area in Central and East Java, products consumed varied from 21 to 85 percent, with an average of 44, while in Stoler's (1975) study, an average of 67 percent was consumed. Stoler noted that the cash value of the produce consumed was, on average, 16 percent of total consumption with a breakdown of 12,17 and 19 percent, respectively, for the smallest, the medium-sized and the largest homegarden groups. Ahmad et al. (1980) reported that the percentage of total produce consumed by the household were as follows: fruit 46, coconut 83.7, vegetables 94.7, medicinal plants 95.5, and tubers and roots 97.3. Other important products for home consumption are fuelwood, construction material and materials for handicraft and home industry.

The socio-cultural functions of homegardens have not received much attention so far. In many areas products for religious rituals and ceremonies are very important, e.g., in Bali and Thailand. Thaman (1985) noted the importance of sacred or fragrant plants in homegardens. He gave a list of 35 plant species considered sacred in Tonga. There is an abundance of examples from all parts of the world of plants and animals considered sacred or supposed to possess magical or mystical powers. They play a very important role in the lives of the people. In areas where the market economy has developed, many of these plants and animals also have economic importance because of the demand for them in daily rituals and particularly during major festivals. In cities, the aesthetic role of homegardens is important.

Studies in villages in West Java have shown that homegardens are an important social-status symbol (Ahmad et al., 1980). People who do not have a homegarden and, hence, have to build their house on someone else's homegarden, are considered of low status. In traditional Indonesian villages, people can freely enter homegardens, e.g., to get water from a well, or just to pass through them. Although there may be fences around them, they are seldom completely closed nor are there locked gates. The concept of trespassing does not exist. Those who close off their gardens completely are considered conceited. Fruits and other products are traditionally shared with relatives and neighbours, and products for religious or traditional ceremonies and medicine are given away freely when requested. However, this equitable social situation is now gradually changing (see a later section in this chapter).

A homegarden is also an important place for children to play and for adults to congregate in their free time. For this purpose a small part of the homegarden is kept clean (not planted), e.g., the so-called baruan in West Java or pelataran in Central Java. It is shaded by some trees planted at the edges. The owner of the homegarden is responsible for the safety of all children who play there. Young people who get married may build their homes on their parents' homegarden. Thus homegardens play an important role in family and community life. They are more than just a production system.


Soil erosion under homegardens

The ecological functions of homegardens have generally been taken for granted and only mentioned in passing. The protective effect of the homegarden on soil erosion has been deduced from its multistorey structure. However, the relationship between canopy structure and soil erosion is complex.

Soil erosion consists of two processes, i.e., splash and surface erosion. Splash erosion is the detachment of soil particles from the soil surface by the kinetic energy of falling water drops, which in homegardens are drops of throughfall, i.e., water dripping from the canopies. The kinetic energy of a falling drop is determined by its mass and velocity (Chapman, 1948). Williamson (1981) showed that the volume of a drop bears a linear relationship to the logarithm of the width of the driptips of leaves — the wider the driptips the larger the volume. Therefore, throughfall from canopies with leaves which on the average have large driptips has high kinetic energy. Due to gravitational force, the velocity of the falling drop is accelerated. However, as the velocity increases the drag force of the air, working in the opposite direction of the movement, also increases until an equilibrium is reached between gravitational force and drag, at which time the velocity becomes constant — the so-called terminal velocity. At a free-fall distance of 7.8 metres, 95 percent of the terminal velocity has been attained (Laws, 1941). Hence, a canopy base that is more than 8 metres high does not have much additional effect on the velocity of the falling drop. However, there are large differences in the drop's velocity with canopy heights lower than 8 metres and, therefore, its kinetic energy. Consequently, when a homegarden has a multilayer structure, with the lowest canopy base being about 3 metres high and the leaves not having narrow driptips, the protective effect on splash erosion will be much less than in a homegarden with tall trees and a very low closed undergrowth with narrow driptips. Litter also provides effective protection against the erosive force of falling drops.

Surface erosion is the removal of soil from the surface by water running over it. The erosive force of running water is also determined by its mass and velocity. Litter reduces the run-off coefficient by increasing infiltration and thus decreases the mass of the overland flow and, hence, also its erosive force.

Measurements of the erosivity of throughfall in multilayered homegardens showed that it was 135 percent of incident rainfall (Ambar, 1986). But splash erosion was only 80 percent of that in an open space. The reduction of splash erosion in the homegardens was due to the low-growing crops and litter. Ambar made similar measurements in monoculture bamboo groves which did not have an undergrowth but with the ground being covered with a mat of litter. The erosivity of throughfall was 127 percent of incident rainfall, but the splash erosion was only 47 percent of that in an open space. Therefore, although the bamboo groves did not have a multilayer structure, the splash erosion was very much reduced by the litter. The lower erosivity of throughfall in the bamboo groves compared to that in the homegardens was the result of the narrower driptips of bamboo leaves compared with the average width of those in homegardens.

In another experiment, splash erosion was measured in a mixed garden, in a bamboo grove and in an open space where the weeds and litter had been removed (Soemarwoto and Soemarwoto, 1984). The erosion in an intact mixed garden and bamboo grove was minimal, while in the open, erosion increased sharply with higher rainfall intensities. When the litter and lower-level crops of the mixed garden and the litter of the bamboo grove were removed, erosion increased significantly in both types of garden. The erosion curve of the mixed garden was a function of rainfall intensity and was greater than that of the bamboo grove. This again was due to the narrow driptips of the bamboo leaves. Clearly the erosive effect of throughfall in agroforestry systems is generally higher than that of incident rainfall and driptip, undergrowth and, especially, litter are important factors in reducing the rate of erosion in such systems (Wiersum, 1984).

Animals in homegardens are important elements in the cycling of matter. In West Javanese villages plants, goats, sheep, horses, chicken and fish, and also man, are components of the recycling of wastes. In non-Muslim regions the pig plays the role offish. Thus man is an integral part of the trophic system from which he obtains nutrients and income (Soemarwoto et al, 1975). Naturally, there is a health hazard attached to this recycling system. Therefore, although the recycling of wastes does present an excellent opportunity for the efficient use of resources and helps in the maintenance of soil fertility, it should not be accepted uncritically.


Prospects and pitfalls

The fact that homegardens have existed for many centuries, if not millenia, in some parts of the world shows that they are ingrained in the tradition and culture of the peoples in those regions. In some parts at least they have not shown any decline, e.g., in Indonesia. Data from Java show that the total area under homegardens has increased from 1.398 million ha in 1933 to 1.417 million ha in 1937 (Terra, 1953)and 1,554 million ha in 1980 (Government of Indonesia, 1982), representing, respectively, 18.0, 18.1 and 18.5 percent of the total agricultural land. In areas which suffer from heavy soil erosion, such as in the upper Solo River basin in Central Java, there are no visual symptoms of soil erosion under homegardens while the fields outside the village are heavily eroded. In a short survey in the Phu Wai watershed in Khon Kaen, north-eastern Thailand, the author and his team also observed that in the plantation forests and upland fields there was slight-to-severe soil erosion, but there was almost no erosion under the homegardens. Homegardens are also a rich genetic resource. Therefore, it seems reasonable to conclude that homegardens are a sustainable production system. Their socio-cultural functions and the recycling systems also contribute to their sustainability. However, in general, this sustainability prevails under conditions of low yield and input. Therefore, if we wish to use the homegardens as a major tool to raise the standard of living of the people to satisfactory levels, the question arises whether the yield and the income from homegardens can be significantly increased without sacrificing their sustainability.

The earlier-mentioned unpublished studies of Stoler showed that homegardens responded positively to higher inputs of labour. Moreover, introduction of new species as well as high yielding and pest/disease-resistant varieties and/or selection of superior lines from the existing stock, adoption of better planting techniques, and post-harvest technology and marketing systems should be able to increase yield and income from homegardens significantly.

Although many homegardens are traditional, especially in the rural areas, they are not static but are capable of responding to new opportunities or adapting to new conditions. For example, in samples of homegardens of transmigrants in Lampung, Sumatra, who migrated from Tulungagung, East Java, we found a total of 138 species, while in the same number of samples in their village of origin we found only 69 species. Of these 138 species, 96 were acquired locally by the transmigrants, and of these, 36 species had been obtained from the local indigenous people. A majority of the transmigrants believed that local plants grew better than the ones they took from their original village.

The homegardeners in the tourist areas who grow ornamental plants to cater to tourist demand and obtain high economic returns are another example of the ability of traditional people to respond to new opportunities.

There are many examples of introduction of higher inputs and improved technology in homegardens, both spontaneously carried out by the people or stimulated by the government. Improved strains of fish, rabbits, chickens, coconut, citrus, cloves and vegetables have been widely introduced in Indonesian homegardens. The trees are planted in rows and the annuals in blocks. Near to and in cities highly valued plants such as orchids and animals such as poultry are raised. Fish and chickens are not left untended, but are fed with high-quality feed, and the chickens are vaccinated against diseases. Chemical fertilizers and pesticides are increasingly being used. Many homegardeners have benefited greatly from this development to the extent that they obtain their income solely or primarily from their gardens — an income which is sufficient to support a relatively comfortable life-style.

However, these developments also carry risks. These risks are associated with the changes in the structure of homegardens, which in turn bring about changes in their ecological as well as socio-economic functions. Paying attention solely to the tangible economic and nutritional gains of homegardens, and agroforestry in general, runs the risk of sacrificing the intangible ecological and social values. For example,. when market demand and price offered for a certain plant product becomes high, the cultivation of that species will spread, often replacing those species and varieties which are of little or no immediate economic value. This causes a reduction in the complexity of the homegarden and degeneration of its forest-like structure. In such processes of commercialization, the highly nutritious, yet commercially less valuable local vegetables are usually the first ones to go. It is not easy to achieve homegarden development with both nutritional and economic advantages.

In homegardens with low species diversity, harvesting becomes more seasonal, instead of continuous. Its multifunctional characteristics decline with a corresponding reduction in its versatility. Its function as a living store from which one can harvest according to the nature and time of need has been more or less lost. In these commercialized homegardens control of species composition and harvest has changed from being internal, i.e., by the gardeners themselves, to external, i.e., by the market forces.

Commercialization causes a decline in the diversity of species and/or varieties, and consequently the process of genetic erosion sets in. For example, in the 1920s, 75 varieties of mango had been reported in the Cirebon area in West Java; but in a recent survey in the same area we found only 48 varieties. In Depok near Jakarta, where homegardeners have specialized in commercial fruit growing, we found only one variety of mango in a sample of 15 homegardens. In the same samples we found four banana varieties, while in the rural interior not far from Jakarta there were 25 varieties in a sample of the same number of homegardens.

The dominance of a certain crop on the farm increases the risk of losses due to its specific pests and diseases. Although sometimes a higher number of plant species can lead to an increase of pest losses, the advantage of a species-rich polyculture is undoubtedly that the risk of losses is spread among many species (Ewel, 1986).

In many homegardens with fruit trees, there are signs of soil erosion, even though these homegardens are not monocultures and to a certain extent they still retain the layered canopy structure but without the low-growing, ground-covering crops. Often there was no litter on the ground which led to severe erosion, for example, where clove trees were dominant and the leaves were collected to be distilled for their oil, and under coffee bushes where clean weeding was practised to facilitate easy collection of the beans which fell on the ground. Erosion was also often observed in homegardens in which vegetables were grown in nutrition campaigns.

The availability of chemical fertilizers has reduced the need for organic manures. Composting is deemed cumbersome and time consuming, and its opportunity cost is considered high, while chemical fertilizers can be bought easily in large or small quantities, as needed. As a result the extent and intensity of the recycling systems are declining. This reduces the efficiency of resource use and in the long run will also affect soil structure and fertility.

The social functions have also been lost to a greater or lesser extent. Fences are now often built for security reasons. Sharing of the harvest and amenities of the garden, such as a well, is no longer a common practice. Households have become more individualized and less community-oriented. The concept of trespass is becoming more widespread. Equitability has declined.

Refusal or resistance to improvement of the homegarden is also common. These failures often happen when the opportunity cost is high, when the growth requirements are not met, or when the introduced species conflict with the interest of the people. For example, many poor people have to use the labour of the whole family, i.e., husband, wife and children, to earn money and cannot spare the time for gardening, even though they know they can get some additional income or nutritious food. Many vegetables, which have been recommended in nutrition campaigns, require full sunlight or intensive cultivation for their growth and, consequently, grow poorly in the shade or half-shade of trees or when they are not well tended. These plants disappear soon after the campaign ends, or more fortunately some of the heliophilic species or varieties end up being planted on the dykes of rice fields.

An example of conflict between an introduced variety and the existing homegarden structure and species association was the introduction of the rapid-growing and high-yielding hybrid coconut. Traditionally the coconut, which requires full sunlight for maximum production, occupied the highest canopy. But the hybrid variety has a low growth habit and, hence, it displaces the fruit trees, while the spatial niche above this low hybrid becomes empty. As a result these plants are only grown at the edges of homegardens, just to please the government officers. (Also see Hoskins, this volume.)

From the above discussion we see that while there are many opportunities and potentials for improving homegardens, there are many risks associated with it if we proceed in a manner which only takes into account the economic and nutrition aspects. A system has properties of productivity, stability, equitability and sustainability (Conway, 1985). These are being changed. High-yielding, but high-input and high-risk, species and varieties increase productivity, but at the expense of stability, equitability and sustainability. Pests and diseases, market-price fluctuations and high seasonality reduce stability. High-yielding, high-input, high-risk species and varieties tend to increase inequitability. Inequitability adversely affects sustainability (Crosson, 1986). Soil erosion and loss of soil fertility and genetic erosion also reduce sustainability. Consequently, sustainability is being jeopardized and, hence, long-term gains become questionable.

Erosion hazards can be dealt with relatively easily since the factors influencing them are well known. Good ground cover and litter, or mulch, are essential. Improved homegardens should be designed and managed in such a way that they have both. Leaves with driptips are beneficial and, as far as possible, efforts should be made to grow plants with such leaves. A multilayer canopy structure is not a sine qua non, if the above requirements are fulfilled. Therefore, we do not have to aim specifically at having such a structure. Usually terracing is not necessary, since land for settlements is made flat to facilitate easy construction of the houses. But terracing will be necessary for homegardens which are located on slopes.

The needs of the people, present and future, and taking into account their traditional and religious beliefs and ecological conditions, should be the bases for the planning of species and varietal composition. Whenever possible a combination of woody perennials and annuals should be attempted since they combine the stability of a mature system and the high productivity of an immature one. This would reduce the need for high inputs which would therefore make such improved homegardens more accessible to the poor. Usually the perennials play the role of cash crops and the annuals food. The perennial trees will also help the poor through lean periods (Chambers and Longhurst, 1986).

Since homegardens are a part of the total agro-ecosystem and linkages exist between them and the other parts of the system, i.e., the rice and the dry fields, their development cannot be considered in isolation. Information on the agricultural calendar and the seasonality of homegarden crops would enable us to design a species composition which would improve the role of the homegarden so that it could fill in the troughs of the lean periods and the seasonability of labour supply and demand. Considering the earlier-mentioned study of Stoler (1975), presumably the smaller farmers would be more responsive to providing more labour for homegarden improvement than the larger ones.

In conclusion we can say that homegardens do have a promising future. However, while it is relatively easy to increase yields and income, there are difficult problems in achieving long-term sustainability. These difficulties are both in the biophysical and in the socio-economic realm. It is recommended that ICRAF look into these problems and stimulate research to seek appropriate solutions.


REFERENCES

Abdoellah, O.S. 1980. Structure of homegardens of Javanese and Sundanese people in Bantarkalong. First degree thesis, Department of Biology, Faculty of Science and Mathematics, Padjadjaran University, Bandung, Indonesia (Indonesian).

Ahmad, H., A. Martadihardja and Suharto. 1980. Social and cultural aspect of homegarden. In J.I.Furtado (ed.), Tropical ecology and 'development. Kuala Lumpur: The International Society of Tropical Ecology, Kuala Lumpur.

Ambar, S. 1986. Aspects of vegetation and land use in the erosion process in the Jutiluhur lake catchment, West Java. Doctoral thesis. Padjadjaran University, Bandung, Indonesia.

Anderson, E. 1954. Reflections on certain Honduran gardens. Landscape 4: 21-23.

Asare, E.O., S.K. Oppong and K. Twung-Ampofo. 1985. Homegardens (backyard gardens) in the humid tropics of Ghana. In Proceedings of the First International Workshop on Tropical Homegardens. Bandung, Indonesia: UN University and Institute of Ecology, Padjadjaran University (in press).

Brierley, J.S. 1985. West Indian kitchen gardens: A historial perspective with current insights from Grenada. Food and Nutrition Bulletin 7: 52-60.

Brownrigg, L. 1985. Homegardening in international development: What the literature shows. Washington, D.C.: The League for International Food Education (LIFE).

Chambers, R. and R. Longhurst. 1986. Trees, seasons and the poor. Institute of Development Studies Bulletin (University of Sussex) 17: 44-50.

Chapman, G. 1948. Size of raindrops and their striking force at the soil surface in a red pine plantation. Transactions of the American Geophysical Union 29: 664-670.

Christanty, L., M. Hadyana, Sigit and Priyono. 1980. Light distribution in a Sundanese homegarden. Paper presented at the seminar on the Ecology of Homegardens III. Institute of Ecology, Padjadjaran University, Bandung, Indonesia (Indonesian).

Conway, G.R. 1985. Agroecosystem analysis. Agricultural Administration 20: 31-55.

Covich, A. and N.H. Nickerson. 1966. Studies of cultivated plants in Chogo dwelling clearings, Darien, Panama. Economic Botany 20: 285-301.

Crosson, P. 1986. Sustainable food production. Food Policy: 143-156.

Danoesastro, H. 1980. The role of homegarden as a source of additional daily income. Paper presented at the seminar on the Ecology of Homegardens III. Institute of Ecology, Padjadjaran University, Bandung, Indonesia (Indonesian).

Ewel, J.J. 1986. Designing agricultural ecosystems for the humid tropics. Ann. Rev. Ecol. Syst. 17: 245-271.

Falanruw, M.V.C. 1985. The traditional food production system of Yap Islands. In Proceedings of the First International Workshop on Tropical Homegardens. Bandung, Indonesia: UN University and Institute of Ecology, Padjadjaran University (in press).

Fernandes, E.C.M. and P.K.R. Nair. 1986. An evaluation of the structure and function of tropical homegardens. Agricultural Systems 21: 179-310.

Fernandes, E.C.M., A. OTctingati and J. Maghembe. 1984. The Chagga homegardens: A multi- storeyed agroforestry cropping system on Mt. Kilimanjaro, Northern Tanzania. Agro-forestry Systems 2: 73-86.

Gonzales-Jacomes, A. 1981. Homegardens in Central Mexico, their relationship with water control and their articulation to major society through productivity, labour, and market. In Prehistoric Intensive Agriculture in the Tropics. Canberra: Australian National University.

Government of Indonesia. 1983. Statistics of Indonesia. Central Bureau of Statistics, Jakarta, Indonesia.

Haryadi, M.M.S.S. 1975. Potential contribution of homegardening to nutrition intervention program in Indonesia. Seminar on Food and Nutrition, Gadjah Mada University, Yogyakarta, Indonesia.

Hutterer, K.L. 1984. Ecology and evolution of agriculture in Southeast Asia. In T.A. Rambo and P.E. Sajise (eds.), An introduction to human ecology research on agricultural systems in Southeast Asia. Los Banos, Philippines: University of the Philippines.

Jacob, V.J. and W.S. Alles. 1987. Kandyan gardens of Sri Lanka. Agroforestry Systems 5:123-137.

Kamtuo, A., K. Lertrat and A. Wilariat. 1985. Traditional homegardening in Hinland village, Khon Kaen Province, Northeast Thailand. In Proceedings of the First International Workshop n Tropical Homegardens. Bandung, Indonesia: UN University and Institute of Ecology, Padjadjaran University (in press).

Karyono, E. 1981. Homegarden structure in rural areas of the Citarum watershed, West Java. Doctoral thesis, Padjadjaran University, Bandung, Indonesia.

Kimber, C.T. 1973. Spatial patterning in the dooryard gardens of Puerto Rico. Geographical Review 6: 6-26.

Laws. J.O. 1941. Measurements of the fall-velocities of water drops and raindrops. Washington, B.C.: U.S. Soil Conservation Service.

Leuschner, W.A. and K. Khalique. 1987. Homestead agroforestry in Bangladesh. Agroforestry Systems 5: 139-151.

Lundgren, B.O. and J.B. Raintree. 1983. Sustained agroforestry. In B. Nestel (ed.), Agricultural Research for Development: Potentials and Challenges in Asia. The Hague: ISNAR.

Nair, M.A. and C. Sreedharan. 1986. Agroforestry farming systems in the homesteads of Kerala, southern India. Agroforestry Systems 4: 339-363.

Ninez, V. 1985. Working at half-potential: Constructive analysis of home garden programmes in the Lima slums with suggestions for an alternative approach. Food and Nutrition Bulletin 7: 6-14.

Noor, Z.M. 1981. The effect of various light intensities on the physiological and morphological characteristics of taleuspadang (Xanthosoma atrovirens). First degree thesis, Faculty of Science and Mathematics, Padjadjaran University, Bandung, Indonesia.

Omta, S.W.F. and F.T.J.M. Fortuin. 1978. The cultivation ofSolanum nigrum L. as a leaf and fruit vegetable in the homegardens of West Java: Surveys and experiments. Bandung; Indonesia: Institute of Ecology, Padjadjaran University, and Groningen, Netherlands:Department of Plant Physiology, Groningen State University.

Ochse, J.J. and G.J.A. Terra. 1937. The economic aspect of the "Koetawinangoen report". Landbouw 13: 54-67 (Dutch).

Penny, D.H. and M. Singarimbun. 1973. Population and poverty in ruralJava: Some arithmetics from Sriharjo. Ithaca, New York: Department of Agricultural Economics, New York State University.

Poerwadarminta, W.J.S. 1976. General dictionary of the Indonesian language. Jakarta: PN Balai Pustaka (Indonesian).

Ramsay, D.M. and K.F. Wiersum. 1974. Problem of watershed management and development in the Upper Solo River Basin. Paper presented at the Conference on Ecological Guidelines for Forest, Land and Water Resources, Institute of Ecology, Padjadjaran University, Bandung, Indonesia.

Simoons, P.J. 1965. Two Ethiopian gardens. Landscape 6: 15-20. Soemarwoto, O. and I.

Soemarwoto. 1984. The Javanese rural ecosystem. In A.T. Rambo and P.E. Sajise (eds.), An introduction to human ecology research on agricultural systems in Southeast Asia. Los Banos, Philippines: University of the Philippines.

Soemarwoto O., I. Soemarwoto, E. Karyono, E.M. Soekartadiredja and A. Ramlan. 1975. The Javanese homegarden as an integrated agroecosystem. In Science for a Better Environment. Tokyo, Japan: Science Council of Japan.

Sommers, P. 1985. Advancing Pacific Island food gardening systems. Some observations and suggestions. Paper presented at the First International Workshop on Tropical Home-garden. Bandung, Indonesia: UN University and Institute of Ecology, Padjadjaran University.

Stoler, A. 1975. Garden use and household consumption patterns in a Javanese village. Department of Anthropology, Columbia University, New York.

Terra, G.J.A. 1954. Mixed garden horticulture in Java. Malayan J. Trop. Geogr. 4: 33-43.

Terra, G.J.A. and S. Satiadiredja. 1941. Results of a third budget-research in the district of Koetowinangoen. Landbouw 17: 137-178 (Dutch).

Thaman, R.R. 1984. Urban agriculture and home gardening in Fiji: A direct road to development and independence. Transactions and Proceedings of the Fiji Society 14.

.1985. Mixed gardening in the Pacific Islands: Present status and future prospects. Paper presented at the First International Workshop on Tropical Homegarden. Bandung, Indonesia: UN University and Institute of Ecology, Padjadjaran University.

Wagner, P.L. 1958. Nicoya: A cultural geography. University of California Publication in Geography, Vol. 12, No. 3.

Wiersum, K.F. 1984. Surface erosion under various tropical agroforestry systems. Paper presented at the Symposium on the Effects of Forest Land Use on Erosion and Slope Stability, East-West Center, Honolulu, Hawaii.

Williamson, G.S. 1981. Driptips and splash erosion. Biotropica 13: 228-231.

Yoda, K. 1974. Three dimensional distribution of light intensity in a tropical rainforest in West Malaysia. Japanese J. Ecol. 24: 247-254.