Agriculture.

Ecosystems and Environment, 41 ( 1992 ) 189-207 189

Elsevier Science Publishers B.V., Amsterdam

Are tropical agroforestry home gardens

sustainable?

Emmanuel Torquebiau
International Centre for Research in Agroforestry (ICRAF). P.O. Box 30677. Nairobi. Kenya
(Accepted 17 January 1992 )

ABSTRACT

Torquebiau, E., 1992. Are tropical agrotbr.=stry home gardens sustainable'? .4gric. Ecosystems Era,i-
ron., 41 : 189-207.

While tropical multistrata home gardens (or tree home gardens) are classically said to be sound,
efficient and sustainable land-use systems, there is little quantitative evidence and detailed analysis
of home gardens in the literature to support this. In order to strengthen the basis of this assumption
and to contribute to an opt cational understanding of sustainability, a literature review was applied to
home g.,rdens, covering different bio-physical and socio-economic conditions, For this, a series of
sustainability descriptors were identified, hypotheses formulated on what should be the effect of home
gardens on these descriptors, and possible indicators confirming the hypotheses mentioned, when
found in the literature.
Home gardens possess a number of sustainability attributes, with regard not only to their ability to
meet a number of farmers' needs without negatively affecting the resource base, and in many cases
even improving it, but also to their potential to meet several economic, social, ecological and institu-
tional conditions which contribute to their sustainability.

INTRODUCTION

Many definitions have been put forward to define sustainability (see Brown
et al., 1987), but there is not yet a clear agreement on what exactly sustaina-
bility is and how to measure it. As home gardens claim to have a number of
attributes relevant to sustainability, it must be possible to identify and quan-
tify these attributes. This would allow us to say whether home g~.rdens are
sustainable or not, and why, and would be a contribution toward an opera-
tional definition of sustainability.
Home gardens are agroforestry land-use systems with multipurpose trees
and shrubs in intimate association with seasonal and perennial agricultural
crops and livestock, within the compound of individual houses, and under

Correspondence to."E. Torquebiau, International Centre for Research in Agroforestry (ICRAF),

P.O. Box 30677, Nairobi, Kenya.

© 1992 Elsevier Science Publishers B.V. All rights reserved 0167-8809/92/$05.00

] 90 E. TORQUEBIAU

the management of family labour (Fernandes and Nair, 1986). It should be

noted that this definition applies only to home g~xdens with trees and that
there are some home gardens which comprise only agricultural crops. A better
name for agroforestry home gardens could be 'tree home gardens'. The term
'forest gardens' (or 'village forest gardens') can also be found in the literature
but applies to gardens which are much larger, less densely planted and not so
well tended as the home garden (see Michon, 1983). In some cases, there is a
continuity between the different types of gardens.
Home gardens which are well known in many tropical countries, are clas-
sically said to be sovnd, efficient and sustainable land-use systems which ful-
fill the basic needs of the local populations, avoid environmental deteriora-
tion, etc. (e.g. Fernandes and Nair, 1986; Landauer and Brazil, 1990). Yet
ti:,'se assets of home gardens are seldom expressed through any quantified or
functional analysis. In other words, one assumes that home gardens are sus-
tainable without really having exa~i~ined the evidence. If they are sustainable,
what makes them so?

A FRAMEWORK FOR SUSTAINABILITY ANALYSIS

The following framework, which will be used in this article, is adapted from
the 'Committee on Agricultural Sustainability for Developing Countries'
( 1987 ). It describes sustainability of an agricultural system as "its ability to
meet evolving human needs without destroying and if possible, improving the
natural resource base on which it depends". A sustainable rural production
system is only one of the elements in the global concept of sustainability that
includes a series of conditions outside the rural system that are classified as
economic, social, ecological, political and institutional.
For an agricultural system to be sustainable, it should fulfill several main
requirements. First, these are: (a) soil conservation, including erosion con
trol and fertility maintenance; (b) the efficient use and conservation of exist-
ing resources (water, light, energy, genetic resources, labour); (c) the use of
biological interactions between the different elements of the agricultural sys-
tem (for exar~ple, mulching, the association of climbing plants ~md supports,
nitrogen fixation, biological control of weeds and diseases); (d) the use of
inputs that are easily available and of inputs and practices that ensure both
human health and environmental conservation.
Second, and especially for the distinct case of small-scale farmers who de-
pend both on cash crops and on subsistence crops, a sustainable system must
fulfill stricter requirements, such as: (a) meet the farmers' energy needs (fuel,
heat, labour); (b) meet the farmers" needs for subsistence, so that they may
be assured of having an adequate and balanced diet; (c) strengthen coopera-
tion between local community members; (d) ensure that social equity, cul-
tural integrity, ethnic and gender issues are adequately considered. These re-
ARE TROPICAL AGROFORESTRY HOME GARDENS SUSTAINABLE? 191

quirements can help households withstand difficult periods caused by climatic

or economic stress, improve living conditions in rural areas while bridging
the gap between production seasons, take care of various social concerns and
ensure the survival of traditional rural systems.
Finally, there are some external parameters that contribute directly to the
sustainability of rural production systems. Some of these are: (a) the quality
of infrastructures available to farmers (roads, irrigation, means of transport,
etc.); (b) credit opportunities, with manageable risk conditions in case of
production failure; (c) access to a minimum of social services (school, health
facilities, family planning); (d) direct or indirect access to national and in-
ternational markets, with prices that are favourable in relation to agricultural
production costs.

SUSTAINABILITY INDICATORS

in order to be able to clearly describe attributes contributing to the sustain-

ability of agricultural systems, it is essential to identify indicators which in
turn would suggest the empirical data to test specific assumptions explaining
sustainability. This procedure would permit evaluating and demonstrating
scientifically whether a given system is sustainable or not and why.
Such indicators should be easy to measure, and be reproducible over time
and across systems. An indicator gives the characteristics of a descriptor, cho-
sen to be of critical importance to sustainability. Hypotheses can then be for-
mulated to check whether the particular production system has a positive ef-
fect on the descriptor. If the system is not sustainable, the effect is negative.
Different consequences can be observed from these positive and negative ef-
fects. Thus, the indicator is a measurement of the effect of the system on the
descriptor.
For example, if one formulates a hypothesis that one of the sustainability
attributes of a given system is the regular distribution of labour throughout
the year, the descriptor is 'labour', and the effect of the system of this descrip-
tor is that it does not induce any peak of labour but distributes it over the
different moiaths. The indicator of such a distribution is a table showing the
labour units required per month over a year.
A first set of sustainability indicators must relate to the resource base. One
descriptor, for example, is the soil: if the system is sustainable, soil fertility
(among other factors) will be positively affected and indicators will be the
organic matter and moisture content of the soil. A second set of indicators
concerns the function of the system itself. These indicators will show whether,
how and why the management and performance of the system are compatible
with the requirements of sustainability. An example of a management indi-
cator is the ratio of endogenous/exogenous inputs; if the system is sustaina-
ble, it will drive the use of inputs towards the endogenous rather than exoge-
192 E. TORQUEBIAI!

nous sources. An example could be the use of farmyard manure, as opposed

to chemical fertilisers.
The last set of indicators focuses on other systems which may be influenced
by the system under study. This is because sustainability is taken here in a
hierarchical context: a sustainable agricultural system cannot exist without
external factors also conforming to the requirements of sustainability, and a
system which is sustainable influences in a positive way other systems with
which it has relationships. For example, if home gardens in a given country
produce a lot of fuel wood, there will be less encroachment into forest areas.
So forests will be a descriptor, positively affected by home gardens, and an
indicator will be the amount of fuel wood collected from forests compared
with that from home gardens.
in the next section, relevant sustainability indicators are applied to home
gardens. These are presented for different descriptors. They have been ob-
tained from a survey of literature and are, unless evidence is sufficient, based
on quantitative data. When no data were found to indicate the effect of the
system on a particular descriptor, this is indicated. This review of literature
does not claim to be exhaustive but is a first attempt to test the usefulness of
sustainability indicators.

SUSTAINABILITY INDICATORS FOR HOME GARDENS

Resource base

Soil

Descriptor 1: soil erosion. Hypotheses: the multilayer plant cover protects the
soil from direct impact of raindrops and the dense root system bonds the soil.
As a positive consequence, erosion is generally low under a home garden. There
can be, however, a negative consequence: if there is no understorey under the
tall trees, the erosive power of rain may increase by coalescence of raindrops
in high canopies.
Indicator: u rate of soil erosion. Low rates of erosion under home gardens
were measured by Wiersum (1988) (0.06 t ha -~ year -I in Java), Krishna-
rajah (1984) in Wiersum (1988) (0.05 t ha -I year -~ in Kandy Sri Lanka).
Such values can be compared with normally accepted values under farming
conditions, e.g. the US Soil Conservation Service sets limits for tolerable ero-
sion mainly in the range 2.2- l 1.2 t ha- t year- t, limits which are often not
achieved (Young, 1989). Lundgren (1980) also reported less erosion under
home garden than in adjacent forest in tanzania. Many authors mention as
an observation, but witt,~out measurements, that there is obv;ously litle ero-
sion under home gardens~ e.g.: Pelzer ( (1948) Java, in Anderson (1979) ),
Anderson ( 1950, Guatemvla), Balasubramanian and Egli ( 1986, Rwanda),
ARE TROPICAL AGROFORESTRY HOME GARDENS SUSTAINABLE? 193

Achutan Nair and Sreedharan (1987, India), Nifiez (1987), Brierley ( 1985,
Grenada), Christanty et al. ( 1986, Java).
Karyono (1990) reported that erosion under home gardens in Java is
strongly ~ a t ~ u lO th,. undergrowth and litter layers: measurements of erosion
with litter and undergrowth, with litter only and without litter and under-
growth differed by factors of 2.5, 15 and 39, respectively. The importance of
the undergrowth and litter to avoid erosion by big drops formed by coales-
cence on high canopies is stressed by Soemarwoto ( 1987, Java).

Descriptor 2: soil organic matter. Hypotheses: the abundant litter and root
system of home gardens increases organic matter content of the soil. The mul-
tilayer root system retrieves nutrients from deep soil layers and litter releases
them on soil surface. The recycling of wastes and making of compost also
improves soil organic matter content.
As a consequence of these effects, biological and physical characteristics of
the soil improve. The soil nutrient statu.s (fertility) is enhanced because of
the good decomposition rate of the orgar, ic matter favoured by high soil mois-
ture content and, under warm climates,, reduced soil temperature (,~ee also
Descriptors 3 and 4). Soil structure is improved by abundant organic matter',
this increases soil porosity and aeration, hence the soil bulk density decreases
and the cation exchange capacity (CEC) increases. A negative consequence
of this mixture of many species is that some allelopathic effects can exist from
litter and roots.
Indicators: m soil organic matter content and soil bulk density. A good soil
organic matter content under a home garden seems obvious and is commonly
mentioned but has never been checked and no data were found in the litera-
tare. The following references are observations rather than measurements.
The ability of home gardens to maintain soil fertility has been mentioned by
Anderson ( 1950, Guatemala), Balasubramanian and Egli ( 1986, Rwanda),
Achuthan Nair and Sreedharan ( 1986, India) and Nifiez (1987). Achuthan
Nair and Sreedharan ( 1986 ) observe that this is owing to good nutrient cy-
cling and use of waste materials in the home gardens and that soil physical
and biological characteristics are improved as a consequence. They also men-
tion that the mixture of other plants with coconuts in a home garden reduces
the effect of the wilt root disease of the coconut. Christanty et al. (1986)
observed in Javanese home gardens that crops with high nutrient require-
ments were planted near the garbage dump. Nair and Rao ( 1977 ) in Achutan
Nair and Sreedharan (1986), observed a high rhizosphere microbial activity
under home gardens in Kerala (India).

Descriptor 3: Sc,ii moisture. Hypotheses: by intercepting and holding water

through litter arid plant cover, home gardens keep soil moisture at high levels.
Run-off and erosion are also reduced by the plant cover and good conditions
i 94 E. TORQU EBIAU

are created for the development of the soil fauna and flora. Climatic hazards
(drought) are buffered by this 'sponge effect'.
A high soil moisture content is closely linked with the reduction of temper-
ature by plant cover (see Descriptor 4) and the soil organic matter content
(see Descriptor 2). A negative consequence is that soil pests and diseases may
be favoured by high soil moisture
Indicator: - - soil moisture status. No measurements of soil moisture under
home 5ardens were found in the literature, but the observation that the soil is
moist is common, e.g. Anderson (1950) and Nifiez (1985).

Descriptor 4: soil temperature. Hypotheses: plant and litter cover reduce soil
temperature by intercepting solar radiation. Soil evaporation is reduced and
the daily variations of soil temperature are buffered. The low soil temperature
is good for the biological activity of the soil: flora, fauna, fungi, symbionts
which would otherwise not grow under high temperature and moisture stress.
Indicator: - - soil temperature. No measurements were found in the litera-
ture. Nifiez (1985) mentions that tropi¢~d multistrata home gardens main-
tain a constant soil temperature.

Climate

Descriptor 5: light. Hypotheses: light is util~sed by several layers of vegetation,

so the overall biomass yield of home gar¢lens is high. Photosynthetically ac-
tive radiation (PAR) decreases downwards in the canopy, and shade-tolerant
crops can grow in the lower layers, but not the light-demanding ones. As a
consequence, temperature decreases downwards and the resulting light and
temperature microclimate is good for the seedlings and saplings of future tall
trees. The albedo (reflected radiation) is low and the biomass/solar energy
ratio is high.
Indicators: - - understorey PAR, albedo, understorey temperature, and bio-
mass yield per hectare. Christanty et al. ( i 980) in Soemarwoto ( 1987 ) noted
the close resemblance of the light interception curve between a home garden
and the tropical rain forest. In a Javanese home garden, Christanty et al.
(1986) found that 20% of the incident light is intercepted by the top layer,
64% by the second layer, 10% by the third layer and that 6% ofthe light reaches
the bottom layer; well-known shade-tolerant crops are found in this bottom
layer.
Omta and Fortuin (1978) and Noor ( 1981 ) both in Soemarwoto (1987)
noted the good photosynthetic rates of Solanum nigrum (leaf and fruit vege-
table in Java) and Xanthosoma atrovirens (cocoyam) under shade in a home
garden. A review of useful shade-tolerant crops and their agroforestry poten-
tial in different parts of the world (especially in Indonesia), has been made
ARE TROPICAL AGROFORESTRY HOME GARDENS SUSTAINABLE? 195

by Bahri (1984). Several shade-tolerant crops (e.g. ginger, turmeric, taro),

are grown in home gardens in Bangladesh (see Abedin ,~t al., 1990).

Descriptor 6: water. Hypotheses: rainfall is efficiently intercepted by the sev-

eral plant layers, so water retention in the aerial parts of the plants is good
and the impact of raindrops on the soil is reduced if there is an understorey
of vegetative cover near the ground. Macroclimatic variations of air wetness
are buffered under the canopy of the home garden. This is closely linked with
water retention in the soil and the control ofrun-offand erosion (see Descrip-
tors 1 and 3 ). Excessive air moisture may, however, favour the proliferation
of pests and diseases.
Indicators:- undercanopy rainfall (throughfall, stemflow) and understo-
rey air humidity. Nair and Balakrishnan ( 1977 ) observed in India the buff-
ering effect of a home garden with coconuts and cocoa on the daily variations
of relative humidity (and temperature) and on the evaporation, compared
with an open field or a coconut monoculture. Baldy (1963) noted a high hu-
midity in oasis as well as a high potential evapotranspiration. Niftez (1987)
mentions that tropical home gardens maintain good air moisture conditions,
but Leiva and L6pez (1985) mention the risk of fungal diseases associated
with moisture in home gardens in Guatemala. High pest incidence in home
gardens is also mentioned by Amir and Aria (1980) in Christanty (1990) for
Java and by Alvarez-BuyUa et al. (1989) for Mexico. Amir and Aria (1980)
in Christanty (1990) noted many insect pests in home gardens in Java, but
they also noted the abundance of insectivorous birds controlling the insects,
Alvarez-Buylla et al. (1989) observed in southeast Mexico that favourable
microclimatic conditions in home gardens allow for the growing of varieties
with different microclimatic requirements and a management calendar much
more independent of climatic fluctuations.

Biological resources

Descriptor 7: useful plants and animals. Hypotheses: a very high diversity of

plants and animals are used in the home gardens. This is valuable for farmers'
everyday life because it provides a high diversity of foods and incomes and
should also be regarded as a valuable asset for future breeding programmes in
the form of germplasm banks, the latter being relevant for long-term sustain-
ability. There is good space and time utilisation by plants and animals with
different biological cycles and growth forms. The high diversity of crops, the
low density per species and the different biological cycles of the crops are all
factors which reduce the risks linked with pests and diseases.
Indicator: - - diversity of useful plants and animals (species numbers, di-
versity indices). The high diversity of plants in home gardens has been re-
corded by a large number of people and data are available from many parts
196 E. TORQUEBIAU

of the world. Record figures are probably for Java where 219 (dry season ) to
270 (wet season) different plant species have been recorded across home gar-
dens of a west Javanese village of 41 households (Soemarwoto, 1987 ) or for
southeast Mexico with 338 plant species growing (not necessarily planted ) in
the home gardens of southern Veracruz (Alvarez-Buylla et al., 1989). In
Bangladesh, farmers grow 52 different tree species in home gardens (Abedin
and Quddus, 1990).
Balasubramanian and Egli (1986) mention that local varieties of plants are
well adapted to home gardens in Rwanda, and Thaman (1990) thinks that
home gardens are the most effective avenues for the introduction and accep-
tance of new species. The preservation of wild fruits and other useful trees is
mentioned by Alvarez-Buylla et al. (1989) for home gardens in southeast
Mexico, where 15% of the woody plants are local wild species. In the tree
gardens of the Yap Islands (Micronesia), Falanruw (1990) found 21 vari-
eties of coconuts, 28 of breadfruit and 37 of bananas. Data on abundance of
medicinal plants in home gardens can be found for many countries, e.g. south-
east Asia, Mexico.
Few studies provide quantitative information on diversity of domesticated
animals in home gardens, other than saying that animals are present. Alvarez-
Buylla et al. (1989) mention an average of seven domesticated animal spe-
cies by household in home gardens in southeast Mexico. Fish ponds are a
common featm,, of home gardens in southeast Asia (Soemarwoto, 1987 ).

System
Technical management

Descriptor 8: inputs. Hypotheses: the use of endogenous inputs (e.g. mulch,

manure, wastes, bio-control of pests) is optimised in home gardens and few
exogenous, cash-demanding inputs are required.
Indicator: m inputs per hectare from farm and off-farm source. The follow-
ing references are mainly of descriptive nature, as no detailed quantitative
analysis of inputs in home gardens was found in the literature. Balasubraman-
ian and Egli (1986) noted that farmers in Rwanda do not use chemical fertil-
isers for their home gardens. The same observation was made by Alvarez-
Buylla et al. (1989) in Mexico, by Achuthan Nair and Srcedharan ( 1986 ) in
India, by Alam et al. (1990) in Bangladesh and by Thaman (1990) in the
Pacific Islands, the latter also for pesticides and other chemical products. Luu
(1989) mentions the importance of coconut husks pits for the management
of soil fertility in home gardens in Sri Lanka, while Leiva and L6pez (1985)
mention the use of home refuse for the same purpose in home gardens in
Guatemala.
Thaman (1990) and Nifiez (1985) observed that plant material for home
ARE TROPICAL AGROFORESTRY HOME GARDENS SUSTAINABLE? 197

gardens is obtained either by self-production of seeds or by vegetative repro-

duction of plants, so that no planting stock needs to be bought. Nifiez also
studied the capital and labour inputs for home gardens in Lima and con-
cluded that they were very low. Finally, Alvarez-Buylla et al. ( 1989 ) mention
that all harvesting in home gardens in Mexico is done by hand or home-made
tools. In a survey of 482 urban home gardens in Papua New Guinea, Vasey
(1990) found that no gardeners used tiller or other power implement for til-
lage, cultivation or any other task; none used herbicides or pesticides and few
used chemical fertilisers.

Socio-economic management

Descriptor 9: labour. Hypotheses: little labour is needed in a home garden; this

labour is well spread over the different seasons (no peak) and can be allo-
cated in a flexible manner. Hired labour is seldom necessary. As a conse-
quence, the farmer can work in the home garden when other activities in his
farm are low. Family labour is used and the use of existing, traditional knowl-
edge is optimised.
Indicators: - - home garden labour requirement per month and flexibility
(cropping calendar), home garden labour requirement per hectare, gender/
age labour allocation for home gardens, use of traditional knowledge and
practices in home gardens.
Several authors mention the low labour demand from home gardens in dif-
ferent countries, e.g. 1 h morning and evening in a 500 m 2 home garden in
the Philippines (Sommers, 1982); similar values in Indonesia (Hariady
(1975) in Christanty (1990)); 50 rain day-' in a 200 m 2 home garden in
Lima (Nifiez, 1985); 13-57 man-days year -I in Sri Lanka (Luu, 1989); 35-
45 days of family labour year- mduring the year of home garden establishment
and 17-22 days during subsequent years in Mexico (Alvarez-Buylla et al.,
1989). Ochse and Terra (1937) in Abdoellah (1990) mention that 7% of
people's time is spent in home gardens in Indonesia for potential production
of 44% and 32% of their total intake of carbohydrates and proteins
respectively.
Ahmad et al. (1978) in Christanty (1990) and Christanty et al. (1986)
observed that the pattern of labour in the home gardens and rice fields in Java
complement each other over the year. It is a traditional saying in Java that
there is no heavy labour in the home gardens (Christanty et al., 1986).
The flexibility of labour in home gardens is illustrated by Penny and Sin-
garibum (1973) in Christanty et al. (1986) who mention that people with no
off-farm work in Java concentrate labour on home gardens and get better yields
than other farmers who usually give home gardens low management because
of the higher priority given to paddy fields. Balasubramanian and Egli (1986)
a~zo mention the flexibility of labour in home gardens in Rwanda, as well as
198 E. TORQUEBIAU

Luu (1989) in Sri Lanka. In southeast Mexico, Alvarez-Buylla et al. (1989)

found that weeding in home gardens is done during farmc "s' spare time and
that the good microclimatic conditions in home gardens allow for a flexible
management calendar.
In India (Kerala), Achuthan Nair and Sreedharan (1986) observed that
labour demand is higher in the home gardens than in the rice fields but labour
utilisation in relation to outputs is higher in the home gardens. In his detailed
functional analysis of home gardens in Sri Lanka, Luu (1989) has found that
the use of hired labour for home gardens is uncommon.
A correlation was found by Alvarez-Buylla et al. (1989) between the diver-
sity of home gardens in Mexico and the labour force available from farmers'
families. This indicates that, although home gardens are not labour demand-
ing, they can easily be improved if labour is not a constraint. In this same
study, mention is made of the use of traditional knowledge of biological pro-
cesses (e.g. selective weeding) for home garden management.
In Bangladesh, it was found that there is a clear share of tasks between
women and men for the management of home gardens: women are mostly
involved in the pre- and post-harvest work of vegetable production while men
play a key role in timber and fruit trees growing activities (Hussain et al.,
1988).

Descriptor 10: cash. Hypothesis: home gardens need few cash inputs, so the
farmers can either use their money for other purposes or survive with very
little cash (subsistence strategy).
Indicators: - - home garden cash inputs per hectare and home gardens cash
input per month. Nifiez ( 1985 ) reports that the capital input for a 200 m s
home garden in Lima iS US$ 2.80 per growing season (5 months). Yang
(1976) in Thaman (1990) mentions that a family offive in Hawaii can save
$ 1-20 d a y ~ in tbod costs from a 35 m 2 home garden. Bompard (1986)
observed thatthe production costs for home gardens in Java are about 10%
of the output while it is 30-50 for wet rice. Ochse and Terra (1937) in Ab-
doellah (1990) report that 8% of farmers' expenses in Indonesia are for home
gardens.
Abdoelah and Marten (1986) calculated for Javanese home gardens that
the ratio of cash output to cash input is higher in home gardens than in rice
fields because ofthe low cash input and high sale value output of home garden
products, although the net cash output m -2 is twice as much in rice~ A similar
remark was made by Danoesastro (1980) in Christanty (1990) who calcu-
alated in Java that the cost of production in home gardens is 15.1% of the
gross income, while it is 55.9% in the rice fields. AIvarez-Buylla et al. (1989)
observed in Mexican home gardens that the whole technological package
(tools, etc. ) used in home gardens has a value of only 0.01-0.04% of the total
annual family cash income.
ARE TROPICAL AGROFORESTRY HOME GARDENS SUSTAINABLE? 199

Technical performance

Descriptor 11: bio-physical outputs. Hypotheses: overall biomass yield is high

and nutritional value of varied products is good. Complete crop failure is very
unlikely because of the high diversity of products in the home gardens.
Indicators: --biomass to solar energy ratio, nutritional analysis of diet pro-
vided with home garden products, and land equivalent ratio. The net primary
production of a home garden in Java is reported by Christanty et al. (1986)
to be about 1250 c a l m -2 year -I. No further references were found about
overall biomass yield in home gardens, and the total bio-physical outputs are
probably difficult to measure. The similarity of structure between home gar-
dens and natural forests as observed by Michon et al. (1983), Torquebiau
( ! 984 ) and Luu (1989) suggests a high biomass to solar energy ratio in home
gardens, as has been measured in forests. Wiersum (1985) noted that a Ja-
vanese home garden can produce 7-9 m a of wood ha- i year- i; this compares
well with growth rates obtained in plantation forestry.
In the Philippines, the nutrient recommended daily allowance is exceeded
by a 500 m 2 home garden (Sommers, 1982). Abdoellah and Marten (1996)
in Marten (1990) and Karyono (1990) found in Java that the nutrients in
shortest supply in rice (vitamins C and A and calcium) are abundant in the
home gardens. Marten (1990) found in Java that the number of crop species
in home gardens is significantly correlated with nutrient production, espe-
cially vitamins and minerals. Thaman (1990) noted that the main staple foods
in the Pacific Islands are obtained from the home gardens, while important
dietary complements for children are obtained from home garden fruits in
Mexico (Alvarez-Buylla et al., 1989). Similarly, the daily per capita con-
sumption of animal products from home gardens ranges from 0.022 to 0.139
kg. In the Solomon Islands, a correlation was observed between vitamin de-
ficiency and people without home gardens (Fitzroy (1981) and Willmot
( 1968 ) in Thaman (1990) ). In Papua New Guinea, a 150 m 2 home garden
supplies enough fresh vegetables for a family of four (Kesavan (1979) in
Thaman (1990) ). In Hawaii, a 35 m 2 home garden can provide 100% of the
vitamins A and C, 50% of iron and 18% of protein for a family of five.
In Java, amounts of calories obtained from the home gardens (in percent
of total calorie intake) vary according to authors and areas: 3.1% (poor peo-
ple) or 8% (rich people) (Christanty et al., 1986) to 10.9% (Haryadi (1977)
in Christanty (1990) ) and to 44% (Ochse and Terra ( 1937 ) in Bompard et
al., 1980); similarly for proteins: 4% or 9% to 5.2% and to 32%. These are
quite high values, for a country where the diet is mainly rice-based.
No calculation of land equivalent ratio for home gardens was found in the
literature. Data for these would be strenuous to collect, because of the high
diversity of crops in home gardens, but would certainly yield interesting
figures.
200 E. TORQUEBIAU

Socio-economic performance

Descriptor 12: socio-economic outputs. Hypotheses: outputs from the home

gardens are diversified and distributed over time, and can be for subsistence,
cash or re-utilisation into the home garden. There is flexibility to favour one
or another kind of output. This provides a high safety factor against market-
ing and seasonality hazards, including the components of the home gardens
which constitute a capital value for the farmer, e.g. timber trees. A drawback
is that there is a small quantity of any given output. Besides, yields may be
low during the early years of establishment of big trees; home gardens are
theft prone and wandering animals may damage crops.
Sociological advantages of having a dense, tree-sheltered garden by a house
relate to factors of socialisation for children and adults as well as opportuni-
ties for exchanges and interactions at community level and the preservation
of traditional customs and beliefs.
lr~dicators: -- value of production per hectare per month or year, value of
production per household per month or year, cash income per hectare per
month or year, and sociological benefits of home gardens.
The high diversity of products is a general feature of home gardens and it a
natural consequence of the diversity of plants being grown (see Descriptor
7). Distribution of the production throughout the year in home gardens is
mentioned by Anderson (1950) for Guatemala, Brierley (1985) for Gren-
ada, Christanty et al. (1986) for Java, Kandaragama (1983) in Christanty
(1990) for Sri Lanka, Alvarez-Buylla et al. ( ! 989) for Mexico. Hysam et al.
(1979) in Karyono (1990) observed a complementarity in time between
home gardens and paddy field productioll in Java, which is probably linked
with labour allocation for rice harvesting and indicates a good flexibility in
home garden management as well as the rcle ef home gardens in time of scarc-
ity ofother commodities. A similar compl~:mentarity is observed in Sri Lanka
by Luu (1989), in terms of food balance, especially starchy products, and
cash income.
Surplus of fruits or other subsistence products, including animal products,
providing extra income to the farmers, is 1lentioned for home gardens in dif-
ferent countries, e.g. Philippines (Sommers, 1982), India (Abdul Khader,
1982 ), Mexico (AIvarez-Buylla et al., 1989 ).
In Lima, a 200 m 2 home garden provides 10% of the family earnings, plus
the subsistence productions (Nifiez, 1985). In Sumatra, the income from some
fruit-producing home gardens is similar to the income from paddy fields
(Bompard, 1986). In Java, Christanty et ai. (1986) observed that the relative
income from home gardens is higher for poor people than for rich people, and
it ranges from 10 to 50% of the family total income (7-56% according to
Soemarwoto ( 1987 ) and 0.84-54% according to AbdoeUah (1990) ), with
50% of the products being directly consumed (Ahmad et al., 1980). Karyono
ARE TROPICAL AGROFORESTRY HOME GARDENS SUSTAINABLE? 201

(1990) reports that in some areas of Java, the income per unit area per year
from home gardens can be higher than that of paddy fields. In home gardens
in Sri Lanka, the self-consumption rate varies from 55 to 88% (Luu, 1989).
The favourable flexibilityof home garden management is illustrated by AI-
varez-Buylla et al. (1989) in Mexico, who found that the proportion of plant
production sold or consumed varied greatly across home gardens in the same
village. Michon and Mary (1990) noted in Indonesia that the price of fruits
(mostly obtained from home gardens) multiplied by 10 from 1975 to 1983,
while the price of rice multiplied by 2.4 only.
Ahmad et al. (1978), and Kimber (1973), both in Christanty (1990) ob-
served that home gardens ix'~Java are used for playgrounds and socialisation.
Thaman (1990) calculated ill the Pacific Islands that 64% of households with
home gardens distribute harvest products to their friends and relatives. Ab-
doeilah (1990) stresses the socio-cultural and aesthetic functions of home
gardens, among which are magical and religious values, status symbols, role
for socialisation and plants being used as weather indicators. Ritual practices
associated with crop management are reported by Alvarez-Buylla et al. (1989)
in southeast Mexico.
In a detailed survey of home gardens in Bangladesh, Abedin and Quddus
(1990) found that the first piece of land developed by functionally landless
farmers (i.e. owing less than 0.2 ha of land) is for a home garden. The same
study showed that farmers never sell their home gardens before their
croplands.

Impact on other systems

Descriptor 13:.forests and reserves (other trees in general). Hypothesis: in coun-

tries with abundant home gardens, there are less encroachments into other
areas for fuel wood collection.
Indicators: - - amount of fuel wood collected from home garden and time
spent in fuel wood collection.
Simon ( 1981 ) noted that 51-90% of the fuel wood collected in java is ob-
tained from home gardens. Byron (1984) in Singh (1987) noted that 70% of
the sawlogs and 90% of harvested fuel wood and bamboos in Bangladesh come
from the homesteads.

Descriptor 14: wildlife. Hypothesis: the complex vegetation structure of home

gardens, microclimatic conditions and recycling processes allow them to har-
bour a number of wild plants and animals. This, nevertheless, can have neg-
ative effects in terms of crop preOation.
Indicator: - - density of wildlife species in home gardens per hectare. The
abundance of spontaneously growing plants in home gardens is mentioned in
a number of references, and it is common that such plants yield secondary
202 E. TORQUEBIAU

useful products and have service roles. In Java, Karyono et al. (1978) in
Christanty ( 1981 ) found 121 different bird species in home gardens in four
villages, out of which 15 were endangered species.

Descriptor 15: rural industries. Hypothesis: the diversity of products from

home gardens provides numerous opportunities for the development of cot-
tage and rural industries. This creates jobs and off-farm employment and
marketing opportunities for different products.
Indicators: - - rural industries and their sources of raw materials. Achuthan
Nair and Sreedharan ( 1986 ) observed in India that there are many opportun-
ities for rural employment (cottage and canning industries) in areas with home
gardens in Kerala.

HOME GARDENS AS AGROFORESTRY SYSTEMS

In agroforestry, there should be positive ecological and economic interac-

tions between the trees and the other agricultural components. If such inter-
actions are properly achieved, agroforestry can combine production and con-
servation objectives, such a combination being the very essence of
sustainability. However, all agricultural research has in some way tried to
control or minimise the risk and uncertainty effects of bio-physical factors,
which is precisely the greatest explicit strength ofagroforestry (Avila, 1989).
Because of the intimate association of trees, crops and animals in home gar-
dens, they qualify a~ a good example of agroforestry, though many of the in-
teractions taking place in them are poorly understood. The different indica-
tors found in the literature which confirm hypotheses about the sustainability
of home gardens show that, complex as they are, home gardens do have a wide
range of sustainability attributes.
These attributes can be summarised as belonging to the following main cat-
egories: (a) soil conservation, both the control of erosion and the mainte-
nance of fertility (see Young, 1989), (b) modification of the microclimate,
(c) diversification and distribution of production more uniformly over the
year, (d) use of endogenous inputs, (e) management flexibility, ( f ) sociolog-
ical roles, and (g) impact on other systems.
Much has yet to be done to improve home garden management in areas
such as improved shade tolerance of different crops, successional stages of
home garden development, increased yields, fertiliE, management, etc. It
would be useful if future research on home gardens could focus on the above-
mentioned areas, as well as on establishment methods for home gardens. This
is because the usual claim that home gardens are sustainable concerns a stage
when they are fully developed, i.e. after several years of existence. There is,
however, a production gap due to the long establishment and gestation period
in home garden development, before large size trees are fully productive. Dif-
ARE TROPICAL AGROFORESTRY HOME GARDENS SUSTAINABLE? 203

ferent space and time combinations of trees and crops for home gardens have
to be found for different agro-ecological and socio-economic conditions. The
key for such combinations probably lies in the appropriate mixture of crops
both annual and perennial) which are fast or slow growing, small or large
and shade tolerant or light demanding. The optimal crop mix for given nutri-
tonal needs is also a topic of importance (Marten, 1990). Recent successful
examples of home garden establishment can be. found in the transmigration
sites in Sumatra, Indonesia (Suriamihardja, 1981 ) or in settlement schemes
for farmers encroaching upon forest land in Bangladesh (Torquebiau and
Abedin, 1991 ).
The role of the animal component in home g,~rdens has also to be carefully
studied, e.g. for fertility management, recycling of nutrients and wastes via
compost, ponds and fish production, etc. Finally, home gardens can be pro-
moted if associated rural industries using raw materials from home gardens
are also developed.
As can be suggested by the comparison between Figs. l and 2, home gar-
dens can exist in a variety of hip-physical conditions. Although they look
extremely different in the two contrasting cases shown in these pictures, home

Fig. 1. A home garden in Sungai Samba, Kalimantan (Indonesian Borneo). Annual rainfall:
about 4000 mm. Conspicuous tree species are: Durio zibethinus, Mang~fera indica, Cocos nuci.
fera, Lansium domesticum, and Citrus sp.
204 E. TORQUEBIAU

Fig. 2. A home garden ('jardin de concession ' or 'jardin de case') in Niger, near Tillab~ri.
Annual rainfall: about 400 ram. Conspicuous tree species are: Eucalyptus camaldulensis and
Prosopisjuliflora.

gardens always display a strong microclimatic difference in relation to their

external environment. Such a difference arises from the intensive manage-
ment of home gardens which is possible because of their small size and the
use of day-to-day family labour. It can hence be suggested that home gardens,
as eminently sustainable agroforestry systems, be promoted for a wide range
of bio-physical conditions and not only for the humid tropics where they are
presently prevalent (see e.g. Nair, 1989).

ACKNOWLEDGEMENTS

The home garden bibliographic review and analysis which led to this article
was done within the activities of an ICRAF Taskforce on Agroforestry and
Sustainability which has been operative since 1989. I would like to thank my
colleagues in the taskforce for the strong intellectual stimulus provided to me,
especially Marcelino Avila for nis advice on home garden system analysis and
sustainability indicators and Peter Huxley for repeated helpful discussions.
Comments of an anonymous referee greatly helped improve the paper.
ARE TROPICAL AGROFORESTRY HOME GARDENS SUSTAINABLE? 205

REFERENCES

Abdoellah, O.S., 1990. Homegardens in Java and their future development. In: K. Landauer
and M. Brazil (Editors), Tropical Home Gardens. United Nations University Press, Tokyo,
pp. 69-79.
Abdoellah, O.S. and Marten, G.M., 1986. The complementary roles of homegardens, upland
fields and rice fields for meeting nutritional feed in Java. In: G.G. Marten (Editor), Tradi-
tional Agriculture in South East Asia. Westview Press, Boulder, CO, pp. 295-325.
Abdul Khader, K.B., 1982. Inter and mixed cropping in Areca gardens. Indian Farming, De-
cember, pp. 20-23.
Abedin, M.Z., Lai, C.K. and Ali, M.O. (Editors), 1990. Homestead Plantation and Agrofores-
try in Bangladesh. Joydebpur, Bangladesh: BARi, Bangkok: FAO, and Dhaka: Winrock in-
ternational, 170 pp.
Abedin, M.Z. and Quddus, M.A., 1990. Household fuel situation, homegardens and agrofores-
try practices at six agro-ecologically different locations of Bangladesh. In: M.Z. Abedin, C.K.
Lai and M.O. Ali (Editors), Homestead Plantation and Agroforestry in Bangladesh. Joydeb-
put, Bangladesh: BARI, Bangkok: FAO, and Dhaka: Winrock International, pp. 19-53.
Achuthan Nair, M. and Sreedharan, C., 1986. Agroforestry farming systems in the homesteads
of Kerala, Southern India. Agrofor. Syst., 4: 339-363.
Ahmad, H., Martadihardja, A. and Suharto, 1980. Social and cultural aspects of the homegar-
den. In: J.L. Furtado (Editor), Tropical Ecology and Development. Proceedings Fifth Inter-
national Symposium of Tropical Ecology, International Sociely of Tropical Ecology, Kuala
Lumpur, Malaysia, pp. 453-458.
Alam, M.S., Haque, M.F., Abedin, M.Z. and Aktar, S., 1990. Homestead trees and household
fuel uses in and around the FSR site, Jessore. In: M.Z. Abedin, C.K. Lai and M.O. Ali (Edi-
tors), Homestead Plantation and Agroforestry in Bangladesh. BARI, Joydebpur, Bangla-
desh: FAO, Bangkok and Winrock International, Dhaka, pp. 106-119.
Alvarez-Buylla Roces, M.E., Lazos Chavera, E. and Garcia-Barrios, J.R., 1989. Homegardens
of a humid tropical region in Southeast Mexico: an example of an agroforestry cropping
system in a recently established community. Agrofor. Syst., 8:133-156.
Anderson, E., 1950. An Indian garden at Santa Lucia, Guatemala. Ceiba, I: 97-103.
Anderson, J.N., 1979. Traditional homegardens in South East Asia: A prolegomenon for second
generation research. In: Proceedings Fifth International Symposium on Tropical Ecology,
International Society of Tropical Ecology, Kuala Lumpur, Malaysia, pp. 1-36.
Avila, M., 1989. Sustainability and agroforestry. In: International Center for Research in Agro-
forestry (Editors), Viewpoints and Issues on Agroforestry and Sustainability. ICRAF, Nai-
robi, 115 pp. (mimeo).
Bahri, S., 1984. Piantes utiles de sous-bois, une perspective en agroforesterie tropicale. DEA,
University of Montpellier, France, 52 pp.
Balasubramanian, V. and Edit, A., 1986. The role of agroforestry in the farming systems in
Rwanda with special reference to the Bugesera-Migongo region. Agrofor. Syst., 4:271-289.
Baldy, C., 1963. Cultures associ6es et prczluetivit6 de reau. Ann. Agron., 14 (4): 489-534.
Bompam, J.M., 1986. Arboriculture fruiti6re en Indon6sie occidentale: traditions et perspec-
tives. Fruits, 41 (9): 531-551.
Bompard, J.M., DucatiUion, C., Hecketsweiler, P. and rv,~.hon, G., 1980. A traditional agricul-
tural system: village forest gardens in West Java. DEA, University of Montpellier, France,
136 PI~ plus figures.
Brierley, J.S., 1985. West Indian kitchen gardens: a historical perspective with current insights
from Grenada. Food Nutr. Bull., 7 (3): 52-60.
Brown, B.J., Hanson, M.E., Liverman, D.M. and Merideth, Jr., R.W., 1987. Global sustainabil-
ity: toward definition. Environ. Manage., I l (6): 713-719.
/-,r~,~
.uu E. TORQUEBIAU

Christanty, L., 1981. An ecosystem analysis of West-Javanese homegardens. Environment and

Policy Institute, East-West Center, Honolulu, Hawaii, 15 pp.
Christanty, L., 1990. Homegardens in tropical Asia, with special reference to Indonesia. In: K.
Landauer and M. Brazil (Editors), Tropical Home Gardens. United Nations University Press,
Tokyo, pp. 9-20.
Christanty, L., Abdoellah, O.L., Marten, G.G. and Iskandar, J., 1986. Traditional agroforestry
in West Java: the Pekarangan (homegarden) and Kebun-Talun (annual-perrennial rota-
tion) cropping systems. In: G.G. Martens (Editor), Traditional Agriculture in South East
Asia. Westview Press, Boulder, CO,pp. 132- i 58.
Committee on Agricultural Sustainability for Developing Countries, 1987. The transition to
sustainable agriculture: an agenda for AID. AID, Washington, DC, 29 pp.
Falanruw, M.V.C., 1990. The food production system of the Yap Islands. In: K. Landauer and
M. Brazil (Editors), Tropical Home Gardens. United Nations University Press, Tokyo, pp.
94-104.
FernanO,es, E.C.M. and Nair, P.K.R., 1986. An evaluation of the structure and function of trop-
ical homegardens. Agric. Syst., 21: 279-310.
Hussain, M.S., Abedin, M.Z., Quddus, M.A., Monowar Hossairi~.M., Banu, T., Ara, S. and
Ahmed, D., 1988. Women's contribution to homestead agricultural production systems in
Bangladesh. Bangladesh Academy for R.-al Development, Dhaka, 344 pp.
Karyono, 1990. Home gardens in Java: their structure and function. In: K. Landauer and M.
Brazil (Editors), Tropical Home Gardens. United Nations University Press, Tokyo, pp. 138-
146.
Landauer, K. and Brazil, M. (Editors), 1990. Tropical Homegardens. United Nations Univer-
sity, Tokyo, 257 pp.
Leiva, J.M. and L6pez, J., 1985. Los sistemas agroforestales de la cuenca del rio Polochic: Com-
posici6n y caracteristicas. Tikalia (Guatemala), I, 2: 48-84.
Lundgren, L., 1980. Comparison of surface runoffand sod loss from plots in forests and small
scale agriculture in the Usambara Mounts of Tanzania. Department of Physical Geography,
University of Stockholm, 97 pp.
Luu, P., 1989. Les syst6mes agroforestiers de la zone humide sri lankaise: agriculture du pass6
ou d'avenir? Th6se de Doctorat, Universit6 des Sciences et Techniques du Languedoc, Mont-
pellier, France, 251 pp.
Marten, G.G., 1990. A nutritional calculus for home garden design: case-study from West Java.
In: K. Landauer and M. Brazil (Editors), Tropical Homegardens. United Nations Univer-
sity Press, Tokyo, pp. 147-168.
Michon, G., i 983. Village forest gardens in West Java. In: P.A. Huxley (Editor), Plant Research
and Agroforestry. International Centre for Research in Agroforestry. Nairobi, pp. 13-24.
Michon, G,, Bompard, J.M., Hecketsweiler, P. and Ducatillion, C., 1983. Tropical forest archi-
tectural analysis as applied to agroforests in the humid tropics: the example of traditional
village agroforests in West Java. Agrofor. Syst., I (2): 118-129.
Michon, G. and Mary, F., 1990. Transforming traditional homegardem~ and related systems in
West Java (Bogor) and West Sumatra (Maninjau). In: K. Landauer and M. Brazil (Edi-
tors), Tropical Home Gardens. United Nations University Press, Tokyo, pp. 169-185.
Nair, P.K.R. and Balakrishnan, T.K., 1977. Ecoclimate of a coconut plus c,acao crop combina-
tion on the West coast of India. Agric. Meteorol., 18: 455-462.
Nair, P.K.R., 1989. Ecological spread of major agroforestry systems. In: P.K.R. Nair (Editor),
Agroforestry Systems in the Tropics. Kluwer, Dordrecht, pp. 63-84.
Nifiez, V., 1985. Working at half-potential: constructive analysis of homegarden programme in
the Lima slums with suggestions for an alternative approach. Food Nutr. Bull., 7 (3): 6-13.
Nii~ez, V., 1987, Household gardens: theoretical and policy considerations. Agric. Syst., 23: 167-
186.
Simon, H., 1981. Wood production and consumption from non-forest areas on Java, Indonesia.
ARE TROPICAL AGROFORESTRY HOME GARDENS SUSTAINABLE? 207

In: K.F. Wiersum (Editor), Observations on Agroforestry on Java, Indonesia. Forestry Fac-
ulty, Gadjah Mada University, Yogyakarta, Indonesia, pp. 56-61.
Singh, G.B., 1987. Agroforestry in the Indian subcontinent: past, present and future. In: H.A.
Steppler and P.K.R. Nair (Editors), Agroforestry, a Decade of Development. International
Centre for Research in Agroforestry, Nairobi, pp. 117-138.
Soemarwoto, O, 1987. Homegardens: a traditional agrotbrestry system with a promising fiJture.
In: H.A. Steppler and P.K.R. Nair (Editor), Agroforestry, a Decade of Development. Inter-
national Centre for Research in Agroforestry~ Nairobi, pp. 157-170.
Sommers, P., 1982. The UNICEF homegarden handbook for people promoting mixed garden-
ing in the humid tropics. UNICEF, New York, 55 pp.
Suriamihardja, S., 1981. Kemungkinan pengembangan agro-perhutanan di pemukiman trans-
migrasi. (Feasibility of agro-forestry development in transmigration areas). Badan Peneli-
tian dan Pengembangan Pertanian, No. 34/1981, Bogor, 47 pp. (mimeo) (in Indonesian,
with English summary).
Thaman, R.R., 1990. Mixed home gardening in the Pacific Islands: present status and future
prospects. In: K. Landauer and M. Brazil (Editors), Tropical Home Gardens. United Na-
tions University Press, Tokyo, pp. 41-65.
Torquebiau, E., 1984. Man-made Dipterocarp forest in Sumatra. Agrofor. SysL, 2 (2): 103-
127.
Torquebiau, E. and Abedin, M.Z. (Editors), 199 I. Agroforestry technologies for an encroached
forest area in Dhaka Division. Bangladesh. Forest Department and BARC, Dhaka, Interna-
tional Centre for Research in Agroforestry, Nairobi, 95 pp.
Vasey, D.E., 1990. On estimating the net social and economic value of urban home gardens. In:
K. Landauer and M. Brazil (Editors), Tropical Home Gardens. Tokyo: United Nations Uni-
versity Press, Tokyo, pp. 203-213.
Wiersum, K.F., 1985. Trees in agriculture and livestock development. In: IX World Forestry,
Congress, Theme !.3, Forestry Plantations. July 1985, Mexico City, 12 pp.
Wiersum, K.F., 1988. Surface erosion in agroforestry systems. In: K.F. Viersum (Editor),
Viewpoints on Agroforestry. Second, renewed edition. "Hinkeloord', Agricultural Univer-
sity, Wageningen, pp. 127-144.
Young, A., 1989. Agroforestry for soil conservation. CAB lnternatic, nal, Wallingford, UK, and
International Center for Research in Agroforestry, Nairobi, 276 pp.