Low External Input Technologies f or Livelihood Improvement i n Subsistence Agriculture

LOW EXTERNAL INPUT TECHNOLOGIES FOR LIVELIHOOD IMPROVEMENT IN SUBSISTENCE AGRICULTURE Anil Graves,1 Robin Matthews1 and Kevin Waldie2 1

Institute of Water and Environment, Cranfield University, Silsoe, Bedfordshire MK45 4DT, United Kingdom 2 International and Rural Development Department, University of Reading, Reading RG6 6AH, United Kingdom

I. Introduction II. The Technologies A. Intercropping B. Alley Cropping C. Cover Crops and Green Manures D. Biomass Transfer Techniques E. Compost F. Animal Manure G. Improved Fallows III. Generic Issues A. Soil Fertility Management B. Socio-economic Issues IV. Discussion A. Integrated Nutrient Management B. A Systems Perspective C. Modelling V. Concluding Remarks Acknowledgements References

I. INTRODUCTION The global population is currently predicted to reach 9.3 billion people by the year 2050, a 50% increase over the current level, after which it will level off due to falling fertility rates and family sizes (UNPD, 2001). Of these, 84% will live 473 Advances in Agronomy, Volume 82 Copyright q 2004 by Academic Press. All rights of reproduction in any form reserved 0065-2113/03 $35.00

474

A. GRAVES, R. MATTHEWS AND K. WALDIE

in countries that are currently classified as being in the developing world (Pretty, 1999). This rise in population, together with a desire for a wider variety in diet brought about by greater purchasing power through a steady improvement in incomes, is predicted to increase food demand over the period 1990 – 2050 by 2.4 times in Asia, 1.9 times in Latin America and the Caribbean, and 5-fold in Africa (FAO, 1996). Moreover, the increase in production required to meet this demand will need to be achieved with less water, less labour, and less land, and without adversely affecting the environment (Dowling et al., 1998). How this is to be achieved is the topic of some debate (e.g., Pretty, 1999; Crosson and Anderson, 2002). Over the 40 years since 1960, the global population has doubled; despite this, food production has more than kept pace, resulting in a 24% increase in per capita world food production and a 40% reduction in food prices in real terms (although these figures do mask some striking imbalances — per capita food production has fallen 20% in Africa, for example). The total number of undernourished people in the world has also fallen significantly over the same period. This has been largely achieved by the use of “Green Revolution” technologies, i.e., high-yielding cereal varieties, together with high levels of inputs such as water from irrigation systems, fertiliser to provide the nutrients needed by the varieties, and pesticides to control any associated weeds, pests and diseases. These technologies generally need a relatively high capital investment, either by, or on the part of farmers, and also need a well-functioning economic and physical infrastructure for effective implementation. However, an estimated 30 –35% of the world’s population (i.e., 1.9 – 2.1 billion people) do not have access to such infrastructures, are remote from markets, practice subsistence agriculture on marginal soils, and lack access to knowledge on how to improve their situation (Pretty, 1999). One school of thought is that a similar high external input agriculture (HEIA) approach as used in the last 40 years can also be used to address the demand for food in the next 50 years by improving the productivity of this group of subsistence farmers, perhaps using new emerging technologies such as genetic modification (e.g., Crosson and Anderson, 2002). A second school of thought is that such an approach is not sustainable, and moreover, is damaging to the environment as the inputs of fertilisers and chemicals accumulate in neighbouring ecosystems. Thus, technologies using low levels of external inputs readily available either on-farm or from nearby off-farm sources are seen by some experts as more appropriate and sustainable (Pretty, 1995). This approach, often referred to as low external input agriculture (LEIA), emphasises the use of techniques that integrate natural processes such as nutrient cycling, biological nitrogen fixation (BNF), soil regeneration and natural enemies of pests, into food production processes (Pieri, 1995; Snapp et al., 1998). Efforts are also made to minimise losses from the system, such as by leaching or removal of crop residues.

LOW EXTERNAL INPUT TECHNOLOGIES

475

The use of non-renewable inputs such as pesticides and fertilisers that can damage the environment or harm the health of farmers and consumers is also minimised, and more emphasis is placed on the use of such techniques as, for example, intercropping, agroforestry, cover crops, or animal manure. Usually, but not always, such technologies are more labour-intensive than the HEIA approach (Deugd et al., 1998). In many cases, LEIA technologies are not new, but are variations of those practised by farmers for generations, who have sought to make use of resources such as vegetation or animal manure that have always been ready to hand. Wolf (1986) has estimated that about 1.4 billion people (25% of the world’s population), depend on this type of agriculture for their livelihood. Thus, the heart of the debate is not about whether either approach “works,” as clearly both do, and have done, under the appropriate conditions and according to their own criteria. Rather, the central question concerns which approach can best address the future demand for food production while protecting the environment as much as possible. Within this general debate, more specific questions relate to whether LEIA technologies really have the capability to maintain or increase productivity per unit area above current levels (e.g., Crosson and Anderson, 2002). Certainly, there is evidence to suggest that the relative rate of increase in crop yields through the use of Green Revolution technologies is slowing (Mann, 1997), although Crosson and Anderson (2002) argue that this is more likely due to the practice of quoting annual percentage increases of a constantly increasing baseline rather than absolute annual growth. Proponents of LEIA technologies often claim that the reliance on local sources of inputs is more sustainable, but the analysis of De Jager et al. (2001) suggests there is little difference between the two approaches in this respect, with both mining similar quantities of soil nutrients to generate farm income. However, despite the continuing debate on the relative performance of the two approaches, there are few studies that compare yields and production under the same soil and climatic conditions and over wide areas. With LEIA technologies in particular, there is little in the literature on the issues that need to be faced in scaling up production from plot level to supplying inputs and meeting food demand on a larger scale. The purpose of this review is to examine research that has been conducted in recent years on a number of LEIA technologies in the context of subsistence agriculture — firstly, to determine our current state of knowledge, and secondly, to assess their potential to improve livelihoods of subsistence farmers and contribute to long-term sustainability of the resource base. The review was prompted by concerns within the United Kingdom Department for International Development (DFID) that, despite considerable research on, and dissemination of, a wide range of LEIA technologies, there had been little clear evidence of widespread uptake by farmers. Although this could partly be ascribed to inadequate attention to promotion pathways and dissemination by research centres, there was some concern that more fundamental reasons may be

476

A. GRAVES, R. MATTHEWS AND K. WALDIE

responsible. In this review, therefore, we have attempted to step back and take a wider view of some of these technologies, examining both their biophysical and socio-economic aspects to try understand the reasons for their variable uptake by farmers. We have focused on those technologies designed to address problems of soil fertility and weed control; these include intercropping, alley cropping, cover cropping and green manuring, biomass transfer, compost, animal manure, and improved and enriched fallows.

II. THE TECHNOLOGIES A. INTERCROPPING Intercropping is the growing of two or more crops on the same piece of land within the same year. Various forms of intercropping have been a central feature of many tropical agricultural systems for centuries. Vandermeer (1989) has proposed that intercropping can be divided into three general categories — full, relay and sequential intercropping — depending on the extent of physical association between the crops. Full intercropping involves complete association between crops planted at the same time, while relay cropping involves only partial association, in which a second crop is planted into an already standing crop before it is harvested. Sequential intercropping, where there is no physical association, is the extreme case where two crops are grown on the same land in the same year but not at the same time. Cover crops are a special case of intercropping and are discussed in more detail in Section II.C; in this section, we discuss cases where the intercrop components are both food crops. The main advantages of intercropping are in reducing the risk of total crop failure, and in product diversification — food crops are often mixed with cash crops to help ensure both subsistence and disposable income (Vandermeer, 1989; Singh and Jodha, 1990). There is some evidence to suggest that in intercropping systems, the microclimate surrounding the lower crop is more conducive to plant growth than in a sole crop (Matthews et al., 1991), and that an intercrop is more efficient at using resources such as light and water (Azam-Ali et al., 1990). Cereals and legumes are often mixed, but probably more for dietary reasons than for any beneficial effect from the nitrogen-fixing ability of the legumes. In Zimbabwe, for example, farmers intercrop sorghum with cowpeas, pumpkins, cucumbers and watermelon to provide nutritional and livelihood benefits (Chivasa et al., 2000). Thus, any soil fertility benefits that can be obtained by intercropping leguminous grain crops with other food crops should probably be seen as a useful spin-off rather than the main purpose of the practice. Moreover, main crop yields can even be reduced by intercropping techniques, both as a result of loss of land to the legume, and also to competition for resources

LOW EXTERNAL INPUT TECHNOLOGIES

477

(Vandermeer, 1989; Snapp et al., 1998). In the long-term, cereal/legume intercrops are still likely to require fertilisers for the provision of phosphorus (P), potassium (K) and micro-nutrients in order to maintain satisfactory yields (Coultas et al., 1996; Kumwenda et al., 1997a). Nevertheless, yield advantages through intercropping are well documented. For example, Rao (2000) found that, compared to sole maize, maize intercropped with pigeonpea (Cajanus cajan) in Kenya yielded 24% more and was 49% more profitable, even though the pigeonpea was affected by pests and diseases. In an interesting variation, sequential intercropping of rose (Rosa damascena) with potatoes, maize and cowpea greatly increased the land equivalent ratio (LER) and provided large economic gains (Yaseen et al., 2001). Similarly, positive effects on soil fertility improvement have been observed. Kumara Rao et al. (1981) estimated that leaf abscission during the growth of a pigeonpea intercrop was equivalent to the addition of between 10 and 40 kg N ha21. The root system of pigeonpea may also recycle N from deeper layers, and in some areas, the build-up of sub-surface nitrates at about 1 –3 m has been observed (Farrell et al., 1996; Hartemink et al., 1996). Morris et al. (1990) observed N transfer from arrow-leaf clover to rye grass, and suggested that in mixed stands of legumes and non-legumes, direct transfer of N during the growing season was possible, although this was likely to be 10% or less of the total N fixed. It is also possible that growing non-legumes with legumes encourages legumes to respond by fixing more N than they might do in a pure stand, so long as the legumes dominate the mixture (Marschner, 1995). Snapp et al. (1998) considered soybean (Glycine max), pigeon pea (C. cajan), groundnuts (Arachis hypogaea), dolichos bean (Dolichos lablab), and cowpea to be among the most promising grain legumes in southern Africa for both food provision and fertility enhancement. Although grain legume intercrops can often help to increase the resource use efficiency and stabilise yields of the main crop under optimal plant growing conditions, this is not always the case. In India, Indonesia, and the Philippines, Ali (1999) found that although intercropping could help to increase the yield of rice, it also increased the variability of yield. Many green manure crops have been selected on the assumption that maximum production is desirable. However, yield stability under adverse conditions may be more important to many farmers than high productivity under good conditions. Ironically, it is usually under adverse conditions that intercrop competition is most intense, and it is during these conditions that the farmer can least afford the technology to fail. Intercropping is most likely to be practised on small farms, in areas where land is scarce, forcing the simultaneous production of different crops on the same area of land. For example, Ali (1999), in a survey of data from India, Nepal and the Philippines, found that the attractiveness of intercropping increased as land and labour costs grew. Lower rainfall and/or a unimodal distribution of rain may encourage intercropping as farmers try to maximise their use of water, although,

478

A. GRAVES, R. MATTHEWS AND K. WALDIE

in the extreme, this can result in competition for a scarce resource. Mixedcropping techniques are also more likely to be used by farmers relying on handheld implements for tillage (Ruthenberg, 1980). The need for simultaneous production of different food crops and/or cash crops can also encourage intercropping. Relatively better-off farmers with large farms are less reliant on intercropping, being able to fallow and/or control production with other inputs such as water and inorganic fertilisers. The Machobane farming system in Lesotho (IIRR, 1998) is an example of a system incorporating intercropping. Developed in the 1950s by an agronomist, James Machobane, and based on experiments on his own farm, the approach is a complex, integrated farming system designed to improve the productivity of small-scale mountain farms in Lesotho. Based on 0.4 ha (1 acre), an application of 7.5 t FW ha21 of a mixture of animal manure and wood ash (from household ash) is made each year, with the proportion of each depending on soil type. Enough for the 0.4 ha can generally be met from household waste (1 – 2 t year21 of animal manure, 2 t year21 of ash, (Pantanali, 1996)). Wheat, peas, and possibly potatoes as a cash crop, are planted as intercrops in April –May for harvesting the following January – March, and summer crops such as maize, beans, sorghum, and possibly pumpkins and water melons, are planted in August –October for harvesting in November –December (Fig. 1). Crop residues are left in the field to allow humus to build up, and the field is ploughed only once every five years. The intercropping/relay-cropping pattern allows food crops to be produced almost all the year round, and there is always some crop cover throughout the year so that erosion, a major problem in Lesotho, is minimised. Despite the crop cover, weeding is essential, and represents a major labour input into the system. Some of the crops (e.g., pumpkins) help to reduce pests, and the use of chemical pesticides is discouraged. Overall, labour inputs are high, and perhaps reflecting this, annual productivity is three times higher than the traditional system, allowing a household of five people to be self-sufficient on 0.4 ha of land (Pretty, 1999). Moreover, the potato crop is a source of cash, and income fluctuations over the year are lower through less reliance on a single crop. Farmers also claim a better resistance of crops to drought. Promotion of the approach emphasises self-reliance, appreciation of the resource base, readiness for hard work, learning by doing, and a duty to help neighbours, on the part of farmers, and lack of success of the system is sometimes blamed on non-compliance with some of those conditions (Pantanali, 1996). Due to the high labour inputs (peak labour demand estimated at 14 days (0.4 ha)21 month21), critics of the system regard it as little more than “gardening,” with limited relevance to Lesotho’s major agrarian problems. Unfortunately, little data appear to have yet been collected to measure the impact of the system in terms of production, gross income and net returns to labour, either compared to traditional cropping methods, or to recommended

LOW EXTERNAL INPUT TECHNOLOGIES

479

Figure 1 Crop calendar for the Machobane farming system in Lesotho (from IIRR, 1998).

“modern” improved technology practices based on monocropping, mechanical power, chemical fertilisers, and pesticides (Pantanali, 1996). Long-term sustainability of the system, however, depends on the production of animal manure, and hence the availability of pasture, which fortunately is available in the mountains. As with many LEIA systems, therefore, there is a reliance on nutrients collected and concentrated from a much wider area. More manure could be made available if less was used for fuel, but in this case, fuelwood would have to be grown, which again could be done on the land saved through intensification of production. Thus, through careful integration of crops, livestock and trees, the long-term sustainability of the system seems possible (Pantanali, 1996). In relation to its applicability to other areas, the system’s economic sustainability rests on being able to grow crops all year round, which will not be feasible in areas with a pronounced dry season unless irrigation is available. Even in Lesotho, the system cannot be practised in all areas due to severe winter conditions with snowfall preventing the growth of many of the crops (Pantanali, 1996).

B. ALLEY CROPPING Alley cropping is an agroforestry practice developed in the 1970s at the International Institute for Tropical Agriculture in Nigeria (Kang et al., 1981), in

480

A. GRAVES, R. MATTHEWS AND K. WALDIE

which hedgerows of trees and shrubs are established and annual crops are cultivated in the alleys between the hedgerows. The hedgerows are pruned before planting the crop and periodically while it is growing to prevent shading, with the prunings being applied to the soil as green manure and/or mulch. Between cropping cycles, hedgerows are usually allowed to grow without pruning. It was originally hoped that by incorporating fast-growing nitrogen-fixing woody perennials with crops, their ability to cycle nutrients, suppress weeds, and reduce erosion would create soil conditions similar to those in the fallow phase of shifting cultivation. In this way, the cropping and fallow phases could take place simultaneously on the same land, allowing the land to be cropped for an extended period when long fallow periods are not feasible under the particular socioeconomic conditions. Researchers saw the technology as the combination of farmers’ accumulated traditional wisdom with the efficiency of modern science (Kang, 1993). Initial results from on-station experiments were promising. In Nigeria, for example, prunings from Leucaena leucocephala increased maize grain yields from 1.9 to 3.5 t ha21 (Kang et al., 1981), while a Gliricidia sepium alley system on a degraded soil increased maize yields from 1.74 to 2.42 t ha21 (Atta-Krah and Sumberg, 1988). Increases in yields of banana were obtained when alleycropped with Enterolobium cyclocarpum, and of cowpea when alley-cropped with Enterolobium cyclocarpum and Dialium guianense (Oko et al., 2000). In the fourth and fifth years of an alley-crop trial in Burundi, Calliandra calothyrsus increased maize yields by 29 – 63%; Leucaena diversifolia by 27 –43%, and Senna spectabilis by 24 –38% (Akyeampong, 1999). However, these results were obtained in humid regions on soils of high base status. Results from semi-arid regions were less positive. Yields of sorghum, castor and cowpea were found to be lower when alley-cropped with Leucaena than when grown alone (Singh et al., 1989), with the magnitude of the yield depression correlating with distance from the hedgerows. This was attributed to competition between trees and crops for water. Similar depressions of yield under alley cropping were found in Peru (Szott, 1987) and Zambia (Matthews et al., 1992a,b), although, in the latter work, Leuceana was found to have a positive effect on maize yields if lime was applied first. In smallholder farms in southern Africa, little benefit was derived from grain legume intercrops, particularly in adverse weather conditions (Mukurumbira, cited in Snapp et al., 1998). Hedgerow intercropping did not increase maize yields in below average rainfall years, indicating that competition by trees predominated over benefits to soil fertility (Snapp et al., 1998). Similarly, Vanlauwe et al. (2001) found reduced crop yields under alley cropping in the absence of mineral fertilizers. Competition, especially for light, is particularly acute with perennial legumes, which are larger than the main crop (Ong, 1994). In general, it appears that where resources are scarce, competition for resources such as water and nutrients is not offset by benefits to fertility (Manu et al., 1994; Rao et al., 1997). Root pruning

LOW EXTERNAL INPUT TECHNOLOGIES

481

has been found to reduce competition for water, but this adds to the labour requirement, and may affect the ability of the tree to access water and nutrients below the crop zone (Ong et al., 2002). In some cases, other negative effects of trees on the crop have been found. For example, Adeorike (2001) found that Inga edulis, Anthonatha macrophylla, and Dactyladenia barterri had allelotrophic effects on maize. In Kenya, decreasing the width between Leucaena hedgerows was found to increase the incidence of angular leaf spot (Phaeoisariopsis griseola) and anthracnose (Colletotrichum lindemuthianum) on beans (Phaseolus vulgaris), with the bean rows closest to the hedgerows the most affected (Koech and Whitbread, 2000). Microclimate changes caused by the hedgerows appeared to best explain the distribution of the diseases on the beans. In some cases, more bird damage has been reported in alley crops than in monocrops (e.g., Hoang Fagerstro¨m et al., 2001). Some control of weeds was achieved in alley crops if the hedgerow canopy was maintained during the fallow period. For example, in Nigeria, uncut hedgerows of Gliricidia and Leucaena decreased the shoot biomass of the weed Imperata cylindrica by about 80% (Anoka et al., 1991). Similarly, Yamoah et al. (1986b) found that unpruned hedgerows of Flemingia macrophylla, Gliricidia sepium, and Cassia siamea were able to reduce weed yields. Shifts in weed composition were also observed — Siaw et al. (1991), for example, reported a significant change towards more broadleaf weeds after alley cropping with Leucaena and Dactyladenia barter. In most alley-cropping systems, the weed suppression effect of the hedgerows probably has not been fully exploited, and further studies of the effect of different hedgerow species, fallowing and manipulation of cutting regimes may improve the effectiveness of the system in reducing weed infestation. Alley cropping does seem to have favourable effects on soil physical and chemical properties through the addition of large amounts of organic matter from the prunings. Levels of organic C, total N, extractable P, Mg and K, and pH have been shown to increase under alley cropping under a range of conditions (e.g., Kang et al., 1985; Lal, 1989b; Dalland et al., 1993). Similarly, lower bulk density and penetration resistance, and higher infiltration rate and pore volume fraction, were found under Leucaena alley crops in Zambia, which was ascribed to increased levels of soil organic matter (SOM) (Dalland et al., 1993). The magnitude of these effects, however, varied with hedgerow species and management — Leucaena prunings increased crop yields more than Flemingia congesta due to the faster release of nutrients as a result of a lower C/N ratio. There is also some evidence that nutrient recycling is enhanced by alley cropping so that the downward displacement of nutrients is reduced. Hauser (1990), for example, attributed higher concentrations of N, K, Ca and Mg in the surface soil than in the subsoil under Leuceana hedgerows to leaf litter fall and nutrient uptake by the trees from the subsoil. Between the rows, there were lower nutrient

482

A. GRAVES, R. MATTHEWS AND K. WALDIE

levels in the surface soil due to crop uptake and higher levels in the subsoil due to leaching. Alley cropping also seems to have been successful in reducing runoff and soil erosion on sloping land. For example, in Nigeria, Lal (1989a) found a reduction of 73 and 83% in soil erosion under alley crops of Gliricidia and Leucaena, respectively. In the Philippines, Paningbatan (1990) recorded soil erosion over a three-month period of 41 t ha21 with Desmanthus hedgerows and contour cultivation, and only 0.2 t ha21 with hedgerows, application of prunings as a mulch, and zero tillage, compared to 127 t ha21 in the control treatment. In northern Vietnam, contour planting of hedgerows on sloping lands reduced soil loss from 18 to 7.4 t ha21 year21 and also produced 2.5 –12 t DM ha21 year21 for green manure (Nguyen The et al., 2001). Farmers were aware that soil loss resulted from cultivating annual crops on sloping lands without adequate protection, but were constrained by lack of labour and capital (Brodd and Osanius, 2002). Initial uptake of alley cropping by farmers in Nigeria and Benin was good, but nearly half of those who adopted the technology subsequently abandoned it. In Nigeria, no new adoption occurred after 1990 (Douthwaite et al., 2002). Followup studies concluded that the initial enthusiasm shown by farmers was probably more related to the incentives offered by researchers (free establishment of the alley fields, weeding, provision of animals, animal vaccination, fertilisers and seed of improved crop varieties), and to the prestige of contact with international researchers, than to the characteristics of the technology itself (Whittome, 1994). The Nigerian farmers gave the high labour demand for establishment and management of the hedgerows, and incorporation of the biomass into the soil, as the main constraints. Other studies have shown similar results (e.g., Reynolds et al., 1991; David, 1995; Craswell et al., 1998), with many adopters specifically citing the labour required for pruning as being the most difficult aspect of alley cropping. Interestingly though, Hoang Fagerstro¨m et al. (2001) note no difference in labour requirements between a monocrop and Tephrosia alley-crop in Vietnam — the extra labour required for hedgerow management was balanced by a reduction in labour associated with crop husbandry, such as weeding. An abundance of available land has also been found to be a factor constraining uptake of alley cropping – Whittome (1994), in his study of farmer experiences in Nigeria and Benin, found that in most cases, land was still sufficiently abundant for them not to consider soil fertility decline as a problem, and therefore not to find alley cropping attractive. In Kenya, Swinkels and Frankel (1997) found that alley farming was most attractive in areas where the population density was high, farms were small, and labour was plentiful. Access to capital is another key factor. For example, Cenas et al. (1996) have shown that adoption may be higher where farmers have off-farm sources of income, relatively large farms, and were interested in cash cropping. This was not only because of the high labour costs involved in establishment of hedgerows (Nelson et al., 1996), but also because

LOW EXTERNAL INPUT TECHNOLOGIES

483

alley cropping alone could not maintain total soil fertility requirements and money had therefore to be spent on fertiliser for full benefits (Wendt et al., 1994). Moreover, benefits from alley cropping take some time to accrue (Carter, 1995), and resource-poor farmers may feel that these are not realised rapidly enough to meet their current needs and that the long-term benefits do not outweigh the more immediate costs of establishment (Nelson and Cramb, 1998). Indeed, traditional cropping practices often create greater net revenue than alley cropping over the first 4– 5 years (Nelson et al., 1996), and this may discourage the use of alley cropping despite greater long-term benefits (Nelson et al., 1998). Security of tenure and long-term access to land are important issues affecting uptake in some countries. Tenant farmers, for example, are unlikely to want to bear the full cost of the technique while the benefit is shared with the landlord (Nelson et al., 1998). Similarly, systems based on revolving cultivation of land amongst family members, short-term tenancy, and share cropping tenancy arrangements may have the same effect. Where farmers have long-term security of tenure over discrete areas of land, alley cropping may be more relevant (Carter, 1995). It is important to note that there are cases of farmers adapting the basic alley cropping practice to fit their own needs. For example, in the Philippines, farmers often increased alley spacing, planted single rather than double hedgerows to reduce planting density, and reduced trimming frequencies and mulch application (Garcia et al., 2002). Some farmers even used alternative tree species so that the hedgerow could be used for other purposes. These modifications may have reduced the value of alley cropping as a soil fertility-enhancing technique, but have allowed it to fit within the constraints of the farmer and to answer a wider set of needs. In some cases, there has even been an evolution from alley cropping into intercropping two crop species — in eastern Indonesia, for example, Harsono (1996) describes the replacement of hedgerows with strips of grain legumes such as soybean which were shown to increase net profits. In other cases, alley cropping has had some success for reasons other than soil fertility enhancement or erosion control. For example, farmers have planted hedgerows for the provision of poles, medicines, plants, fibre, fruit, and fuel (Cenas et al., 1996), while in the Amarasi district of Indonesia, Field et al. (1992) noted that Leucaena alley crops could provide fodder to allow farmers to develop intensive livestock systems. The return from the sale of cattle could be used for the purchase of food, and as farmers become less reliant on annual crops for subsistence, more appropriate perennial plant systems could be established on steeper land prone to erosion. These modifications by farmers illustrate an important point regarding adoption of alley cropping, or any technique for that matter — if it is to be successful, it must address a range of farmers’ requirements, some of which may not necessarily be related to the researchers’ original intentions, in this case, those of soil fertility enhancement or erosion control (Douthwaite et al., 2002).

484

A. GRAVES, R. MATTHEWS AND K. WALDIE

Alley cropping (and agroforestry technologies in general) must become more “versatile,” capable of meeting a range of needs in response to changes in the socio-economic circumstances of the farmer (Vosti et al., 1998). Gender issues are also important — adoption is more likely if it is able to meet the needs of both men (e.g., for poles or fodder) and women (e.g., for fibre, fuel-wood and mulch) (Rocheleau and Rocheleau, 1990).

C. COVER CROPS

AND

GREEN MANURES

A cover crop is a crop grown to provide soil cover to prevent erosion by wind and water, regardless of whether it is later incorporated. Green manuring involves the incorporation of a crop while it is still mainly green into the soil for the purpose of soil improvement. Cover crops and green manures are generally annual, biennial, or perennial herbaceous plants grown in a pure or mixed stand during all or part of the year, and as such can be seen as a special case of intercropping. In addition to providing ground cover and, in the case of a legume, producing N, they may also help suppress weeds and reduce insect pests and diseases. Catch crops are cover crops that have been planted specifically to reduce losses of nutrients by leaching following a main crop. To compete with weeds, which by definition are aggressive plants, cover crops need to have an appropriate canopy architecture. A spreading cover crop is more likely to suppress weeds than a cover crop with an erect habit. For example, in trials in rice systems, in the Ichilo Sara area of Bolivia, Pound et al. (1999) noted that the performance of Arachis pintoi as a cover crop was highly variable and rejected by farmers due to its inability to suppress weeds (especially Imperata contracta), largely as a result of poor growth and lack of full cover. Similarly, in Ghana, Jackson et al. (1999) found that C. cajan (pigeonpea) was slow-growing, low yielding, and incapable of suppressing weeds due to its poor ground coverage. However, even the use of aggressive cover crops that spread and shade well may fail to suppress the growth of certain shade-tolerant weed species. Pound et al. (1999) found that Mucuna pruriens and Calopogonium mucunoides could not suppress the growth of weed species such as Axonopus compressus, Cyperaceas, Panicum spp., Leersia spp., and shade-tolerant species such as Drymaria, Commelina and Talinum. The duration of the cover crop may also determine its effectiveness in controlling weed growth. If its duration is less than the period between harvest of the main crop and the planting of the next main crop, weeds may proliferate during the time gap, and nutrients may even be released for uptake by weeds rather than by the subsequent crop. For example, in experiments on winter cover crops grown in rotation with rice, Pound et al. (1999) found that certain cover crops with relatively short durations could actually increase the number of weeds in comparison with the traditional practice of leaving a winter fallow. In these

LOW EXTERNAL INPUT TECHNOLOGIES

485

cases, the weeds had time to multiply before the planting date of the next rice crop, possibly taking advantage of N that was fixed by the leguminous cover crops. Cover crops such as alfalfa, cowpea, carioca beans or Mucuna deeringiana had very short durations, and tended to produce comparatively little biomass as a result. Weed counts in these plots were substantially greater than for the traditional fallow. Species with a longer duration, such as Mucuna pruriens, C. cajan, and Canavalia ensiformis, continued growing up to the first planting of the rice, produced more biomass, and continued to suppress the growth of weeds over this time, although not significantly more than the traditional fallow treatment as far as grass and broadleaf weeds were concerned. Rice yields also showed no significant differences between different cover-crop treatments, despite the large variation in grass weed density (Pound et al., 1999). A cover crop with too long a duration may also cause problems. If the use of cover crops in the system is to be sustained, particularly in isolated areas, farmers need to be able to collect seed for the next season. If the crop fails to flower and seed before the next main crop planting, then seed must be obtained externally. Most legume genotypes appear to be adapted to quite narrow biophysical conditions, and, therefore, must be tailored to specific environments if they are to be successful. Keatinge et al. (1998) showed that Vicia faba, Vicia villosa ssp. dasycarpa, and Lupinus mutabilis would be suitable as autumn-sown cover crops across most of the mid-hills of Nepal if early sowing was possible. Vicia sativa and Trifolium resupinatum, on the other hand, were only likely to mature soon enough at lower elevations. Similar exercises were conducted for hillside regions in Bolivia (Wheeler et al., 1999) in which potential cover crops, not grown locally, were recommended for further trials, and also in Uganda (Keatinge et al., 1999). In some cases, the growth of cover crops grown in association with another crop has been found to be too aggressive, resulting in undue competition with the main crop for nutrients, space, water, and light. If left unchecked this can lead to complete domination of the main crop by the cover crop, i.e., the latter is beginning to behave like a weed itself. For example, Pound et al. (1999) found that Mucuna pruriens was rejected by some farmers in the Ichilo-Sara area of Bolivia because it dominated Bactris gasipaes, a local palm, and banana, limiting their growth and development. Similarly, in on-farm trials of a Calopogonium/ rice intercropping system, farmers found that Calopogonium tended to climb over the rice and cause it too lodge. This occurred especially when rice and Calopogonium were sown simultaneously or when long-duration rice varieties were used (Pound et al., 1999). Nevertheless, cover crops have had some success in addressing problems of soil fertility and weed control. It has been shown that short-term fallows of herbaceous crops such as Mucuna pruriens (velvet bean) and Stylosanthes hamata (Stylo) can help increase main crop yields compared with continuous cropping, and that weed densities can be reduced (Tarawali et al., 1999).

486

A. GRAVES, R. MATTHEWS AND K. WALDIE

Farmers seem to be well aware of these benefits (Buckles and Triomphe, 1999; Franzel, 1999), and because of this, the adoption of cover crops by farmers has been relatively widespread (Sanchez, 1999). In Benin, for example, the ability of Mucuna grown as a cover crop to suppress the weed Imperata cylindrica was discovered almost by accident by researchers with the Recherche Appliquee´ Milieu Ree´l (RAMR) Project, and was subsequently promoted amongst farmers by various formal and informal sector organisations. It was estimated that about 10,000 farmers tested Mucuna between 1988 – 1996 (Tarawali et al., 1999), although Elbasha et al. (1999) report that this was on only 1000 ha. Other estimates put the adoption figure as high as 100,000 farmers (Versteeg et al., 1998). An adoption study by RAMR showed that about 25% of farmers had adopted the technique (defined as having used it at least twice), whilst about 35% had rejected it despite still having an Imperata problem on their fields (Versteeg et al., 1998). Adopters cited the need to control Imperata infestations as the primary motivation for their using Mucuna, rather than soil fertility enhancement, although apparently some reported benefits from higher maize yields (3 – 4 t ha21 without application of nitrogen fertilizer and with less labour input for weeding, compared to 1.3 t ha21 without Mucuna (Pretty, 1999)). Non-adopters said that leaving the field unproductive during the minor season was a major disincentive, as well as the lack of a use for the grain produced by Mucuna, which is toxic (due to the presence of l-Dopa) unless treated properly (Versteeg et al., 1998). However, in reality, lack of a market was not always a problem, as demand for Mucuna seed grew as use of the technique spread. Interestingly, adoption was lower in areas where land was less scarce, although some farmers in these areas discovered that Mucuna made good fodder for livestock, and could also be used to suppress the parasitic weed Striga hermonthica. The benefit/cost analysis over a period of eight years indicated a ratio of 1.24 when Mucuna was included in the system, and 0.62 for the system without Mucuna, with the ratio as high as 3.56 if Mucuna seeds were sold (UNEP, 1999). Positive returns were achieved in the second year of establishment at both the farm and regional levels. It has been estimated that Mucuna grown as a cover crop can provide more than 100 kg N ha21 to a following maize crop, and, as such, its adoption throughout the Mono Province of Benin would result in savings of about 6.5 million kg of N, or about US$1.85 million year21. However, yearly analysis of the benefit/cost ratio indicated a declining trend over time, suggesting that addition of external inputs (probably P and K fertilizer) are required in order to achieve full sustainability (Pretty, 1999). Widespread adoption of Mucuna as a cover crop has also been evident in Honduras, where, without any extension support, it spread from farmer to farmer since the 1970s, when, due to population pressure on the plains, hillside areas were first used for agricultural production (Buckles and Triomphe, 1999). Traditional practice was to crop maize twice a year in the wet and dry seasons for

LOW EXTERNAL INPUT TECHNOLOGIES

487

two years, then return the field to fallow for four years, a system which was labour intensive (land preparation, planting, weed control), land extensive (due to the need to have additional plots to meet food requirements during the fallow), and capital intensive (due to outlay on herbicides and labour for weeding). In the new system, farmers plant Mucuna as a relay crop into the maize in the dry season so that it grows alone over the wet season, suppressing other weed species at the same time. At the beginning of the next dry season, the farmer slashes the Mucuna and sows the following maize crop directly into the decaying mulch, and so the cycle continues. The system gives benefits in terms of reduced labour (15 – 20% less) and increased yields after the second year — while the traditional system provided four harvests over six years from a single plot, the maize – Mucuna system produces six harvests with yields 50 –100% higher. On average, those who have adopted the Mucuna technique planted twice as much maize as those who did not. Despite this, the total amount of land occupied by their cropping system was less, as they no longer needed large areas to fallow, although, interestingly, overall deforestation rates continued to increase due to an influx of migrants into the area (Humphries, 1996). Experimental evidence indicated that the system was capable of maintaining soil N and OC, Ca, pH and P levels. This was largely achieved through a large biomass production of about 10– 12 t ha21, and large amounts of N (about 300 N kg ha21) being contributed through this biomass, although, of course, only a proportion of this represents a net addition to the system through nitrogen fixation. By 1992, 65% of farmers were using the maize – Mucuna system, with a further 19% having used it in the past (Buckles and Triomphe, 1999). However, more recently there has been a widespread decline in the number of farmers using the system; such that with abandonment running at 10% per year, only 39% of farmers were still using it by 1997. Neill and Lee (2001) provide interesting insights into why this has occurred. Their surveys showed that there was no single overriding factor involved. Firstly, a minimum farm size of 2– 3 ha is required to meet household food requirements during the wet season while some fields are under Mucuna, and farmers must have security of land tenure to adopt the system. However, over the last 30 years, there has been a steady decline in farm size and insecurity of tenure has increased. Secondly, the increase in extensive cattle production in the region has decreased the availability of land for rent and off-farm work, both necessary requirements for the smallest farmers if they are to adopt the maize –Mucuna system. Thirdly, improved road access in the region has made other alternatives, such as fruit-trees and off-farm work, more attractive than maize growing. Fourthly, the incursion of the noxious weed Rottboellia cochinchinensis into the area in the last few years has drastically reduced maize yields (by 50 –72%) by establishing in gaps within the Mucuna caused by farmers relying on it to regenerate itself rather than deliberately reseeding it at the beginning of each new season. The presence of Rottboellia has negated the two advantages of the maize – Mucuna system — that of lower

488

A. GRAVES, R. MATTHEWS AND K. WALDIE

labour requirements and higher productivity. Thus, it is clear that the sustainability of a technique is a function both of the agronomic performance of a system, and of the socio-economic context in which it is located, including the knowledge and understanding of the farmers. Neill and Lee (2001) make the point that farmers’ inadequate grasp of agronomic aspects such as reseeding, nutrient cycles, nitrogen fixation and herbicide effects, brought about by too rapid adoption of the system through hearing about it from neighbours, may also have contributed to the decline of the system. In both the West African and Central American cases, adoption of Mucuna cover cropping was due to a specific set of biophysical and socio-economic circumstances, which may not always be the same in other countries. In Benin, the critical conditions associated with its adoption were a decline in soil fertility, the lack of fertilizer, land scarcity, and weed encroachment, all of which combined to induce farmers to adopt a new technique which might not otherwise have been accepted (UNEP, 1999). In semi-arid Africa, however, establishment and management of cover crops has proven difficult where rainfall is less than 800 mm or on very clayey soils, so the practice is not likely to have much success there (Ganry et al., 2001). In Hondurus, adoption was widespread as farmers appreciated the benefits of the technique, which included higher maize yields, improved soil fertility with ease of land preparation, and moisture conservation, with weed control and erosion control of lesser importance (Buckles and Triomphe, 1999). The seasonality of maize prices also encouraged the uptake of the technique, as maize planted during the dry season commanded a higher than average value. These factors all helped to improve productivity both to land and to labour. The benefits were such that farmers who rented land were even willing to pay a premium on land that had been under the Mucuna technique. This also encouraged landlords who owned more land than they were able to cultivate under maize to invest in the technique. Thus, while rental of land generally discourages investment in techniques by tenants, the availability of land for rent and the value placed on Mucuna-treated fields encouraged the initial spread of the technique in this particular case. In general, therefore, intermediate intensities of land use are likely to be the best context for cover crops, as the opportunity cost of land will be high in land extensive systems, while the opportunity cost of labour and capital will be high in land extensive systems (Tarawali et al., 1999). Security of land tenure is also important as several years are required to reduce weeds and enhance soil physical and chemical properties to a level that might sustain another arable rotation (Tarawali et al., 1999). Some financial and labour investment in herbicides and fertiliser may also be necessary as evidence suggests that cover crops are not capable of indefinitely sustaining arable crop production, even in rotations (Tarawali et al., 1999). Purchase of seed can also be costly and this has been identified as a major constraint in some areas. Those farmers with intermediate levels of wealth and/or off-farm incomes may prefer to use cover crops in

LOW EXTERNAL INPUT TECHNOLOGIES

489

improved fallows (Tarawali et al., 1999), as the opportunity cost of their labour is relatively high (Franzel, 1999). Poorer farmers are likely to make use of natural fallow unless they are provided with credit facilities and/or other incentives (Tarawali et al., 1999) or have a large labour pool. However, the opportunity cost of capital, land, and labour may be relatively high. Wealthier farmers may decide to use inorganic fertilisers during the cropping cycle, although as they often have fairly large farms, part of the farms may be under fallow. Labour demand and the timeliness of that demand may also be problematic and cover crops will probably have the best chance of being adopted by households with sufficient surplus labour. This can sometimes be difficult as labour availability in rural areas often declines as farmers attempt to broaden their livelihood strategies with off-farm work. In other areas, cover crops have been found in niche uses with high-value tree crops, even though little extension effort has been made for this. For example, Thomas et al. (1991) discuss the use of cover crops as green manures as a source of nutrients for coconut. Similarly, Lehmann (2000) found that Pueraria phaseoloides cover crops in tropical fruit-tree systems enhanced nutrient cycling and reduced leaching, although competition for other nutrients (e.g., P) was a problem. In the tropical rainforest of Nigeria, farmers found that cover cropping was the most effective way of controlling soil erosion (Akinyemi and Kalejaiye, 1998). Another possible use is for feed meal production (Stur et al., 2002). It should be noted that cover crops have not always controlled weeds or increased crop yields, primarily due to competition with the main crop. For example, maize intercropped with the cover crops Pueraria phaseoloides, Vigna unguiculata (cowpea), and Mucuna pruriens in SW Nigeria yielded 2.15 t ha21 (þ 15% increase), 1.92 t ha21 (þ 3%), and 1.71 t ha21 (2 9%), respectively, compared to the 1.87 t ha21 for maize grown with no cover crop, the differences reflecting the relative competitive ability of the legume crops (Kirchhof and Salako, 2000). In some cases, the decomposition of green manure (particularly legumes) can release organic acids and/or other compounds which may affect germination, seedling growth, and yield of the following crop (Boddey et al., 1997). In Costa Rica, intercropping maize with Mucuna deeringiana and Canavalia ensiformis helped to control itchgrass (Rottboellia cochinchinensis) infestation to some extent, which usually resulted in improved maize yields, although on occasions yields were lower due to competition from the cover crop, especially Mucuna (Valverde et al., 1999). Similarly, in Benin, although the main reason for the widespread adoption of Mucuna cover cropping was its ability to control weeds, it was found that Imperata, in particular, was suppressed but not eliminated, and that it regained its original strength after one or two years of maize cropping (UNEP, 1999). This was because Mucuna has a limited effect on the rhizomes of Imperata (Akobundu et al., 2000), which suggests that eradication of Imperata may require a more integrated approach

490

A. GRAVES, R. MATTHEWS AND K. WALDIE

using cover crops in combination with other management techniques such as tillage or herbicides. The use of cover crops (traditional or new) may also sometimes encourage the introduction of pests, which may hamper the growth of the main crop or cause harm to the farmers themselves. In Bolivia, Pound et al. (1999) noted the increased presence of rats, snakes, and red spider mites with the use of Pueraria phaseoloides and Arachis pintoi as cover crops, while in West Africa, one of the constraints to adoption of mucuna cover crops was the fear possessed by farmers for the presence of snakes within the canopy (Galiba et al., 1998). Another concern with the use of green manures in flooded systems such as rice agriculture, which is receiving increased attention, is its influence on methane emissions into the atmosphere and subsequent contribution to global warming (e.g., Matthews et al., 2000).

D. BIOMASS TRANSFER TECHNIQUES In an effort to relocate nutrients from forests to agricultural land, tropical farmers have traditionally used a variety of biomass transfer techniques (e.g., Young, 1987; Nyathi and Campbell, 1993). In most cases, this has involved the use of naturally occurring biomass (i.e., tree or grass material), and rarely biomass that has been specifically planted for that purpose. Recently, however, the attention of researchers has focused on transfer of biomass from deliberately planted “biomass banks” of species such as Tithonia diversifolia (ICRAF, 1997; Gachengo et al., 1999; Jama et al., 2000), Gliricidia sepium (Rao and Mathuva, 2000), Calliandra calothyrsus, and L. leucocephala (Mugendi et al., 1999) as a means of providing nutrients for crop growth, and organic material for physical improvement of the soil. The use of so-called cut-and-carry grasses is another technique where biomass is harvested and transported, in this case specifically to provide fodder for animals (e.g., Tanner et al., 1993; Stur et al., 2002). While similar in principle to alley cropping in that plant biomass is cut and incorporated into the soil to release nutrients for crops and to help improve SOM levels, one of the potential advantages of biomass banks is that direct competition between the main crop and that used to supply the biomass is minimised, if not eliminated altogether. The evidence so far suggests that biomass transfer techniques can help to increase soil fertility and sustain or increase crop yields (Mugendi et al., 1999; Rao and Mathuva, 2000). However, for the technique to be successful, the quality of this biomass needs to be high (Snapp et al., 1998), and very large amounts of biomass are required to supply “ideal” quantities of nutrients to crops (Gachengo et al., 1999). Labour for the collection, transportation, and incorporation of the OM into the soil must also be plentiful (Snapp et al., 1998; Gachengo et al., 1999; Jama et al., 2000). For example, in the

LOW EXTERNAL INPUT TECHNOLOGIES

491

subhumid highlands of Kenya, biomass transfer of Caliandra calothyrsus and Leucaena leucocephela prunings was found to increase maize yields compared to both unfertilised and fertilised sole maize (Mugendi et al., 1999). Similarly, Rao (2000) found that transfer of biomass from a Gliricidia stand increased maize yields on an equivalent area by 27%, but the high labour cost for harvesting, transfer, and application of prunings, made the technique uneconomical. These issues are discussed in more detail later in this review. In Vietnam, Hoang Fagerstro¨m et al. (2001) found that Tephrosia grown in a “biomass bank” and used as mulch on rice fields increased production over a four-year period by 55% compared to the monocropped control, if the extra land used for growing the Tephrosia was not included. This increase dropped to 29% if the total land area was taken into account. The use of the shrub Tithonia diversifolia (Mexican sunflower) as a source of P has also received some interest recently (e.g., Jama et al., 2000). Tithonia appears to have the ability to extract relatively high quantities of P from the soil giving it a high leaf P content (. 0.25%), and although it is a non-leguminous species, it also has a relatively high biomass N content of about 3.4% (Gachengo et al., 1998), above the level required to prevent net immobilisation of N (Palm et al., 1997b). In addition, its leaf lignin and polyphenol contents are less than 15 and 4%, respectively, so that it decays relatively easily. As it grows abundantly in the wild in many areas, and indeed, is often cultivated as a farm boundary hedge in Africa and Asia, it has been suggested for use as a biomass transfer technique (ICRAF, 1997; Jama et al., 2000). Experimental evidence suggests that addition of N and P through the application of Tithonia biomass may increase yields more than the use of equivalent quantities of mineral N and P (Jama et al., 2000). This has been attributed to the presence of K, Ca, and Mg in the biomass which might ameliorate deficiencies of these nutrients in the soil, and possibly also because of an improvement to soil physical characteristics. In addition to providing nutrients, Tithonia has been shown to reduce P sorption and increase soil microbial biomass (Jama et al., 2000). Various techniques have been found to increase Tithonia biomass production and therefore the quantity of nutrients extracted from the soil and made available for transfer elsewhere. The use of woody cuttings for propagation instead of soft stem cuttings, for example, was found to increase its biomass production, as was the application of mineral P fertiliser (Jama et al., 2000). However, its ability to extract P appears to be less if it is grown as a stand rather than as a hedge (George et al., 2001), suggesting that it is taking up P from a wider area than just that occupied by the hedge. It may, therefore, not be so effective if grown as a biomass bank. The P content of the soil also influences the concentration in the leaves, and continual P-mining through removal of Tithonia prunings is likely to eventually reduce the quantity and quality of the biomass to the level below which immobilisation of P is likely to occur (, 2.4 g P kg21). Moreover, it needs to be remembered that only the importation of Tithonia biomass will result in a net

492

A. GRAVES, R. MATTHEWS AND K. WALDIE

increase of P on the farm — that taken from boundary hedges will merely redistribute it within the farm. Most literature on established biomass banks appears to be research oriented in nature, and little evidence exists of their deliberate and successful introduction onto farm land, probably for the reasons mentioned above. As with many other LEIA technologies, if biomass transfer techniques are to be used, considerable areas of land will be needed to grow sufficient quantities of plant material, which is clearly a limitation if land is scarce and farms are small. Consequently, it is unlikely that biomass banks for the sole purpose of soil fertility enhancement are likely to be widely adopted by farmers. There is also the disadvantage that there is a relatively long time-lag for benefits on soil fertility of such techniques to accrue (Snapp et al., 1998). Farmers may be interested in biomass banks if they cannot effectively use all their land for cultivation. However, in such cases, it is more likely that they may opt to fallow this land to regenerate soil fertility. Biomass banks could be established on strategically located common land, which would be especially valuable for poor farmers, but for this to be workable, it would be necessary for a system of access agreements to be developed. The question of who was responsible for the establishment and maintenance of such stands would also need consideration, particularly as there are costs involved in purchase of seeds and seedlings, planting, and the care required during the initial establishment (i.e., weeding, etc.). Similarly, the sustainability of the system, with constant removal of nutrients in the biomass, would need to be addressed. Snapp et al. (1998) have noted that even large naturally occurring woodlands, such as the miombo woodlands in southern Africa, cannot be indefinitely mined for nutrients. In most cases, it is unlikely that biomass transfer techniques will be capable of supplying the full fertility needs of a farm, and as with other LEIA fertility enhancing techniques, it may be best to see them as a component of an integrated nutrient management (INM) system involving external supplies of inorganic nutrients. It may be that there are specific niche roles which will make them useful on small areas within a single farm (e.g., for home gardens), or on degraded common land, although it is more likely in this case that they will fulfil other important needs, such as the provision of fuel-wood and/or fodder. Here they would move closer to the role played by natural forests, in which case, they may help relieve some of the pressure on the latter. However, there is a greater likelihood of adoption of biomass transfer techniques where there is some immediate benefit to be obtained by the farmer. The use of cut-and-carry grasses as animal fodder, for example, has the advantage that some animal products, including milk, meat and draught power, are available almost immediately. Improvements to soil fertility through the application of manure and urine may be a secondary result. Farmers are usually well aware of the beneficial effects of manure on soil fertility — there is some evidence that they feed their livestock much more than required for optimal live-weight gain, in

LOW EXTERNAL INPUT TECHNOLOGIES

493

order to provide manure for arable crop production (Tanner et al., 1993). Nevertheless, the essential difference of “processing” plant biomass through animals first to gain immediate benefits (e.g., meat, milk, and draft power), rather than using the biomass directly to improve soil fertility, is likely to be a determining factor of whether biomass banks are adopted by farmers or not. Of course, the intensification of agriculture with cut-and-carry grasses and/or fodder banks is most likely to occur where animals are already a major component of the agricultural system and where satisfactory and alternative feeding strategies do not already exist (Reynolds et al., 1991; Rachmat et al., 1992). In some areas, the importance of cut-and-carry and zero-grazing techniques may increase as population increases and less and less land is available for free-grazing on communal land (Murwira et al., 1995), or where access to naturally occurring vegetation is limited, either because it does not exist or because access to it is blocked. However, in either case, sufficient land still needs to be found somewhere to grow the biomass (Gashaw et al., 1991). As with other biomass transfer technologies, the evidence indicates that the amount of labour required for cut-and-carry techniques for fodder provision is often a disincentive to adoption (Mogaka, 1993; Wandera et al., 1993; Sanchez and Rosales Mendez, 1999). Cut-and-carry grasses may also not supply the full fodder requirements of livestock, necessitating supplementary feeding. Capital will be needed to pay for this, and for the inevitable veterinary fees associated with keeping livestock healthy (Mogaka, 1993). In addition, the decline of productivity of the fodder banks, as nutrients are removed may require investment in fertilisers to maintain productivity (Wandera et al., 1993). Finding suitable grasses, as well as issues related to land tenure are other important considerations. On the whole, the deliberate establishment and maintenance of fodder banks and cut-and-carry systems is probably most likely where farmers have some surplus capital and labour, but where land scarcity has resulted in limited access to natural vegetation. The ownership of cattle and the establishment of biomass stands involve costs that very poor farmers are unlikely to be able to meet.

E. COMPOST Compost is the aerobic, thermophilic decomposition of organic wastes to a relatively stable humus. Although it makes use of the same decomposition processes occurring naturally, the aim in compost-making is to control the conditions to a level that allows faster decomposition. The biophysical conditions that are required for effective composting are generally those that are required by the micro-organisms at various stages of the composting process, i.e., good moisture levels, moderate temperatures, mixed quality organic matter, and a fairly neutral pH range.

494

A. GRAVES, R. MATTHEWS AND K. WALDIE

Composting is not a new technique for the improvement of soil fertility and structure, and tropical farmers have been aware for centuries of its impact on crop yields, soil structure and fertility, crop growth and vigour (Diop, 1999; Onduru et al., 1999). For example, Oue´draogo et al. (2001) observed that yields of sorghum in Burkino Faso could be tripled by the application of 10 t ha21 of compost. Another benefit noted is the reduced need for capital inputs (Onduru et al., 1999), although some capital may be necessary for farmers to adopt the technology (Girish and Chandrashekar, 2000; Slingerland and Stork, 2000). Compared to techniques such as alley cropping, composting in relation to subsistence agriculture seems to have received little attention by researchers. Much of the literature available tends to be written from a purely technical standpoint, and is often from the perspective of agriculture in developed countries. Relatively few studies have considered the on-farm issues of using compost, although the few that do give a good indication of the constraints involved. The major problem associated with the use of compost is the high labour requirements (Onduru et al., 1999). In particular, female-headed households can have considerable difficulty in undertaking some of the heavier tasks involved in composting, such as preparing compost pits (Diop, 1999). In Uganda, Briggs and Twomlow (2002) found that the poorest households did not make compost at all because of the labour and time requirements. Transportation of biomass and compost is also problematic (Apiradee, 1988; Adeoye et al., 1996). Also, like the other LEIA techniques already discussed, large quantities of biomass are required, and questions arise as to where farmers can obtain this (Onduru et al., 1999; Oue´draogo et al., 2001). This is particularly relevant where there are competing demands for such resources, for example as mulch, fuel or fodder (Drechsel and Reck, 1998), and where land to produce the biomass is scarce. Resource-poor farmers may have problems providing land for processing of “ideal” quantities of biomass, although this is not generally cited as a limitation, probably because fairly small quantities of compost are usually produced. On the other hand, some systems seem capable of producing quite large quantities; Briggs and Twomlow (2002) found that smallholder households in Uganda produced 40 kg of fresh organic waste per day, or about 9.2 t DM year21, 25% of which was used to make compost, with the rest either fed directly to livestock or applied directly to household plots. Composting may sometimes be constrained by lack of water (Apiradee, 1988; Diop, 1999), which is needed to aid decomposition, and by lack of biostarter (Apiradee, 1988), although it appears that animal manure and inorganic fertiliser may be used. Lack of tools for composting may also be a problem in some areas (Oue´draogo et al., 2001). An example of the use of compost in subsistence agriculture is provided by the work of the Rodale Institute, which has been working since 1987 with farmers in the Peanut Basin region of Senegal to help offset the rising cost of fertiliser due to removal of subsidies (Diop, 1999). Composting is not a new technique in

LOW EXTERNAL INPUT TECHNOLOGIES

495

Senegal, so research has focused on attempting to improve existing techniques involved in its production and use. Farmers are encouraged to collect crop residues for compost making rather than burning them, and to incorporate the resulting compost in the soil rather than leaving it lying on the surface. Pit composts are also being developed — these are about 1 m deep, lined with cement and bricks, and covered to prevent contamination from wind-blown sand. In the simplest and most efficient approach, which does not require turning, rainfall in the wet season is used as a water source, with compost being produced within 18 months. In a more labour-demanding approach during the dry season, water is added manually when the compost humidity drops below a certain level. The compost must be turned every 15 days, but is ready within 45 days. For larger-scale compost making with this method, three pits are used with compost being transferred from one pit to the next every 15 days. Yields of both groundnut and millet, the traditional crops, have been approximately tripled through the application of 2 t ha21 of compost (Diop, 1999), although the area of crop involved in this case is not stated. Another example is the Zai system, traditionally used by farmers in Burkina Faso in times of drought (Roose et al., 1999). The system involves digging holes (20,000– 25,000 ha21), typically 30 cm wide and 20 cm deep, filling them with compost, and planting seeds of sorghum, millet, and cowpea into them. The compost is made from farmyard manure, plant residues, garbage, and rock phosphate — a natural product from mines in Burkina Faso. Crop yields can be more than 10-fold higher using the system than otherwise (from 150 to 1700 kg ha21), and the holes can be reused for 3 years. Trees may also sprout spontaneously in the holes. As composting is labour intensive, it is probably most appropriately used close to the homestead, on specific crops (e.g., Briggs and Twomlow, 2002). Evidence also suggests that progress may be made by improving the technical knowledge of farmers, so that composting practice is improved (e.g., Sutihar, 1984; Adeoye et al., 1996; Wakle et al., 1999). For example, the quality of compost can be greatly enhanced by mixing it with a combination of inorganic chemicals (van den Berghe et al., 1994), or by combining it with manure (Onduru et al., 1999). During processing, protecting it from heat and direct light may reduce volatilisation of the nutrients, while protecting it from rain may prevent similar losses by leaching (Diop, 1999). It seems unlikely that composting would be a new technique in many areas, unless a rapid transition from an abundance of land to scarcity is occurring. Progress is more likely to be made by determining whether composting practices can be improved within the socio-economic constraints of such farmers, such as with the Rodale Institute work. Recommending the use of animal manure in compost, for example, would not be appropriate if animal manure is the only source of fuel. Thus, as with other OM management techniques, composting is likely to provide only a partial solution to the problem of decline in soil fertility and soil

496

A. GRAVES, R. MATTHEWS AND K. WALDIE

structure in subsistence agriculture, but should be an important component in the basket of options that a farmer could use. Since composting is very labour intensive, it will probably continue to be most used close to the homestead, or on small areas of high-value crops. Extending the use of composting to large or distant fields may require unrealistic levels of labour for most farmers.

F. ANIMAL MANURE As with several of the other LEIA technologies, the beneficial impacts of applying animal manure to land are well known to many tropical farmers and are also well documented in the scientific literature (e.g., Prudencio, 1993). As already mentioned in relation to the Machobane System, the use of animal manure as a source of crop nutrients is often a nutrient-harvesting technique in which nutrients are gathered through grazing of a relatively large area and concentrated on a smaller area where crops are grown. Even where animals are stall-fed, the nutrients they consume must be brought to them from elsewhere, either as collected fodder or as purchased concentrates. The rumen provides an ideal environment for the decomposition of organic matter, and is a way of improving the rate of decomposition in suboptimal environments, such as those with low temperatures or dry conditions. Livestock in developing countries generally feed on low-quality grasses and/or crop residues, which have significant fractions of indigestible materials such as lignin and cellulose – lignin complexes, as well as low N contents (Leng, 1990). Resource-poor farmers often cannot afford to purchase supplements, and instead use tree and forage legumes to improve poor-quality diets (Coppock and Reed, 1992). The quality of a diet can influence feed intake, feed digestibility, and the partitioning of N between the urine and faeces (e.g., Somda et al., 1995). This latter characteristic is important, as a large proportion of the N excreted in the form of urine can be lost from the system through volatilisation and leaching, whereas that in faeces can be collected, stored and used to benefit crop growth (Delve et al., 2001). The manure of animals which are fed highly digestible diets is more susceptible to N losses than that of animals fed a greater amount of roughage (Powell and Williams, 1993), while feeding on browse species has been found to shift the balance of excreted N away from the urine more towards the faeces (e.g., Reed, 1986). For example, Coppock and Reed (1992) found that animals fed Acacia tortilis pods excreted only half as much N in urine as those fed cowpea hay. Similarly, inclusion in the diet of Flemingia congesta, which has a high tannin content, was found to reduce N excreted in the urine by up to 35% (Fassler and Lascano, 1995). Hypothetically, a farmer with access to organic material is faced with a choice of whether to apply it directly to a crop as an organic fertiliser, or to use it to feed livestock first with the manure produced then being applied to the crop.

LOW EXTERNAL INPUT TECHNOLOGIES

497

The optimal choice depends on the quality of the biomass available. It is generally accepted that material with a C/N ratio of more than 20 kg C (kg N)21 results in immediate immobilisation of N by microbial biomass when incorporated into the soil (Senesi, 1989) which is likely to reduce crop yields by reducing the amount of mineral N available. Passing this type of material (e.g., cereal straws) through an animal rumen first can result in a large reduction in its C/N ratio, thereby reducing the amount of immobilisation. For example, Delve et al. (2001) found that cattle faeces from a diet of barley straw had a C/N ratio of 27 kg C (kg N)21 compared to that of 86 kg C (kg N)21 of the plant material only; consequently on incorporation, the faeces had a faster mineralisation rate and showed less net N immobilisation than did the fresh barley straw alone. Crop uptake of N was also greater from incorporated faeces than from the fresh straw. On the other hand, incorporation of high-quality plant material usually results in a higher recovery of N than of faeces derived from the same plant materials (e.g., Catchpoole and Blair, 1990). Delve et al. (2001) found that leaves of Calliandra calothyrsus with a C/N ratio of 13 kg C (kg N)21 resulted in steady net N mineralisation, but that manure from a diet containing 30% Calliandra calothyrsus resulted in net immobilisation between 3 and 16 weeks. The results from these studies indicate that the effect of passing material with C/N ratios ranging from 13 to 86 kg C (kg N)21 through an animal rumen is to make the C/N ratio converge to a value between 20 and 27 kg C (kg N)21. Thus, from a nitrogen supply point of view, it would seem better to apply high-quality organic matter (C/N ratio , 20 kg C (kg N)21) directly to the soil, but to pass low-quality biomass through an animal first, and then apply the manure to the soil. Of course, this does not take into account other benefits such as a better supply of milk and meat that might come from feeding animals high-quality fodder. The beneficial effects of animal manures on crop yields, when applied in sufficient quantities, are well documented. Not only can they enhance immediate crop yields (Selvarajan and Krishnamoorthy, 1990; Ali, 1996; Drechsel and Reck, 1998), they can also provide residual benefits for following years (Singh and Desai, 1991). However, Giller et al. (1997) note that, although animal manures are of major importance in nutrient cycling, they are generally of poor quality as a supply of plant nutrients. In Rwanda, for example, Roose et al. (1997) found that the application of manure at 10 t ha21 had no effect on maize yields. Even the effect of applying manure at the rate of 20 t ha21 was limited to a single season, as the yield of the second crop (sorghum), showed no significant difference compared to the other treatments, which was ascribed to the low amount of P supplied in the manure. The benefits to the soil from applying manure, therefore, may be more related to improvements in physical characteristics rather than the provision of nutrients, especially in the quantities that farmers can supply (Singh and Desai, 1991). It is important to remember that the use of manure is related to its role as part of a larger system. In many subsistence-farming systems, there is a close

498

A. GRAVES, R. MATTHEWS AND K. WALDIE

interdependence between crop production, animal production, and possibly neighbouring forests and/or rangelands. In Nepal, for example, animals are grazed in the forest, crop residues and forage collected from the forest are used to feed housed animals, animal manure is applied to cropped areas, crop residues and forest litter are used for animal bedding, and animals are used to provide tillage and transport (Pilbeam et al., 1999b). In certain areas of western Africa, some arable farmers make arrangements with itinerant herdsmen to corral livestock on their land (Waldie, 1990; Enyong et al., 1999). Farmers may also move their homesteads from place to place so that crops can be grown on that land to benefit from the manure left over by livestock (Ruthenberg, 1980). In upland Java, livestock may be fed with far greater amounts of biomass than is needed for optimal live-weight gain, the rationale being the production of manure-compost that is collected for intensive upland agriculture (Tanner et al., 1993). In smallholder farms in Uganda, Briggs and Twomlow (2002) calculated that an average of 4.6 t of manure was produced annually from a household with 2.2 ha of land — this was mainly from goats grazing on communal land during the day, but tethered near the household at night — a figure that could be increased if better manure management was practised. In many cases, there may be a high opportunity cost of using manure as a fertiliser, and farmers may often value it more for uses other than soil fertility maintenance. Benefits obtained from manure include the provision of material for plastering and building, and fuel for heating and cooking (Jeffery et al., 1989; Murwira et al., 1995). Provision of milk for human consumption from keeping livestock has already been mentioned. Many of these activities (e.g., the production of fuel cakes or milk) have direct economic value in themselves (Jeffery et al., 1989). As discussed for the Machobane system, competing uses of manure inevitably reduce the amount available for soil chemical and physical improvement. If manure is to be extensively used to enhance soil fertility, it will need to be culturally acceptable to farmers, which is most likely to occur where livestock are an integral part of the farming system already. This is more likely to be in areas where population pressure is higher, labour availability is higher, and land is scarce, such as in Nepal (Murwira et al., 1995). Where land is not scarce, farmers may find it more practical to improve soil fertility and structure through natural or improved fallow, although they may still use manure on plots or kitchen gardens close to the homestead if they own poultry, smallstock, or cattle. Both the production and use of manure are labour intensive (Enyong et al., 1999). Households without adequate labour, or the means to procure it (e.g., in communal work groups or through purchase), may only be able to use small amounts of manure. In general, investment in manure to improve soil increases where an enabling environment for agricultural production is provided. For example, in areas where there are market outlets and effective extension networks providing technical guidelines on good manure practice, farmers may be

LOW EXTERNAL INPUT TECHNOLOGIES

499

encouraged to increase productivity, stimulating the use of manure as one of the various options available for increasing crop production (Enyong et al., 1999). However, the fragmentation of fields occurring in many developing countries may also make it more difficult to transport manure, reducing farmers’ willingness to apply it to fields at a distance from the homestead (Enyong et al., 1999). In general, as with other LEIA techniques, the use of manure should be seen as a component in an INM system. Indeed, Sherchan et al. (1999) found that the highest yields of maize and wheat in Nepal were obtained from a combination of manure and inorganic fertiliser. Alternatively, it could be used in certain niches, say around the homestead with specialised high-value crops, where the distance to transport it is not a major constraint.

G. IMPROVED FALLOWS Shifting cultivators have traditionally alternated periods of crop production with periods of fallow in order to restore soil fertility and suppress weeds. In some cases, the cropping period only lasts 4– 5 years while the fallow period may be as long as 30 years (e.g., Matthews et al., 1992a), during which time the land is usually unproductive in terms of generating a livelihood. In recent years, researchers have focused on ways to shorten this period, and/or to make some use of the land while it is fallow. Thus, an “accelerated fallow” is where specific fastgrowing leguminous or non-leguminous trees, shrubs, legumes, and other plants are used to improve soil fertility faster than would occur otherwise, while an “enriched fallow” is where trees or shrubs of economic value are planted into the fallow so that the farmer can derive some income from them while the land is regenerating (Garrity and Lai, 2000). Szott et al. (1999) have reviewed the processes involved in fallowing from an ecological perspective. Under a natural fallow, the time needed to restore N to original levels can be less than two years, but can be up to 20 years for other nutrients such as Ca. The rapid restoration of N is likely due to N fixation (up to 300 kg N ha – 1 year21 on high base status soils, (Giller, 2001)) and retention in the growing biomass. Most evaluations of the performance of improved fallows have compared crop yields after the fallow period with those after a natural fallow of the same duration. In general, it has been found that short-term fallows (i.e., , 3 years) growing leguminous trees or shrubs can increase crop yields compared to the natural fallow control (Szott et al., 1999). Nitrogen fixation is the main reason for this, but the retrieval of subsoil N by deep-rooted species from below the crop root depth can also be important, and may be of the same magnitude as N-fixation, particularly if nitrate has accumulated in the subsoil during the cropping phase before the fallow (e.g., Mekonnen et al., 1997). Improvement in soil structure may also be important — Salako et al. (2001), for example, found that soil physical properties improved under fallows of

500

A. GRAVES, R. MATTHEWS AND K. WALDIE

Pueraria phaseoloides, Acacia auriculiformis, Leucaena leucocephela, Senna siamea, and Acacia leptocarpa, as well as under a natural fallow. Grass fallows do not usually increase crop yields to the same degree, if at all, probably due to immobilisation of N (Barrios et al., 1997) although in some cases, they do improve soil structure (Padwick, 1983). It is noteworthy that crop yields were significantly lower following fallows of Ageratum conizoides, Tithonia diversifolia, Sida rhombifolia, or Rottboellia cochichinensis than following a natural fallow, despite having more N and P in the fallow biomass (Szott et al., 1999). On soils that are severely deficient in P, short-duration improved fallows have been shown to need additions of fertiliser P to maintain crop yields (Jama et al., 1998). Benefits of short-duration fallows to crop yields are sometimes related to the amount of biomass accumulated during the fallow. Maroko et al. (1998), for example, reported no increase in yields of the second maize crop after a Sesbania fallow, which had accumulated 14 t ha21, but a significant increase in crop yield with a 23 t ha21 Sesbania fallow. In most cases, there is little residual effect of the fallow biomass after two crops (Drechsel et al., 1996). Interestingly, the use of short-term cover crops such as Mucuna in fallows may be detrimental to long-term accumulation of nutrients and biomass by impeding the establishment of trees and shrubs (e.g., Szott and Palm, 1996). Kang (1997) has recommended that short fallows be used where the duration of the cropping period is at least half of the total cropping/fallow cycle. There may also be other benefits derived from improved fallows. For example, Brodd and Osanius (2002) report reduced soil erosion under an improved fallow of Tephrosia candida in northern Vietnam compared with shifting cultivation. Similarly, Gallagher et al. (1999), in a review of literature, have shown that improved fallows of woody perennials and herbaceous cover crops could suppress weeds, particularly over a number of years, and might be an important component of integrated weed management (IWM) strategies. Perhaps the main advantage of accelerated and enriched fallow systems is that that they are modifications of an existing system, requiring only minor changes to existing farmer practice. Accelerated fallows can be seen as a natural progression from shifting cultivation and other long-fallow techniques, and may therefore be an appropriate choice where these are no longer sufficient to maintain productivity due to population pressure (Sanchez, 1999). From the biophysical point of view, due to the deeper-rooting characteristics of the woody species usually used in accelerated fallows, nutrients from below the rooting depth of arable crops can be made available again, and the use of appropriate leguminous species can result in improved rates of addition of N to the system. Also, there is no direct competition for resources with main crops as with some of the other LEIA techniques such as intercropping or alley cropping (Sanchez, 1999). The downside, compared with continuous cropping systems at least, is that production is lost from the land set aside for fallow. Hoang Fagerstro¨m et al. (2001) found that the overall rice production from both a natural fallow and a Tephrosia fallow

LOW EXTERNAL INPUT TECHNOLOGIES

501

with 2-year crop/2-year fallow cycle were 18 and 6% less, respectively, than production from continuous cropping over the four-year period, despite annual yields in the latter being considerably lower than in the fallow treatments. Similar results were obtained by Swinkels et al. (1997) in western Kenya. Accelerated fallows are more likely to be used by farmers in areas where increases in the population density is starting to make the long periods required by natural fallow impracticable. At higher population densities, however, scarcity of land means that there is a high opportunity cost in putting land to fallow, and intensive continuous cultivation systems may dominate (Drechsel et al., 1996). Accelerated fallows are, therefore, most relevant in the intermediate stage between extensive and intensive land use (Franzel, 1999). However, if they are to be adopted more widely, farmers need to be aware that there is a problem to be addressed. This may be declining yields (Franzel, 1999) or fertility (Degrande and Duguma, 2000), or controlling weeds (Tarawali et al., 1999). Security of land tenure is also an important consideration, as farmers are unlikely to be willing to invest time and effort in establishing accelerated fallows if they are not the ones to receive the benefits (Seif El Din and Raintree, 1987; Long and Nair, 1999; Tarawali et al., 1999). Institutional support, in the form of seed programmes and training of extension agents and farmers, has been found to be important in improving the adoption of accelerated fallow techniques (Franzel, 1999). Other important requirements may also be to provide adequate and/or improved germplasm (Place and Dewees, 1999). A major disadvantage of both natural and accelerated fallow systems is the length of time it takes for any financial benefits to accumulate (Grist et al., 1999), particularly with the latter, since natural fallows may provide resources that accelerated fallows do not. Kaya et al. (2000) concluded that improved fallows in Mali were not attractive to farmers if their sole purpose was soil fertility improvement. Enriched fallows address this problem to some extent, in that species that are able to provide some economic benefit, such as fruit or nuts, are planted in preference to species that only improve soil fertility (Cairns and Garrity, 1999; Franzel, 1999; Sanchez, 1999). Other practical benefits to farmers may include production of fodder (Kaya et al., 2000), honey (Franzel, 1999), firewood, or bean poles (Drechsel et al., 1996), or light timber for construction (Franzel, 1999). From an ecological viewpoint, Styger et al. (1999) have suggested that the judicial identification, selection, and domestication of preferred forest fruit trees could be used as a means of preserving biodiversity in forest margin areas that are under pressure. Where such multipurpose tree (MPT) techniques are successful, farmers may be encouraged enough to develop them into permanent agroforestry systems (Cairns and Garrity, 1999), and, indeed, in several places many have done this already. Farmers in Benin, for example, plant oil-palm trees in a fallow of about 12– 15 years (Versteeg et al., 1998). This restores soil fertility, but also provides subsistence and cash income, even upon clearance of the trees when “palm wine”

502

A. GRAVES, R. MATTHEWS AND K. WALDIE

is produced. However, removal of harvestable products (and the nutrients they contain) lengthens the time taken for regeneration of soil fertility, and in extreme situations, may eventually mine the soil of nutrients (Franzel, 1999). These deficits will have to be made up with other sources of nutrients. Pradeepmani (1988) has discussed some of the issues affecting farmers’ decisions to plant MPT species. These include having adequate land, time, labour, knowledge, and inputs, being able to protect trees properly, and success with tree survival. Security of tenure and access to markets are also important (Mahamoudou and Meritan, 1998; Hellin et al., 1999). A certain level of access to capital also appears to encourage adoption. Where these factors are not in place, farmers tend to increase the rate at which they discount future benefits, making such techniques socio-economically unviable, and reducing both the “action time horizon” and the “planning time horizon” (Vosti and Witcover, 1996). Efforts to encourage planting of MPTs species through training visits, extension of effective methods for protecting trees (which is often expensive), and government land tax incentives, were also noted as important factors (Pradeepmani, 1988). Although the multipurpose nature of many trees may serve as “pull” factors, strong “push” factors can also operate at the localised scale. For example, shortages of agricultural labour, high cost of agricultural inputs, and shortages of power and water have all been reported to encourage farmers to plant MPTs (Dasthagir et al., 1996). In most cases, economic factors are more important than ecological factors in influencing farmers’ decisions to plant MPTs (Mahamoudou and Meritan, 1998). Certainly, in a study of the reasons for the adoption of MPTs on homestead land by farmers in Bangladesh, direct economic concerns were reported as being uppermost in the minds of farmers (Salam et al., 2000). Other factors in order of importance were the provision of fruit, firewood, and building materials for subsistence, emergency cash needs, maintenance of ecological balance, and protection from strong winds. Further analysis showed that tree planting increased with increases in the amount of land owned, the level of non-agricultural income, the market costs of fuel-wood, the male membership of the household, and the knowledge of extension activities. Compared to natural fallow systems, more labour is required for both accelerated and enriched fallows, primarily for planting of the fallow species and weeding to ensure their establishment (Drechsel et al., 1996; Grist et al., 1999). Hoang Fagerstro¨m et al. (2001) give values of 386 and 600 labour days ha21 over a 4 year period for a natural fallow and a Tephrosia fallow, respectively. There is a clear danger that this additional labour demand could clash with those for planting and management of other crops (Franzel, 1999). On the other hand, the labour demand of improved fallows is relatively flexible, certainly compared to alternative systems such as alley cropping where the timeliness of pruning is very important. Some capital input is needed for improved fallows, mainly for the purchase of seeds or seedlings of the species to be planted. If farmers have insufficient capital, they tend to use natural fallows for soil fertility improvement,

LOW EXTERNAL INPUT TECHNOLOGIES

503

whereas better-off farmers will tend to purchase inorganic fertilisers (Franzel, 1999). Improved fallows may therefore, be most appropriate for farmers at an intermediate level with some disposable income. Adoption of improved fallow techniques by farmers has been significant in some regions. Citing various sources, Sanchez (1999) claims that there is largescale adoption of improved short-term fallows (i.e., , 5 years in duration) occurring in Central America, Brazil, Southeast Asia, East Africa, and southern Africa, with perhaps hundreds of thousands of farmers using the technique. The majority of species used are Sesbania, Leucaena, Mucuna, Centrosema, Pueraria, Crotalaria, Cajanus, Indigofera, and Mimosa. An example is the “Qezungual System,” which has been indigenously developed in western Honduras (Hellin et al., 1999). The technique can be described as a triple-level agroforestry system, combining crops such as maize, sorghum and beans, numerous pollarded trees and shrubs (about 1.5 m high) and high-value trees, particularly fruit trees and timber trees. The system has generally developed on land that has been under secondary vegetation, or less commonly, on land that is under primary forest. A main characteristic of the technique is the reduction of labour requirements for the establishment of the valuable fruit and timber trees species, as these are simply selected when the land is cleared for agriculture. Other less valuable trees and shrubs are pollarded and the land is prepared for cultivation. This also reduces the time required for benefits to accrue to the farmer, and may reduce the need for inputs requiring capital (for example, seedlings) and labour (maintenance of vulnerable seedlings). Competition between perennial plants and food crops is greatly reduced by the pollarding, and can be manipulated by gradual clearing of perennial plants, if necessary. Natural regeneration is managed by selecting specific trees for production, whilst others are pollarded. Within the pollarded areas, crops may be rotated and areas left fallow to control pests. From discussions with farmers, there appear to be several benefits from the use of this technique. Pollarded plots have higher agricultural production and can also be cultivated for a longer time than unpollarded plots. Soil moisture is conserved because of reduced soil evaporation due to the mulch from the pollards, and perhaps because the pollards improve soil physical structure and allow increases in WHC. The system provides multiple benefits for subsistence and cash income (fruit, food crops, timber and firewood). Disadvantages are that moisture levels can become too high in times of high rainfall, leading to fungal attacks on crops. Birds may also be attracted by the trees and pollards, causing reduction of food crop yields. Animal or mechanised traction cannot be used to cultivate the land due to the random presence of the trees and pollards. Various factors have encouraged farmers to adopt the Qezungual system. Land scarcity is important (most farms are about 2.5 ha) as, where land is abundant, farmers generally continue to use natural fallows. Absence of fire as a management tool is also important, otherwise the trees are destroyed. Lack of

504

A. GRAVES, R. MATTHEWS AND K. WALDIE

animal or mechanised traction is another factor, as the pattern of trees and pollards is fairly random, although with selective thinning, it might also be possible to develop pathways for animal and mechanical traction tillage. Possibly the most important factor is that it addresses a problem that the farmers find important — soil moisture. In dissemination of the use of the system, it is promoted as a soil moisture technology, rather than an erosion control technology, although erosion is definitely a problem in some areas.

III. GENERIC ISSUES A. SOIL FERTILITY MANAGEMENT 1.

Biomass Quantity

Relatively little quantitative information exists on the ideal level of SOM. Brady (1990) suggests that it should be around 5%, corresponding to a soil organic carbon fraction (SOC) of about 3%. In sandy southern African soils, 1– 1.5% SOC has been recommended as the long-term agroecologically viable minimum (Araki, 1993). Similarly, research from western African countries suggests that when SOC levels fall below 1%, severe physical soil degradation can be expected to take place (Pieri, 1995). The threshold level of SOC required to prevent severe physical degradation of a soil is also related to the soil’s texture — for soils with a low-sand content, an SOC content of 0.9% may be adequate, but for sandy soils, this may have to be as high as 1.5% (Araki, 1993). Thus, although there appears to be some variation as to an “ideal” level, somewhere between 1 and 3% SOC is what many researchers would consider necessary. Young (1989) has estimated indicative quantities of plant biomass required for maintenance of SOC in soils in various agroecological zones, based on typical topsoil organic carbon levels and approximate oxidation and erosion losses. Above-ground plant biomass required inputs were estimated to be about 8.4 t DM ha21 year21 for humid regions, 4.2 t DM ha21 year21 for subhumid regions, and about 2.1 t DM ha21 year21 for semi-arid areas, with below-ground biomass inputs about 70% of these figures. These are broad estimates and other evidence suggests different quantities are required. For example, in southern Africa, Snapp (1998) estimated that annual applications of about 10 t DM ha21 year21 of high-quality plant biomass (see below for discussion on biomass quality), or 7 t DM ha21 year21 of low-quality residue, were necessary to maintain a minimum level of 1% SOC in sandy loam soils in the subhumid tropics, assuming a decomposition rate of 0.05 year21 (Janssen, 1993). Thus, assuming a linear relationship between SOC levels and biomass inputs, the “ideal” level of 3% SOC mentioned previously may require as much as 30 t DM ha21 year21 in

LOW EXTERNAL INPUT TECHNOLOGIES

505

similar conditions. Even if the relationship is not linear, the data suggest that on the whole, large amounts of biomass are required to maintain the physical condition of the soil at a level that can support continuous and sustained crop production. The supply of N and P to meet crop nutrient requirements also requires large quantities of biomass, not only because of the relatively low concentration of nutrients in biomass, but also because the fraction actually recovered by crops is generally low for organic inputs due to losses to the atmosphere, surface waters, or ground water (Gregory et al., 2002), although this can vary substantially depending on biophysical conditions. Giller and Cadisch (1995) have suggested a value of 20% for most organic inputs. Gachengo et al. (1999) found that the recovery of N in Tithonia prunings was about 25% by a first maize crop in western Kenya. Other evidence suggests that N recovery by the first crop after OM incorporation is generally between 9 and 28% of the N supplied in the OM, but may be as low as 2 –10% in a second crop (Snapp et al., 1998). However, the rate of N recovery can be improved by the addition of limiting nutrients to ensure that the growth of the main crop is not limited. For example, Snapp et al. (1998) found that the recovery fraction of organic N increased from 25 to 46% in the first year when 25 kg P ha21 was applied to correct the P deficiency at the site. To supply adequate N for a typical 5 t ha21 maize crop removing about 100 kg N ha21, therefore, more than 14 t DM ha21 of biomass would be required (assuming a 20% recovery rate and 3.5% leaf biomass N content (Jama et al., 2000)). For animal manure with 1.5% N content (Lekasi et al., 1998), 33 t DM ha21 would be required. These figures would equate to 60 and 67 t ha21 of fresh material, respectively, assuming a 20% dry matter content for fresh leaf biomass and a 40% dry matter content for fresh manure. For P, Jama et al. (2000) calculated that 5 t DM ha21 would be required to supply the 18 kg P ha21 needed to overcome moderate P deficiencies. With a 20% recovery rate, this is equivalent to an application of 25 t DM ha21 or 125 t ha21 fresh weight. In severely P-deficient soils, even more would be required. The quantities required for the effective management of other soil nutrients is also large. For example, Jackson et al. (1999) found that to correct zinc deficiencies of soils near Wenchi in Ghana, about 20 t ha21 (dry or fresh weight not specified) of poultry manure was required. The amount of sheep or cattle manure required was estimated to be between 40 and 60 t ha21. If these levels of biomass are required to have an appreciable effect on SOC, the question arises as to how easily these quantities can be produced by smallholders. In the studies where alley cropping has been shown to benefit crop yields, tree biomass production has been in the order of 6 –8 t ha21 year21, using L. leucocephala (Kang et al., 1985). For Flemingia congesta, Budelman (1988) recorded an annual dry matter production of 12.4 t ha21 year21 in the Ivory Coast, while Yamoah et al. (1986a) measured 16.9 t DM ha21 year21 for Flemingia congesta in Nigeria. However, in many cases, biomass production is not this high. In an alley cropping system in Costa Rica (humid), the net primary

506

A. GRAVES, R. MATTHEWS AND K. WALDIE

production of Calliandra calothyrus was about 4.4 t DM ha21 year21. Of this, about 2.8 t DM ha21 year21 was estimated to be leaf production (Baggio and Heuveldorp, 1988) and of possible use as a green manure. In Nepal, C. cajan was able to produce about 3– 4 t DM ha21 year21, Centrosema pubescens about 5 t DM ha21 year21, and Calopogonium mucunoides about 4 t DM ha21 year21 (Pande, 1997). In Malawi, Saka et al. (1995) showed that the leaf biomass production of three hedgerow species (Gliricidia sepium, Leuceana leucocephala, and Senna spectabilis) varied between 0.5 and 2 t DM ha21 year21, and did not affect the SOC level over a 1 year period. Also in Malawi, Kanyama-Phiri et al. (1997) found that Sesbania sesban produced about 2– 3 t DM ha21 year21 of high-quality leaf biomass. This was in addition to the fuel-wood produced during the 10 months of growth between January and October. In Zambia, although Flemingia congesta produced a maximum of 3 t DM ha21 year21 in one trial (Table I), the mean production of all the species was only 1.3 t DM ha21 year21 (Matthews et al., 1992a). In many cases, therefore, biomass production does not seem high enough to have any appreciable effect on SOC levels. Cover crops are likely to produce even less biomass annually due to their shorter duration of growth compared to woody perennials. Increasing SOC to “ideal” levels, therefore, will in most cases, necessitate the importation of additional amounts of organic material. The question is, therefore, where is this biomass to come from? If the farmer is to grow it, can it be produced in sufficient quantities to have an appreciable effect

Table I Mean Annual Biomass Production (t ha21 y21) of Different Tree Species in Agroforestry Trials at Kasama, Northern Province, Zambia. (Developed from Matthews et al., 1992a,b) Trial no. D11 D21 D22

D31

D32

D33

Species Flemingia congesta Flemingia congesta Flemingia congesta Tephrosia vogelii Cassia spectabilis Calliandra calothyrsus Leucaena leucocephala Albizzia falcataria Flemingia congesta Gliricidia sepium Flemingia congesta Cassia spectabilis Sesbania sesban Flemingia congesta Cassia spectabilis Sesbania sesban

1987

2.46 0.67 0.94 0.83 0.54 0.72 1.05 0.56 0.86

1988

1989

1990

1.40 0.28 1.06 2.19

2.45 2.22 1.89 0.33

2.91 1.09 1.41

2.23

0.64 2.47

0.44 0.79 0.74 1.31 0.44 1.13 2.14 0.66

1.09 1.08 0.51 0.84 0.07 1.36 1.40 0.81

0.56 1.51 1.48 0.73 0.47 0.60 1.12 0.82 0.97

LOW EXTERNAL INPUT TECHNOLOGIES

507

on SOC levels? Assuming an annual biomass production of 2 t DM ha21 year21, 3 ha of land would be required to supply 1 ha of cropped land area with enough biomass to maintain the SOC level at just 1%. Significantly more would be required to raise it to the 3% level indicated earlier. In the initial years of hedgerow intercrop systems, when farmers are most likely to reject or accept a new technique, tree biomass production is likely to be well below 2 t DM ha21 (e.g., Matthews et al., 1992a). This is clearly insufficient to maintain SOC at even the 1% level. From the farmer’s perspective, if this biomass is to be grown on-farm, the area available for food crop production must necessarily be less. If it is to be supplied from outside the farm, the transport of such large quantities of organic material would require considerable labour and physical effort (Rao and Mathuva, 2000; Hoang Fagerstro¨m et al., 2001). Furthermore, labour will be required for incorporation of the biomass into the soil. In many situations, therefore, it may simply not be possible to use plant biomass to increase SOC within farmer constraints. Supplying adequate amounts of organic matter through animal manure is also difficult. The dry matter quantities required for soil physical improvement are similar to the amounts required when using plant biomass (Euroconsult, 1989). Pilbeam et al. (1999b), in deriving an N balance for a hypothetical household with 1 ha of agricultural land in the mid-hills of Nepal, estimated that the total feed requirement for buffalo, assuming a live-weight of 450 kg, was about 2.6 t DM year21. For cattle, assuming a live-weight of 250 kg, feed requirements were estimated to be 1.8 t DM year21. Thus, to balance the SOM losses given by Young (1989), about 3 buffaloes or 5 cows would need to be kept for every 1 ha in humid regions and one buffalo or 1.5 cows per hectare in semi-arid regions, assuming that nearly 100% of the consumed biomass passes through the animals. For comparison, Nandwa and Bekunda (1998) calculated that between 2 and 8 cattle would be needed to supply enough nutrients for a 2—3 t ha21 maize crop. For a typical farm in the mid-hills of Nepal, the analysis by Pilbeam et al. (1999b) suggests that about 2.5 t of animal feed comes from dry and green crop residues, presumably from on-farm sources. However, to maintain SOC levels of one hectare of land through the use of animal manure alone, this still leaves a requirement of about 6 t DM of fodder from off-farm sources. These biomass requirements are rough guides, but serve to show that the quantity of animal manure needed to maintain soil physical characteristics and nutrient levels effectively are substantial. In most cases, it is unlikely that resource-poor farmers would have access to on-farm sources of manure in sufficient quantities to supply the total OM requirements of their cropped land, particularly as there are competing demands for its use, such as for building material or fuel (e.g., Nandwa and Bekunda, 1998). In such cases, reliable access to off-farm land for fodder collection will be a major requirement for the use of animal manure. In the Embu District of the Kenyan highlands, application rates of 5– 8 t FM ha21 (~2 –3 t DM ha21) are recommended to farmers

508

A. GRAVES, R. MATTHEWS AND K. WALDIE

(Lekasi et al., 1998), but average rates applied by farmers are often much higher at 11 t FM ha21 (~4.4 t DM ha21), and could even exceed 17 t FM ha21 (~7 t DM ha21). Despite these relatively high rates of application, many farmers felt that they would use even more if it were available. In general, the smaller the animal, the higher is the nutrient concentration in its manure. For example, poultry manure has much higher levels of N and P (up to 4.8 and 1.8%, respectively) than cattle manure (1.5% N and 0.14% P) (Reddy et al., 2000). The amount of poultry manure needed to supply the requisite amount of N and P to a crop would, therefore, be correspondingly lower, as would the labour required for transporting it. For example, Swift et al. (1994) estimated that 1 –2 tonnes of poultry manure would be required to fertilise a 2 tonne maize crop, compared to 7 tonnes of low-quality animal manure or 10 tonnes of straw. On the other hand, smaller animals produce lower quantities of manure per animal, and finding poultry manure in sufficient quantities could prove difficult, as it is unlikely that the numbers of poultry typically found on resource-poor farms would supply more than a few kilograms of manure annually. Overall, therefore, the quantity of biomass available to small farmers, either from plants or from animal manure, is likely to constrain the degree to which soil fertility can be maintained or even improved. Production of this biomass on farm must be weighed against the loss of any other use the land may be put to, particularly the growing of crops. Off-farm sources may be an option, as in Nepal, but this will not always be the case. However, an important effect of the addition of OM to the soil, and one that may be immediately appreciated by farmers, is an improvement in the workability of the soil, so that farming operations, particularly ploughing or hand hoeing, are eased. This is likely to be particularly important where continuous cultivation of land is already practised and SOC has decreased, and bulk density has increased as a result. For example, in on-farm trials in Ghana and Bolivia, farmers reported the benefits in terms of the “softness,” “looseness” or “coolness” of soils after using OM from cover crops and animal manure (Kiff et al., 1999; Pound et al., 1999). The extent, to which these benefits were a result of the incorporation of the OM in the soil, or the growth of cover crop, was not reported. However, what is important in that soil workability was a key factor by which farmers appreciated the effect of OM techniques in the short-term rather than in the long term. 2. Biomass Quality The “quality” of biomass is a function of its nutrient content and its resistance to breakdown. Biomass quality has two opposing effects, in that lower quality OM is likely to have a larger impact on SOM levels than high-quality material, which mineralises more rapidly. On the other hand, higher quality organic material

LOW EXTERNAL INPUT TECHNOLOGIES

509

contributes more to the nutrient status of the soil (N, P, K, and micronutrients), and is important for maintaining soil microbial activity and the soil buffering capacity. Successful OM management depends on finding the correct balance between these two effects. This applies to plant OM, whether green manure or crop residues, as well as to animal manure. Crop residues and other low-quality organic material, particularly if added in large quantities, may temporarily induce N or P deficiencies in the soil due to microbial immobilisation, thereby reducing crop yields. Palm et al. (1997b), for example, have shown that addition of OM containing less than about 0.25% P to the soil is likely to cause net immobilisation of P. Such deficiencies may have to be overcome through the use of inorganic fertilisers (Muriwara and Kirchmann, 1993). The ratio of carbon to nitrogen (i.e., the C/N ratio) in organic material is often used as a measure of its quality. More recently, the concentrations of lignin and polyphenols have also been found to be important, particularly the lignin/N, polyphenol/N, and (lignin þ polyphenol)/N ratios (Snapp et al., 1998). Highquality organic inputs are low in lignin and polyphenol and high in N (Palm et al., 1996), and as such decompose more quickly. It has been estimated that they release between 70 and 95% of their N within a season under tropical conditions (Giller and Cadisch, 1995). The quality of plant biomass is not constant, but varies with age and whether a plant is leguminous or non-leguminous. In general, young plant material has low C/N ratios ensuring that its nutrients will be released quickly when it is incorporated into the soil, while material from older plants (or plant organs) of the same species generally has a higher C/N ratio. Data on C, N, P, and K contents for a range of organic materials used in lowexternal input agriculture are available in the Organic Resources Database developed as part of the Tropical Soils Biology and Fertility Programme (TSBF) (Palm et al., 2001). The nutrient content of animal manure depends on the species (Table II), the diet of the animal, and on how the manure is collected, stored, and applied. Diet is particularly important in relation to the partitioning of N between the faeces and the urine (Snapp et al., 1998). High-quality diets (low in lignin and polyphenols) result in more N being excreted in the urine than in the faeces (Somda et al., 1995). N that is excreted in the urine is much more quickly volatilised, and urine is also more difficult to collect. Animals fed with a tannin-rich diet tend to excrete more of the N in their faeces. However, recent results suggest that this kind of N is very resistant to mineralisation (Mafongoya et al., 1997a). The beneficial 2 effects of N released from manure as NHþ 4 and NO3 appear to be directly after application. However, poor-quality manure has been found to result in prolonged periods of N immobilisation, and under these conditions, N availability has been increased with inorganic sources (Muriwara and Kirchmann, 1993). Table II shows that for manures from most domesticated animals, P content is above the threshold at which P immobilisation is likely to occur, but that N content is often below the critical limit required to prevent N immobilisation.

510

A. GRAVES, R. MATTHEWS AND K. WALDIE

Table II Mean Content of C, N, P and K in Manure from Various Domestic Animals in Muranga and Kiambu Districts, Kenya

Manure type

C (%)

N (%)

P (%)

K (%)

C/N ratio (kg C (kg N)21)

Cattle Cattle and compost Goat Pig Poultry (broilers) Poultry (local) Rabbit Sheep

35 25 32 40 41 22 33 33

1.4 1.3 1.5 2.0 2.4 1.2 1.6 1.5

0.60 0.44 0.40 1.19 1.60 0.91 0.40 0.33

0.59 0.36 0.53 0.49 0.41 0.26 0.50 0.44

26 21 22 21 17 19 20 22

(From Lekasi et al., 1998).

In the absence of inherently fertile soils and inorganic fertilisers, improved crop yields are usually achieved only with high-quality OM. Low- or even medium-quality residues have generally been unsatisfactory (Snapp et al., 1998). In Kenya, for example, Nandwa (1995) found that incorporation of maize stover (low-quality OM) reduced maize grain yield by between 3 and 30%. Maize stover has also been shown to reduce crop yields in Zimbabwe (Rodell et al., 1980; Muriwara and Kirchmann, 1993). In India, Goyal et al. (1992) found that a combination of wheat straw (low-quality OM) and urea reduced yields, while a combination of Sesbania green manure (high-quality OM) and urea increased yields compared with the application of urea alone. The reduction in yields with low-quality OM is generally attributed to the immobilisation of N by microbial growth. Added organic material with a high C/N ratio (e.g., . 20 kg C (kg N)21) provides adequate C substrate, but as the C/N ratio of microbial biomass is lower (3—14 kg C (kg N)21), N is taken up from the mineral N pool in the soil to meet the shortfall, thereby reducing that available for crop uptake. This can be a major problem if it coincides with critical growth stages of the crop. Mafongoya et al. (1997b) have shown that N immobilisation also occurs when the lignin and polyphenol content of the residues incorporated into the soil were over 15 and 3%, respectively. Interestingly, N immobilisation resulting from high polyphenol levels seems to last much longer than that resulting from low C/N ratios (Palm et al., 1996). As the C content of dry biomass is usually around 0.4 kg C (kg DM)21, the critical C/N ratio of 20 kg C (kg N)21 corresponds to a biomass N content of between 2.0 and 2.5% (Palm et al., 1997b). Table III shows the N contents for a range of organic materials, from which it can be seen that some cover crop species, the leguminous tree species, and Tithonia diversifolia all contain N levels above 2.5%. The animal manures, on the other hand, particularly

LOW EXTERNAL INPUT TECHNOLOGIES

511

Table III Mean, Maximum and Minimum N Percentages for Various Organic Materials Organic matter source Non-legume Legumes (cover crops)

Legumes (trees)

Animal manure

Crop residues

Tithonia diversifolia Crotalaria juncea Canavalia ensiformis Mucuna pruriens Leuceana leucocephala Gliricidia sepium Sesbania sesban Cattle Cattle (Lekasi et al., 1998) Pigs Pigs (Lekasi et al., 1998) Poultry Poultry (Lekasi et al., 1998) Maize Sorghum

Mean

Max

Min

3.38 3.47 2.89 3.23 3.68 3.38 3.54 1.04 1.40 3.79 2.00 4.02 2.40 1.01 0.63

4.59 6.30 4.74 6.05 6.32 5.33 4.81 4.15 2.00 4.25 2.20 6.73 2.60 3.07 0.63

1.10 0.80 0.23 0.83 1.04 1.33 1.39 0.30 0.50 3.08 1.50 1.85 2.30 0.25 0.63

Compiled from the Organic Resources Database (Palm et al., 2001) and Lekasi et al. (1998).

from cattle, tend to have mean N contents that are likely to lead to net immobilisation of N in the soil. Data from Lekasi et al. (1998) also suggest the pig and poultry manures are marginal in terms of N levels, although the data from the Organic Resources Database show more favourable levels of mean N for these animals. The low N contents in most cereal crop residues also suggest that there may be negative yield effects on the crops to which they are added, due to net N immobilisation. These issues are important when adding OM to the soil, whether it is to improve the physical or chemical characteristics of the soil. Where long-term improvement to soil physical characteristics is important, the requirement will be for moderate- to low-quality OM to be applied. However, this may result in N and P immobilisation, particularly without supplementary use of inorganic fertilisers, and crop production may therefore decline. Where an improvement to the soil nutrient status in the short-term is needed, high-quality OM with low C/N ratios should be applied.

3. Nutrient Mining In view of the large quantities of OM required and also because of the resulting uptake of nutrients during the production of this biomass, there is the question of whether such production is sustainable in the context of subsistence agriculture in the tropics. For example, biomass transfer systems,

512

A. GRAVES, R. MATTHEWS AND K. WALDIE

such as cut-and-carry grasses or fodder banks, may function effectively for a certain period of time and relieve pressure on other sources of biomass, in particular, forest biomass and on-farm sources of biomass. However, it is unlikely that such systems can exist long without additional inputs, as nutrients are “mined” from the soil in which the biomass is being produced. In Kenya, it was found that the introduction of fodder banks was effective for a short while, but it soon became apparent that their productivity was declining due to the continuous removal of biomass and hence of nutrients from the soil (Wandera et al., 1993). Similar considerations apply to the use of Tithonia diversifolia as a supply of P to crops. Tithonia is able to scavenge relatively large quantities of P from the soil and to provide biomass with relatively high concentrations of P for incorporation as OM. However, this does not add to the net amount of P in the soil, but rather is a way of redistributing it spatially. Clearly, Tithonia will eventually mine the soil of P and other nutrients, which is unsustainable at the level of the whole farm. Unless resource-poor farmers have access to large areas of off-farm land growing Tithonia, it is unlikely that the system can be sustained over long periods of time. One solution might be to fertilise the on-farm biomass banks or hedges, but in this case farmers may as well fertilise the crops directly. The sustainability of the System for Rice Intensification (SRI) technology developed in Madagascar in the 1980s by Fr. Henri de Laulani´e (Stoop et al., 2002) and promoted as a successful example of low-external input agriculture (e.g., Pretty and Koohafkan, 2002) can also be questioned from this point of view. The approach focuses on early transplanting, planting of single plants rather than hills, a square arrangement of plants rather than in rows, frequent weeding, and non-flooding of the soil during the vegetative period, resulting in less seed and water being required. Compost can be added as a source of nutrients, although this is not an essential part of the package. Huge increases in rice yields up to 21 t ha21 without the use of purchased inputs of fertiliser and pesticide have been claimed (Uphoff, 1999). Some of the individual components (e.g., adequate spacing and weeding) are commonly recommended practices, but the remarkable yield increases are claimed to arise from the “synergistic” effect of all of these used in combination (Stoop et al., 2002). Part of the success of the approach appears to be in maintaining the soil in an aerobic state rather than completely flooded as in traditional irrigated rice cultivation, resulting in higher rates of organic matter mineralisation (Stoop et al., 2002), so that more nutrients can be supplied from the soil rather than from fertiliser applications. Certainly, anaerobic conditions are known to depress mineralisation rates, and there are also reports of increases in rice yields in well-drained soils compared to flooded soils (e.g., Ramasamy et al., 1997). While the yields of 21 t ha21 are undoubtedly an extreme (and indeed questionable physiologically), average yields of 8.8 t ha21 in farmers fields have been reported (Stoop et al., 2002). Such yields will remove about 130 kg N ha21 along with other nutrients at each harvest.

LOW EXTERNAL INPUT TECHNOLOGIES

513

Even if such low-fertility soils as reported (Uphoff, 1999) are able to supply such quantities of N, it is difficult to see how such yields could be sustained for long without further addition of nutrients. If compost is to be the main source of nutrients, then quantities in the order of 33 t DM ha21 (at 2% N content and 20% recovery rate) would be required to replace the N removed by the rice crop. The 21 t ha21 rice yields were reportedly achieved with applications of 40 t FW ha21 (~16 t DM ha21) compost made from leaves of Tephrosia, Crotalaria, and banana along with rice straw (Uphoff, 1999). It is not easy to see how such quantities of compost could be produced and transported by a single household on a sustainable basis, except for very small areas of crop. Estimates of annual household waste production (potentially available for making compost) vary widely from ~100 kg DM year21 in Uganda (Wortmann and Kaizzi, 1998) to as high as 9.8 t DM year21 for households of 8 –14 people, also in Uganda (Briggs and Twomlow, 2002), although at between 700 and 1225 kg DM person21 year21, this last value appears somewhat high. Pantanali (1996) gives an intermediate value of 2 t DM year21 for the typical annual production of household waste in Lesotho, which, assuming a family of eight persons, represents about 250 kg DM person21 year21. Estimates for low-income households in South Africa are as low as 40 kg person21 year21 (Durban Metropolitan, 1999). For comparison, average household waste produced per person in England is about 500 kg person21 year21. Whatever figure is used, it would seem that household waste alone is unlikely to be able to meet the nutrient requirements of such highyielding crops, and that nutrients, either in inorganic or organic form, would have to be collected from a wider area for the enhancement of the cropped area.

4. Biological Nitrogen Fixation Biological nitrogen fixation (BNF) by legumes is a key process in LEIA technologies as it potentially results in a net addition of N to the system. However, the quantity of N fixed by legumes is difficult to quantify and varies according to the species involved and location. Webster (1998) noted that estimates of the amount of N fixed by groundnuts and grain legumes range from about 25 to 200 kg N ha21 during growing seasons of 60 –120 days. Bouldin et al. (1979) found that some legumes seemed to fix N at relatively high rates — for example, values up to 535 kg N ha21 have been recorded for Crotalaria spp. and 400 kg N ha21 for other pure legume green manure crops. Moore (1962) reported that Centrosema pubescens (star grass) fixed N at the rate of 280 kg ha21 year21. Reviewing the contribution of BNF in alley cropping, Sanginga et al. (1995) quote measured values of BNF of between 100 and 300 kg N ha21 year21 for some tree species such as L. leucocephala, Gliricidia sepium, and Acacia mangium, but values as low as 20 kg N ha21 year21 for others such as

514

A. GRAVES, R. MATTHEWS AND K. WALDIE

Faidherbia albida and Acacia senegal. Values compiled by Brady (1990) indicate that rates of N fixation by a range of legumes vary between 5 and 300 kg N ha21 year21, with an average of about 100 kg N ha21 year21. The amount of N fixed biologically that is actually taken up by the main crop is difficult to determine with accuracy. In their review, Sanginga et al. (1995) estimate the contributions of N from trees in alley crops to an associated food crop to range from 25 –102 kg N ha21 season21 from root turnover and nodule decay, with a further 40 –70 kg N ha21 season21 being supplied through prunings. This represents recoveries of N of about 30%, although values as low as 5 –10% were reported. Thus, between 65 and 172 kg N ha21 season21 could potentially be supplied to the main crop over a season, representing enough N to support maize crops yielding between 3.5 and 8.5 t ha21. These figures, however, are based on BNF rates towards the high end of the range. Taking the average value of 100 kg N ha21 year21, and assuming that the recovery by the crop was the same 20% suggested by Giller and Cadisch (1995), with the rest lost through immobilisation, adsorption, volatilisation, leaching, and denitrification, BNF would provide around 20 kg N ha21, or roughly the N requirements of a 1 t ha21 maize crop. Moreover, this figure assumes a pure stand of the legume — if it is a component of mixed species system, then the addition of N to the system would be correspondingly less (Snapp et al., 1998). Kumara Rao et al. (1981) showed that both relay and sequential intercrop systems in southern Malawi produced insufficient quantities of biologically fixed N to maintain a maize biomass yield of 4 t ha21 on infertile land. They estimated the N contribution of Sesbania sesban when relay intercropped in low-fertility areas to be 28 kg N ha21, and that from a pigeonpea/groundnut intercrop to be about 20 kg N ha21. In comparison, the quantity of N required to sustain maize yields was estimated to be about 100 kg N ha21. Similar calculations were made by Matthews et al. (1992a) for alley crops in northern Zambia — a net amount of 14 kg N ha21 year21 was estimated to be added to the system through BNF, only 13% of the local recommended fertiliser N application rate for maize. Despite this, maize crop yields of 2 – 2.5 t ha21 were achieved for 4 years, which, with an annual removal of 30 – 35 kg N ha21 year21 in the harvest, suggests that some nutrient mining may have been occurring. The growth rate of the legume is a strong determinant of the rate of BNF. Any factor that reduces this growth, therefore, will also reduce the introduction of N into the system through BNF. In particular, competitive main crops such as maize may reduce the growth, and hence BNF rates, of legumes growing in an intercrop (Shumba et al., 1990; Muza, 1995; Kumwenda et al., 1997b). For example, Patra and Poi (1998) noted that intercropping maize with various legumes caused the number of nitrogen-fixing nodules on the legume crops to decrease due to shading, while Fujita et al. (1993) found that artificial shading reduced BNF in several pasture legume species. In agroforestry systems, competition for light can be managed with timely pruning of the perennial species. It has been suggested

LOW EXTERNAL INPUT TECHNOLOGIES

515

that competition for water and nutrients can also be managed with pruning, although the effectiveness of this is still debated (Snapp et al., 1998). Nair (1993) has pointed out that controlling the below-ground competition of roots is far more difficult than controlling the above-ground competition of canopies for light. Reduced light intensity as a result of interception from a dominant main crop has also been found to reduce nodulation (Webster and Wilson, 1998). Another factor influencing the net contribution of N to the system is the amount that is removed if the legume crop is harvested for grain or forage. Where the legume produces a useable product, farmers are likely to harvest and use the product. In the case of edible grain legume crops grown either as intercrops or in rotation, harvesting the grain can result in lower than expected residual N effects on the following crop, particularly if the grain legume has a high harvest index. In Brazil, residual N benefits from using soybean as a leguminous green manure were greatly reduced by harvesting (Boddey et al., 1997). Farmers are likely to leave grain for incorporation as a green manure only where they do not market or consume it themselves. Boddey et al. (1997) have suggested that the solution may be to develop or use varieties of soybean that produce extra biomass while producing the same grain yield, thereby reducing the harvest index and allowing more N to be incorporated into the soil with the extra plant biomass. Grain legumes with a naturally low harvest index, such as pigeonpea or groundnut, or plants with no commercial or subsistence value could also be used. However, such solutions are only likely where improved N management is more important than short-term subsistence or cash benefits. Legumes may also have other important functions within the smallholder’s strategy. For example, in the semi-arid areas of northern Namibia, McDonagh and Hillyer (2000) found that for cowpea, bambara and groundnut intercrops to be able to make any contribution of N to the system, there should be no grazing or burning of legume residues; but as cattle were an integral part of the system, this was unlikely to be a popular option with the farmers. Increasing the legume plant density just to the point where it began to affect the growth of the pearl millet contributed only about 4 kg N ha21 to the system, which increased millet yields by 80 kg ha21, a somewhat insignificant amount. They concluded that grain legumes alone are unlikely to be able to improve soil fertility in the area substantially, and that external fertiliser inputs would be necessary, although the uncertain rainfall makes investment in soil fertility unattractive for farmers there. It would appear, therefore, that in many cases BNF approaches are unlikely to provide more than a small fraction of the N requirements of main crops, unless low yields are accepted, or unless the farmer has access to other sources of leguminous biomass. Certainly, the use of these techniques to supply the full N requirements of the crop will require more land under the legume than under the main crop, which deprives the farmer of land that could otherwise be used for subsistence or cash crops. The use of mixed species intercropping is likely to decrease main crop yields through competition, particularly in difficult

516

A. GRAVES, R. MATTHEWS AND K. WALDIE

biophysical conditions, where legume growth is limited, for example, by lack of P, or where lack of rainfall exacerbates competition for water. Nevertheless, Snapp et al. (1998) suggest that farmers may contemplate the use of such systems as part of the solution to N management, providing that the area available from growing the main crop is not reduced too much.

B. SOCIO-ECONOMIC ISSUES As the technologies we have discussed are ultimately used by people, social and economic factors are likely to be as important, if not more so, than the biophysical characteristics of the technologies. Amede (2001), for example, noted that in the Ethiopian highlands, adoption of various techniques such as cover crops and crop residue incorporation related as much to farm size and availability of labour as to the conditions of the soil. In this section, we discuss various socio-economic factors that can influence the successful uptake of LEIA technologies.

1.

Land

The amount of land required to produce enough biomass to maintain or improve the SOC level of cropped land or to supply sufficient nutrients to meet crop requirements has already been discussed. The calculations described previously indicate that in general, a minimum of 3 ha of land is required to produce enough plant biomass to maintain SOC at 1% or to meet the nutrient requirements of 1 ha of cropped land. Essentially, in such a system, nutrients are being harvested from a wider area to enhance productivity within a smaller area, as in some slash-and-burn cultivation systems (e.g., chitemene and fundikila in Zambia, (Matthews et al., 1992a)). In many cases, depending on the type of biomass, climate and soil conditions, this ratio may be considerably more than 3:1. In the case of animal manures, where animals are used to gather nutrients from a wider area, the ratio is likely to be higher because of the losses of nutrients during manure storage and N losses in the urine not incurred in direct incorporation of biomass in the soil. Palm et al. (1997a) estimated that between 14 and 42 ha of miombo woodland would need to be grazed to provide enough N in manure for a 2 t ha21 maize crop, while ratios of up to 45:1 have been estimated for other extensive systems (Turner, 1995). In many situations, it is very unlikely that resource-poor farmers will have access to sufficient land, over and above that needed for crop production, to produce the appropriate quantities of biomass. Thus, if organic materials are to be used solely for soil fertility maintenance or improvement, in most cases, it would seem that they must be

LOW EXTERNAL INPUT TECHNOLOGIES

517

obtained from off-farm sources such as forests or, possibly, as waste from urban areas. Even with intercropping or alley cropping, where a leguminous species is grown on the same piece of land as the main crop, there are costs; firstly, in that the area taken up by the legume cannot be used by the main crop, and secondly, due to potential competition for resources such as light, water, and nutrients. Few studies have demonstrated the possibility of achieving increases in main crop yield per unit of total system area while maintaining soil fertility. Relay cover cropping with annual legumes, or perennial legumes like C. cajan for a single season, may address the issue of direct competition, but is often not feasible due to lack of rainfall outside the main cropping season. Thus, legume growth and hence BNF rates do not compensate for the N that is removed in the harvest of the main crop. On the other hand, if rainfall does continue, farmers may be more likely to grow a second food crop rather than an unproductive legume. Improved fallows may be an option for resource-poor farmers where fallows are already an accepted part of the cropping system. However, where land is limited and continuous cropping exists, improved fallows would have to replace a main crop, which is not feasible for most resource-poor farmers. Furthermore, resource poor-farmers are unlikely to be able to meet the financial and labour costs required for the establishment of improved fallows. A natural fallow, particularly one of several years, may provide resources that an improved fallow cannot. Plants, animals, or residual germination and growth of cultivated crops provide important products for farmers and soil regeneration under a natural fallow may not be very different compared to that under an improved fallow, especially if the natural fallow contains leguminous plants. Suggestions have been made that field boundaries could be used to source nutrients, particularly P, from biomass transfer species such as Tithonia diversifolia (e.g., Briggs and Twomlow, 2002). While this may make some contribution, a quick calculation suffices to show how much P, for example, might be supplied in this way. If we assume that a landholding is 1 ha in area consisting of 3 –4 fields, (typical for many small-holdings in Nepal and Ghana), there would be about 600 m of field boundaries. Tithonia growing on these boundaries producing 1 kg DW m21 from biannual pruning (Jama et al., 2000), would therefore supply only 0.6 t DM ha21, or about 2.2 kg P ha21, equivalent to about 15% of the 15 kg P ha21 removed in the harvest of a typical 5 t ha21 maize crop. Alternatively, this amount of P would only support a 0.7 t ha21 maize crop in P-deficient soils, and probably far less if due consideration is given to the usual crop recovery rates of nutrients from organic material. It must also be borne in mind that the Tithonia hedge has probably extracted some of this P from the adjoining cropped fields in the first place. In some situations, there may be ways of making better use of land that is unproductive at certain times of the year. For example, in the Barind Tract of Bangladesh, land is often left fallow during the dry season following the main

518

A. GRAVES, R. MATTHEWS AND K. WALDIE

rice crop, as it is often difficult to establish a second crop during this time due to drying of the soil surface layers (Musa et al., 2001). This is despite the fact that there is usually sufficient residual water further down in the soil profile left from the irrigated rice crop. In recent years, drought-tolerant crops such as chickpea have been introduced, but establishment is not certain, and complete crop failure may result. Recent research has shown that “priming” chick-pea seed, by soaking it overnight in water, can result in a marked improvement in crop establishment, making the difference between a healthy crop and no crop at all (Musa et al., 2001). Attractions of the technique are that it is simple for farmers to implement, requires no expensive inputs, and is aimed at resource-poor farmers rather than at those with mechanised systems (Harris et al., 1999). Farmers welcome the ability to gain an extra crop in the sequence, particularly of chickpea, which currently commands good prices in the market, at little extra cost in land, labour, or capital. Of course, there are questions of whether such systems of increased intensification are sustainable, or whether soil fertility decline and possibly weed encroachment is enhanced, although the experience in Bangladesh would suggest that this is not the case. The area in question was converted from forest about 150 years ago, and although current SOM levels are very low (0.5 –0.8%), reasonable main crop yields seem to be obtained year after year with appropriate inputs of inorganic fertilisers. Whether a further crop in the sequence will reduce SOM levels even lower remains to be seen. Land tenure is also an important issue influencing the use of LEIA technologies. It is generally thought that land users who do not own their land have less incentive to invest in technologies that take some time for soil fertility benefits to accrue. Share-cropping is one such example, where a farmer exchanges a proportion of farm output in exchange for the right to crop an area of land (Ellis, 1988). The system has generally been considered as being less economically efficient than if the land is owned, as share-croppers are thought to input labour only at a level that maximises their own perceived share of farm production, which is less than what they would be prepared to give if they receive the total production from the farm (Todaro, 1999). Nelson et al. (1998) analysed the economics of upland agriculture in the Philippines, concluding that share-cropping would reduce the economic attractiveness of alley cropping techniques compared with alternative techniques (Fig. 2). This was because it was assumed that, under the particular share-cropping arrangement, landlords would not contribute to the establishment costs of the hedgerows, while a portion of the main crop would be given to them as part of the tenancy agreement. However, other analysis suggests that share-cropped farms are not necessarily inherently less efficient than ownermanaged farms (Reid, 1976). Also, share-cropping may at least give very poor farmers the opportunity to farm, and some evidence suggests that it is the poorest and most unskilled who stand to benefit from share cropping (Reid, 1976), especially where the landlord also wishes to maximise the productivity of the farm. Similarly, Ayuk (2001) in reviewing a number of studies, concluded that

LOW EXTERNAL INPUT TECHNOLOGIES

Figure 2 (a) The impact of sharecropping on the net present value of open field, fallow and hedgerow intercropping in upland Philippine agriculture at a discount rate of 25%. (b) The net present value without the impact of sharecropping is also shown for all three systems at a discount rate of 25%. (Source: Nelson et al., 1998).

519

520

A. GRAVES, R. MATTHEWS AND K. WALDIE

the relationship between land ownership and willingness to invest in technologies to improve SOM was not a simple one. While in some cases, there was a strong correlation between long-term use rights and soil fertility improvement in the form of manuring and mulching (e.g., Blarel, 1994), in other cases, there was no correlation at all (e.g., Golan, 1994). Interestingly, a number of studies indicate that traditional tenure systems do not inhibit land users from investing in improving their land, and that a greater danger may be the granting of titles to urban dwellers who have no interest in agriculture (Ayuk, 2001). Fragmentation of land is another factor that may influence the adoption of certain LEIA technologies. Fragmented landholdings may result in a single farmer having to transport inputs to several isolated plots of land in several different locations. This difficulty is particularly great with the use of biomass transfer techniques, where several tonnes of biomass per hectare may be required. In the mountainous terrain found in Nepal or Bolivia, for example, transporting heavy loads to small isolated plots of land is extremely arduous. This is an extremely labour-intensive process if the farmer has to do it alone, or expensive if he/she hires labourers. Moreover, land fragmentation may result in decreasing field sizes, which makes the implementation of certain techniques impractical. Trees used for green manure or fodder, for example, may shade out the crop if planted on the borders of small fields.

2.

Labour

The labour required to make use of LEIA technologies may also constrain their adoption. Often household labour may need to be supplemented with that purchased from off the farm. While such contributions are seldom accounted for in analyses of technology costs and benefits, especially when female labour is involved, it is important to recognise that the use of domestic labour represents a real opportunity cost. The need for external labour, which will generally involve a cash transaction and therefore directly affect household finances, is more readily acknowledged. Both aspects are important. Ali (1999), in a study of farmers in Asia, found the cost of labour to be partly responsible for making nutrient supply through organic matter less cost-effective than through mineral fertilisers, a situation which is likely to get worse due to rapidly rising wages. In India, the requirement for a pair of bullocks and a ploughman was 10.5 days ha21 in a rice/green manure/rice rotation, while in Nepal, the number of days needed in a wheat/green manure/wheat system was 11 days ha21. In both cases, the cost of this was about US$40 ha21 which largely accounted for the differences in economic performance between green manure and mineral fertilisers. This occurred despite the reduced need for weeding and the increased yields obtained. Ali (1999) also analysed the economic

LOW EXTERNAL INPUT TECHNOLOGIES

521

performance of using grain legumes rather than green manures, and found that these were more profitable, despite increased labour requirements, because the legume produced grain that could be sold. However, acceptance was not as high as expected because the incorporation of the legume biomass delayed the planting of the monsoon crop and increased the burden of labour at a time when it was already high. The grain legume was also susceptible to pests and diseases. Similarly, the cutting and carrying of biomass, such as from Tithonia, is extremely labour intensive, particularly if it is to supply the full crop P requirements in a P-deficient soil (Buresh and Niang, 1997). For a typical crop requirement of 15 kg P ha21, the application of about 20 t DM ha21 of Tithonia biomass would be needed (assuming a 20% recovery rate of the P by the crop), equivalent to about 100 t ha21 of fresh biomass. Harvesting at the rate of 100 kg FW man-day21 (ICRAF, 1997), 1000 man-days of labour would be needed just to harvest the biomass! These are not the only costs. Tithonia also needs to be propagated and prepared for incorporation in to the soil (ICRAF, 1997), and, although it does not have thorns, it is difficult to handle because it is sticky and exudes a pungent smell (Jiri and Waddington, 1998). In addition, because of its ability to regenerate, it may invade farmland (Jama et al., 2000), thereby increasing the labour required by a farmer to control it. The implication is that either labour must be plentiful and cheap, or that the crops fertilised with Tithonia should be high-value crops. As an example, Jama et al. (2000) cite data from ICRAF showing that under farmer-managed conditions in western Kenya, investing in Tithonia fertilisation was viable for high-value kale (Brassica olecacea), but uneconomical when used with a low-value crop such as maize. Supplying sufficient P through animal manure also requires substantial labour. For example, the supply of 15 kg P ha21 through cattle manure (using a mean concentration of P of 0.138% and a 20% recovery rate of P by the crop) would require 55 t DM ha21 of manure, equivalent to 275 t FW ha21. Assuming that the farmer had to transport the manure manually, that 20 kg of manure per load could be carried at 5 km hr21, and that this load had to be transported 100 m from the source, transporting this 15 kg P would require about 550 man-hours of labour. In comparison, transporting the same amount of P in poultry manure, the same distance would take only about 43 man-hours. Loads may often have to be carried much further than this, and where large amounts of manure have to be transported long distances, farmers may struggle to provide the labour required. Where the manure has to be transported in mountainous terrain, such as in Nepal, the amount of time required to transport manure will be even greater. For many farmers, weeding is one of the most labour-demanding activities undertaken. Gill (1982) noted that hand-hoe weeding in India required between 200 and 400 man-hours ha21 and that two weedings were needed during the growth and development of many field crops. Van Tienhoven et al. (1982) found that between 13 and 37 man-days ha21 of labour were required to weed a maize – bean production system in the Jinotega region of Nicaragua.

522

A. GRAVES, R. MATTHEWS AND K. WALDIE

This accounted for between 21 and 35% of family labour. Ruthenberg (1980) has compiled data of labour requirements for weeding from various sources — in Ghana, for example, weeding in a maize system required 31% of the total labour (about 186 man-hours ha21), while in Columbia, weeding a cassava crop required about 55% of total farm labour (about 408 man-hours ha21). Other traditional agricultural systems cited were less intensive, although they all required at least 20% of total labour requirements for weeding. In the Ichilo-Sara area of Bolivia, Pound et al. (1999) found that weeding could require from 35 man-days ha21 of labour in the first year of a cropping cycle to 53 man-days ha21 in the third year as weeds started to dominate the system. The increase in weed cover was associated with large decreases in rice yield, with yields in the third year only about 30% of those in the first year. This was probably due to the combined effect of weed growth and declining soil fertility. The labour requirement for weeding was reduced when Calopogonium was sown as an intercrop 25 days after the rice planting, but not if the Calopogonium was sown 45 days after the rice planting or as a cover crop after the rice was harvested. Pound et al. (1999) make the point that these reductions in labour were probably not great enough to be of practical significance to subsistence farmers, and clearly, would not make a substantial difference to their livelihoods. The system for rice intensification (SRI) discussed above (Stoop et al., 2002), provides an interesting example of how labour requirements can limit the uptake of an improved practice. Although labour requirements are high (38 – 54% more than traditional methods), returns to labour are also high ($3.87 day21 compared to $2.61 day21), a characteristic which has been seen as a major advantage of the technology compared to traditional approaches. Despite these apparent advantages, farmer adoption of SRI in the areas where it was promoted has been low, abandonment of the method by those farmers who originally adopted it has been high, and those who continue to practice the method rarely do so on more than half of their land (Moser and Barrett, 2002). Participatory surveys showed that this was because the recommended technological package required significant additional labour inputs due to the extra weeding and water management involved (the latter on a daily basis), during the time of year when poorer farmers need to seek employment with other farmers to earn cash to meet immediate consumption needs. Those who did try SRI were less likely to rely on agricultural day labour as a source of income, and were also more likely to have larger farms due to the economies of scale in offsetting the initial costs involved in levelling their fields and in redesigning their irrigation systems to allow more precise water control. Interestingly, adopters were also less likely to have a relatively high salaried income as the opportunity cost of foregoing their salary to supervise hired labour for the SRI was greater than the returns gained. Thus, poorer farmers for whom the technology was designed to help were less able to take advantage of it. This example highlights the need to evaluate new technologies in the context of the whole agricultural system — even

LOW EXTERNAL INPUT TECHNOLOGIES

523

considering returns to labour in this case was not sufficient to predict the uptake of SRI, as the timing of labour and income requirements throughout the year was the most important factor. Assertions by the proponents of SRI that “farmer perceptions and practices are not necessarily wise or optimal” (Uphoff, 1999) may, therefore, be somewhat premature when the wider picture is taken into account. Moser and Barrett (2002) make the point that opportunity cost is often overlooked in evaluating LEIA technologies, i.e., even though the financial cost of inputs may be low, this does not mean that the technology is without other costs. They offer an interesting comparison with the technique of off-season cropping (OSC), which was introduced in the same areas in Madagascar as SRI, in which crops such as potatoes are grown in the winter season after the rice harvest. Adoption of the OSC technique has been high at 84% of households (across a range of wealth classes), and with no disadopters to date (2002). The key to its success appears to be the ease with which it fits into the existing agricultural system. Despite relatively high input costs for the purchase of seed and fertiliser, these occur at a time when farmers have just completed the rice harvest, and have more time and money. Moreover, the OSC harvest provides them with food and working capital at the beginning of the rice season, freeing the household from needing to do off-farm work to earn wages to purchase food. Farmers also perceive a carry-over benefit of the fertiliser applications on soil fertility in the following rice crop. A large proportion of the farmers adopting the technique had learned of it from other farmers, suggesting that it was easy to learn. The importance of the labour profiles of new technologies fitting in with existing labour patterns has also been noted by Hoang Fagerstro¨m et al. (2001) — biomass banks of Tephrosia were not popular with farmers in Vietnam as the labour required for cutting and transporting the biomass coincided with busy periods for other farm activities. On the other hand, a two-year crop/twoyear fallow cycle for upland rice fitted in well with the existing split of work between upland and lowland cultivation.

3.

Economics

In addition to its labour requirements, the adoption of a particular technique will also depend on its economic benefits in relation to other options, regardless of its biophysical attributes. For example, in a study of the economic viability of combined fertiliser, green manure and grain legume techniques, Ali (1999) found that the short-term benefits of green manure were “negative or trivial” compared with inorganic fertiliser, despite on-farm experiments showing that yields were higher with combined fertiliser and green manure treatments. On the other hand,

524

A. GRAVES, R. MATTHEWS AND K. WALDIE

farmers using grain legumes obtained short-term benefits from the sale of the grain as well as long-term benefits, compared to inorganic fertiliser systems. Ali (1999) concluded that recommended techniques need to have short-term economic benefits, or risk being rejected by farmers despite any long-term advantages they might have. The relative costs of inorganic sources of nutrients is a major determinant of the use of LEIA techniques. In Nepal, for example, Pilbeam et al. (1999a) showed that margins were generally negative when manure was applied either alone or in combination with fertilizer, but positive with applications of fertilizer. Similarly, in Zambia, subsidies on N fertiliser in the 1980s stimulated many subsistence farmers to adopt high-input maize production based on mineral fertiliser inputs (Matthews et al., 1992a). An economic analysis showed that under these conditions, the best net returns were obtained from applying as much mineral fertiliser N as possible (up to 120 kg N ha21 in the study) rather than using alley cropping (Matthews et al., 1992b). However, when the fertiliser subsidies were removed in 1990 through donor pressure, farmers reverted to their traditional low-input chitemene and fundikila shifting-cultivation systems. Under these conditions, alley cropping with Leucaena always gave a small positive return, and indeed, the highest net return was obtained from Leucaena and 60 kg N ha21 fertiliser. This analysis did not, however, include the cost of labour; if it had, the positive net returns from alley cropping would probably have disappeared. Sometimes, however, the yield increase under alley cropping may be economically viable. Chianu (2002) used partial budget analysis to show that compared to a bush fallow, alley cropping L. leucocephala becomes advantageous during longer fallow periods due to the production of fuelwood. However, it was noted that yield variability, labour scarcity, and risk aversion could influence the technology choice of the farmer. Similarly, novel alley cropping systems may be economically viable to farmers. In India, introducing geranium (Pelargonium spp.) into alleys of Eucalyptus citriodora did not affect the essential oil yield of the latter, but resulted in higher monetary benefit over sole Eucalyptus plots (Singh et al., 1998). Intercropping with Java citronella and lemongrass also resulted in higher net benefits than from Eucalyptus alone, although lemongrass did reduce Eucalyptus yields. In India, Pakistan, the Philippines, and Indonesia the prices of both land and labour have increased relative to inorganic fertiliser since the 1970s; consequently, the use of the latter has increased at the expense of labourintensive, land-extensive, organic matter approaches to maintaining soil fertility (Ali, 1999). In Taiwan, for example, green manure crops have decreased from an area of 153,000 ha in 1954 to 11,000 ha in 1991, while in India, Nepal, and Pakistan, green manure is no longer widely used. Ali (1999) calculated the economics of using Sesbania, Azolla, and rice straw as green manures in India, Indonesia, and the Philippines, assuming that a Sesbania green manure crop would provide 70 kg N ha21, Azolla 30 kg N ha21, and rice straw 18 kg N ha21.

LOW EXTERNAL INPUT TECHNOLOGIES

525

Ignoring the opportunity cost of growing green manure on productive land, he found that the benefit/cost ratio of all three techniques was less than 1.0. This indicated that at these labour/fertiliser price ratios, green manure was not a costefficient option in any of the three countries compared to inorganic fertiliser. Ali (1999) calculated that for green manure to be competitive with fertiliser and with a zero opportunity cost of land, the ratio of wages (US$ day21) to fertiliser price (US$ kg21 N) should not exceed 3.0 for Sesbania and Azolla, and 2.0 for rice straw. As the ratio was higher than this in India, Indonesia, and the Philippines, this explained the low use of green manure techniques there, while a more favourable ratio accounted for the greater use of green manure in Myanmar and China. Interestingly, Swinkels et al. (1997) found that farmers in the high population density areas of western Kenya were more likely to practice fallowing due to the availability of off-farm income, contradicting the view that continuous cropping on depleted soils is the final stage in the land intensification process (e.g., Ruthenberg, 1980). Cropping depleted land gave farmers poorer returns to labour, and, therefore, it was more rational to take work to buy food and allow the soil fertility to regenerate. Only a 21% increase in crop yield following a one-year fallow was necessary to compensate for the loss of returns during the fallow, mainly due to the savings in crop husbandry costs. Similar behaviour has been observed in Zambia (Low, 1988), Kenya and India (Dewees and Saxena, 1995), and Indonesia (Nibbering, 1991). As land and labour prices rise, there is pressure to shift away from green manure towards intercropping, grain legumes, compost, and animal manure. In India, farmers growing Crotalaria juncea and Tephrosia purpurea as green manure after winter rice moved to grain legumes such as C. cajan and Vigna spp., particularly if irrigation water was available. In the central terai of Nepal, Crotalaria juncea and Sesbania rostrata have given way to grain legumes such as mung-bean, particularly if water is available during the dry season. In many cases, only dramatic increases in fertiliser prices due to scarcity of fossil fuels may make green manure viable in these countries, because transport networks are well developed (Ali, 1999). Grain legumes, on the other hand, have potential if the grain has economic value, although the removal of nutrients in the grain will largely undermine their utility as a source of N. Economic analyses of improved fallow systems have given conflicting results. In upland northern Vietnam, a Tephrosia fallow had a negative net present value (NPV) and was judged not to be a rational choice where natural fallow was still viable (Hoang Fagerstro¨m et al., 2001). In Mali, Kaya et al. (2000) concluded that improved fallows were only financially attractive if fodder had a value and if subsequent crop yields exceeded the regional average of 2500 kg ha21. This is in contrast to the example in western Kenya, where improved fallows were seen to be a rational choice because of the relatively small increase in crop yields required to break even, provided labour demands of establishing the fallow were low (Swinkels et al., 1997). A study in West Africa showed that cumulative net

526

A. GRAVES, R. MATTHEWS AND K. WALDIE

Figure 3 Cumulative discounted net benefits for continuous maize with no fertiliser, continuous maize with recommended, one-year Sesbania fallow fertiliser, two-year Sesbania fallow, and threeyear Sesbania fallow. (Drawn from data of Kwesiga et al., 1999).

benefits were $1664 ha21 under a Pueraria fallow, $1121 ha21 under Chromlaena odorata natural fallow and cover crops, and $1113 ha21 under continuous cropping at farmer input levels (Tian et al., 2001). In Cameroon, Adesina and Coulibaly (1998) found that improved fallows of Tephrosia, Sesbania, Mucuna, and Calliandra were all profitable whether used alone or in conjunction with inorganic fertilisers. In Zambia, Kwesiga et al. (1999) showed that one- and two-year Sesbania fallows gave higher net returns than continuous unfertilised maize crops, while a three-year Sesbania fallow was the same (Fig. 3). Fertilised continuous maize, however, gave more than twice the net return of the best fallow treatment. These authors make the point that the timing of cash flow from improved fallows is important for subsistence farmers — even though the net return of the 2-year fallow was higher than the unfertilised maize after 4 years, for the first 2 years the net returns were negative, and the farmer would need to have other sources of food and income.

IV. DISCUSSION A. INTEGRATED NUTRIENT MANAGEMENT In general, LEIA technologies aim at reducing losses, while also introducing inputs from organic sources, while HEIA technologies aim at ensuring that

LOW EXTERNAL INPUT TECHNOLOGIES

527

agronomic inputs (i.e., fertilisers, pesticides, water, energy) into the production system are maintained at a high level, with less attention being paid to reducing losses from the system. Gregory et al. (2002) refer to these two approaches as Type I and Type II intensification, respectively. It is clear from the analysis in this review that the LEIA techniques we have discussed cannot alone meet crop nutrient requirements if potential yields are to be obtained. On the other hand, sole reliance on HEIA technologies is also not an option for many resource-poor farmers due to cost and availability of the inputs required, and is also not desirable because of the associated health and environmental pollution problems. There is, therefore, a growing consensus amongst researchers that the debate on whether HEIA or LEIA technologies are the most appropriate is largely irrelevant, and that the best way forward is through an INM approach, in which a combination of techniques from the two extremes is used (e.g., Sanchez et al., 2001). INM has been defined as the “judicious manipulation of nutrient stocks and flows, in order to achieve a satisfactory and sustainable level of agricultural production” (e.g., Deugd et al., 1998). Certainly, there is ample evidence (e.g., Palm et al., 1997a; Jadhao et al., 1999; Prasad et al., 2002) that the highest yields and returns can be obtained by a combination of maximising inputs into the system, both from organic and inorganic sources, and at the same time, reducing losses. Inorganic fertilisers have the advantage that nutrient concentrations are much higher than in organic material, so that handling and incorporation into the soil is greatly facilitated. On the other hand, organic matter in the soil acts as both a “binder” for added inorganic nutrients so that they are less likely to be lost by leaching and volatilisation and more likely to be taken up by a crop, and also as a source of nutrients in its own right through decomposition. It is also critical in improving the physical structure of the soil. The resulting greater efficiency of nutrient use through the combined use of both inorganic and organic sources of nutrients and reduced losses is referred to as Type III intensification by Gregory et al. (2002). The challenge, therefore, is to identify those techniques from the continuum between the HEIA and LEIA extremes, which can complement each other to achieve this goal of sufficient and sustainable production, regardless of whether they are labelled “organic” or “inorganic.” In general terms, Sanchez et al. (2001) have proposed a combination of (1) biological N fixation by short-term leguminous fallows, (2) applications of mineral P fertilisers, (3) enhanced P cycling using Tithonia, (4) use of trees to maximise nutrient cycling, (5) return of crop residues, (6) soil erosion prevention, (7) improved crop management practices such as the use of better varieties, and (8) improved availability and timeliness of supply of inorganic fertilisers. For this approach to be feasible, ways need to be found to extract the indigenous deposits of rock phosphate present in many countries, particularly those in sub-Saharan Africa (Ayuk, 2001), and make it available to farmers at reasonable cost. At the farm level, improved management of organic resources, such as in the storage and application of

528

A. GRAVES, R. MATTHEWS AND K. WALDIE

compost and manure, may help reduce losses by volatilisation (Pilbeam et al., 1999b). Similarly, the development of alternative energy sources such as woodlots and the introduction of more efficient stoves may help to free up crop residues currently used as fuel sources, which could then become available for SOM maintenance and improvement (Ayuk, 2001). Many of the nutrient management approaches in subsistence systems involve moving nutrients from one part of the landscape at various scales to another where they are more useful. For instance, several of the techniques we have reviewed involve collection of nutrients from a wider area through grazing or biomass harvesting and concentrating these on a smaller cropped area. In Nepal, for example, the fertility built up in forested areas over long periods of time is transferred gradually to nearby farms through collection of fodder for animals and biomass for enhancing on-farm soil fertility (Pilbeam et al., 1999b). In much of the analysis of farming systems, however, this heterogeneity is often ignored and the fertility of a farm is assumed to be constant across all fields or parts of the farm. In reality, farmers are usually very adept (consciously or subconsciously) at manipulating this heterogeneity to improve their livelihoods. Both Wortmann and Kaizzi (1998) and Briggs and Twomlow (2002) describe flows of organic material from distant maize fields to higher value banana plantations nearer the household on smallholder farms in Uganda. In many countries, the soil fertility of small areas used as home gardens is enhanced by incorporating household waste containing nutrients gathered from a wider area, both from other parts of the farm through consumption of crops harvested there, and from off-farm sources such as food bought in a market. Higher-value crops such as vegetables or fruit, which would not grow well elsewhere on the farm, are often grown in these highfertility areas. Indeed, Sanchez et al. (2001) have suggested that the growing of high-value crops may be the most direct way out of poverty. For example, they quote highvalue vegetables such as kale, tomatoes, and onions in Kenya as being able to increase net profits from US$91 to US$1665 ha21 year – 1. Whether this is viable on a large scale will depend on broader economic development and the availability of markets, storage and processing facilities, and urban population growth rates. High-value tree crops may also be promising. For example, extractions from the bark of Prunus africana can be used to treat prostate glandrelated diseases, and has an annual market value of $220 million per year (Sanchez et al., 2001). The demand has been so high that the species is now on the CITES (Convention on International Trade in Endangered Species) list, but is now in the process of being domesticated. Other examples include bush mango (Irvingia gabonensis) in West Africa, and Sclerocarya birrea from the miombo woodlands of southern Africa, which are used to make liqueurs. The danger is, however, that without replenishment in some form or another, eventually the fertility of those parts of the farm providing nutrients for these high-value crops will decline to a point where it is no longer possible for them to act as a nutrient source.

LOW EXTERNAL INPUT TECHNOLOGIES

529

Clearly, such systems are complex, and there are questions regarding overall sustainability, but with the advent of new crops, new markets, and better infrastructure in many regions, it may be worthwhile re-examining these traditional systems of nutrient redistribution carefully to see if there are any possibilities of adding value to the overall production system through innovative practices that farmers have not yet discovered. The NUTMON project is an example of an approach to develop INM strategies at the farm level (de Jager et al., 1998). The prime objective of the project was to investigate technologies that mitigate against nutrient depletion and possibly add nutrients which are economically viable and socially acceptable. This was done by monitoring nutrient inflows and outflows of a farm, calculating the balances, and quantifying the impact of INM practices on soil fertility, and hence agricultural production and sustainability. The approach included a diagnosis phase during which both qualitative and quantitative assessments of nutrient management and stocks were made from farmer interviews and by taking and analysing soil samples. Then, the most appropriate technologies for a specific farming system were determined from a combination of the quantitative nutrient flows and balances, the economic performance indicators, and farmers’ perceptions, which were subsequently tested through onfarm trials. Existing indigenous or science-based technologies as well as any new ideas or modifications of existing technologies were considered. The approach allowed the study of a farm in a holistic way, taking into account the effects of many different household activities on the nutrient stocks and flows and the economic performance of the farm, providing a way to assess the constraints to adoption of alternative INM technologies in relation to economic viability and demand for labour. Briggs and Twomlow (2002) followed a similar approach in determining flows of organic material within smallholder farms in southwest Uganda. By helping farmers to conceptualise and draw diagrams of these flows within their farms, several sources of organic material, such as hedgerows, weeds, fallowed areas, and ash residues, were identified which farmers had not previously recognised as potentially contributing to the household’s fertility management practices. At the national level, Cuba provides an interesting example of how INM practices can work (Carney, 1993). With the collapse of the Soviet empire at the end of the Cold war in 1989, the country was deprived of a source of imported fertilisers, pesticides, and fossil fuel. To cope, it was forced to look for other ways of sustaining its agricultural production, and focused on more efficient nutrient recycling and reuse of organic urban waste, and biological pest control. Although food production initially dropped by 30%, it has since risen steadily, and to date (2003) there are no food shortages despite soils being severely degraded, a high population, and a continuing economic blockade. Better integration between rural and urban areas may be a key focus for other countries — in general in developing countries, there is a net flow of nutrients from the former to the latter

530

A. GRAVES, R. MATTHEWS AND K. WALDIE

in the form of produce for human and animal consumption, which is not often returned due to the inability or reluctance to process human and other waste. In Ghana, for example, poultry manure produced in urban areas is considered to have little value and is often dumped on the roadside (Quansah et al., 2001). Thus, ways in which these nutrients can be retained and returned to the rural areas, rather than being lost to the groundwater or sea as usually occurs, need to be developed, taking into account, of course, the costs involved in processing and transport, and possible disease risks. Currently, the economics are not favourable — Palm et al. (1997a) showed that in Kenya, N and P in urban compost cost US$0.5 kg21 and US$1.2 kg21, respectively, compared to US$0.42 kg21 and US$0.18 kg21 in purchased inorganic fertilisers. It is unlikely that there are any techniques that will provide universal solutions — it is much more likely that progress will be made by farmers taking an idea and adapting it to their own particular “microniche.” These microniches will be unique for every farmer — not only will the biophysical environment vary, but each farmer will also have different perspectives based on their interests and experiences and be influenced by his/her own particular worldview (Scoones and Toulmin, 1998). It is perhaps more important that farmers have a good understanding of the principles involved in nutrient management, pest management, and crop interactions, rather than a detailed knowledge of a particular technique, and know how to apply this understanding to their own situation. Deugd et al. (1998) have emphasised the importance of any improvements to nutrient management being through farmer participation and learning. Thus, improvements should not be seen as optimal solutions to scientifically welldefined problems, but more as stages in an adaptive learning process within a complex and changing environment.

B. A SYSTEMS PERSPECTIVE It is evident from many of the examples we have reviewed, that a major reason for the lack of widespread uptake of particular technologies is the failure to see them as part of a larger system. For example, the decline in numbers of farmers using the maize –Mucuna system in Honduras was found to be due largely to external socio-economic factors independent of the agronomic performance of the system itself (Neill and Lee, 2001). Similarly, despite the large increases in rice yields and high returns to labour claimed for the System of Rice Intensification, it has not been adopted widely by farmers, and a sizeable fraction of those who did adopt it are now in the process of abandoning it (Moser and Barrett, 2002). This was ascribed to the large amount of labour required by the technology at the time of the year when poor farmers, facing cash shortages, need to work for other farmers to earn enough to buy food. It remains to be seen whether provision of credit facilities can ease this financial pressure to seek

LOW EXTERNAL INPUT TECHNOLOGIES

531

off-farm work so that technologies such as SRI can be adopted. It is also questionable whether the large quantities of biomass required for making the compost used in the system would be available for the sustainable uptake of the technology on a large scale, both in terms of the numbers of farmers involved, and in terms of the actual cropped area under the technology. This biomass must of necessity, come from a wider area than that actually cropped. Widespread uptake of the Machobane system in Lesotho (Pantanali, 1996), and the Zai system in Burkino Faso (Roose et al., 1999), both of which also rely on large supplies of compost for their success, is likely to face the same constraint. Thus, failure to consider the whole livelihood system, including the influence of temporal flows of labour and money on farmer decision making, can lead to false conclusions and unrealistic expectations on the part of the scientists about the appropriateness of a particular technology. Ironically, the SRI technology has been more widely adopted by middle-income farmers so far than by the poor farmers it was designed to help (Moser and Barrett, 2002). It is important, therefore, for researchers to try and view the situation from a farmer’s perspective (Douthwaite et al., 2002). In general, most researchers conceptualise problems associated with LEIA systems in developing countries in purely biophysical terms, particularly in relation to soil fertility decline and weed encroachment; less attention has been paid to the socio-economic issues (Ayuk, 2001). While this approach is useful from the perspective of scientific research, farmers seldom think in these terms. Rather, they are more concerned with how particular practices relate to their broader livelihoods. In considering whether or not to adopt a particular research product such as alley cropping, they are more likely to be influenced by how their livelihood will benefit in terms of the extra food, cash, or quality of life it is likely to provide, than they are from technically framed arguments concerned with nitrogen fixation and the like. This is not an argument about a “reductionist” versus an “holistic” approach. The fallaciousness of drawing a dichotomy between the two approaches has been pointed out by Kline (1995). Neither is superior to the other, and we would argue that a “reductionist” approach is essential, provided it is contextualised within a broader framework of analysis. An example of such a framework is the sustainable livelihoods (SL) approach currently being promoted by a number of aid agencies as a way of thinking more broadly about the objectives, scope and priorities for development, in order to enhance progress on the elimination of poverty (Ashley and Carney, 1999). The main feature of the SL approach is that it places people at “centre stage” of these discussions, rather than natural resources or commodities as has been the case in the past, and considers their assets (natural, human, financial, physical, and social capital) and their external environment (trends, shocks, and transforming structures and processes). A key concept is that of “sustainability” — a livelihood is defined as sustainable where it can cope with (and recover from) stresses and shocks, and maintain or enhance its capabilities and assets both now and in the future, while not undermining

532

A. GRAVES, R. MATTHEWS AND K. WALDIE

the natural resource base (Carney, 1998). An important point is that agriculture is seen as a system, rather than as simplified issues arising from single discipline perspectives. We would argue, therefore, that a more appropriate approach might be to consider the farmers’ problems from a livelihoods perspective, identify potential solutions to these problems, and match or develop relevant techniques in order to achieve these solutions. An illustrative example is provided in Table IV. The three broad problems we have defined in Table IV are included in the livelihood outcomes of the SL framework, while the various techniques we have been reviewing contribute to the various livelihood strategies that subsistence farmers can adopt to achieve these outcomes. We do not claim that Table IV is exhaustive, but have attempted to present a different way of looking at possible interventions, one that perhaps corresponds more closely to problems perceived by farmers. Whether or not a particular technique is adopted will depend on the balance between the perceived benefits and the costs of obtaining these benefits, particularly, but not exclusively, in terms of the land, labour, and capital that is required. Although the abilities of a technique to meet researchers’ expectations, such as soil fertility enhancement or better weed control, is important, these abilities are not necessarily how the farmers value them. This approach also allows the consideration of other options besides natural resource management techniques. For example, a cash-generating activity might be for some of the household members to seek work in a local town or abroad. In many cases, this could be a better option than trying to grow a cash crop for this purpose, as returns

Table IV Possible Approach to Addressing Relevant Problems of Subsistence Farmers and Matching of LEIA Techniques and Practices to Problem Solutions Problem being addressed More food for the household

Solution Increased yields Extra food source

More cash generated for the household

Increase sales of surplus produce

Enhanced quality of life for members of the household

Reduce labour

More varied diet Ease of cultivation Fuel source Aesthetic value

Technique Improved varieties, intercropping, animal manure, composting Multipurpose trees, cover crops, intercropping Enriched fallow, cover crops, animal manure, multipurpose trees, composting, crop residues Improved fallows

Crop diversification, cover crops, multipurpose trees Animal manure, cover crops Animal manure, multipurpose trees, crop residues Tithonia hedgerows

LOW EXTERNAL INPUT TECHNOLOGIES

533

to labour may be greater. Thus, not only is there a much wider range of possible solutions that can be evaluated, but also any improved solutions developed are much more likely to be adopted by farmers. Gender issues are also important. Women make up a sizeable proportion of tropical farmers, and it is they who most often focus on subsistence crops, generally using lower inputs of organic and inorganic fertilisers than men (Gladwin et al., 1997). In a study on constraints faced by women using organic agriculture, Gladwin et al. (1997) found that lack of capital prevented them from investing in either organic or inorganic fertilisers, lack of land limited their use of low-input organic techniques, and lack of labour limited their ability to undertake the activities that were required to implement such techniques, particularly as most women were also solely responsible for household duties and child care. In Senegal, women have even less rights to long-term use of land than do men (Golan, 1994), and therefore have no incentive to make long-term investments to improve soil productivity. Another gender-related problem that may act against the uptake of any improved technique is that additional incomes arising from sales of produce may go directly to the men in households, who are less likely than women to invest in children and the household as a whole (Pretty and Koohafkan, 2002). Cultural traditions may also restrict the uptake and use of particular techniques, despite any other benefits they may have. For example, in relation to the potential use of animal manure in Ghana, Kiff et al., (1997) found from surveys in a number of villages that farmers knew that manure could be used, but generally found its use unattractive due to its supply being unreliable, too much effort involved in its collection, and because they felt that manuring techniques were regressive and old-fashioned. In certain areas, there is also the cultural problem of persuading farmers who have no tradition of cattle husbandry to develop the knowledge and interest to keep them on their farms (Dickson and Benneh, 1995). In Nepal on the other hand, there is a long tradition of keeping livestock, and the integration of animal manure into farm nutrient management is well developed. Animals are multifunctional in Nepalese agricultural systems, and provide meat, milk, ghee, curd, and draught power, as well as manure. Many families may own more than a single species of animal, with the most common combination (60% of those owning animals) being cattle, goats, chickens and buffaloes (Pilbeam et al., 1999b). Often there are competing demands for manure produced by livestock, particularly for use as fuel or in construction, which may make its availability as a nutrient source scarce. The use of the livelihoods-oriented approach described above may help to identify the real limitations of agricultural production systems more clearly. For example, the destruction of primary forest at the forest margins in Bolivia and Brazil is driven by other more powerful factors than soil fertility or weed encroachment issues. There, the underlying causes are economic or political in nature — aided by government subsidies, wealthy landowners buy out the

534

A. GRAVES, R. MATTHEWS AND K. WALDIE

frontier colonists to obtain land for cattle ranching, so that the latter then move on and clear new land. Converting primary forest to pasture is therefore an effective “cash generating” option for resource poor farmers that fits neatly within wider economic and political realities. Consequently, the introduction of LEIA soil fertility and weed management techniques in an attempt to stabilise agricultural production at the forest margin, whilst technically feasible, is rendered almost irrelevant by these broader political and economic factors (Muchagata, 1997).

C. MODELLING To integrate all the various enterprises on a farm and to understand the often complex differences between farms and farmers, tools are needed. The SL framework just described is a useful start in this direction, but is limited in that it does not capture the dynamic and spatial nature of farming systems, particularly in relation to nutrient flows (Scoones and Toulmin, 1998). Simulation modelling is a tool that offers the potential to integrate knowledge from a range of different disciplines and to explore the complex relationships between them in a dynamic and spatial way. For this reason, we believe that it is important that effort is made in developing simulation models of subsistence agricultural systems so that the processes involved are made explicit and to identify gaps in our knowledge. Because of the long-term nature of many of the underlying processes, such modelling offers a cost-effective and relatively quick way of obtaining answers to questions regarding potential interventions. Such models could also help to explore some of the wider global issues such as climate change, deforestation, and desertification from the livelihoods perspective discussed above. The type of modelling we propose to be the most appropriate for this task is an integration of the key biophysical and socio-economic processes at the level of a household. A large number of household models incorporating these aspects already exist (e.g., de Jager et al., 1998; Pagiola and Holden, 2001), but most of these are static models providing only a snapshot of the state of a household at an instant in time and do not capture the dynamic characteristics of household activities (Scoones and Toulmin, 1998). In a recent attempt to make progress in this area, Shepherd and Soule (1998) developed a farm simulation model to assess the long-term impact of existing soil management strategies on the productivity, profitability, and sustainability of farms in west Kenya. The model ran in time-steps of one year and linked soil management practices, nutrient availability, crop and livestock productivity, and farm economics. Crop types included weeds, fodder, grass, shrubs, and two types of grain crops. Growth of the plants was determined by N and P availability. A wide range of soil management options were simulated, including crop residue and manure management, soil erosion control measures, biomass transfer, improved fallow, green manuring, crop rotations, and N and P fertiliser application.

LOW EXTERNAL INPUT TECHNOLOGIES

535

The model was used to examine the sustainability of existing farming systems in the Vihiga district in western Kenya. Three household types were simulated to represent the range of resource endowments of households in the study area. Data to initialise the model for the different household types were collected through participatory research in the area, which indicated large differences in farm size, quantity and quality of livestock, soil and plant management, food consumption patterns, and sources of income. It was shown that the low and medium resource endowment farms had declining SOM, negative C, N, and P budgets, and low productivity and profitability. The high resource endowment farms, on the other hand, had increasing SOM, low soil nutrient losses, and were productive and profitable. This disaggregation into household types according to resource endowments highlighted the dangers of relying on nutrient balances of an “average” farm type — most previous studies in Africa have generally shown negative nutrient balances using this approach. Nevertheless, in this particular case, the low and medium resource households represented around 90% of the total in the study area, suggesting that overall a negative nutrient budget was likely. There was also the question of where the increasing nutrients of the high resource households were coming from — their greater purchasing power may have meant that there was nutrient flow from the poorer households to the richer ones. Shepherd and Soule (1998) concluded that the ability of the high resource households to manage their farms profitably and sustainably indicates that it is possible, but that capital is required. Strategies they suggested to improve livelihoods included: (1) an increase in the value of farm input, (2) an increase in high quality nutrient inputs at low cash and labour costs to the farmer, and (3) an increase in off-farm income. In another example, as part of a project evaluating integrated agriculture – aquaculture farming systems in the Philippines, Schaber (1997) developed a whole-farm model called FARMSIM to quantify flows of nutrients between the different farming enterprises. The ORYZA_0 rice simulation model was combined with a fish-pond model (for Nile Tilapia Oreochromis niloticus (L.)) and models of pigs, chickens and buffaloes. After the rice was harvested, the straw was assumed to be composted, the bran to be fed to the pigs and to the fish, and broken rice to be used as chicken food. If this supply of food was less than the demand, then more had to be purchased externally. Weeds from the field bunds were fed to the buffaloes. Manure from the pigs and buffaloes was fed to the fish, although buffalo manure could also be applied to the rice fields. The model was then used to evaluate three different scenarios in terms of the efficiency of N use (defined as the output N as a ratio of the input N) of the farm as a whole. In the first scenario, a conventional farming system with a mono-cultural rice field, two buffaloes and fifteen chickens was assumed. High levels (200 kg N ha21) of commercial fertiliser were applied to the rice field, and the output of rice grain was high. In the second scenario, diversification increased as pigs were introduced. The third scenario represented a fully integrated, diversified farming

536

A. GRAVES, R. MATTHEWS AND K. WALDIE

system, including a fish-pond, with all reusable products from each farm enterprise being used as inputs to other enterprises where appropriate. The predicted efficiencies of N use in each of the three scenarios were 13.3, 18.7, and 21.6% respectively. It was concluded that the greater the number of enterprises there were on the farm, the greater the efficiency of N use of the farm as a whole. A similar household model is currently being developed as a tool to help evaluate the relevance of potential soil fertility-enhancing techniques to livelihoods of farmers in the mid-hills of Nepal (Matthews, 2000). In addition to the biophysical processes of crop and animal growth, and water and nitrogen fluxes through the household, economic and labour flows have also been incorporated, along with household resources such as food, money, manure, fodder, and fertiliser. The model, therefore, incorporates elements of the natural, human, and financial capitals in the SL framework described above. Various types of household can be accommodated, ranging from resource poor to resource rich. The model will be used in the first instance to evaluate potential interventions in the existing system and the likelihood of uptake of these interventions, using criteria such as their contribution to household finances, food production, alleviation of risk, and labour demands in relation to other farm enterprises. So far, these models are not spatially explicit, but they do have the capability to become so. This is important because, as we have already discussed, the spatial relationship between different parts of the landscape is often central to the functioning of a farming system, with some areas acting as nutrient sources and others as nutrient sinks. The net movement of nutrients from fields far from the household to those close by on smallholder farms in Uganda (Wortmann and Kaizzi, 1998; Briggs and Twomlow, 2002) has already been mentioned, while in Burkino Faso, Prudencio (1993) found that 85% of household manure was applied to nearby plots and only 15% to distant plots. Similarly, farmers often consciously manipulate erosion and run-off from dry toplands as a means of concentrating moisture and nutrients on the wetter and more accessible bottomlands where they can benefit from them (Scoones and Toulmin, 1998). Thus, extrapolation of results from measurements at a single site can be highly misleading. The household models described above could take this into account by incorporating a number of associated land management units (e.g., fields) that are linked spatially, both in the horizontal and vertical directions if necessary. Heterogeneity at the village level could then be described by linking a number of such individual household models together (Matthews and Stephens, 2002). Such models would then provide insights into the movement of nutrients within the landscape, and possibly suggest ways in which this could be optimised to benefit people’s livelihoods. At some point in the future, these models should also incorporate household decision-making processes, based on a labour and economic analysis each year of the various household enterprises (crops, livestock, off-farm work, etc.), also

LOW EXTERNAL INPUT TECHNOLOGIES

537

taking into account subsistence needs and attitude to risk. However, considerable thought needs to be given to the dynamic processes involved in household decision-making, and how these are influenced by the biophysical environment. Some progress has been made by Pagiola and Holden (2001) and Angelsen and Kaimowitz (2001) in determining when forest clearing is likely to be a rational decision for farmers. The emerging field of multi-agent systems simulation, in which artificial intelligence approaches are incorporated and interactions between individuals are central (e.g., Bousquet et al., 1999), may be one way in which further progress can be made in this direction. This is one area where multidisciplinary research involving biophysical scientists and social scientists is likely to be fruitful. It is important to emphasise that while such models cannot be used to predict the behaviour of specific households precisely, they are useful as tools for understanding and testing hypotheses regarding the processes involved in interactions between the biophysical and socio-economic environments of subsistence farmers, and how these processes relate to their livelihoods and poverty. Exploration of viable pathways out of poverty is more important than the prediction of final endpoints. In the context of LEIA technologies, some of the types of questions that can be addressed with such models are as follows: 1. The potential of LEIA techniques: While LEIA techniques such as the ones we have discussed in this review can make a useful contribution to maintenance of soil fertility, they are unable to supply enough nutrients required to achieve the genetic potential of high-yielding crops. However, it would be useful to know what level of crop yields could be sustained by the sole use of such techniques in different environments and contexts, and how farmer livelihoods are affected by this. 2. Evaluation of fallow types: Natural fallows offer a way of regenerating soil fertility, but land must be set aside for long periods of time. Where land is relatively plentiful, natural fallowing is a rational strategy. However, where population increases and the availability of land decreases, the opportunity cost of setting aside land for long periods of time rises significantly. Improved fallows may be able to speed up the regeneration process, but at what level of land availability does it become worthwhile for a farmer to consider the technique? Similar questions can be asked for enriched fallows, where the regenerative process is accompanied by income generation, taking into account the possibly slower regeneration rate due to removal of harvested material. Experimental determination of these issues is time-consuming and expensive. 3. Managing variation in natural resources: By concentrating resources in one area at the expense of another, higher value crops may be grown, leading to an improvement in cash income for the household, some of which could be re-invested in the poorer areas of the farm, thereby improving the overall

538

A. GRAVES, R. MATTHEWS AND K. WALDIE

fertility of the farm in the long term. Many of the LEIA systems we have discussed (e.g., the Machobane system, SRI, the Zai system) essentially concentrate nutrients from a wider area into a smaller area that can be cropped. What is the optimum way to do this for a given biophysical and socioeconomic situation? 4. Effect of a change in farmer perceptions: Recent work in Ghana has shown that in evaluating different practices, farmers do not always value the opportunity cost of their own labour, though they more readily assess the financial cost of hiring others to work on their farms (Galpin et al., 2000). Participatory interaction, however, has brought some of them to consider that their own time and labour should be factors taken account of in the evaluation. It would be interesting to compare the likelihood of adoption of various techniques (both traditional and researcher-generated) with and without consideration of the labour involved. Would patterns of development be different in each case? Do more sustainable practices result from taking labour into account? Or is the concept of opportunity cost of labour meaningless when there are so few options available in which it could be deployed, anyway? 5. Effect of current socio-economic trends: In Nepal in recent years, a decline in soil fertility has been ascribed by farmers themselves to a decline in manure applications, due in turn to a decline in livestock numbers brought about by a reduction in the household labour pool with more and more children going to school (Ellis-Jones, pers. comm.). School leavers are not interested in returning to work on the farm, preferring to find jobs in the towns and cities. What effect is this likely to have on the fertility of the soil in the first instance, and on the overall livelihood of the household, bearing in mind that urban jobs represent a potential source of cash income for the household in the future? Is it a good livelihood strategy to invest in the education of one’s children, and at what cost is this to the biophysical environment? Should government policies aim to encourage the educated to take up farming, or is it desirable that hill agriculture continues to decline? 6. Trajectories out of poverty: The question of whether there are “natural” processes (in the broadest sense, including both biophysical and socioeconomic processes) that can lead to the evolution of one agricultural system into another needs to be explored. Given that it is perfectly rational for poor people to adopt short-term strategies that attempt to maximise their livelihood outcomes, can improved (or even new) strategies be developed or promoted to hasten the change from shifting cultivation systems to more settled patterns of agriculture? Can LEIA techniques improve livelihoods even if they are used efficiently, or are external inputs essential? Could governments adopt particular policies that would facilitate the transition process? How is the distribution of wealth influenced by different processes of transition? What are the long-term environmental consequences of such transitions?

LOW EXTERNAL INPUT TECHNOLOGIES

539

V. CONCLUDING REMARKS For any agricultural system to be sustainable, regardless of whether it is classified as LEIA or HEIA, inputs must not be less than outputs — if they are, then mining of existing stocks occurs, and the system must deteriorate in time. Where there is constant removal of nutrients in crops harvested for human consumption and these are not returned to the system, then no such system can be sustainable without some further input from outside. No group of technologies, whether based on organic or inorganic sources of nutrients, can contravene this basic law of mass conservation. In this review, we have attempted to analyse the biophysical and socioeconomic characteristics of a number of LEIA techniques that have been evaluated as potential improvements to subsistence agricultural systems. These techniques included intercropping, alley cropping, cover crops and green manures, biomass transfer, compost, use of animal manure, and improved and enriched fallows. It is important to remember that farmers have been practising variations of these techniques for generations, using organic materials that have been available within their immediate vicinity. While population density has been low, and land has been abundant, these systems have been sustainable, but with the population explosion in the 20th century, the parameters have now changed, and there is no guarantee that these systems will continue to function in a sustainable way. Indeed, many are beginning to break down as cropping periods are extended and fallow periods are shortened (e.g., Chidumayo, 1987). There is no doubt that there is an urgent need to find ways to make these systems more productive and sustainable, since it is upon the outputs of such systems that the lives and livelihoods of so many of the world’s poor ultimately depend. While there may be some scope in making existing systems more efficient by reducing losses of nutrients, identifying new sources of organic material, or spatially manipulating nutrient concentrations within farms (Wortmann and Kaizzi, 1998; Briggs and Twomlow, 2002), most of the LEIA techniques we have reviewed, when used alone, appear to have limited potential to increase food production dramatically. In the case of those techniques aimed at maintaining or improving soil fertility, the nutrient content and the quantity of biomass that can be produced within the resources available to such farmers is insufficient to meet the requirements of most crops, certainly at a reasonable yield level, although where existing yields are very low, substantial relative yield increases may be possible (e.g., Pretty et al., 2003). The contribution that these large relative, but small absolute, increases in yields make to overall global food production needs to be determined by weighting each yield level with an appropriate land area. Similarly, those LEIA techniques aimed at weed control in the studies reviewed, while often being able to reduce weed populations, had little effect on crop yields. In both cases, there is also the possibility that crop yields may actually be depressed,

540

A. GRAVES, R. MATTHEWS AND K. WALDIE

either as a result of competition for resources, particularly in difficult conditions, or as a result of N or P immobilisation if low-quality organic material is incorporated into the soil. While LEIA technologies are often promoted as “natural” (e.g., Pretty, 1999), as we have seen, in reality many of them have very high demands on land and labour. Thus, the debate over the relative merits of LEIA and HEIA technologies essentially hinges on whether the inputs should be capital intensive or land and/or labour intensive. Ironically, all of these are assets that by definition, resource-poor farmers have in short supply. Any hopes of promotion of their use to bring about radical changes to existing cultivation systems to address the problem of future global food demand, and to provide impetus to lift such farmers out of the poverty spiral, is, therefore, likely to be more utopian than realistic. A similar conclusion was also reached by Sanchez et al. (2001), who argued that sole use of LEIA technologies is only likely to perpetuate food insecurity and poverty. The future, therefore, must surely lie in the integration of the two approaches, using LEIA technologies when organic sources of nutrients are available, but also being prepared to supplement these with external supplies if necessary and when it is economic to do so. Taking a systems perspective is essential — there is no doubt that higher inputs of either inorganic or organic forms of nutrients can increase crop yields of individual plots or fields — the real issue, however, is whether the supplies of these inputs are sustainable at higher scales, and whether there is sufficient labour and capital in the system for their transport and handling. There is little point in increasing crop yields per se if these cannot be feasibly scaled up or maintained at that level for long. This applies to both LEIA and HEIA technologies — while promotion of the use of external supplies of fertiliser has been rightly criticised for being out of reach for many subsistence farmers, so might the large amounts of land required to provide sufficient nutrients in biomass, or the large amounts of labour at the right time to harvest, transport and incorporate this biomass. There is a tendency by researchers and proponents of both LEIA and HEIA approaches to quote only yield increases and to ignore the overall constraints of the system required to produce these increases (e.g., Uphoff, 1999; Singh and Sharma, 2001), which can be misleading. We would make a plea, therefore, that as well as providing adequate biophysical information on soils and climate, all future reports of yield increases in LEIA techniques state the area of land on which the crop yield has been measured, accompanied by estimates of the amount of land from which the particular technique has gathered its nutrients from, and the amount of labour required to harvest and transport these nutrients to the cropped area. Similar analyses should be made for existing farmer practice, and for HEIA techniques in terms of capital costs of purchasing, transporting, and applying the external inputs. The economic costs and benefits of the technique should also be determined, preferably dynamically, so that it is possible to see whether resource demand matches resource availability throughout the year. In this way, a more objective appraisal of all techniques may be possible.

LOW EXTERNAL INPUT TECHNOLOGIES

541

Future research to improve subsistence agriculture needs, therefore, to focus more on developing interventions to meet farmer realities, such as increased food security, improved cash generation, reduced risk, and enhanced quality of life. This may necessitate consideration of a broader range of interventions than just natural resource management options. In this process, both “holistic” and “reductionist” approaches are necessary and valuable in an iterative process of investigation and analysis. The starting point should be from a holistic viewpoint, the analysis of problems in the system and development of solutions should be reductionist, and any successful solutions to the problems should be evaluated holistically again. Many of the processes, both biophysical and socio-economic and their interactions, are poorly understood, and it is essential that future research addresses this. This lack of knowledge is compounded by the large degree of heterogeneity of the production systems involved, both at the system level with different cultivation systems in the different countries, and also at the individual farm level with between-farm variability in terms of farmer aspirations and attitudes, and withinfarm variability in resources. However, it is important that this heterogeneity is preserved as it contributes to the resilience, and hence the sustainability, of the production system. This was highlighted in the example of the maize – Mucuna system in Honduras in which over-reliance on one cover crop species was probably a major factor allowing the incursion of Rottboellia as Mucuna as more effective against broad-leaf weeds than grasses (Neill and Lee, 2001). Of all of the LEIA and HEIA technologies on offer, there is no single one that is a panacea for the problems faced by subsistence farmers. Each has particular strengths and weaknesses, and the challenge is to identify these strengths and combine them into integrated systems capable of adapting to changing circumstances when necessary.

ACKNOWLEDGMENTS We are grateful to the Natural Resources Systems Programme of the United Kingdom’s Department for International Development (DFID) for funding this work. The review is condensed and updated from the Final Report of DFID Project R7560.

REFERENCES Adeorike, V., Ogburia, M. N., and Anegbeh, P. (2001). Evaluation of the allelopathic influence of selected multipurpose tree species on maize (Zea mays) under a simulated field condition. Tropicultura 19(4), 191 –193.

542

A. GRAVES, R. MATTHEWS AND K. WALDIE

Adeoye, G. O., Sridhar, M. K. C., and Oyenekan, O. T. (1996). A survey of compost utilization for ornamental plants in Nigeria. Compost Sci Utilization 4(4), 81–83. Adesina, A., and Coulibaly, N. (1998). Policy and competitiveness of agroforestry-based technologies for maize production in Cameroon: an application of policy analysis matrix. Agric. Econ. 19, 1–13. Akinyemi, J. O., and Kalejaiye, A. S. (1998). The strategies for implementing the soil conservation practices among the rural farmers in the tropical rain forest zone of Nigeria. Int. J. Tropic Agric. 14(1–4), 35–48. Akobundu, I. O., Udensi, U. E., and Chikoye, D. (2000). Velvetbean (Mucuna spp.) suppresses speargrass (Imperata cylindrica (L.) Raeuschel) and increases maize yield. Int. J. Pest Manag. 46(2), 103–108. Akyeampong, E. (1999). Suitability of six hedgerow species for alley cropping on an acid infertile soil in Burundi. Tropic. Agric. 76(2), 82 –87. Ali, M. (1996). Quantifying the socio-economic determinants of sustainable crop production: an application to wheat cultivation in the Tarai of Nepal. Agric. Econ. 14(1), 45–60. Ali, M. (1999). Evaluation of green manure technology in tropical lowland rice systems. Field Crops Res. 61, 61 –78. Amede, T., Belachew, T., and Geta, E. (2001). Reversing the degradation of arable land in the Ethiopian Highlands. Managing Africa’s Soils 23, 23. Angelsen, A., and Kaimowitz, D. (2001). When does technological change in agriculture promote deforestation? In “Tradeoffs or Synergies? Agricultural Intensification, Economic Development, and the Environment” (D. R. Lee and C. B. Barrett, Eds.), pp. 89 –114. CAB International, Wallingford, UK. Anoka, U. A., Akobundu, I. O., and Okonkwo, S. N. C. (1991). Effects of Gliricidia sepium and Leucaena leucocephala on growth and development of Imperata cylindrica. Agrofor. Syst. 16, 1–12. Apiradee, K. (1988). Factors affecting the adoption of farmers’ compost making and application in Amphoe Khao Yoi, Changwat Phetchaburi, Thailand. Kaesart University, Bangkok. Araki, S. (1993). The effect of burning, ash fertilisation, and soil organic matter on productivity of the chitemene shifting cultivation in northern Zambia. In “Soil Organic Matter Dynamics and Sustainability of Tropical Agriculture” (K. Mulongoy and R. Merckz, Eds.), pp. 367–375. John Wiley & Sons, New York. Ashley, C., and Carney, D. (1999). Sustainable Livelihoods: lessons from early experience, p. 55. Department for International Development, London, UK. Atta-Krah, A. N., and Sumberg, J. E. (1988). Studies with Gliricidia sepium for crop/livestock production systems in West Africa. Agrofor. Syst. 6, 97 –118. Ayuk, E. T. (2001). Social, economic and policy dimensions of soil organic matter management in sub-Sahara Africa: challenges and opportunities. Nutrient Cycling in Agroecosystems 61, 183–195. Azam-Ali, S. N., Matthews, R. B., Williams, J. H., and Peacock, J. M. (1990). Light use, water uptake and performance of individual components of a sorghum/groundnut intercrop. Exp. Agric. 26, 413–427. Baggio, A., and Heuveldorp, J. (1988). Initial performance of Calliandra calothyrus Messn. in live fences for the production of biomass. Agrofor. Syst. 2, 19–29. Barrios, E., Kwesiga, F., Buresh, R. J., and Sprent, J. I. (1997). Light fraction soil organic matter and available nitrogen following trees and maize. Soil Sci. Soc. Am. J. 61, 826–831. van den Berghe, C., Sota, P., and Mujawayezu, A. (1994). Study on mineral complementation in the composting process in farmers conditions of Mugamba, Burundi. Tropicultura 12(2), 55– 60. Blarel, B. (1994). Tenure and agricultural production under land scarcity: the case of Rwanda. In “Searching for Land Tenure Security in Africa” (J. W. Bruce and S. N. Migot-Adholla, Eds.), pp. 71– 96. Kendall/Hunt Publishing Company, Dubuque, Iowa, USA.

LOW EXTERNAL INPUT TECHNOLOGIES

543

Boddey, R. M., de Sa, J. C. M., Alves, B. J. R., and Urquiaga, S. (1997). The contribution of biological nitrogen fixation for sustainable agricultural systems in the tropics. Soil Biol. Biochem. 29(5/6), 287 –799. Bouldin, D. R., Mughogho, S., Lathwell, D. J., and Scott, T. W. (1979). Nitrogen Fixation by Legumes in the Tropics. (International Agricultural Mimeograph). Report No. 75. Department of Agronomy, Cornell University, New York. Bousquet, F., Barreteau, O., Le Page, C., and Weber, J. (1999). An environmental modelling approach: the use of multi-agent simulations. In “Advances in Environmental and Ecological Modelling” (F. Blasco and A. Weill, Eds.), pp. 113–122. Elsevier, Paris, France. Brady, N. (1990). In “The Nature and Properties of Soils”. Macmillan Publishing Company, New York. Briggs, L., and Twomlow, S. J. (2002). Organic material flows within a smallholder highland farming system of south west Uganda. Agric. Ecosyst. Environ. 89, 191–212. Brodd, P., and Osanius, M. (2002). Erosion and land use at plot and catchment levels — a case study in northern Vietnam, p. 62. Minor Field Studies International Office, Swedish University of Agricultural Sciences, Uppsala, Sweden. Buckles, D., and Triomphe, B. (1999). Adoption of mucuna in the farming systems of northern Honduras. Agrofor. Syst. 47, 67 –91. Budelman, A. (1988). Leaf dry matter productivity of three selected perennial leguminous species in humid tropical Ivory Coast. Agrofor. Syst. 7, 42–62. Buresh, R. J., and Niang, A. I. (1997). Tithonia diversifolia as a green manure: awareness, expectations and realities. Agrofor. Forum 8(3), 29–31. Cairns, M., and Garrity, D. P. (1999). Improving shifting cultivation in Southeast Asia by building on indigenous fallow management strategies. Agrofor. Syst. 47, 37–48. Carney, J. (1993). Low input sustainable agriculture in Cuba. Agric. Human Values 10(3) (Special Issue). Carney, D. (1998). Implementing the sustainable rural livelihoods approach, Sustainable Rural Livelihoods — what contribution can we make?, pp. 3–23. Department for International Development, London, UK. Carter, J. (1995). Alley farming: have resource-poor farmers benefited? Natural Resources Perspectives No. 3. Overseas Development Institute, London. Catchpoole, D. W., and Blair, G. J. (1990). Forage tree legumes. III. Release of nitrogen from leaf, faeces and urine derived from Leucaena and Gliricidia leaf. Aust. J. Agric. Res. 41, 539–547. Cenas, P. A. A., Pandey, S., and Pandey, S. (1996). Determinants of adoption and retention of contour hedgerows in Claveria, Misamis Oriental. Philippine J. Crop Sci. 21(1–2), 34 –39. Chianu, J. N., Akintola, J. O., and Kormawa, P. M. (2002). Profitability of cassava-maize production under different fallow systems and land-use intensities in the derived savanna of southwest Nigeria. Exp. Agric. 38(1), 51 –63. Chidumayo, E. N. (1987). A shifting cultivation land use system under population pressure in Zambia. Agrofor. Syst. 5, 15 –25. Chivasa, W., Harris, D., Chiduza, C., Nyamudeza, P., and Mashingaidze, A. B. (2000). Biodiversity on-farm in semi-arid agriculture: a case study from a smallholder farming system in Zimbabwe. Zimbabwe Sci. News 34(1), 13–18. Coppock, D. L., and Reed, J. D. (1992). Cultivated and native browse legumes as calf supplements in Ethiopia. J. Range Manag. 45, 231 –238. Coultas, C. L., Post, T. J., Jones, J. B. Jr, and Hsieh, Y. P. (1996). Use of velvet bean to improve soil fertility and weed control in corn production in northern Belize. Commun. Soil Sci. Plant Anal. 27(9– 10), 2171–2196. Craswell, E. T., Sajjapongse, A., Howlett, D. J. B., and Dowling, A. J. (1998). Agroforestry in the management of sloping lands in Asia and the Pacific. Agrofor. Syst. 38, 121–137. Crosson, P., and Anderson, J. R. (2002). Technologies for meeting future global demand for food. Discussion Paper 02-02. Resources for the Future, 1616 P Street, NW, Washington D.C. 20036.

544

A. GRAVES, R. MATTHEWS AND K. WALDIE

Dalland, A., Va˚je, P. I., Matthews, R. B., and Singh, B. R. (1993). The potential of alley cropping in improvement of cultivation systems in the high rainfall areas of Zambia. III. Soil chemical and physical properties. Agrofor. Syst. 21, 117 –132. Dasthagir, M. G., Kohli, R. K., Arya, K. S., and Atul (1996). Review of research on tropical agroforestry. In “IUFRO-DNAES International Meeting; Resource Inventory Techniques to Support Agroforestry and Environment”. pp. 306 –309. DAV College, Chandigarh, India. David, S. (1995). What do farmers think? Farmer evaluations of hedgerow intercropping under semiarid conditions. Agrofor. Syst. 32, 15–28. Degrande, A., and Duguma, B. (2000). Adoption potential of rotational hedgerow intercropping in the humid lowlands of Cameroon, p. 8. Agricultural Research and Extension Network, Network Paper No. 103. Overseas Development Institute, London, UK. Delve, R. J., Cadisch, G., Tanner, J. C., Thorpe, W., Thorne, P. J., and Giller, K. E. (2001). Implications of livestock feeding management on soil fertility in the smallholder farming systems of sub-Saharan Africa. Agric. Ecosyst. Environ. 84, 227–243. Deugd, M., Ro¨ling, N., and Smaling, E. M. A. (1998). A new praxeology for integrated nutrient management, facilitating innovation with and by farmers. Agric. Ecosyst. Environ. 71, 269– 283. Dewees, P. A., and Saxena, N. C. (1995). Tree planting and household land and labour allocation: case studies from Kenya and India. In “Tree Management in Farmer Strategies: Responses to Agricultural Intensification” (J. E. M. Arnold and P. A. Dewees, Eds.), pp. 198– 241. Oxford University Press, Oxford. Dickson, K. B., and Benneh, G. (1995). In “A New Geography of Ghana”. Longman, Harlow. Diop, A. M. (1999). Sustainable agriculture: new paradigms and old practices? Increased production with management of organic inputs in Senegal. Environ. Develop. Sustainability 1, 285–296. Douthwaite, B., Manyong, V. M., Keatinge, J. D. H., and Chianu, J. (2002). The adoption of alley farming and Mucuna: lessons for research, development and extension. Agrofor. Syst. 56(3), 193–202. Dowling, N. G., Greenfield, S. M., and Fischer, K. S. (1998). Sustainability of Rice in the Global Food System, p. 404. Pacific Basin Study Centre, Davio, California, and International Rice Research Institute, Manila. Drechsel, P., and Reck, B. (1998). Composted shrub-prunings and other organic manures for smallholder farming systems in southern Rwanda. Agrofor. Syst. 39, 1–12. Drechsel, P., Steiner, K. G., and Hagedorn, F. (1996). A review on the potential of improved fallows and green manure in Rwanda. Agrofor Syst. 33(2), 109 –136. Durban Metropolitan, (1999). Durban Metro State of the Environment and Development: 1999 (Accessed 12-Dec-2002 online at http://www.ceroi.net/reports/durban/issues/waste/state.htm). Elbasha, E., Thornton, P. K., and Tarawali, G. (1999). An ex post economic impact assessment of planted forages in West Africa. ILRI Impact Assessment Series, No. 2. International Livestock Research Institute (ILRI), Nairobi. Ellis, F. (1988). In “Peasant Economics: Farm Households and Agrarian Change”. Cambridge University Press, Cambridge. Enyong, L. A., Debrah, S. K., and Bationo, A. (1999). Farmers perceptions and attitudes towards introduced soil fertility enhancing technologies in western Africa. Nutrient Cycling in Agroecosystems 53, 177 –187. Euroconsult, (1989). “Agricultural Compendium for Rural Development in the Tropics and Subtropics”. The Netherlands Ministry of Agriculture and Fisheries, Amsterdam. FAO, (1996). “Food Requirements and Population Growth”. Food and Agriculture Organisation, Rome, Italy. Farrell, R. E., Sandercock, P. J., Pennock, D. J., and van Kessel, C. (1996). Landscape-scale variations in leached nitrate: relationship to denitrification and natural N15 abundance. Soil Sci. Soc. Am. J. 60, 1410–1415.

LOW EXTERNAL INPUT TECHNOLOGIES

545

Fassler, O. M., and Lascano, C. E. (1995). The effect of mixtures of sun-dried tropical shrub legumes on intake and nitrogen balance by sheep. Trop. Grasslands 29, 92–96. Field, S. P., Momuat, E. O., Ayre-Smith, R., and Bamualim, A. (1992). Developing land rehabilitation technologies compatible with farmer needs and resources. J. Asian Farm. Syst. Assoc. 1(3), 385 –394. Franzel, S. (1999). Socioeconomic factors affecting the adoption potential of improved tree fallows in Africa. Agrofor. Syst. 47, 305–321. Fujita, K., Matsumoto, K., Ofosu-Budu, G. K., and Ogata, S. (1993). Effect of shading on growth and dinitrogen fixation of kudzu and tropical pasture legumes. Soil Sci. Plant Nutr. 39(1), 43–54. Gachengo, C., Palm, C., Adams, E., Giller, K. E., Delve, R. J., and Cadisch, G. (1998). Organic Resources Database. Tropical Soil Biology and Fertility Programme (TSBF), Nairobi and Wye College, University of London. Gachengo, C. N., Palm, C. A., Jama, B., and Othieno, C. (1999). Tithonia and Senna green manures and inorganic fertilizers as phosphorus sources for maize in western Kenya. Agrofor. Syst. 44(1), 21 –36. Galiba, M., Vissoh, P., Dagbenonbakin, G., and Fagbohoun, F. (1998). The reactions and fears of farmers utilizing velvet bean (Mucuna pruriens). Cover crops in West Africa: contributing to sustainable agriculture. In “Proceedings of a conference 1–3 October 1996, Cotonou, Benin” (D. Buckles, A. Eteka, O. Osiname, and G. Galiano, Eds.), pp. 55–64. International Development Research Centre (IDRC), Ottawa, Canada. Gallagher, R. S., Fernandes, E. C. M., and McCallie, E. L. (1999). Weed management through shortterm improved fallows in tropical agroecosystems. Agrofor. Syst. 47, 197–221. Galpin, M., Dorward, P., and Shepherd, D. (2000). Participatory farm management methods for agricultural research and extension: a training manual. Agricultural Extension and Rural Development Department, Reading University, Reading, UK. Ganry, F., Feller, C., Harmand, J. M., Guibert, H., Martius, C., Tiessen, H., and Vlek, P. L. G. (2001). Management of soil organic matter in semiarid Africa for annual cropping systems. Nutrient Cycling in Agroecosystems 61(1– 2), 105–118. Garcia, J. N. M., Gerrits, R. V., and Cramb, R. A. (2002). Adoption and maintenance of contour bunds and hedgerows in a dynamic environment: experience in the Philippine uplands. Mountain Res. Develop. 22(1), 10 –13. Garrity, D. P., and Lai, C. K. (2000). Innovation in shifting cultivation in Asia: indigenous fallow management. Ileia Newslett. 16(3), 5–6. Gashaw, G., Varvikko, T., and Khalili, H. (1991). Smallholder Dairy Production in Northern Shoa Province, Central Highlands of Ethiopia. Assessment of Feed Resources in Relation to the Requirements on a Smallholder Farm (Final Report), p. 25. International Livestock Centre for Africa (ILCA), Addis Ababa. George, T. S., Gregory, P. J., Robinson, J. S., Buresh, R. J., and Jama, B. A. (2001). Tithonia diversifolia: variations in leaf nutrient concentration and implications for biomass transfer. Agrofor. Syst. 52(3), 199–205. Gill, H. S. (1982). The role of hand and mechanical weeding in weed management in the advancing countries. Improving Weed Management (Plant Production and Protection Paper 44). FAO, Rome, Italy. Giller, K. E. (2001). “Nitrogen Fixation in Tropical Cropping Systems”, p. 313. CAB International, Wallingford, UK. Giller, K. E., and Cadisch, G. (1995). Future benefits from biological nitrogen fixation: an ecological approach to agriculture. Plant Soil 174, 255 –277. Giller, E. K., Cadisch, G., Ehaliotis, C., Adams, E., Sakala, W. D., and Mafongoya, P. L. (1997). Building soil nitrogen capital in Africa. In “Replenishing Soil Fertility in Africa”. SSSA Special Publication No. 51 (R. J. Buresh, P. A. Sanchez, and F. Calhoun, Eds.), pp. 151 –192. Soil Science Society of America, Madison, Wisconsin.

546

A. GRAVES, R. MATTHEWS AND K. WALDIE

Girish, G. K., and Chandrashekar, C. P. (2000). Farmers practice of nutrient management for sugarcane in GLBC area of Karnataka — a status paper. Cooperative Sugar 31(6), 477 –484. Gladwin, H., Buhr, K. L., Goldman, A., Hiebsch, C., Hildebrand, P. E., Kidder, G., Langham, M., Lee, D., Nkedi-Kizza, P., and Williams, D. (1997). Gender and Soil Fertility in Africa. In “SSSA Special Publication No. 51: Replenishing Soil Fertility in Africa” (R. J. Buresh, P. A. Sanchez, and F. Calhoun, Eds.), pp. 219–236. Soil Science Society of America, Wisconsin. Golan, E. H. (1994). Land tenure reform in the Peanut Basin of Senegal. In “Searching for Land Tenure Security in Africa” (J. W. Bruce and S. N. Migot-Adholla, Eds.), pp. 231– 249. Kendall/ Hunt Publishing Company, Dubuque, Iowa, USA. Goyal, S., Mishra, M. M., Hooda, I. S., and Singh, R. (1992). Organic matter– microbial biomass relationship in field experiments under tropical conditions: effects of inorganic fertilization and organic ammendments. Soil Biol. Biochem. 24, 1081–1084. Gregory, P. J., Ingram, J. S. I., Andersson, R., Betts, R. A., Brovkin, V., Chase, T. N., Grace, P. R., Gray, A. J., Hamilton, N., Hardy, T. B., Howden, S. M., Jenkins, A., Meybeck, M., Olsson, M., Ortiz-Monasterio, I., Palm, C. A., Payn, T. W., Rummukainen, M., Schulze, R. E., Thiem, M., Valentin, C., and Wilkinson, M. J. (2002). Environmental consequences of alternative practices for intensifying crop production. Agric. Ecosyst. Environ. 88, 279–290. Grist, P., Menz, K., and Nelson, R. (1999). Multipurpose trees as improved fallow: an economic assessment. Int. Tree Crops J. 10(1), 19–36. Harris, D., Joshi, A., Khan, P. A., Gothkar, P., and Sodhi, P. S. (1999). On-farm seed priming in semiarid agriculture: development and evaluation in maize, rice and chickpea in India using participatory methods. Exp. Agric. 35, 15–29. Harsono, A. (1996). Potency and opportunity of soybean farming system on upland in Sumbawa (NTB) and Sikka (NTT) districts. In “Proceedings of a Seminar on Legumes and Tuber Crops Research Results in 1995” (N. Marwoto, A. Saleh, Kasno, and Sunardi, Eds.). Malang, Indonesia. Hartemink, A. E., Buresh, R. J., Jama, B., and Janssen, B. H. (1996). Soil nitrate and water dynamics in Sesbania fallows, weed fallows and maize. Soil Sci. Soc. Am. J. 60, 568–574. Hauser, S. (1990). Water and nutrient dynamics under alley cropping versus monocropping in the humid–subhumid transition zone. Transactions of the 14th International Soil Science Society Congress, Kyoto, Japan, 204– 209. Hellin, J., Welchez, L. A., and Cherrett, I. (1999). The Quezungual system: an indigenous agroforestry system from western Honduras. Agrofor. Syst. 46(3), 229 –237. Hoang Fagerstro¨m, M. H., van Noordwijk, M., Thai, Phien, and Nguyen Cong, Vinh (2001). Innovations within upland rice-based systems in northern Vietnam with Tephrosia candida as fallow species, hedgerow, or mulch: net returns and farmers’ response. Agric. Ecosyst. Environ. 86(1), 23–37. Humphries, S. (1996). In Milk cows, migrants, and land markets: unravelling the complexities of forest to pasture conversion in northern Honduras. University of Guelph, Guelph, Ontario, Canada. ICRAF, (1997). Using the Wild Sunflower, Tithonia, in Kenya for Soil Fertility and Crop Yield Improvement. International Centre for Research in Agroforestry, Nairobi, Kenya. IIRR, (1998). The Machobane farming system in Lesotho, Sustainable Agricultural Extension Manual for Eastern and Southern Africa. pp. 171 –176. International Institute of Rural Reconstruction, Africa Regional Office, PO Box 66873, Nairobi, Kenya. Jackson, D., Kiff, E., and Chan, M.-K. (1999). Integrated Food Crop Systems Project, Ghana: Development and Promotion of Improved Techniques of Water and Soil Fertility Management for the Sustainable Production of Crops on Land in the Humid Forest Belt. (Progress Against Milestones Review). Report No. 4242. Natural Resources Institute, London. Jadhao, S. L., Giri, D. G., Kharat, B. S., Deshmukh, J. P., and Kubde, K. J. (1999). Effect of integrated nutrient management on productivity of sorghum-wheat cropping system. Ann. Plant Physiol. 13(2), 128–132.

LOW EXTERNAL INPUT TECHNOLOGIES

547

de Jager, A., Nandwa, S. M., and Okoth, P. F. (1998). Monitoring nutrient flows and economic performance in African farming systems (NUTMON) I. Concepts and methodologies. Agric. Ecosyst. Environ. 71, 37– 48. de Jager, A., Onduru, D., van Wijk, M. S., Vlaming, J., and Gachini, G. N. (2001). Assessing sustainability of low-external-input farm management systems with the nutrient monitoring approach: a case study in Kenya. Agric. Syst. 69, 99–118. Jama, B., Buresh, R. J., and Place, F. M. (1998). Sesbania tree fallows on phosphorus-deficient sites: maize yield and financial benefit. Agron. J. 90, 717–726. Jama, B., Palm, C. A., Buresh, R. J., Niang, A., Gachengo, C., Nziguheba, G., and Amadalo, B. (2000). Tithonia diversifolia as a green manure for soil fertility improvement in western Kenya: a review. Agrofor. Syst. 49, 201 –221. Janssen, B. H. (1993). Integrated nutrient management: the use of organic matter and mineral fertilizers. In “The Role of Plant Nutrients for Sustainable Food Crop Production in sub-Saharan Africa” (H. Van Reuler and W. Prins, Eds.), pp. 89 –106. VKP, Leidschendam. Jeffery, R., Jeffery, P., and Lyon, A. (1989). Cattle-dung and dung-work: an aspect of the value of women’s work in rural north India. Int. Asienforum 20(1–2), 21–42. Jiri, O., and Waddington, S. R. (1998). Leaf prunings from two species of Tithonia raise maize grain yields in Zimbabwe, but take a lot of labour! pp. 4–5. In “Newsletter of Soil Fertility Network”, Harare, Zimbabwe. Kang, B. (1993). Alley cropping: past achievements and future directions. Agrofor. Syst. 23, 141 –155. Kang, B. T. (1997). Alley cropping — soil productivity and nutrient recycling. For. Ecol. Manag. 91, 75 –82. Kang, B. T., Wilson, G. F., and Sipkens, L. (1981). Alley cropping maize (Zea mays L.) and Leucaena (Leucaena leucocephala LAM) in Southern Nigeria. Plant Soil 63, 165–197. Kang, B. T., Grimme, H., and Lawson, T. L. (1985). Alley cropping sequentially cropped maize and cowpea with leucaena on a sandy soil in southern Nigeria. Plant Soil 85, 267–276. Kanyama-Phiri, G., and Snapp, S. S. (1997). Developing improved organic matter technologies in partnership with smallholder farmers in southern Malawi. In “Africa Crop Science Conference, Pretoria, S. Africa”. Kaya, B., Hildebrand, P. E., and Nair, P. K. R. (2000). Modelling changes in farming systems with the adoption of improved fallows in southern Mali. Agric. Syst. 66, 51–68. Keatinge, J. D. H., Qi, A., Wheeler, T. R., Ellis, R. H., and Summerfield, R. J. (1998). Effects of temperature and photoperiod on phenology as a guide to the selection of annual legume cover and green manure crops for hillside farming systems. Field Crops Res. 57, 139–152. Keatinge, J. D. H., Qi, A., Wheeler, T. R., Musitwa, F., Franks, N. A. P., Kuribuza, D., Ellis, R. H., and Summerfield, R. J. (1999). Potential annual-sown legumes for low-input systems in the East African Highlands of southwestern Uganda. Mountain Res. Develop. 19(4), 345 –353. Kiff, E., Chan, M.-K., and Jackson, D. (1997). Integrated Food Crop Systems Project, Ghana: Development and Promotion of Improved Techniques Of Water and Soil Fertility Management for the Sustainable Production of Crops on Land in the Humid Forest Belt. (Inception Report). Report No. R6789 AO683. Natural Resources Institute, London. Kiff, E., O’Connell, N., Musa, H., and Nelson, D. (1999). Integrated Food Crop Systems Project, Ghana: Development and Promotion of Improved Techniques of Soil and Fertility Management for the Sustainable Production of Crops on Land in the Humid Forest Belt. Report No. R6789 AO683. Natural Resources Institute, London. Kirchhof, G., and Salako, F. K. (2000). Residual tillage and bush-fallow effects on soil properties and maize intercropped with legumes on a tropical Alfisol. Soil Use Manag. 16(3), 183–188. Kline, S. J. (1995). In “Conceptual Foundations for Multi-disciplinary Thinking”. Stanford University Press, Stanford, California.

548

A. GRAVES, R. MATTHEWS AND K. WALDIE

Koech, E. K., and Whitbread, R. (2000). Disease incidence and severity on beans in alleys between Leucaena hedgerows in Kenya. Agrofor. Syst. 49(1), 85–101. Kumara Rao, J. V. D. K., Dart, P. J., Matsumato, T., and Day, J. M. (1981). Nitrogen fixation by pigeonpea. In “International Workshop on Pigeonpeas” (Y. I. Nene, Ed.), pp. 190– 199. International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Patancheru, Andhra Pradesh, India. Kumwenda, J. D. T., Saka, A. R., Snapp, S. S., Ganunga, R. P., and Benson, T. (1997a.). Effects of organic legume residues and inorganic fertilizer nitrogen on maize yield in Malawi. In “Soil Fertility Research for Maize-Based Farming Systems in Malawi and Zimbabwe” (S. R. Waddington, H. K. Murwira, J. D. T. Kumwenda, and D. Hikwa, Eds.), CIMMYT, Mutare. Kumwenda, J. D. T., Waddington, S., Snapp, S. S., Jones, R. B., and Blackie, M. J. (1997b). Soil fertility management in the smallholder maize-based cropping systems of eastern and southern Africa. In “The Emerging Maize Revolution” (C. Eisher, Ed.), pp. 153– 172. Michigan State University, Michigan. Kwesiga, F. R., Franzel, S., Place, F., Phiri, D., and Simwanza, C. P. (1999). Sesbania sesban improved fallows in eastern Zambia: their inception, development, and farmer enthusiasm. Agrofor. Syst. 47, 49–66. Lal, R. (1989a). Agroforestry systems and soil surface management of a tropical alfisol. II. Water runoff, soil erosion and nutrient loss. Agrofor Syst. 8, 97–112. Lal, R. (1989b). Agroforestry systems and soil surface management of a tropical alfisol. III. Changes in soil chemical properties. Agrofor. Syst. 8, 113–132. Lehmann, J., da Silva, J. P., Trujillo, L., Blanke, M., and Pohlan, J. (2000). Legume cover crops and nutrient cycling in tropical fruit tree production. Acta Hortic. 531, 65–72. Lekasi, J. K., Tanner, J. C., Kimani, S. K., and Harris, P. J. C. (1998). Manure Management in the Kenya Highlands: Practice and Potential, p. 32. HDRA Publications, UK. Leng, R. A. (1990). Factors affecting utilisation of poor quality forages by runinants under tropical conditions. Nutr. Res. Rev. 3, 277 –303. Long, A. J., and Nair, P. K. R. (1999). Trees outside forests: agro-, community, and urban forestry. New Forests 17, 145–174. Low, A. (1988). Farm household economics and the design and impact of biological research in southern Africa. Agricultural Administration and Extension 29, 23–34. Mafongoya, P. L., Barak, P., and Reed, J. D. (1997a). Carbon, nitrogen and phosphorus mineralization of tree leaves and manure. Biol. Fert. Soils 30(4), 298 –305. Mafongoya, P. L., Dzowela, B. H., and Nair, P. K. (1997b). Effect of multi-purpose tree, age of cutting, and drying method on pruning quality. In “Driven by Nature; Plant Litter Quality and Decomposition” (G. Cadisch and K. E. Giller, Eds.), pp. 167–174. CAB International, Wallingford, UK. Mahamoudou, S., and Meritan, R. (1998). Agricultural intensification and agroforestry in the Comoros Islands. Cahiers Agric. 7(3), 207–212. Mann, C. (1997). Reseeding the Green Revolution. Science 227(August), 1038–1043. Manu, V. T., Halavatau, S., and Craswell, E. T. (1994). Agroforestry in the food production systems in the South Pacific. In Soil Organic Matter Management for Sustainable Agriculture. “Proceedings of a Conference held at Ubon, Thailand, 24 –26 August 1994” (R. D. B. Lefroy and G. J. Blair, Eds.), pp. 63–68. ACIAR Proceedings No. 56. Maroko, J., Buresh, R. J., and Smithson, P. C. (1998). Soil nitrogen availability as affected by fallow– maize systems on two soils in Kenya. Biol. Fert. Soils 26, 229 –234. Marschner, H. (1995). In “Mineral Nutrition of Higher Plants”. Academic Press Ltd, London. Matthews, R. B. (2000). The role of quantitative methods in integrating biophysical and socioeconomic aspects of soil fertility management. In “Proceedings of a workshop on ‘Biophysical and socio-economic tools for assessing soil fertility’ held at ARS-Lumle, Nepal, 17 –18 October 2000” (J. Ellis-Jones, Ed.). Silsoe Research Institute, UK, Silsoe.

LOW EXTERNAL INPUT TECHNOLOGIES

549

Matthews, R. B., Azam-Ali, S. N., Saffell, R. A., Peacock, J. M., and Williams, J. H. (1991). Plant growth and development in relation to the microclimate of a sorghum/groundnut intercrop. Agric. For. Meteorol. 53, 285–301. Matthews, R. B., Holden, S. T., Lungu, S., and Volk, J. (1992a). The potential of alley cropping in improvement of cultivation systems in the high rainfall areas of Zambia. I. Chitemene and fundikila. Agroforestry Systems 17, 219– 240. Matthews R. B., and Stephens, W. (Eds.), (2002). In “Crop–Soil Simulation Models: applications in Developing Countries.” p. 277. CAB International, Wallingford, UK. Matthews, R. B., Volk, J., Lungu, S., Holden, S. T., and Solberg, K. (1992b). The potential of alley cropping in improvement of cultivation systems in the high rainfall areas of Zambia. II. Maize production. Agrofor. Syst. 17, 241–261. Matthews, R. B., Wassmann, R., Knox, J. W., and Buendia, L. (2000). Using a crop/soil simulation model and GIS techniques to assess methane emissions from rice fields in Asia. IV. Upscaling of crop management scenarios to national levels. Nutrient Cycling in Agroecosystems 58, 201–217. McDonagh, J., and Hillyer, A. (2000). Soil fertility, legumes and livelihoods in northern Namibia. In DFID Plant Sciences Research Programme Annual Report. 1999/2000 (J. Witcombe and D. Harris, Eds.), pp. 39–47. Centre for Arid Zone Studies, University of Bangor, Bangor, UK. Mekonnen, K., Buresh, R. J., and Jama, B. (1997). Root and inorganic nitrogen distributions in sesbania fallow, natural fallow, and maize fields. Plant Soil 188, 319 –327. Mogaka, L. M. (1993). An economic evaluation of zero-grazing feeding system for high yielding cows on smallholder farms in Kenya. In “Sustainable Feed Production and Utilisation for Smallholder Livestock Enterprises in sub-Saharan Africa” (J. Ndikumana and W. d. Leeuw, Eds.), AFRNET, Nairobi. Moore, A. W. (1962). The influence of a legume on soil fertility under a grazed tropical pasture. Empire J. Exp. Agric. 30, 239– 248. Morris, D. R., Weaver, R. W., Smith, G. R., and Rouquette, F. M. (1990). Nitrogen transfer from arrowleaf clover to ryegrass in field plantings. Plant and Soil 128, 293 –297. Moser, C. M., and Barrett, C. B. (2002). The system of rice intensification in practice: explaining low farmer adoption and high disadoption in Madagascar. Paper presented at the Water-wise Rice Production Workshop, 8 –11 April 2002. International Rice Research Institute, Los Ban˜os, Philippines. Muchagata, M. G. (1997). Forests and people: the role of forest production in frontier farming systems in eastern Amazonia. (Development Studies Occasional Paper 36). School of Development Studies, University of East Anglia, Norwich, UK. 79 pp. Mugendi, D. N., Nair, P. K. R., Mugwe, J. N., MK, O. N., and Woomer, P. L. (1999). Alley cropping of maize with Calliandra and Leucaena in the subhumid highlands of Kenya. Part 1. Soil-fertility changes and maize yield. Agrofor. Syst. 46(1), 39–50. Muriwara, H. K., and Kirchmann, H. (1993). Nitrogen dynamics and maize growth in a Zimbabwean sandy soil under manure fertilization. Commun. Soil Sci. Plant Anal. 24, 2343–2359. Murwira, K. H., Swift, M. J., and Frost, P. G. H. (1995). Manure as a key resource in sustainable agriculture. In International Conference on Livestock and Sustainable Nutrient Cycling in Mixed Farming Systems of Sub-Saharan Africa (J. M. Powell, S. Fernandez-Rivera, T. O. Williams, and C. Renard, Eds.), pp. 131–148. ILCA, Addis Ababa. Musa, A. M., Harris, D., Johansen, C., and Kumar, J. (2001). Short duration chickpea to replace fallow after Aman rice: the role of on-farm seed priming in the high Barind Tract of Bangladesh. Exp. Agric. 37(4), 509–521. Muza, L. (1995). Green manuring through legume relay cropping in maize in communal areas of Zimbabwe. In “Second South Eastern Africa Soil Fertility Network Workshop”. CIMMYT, Kadoma, Zimbabwe. Nair, P. (1993). An Introduction to Agroforestry. Kluwer Academic Publishers, Dordrecht, Netherlands.

550

A. GRAVES, R. MATTHEWS AND K. WALDIE

Nandwa, S. M. (1995). Synchronization of nitrogen mineralization with N uptake through maize stover placements and N fertilization under continuous maize monocropping systems in Kenya, p. 344. University of Exeter, Exeter, UK, Ph.D. Thesis. Nandwa, S. M., and Bekunda, M. A. (1998). Research on nutrient flows and balances in East and Southern Africa: state-of-the-art. Agric. Ecosyst. Environ. 71, 5–18. Neill, S. P., and Lee, D. R. (2001). Explaining the adoption and disadoption of sustainable agriculture: the case of cover crops in northern Honduras. Econ. Develop. Cult. Change 49(4), 793–820. Nelson, R. A., and Cramb, R. A. (1998). Economic incentives for farmers in the Philippine uplands to adopt hedgerow intercropping. J. Environ. Manag. 54(2), 83 –100. Nelson, R. A., Grist, P. G., Menz, K. M., Paningbatan, E. P., and Mamicpic, M. A. (1996). Cost-benefit analysis of hedgerow intercropping in the Philippine uplands using SCUAF (Soil Changes Under Agroforestry). SEARCA/SEAMEO Regional Center for Graduate Study and Research in Agriculture, College, Laguna. Nelson, R. A., Cramb, R. A., and Mamicpic, M. A. (1998). Erosion/productivity modelling of maize farming in the Philippine uplands. Part III: Economic analysis of alternative farming methods. Agric. Syst. 58(2), 165– 183. Nguyen The, D., Klinnert, C., Martius, C., Tiessen, H., and Vlek, P. L. G. (2001). Problems with and local solutions for organic matter management in Vietnam. Nutrient Cycling in Agroecosystems 61(1–2), 89–97. Nibbering, J. W. (1991). Crisis and resilience in upland land use in Java. In “Indonesia: Resources, Ecology and Environment” (J. Hardjono, Ed.), pp. 104–132. Oxford University Press, New York. Nyathi, P., and Campbell, B. M. (1993). The acquisition and use of miombo litter by small-scale farmers in Masvingo, Zimbabwe. Agrofor. Syst. 22, 43–48. Oko, B. F. D., Shiyam, J. O., Eneji, A. E., Honna, T., and Yamamoto, S. (2000). Yield of selected food crops under alley-cropping with some hedgerow species in humid tropical southeastern Nigeria. Jpn. J. Tropic. Agric. 44(3), 167–170. Onduru, D. D., Gachini, G. N., de Jager, A., and Diop, J. M. (1999). Participatory research on compost and liquid manure in Kenya. “Managing Africa’s Soils”, Vol. 8, p. 23. Drylands Programme, International Institute for Environment and Development, London, UK. Ong, C. (1994). Alley cropping — ecological pie in the sky? Agrofor. Today 6, 8–10. Ong, C. K., Wilson, J., Deans, J. D., Mulayta, J., Raussen, T., Wajja Musukwe, N., Turner, N. C., and Ward, P. R. (2002). Tree– crop interactions: manipulation of water use and root function. Agric. Water Manag. 53(1–3), 171–186. Oue´draogo, E., Mando, A., and Zombre´, N. P. (2001). Use of compost to improve soil properties and crop productivity under low input agricultural system in West Africa. Agric. Ecosyst. Environ. 84, 259–266. Padwick, G. W. (1983). Fifty years of Experimental Agriculture in the maintenance of soil fertility in tropical Africa: A review. Exp. Agric. 19, 293–310. Pagiola, S., and Holden, S. (2001). Farm household intensification decisions and the environment. In “Tradeoffs or Synergies?” (D. R. Lee and C. B. Barrett, Eds.), pp. 73–88. Agricultural Intensification, Economic Development, and the Environment. CAB International, Wallingford, UK. Palm, C., Wangari, N., and Gichuru, M. (1996). Selection of organic inputs for soil fertility management. In “Soil Fertility Network for Maize-based Farming Systems in Selected Countries of southern Africa. Research Results and Network Outputs in 1994 and 1995” (S. R. Waddington, Ed.), pp. 19 –24. CIMMYT, Harare, Zimbabwe. Palm, C. A., Myers, R. J. K., and Nandwa, S. M. (1997a). Combined use of organic and inorganic nutrient sources for soil fertility maintenance and replenishment. In “Replenishing Soil Fertility in Africa” (R. J. Buresh, P. A. Sanchez, and F. Calhoun, Eds.), pp. 193–217. Soil Science Society of America, Madison, Wisconsin, USA. Palm, C. A., Myers, R. J. K., and Nandwa, S. M. (1997b). Organic–inorganic nutrient interactions in soil fertility replenishment. In “Replenishing Soil Fertility in Africa”

LOW EXTERNAL INPUT TECHNOLOGIES

551

(R. J. Buresh, P. A. Sanchez, and F. Calhoun, Eds.), pp. 193–218. Soil Science Society of America, Madison. Palm, C. A., Gachengo, C., Delve, R., Cadisch, G., and Giller, K. E. (2001). Organic inputs for soil fertility management in tropical agroecosystems: application of an organic resource database. Agric. Ecosyst. Environ. 83, 27–42. Pande, R. S. (1997). Fodder and Pasture Development in Nepal. Udatya Research and Development Services, Kathmandu. Paningbatan, E. P. Jr. (1990). Alley cropping for managing soil erosion in sloping lands, Transactions of the 14th International Soil Science Society Congress, Kyoto, Japan. pp. 376 –377. Pantanali, R. (1996). Lesotho: a note on the Machobane System. Report No. TCI Occasional Paper Series No. 7. Food and Agricultural Organisation, Rome, Italy. Patra, B., and Poi, S. C. (1998). Influence of intercropping on the nodulation in legumes. Environ. Ecol. 16(2), 418–423. Pieri, C. (1995). Long-term soil management experiments in Semi-arid Francophone Africa. In “Soil Management Experimental Basis for Sustainability and Environmental Quality. Advances in Soil Science” (R. Lal and B. A. Stewart, Eds.), pp. 225 –266. CRC, Lewis. Pilbeam, C. J., Tripathi, B. P., Munankarmy, R. C., Gregory, P. J., Ellis Jones, J., Mason, T., and Keatinge, D. (1999a). Productivity and economic benefits of integrated nutrient management in three major cropping systems in the mid-hills of Nepal. Mountain Res. Develop. 19(4), 333–344. Pilbeam, C. J., Tripathi, B. P., Sherchan, D. P., Gregory, P. J., and Gaunt, J. (1999b). Nitrogen balances for the households in the mid-hills of Nepal. Agric. Ecosyst. Environ. 79, 61–72. Place, F., and Dewees, P. (1999). Policies and incentives for the adoption of improved fallows. Agrofor. Syst. 47, 323 –343. Pound, B., Webb, M., Carreno, B., and Flores, A. (1999). Weed Management for Sustainable Agriculture in Forest Margins; Investigacion Adaptiva en Margenes de Bosque. (Joint Final Technical Report). Natural Resources Institute, London, and CIAT, Cali Bolivia. Powell, J. M., and Williams, T. O. (1993). Livestock, nutrient cycling and sustainable agriculture in West African Sahel. International Institute for Environment and Development, London, Gatekeeper Series No. 37. Pradeepmani, D. (1988). Introducing multipurpose trees on small farms in Nepal. In “Proceedings of an international workshop held November 2–5, (1987) in Pattaya, Thailand” (D. Withington, K. G. MacDicken, C. B. Sastry and N. R. Adams, Eds.), pp. 197–203. Winrock International Institute for Agricultural Development, Arlington, Virginia, USA, and International Development Research Centre, Canada. Prasad, P. V. V., Satyanarayana, V., Murthy, V. R. K., and Boote, K. J. (2002). Maximizing yields in rice–groundnut cropping sequence through integrated nutrient management. Field Crops Res. 75(1), 9– 21. Pretty, J. (1995). Regenerating Agriculture: Policies and Practice for Sustainability and Self-reliance, p. 336. Earthscan, London. Pretty, J. (1999). Can sustainable agriculture feed Africa? New evidence on progress, processes and impacts. Environ. Develop. Sustainability 1, 253–274. Pretty, J., and Koohafkan, P. (2002). Land and Agriculture: a compendium of recent sustainable development initiatives in the field of agriculture and land management, p. 76. Food and Agriculture Organization of the United Nations, Rome. Pretty, J. N., Morison, J. I. L., and Hine, R. E. (2003). Reducing food poverty by increasing agricultural sustainability in developing countries. Agric. Ecosyst. Environ. 95, 217– 234. Prudencio, C. F. (1993). Ring management of soils and crops in the West African semi-arid tropics: the case of the mossi farming system in Burkina Faso. Agric. Ecosyst. Environ. 47, 237 –264. Quansah, C., Drechsel, P., Yirenkyi, B. B., and Asante-Mensah, S. (2001). Farmers’ perceptions and management of soil organic matter — a case study from West Africa. Nutrient Cycling in Agroecosystems 61, 205–213.

552

A. GRAVES, R. MATTHEWS AND K. WALDIE

Rachmat, R., Stur, W. W., and Blair, G. J. (1992). Cattle feeding systems and limitations to feed supply in South Sulawesi, Indonesia. Agric. Syst. 39(4), 409 –419. Ramasamy, S., ten Berge, H. F. M., and Purushothaman, S. (1997). Yield formation in rice in response to drainage and nitrogen application. Field Crops Res. 51, 65–82. Rao, M. R., and Mathuva, M. N. (2000). Legumes for improving maize yields and income in semi-arid Kenya. Agric. Ecosyst. Environ. 78(2), 123–137. Rao, M. R., Nair, P. K. R., Ong, C. K., and Latt, C. R. (1997). Biophysical interactions in tropical agroforestry systems. Agrofor. Syst. 38(1–3), 3 –50. Reddy, D. D., Rao, A. S., and Rupa, T. R. (2000). Effects of continuous use of cattle manure and fertiliser phosphorus on crop yields and soil organic phosphorus on a vertisol. Bioresour. Technol. 75, 114– 118. Reed, J. D. (1986). Relationships among soluble phenolics, insoluble proanthocyanins and fibre in East African browse species. J. Range Manag. 39, 5–7. Reid, J. D. Jr. (1976). Sharecropping and agricultural uncertainty. Econ. Develop. Cult. Change 24(3), 549–576. Reynolds, L., Di Domenico, C., Atta-Krah, A. N., and Cobbina, J. (1991). Alley farming in southwestern Nigeria: the role of farming systems research in technology development. In “Planned Change in Farming Systems: Progress in On-farm Research” (R. Tripp, Ed.), pp. 85–108. John Wiley & Sons Ltd, London. Rocheleau, D., and Rocheleau, D. E. (1990). Gender complementarity and conflict in sustainable forestry development: a multiple user approach “19th IUFRO World Congress”, pp. 432– 448. Canadian IUFRO World Congress Organizing Committee, Montreal. Rodell, M. G. W., Hopley, J. D. H., and Boaltwood, J. N. (1980). Effects of applied nitrogen, kraal compact and maize stover on yields of maize growth on a poor granite soil. Zimbabwe Agric. J. 77, 229 –232. Roose, E., and Ndayizigiye, F. (1997). Agroforestry, water and soil fertility management to fight erosion in tropical mountains of Rwanda. Soil Technol. 11, 109 –119. Roose, E., Kabore, V., and Guenat, C. (1999). Zai practice: a West African traditional rehabilitation system for semiarid degraded lands, a case study in Burkina Faso. Arid Soil Res. Rehabil. 13(4), 343–355. Ruthenberg, H. (1980). Farming Systems in the Tropics, p. 424. Clarendon Press, Oxford. Saka, A. R., Phombeya, H. S. K., Itimu, O. A., Jones, R. B., Minae, S., and Snapp, S. S. (1995). Agroforestry Commodity Team Annual Report. Department of Agricultural Research, Malawi Ministry of Agriculture, Lilongwe. Salako, F. K., Hauser, S., Babalola, O., and Tian, G. (2001). Improvement of the physical fertility of a degraded Alfisol with planted and natural fallows under humid tropical conditions. Soil Use Manag. 17(1), 41– 47. Salam, M. A., Noguchi, T., and Koike, M. (2000). Understanding why farmers plant trees in homestead agroforestry in Bangladesh. Agrofor. Syst. 50, 77–93. Sanchez, P. A. (1999). Improved fallows come of age in the tropics. Agrofor. Syst. 47, 3–12. Sanchez, M. D., and Rosales Mendez, M. (1999). Agroforestry systems for intensifying sustainable livestock production in tropical Latin America. “FAO Anim Prod Health Paper”, 143, pp. 1 –13. Sanchez, P. A., Jama, B., Niang, A., and Palm, C. (2001). Soil fertility, small-farm intensification, and the environment in Africa. In “Tradeoffs or Synergies? Agricultural Intensification, Economic Development, and the Environment” (D. R. Lee and C. B. Barrett, Eds.), pp. 325–344. CAB International, Wallingford, UK. Sanginga, N., Vanlauwe, B., and Danso, S. K. A. (1995). Management of biological N2 fixation in alley cropping systems: estimation and contribution to N balance. Plant Soil 174(1–2), 119– 141. Schaber, J. (1997). “FARMSIM: a dynamic model for the simulation of yields, nutrient cycling and resource flows on Philippine small-scale farming systems”, MSc Thesis. p. 115. Universita¨t Osnabru¨ck, Osnabru¨ck.

LOW EXTERNAL INPUT TECHNOLOGIES

553

Scoones, I., and Toulmin, C. (1998). Soil nutrient balances: what use for policy? Agric. Ecosyst. Environ. 71, 255–267. Seif El Din, A. G., and Raintree, J. B. (1987). Gum hashab and land tenure in western Sudan, Land, trees and tenure. In “Proceedings of an international workshop on tenure issues in agroforestry, Nairobi, May 27 –31, 1985”, pp. 217–224. International Council for Research in Agroforestry (ICRAF), Nairobi, Kenya, and Land Tenure Center (LTC), Madison, Wisconsin, USA Selvarajan, K. N., and Krishnamoorthy, S. (1990). Cost benefit analysis of potato in the Nilgiris district, Tamil Nadu. Artha Vikas 26(1–2), 59– 66. Senesi, N. (1989). Composted materials as organic fertilisers. Sci. Total Environ. 81/82, 521 –542. Shepherd, K. D., and Soule, M. J. (1998). Soil fertility management in west Kenya: a dynamic simulation of productivity, profitability and sustainability at different resource endowment levels. Agric. Ecosyst. Environ. 71, 131–145. Sherchan, D. P., Pilbeam, C. J., and Gregory, P. J. (1999). Response of wheat-rice and maize/millet systems to fertilizer and manure applications in the mid-hills of Nepal. Exp. Agric. 35(1), 1–13. Shumba, E. M., Dhilwayo, H. H., and Mukoko, O. Z. (1990). The potential of maize–cowpea intercropping in low rainfall areas of Zimbabwe. Zimbabwe J. Agric. Res. 28, 33–38. Siaw, D. E. K. A., Kang, B. T., and Okali, D. U. U. (1991). Alley cropping with Leucaena leucocephala (Lam.) de Wit and Acioa barteri (Hook f.). Engl. Agrofor. Syst. 14, 219–231. Singh, R. P., and Desai, G. M. (1991). Farmers’ fertilizer practices in the semi-arid tropics of India. (Progress Report). Report No. 100, p. 54. Economics Group, Resource Management Program, ICRISAT, Patancheru, Andhra Pradesh, India. Singh, R. P., and Jodha, N. S. (1990). Determinants of intercropping in the semi-arid tropics. (Progress Report). Report No. 95, p. 14. Economics Group, Resource Management Program, ICRISAT, Patancheru, Andhra Pradesh, India. Singh, K. N., and Sharma, D. P. (2001). Response of wheat to nitrogen and potassium in saline soils. Exp. Agric. 37(3), 417 –427. Singh, R. P., Ong, C. K., and Saharan, N. (1989). Above and below ground interactions in alley cropping in semi-arid India. Agrofor. Syst. 9, 259–274. Singh, K., Rao, B. R. R., Singh, C. P., Bhattacharya, A. K., Kaul, P. N., and Rajeswara Rao, B. R. (1998). Production potential of aromatic crops in the alleys of Eucalyptus citriodora in semi-arid tropical climate of south India. J. Medicinal Aromatic Plant Sci. 20(3), 749–752. Slingerland, M. A., and Stork, V. E. (2000). Determinants of the practice of Zai and mulching in North Burkina Faso. J. Sustainable Agric. 16(2), 53– 76. Snapp, S. S., Mafongoya, P. L., and Waddington, S. (1998). Organic matter technologies for the integrated nutrient management in smallholder cropping systems of southern Africa. Agric. Ecosyst. Environ. 71, 185 –200. Somda, Z. C., Powell, J. M., Fernadez-Rivera, S., and Reed, J. (1995). Feed factors affecting nutrient excretion by ruminants and fate of nutrients when applied to soil. In “Livestock and Sustainable Nutrient Cycling in Mixed Farming Systems of sub-Saharan Africa” (J. M. Powell, S. FernadezRivera, J. Reed, S. Williams, and C. Renard, Eds.), pp. 227–246. ILCA, Addis Abbaba. Stoop, W. A., Uphoff, N., and Kassam, A. (2002). A review of agricultural research issues raised by the system of rice intensification (SRI) from Madagascar: opportunities for improving farming systems for resource-poor farmers. Agric. Syst. 71, 249 –274. Stur, W. W., Horne, P. M., Gabunada, F. A. Jr, Phengsavanh, P., and Kerridge, P. C. (2002). Forage options for smallholder crop–animal systems in Southeast Asia: working with farmers to find solutions. Agric. Syst. 71(1–2), 75–98. Styger, E., Rakotoarimanana, J. E. M., Rabevohitra, R., and Fernandes, E. C. M. (1999). Indigenous fruit trees of Madagascar: potential components of agroforestry systems to improve human nutrition and restore biological diversity. Agrofor. Syst. 46, 289 –310. Sutihar, D. N. (1984). An econometric analysis of wheat production in Nepal: a case study of Saptari district. Econ. J. Nepal 7(1), 56– 66.

554

A. GRAVES, R. MATTHEWS AND K. WALDIE

Swift, M. J., Seward, P. D., Frost, P. G. H., Qureshi, J. N., and Muchena, F. N. (1994). Long-term experiments in Africa: developing a database for sustainable land use under global change. In “Long-term Experiments in Agricultural and Ecological Sciences.” (R. A. Leigh and A. E. Johnson, Eds.), pp. 229 –251. CAB International, Wallingford, UK. Swinkels, R., and Franzel, S. (1997). Adoption potential of hedgerow intercropping in maize-based cropping systems in the highlands of western Kenya. 2. Economic and farmers’ evaluation. Exp. Agric. 33, 211–223. Swinkels, R. A., Franzel, S., Shepherd, K. D., Ohlsson, E., and Ndufa, J. K. (1997). The economics of short rotation improved fallows: evidence from areas of high population density in western Kenya. Agric. Syst. 55(1), 99–122. Szott, L. T. (1987). “Improving the production of shifting cultivation in the Amazon Basin of Peru through the use of leguminous vegetation”, Ph.D. Thesis. North Carolina State University, Raleigh, North Carolina, p. 168. Szott, L. T., and Palm, C. A. (1996). Nutrient stocks in managed and natural humid tropical fallows. Plant Soil 186, 293–309. Szott, L. T., Palm, C. A., and Buresh, R. J. (1999). Ecosystem fertility and fallow function in the humid and subhumid tropics. Agrofor. Syst. 47, 163–196. Tanner, J. C., Holden, S. J., Winugroho, M., Owen, E., and Gill, M. (1993). Feeding livestock for compost production: a strategy for sustainable upland agriculture on Java. In “International Conference on Livestock and Sustainable Nutrient Cycling in Mixed Farming Systems of SubSaharan Africa” (J. M. Powell, S. Fernandez-Rivera, T. O. Williams and C. Renard, Eds.), pp. 115–128. Addis Ababa. Tarawali, G., Manyong, V. M., Carsky, R. J., Vissoh, P. V., Osei-Bonsu, P., and Galiba, M. (1999). Adoption of improved fallows in West Africa: lessons from mucuna and stylo case studies. Agrofor. Syst. 47, 93–122. Thomas, G. V., Rohini, I., Bopaiah, B. M., and Iyer, R. (1991). Beneficial microbes in the nutrition of coconut. J. Plantation Crops 19(2), 127–138. Tian, G., Hauser, S., Koutika, L. S., Ishida, F., Chianu, J. N., Tian, G., Keatinge, D., Carsky, R., and Wendt, J. (2001). Pueraria cover crop fallow systems: benefits and applicability, sustaining soil fertility in West Africa. In “Proceedings of a symposium sponsored by the Soil Science Society of America and the American Society of Agronomy, 5– 9 November 2000”, pp. 137–155. Soil Science Society of America Inc. Madison; USA, Minneapolis, USA. van Tienhoven, N., Icaza, J., and Lagemann, J. (1982). Farming Systems in Jinotega Nicaragua. CATIE Technical Series Technical Report No. 31. CATIE, Turrialba, Costa Rica. Todaro, M. P. (1999). “Economic Development”, Addison-Wesley Publishing, Boston, USA. p. 783. Turner, M. (1995). The sustainability of rangeland to cropland nutrient transfer in semi-arid west Africa: ecological and social dimensions neglected in the debate. In “Livestock and Sustainable Nutrient Cycling in Mixed Farming Systems of sub-Saharan Africa”. Volume II: Technical papers. Proceedings of an International Conference held in Addis Ababa, Ethiopia, 22 –26 November (1993) (J. Powell, S. Fernandez-Rivera, T. Williams and C. Renard, Eds.), pp. 435 – 452. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. UNEP, (1999). “Use of Mucuna cover cropping to control Imperata cylindrica”. United Nations Development Programme, New York, NY 10017, Accessed 14-Nov-2002 online at: http://www. undp.org/tcdc/bestprac/agri/cases/benin15.htm. UNPD, (2001). World Population Prospects: The 2000 Revision “Highlights”, Population Division, Department of Economic and Social Affairs, United Nations, New York. p. 22. Uphoff, N. T. (1999). Agroecological implications of the system of rice intensification (SRI) in Madagascar. Environ. Develop. Sustainability 1(3–4), 297– 313. Valverde, B. E., Merayo, A., Reeder, R., and Riches, C. R. (1999). Integrated management of itchgrass (Rottboellia cochinchinensis) in maize in seasonally-dry Central America: facts and perspectives.

LOW EXTERNAL INPUT TECHNOLOGIES

555

“Proceedings of International Crop Protection Conference, Brighton, UK, 15–18 November 1999”, pp. 131– 140. British Crop Protection Council, Farnham, UK. Vandermeer, J. (1989). “The Ecology of Intercropping”. University of Cambridge, Cambridge. Vanlauwe, B., Aihou, K., Houngnandan, P., Diels, J., Sanginga, N., Merckx, R., Jensen, E. S., and Recous, S. (2001). Nitrogen management in ‘adequate’ input maize-based agriculture in the derived savanna benchmark zone of Benin Republic. Plant and Soil 228(1), 61–71. Versteeg, M. N., Amadji, F., Eteka, A., Gogan, A., and Koudokpon, V. (1998). Farmers’ adoptability of Mucuna fallowing and agroforestry technologies in the coastal savanna of Benin. Agric. Syst. 56(3), 269 –287. Vosti, S. A., and Witcover, J. (1996). Slash and burn agriculture — household perspectives. Agric. Ecosyst. Environ. 58, 23– 38. Vosti, S. A., Witcover, J., Oliveiara, S., and Faminow, M. (1998). Policy issues in agroforestry: technology adoption and regional integration in the western Brazilian Amazon. Agrofor. Syst. 38, 195 –222. Wakle, P. K., Khandare, V. S., Kalaskar, S. S., Karanjikar, P. N., and Daunde, A. T. (1999). Constraints faced by farmers in adoption of banana cultivation technology. J. Soils Crops 9(2), 182 –184. Waldie, K. J. (1990). Cattle and concrete: some aspects of social organisation among the Fula in and around Kabala, northern Sierra Leone, Ph.D. Thesis, Manchester University, Manchester, UK. (Available online at http://csac.anthropology.ac.uk/CSACMonog/Waldie/). Wandera, J. L., Kiruiro, E. M., Chemitei, V. C., and Bulimu, B. O. (1993). Factors limiting the use of fodders by smallholder dairy farmers in Kenya: experiences from on-farm, farmer-managed, animal feeding trials. In “Sustainable Feed Production and Utilisation for Smallholder Livestock Enterprises in Sub-Saharan Africa” (J. Ndikumana and W. de Leeuw, Eds.), AFRNET, Nairobi, Kenya. Webster, C., and Wilson, P. (1998). “Agriculture in the Tropics”. Blackwell Science, Oxford. Wendt, J. W., Jones, R. B., and Itimu, O. A. (1994). An integrated approach to soil fertility improvement in Malawi, including agroforestry. Soil fertility and climatic constraints in dryland agriculture. In “Proceedings of ACIAR-SACCAR Workshop, Harare, Zimbabwe, 30 August–1 September 1993” (E. T. Craswell and J. Simpson Eds.), pp. 74–79. Australian Centre for International Agricultural Research (ACIAR), Canberra, Australia. Wheeler, T. R., Qui, A., Keatinge, J. D. H., Ellis, R. H., and Summerfield, R. J. (1999). Selecting legume cover crops for hillside environments in Bolivia. Mountain Res. Develop. 19(4), 318 –324. Whittome, M. P. B. (1994). “The Adoption of Alley Farming in Nigeria and Benin: the On-Farm Experience of IITA and ILCA.” Ph.D. Thesis. University of Cambridge, Cambridge, England. Wolf, E. C. (1986). “Beyond the green revolution: new approaches for third world agriculture”, Worldwatch Institute, Washington DC, USA. p. 46. Wortmann, C. S., and Kaizzi, C. K. (1998). Nutrient balances and expected effects of alternative practices in farming systems of Uganda. Agric. Ecosyst. Environ. 71, 115– 129. Yamoah, C., Agboola, A. A., and Wilson, G. F. (1986a). Nutrient contribution and maize performance in alley cropping systems. Agrofor. Syst. 4, 247 –254. Yamoah, C. F., Agboola, A. A., and Mulongoy, K. (1986b). Decomposition, nitrogen release and weed control by prunings of selected alley-cropping shrubs. Agrofor. Syst. 4, 239 –246. Yaseen, T. M., Kothari, S. K., Singh, U. B., Singh, K., Sattar, A., and Roy, S. K. (2001). Bio-economic evaluation of scented rose (Rosa damascena) based intercropping systems in north Indian plains. Journal of Medicinal and Aromatic Plant Sciences 23(2), 69–74. Young, A. (1987). The potential of agroforestry for soil conservation and sustainable land use, (ICRAF Reprint No. 39). In “Collected papers” (J. Kozub, Ed.), p. 16. Economic Development Institute of the World Bank, Washington, DC, USA. Young, A. (1989). “Agroforestry for Soil Conservation”, CAB International, Wallingford, UK. p. 284.