Trees, soils, and food security

Trees, soils, and food security P E D R O A. S A N C H E Z , R O L A N D J. B U R E S H    R O G E R R. B. L E A K E Y International Centre for Research in Agroforestr’, P.O. Box 30677, Nairobi, Ken’a (icraf!cgnet.com) SUMMARY Trees have a different impact on soil properties than annual crops, because of their longer residence time, larger biomass accumulation, and longer-lasting, more extensive root systems. In natural forests nutrients are efficiently cycled with very small inputs and outputs from the system. In most agricultural systems the opposite happens. Agroforestry encompasses the continuum between these extremes, and emerging hard data is showing that successful agroforestry systems increase nutrient inputs, enhance internal flows, decrease nutrient losses and provide environmental benefits—when the competition for growth resources between the tree and the crop component is well managed. The three main determinants for overcoming rural poverty in Africa are (i) reversing soil fertility depletion, (ii) intensifying and diversifying land use with high-value products, and (iii) providing an enabling policy environment for the smallholder farming sector. Agroforestry practices can improve food production in a sustainable way through their contribution to soil fertility replenishment. The use of organic inputs as a source of biologically-fixed nitrogen, together with deep nitrate that is captured by trees, plays a major role in nitrogen replenishment. The combination of commercial phosphorus fertilizers with available organic resources may be the key to increasing and sustaining phosphorus capital. High-value trees—‘ Cinderella ’ species—can fit in specific niches on farms, thereby making the system ecologically stable and more rewarding economically, in addition to diversifying and increasing rural incomes and improving food security. In the most heavily populated areas of East Africa, where farm size is extremely small, the number of trees on farms is increasing as farmers seek to reduce labour demands, compatible with the drift of some members of the family into the towns to earn off-farm income. Contrary to the concept that population pressure promotes deforestation, there is evidence that demonstrates that there are conditions under which increasing tree planting is occurring on farms in the tropics through successful agroforestry as human population density increases. Although food insecurity occurs throughout the developing world, it is most acute in sub-Saharan Africa—hereinafter referred to as Africa—where per capita food production continues to decrease, in contrast with increases in other parts of the developing world (FAO 1996). Africa has the highest rate of population growth of any region in the world (2.9 % per year) and the highest rate (30 %) of degradation of usable land (Cleaver & Schreiber 1994). Deficiencies in vitamin A and micronutrients are also acute on this continent (IFPRI 1996). The Malthusian nightmare, although unrealistic at the global scale, could become a reality in Africa. The bulk of food in Africa is produced on small-scale farms by women. The three main determinants for overcoming rural poverty under these conditions are (i) an enabling policy environment for the smallholder farming sector ; (ii) reversing soil fertility depletion and (iii) intensifying and diversifying land use with highvalue products (Sanchez & Leakey 1997). Attaining these three goals can only be achieved in Africa with modern agricultural practices based on traditional rock-phosphate approaches (Borlaug & Dowswell 1994 ; Borlaug 1996) if fertilizers and other farming inputs are available at a price affordable by 1. I N T R O D U C T I O N The continuing threat to the world’s land resources is exacerbated by protracted rural poverty and food insecurity in the Third World, and wider climatic variations resulting from global warming. During the last decade food security was not a global priority, but studies such as the 2020 Vision (IFPRI 1996) show that rural poverty in the Third World is one of the main global concerns of our time, and that food insecurity is a major factor in rural poverty. Access for all to sufficient and nutritious food for all is the key to poverty alleviation—this was one of the main outcomes of the 1996 World Food Summit (FAO 1996). Food security encompasses both food production and the ability to purchase food. However, calories and protein are not the only factors : nutritional security includes overcoming deficiencies of vitamin A, iron, zinc, iodine and selenium (IFPRI 1996). It is also recognized that the attainment of food security is intrinsically linked with safeguarding the natural resource base (IFPRI 1996). Therefore, the three interlinked factors for reversing rural poverty are (i) income generation, (ii) increasing food and nutritional security and (iii) protecting the environment. Phil. Trans. R. Soc. Lond. B (1997) 352, 949–961 Printed in Great Britain 949 # 1997 The Royal Society 950 P. A. Sanchez and others Trees, soils, and food securit’ resource-poor farmers. They can also be achieved with agroforestry—the deliberate use of trees on farms as a low input system—a common feature of small-scale farming throughout the tropics. The purpose of this contribution is to discuss the added value of tree-based agricultural systems and to link them to the three determinants for poverty alleviation. their biomass. Biomass transfers from one site to another also provide nutrient inputs. These nutrients become inputs to the soil when the tree biomass is added to and is decomposed in the soil. The main processes are BNF, deep nitrate capture and biomass transfer. (i) Biological nitrogen fixation 2. I M P A C T O F T R E E S O N S O I L F U N C T I O N S Trees have different impacts than annual crops on soil properties, because of their longer residence time, larger biomass accumulation and continuous and more extensive root systems. In natural forest stands, nutrients are efficiently cycled with very small inputs and outputs from the system, and the soil surface is continuously protected by one or more plant canopies. In most agricultural systems, the opposite happens ; nutrient cycling is limited, while inputs and outputs are large, and the soil is not continuously protected by a plant canopy. Agroforestry encompasses the continuum between these two extremes, and emerging hard data show that specific agroforestry systems provide added value to soil processes when the competition for growth resources between the tree and the crop component is adequately managed (Ong & Huxley 1996). Such added value occurs more commonly in sequential, as opposed to simultaneous, agroforestry systems, because the competition for water, nutrients and light between the crop and tree component is separated over time (Sanchez 1995). While the effects of trees on soil functions in agroforestry systems are generally positive, the effects on crop production are often negative. This happens when the competition for light, water or nutrients is intense (Sanchez 1995). In such cases, trees decrease crop yields (van Noordwijk et al. 1996). Before considering the effects of agroforestry trees on soil properties it is imperative to deal with agronomically successful agroforestry systems. There are four ways in which trees can have beneficial effects on soil properties, crop production, and environmental protection. Trees in effective agroforestry systems (a) increase nutrient inputs to the soil, (b) enhance internal cycling, (c) decrease nutrient losses from the soil, and (d) provide environmental benefits. These ways are summarized below, based largely on reviews by the authors (Sanchez et al. 1985, 1997 ; Leakey & Newton 1994 b ; Sanchez 1995 ; Leakey et al. 1996 ; Buresh & Tian 1997 ; Sanchez & Leakey 1997). We focus on nitrogen (N) and phosphorus (P), because these are the main limiting nutrients in smallholder farms in Africa. In contrast to other continents, soil acidity and aluminum toxicity are not widespread constraints in cultivated areas of Africa (Sanchez & Leakey 1997). (a) Increased nutrient inputs Trees can provide nutrient inputs to crops in agroforestry systems by capturing nutrients from atmospheric deposition, biological nitrogen fixation (BNF), and deep in the subsoil, and storing them in Phil. Trans. R. Soc. Lond. B (1997) Although the magnitude of BNF is methodologically difficult to quantify, overall annual estimates are in the order of 25–280 kg N ha−" yr−" for leguminous trees (Giller & Wilson 1991). Woody and herbaceous legumes can provide practical means of capturing nitrogen via BNF when grown as fallows in rotation with annual crops, taking advantage of the dry season in subhumid environments when no crops can be grown. Two years of Sesbania sesban fallows in Zambia overcame nitrogen deficiencies for three subsequent maize crops (Kwesiga & Coe 1994). There is high genetic variability within tree species in their effectiveness at BNF (Sanginga et al. 1990, 1991, 1994). Phosphorus deficiencies can limit N # fixation and growth of N -fixing trees. Sanginga et al. # (1994, 1995) found large differences in early growth and P-use efficiency among and within N -fixing tree # species. These results highlight the merit of selecting provenances of N -fixing trees that are tolerant to low # available P at an early growth stage. (ii) Deep nitrate capture The uptake of nutrients by tree roots at depths where crop roots are not present can be considered an additional nutrient input in agroforestry systems. Such nutrients become an input upon being transferred to the topsoil via tree litter decomposition. Tree roots frequently extend beyond the rooting depth of crops. An exciting dimension has recently been discovered in nitrogen-deficient Nitisols of western Kenya, where mean nitrate levels in six farmers’ fields ranged from 70 to 315 kg N ha−" at 0.5–2.0 m depth (Buresh & Tian 1997). The accumulation of subsoil nitrate is attributed to greater formation of nitrate by soil organic matter (SOM) mineralization in the topsoil than the crop can absorb (Mekonnen et al. 1997). The excess nitrate then leaches to the subsoil where it is sorbed on positively charged clay surfaces, retarding the downward movement and leaching loss of nitrate (Hartemink et al. 1996). Nitrate sorption is well documented in subsoils rich in iron oxides (Kinjo & Pratt 1971). Sesbania sesban fallows deplete this pool, thus capturing a resource that was unavailable to the maize crop (Mekonnen et al. 1997). These relationships are shown in figure 1. In soils with high quantities of subsoil nitrate, a N # fixing tree should, ideally, be able to rapidly take up the subsoil nitrate before it can be leached. When the tree has depleted subsoil nitrate, it should then ideally meet a substantial proportion of its N requirements through BNF. Under such conditions, agroforestry trees become a biological safety net. How extensive are these soils ? There are 260 million hectares of Nitisols (oxic or rhodic Alfisols and Oxisols) and similar soils in Africa Trees, soils, and food securit’ P. A. Sanchez and others 951 nitrate N (kg ha–1 0.5 m layer–1) 0 20 40 60 80 0 20 40 60 80 0.5 0.5 depth (m) start start 1.0 1.0 1.5 1.5 (b) (a) 2.0 2.0 Figure 1. Nitrate accumulates in the subsoil of this Oxisol, near Maseno, Western Kenya. (a) Maize is unable to access this pool, while Sesbania sesban (b) depletes it. (Adapted from Hartemink et al. and Buresh, unpublished data.) that have anion exchange capacity in the subsoil, where roots of Sesbania and similar agroforestry trees can penetrate (Sanchez et al. 1997). Assuming that one-tenth of them are under cultivation, the magnitude of this resource could be in the order of 3 million tonnes (Mt) of nitrate nitrogen, much more than the annual nitrogen fertilizer consumption rate, 0.8 Mt of nitrogen in sub-Saharan Africa, excluding South Africa (FAO 1995). We do not yet know the extent to which this resource is renewable. Nevertheless, the utilization of this hitherto unrecognized nitrogen source via its capture by deep-rooted trees is an exciting area of research in Africa, as well as in other regions with similar oxidic subsoils. for such high concentrations remain speculative but members of the Compositae family, to which Tithonia belongs, have a reputation for being nutrient scavengers. The processes involved are not presently identified, but may involve the dissolution of inorganic phosphorus, desorption of fixed soil phosphorus by root exudates, organic acids and}or extremely effective mycorrhizal associations. Woody species grown in hedges outside the cultivated fields, therefore, may be able to transform less available inorganic forms of phosphorus into more available organic forms, as well as supply significant quantities of N and K, when their leaves are incorporated into the soil as biomass transfers. (iii) Biomass transfer The leafy biomass of trees is frequently cut from hedges or uncultivated areas and incorporated into crop fields as a source of nutrients in Africa. While the quantities of biomass farmers are able to apply are often sufficient to supply N to a maize crop with a moderate grain yield of 4 t ha−", they seldom can supply sufficient P to that crop (Palm 1995). Leguminous trees are most frequently used as biomass transfer systems, but there is increasing evidence that some non-leguminous shrubs may also accumulate high concentrations of nutrients in their biomass. Tithonia diŠersifolia, a common hedge species found at middle elevations throughout East Africa and SouthEast Asia has unusually high nutrient concentrations (3.5 % N ; 0.38 % P and 4 % K) in its leaf biomass (Gachengo 1996 ; Niang et al. 1996). These P and K (potassium) levels are higher than those of commonly used legumes in agroforestry (Palm 1995). Reasons Phil. Trans. R. Soc. Lond. B (1997) (b) Enhanced nutrient cycling Trees in agroforestry systems can increase the availability of nutrients in the soil through the conversion of nutrients to more labile forms of soil organic matter (SOM). Plants convert inorganic forms of N and P in the soil solution into organic forms in their tissues. The addition of in situ-grown plant material to the soil as litterfall, root decay, green manures, crop residue returns (or animal manures in grazing systems), and its subsequent decomposition results in the formation of organic forms of soil N and P. Mineralization of soil organic N or P converts them once again to nitrates and orthophosphate ions in the soil solution which are readily available to plants. This is the process of cycling. It is important to distinguish organic cycling from organic inputs. Cycling involves organic materials 952 P. A. Sanchez and others Trees, soils, and food securit’ grown in situ, such as those described in the previous paragraph. They do not add N or P to the soil–plant system, except for additional biological N fixation and # capture from below the crop rooting depth, and therefore do not constitute inputs from outside the system. Biomass transfers, composts, and manures produced outside the field are the true organic inputs. Total SOM generally does not relate to crop yields (Sanchez & Miller 1986). Nutrient release from SOM is normally more dependent on its biologically active fractions than on total SOM quantity. Microbial biomass P, light fraction organic N and P, and NaOHextractable organic P appear to be relevant fractions in agroforestry systems (Buresh & Tian 1997). (i) Soil organic nitrogen Agroforestry tree species vary greatly in their quality, usually measured by the (lignin­phenolics)}N ratio of their leaves (Palm & Sanchez 1991 ; Constantinides & Fownes 1994 ; Schroth et al. 1995 ; Tian et al. 1995 ; Jonsson et al. 1996). High-quality materials are readily mineralized, while low-quality ones decompose slowly and may eventually form part of soil organic pools. For example, Barrios et al. (1997) found that N availability, as determined by inorganic soil N, N in light fraction SOM, and N mineralization in topsoil was higher in maize plots following improved fallow species with the lowest (lignin­polyphenol)}N ratios in leaf litter in an N-deficient Alfisol in eastern Zambia. Sesbania sesban fallows and fertilized maize monocultures resulted in similar inorganic soil N levels, but N mineralization and light fraction N were greater after S. sesban. The amount of light fraction N appears to be a sensitive measure of SOM differences among cropping systems and is correlated with N mineralization of the whole soil (Barrios et al. 1996 a, b). Light fraction SOM can be increased by addition of tree biomass to maize (Barrios et al. 1996 a) and by rotation of maize with planted tree fallows (Barrios et al. 1997). Appropriate agroforestry systems, therefore, seem to enhance internal N flows. (ii) Soil organic phosphorus Most studies have found little or no benefit of trees in agroforestry systems on inorganic soil P tests (Drechsel et al. 1991 ; Siaw et al. 1991 ; Kang et al. 1994, 1997). Methods related to labile soil organic P fractions seem more appropriate for agroforestry systems with little or no inorganic P inputs. For example, S. sesban fallows, as compared to continuous unfertilized maize, increased soil P availability, measured by chloroformextractable P and P in light fraction SOM, but had no effect on extractable inorganic soil P (Maroko et al. 1997). Sesbania sesban fallows, compared with continuous unfertilized maize, increased maize yields when P was the limiting nutrient, but they did not eliminate the need for external P inputs to completely overcome the P deficiency. Some trees and shrubs, but apparently few crop species, have the ability to exude organic acids from their roots or mycorrhizal associations and dissolve inorganic soil phosphates not otherwise available to roots of crop plants (Lajtha & Harrison 1995). Pigeon Phil. Trans. R. Soc. Lond. B (1997) pea (Cajanus cajan) secretes pisidic acid in calcareous soils (Ae et al. 1990 ; Otani et al. 1996), increasing the plant’s phosphorus uptake, while Inga edulis is believed to have access to phosphorus not available to maize and beans (Hands et al. 1995). Both these species are legumes, which are known to acidify their rhizosphere in the process of nitrogen fixation. In such cases, organic cycling has the advantage of transforming otherwise unavailable inorganic soil phosphorus into more available organic forms. Agroforestry will not eliminate the need for P fertilizers on P-deficient soils (Buresh et al. 1997). The integration of organic materials with inorganic P fertilizers is likely to enhance the availability of P added from inorganic fertilizers (Palm et al. 1997). There are, at present, no methods for quantifying nutrient cycling efficiency in agroecosystems and its effects on productivity and sustainability. This is an area that requires further conceptualization, and a start has been made by van Noordwijk (1997) who describes possibilities at various spatial and temporal scales. (c) Decreased nutrient losses from the soil Losses caused by runoff, erosion and leaching account for about half of the N, P and K depletion in Africa (Smaling 1993). Agroforestry systems have been found to decrease nutrient losses by runoff and erosion to minimal amounts (Lal 1989 a ; Young 1989). The evidence for decreased leaching losses is less comprehensive. Horst et al. (1995) reported that Leucaena leucocephala hedgerows reduced nitrate leaching as compared with a no-tree control on a sandy Ultisol in the Benin Republic. Lower subsoil water provided indirect evidence of reduced leaching loss of nutrients under trees in agroforestry systems of western Kenya (ICRAF 1996). Subsoil water in S. sesban fallows seldom exceeded field capacity in a clayey Oxisol despite a mean annual rainfall of about 1800 mm. Subsoil water in the natural uncultivated fallow and maize monoculture at the same site occasionally exceeded field capacity, indicating that mobile water was present to transport nitrate downward. Low subsoil water and nitrate content under S. sesban were attributed to high water and N demand by the fast-growing tree. (d) Environmental benefits Trees protect the soil surface via two canopies : the litter layer and the leaf canopy, thereby decreasing runoff and erosion losses, dampening temperature and moisture fluctuations and in most cases, maintaining or improving soil physical properties (Sanchez et al. 1985 ; Lal 1989 b, c ; Hulugalle & Kang 1990 ; Hulugalle & Ndi 1993 ; Rao et al. 1997). In agroforestry systems, the beneficial effects of protecting the soil surface depend on the spatial and temporal coverage of the tree component. Also, tree roots can loosen the topsoil by radial growth, and improve porosity in the subsoil when roots decompose. The perennial nature of tree root systems provides a dependable source of carbon Trees, soils, and food securit’ P. A. Sanchez and others 953 substrate for microorganisms in the rhizosphere ; microbial mucilage binds soil particles into stable aggregates, which results in improved soil structure (Tisdall & Oades 1982). These two processes, surface soil protection and root penetration, take place continually in agroforestry systems instead of temporarily, as in agricultural systems. Due to these, three major kinds of environmental benefits ensue : soil conservation, biodiversity conservation, and carbon sequestration. (i) Soil conserŠation Many agroforestry systems help keep the soil in place by biological instead of engineering means (Lal 1989 a ; Young 1989 ; Kiepe & Rao 1994 ; Juo et al. 1995 ; Rao et al. 1997). While contour hedges do require management, although certainly less than earth terraces, they also become a productive niche on the farm while conserving the soil. Controlling soil erosion biologically has an additional advantage : the slope between the hedges becomes less steep and even flat in some cases (Kiepe & Rao 1994 ; Garrity 1996). These ‘ biological terraces ’ are produced by taking advantage of the erosion process within the contour hedges, with the vegetative growth keeping up with the higher soil surface at the lower end, something non-biological terraces cannot do. Trees, however, do not conserve the soil until they are well established and have developed a litter layer (Sanchez et al. 1985). Once established, most trees protect the soil constantly, provided they are healthy and the litter layer is not removed. Biomass transfer of tree leaf litter to cropped fields undermines this process (Nyathi & Campbell 1993). (ii) BiodiŠersit’ conserŠation All agroforestry systems are more diverse than crop or forest plantation monocultures, while some, such as the complex agroforests of South-East Asia, are nearly as diverse as natural forests (Thiollay 1995). But, importantly, agroforestry also helps to conserve plant and animal biodiversity by reducing the further clearance of tropical forests through viable alternatives to slash-and-burn agriculture (Sanchez 1994 ; Schroeder 1994). Precise estimates of these substitution values do not exist for agroforestry systems, although figures of 7.1 and 11.5 hectares saved for each hectare put into successful agroforestry have been reported (Schroeder 1993). Multistrata or complex agroforests are one such alternative to slash-and-burn. In these systems, annual food crops are planted along with trees, and cover the ground quickly until they are shaded out by these trees, which in turn eventually occupy different strata and produce high-value products such as fruits, resins, medicinals and high-grade timber (de Foresta & Michon 1994 ; Michon & de Foresta 1996). Plant diversity is in the order of 300 species ha−" in the mature, complex rubber agroforests of Sumatra, Indonesia. This level of plant biodiversity by far exceeds that of rubber plantations (5 species ha−") and approximates to that of adjacent undisturbed forests with 420 plant species ha−". The richness of bird species Phil. Trans. R. Soc. Lond. B (1997) in mature damar (Shorea jaŠanica)-based agroforests is approximately 50 % that of the original rainforest (Thiollay 1995), and almost all mammal species are present in the agroforest (Sibuea & Herdimansyah 1993). This is possible because such agroforests, composed of hundreds of small plots managed by individual families, occupy contiguous areas of several thousand hectares in Sumatra. Tracks of the rare Sumatran rhino (Dicerorhinus sumatrensis) were recently discovered in one of these rubber agroforests, implying that they may provide a habitat similar to the natural rainforest (Sibuea 1995). Such high biodiversity levels, however, cannot be expected of shorter duration agroforestry systems, such as improved fallows, or in systems that are less geographically extensive. Agroforestry plays a major role in the reclamation of degraded and abandoned lands, and is generally considered the most workable approach to mimic natural forest succession and increase biodiversity (Anderson 1990). Hard data on increasing biodiversity in degraded lands through agroforestry, however, are practically non-existent (Sanchez et al. 1994). Below-ground biodiversity is also higher in agroforestry systems than in crop monocultures, approximating the levels of the natural forest in the Amazon (Lavelle & Pashanasi 1989). Soil macrofauna and microflora are key regulators of the basic decomposition processes that provide nutrients to higher plants and animals. While they are not as attractive as ‘ furry and feathered creatures ’, soil communities are a major component of biodiversity conservation and ecosystem functioning. (iii) Carbon sequestration Agroforestry systems help keep carbon in the terrestrial ecosystem and out of the atmosphere by preventing further deforestation and by accumulating biomass and soil carbon (Schroeder 1994). As with biodiversity conservation, the main contribution of improved agroforestry systems to terrestrial carbon conservation comes from its preventive effect, i.e. the area of natural forests that will not be cleared because farmers can make continuous use of already cleared land through improved agroforestry systems (Schroeder 1993 ; Unruh et al. 1993 ; Sanchez 1994). One hectare of humid tropical forests contains on average 160 t C (carbon) ha−" in the above-ground biomass (Houghton et al. 1987). When it is slashed and burned, most of it is emitted to the atmosphere, either immediately during the burn, or gradually through the decomposition of unburned logs and branches. Keeping this carbon resource (some 96 billion tonnes of C in the remaining humid tropical forest biomass) in situ is of critical importance. Complex agroforestry systems of long duration, such as the jungle rubber and damar agroforests of Sumatra and multistrata systems throughout the humid tropics, can sequester carbon in their tree biomass, where it remains for decades. In addition, complex agroforests act as sinks for methane emitted by adjacent paddy fields, thereby neutralizing these greenhouse gas emissions at the landscape scale (Murdiyarso et al. 1996). P. A. Sanchez and others Trees, soils, and food securit’ The greatest potential for carbon sequestration is probably in soils that have been depleted of carbon and nutrients and have the potential to regain their original SOM levels. Woomer et al. (1997) estimate that 66 tonnes ha−" of carbon can be sequestered in woody biomass and nutrient-depleted soils in Africa over a 20year period by a combination of nutrient recapitalization, erosion control, boundary tree plantings and woodlot or orchard establishment. The overall magnitude of carbon sequestration by agroforestry is considered among the highest compared with other land-use systems by climate change researchers. Unruh et al. (1993) performed complex calculations of agroforestry systems in Africa, their biomass accumulation, and their potential distribution using GIS techniques. Their results suggest that a huge amount of carbon can be sequestered, ranging from 8–54 Gt (billion tonnes) of C in a total of 1.55 billion hectares where agroforestry could potentially be practised. This represents the theoretical upper limit. Above- and below-ground carbon sequestration values, however, need to be generated locally, taking into account the duration of each agroforestry system, and extrapolated geographically in a realistic fashion, based on actual rates of agroforestry adoption. 3. T R E E S A N D O V E R C O M I N G R U R A L POVERTY IN AFRICA While agroforestry trees may improve soil fertility, nutrient use efficiency, and provide major environmental benefits, they are not likely to have a significant impact on food security or alleviate poverty by themselves. Successful agroforestry can contribute to (a) food security from the production point of view through soil fertility replenishment, along with fertilizers, and (b) poverty alleviation and access to enough and nutritious food through the domestication of indigenous trees, and (c) enabling policies. This section examines these possibilities. (a) Soil fertility replenishment Soil fertility depletion in smallholder farms in Africa is beginning to be recognized as the fundamental biophysical limiting factor responsible for the declining per capita food production of the continent (IFPRI 1996 ; Sanchez et al. 1996, 1997). The magnitude of nutrient mining is huge, as evidenced by nutrient balance studies. An average of 660 kg of N, 75 kg of P and 450 kg of K ha−" has been lost during the last 30 years from about 200 million ha of cultivated land in 37 African countries. The total annual nutrient depletion in sub-Saharan Africa is equivalent to 7.9 Mt yr−" of N, P and K, six times the amount of annual fertilizer consumption to the region, excluding South Africa (Sanchez et al. 1997). Nutrient capital has gradually been depleted by crop harvest removals, leaching and soil erosion. This is because farmers did not sufficiently compensate these losses by returning nutrients to the soil via crop residues, manures and inorganic fertilizers. The consequences of nutrient depletion are felt at the farm, watershed, national and global scales, and include major economic, social and environmental Phil. Trans. R. Soc. Lond. B (1997) 12 y = 1.0+0.018x r2 = 0.64 10 planted trees (m3 ha–1) 954 8 6 4 2 0 0 100 200 300 400 population density (people 500 600 km–2) Figure 2. The effect of nitrogen source, as either urea or Tithonia diŠersifolia biomass transfer (1.8 t ha−" of dry mass), with Minjingu rock-phosphate (RP) and triple superphosphate (TSP). Both applied at a recapitalization rate of 250 kg ha−" of P, on maize grain yield on an acid soil near Maseno, Kenya. The amounts of N supplied by urea and T. diŠersifolia were the same, 60 kg ha−" of N. (Adapted from Buresh et al. 1997.) externalities. Sanchez et al. (1996, 1997) suggested that soil fertility replenishment should be considered as an investment in natural resource capital. Phosphorus replenishment strategies are mainly fertilizer-based, with biological supplementation, while N-replenishment strategies are mainly biological, with chemical supplementation. Replenishing phosphorus capital can be accomplished by large applications of P fertilizers in high P-fixing soils. Africa has ample rockphosphate deposits that could be used directly or as superphosphates to reverse phosphorus depletion. One of the problems is the need to add acidifying agents to rock-phosphates, in order to facilitate their dissolution in many P-depleted African soils that have pH values above 6.0, which are too high for acidification to occur at a rapid rate. Decomposing organic materials produce organic acids that may help acidify rock-phosphate. Mixing rock-phosphates with compost has shown promise in increasing the availability of rock-phosphate at sites in Burkina Faso (Lompo 1993) and Tanzania (Ikerra et al. 1994). Organic acids produced during the decomposition of plant materials may temporarily reduce the P-fixation capacity of the soils by binding to the oxides and hydroxide surfaces of clay particles (Iyamuremye & Dick 1996). Through this process P availability and nutrient-use efficiency are temporarily increased. Research in western Kenya with Minjingu rockphosphate and triple superphosphate indicates higher maize yields following incorporation of P with T. diŠersifolia, rather than urea, at an equivalent N rate (figure 2). The benefit from T. diŠersifolia was partially attributed to the addition of K and about 5 kg of P ha−" (Buresh et al. 1997). Subsequent research confirmed higher maize production with sole application of T. diŠersifolia biomass than with an equivalent rate of NPK mineral fertilizer on a P- and K- deficient soil Trees, soils, and food securit’ P. A. Sanchez and others 955 (Bashir Jama et al., unpublished data). The integration of available organic resources, such as T. diŠersifolia, with commercial P fertilizers may be important in increasing and sustaining soil phosphorus capital (Palm et al. 1997). Given the largely biological nature of the nitrogen cycle, the use of organic inputs, as a source of biologically-fixed nitrogen and deep nitrate capture, plays a crucial role in N replenishment. Agroforestry trees and herbaceous leguminous green manures play a major role in internal cycling. Organic inputs have an important advantage over inorganic fertilizers with regard to fertility replenishment ; they provide a carbon source for microbial utilization, resulting in the formation of soil organic N. Inorganic fertilizers do not contain such carbon sources ; therefore, most of the fertilizer N not used by crops is subject to leaching and denitrification losses, while much of the N released from organic inputs and not utilized by crops could build soil organic N capital (Sanchez & Palm 1996). Nitrogen fertilizers are likely to be needed to achieve high crop yields on top of the nutrient contributions of agroforestry (Sanchez et al. 1996). Accompanying technologies and enabling policies are needed to make recapitalization operational. Soil conservation technologies must be present in order to keep the nutrient capital investment in place, and to avoid polluting rivers and groundwaters. Policy improvements are needed to provide the timely availability of the right types of fertilizers at reasonable cost, better infrastructure, credit, timely access to markets, adaptive research and extension education— particularly in the combined use of organic and inorganic sources of nutrients. The issue of who should pay for this recapitalization is based on the principle that those who benefit from a course of action should incur the costs of its implementation. On-farm maintenance costs should be borne by farmers, whereas national and global societies should share the more substantial costs of actual phosphorus applications. This sharing should reflect the ratio of national to global benefits (Sanchez et al. 1997). (b) Intensifying and diversifying land use though tree domestication Soil fertility replenishment can go a long way in boosting agricultural production in Africa. However, although it is necessary it is not sufficient for attaining food security and eliminating rural poverty— particularly considering the economic constraint on farmers’ affording fertilizers. Numerous other factors have to come together as well, such as post-harvest losses, pests and disease attacks, the declining size of land holdings and declining human health. The last two factors have an impact on the availability of field labour that is also a consequence of family members moving to the town to secure off-farm income. What is needed is a paradigm shift from policies directed only at increasing yields of the few staple food crops to one geared at ‘ putting money in farmers’ pockets ’. This rock-phosphate approach has played, and will continue to play, an important part in meeting the needs of the Phil. Trans. R. Soc. Lond. B (1997) rural poor, but additional steps must also be taken. It is in this vein that Sanchez & Leakey (1997) suggest that a further transformation is needed in the long run : intensifying and diversifying land of smallholder farms in Africa in ways that generate income for farmers so that they have the option to invest in farm inputs. President Yoweri Museveni of Uganda, in his opening address to a SPAAR (Special Program for African Agricultural Research) meeting in Kampala, 6 February 1996, articulated this idea very clearly. He stated : ‘ It does not make sense to grow low-value products (maize and beans) at a small-scale ; instead, high-value products should be grown at a small-scale, while low-value products should be grown on a largescale ’. The obvious implication is that small-scale farming in Africa must diversify by producing a combination of high-value, profitable crops along with the basic food crops. Examples of this strategy occur in western Kenya, where small patches—in the order of 100 m#— of French beans are grown by smallholders contracted by an exporting company for fresh consumption in Europe. The market is assured, and farmers intensively water, fertilize and weed these islands of wealth among their lower value crops. But the largest opportunities for farm diversification come from trees producing an array of marketable products. Traditionally, people throughout the tropics have depended on indigenous plants for fruits and everyday household products, from medicines to fibres. These products have also provided the essential vitamins and minerals for family health, and through local and regional trading have generated cash to meet household needs for purchased products and services. Maybe it is here, in peoples’ own backyard, that the solution lies. But sadly, through deforestation, the forest or woodland that used to be in the farmers’ backyard has now all but disappeared for the vast majority of people in Africa. This is where tree domestication as part of agroforestry becomes so important. Already there is a body of biophysical information on the techniques available to domesticate a wide range of wild tree species (Leakey & Newton 1994 a, b ; Newton et al. 1994 ; Franzel et al. 1995 ; Leakey et al. 1996). Furthermore, guidelines have been developed for determining the species priorities of farmers (Franzel et al. 1996 ; Jaenicke et al. 1996). These ‘ Cinderella ’ species—so called because their value has been largely overlooked by science although appreciated by local people—include indigenous fruit trees and other plants that provide medicinal products, ornamentals, or high-grade timber. Some examples are shown in table 1. Techniques being developed to convert some of these wild species into domesticated crops in agroforestry systems include vegetative propagation and clonal selection designed to capture genetic diversity (Leakey & Jaenicke 1995). Domestication involves the formulation of a genetic improvement strategy for agroforestry trees and a strategy for the use of vegetative propagation to capture the additive and non-additive variation of individual trees in tree populations (Simons 1996). The domestication strategy 956 P. A. Sanchez and others Trees, soils, and food securit’ Table 1. Examples of ‘ Cinderella ’ tree species With high potential for domestication (Leake’ et al. 1996) species common names ecoregion products IrŠingia gabonensis Uapaca kirkiana Sclerocar’a birrea Bactris gasipaes bush mango, mango de souvage miombo of Southern Africa miombo of Southern Africa peach palm, pejibaye, pupunha, pijuayo, chontaduro karite! , shea nut pigeum johimbe humid West Africa fruit fruit, beverage Western Amazonia fruit, kernels Viterallia paradoxa Prunus africana Pausin’stalia johimbe for these indigenous fruit tree species, as well as for Prunus africana and Pausin’stalia johimbe, two priority trees for medicinal products, is to conserve their genetic resource in living-germplasm banks and subsequently to develop cultivars for incorporation into multistrata agroforests (Leakey & Simons 1997). High-value trees can fit in specific niches on farms, making the system ecologically stable and more rewarding economically, thus diversifying and increasing rural incomes and improving food security. Timber trees can also be grown on farm boundaries with leguminous fodder trees under them. Similarly, fuelwood trees can be grown on field boundaries or as contour hedges on sloping lands. In such a scheme, improved fallows become a crucial part of the crop rotation. The result is that farm income is increased and diversified, providing resilience against weather or price disruptions, soil erosion is minimized, nutrient cycling is maximized and above- and below-ground biodiversity is enhanced. The farm truly approximates a functioning ecosystem. The latest definition of agroforestry summarizes this approach : a dynamic, ecologically-based, natural resource management system that, through the integration of trees in farms and in the landscape, diversifies and sustains smallholder production for increased social, economic and environmental benefits (Leakey 1996). Through domestication these tree crops could become higher yielding, produce higher quality products, be more attractive commercially, and diversify diets (Leakey et al. 1996). Such progress could improve household welfare by providing traditional food and health products, boosting trade, generating income and diversifying farming systems, both biologically and economically, beyond the production of basic food crops. Generally, tree crops have lower labour requirements than basic food crops, and could thus allow farmers time for off-farm income generation. A new paradigm for smallholder farming in Africa emerges : one that instead of being based on a limited number of highly domesticated crops, often grown in monoculture, is based on a much greater diversity of commercially important plants that together produce food and high-value products (Leakey & Izac 1996). (c) Enabling policies Current policy recommendations place a high priority on the revitalization of the agricultural sector in Africa (FAO 1996 ; IFPRI 1996), and some success stories are beginning to emerge (Cleaver & Schreiber Phil. Trans. R. Soc. Lond. B (1997) Sahel montane tropical Africa humid West Africa fruit, heart of palm, parquet floors, fibres cosmetics, oils medicinal medicinal 1994). The fact that most food in Africa is produced by smallholders, often female farmers, is frequently considered a major constraint to agricultural development. In contrast, we believe that small-scale farms can be an asset rather than a liability when supported by appropriate policies. The agricultural production boom in Asia is a product of smallholder farms and not of a shift from small- to large-scale farming. The policies include improvements in land tenure, infrastructure, marketing information, credit, research, extension and access to inputs and markets at reasonable prices (Place 1996). Public investment to increase access to education of girls and improve public health services in rural areas also plays an important role in this transformation process. Policy reform to seize opportunities for smallholder development and to eliminate policies that discriminate against the smallholder agricultural sector therefore remains a top priority. Indeed, policy reform is a necessary, but not a sufficient condition for food security and environmental conservation. In order for enabling policies to work in most of Africa, the twin issues of soil fertility depletion and land-use intensification and diversification have to be tackled. Therefore, the vision now is of agroforestry as an integrated land use policy that combines increases in productivity and income generation with environmental rehabilitation and the diversification of agroecosystems. Such a vision can be fitted to the range of situations found in the major ecoregions of Africa. According to Cooper et al. (1996) and Sanchez et al. (1997), the realization of this vision, however, is going to be dependent on (i) the appreciation by the international community of the importance of soil fertility replenishment and high-value indigenous species in the lives and welfare of local people, as well as incentives (or the removal of disincentives) for local people to plant trees on their farms ; (ii) replenishment of plant nutrients, that can also be viewed as an investment in natural resource capital, similar to investments in dams and irrigation ; (iii) the domestication of commercially-important indigenous tree species producing high-value products ; and (iv) the development of processing infrastructure at the rural scale and a dynamic market perspective at the national and global scales. Commercialization is both necessary and potentially harmful. It is necessary because without it the market for products is small, and the opportunity for rural people to make money would not exist. A degree of product domestication is therefore essential. On the Trees, soils, and food securit’ P. A. Sanchez and others 957 other hand, commercialization is potentially harmful to rural people if it expands to the point where outsiders with capital to invest come in and develop large-scale monocultural plantations. However, from the experience of the complex agroforests in South-East Asia (de Foresta & Michon 1994 ; Michon & de Foresta 1996), smallholder units producing non-timber forest products that are also biologically diverse and economically viable, indicate that the intensification and diversification of land use is not a pipe-dream. 4. T H E W A Y F O R W A R D While land use intensification caused by demographic pressure is generally associated with environmental degradation, the long-term relationship between land resource degradation and demographic pressure is not necessarily negative and linear (Harwood 1994 ; Scherr & Hazel 1994). With further increases in population pressure, however, a point is reached where degradation is reversed, with further land intensification and incorporation of trees within the farm. This has happened in the semi-arid Machakos District of Kenya, where despite increasing population pressure since the 1930s, farmers were able to reverse land degradation through an indigenous soil conservation technology that improved both crop and livestock productivity (Pagiola 1994 ; Tiffen et al. 1994). This technology did not have a major agroforestry component, but recent evidence in eastern Africa indicates that the same is true with agroforestry. In the more heavily populated areas of Burundi (Place 1995), Kenya (Holmgren et al. 1994 ; Bradley et al. 1995 ; Patel et al. 1995) and Uganda (Place & Otsuka 1997) where farm size is extremely small, the number of trees on farms is also expanding as farmers increasingly recognize their value (figure 3). In fact, much of the reforestation in the tropics is taking place on farms, though agroforestry, and not as plantations (J. Spears, personal communication). Most of the planted trees are generally of low value and used for fodder, fuelwood, boundary delineation and exotic fruits like avocado and mango. The next step is to incorporate high-value domesticated trees into these farms. If the three determinants are realized—replenished soils, high value trees and enabling policies—Africa will be facing a win-win-win situation (socially, economically and ecologically) where poverty alleviation, food security and environmental protection go hand in hand. REFERENCES Ae, N., Akihara, J., Okada, K., Yoshinara, T. & Johansen, C. 1990 Phosphorus uptake by pigeon pea and its role in cropping systems in the Indian subcontinent. Science 248, 477–480. Anderson, A. B. (ed.) 1990 AlternatiŠes to deforestation : steps toWard sustainable use of the Ama“on rainforest. New York : Columbia University Press. Barrios, E., Buresh, R. J. & Sprent, J. I. 1996 a Organic matter in soil particle size and density fractions from maize and legume cropping systems. Soil Biol. Biochem. 28, 185–193. Barrios, E., Buresh, R. J. & Sprent, J. I. 1996 b Nitrogen mineralization in density fractions of soil organic matter from maize and legume cropping systems. Soil Biol. Biochem. 28, 1459–1465. Barrios, E., Kwesiga F., Buresh, R. J. & Sprent, J. I. 1997 Light fraction soil organic matter and available nitrogen following trees and maize. Soil Sci. Soc. Am. J. (In the press). Borlaug, N. & Dowswell, C. R. 1994 Feeding a human population that increasingly crowds a fragile planet. Supplement to Transactions of the 15th World Congress of Soil Science, Acapulco, Mexico. Chapingo, Mexico : International Society of Soil Science. a maize grain yield (tons ha–1) 4 a 3 b 2 1 0 bc c d No P Urea No P Tithonia TSP Urea TSP Tithonia RP Urea RP Tithonia Figure 3. Correlation between high population density and planted woody biomass for districts in high potential areas of Kenya (Holmgren et al. 1994). Phil. Trans. R. Soc. Lond. B (1997) 958 P. A. Sanchez and others Trees, soils, and food securit’ Borlaug, N. 1996 Mobilizing science and technology for a rock-phosphate in African agriculture. In AchieŠing greater impact from research inŠestments in Africa (ed. S. A. Breth), pp. 209–217. Mexico City : Sasakawa Africa Association. Bradley, P. N., Chavangi, N. & van Gelder, A. 1995 Development research and energy planning in Kenya. Ambio 14, 228–236. Buresh, R. J., Smithson, P. C. & Hellums, D. 1997 Building up soil P capital in sub-Saharan Africa. In Replenishing soil fertilit’ in Africa. ASA-SSSA Special Publication (In the press.) Buresh, R. J. & Tian, G. 1997 Soil improvement by trees in sub-Saharan Africa. Agroforestr’ s’stems (In the press.) Cleaver, K. M. & Schreiber, G. A. 1994 ReŠersing the spiral ; the population, agriculture and enŠironment nexus in sub-Saharan Africa. Washington, DC : World Bank. Constantinides, M. & Fownes, J. H. 1994 Nitrogen mineralization from leaves and litter of tropical plants : relationships to nitrogen, lignin and soluble polyphenol concentrations. Soil Biol. Biochem. 26, 49–55. Cooper , P. J. M., Leakey, R. R. B., Rao, M. R. & Reynolds, L. 1996 Agroforestry and the mitigation of land degradation in the humid and sub-humid tropics of Africa. Expl Agric. 32, 235–290. de Foresta, H. & Michon, G. 1994 Agroforests in Sumatra : where ecology meets economy. Agroforestr’ Toda’ 6, 12–13. Drechsel, P., Glaser, B., & Zech, W. 1991 Effect of four multipurpose tree species on soil amelioration during tree fallow in Central Togo. Agroforestr’ S’stems 16, 193–202. FAO 1995 FAO fertili“er ’earbook Šolume 44—1994. Rome : Food and Agriculture Organization of the United Nations. FAO 1996 World food summit : s’nthesis of the technical background documents. Rome : Food and Agriculture Organization of the United Nations. Franzel, S., Jaenicke, H. & Janssen, W. 1996 Choosing the right trees : setting priorities for multipurpose tree improŠement. ISNAR Research Report No. 8. The Hague, Netherlands : ISNAR. Gachengo, C. N. 1996 Phosphorus release and aŠailabilit’ on addition of organic materials to phosphorus fixing soils. M.Sc. thesis, Moi University, Eldoret, Kenya. Garrity, D. P. 1996 Tree–soil–crop interactions on slopes. In Tree–crop interactions, a ph’siological approach (ed. C. K. Ong & P. A. Huxley), pp. 299–318. Wallingford, UK : CAB International. Giller, K. E. & Wilson, K. J. 1991 Nitrogen fixation in tropical cropping s’stems. Wallingford, UK : CAB International. Hands, M. R., Harrison, A. F. & Bayliss-Smith, T. 1995 Phosphorus dynamics in slash-and -burn and alley cropping systems of the humid tropics. In Phosphorus in the global enŠironment (ed. H. Tiessen), pp. 155–170. Chichester, UK : John Wiley & Sons. Hartemink, A. E., Buresh, R. J., Jama, B. & Janssen, B. H. 1996 Soil nitrate and water dynamics in Sesbania fallow, weed fallows, and maize. Soil Sci. Soc. Am. J. 60, 568–574. Harwood, R. R. 1994 Agronomic alternatives to slash-andburn in the humid tropics. In AlternatiŠes to slash-and-burn agriculture (ed. P. A. Sanchez & H. van Houten), pp. 93–106. Symposium ID-6, 15th World Congress of Soil Science Acapulco, Mexico. Chapingo, Mexico : International Society of Soil Science. Holmgren, P., Masakha, E. J. & Sjo$ holm, H. 1994 Not all African land is being degraded : a recent survey of trees on farms in Kenya reveals rapidly increasing forest resources. Ambio 23, 390–395. Horst, W. J., Ku$ hne, R. & Kang, B. T. 1995 Nutrient use in Leucaena leucocephala and Cajanus cajan in maize}cassava alley cropping on Terre de Barre, Benin Republic. In Alle’ Phil. Trans. R. Soc. Lond. B (1997) farming research and deŠelopment (ed. B. T. Kang, A. O. Osiname & A. Larbi), pp. 122–136. Ibadan, Nigeria : Alley Farming Network for Tropical Africa. Houghton, R. A., Boone R. D. & Fruci, J. R. 1987 The flux of carbon from terrestrial ecosystems to the atmosphere in 1980 due to changes in land use : geographic distribution of the global flux. Tellus 39B, 122–139. Hulugalle, N. R. & Kang, B. T. 1990 Effect of hedgerow species in alley cropping systems on surface soil physical properties of an Oxic Paleustalf in southwestern Nigeria. J. Agric. Sci. (Cambridge, UK) 114, 301–307. Hulugalle, N. R. & Ndi, J. N. 1993 Effects of no-tillage and alley cropping on soil properties and crop yields in a Typic Kandiudult of southern Cameroon. Agroforestr’ S’stems 22, 207–220. ICRAF 1996 1995 annual report. Nairobi, Kenya : International Centre for Research in Agroforestry. IFPRI 1996 Feeding the World, preŠenting poŠert’ and protecting the Earth : a 2020 Šision. Washington : International Food Policy Research Institute. Ikerra, T. W. D., Mnekeni P. N. S. & Singh B. R. 1994 Effects of added compost and farmyard manure on P release from Minjingu phosphate rock and its uptake by maize. NorWegian J. Agric. Sci. 8, 13–23. Iyamuremye, F. & Dick, R. P. 1996 Organic amendments and phosphorus sorption by soils. AdŠ. Agron. 56, 139–185. Jaenicke, H., Franzel S. & Boland D. J. 1996 Towards a method to set priorities among species for tree improvement research : a case study from West Africa. J. Trop. Forest Sci. 7, 490–506. Jonsson, K., Sta/ hl, L. & Ho$ gberg, P. 1996 Tree fallows : a comparison between five tropical tree species. Biol. Fert. Soils 23, 50–56. Juo, A. S. R., Franzluebbers, K., Dabiri, A. & Ikhile, B. 1995 Changes in soil properties during long-term fallows and continuous cultivation after forest clearing in Nigeria. Agric. Ecos’st. EnŠiron. 56, 9–18. Kang, B. T., Akinnifesi, F. K. & Ladipo, D. O. 1994 Performance of selected woody agroforestry species grown on Alfisol and Ultisol in the humid lowland of West Africa, and their effects on soil properties. J. Trop. Forest Sci. 7, 303–312. Kang, B. T., Salako, F. K., Chianu, J. N., Akobundu, I. O. & Pleysier, J. L. 1997 Role of some perennial leguminous and natural fallow species in the amelioration of a degraded Oxic Paleustalf—effect on soil properties and crop performance. Land Use Mngt. (In the press). Kiepe, P. & Rao, M. R. 1994 Management of agroforestry for the conservation and utilization of land and water resources. Outlook Agric. 23 (1), 17–25. Kinjo, T. & Pratt, P. F. 1971 Nitrate adsorption. Soil Sci. Soc. Am. Proc. 35, 722–732. Kwesiga, F. & Coe, R. 1994 The effect of short rotation Sesbania sesban-planted fallows on maize yields. Forest Ecol. Mngt 64, 199–208. Lajtha, K. & Harrison, A. F. 1995 Strategies of phosphorus acquisition and conservation by plant species and communities. In Phosphorus in the global enŠironment (ed. H. Tiessen), pp. 139–148. Chichester, UK : John Wiley & Sons. Lal, R. 1989 a Agroforestry systems and soil surface management of a tropical Alfisol. II. Water runoff, soil erosion, and nutrient loss. Agroforestr’ S’stems 8, 97–238. Lal, R. 1989 b Agroforestry systems and soil surface management of a tropical Alfisol. IV. Effect on soil physical and mechanical properties. Agroforestr’ S’stems 8, 197–215. Lal, R. 1989 c Agroforestry systems and soil surface management on a tropical Alfisol. V. Water infiltrability, Trees, soils, and food securit’ P. A. Sanchez and others 959 transmissivity and soil water sorptivity. Agroforestr’ S’stems 8, 217–238. Lavelle, P. & Pashanasi, B. 1989 Soil macrofauna and land management in Peruvian Amazonia. Pedobiologia 33, 283–291. Leakey, R. R. B. 1996 Definition of agroforestry revisited. Agroforestr’ Toda’ 8(1), 5–7. Leakey, R. R. B. & Newton, A. C. 1994 a Domestication of ‘ Cinderella ’ species as a start of a woody plant revolution. In Tropical trees : potential for domestication and the rebuilding of forest resources (ed. R. R. B. Leakey & A. C. Newton), pp. 3–6. London : Her Majesty’s Stationery Office. Leakey, R. R. B. & Newton, A. C. (eds) 1994 b Domestication of timber and non-timber forest products. MAB Digest 17. Paris : UNESCO. Leakey, R. R. B. & Jaenicke, H. 1995 The domestication of indigenous fruit trees : opportunities and challenges for agroforestry. In Proceedings of the 4th international BIOREFOR Workshop (ed. K. Suzuki, S. Sakurai, K. Ishii & M. Norisada), pp. 15–26. Tokyo : BIO-REFOR. Leakey, R. R. B. & Izac, A.-M. N. 1996 Linkages between domestication and commercialization of non-timber forest products : implications for agroforestry. In Domestication and commerciali“ation of non-timber forest products for agroforestr’. Non-Wood forest products 9 (ed. R. R. B. Leakey, A. B. Temu & M. Melnyk), pp. 1–8. Rome : Food and Agriculture Organization of the United Nations. Leakey, R. R. B. & Simons A. J. 1997 The domestication and commercialization of indigenous trees in agroforestry for the alleviation of poverty. Agroforestr’ S’stems (In the press.) Leakey, R. R. B., Temu, A. B. & Melnyk, M. (ed.) 1996 Domestication and commerciali“ation of non-timber forest products for agroforestr’. Non-Wood forest products 9. Rome : Food and Agriculture Organization of the United Nations. Lompo, F. 1993 Contribution a[ la Šalorisation des phosphates natureles du Burkina Faso : etudes des effets de l’interaction phosphates natureles–materies organiques. These Docteur Ingenieur, Faculte des Sciences et Techniques de L’Universite! Nationale de Cote d’Ivoire, Abidjan. Maroko, J., Buresh, R. J. & Smithson, P. C. 1997 Soil phosphorus pools in unfertilized fallow–maize systems and relationships to maize yield. Soil Sci. Soc. Am. J. (Submitted.) Mekonnen, K., Buresh, R. J. & Jama, B. 1997 Root and inorganic nitrogen distributions in Sesbania fallow, natural fallow and maize. Plant and Soil. (In the press). Michon, G. & de Foresta, H. 1996 Agroforests as an alternative to pure plantations for the domestication and commercialization of NTFPs. In Domestication and commerciali“ation of non-timber forest products for agroforestr’. NonWood forest products 9 (ed. R. R. B. Leakey, A. B. Temu & M. Melnyk), pp. 160–175. Rome : Food and Agriculture Organization of the United Nations. Murdiyarso, D., Hariah, K., Husin, Y. A. & Wasrin U. R. 1996 Greenhouse gas emisions and carbon balance in slash and burn practices. In AlternatiŠes to slash-and-burn in Indonesia (ed. M. van Noordwijk, T. P. Tomich, D. P. Garrity, D. P. & A. M. Fagi), pp. 15–38. Bogor, Indonesia : AARD. Newton, A. C., Moss, R. & Leakey, R. R. B. 1994 The hidden harvest of tropical forests : domestication of nontimber products. Ecodecision 13, 48–52. Niang, A., Amadalo, B. & Gathumbi, S. 1996 Green manure from the roadside. Miti Ni Maendeleo 2, 10. Kisumu, Kenya : Maseno Agroforestry Research Centre. Nyathi, P. & Campbell, B. M. 1993 The acquisition and use of miombo litter by small-scale farmers in Masvingo, Zimbabwe. Agroforestr’ S’stems 22, 43–48. Phil. Trans. R. Soc. Lond. B (1997) Ong, C. K. & Huxley, P. A. (ed.) 1996 Tree–crop interactions, a ph’siological approach. Wallingford, UK : CAB International. Otani, T., Ae, N. & Tanaka, H. 1996 Phosphorus (P) uptake mechanisms of crops grown in soils with low P status. II. Significance of organic acids in root exudates of pigeonpea. Soil Sci. Plant Nutr. 42, 553–560. Pagiola, S. 1994 Soil conservation in a semi-arid region of Kenya : rates of return and adoption by farmers. In Adopting conserŠation on the farm (ed. T. L. Napier, S. M. Camboni & S. A. El-Swaify), pp. 171–187. Alkeny, Iowa : Soil and Water Conservation Society. Palm, C. A. & Sanchez P. A. 1991 Nitrogen release from the leaves of some tropical legumes as affected by their lignin and polyphenolic contents. Soil Biol. Biochem. 23, 83–88. Palm, C. A. 1995 Contribution of agroforestry trees to nutrient requirements of intercropped plants. Agroforestr’ S’stems 30, 105–124. Palm, C. A., Myers, R. J. K. & Nandwa, S. 1997 Organic– inorganic nutrient interactions in soil fertility replenishment. In Replenishing soil fertilit’ in Africa. ASA-SSSA Special Publication. (In the press.) Patel, S. H., Pinckney, T. C. & Jaeger, W. K. 1995 Smallholder wood production and population pressure in East Africa : evidence of an environmental kuznets curve ? Land Econ. 71, 516–530. Place, F. 1995 The role of land and tree tenure on the adoption of agroforestr’ technologies in Zambia, Burundi, Uganda and MalaWi : a summar’ and s’nthesis. University of Wisconsin, Madison : Land Tenure Center. Place, F. (ed.) 1996 ToWards improŠed polic’ making for natural resources and ecos’stem management in sub-Saharan Africa. Nairobi, Kenya : ICRAF. Place, F. & Otsuka, K. 1997 Population densit’, land tenure and resource management in Uganda. Manuscript. Nairobi, Kenya : ICRAF and IFPRI. Rao, M. R., Nair, P. K. R. & Ong, C. K. 1997 Biophysical interactions in tropical agroforestry. Agroforestr’ S’stems. (In the press.) Sanchez, P. A. 1994 Alternatives to slash-and-burn : a pragmatic approach for mitigating tropical deforestation. In Agricultural technolog’, polic’ issues for the international communit’ (ed. J. R . Anderson), pp. 451–480. Wallingford, UK : CAB International. Sanchez, P. A. 1995 Science in agroforestry. Agroforestr’ S’stems 30, 5–55. Sanchez, P. A., Palm, C. A., Davey, C. B., Szott, L. T. & Russell, C. E. 1985 Trees as soil improvers in the humid tropics ? In Trees as crop plants (ed. M. G. R. Cannell & J. E. Jackson), pp. 327–358. Huntingdon, UK : Institute of Terrestrial Ecology. Sanchez, P. A. & Miller, R. H. 1986 Organic matter and soil fertility management in acid soils of the tropics. In Transactions of the 13th International Congress on Soil Science, Hamburg 6, 609–625. Sanchez, P. A., Woomer, P. L. & Palm, C. A. 1994 Agroforestry approaches for rehabilitating degraded lands after tropical deforestation. In Rehabilitation of degraded forest lands in the tropics–technical approach. JIRCAS International S’mposium Series 1, pp. 108–119. Tsukuba, Japan : JIRCAS. Sanchez P. A., Izac, A.-M., Valencia, I. M. & Pieri, C. 1996 Soil fertility replenishment in Africa. In AchieŠing greater impact from research inŠestments in Africa (ed. S. A. Breth), pp. 200–208. Mexico City : Sasakawa Africa Association. Sanchez, P. A. & Palm, C. A. 1996 Nutrient cycling and agroforestry in Africa. Unas’lŠa 185(47), 24–28. Sanchez, P. A. & Leakey, R. R. B. 1997 Land-use trans- 960 P. A. Sanchez and others Trees, soils, and food securit’ formation in Africa : Three determinants for balancing food security with natural resource conservation. Fourth Congress, European Societ’ of Agronom’. (In the press.) Sanchez, P. A., Izac A.-M. N., Buresh, R. J. et al. 1997 Soil fertility replenishment in Africa as an investment in natural resource capital. In Replenishing soil fertilit’ in Africa. ASA-SSSA Special Publication. (In the press.) Sanginga, N., Bowen, G. D. & Danso, S. K. A. 1990 Assessment of genetic variability for N fixation between # and within provenances of Leucaena leucocephala and Acacia albida estimated by "&N labelling techniques. Plant and Soil 127, 169–178. Sanginga, N., Danso, S. K. A., Zapata, F. & Bowen, G. D. 1994 Field validation of intraspecific variation in phosphorus use efficiency and nitrogen fixation by provenances of Gliricidia sepium grown in low P soils. Appl. Soil Ecol. 1, 127–138. Sanginga, N., Manrique, K. & Hardarson, G. 1991 Variation in nodulation and N fixation by the Gliricidia # sepium}Rhi“obium spp. symbiosis in a calcareous soil. Biol. Fert. Soils 11, 273–278. Sanginga, N., Vanlauwe, B. & Danso, S. K. A. 1995 Management of biological N fixation in alley cropping # systems : Estimation and contribution to N balance. Plant and Soil 174, 119–141. Scherr, S. J. & Hazell, P. A. 1994 Sustainable agricultural development strategies in fragile lands. EnŠironment and production technolog’ diŠision discussion paper 1. Washington, DC : IFPRI. Schroeder, P. 1993 Agroforestry systems : integrated land use to store and conserve carbon. Climate Res. 3, 53–60. Schroeder, P. 1994 Carbon storage benefits of agroforestry systems. Agroforestr’ S’stems 27, 89–97. Schroth, G., Kolbe, D., Pity, B. & Zech, W. 1995 Searching for criteria for the selection of efficient tree species for fallow improvement, with special reference to carbon and nitrogen. Fertili“er Res. 42, 297–314. Siaw, D. E. K. A., Kang, B. T. & Okali , D. U. U. 1991 Alley cropping with Leucaena leucocephala (Lam.) De Wit and Acioa barteri (Hook. f.) Engl. Agroforestr’ S’stems 14, 219–231. Sibuea, T. Th. 1995 Short notes on the Sumatran rhino (Dicerorhinus sumatrensis) in the agroforest areas (damar garden) in Krui, Lampung. Bogor, Indonesia : AWB. Sibuea, T. Th. & Herdimansyah, D. 1993 The variety of mammal species in the agroforest areas of Krui (Lampung), Muara Bungo (Jambi) and Maninjau (West Sumatra). Internal report OSTROM}Himbio. 62 pp. Bandung, Indonesia : Universitas Padjajaran. Simons, A. J. 1996 ICRAF’s strategy for domestication of non-wood tree products. In Domestication and commerciali“ation of non-timber forest products for agroforestr’. Non-Wood forest products 9 (ed. R. R. B. Leakey, A. B. Temu & M. Melnyk). Rome : Food and Agriculture Organization of the United Nations. Smaling, E. 1993 An agroecological framework for integrated nutrient management with special reference to Kenya. Ph.D. thesis, Agricultural University, Wageningen, The Netherlands. Thiollay, J. M. 1995 The role of traditional agroforests in the conservation of rainforest bird diversity in Sumatra. ConserŠation Biol. 9, 335–353. Tian, G., Brussard, L. & Kang, B. T. 1995 An index for assessing the quality of plant residues and evaluating their effects on soil and crop use in the subhumid tropics. Appl. Soil Ecol. 2, 25–32. Tiffen, M., Mortimer, M. & Gichuki, F. 1994 More people, less erosion. EnŠironmental recoŠer’ in Ken’a. Chichester, UK : John Wiley & Sons. Phil. Trans. R. Soc. Lond. B (1997) Tisdall, J. M. & Oades J. M. 1982 Organic matter and water-stable aggregates in soils. J. Soil Sci. 33, 141–163. Unruh, J. D., Houghton, R. A. & Lefebvre, P. A. 1993 Carbon storage in agroforestry : an estimate for subSaharan Africa. Climate Res. 3, 39–52. van Noordwijk, M. 1997 Nutrient cycling in ecosystems versus nutrient budgets in agricultural systems. In Nutrient c’cles and nutrient budgets in global agro-ecos’stems (ed. E. Smaling, O. Oenema & L. Fresco) Wallingford, UK : CAB International (In the press.) van Noordwijk, M., Lawson, G., Soumare, A., Groot, J. J. R. & Hairiah, K. 1996 Root distribution of trees and crops : competition and}or complementarity. In Tree–crop interactions : a ph’siological approach (ed. C. K. Ong & P. A. Huxley), pp. 319–364. Wallingford, UK : CAB International. Woomer, P. L., Palm, C. A., ”ureshi J. N. & Kotto-Same, J. 1997 Carbon sequestration and organic resource management in African smallholder agriculture. AdŠ. Soil Sci. (In the press.) Young, A. 1989 Agroforestr’ for soil conserŠation. Wallingford, UK : CAB International. Discussion L. T. E (CSIRO DiŠision of Plant Industr’, Australia). I agree with your emphasis on the crucial need to raise the available soil phosphorus level in Africa, especially given their high capacity. But I missed one key word in your presentation, namely tenure of land. Small farmers can hardly be expected to invest in fertilizers unless they have security of tenure to guarantee the return on their investment, and small farmers in Africa often lack that essential ingredient. P. A. S. Land tenure, according to ICRAF research in several African countries (Kenya, Uganda, Burundi, Zambia, Malawi), is less of a constraint in smallholder farms than previously thought, because most farmers have either formal land tenure or customary rights, which are well respected. The key tenure-related issues are women’s access to land and trees even though the husband owns them, and land fragmentation. Therefore, effective tenure in many smallholders in Eastern and Southern Africa paves the way for a fertility replenishment strategy. G. D. A. Is there a function for lime in releasing P from organic form in the high altitude forested or former forested soils in Africa ? P. A. S. Aluminium toxicity is not an extensive soil constraint in subhumid and semi-arid Africa, where most of the soil fertility depletion takes place ; therefore, there is seldom a need to apply lime to correct soil acidity. There will be little advantage in applying lime to increase P availability. P. W (CommonWealth Forestr’ Association, Oxford, UK). (1) How far do you see farmers’ decisions based on financial profit (for cash crops, for example) likely to favour practices that do not improve the soil ? (2) Farm sizes in sub-Saharan Africa are small and decreasing. How far to you think privatization, development of ‘ free ’ markets and structural adjustment may lead to increases in farm size, as has happened in the capitalist West ? P. A. S. (1) Farmers everywhere base most of their decisions on economic considerations. African farmers are no exception. (2) I think farm size will not increase with improvements in privatization and structural adjustments Trees, soils, and food securit’ P. A. Sanchez and others 961 programmes. Smallholder areas in Africa are likely to follow the pattern in Asia, where the impact of the Green Revolution did not appreciably change farm size, but it sure alleviated poverty and improved the standard of living of smallholder farmers. There are many social considerations that will likely maintain farm size pretty much as is. P. V (Institute of Agriculture in the Tropics, UniŠersit’ of Goettingen, German’). When you advocate recapitalization of soils with P in Africa, do you believe the infrastructure} marketing system can be put in place in time to avoid a calamity ? P. A. S. Good question. Fertility replenishment must be accompanied by improved road and marketing infrastructure. The answer depends on the degree of commitment governments have to liberalizing their markets and providing an enabling policy environment to smallholder farmers who produce most of the food in the countries. In some countries, it is likely to happen, but not so in others. The point is that soil fertility replenishment is a necessary, but it is not sufficient condition for Africa’s food security. D. S. P (Institute of Arable Crops Research–Rothamsted, Harpenden, UK). What is the mechanism for the increased availability of P to crops relative to the use of organic inputs ? Is it that more inorganic P is made soluble, or are organic forms of P that are less readily fixed put into solution ? P. A. S. The mechanisms involved are (i) the mineralization of organically bound P in the organic inputs ; (ii) the transformation of less available pools of inorganic P into more readily available organic P that is mineralized, when plants convert inorganic P in their tissues, and those Phil. Trans. R. Soc. Lond. B (1997) are cycled back to the soil ; and (iii) organic C radicals can block P-sorption sites. The relative importance of these three processes has not been quantitatively determined in inorganic–organic nutrient interaction studies. E. B. B (UniŠersit’ of York, UK). (1) Comment on the role of land and tree tenure on the farmers’ decisions and (2) the role of risk in influencing farmers’ decisions to invest in agroforestry systems. P. A. S. (1) Land and tree tenure are critical prerequisites for soil fertility replenishment. (See my comments in response to Lloyd Evans’s question, which are relevant to this one.) (2) Risk is also a key determinant ; therefore, agroforestry decisions must entail relatively low risk. The time lag until an agroforestry intervention begins to produce results is perhaps a more important consideration ; policies must address this issue of short-term gains versus a delayed return. Some of our farmer surveys, however, indicate that African farmers are very aware of the time lag and are willing to wait. In other cultures this is not the case. M. W (UniŠersit’ of Reading, UK). Could I offer some of our own evidence in support of Dr Sanchez’s strategy for soil fertility replenishment in Africa. Our detailed "&N studies on the recovery of nitrogen fertilizer in maize in central Kenya have shown that the recovery of fertilizer in the crop is very low (20 %), and indicate that a significant proportion of the fertilizer is leached below the rooting zone. This may contribute to the pool of nitrate at depth, which, as described by Dr Sanchez and others, may be captured by deeper rooting plants such as trees. P. A. S. Thank you. Could you send me the data ?