Soil fertility management research for the smallholder maize-based cropping systems of southern Africa

Material Information

Soil fertility management research for the smallholder maize-based cropping systems of southern Africa a review
Series Title:
Network research working paper
Kumwenda, John David Tabangoni, 1952-
International Maize and Wheat Improvement Center -- Maize Improvement Program
Soil Fertility Network for Maize-Based Cropping Systems in Countries of Southern Africa
Place of Publication:
Harare Zimbabwe
CIMMYT Maize Programme
Publication Date:
Physical Description:
34 p. : ill. ; 30 cm.


Subjects / Keywords:
Soil fertility -- Research -- Africa, Southern ( lcsh )
Corn -- Soils -- Research -- Africa, Southern ( lcsh )
Cropping systems -- Research -- Africa, Southern ( lcsh )
bibliography ( marcgt )
non-fiction ( marcgt )
Spatial Coverage:


Includes bibliographical references (p. 27-34).
General Note:
Cover title.
General Note:
"December 1995."
General Note:
At head of title: Soil Fertility Network for Maize-Based Cropping Systems in Countries of Southern Africa.
Statement of Responsibility:
by John D.T. Kumwenda ... et al..

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright CIMMYT International Maize and Wheat Improvement Center. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
43466321 ( OCLC )
97980344 ( LCCN )

Full Text
22. /7
Soil Fertility Network for Maize-Based Cropping Systems in Countries of Southern Africa
Soil Fertility Management Research for the Smallholder
Maize-Based Cropping Systems of Southern Africa: A Review
John D.T. Kumwenda, Stephen R. Waddington,
Sieglinde S. Snapp, Richard B. Jones and Malcolm J. Blackie
CM CIMMYT CIMMYT Maize Programme

John D. T. Kumwenda, Senior Agronomist, Maize Commodity Team, Ministry of Agriculture and Livestock Development, Malawi
Stephen R. Waddington, Agronomist, CIMMYT Maize Program, Zimbabwe
Sieglinde S. Snapp, Postdoctoral Fellow, The Rockefeller Foundation, Southern Africa Agricultural Sciences Program, Malawi
Richard B. Jones, Agroforestry Commodity Team/Washington State University, Ministry ofAgriculture and Livestock Development, Malawi, and
Malcolm J. Blackie, Senior Scientist, The Rockefeller Foundation, Southern Africa Agricultural Sciences Program, Malawi
Executive Summary
Tze dominant smallholder cropping systems of southern Africa are based on maize. Increasing human population density and declining land availability have made shifting cultivation obsolete. Maize is now grown in continuous cropping rather than as part of a fallow which traditionally restored soil fertility and reduced the build-up of pests and diseases. The soil resource is now being degraded with a consequent reduction in crop yield.
There have been notable successes in the adoption of high yielding maize by smallholders in Africa. Although improved germplasm is now grown on 33-50% of Africa's maize area, national per-hectare increases in maize productivity are disappointing. Losses of mineral nutrients from soil generally exceed nutrient inputs. The productivity-with-sustainability challenge is so large that farmers will need to combine gains from improved germplasm with complementary improvements in their management of soil fertility. This now requires a shift of research and extension emphasis onto the complex issues involved with the build-up and maintenance of soil fertility under the income and other constraints faced by smallholders.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems

Inorganic fertilizers are expensive and their use, frequently recommended as blanket applications even in semi-arid areas, is not very profitable. For farmers that can afford fertilizer, there is an urgent need to increase the profitability of this input by better targeted recommendations and improved fertilizer management techniques that are appropriate for smallholders. The efficiency of fertilizer use is often low because of the declining level of organic matter in tropical soils. The proportion of locally produced organic materials must be increased to maintain soil organic matter and halt the downward spiral of soil fertility. Improving the efficiency of inorganic fertilizer use will consolidate and expand the base of fertilizer users.
For many households the cash requirement needed to buy inorganic fertilizer far exceeds their total annual cash income. The lack of cash dominates decision making at the household level and is central to the development of adoptable technologies. Many households are often forced to sell their labor in return for food or cash which, in turn, compromises their own agricultural efforts. For this expanding group more emphasis on organic sources of nutrients, especially legumes, that capitalize on the freely available nitrogen in the atmosphere, is the best strategy for increased soil fertility.
Legumes are not new to farming systems. Grain legumes, intercropping, rotations, green manures, improved allows, agroforestry, cereal residues and animal manures are all technologies that can enhance soil fertility and sustain the resource base. However, in broad terms, the larger the likely soilfertility benefit from a legume technology, the larger the initial investment required in labor and land, and the fewer short tern food benefits it has. The potential of such technologies is rarely realized on farmers' fields. Combinations of low rates of several inputs show promise, especially those that combine inorganic with organic fertilizer. But such combinations still involve a cash cost. Innovative mechanisms are needed to help farmers begin to access such inputs. One promising approach involves providing farmers with start-up grants of cash paid into savings schemes, from which farmers can access loans.
Process-based research provides the foundation for extrapolating from site-specific trials to agronomic recommendations for specific agro-ecological zones and farmer groups. Past crop husbandry research is often neglected because the results are distilled into a few recommendations that ignore the important interactions in the system and fail to address the widespread diversity that exists among smallholders. Institutional memory needs to be maintained and expanded to a wider group, in part through computer databases and effective networks. Emphasis in both research and extension needs to move away from the rigid and prescriptive to a flexible problem-solving format leading to conditional recommendations. This will facilitate the evolution of a demand-driven technology development process by smallholders. Without concerted action to develop such a process through better interaction, the consequence will be a degrading natural resource base and a continuing decline in the standard of living of rural communities reliant on agriculture in southern Africa.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 2

1. Introduction
In this paper we contend that varietal maize (Zea mays L.) improvement will have a transitory impact on smallholder farming in Africa unless widespread declines in soil fertility are addressed by research and farmers. There is a relative scarcity of adoptable technology to sustain soil fertility. But without such technology, the productivity of smallholder maize-based farming systems in Africa will fail to improve. We use examples from southern Africa, particularly Malawi and Zimbabwe. Those countries have been chosen because of our experience there, and also because they contrast an agriculture in crisis, increasingly unable to meet national food needs (Malawi), with a country widely regarded to have one of the most successful agricultural bases in sub-Saharan Africa (Zimbabwe). These experiences are applicable more widely in sub-Saharan Afica. We propose a soil fertility research and extension model that combines organic with inorganic sources of soil fertility and actively involves farmers and other clients in an integrated, long-term process.
2. Soil Fertility Decline Against A Background Of Widespread Adoption Of Improved Maize
2.1 The Dominance Of Maize And Characteristics Of Smallholder Maize Systems
Today, the dominant smallholder cropping systems of southern (and eastern) Africa are based on maize. Maize is grown from sea level (the coastal zones of Mozambique and Tanzania) to elevations above 2400 m. Overlaid on altitude is rainfall which varies from less than 400 mm (southern Zimbabwe, southern Mozambique) to well watered (higher elevation equatorial zones). Maize accounts for 60 % or more of the cropped area in Malawi, Zimbabwe and Zambia and is almost as dominant in countries such as Mozambique, Tanzania and Kenya. Overall, maize is the staple food of the region, accounting for about 50 % of the calories consumed. In Malawi, with an average population density reaching 215 persons per square kilometer in the southern region, maize is the main crop in nearly 90 % of the area and contributes 80 % of daily food calories. Carr (199$1) attributes the increase in popularity of maize in Malawi to its efficiency as a per-hectare calorie producer compared to the other available food plants. As land availability declines, so the efficiency of calorie production per hectare becomes of greater importance to the farmer. Thus the food security of resource-poor households is critically dependent on the productivity and sustainability of maize-based cropping systems.
Most farm households in southern Africa use family resources, especially labor, to produce maize on 0.5 3.0 ha of land held under traditional tenure arrangements. In parts of the region, particularly southern Zambia, Botswana and Zimbabwe, animal (mainly ox) traction is used to prepare the land and help with weeding. In Malawi, northern and eastern Zambia, northern Mozambique, and many parts of Tanzania, human labor using hand-held hoes is the predominant power source (Low and Waddington, 1990; Waddington, 1994). Maize is grown in loose rotations with such crops as groundnuts and sunflowers. Increasingly it is grown on the same land year after year, often sparsely intercropped with beans, groundnuts, cowpeas or pumpkins. Formal intercropping is common only where high human population density means average land holdings per family are small, e.g., 0.8 ha per family in parts of southern Malawi.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 3

The last 25 years have seen farmers in many parts of Africa, including southern Africa, switch from traditional maize landraces to improved germplasm. Byerlee (1994) calculated that improved varieties and hybrids now cover 33-50 % of the maize area in Africa. But there is little evidence of an increase in maize productivity per hectare (CIMMYT, 1992, 1994; Lele, 1992). Even in countries where there has been a substantial increase in the use of fertilizer by smallholders, productivity increases are disappointing (see Conroy, 1993).
2.2 Soil Fertility Constraints To Maize Production
The most widespread and dominant limitation on yield, and on the sustainability of the maizebased systems of southern and eastern Africa, is the decline in soil fertility. Although the increase in maize area in drought prone semi-arid areas has contributed to reduced average yields per unit of land (e.g., Gilbert et al., 1993), the greater influence is a decline in soil fertility in the wetter, higher yield potential areas. The more densely populated and erosion prone countries in eastern and southern Africa (where maize is a major food crop), tend to be those with the greatest aggregate nutrient loss. Smaling (1993) and Stoorvogel et al. (1993) estimated annual net nutrient depletion exceeding 30 kg nitrogen (N) and 20 kg potassium (K) ha1 of in Ethiopia, Kenya, Malawi, Nigeria, Rwanda and Zimbabwe.
The old and already highly leached soils of humid and sub-humid zones in Africa have inherently low nutrient levels. Sandy and sandy loam soils derived from granite, with organic matter <0.5% and very low cation exchange capacities are widespread in Zimbabwe, southern Zambia and western and southern Mozambique. Deficiency of N is ubiquitous on these soil types, while deficiencies of phosphorus (P), sulfur (S), magnesium (Mg) and zinc (Zn) are common (Grant, 1981). On the common sandy loam and clay loam soils in Malawi, in addition to chronic deficiencies of macronutrients, micro nutrients such as S, Zn and boron (B) are reported as limiting at many sites (Wendt el al., 1994). In Zambia and Mozambique there are large areas of acidic soils with free aluminum (Al3+).
Traditional African agricultural systems were largely based around extended fallows and the harvesting of nutrients stored in woody plants. The site to be cultivated was cleared by cutting and slashing plant growth, ant then by burning the dried plant material (e.g., Araki, 1993; Blackie and Jones, 1993; Blackie, 1994). Extensive exploitation of land by few families characterized these systems. For example, in the chitemene slash-and-bum system in the miombo woodland of northern Zambia, land was fallowed for 50-70 years, followed by cropping in the center of the clearing for a few years (Araki, 1993). Today, in most arable areas of Malawi, Zimbabwe and Kenya, fallowing has almost disappeared from the now sedentary agricultural system. The length of the fallow period continues to decline significantly in Zambia, Mozambique and Tanzania. Continuous cropping of maize, with a consequent downward spiral of soil fertility, has led to a decline in crop yields and soil erosion (Araki, 1993). In some areas, such as the wetter communal lands of northern Zimbabwe, soil depletion is so severe that commonly maize will give near zero grain yield without fertilizer.
Buddenhagen (1992) estimates that, discounting erosion effects, the weathering of minerals and biological nitrogen fixation will enable, at most, 1000 kg grain ha-1 year"! on a sustainable basis in the tropics. It will be less in hot lowland areas, such as Ghana, where this equilibrium was estimated at 600-800 kg of maize grain ha1 year-' (Greg Edmeades, pers. comm.). Loss of soil through erosion (and soils cultivated with annual crops in the upland tropics are very prone to
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 4

erosion) will reduce this level considerably. With farmers locked into low crop productivity per unit of land from a degrading natural resource base, there is increased pressure to open new land for agriculture. Farming will increasingly encroach on ecologically fragile environments such as the Zambezi and Luangwa. valleys (with their unique ecosystems, and remnant habitats for several endangered large mammals) and on the many hilly areas where loss of protective vegetation will quickly lead to severe soil erosion.
Two approaches can be recognized in managing soil fertility (Sanchez, 1995). The best known is that of overcoming soil constraints to fit plant requirements through the application of purchased inputs. The bulk of food produced in the world is based on this approach and its application in the.Green Revolution is responsible for much of the increased yield especially in Latin America and Asia (Pinstrup-Andersen, 1993). The second approach relies more on biological processes to optimize nutrient cycling, minimize external inputs and maximize the efficiency of their use. Knowledge and understanding of the principles underlying the second more complex approach are not well developed, and understanding from temperate areas may be inappropriate for smallholder agriculture in the tropics. These two approaches, and a more sustainable but practicable middle way that combines the best features of both, will be discussed in detail.
2.2.1 Low Use Of Inorganic Fertilizer
In most parts of the world, chemical fertilizers play a major role in maintaining or increasing soil fertility. However, in sub-Saharan Africa chemical fertilizer use is very low. FAO (1988) gave seven kg of fertilizer nutrients ha- of arable land plus permanent crops year- I as the average for sub-Saharan Africa. More recent calculations (for 1993) by Heisey and Mwangi (1995), using similar criteria, give an average of ten kg of fertilizer nutrients ha year- Use is higher in some countries of southern Africa (notably Zimbabwe, Zambia and Malawi) where the commercial fanning sector is relatively well developed and fertilizer-responsive maize an important crop. For maize, Heisey and Mwangi (1995) calculated application rates of 70 kg fertilizer nutrients ha&' of maize crop year-1 in Zambia, 55 kg in Zimbabwe and 26 kg in Malawi. Nevertheless, this is well below crop and soil maintenance requirements and is likely to remain so since fertilizer is prol5ably the most costly cash input used by the typical smallholder in southern Africa.
Fertilizer recommendations are frequently unattractive to smallholders. They often ignore soil and climatic variation found in smallholder farming areas, are incompatible with farmer resources, or are inefficient. This leads to farmers incurring unnecessary expense. For example, the Zimbabwe recommendation for basal fertilizer on maize is to apply it in the planting hole at planting. Yet farmers almost always apply it just after crop emergence because that is easier and less risky, with negligible loss of yield under farm conditions (Shumba, 1989). Their practice allows them, on a given rain, to plant more area more quickly (important on a drying sandy soil), get better crop emergence and have more labor available for other operations. There is rarely a response to K on the predominant granite soils in smallholder areas of Zimbabwe (Mashiringwani, 1983; Hikwa and Mukurumbira, 1995), yet the compound fertilizer recommended contains K, as well as N and P. There is some potential for applying cheaper straight N just after planting and cheaper forms of P at other times.
Soil Fertility Network Results Working Paper 1: Sail Fertility Research for Maize-Based Systems 5

The efficiency of fertilizer use on-fann, as measured by the grain yield response to the addition of chemical N and P fertilizers, is often poor (e.g., Jones and Wendt, 1995). Those authors, summarizing on-farm data from Malawi and Zambia, calculated that farmers using current fertilizer practices can expect just 9.5-16 kg of maize grain per kg of nutrient applied when growing local unimproved maize and 17-19 kg of maize grain per kg of nutrient applied when using hybrids. Research on farmers' fields in Malawi shows that, at farmers' levels of fertilizer application, with improved timing and application methods the maize response to nitrogen can be increased from 15 to 20 kg grain per kg N applied for unimproved maize and from 17.4 to 25 for hybrid. However, the poor profitability of inorganic fertilizer reduces the attractiveness of this crucial input to smallholder maize production. In Malawi, value:cost ratios of 1.8 for fertilizer use on hybrid maize and 1.3 for unimproved maize have been reported for moving from zero fertilizer applied to the current recommendation (Conroy and Kumwenda, 1995). An economic analysis of fertilizer policy in Malawi (see HIID, 1994) concluded that these improvements in fertilizer use efficiency could substantially outweigh feasible price changes in either fertilizer or maize in making fertilizer financially attractive to smallholders.
2.2.2 Declines In Soil Organic Matter
Soil organic matter (SOM) maintenance and management is central to the sustainability of soil fertility on smallholder farms in the tropics (see Swift and Woomer, 1993; Woomer et~ al, 1994). In low input agricultural systems in the tropics, SOM helps retain mineral nutrients (N, 5, micronutrients) in the soil and make them available to plants in small amounts over many years as SOM is mineralized. SOM increases soil flora and fauna (associated with soil aggregation, improved infiltration of water and reduced soil erosion), complexes toxic Al and manganese (Mn) ions (leading to better rooting), increases the buffering capacity on lowactivity clay soils and increases water holding capacity (see Woomer et aL, 1994). Current SOM inputs are insufficient to maintain SOM levels in tropical agricultural soils. Continuous cropping of land with its associated tillage are well known to lead to initial rapid declines in SOM which then stabilizes at a low level (e.g., see Woomer et al., 1994).
The conventional mechanisms for addressing loss of SOM in tropical, rainfed, low input systems are fallowing, rotation; (especially involving legumes), the addition of animal manures, forms of intercropping (including with hedgerow legumes) and reduced tillage. Some of these techniques are used in temperate agriculture but there is more reliance on purchased inputs in those places. As pressure for arable land rises in tropical areas, cropping encroaches into areas previously used for grazing. Agricultural intensification in southern Africa is often associated with a decline in the availability of animal manures as livestock are squeezed out. This is more of a problem in the unimodal rainfall areas of southern Africa where the long dry season makes zero grazing techniques difficult or impossible for smallholders than in the bimodal rainfall areas of eastern Africa. Manure from cattle and other animals is very important for most farmers in Zimbabwe, less so in Zambia, but is rarely available in Malawi (where animals are scarce). But even in the best areas, its supply (and, as importantly, its quality) is inadequate to maintain soil fertility on its own.
Where animals are scarce, farmers have turned to other sources of SOM. Leaf litter from trees can make significant contributions in areas close to woodlands but deforestation associated with the demand for arable land and for building and firewood work against this option as population rises. Composted crop residues are used in wetter areas and where crop biomass
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 6

production is relatively high but composts are rarely sufficient for more than a modest part of the cultivated area, and, like manures, quality is often poor. These technologies require a substantial labor commitment on the part of farmers (e.g., in Zimbabwe, Huchu and Sithole, 1994; Carter, 1993). The reality is that there is rarely sufficient organic matter to maintain SOM, and in the low rainfall marginal areas it is impossible to grow enough biomass to maintain SOM. In addition, organic manures alone will only rarely provide the productivity boost needed by smallholders. They will need to be combined with the judicious use of chemical fertilizers, improved pest and weed management techniques, and high yielding crop varieties.
The success of the germplasm led Green Revolution in Asia, under very different conditions and with more fertile and uniform soils, has biased the research agenda in Africa away from crop nutrition studies towards emphasis on gains from plant breeding. Improved maizes make better use of available nutrients, but in the absence of added nutrients, the gains from genetic improvement alone are small. Under low N conditions maize scavenges the N that is available in the soil very effectively and there is little left to take up by the time flowering is reached (Greg Edmeades, pers. commn.). As an example, the results from an extensive program of onfarm demonstrations conducted over four seasons in the major maize growing areas of Malawi showed that on relatively fertile soils and under good management, hybrids without fertilizer yielded just 1.4 t ha&' compared with 0.9 t ha-1 for unimproved maize (Jones and Wendt, 1995; Conroy and Kumwenda, 1995; Zambezi et al, 1993). The productivity-with-sustainability challenge is so large that farmers need to combine gains from improved germplasm with complementary improvements in their management of soil fertility.
The central issue is how to build up and maintain soil fertility under the income and, increasingly, land and labor constraints faced by smallholders. We now look at what those technology improvements to raise fertilizer efficiency might entail.
3. Improving Soil Fertility And Productivity What Are The Practicable Technology Options For Farmers?
Concurrent adoption of~ both improved germplasm and better management practices are essential for sustainable and long term productivity gains. The analysis in Figure 1 shows three possible scenarios, based on "best bet" estimates of technology adoption for Malawi. For each scenario, a maize deficit or surplus is calculated based upon the balance between maize production and consumption in a year. Assumptions include a 3.2% annual population growth rate and a per capita annual maize consumption of 230 kg. A constant area of 1.4 million ha is planted to maize (Malawi already has almost all of the arable land available to smallholders under continuous maize cultivation so the opportunities for an expansion of maize area are limited), of which, at the start of each scenario, 20% is planted to improved maize. Maize grain yield data used are national average yields of hybrid and unimproved maize for the period 19821992, including fields that received fertilizer and those that did not (Ministry of Agriculture Crop Estimates, quoted by Conroy, 1993). In all scenarios, unimproved maize yields 1000 kg ha' In scenario one, improved maize yield is held constant at 2500 kg ha- and the area planted to improved maize does not change. This represents a continuation of the status quo where national maize production is already in a widening deficit. In scenario two, the situation is modeled where hybrid adoption increases at 20% (compounded) each year but hybrid yields remain at 2500 kg ha- because there is little change in soil fertility and other management. Here maize production will move into modest surplus by 2002, moving back into deficit just six
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 7

years later. And, finally, scenario three represents the combination of the adoption of improved (hybrid) maize (from scenario two) and achieved efficiency gains from better fertilizer management on-farm of about 44% (as described in section 2.2.1) combined with other likely management improvements, particularly timely weeding (Kabambe and Kumwenda, 1995). We assumed a yield increase of 68% when farmers use this set of improved management practices ,Aritb their hybrid maize. In this scenario large surpluses are achieved to 2015. Likely declines in the rate of human population growth will enable grain surpluses to be maintained for longer than given in that scenario. Thus this analysis is very positive that productivity improvements will make a difference well into the 21st century.
So, while there is little argument about the need for more sustainable soil fertility management, there is less clarity on how it might be achieved. In a major overview of past impacts and future prospects for maize research in sub-Saharan Africa, Byerlee et al. (1994) contrast the high adoption of improved maize by farmers against a relative lack of resource management technology for maintaining soil fertility and increasing labor productivity. Many technologies are potentially available but few are easy for farmers to adopt.
We first look at the circumstances under which many smallholder farmers in Africa must operate, then review existing soil fertility-enhancing technology to suggest how the technology might be made more attractive to farmers. We then proceed to outline, based on examples drawn from practice, improvements in technology development and the transfer process to facilitate the adoption of better soil fertility management practices.
3.1 Smallholder Diversity, Labor, Cash and Soil Fertility Management
The smallholder farming sector is characterized by considerable variability in individual access to land (in quality and quantity), resources and skills. Levels of literacy and numeracy are poor. The population involved in agriculture goes from small-scale commercial farmers, through progressive farmers, to the vast majority of poorer people who endeavor to subsist, with varying levels of success, in hostile ecologies where the sustainability of life, livelihood, and environment is hazardous.
The two major costs faced by smallholders in producing maize are labor and fertilizer. The lack of cash dominates the choices available to the fanner. Labor may be provided by the family, it may be bought in from other farmers, and it may be sold to others for food or cash. Often the household is headed by a woman, with small children. The older children may be at school, or have moved to town. If she is fortunate, her husband and children living away will send cash or kind to help support the rural household. If not, she will be attempting to support herself and her children from what she can grow or sell. She will be living on a piece of land that has been cultivated many times before. What inherent fertility was there has long been extracted from the soil. Weeds, increasing including the parasitic weed Striga, will have established themselves and will compete strongly with whatever she plants for light, water, and nutrients.
Given a hectare or more of land, she may be self-sufficient in food if her health is good and the weather favorable. But the start of the rains bring diarrhea and malaria. Often, illness of herself or her children will result in late planting. With a poor rainy season (and to many Africa and drought are synonymous) her crop may fail. The odds are that in some, often many, years, she will find herself unable to produce enough food for her family's needs. This requires her to
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 8

work for neighboring farmers who will then feed or pay her (and any children that work with her) for the days that she puts in. Typically this work will be planting, weeding, or fertilizing the neighbor's crop which means that her own is left unplanted, unneeded, and unfertilized until later in the season. Late planting and poor weeding mean a poor harvest and once again she finds herself without food before the crop comes in.
An increasing number of rural families in Africa are in this position; unable to produce sufficient food for their own annual needs (Weber et aL, 1988). This is the downward spiral that the next maize revolution has to break. Malawi illustrates an advanced case of this scenario with some 60% of rural households (and 41% of the total population) producing less than they need to feed themselves through the year. Of these households, the cash requirement needed to buy inorganic fertilizer far exceeds their total annual cash income (HIID, 1994). African smallholders, almost wherever they live, face conditions of difficulty and stress for which both tradition and science have few real answers. While there are technically sound solutions to many of the problems faced by most smallholders, all too often these turn out to be financially or managerially unsound.
3.2 Increasing Fertilizer Use And Fertilizer Use Efficiency
3.2.1 Increasing The Number of Smallholders Using Fertilizers Since the 1960s, fertilizer use has been growing in sub-Saharan Africa at around 6.7% annually (see Heisey and Mwangi, 1995), but many farmers do not use fertilizers. That base of fertilizer users needs to be consolidated and expanded. But fertilizer will remain a high cost item for African farmers for the foreseeable future. Expansion will depend on generating better returns from fertilizer through greater efficiencies in its use. Much greater use will need to be made of organic sources of fertility to raise fertility management and reduce cost. We continue, therefore, by considering how to help existing fertilizer users, and those who are likely to be able to afford some fertilizer, to achieve increases in fertilizer use efficiency. A three pronged technology strategy is envisaged:
I careful tailoring. of inorganic fertilizer use and type to the conditions faced by
2. increasing the proportion of locally produced organic materials. Organic fertilizer not
only reduces the cash cost of fertilizer to the fanner but also increases the efficiency of
inorganic fertilizer use; and,
3. making greater use of agronomic and economic factors in breeding priorities for maize
and legumes to fit future improved materials to the circumstances of smallholders.
3.2.2 Increasing Inorganic Fertilizer Use Efficiency Central to expanding the base of inorganic fertilizer users will be the task of making best use of the limited amounts of fertilizer that the typical smallholder is able to purchase. Increasing fertilizer use efficiency (FUE) will require a significant shift in thinking for both researchers and policy makers in Africa. Experience from Zimbabwe illustrates this point. Zimbabwe has a history of agricultural research on inorganic fertilizers going back over 50 years but most was geared towards users who could afford relatively large quantities of fertilizer (the large commercial farms and growers of cash crops). This research ignored the needs of those who
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 9

could not. Although, since Independence, a great deal of work has been done to look at the appropriateness of inorganic fertilizer types, amounts, timing and placement for food crops produced by smallholders (see Grant, 1981; Metelerkamp, 1988; and summarized recently by Hikwa and Mukurumbira, 1995), inorganic fertilizer recommendations fail to take adequately into account the cash constraints and the risk faced by resource poor farmers in marginal areas. In Zimbabwe, of the 32% of farmers applying the recommended package of fertilizer to their maize crop in the near average 1990/91 season, 48% failed to recover the value of the fertilizer (Page and Chonyera, 1994).
Micronutrient Supplementation
By supplementing soil micronutrients, the yield response to macronutrients (N and P) can be improved on deficient soils. Relatively cheap additions of nutrients such as Zn, B, S and Mg, that can often be included in existing fertilizer blends, when targeted to deficient soils can dramatically improve fertilizer use efficiency and crop profitability. During the 1950s, 60s and 70s S, Mg, and, less commonly, Zn and B deficiencies were detected for maize on sandy soils in Zimbabwe (Grant, 1981; Metelerkamp, 1988). Enhanced yields were obtained by including selected micronutrients in fertilizer blends (Grant, 1981). Recent experience in Malawi provides a striking example of how N fertilizer efficiency for maize can be raised by appropriate micronutrients on a location specific basis. Supplementation by S, Zn, B and K increased maize yields by 40% over the standard NP recommendation alone (Wendt, et al., 1994). New basal fertilizers with S and Zn were formulated and verified through widespread testing on-farm. Region specific recommendations on their use are now in place with extension. This experience is developed in more detail in section 4.1. to illustrate an integrated long-term research process necessary to improve soil fertility and maize production.
Organic x Inorganic Fertilizer Interactions
There is evidence that the most promising route to improving inorganic fertilizer efficiency in smallholder cropping systems is through the addition of small amounts of high quality organic matter to tropical soils (Ladd and Amato, 1985; Snapp, 1995). High quality organic manures (narrow C/N ratio, low percent lignin) provide readily available N, energy (carbon), and nutrients to the soil ecosystem, and build soil fertility and structure over the long-term. Their use will increase soil microbial activity and nutrient cycling, with reduced nutrient loss from leaching and denitrification (De Ruiter et al., 1993; Doran et al., 1987; Granastein et al., 1987; Snapp, 1995). Two long term cropping studies carried out in Kenya and Nigeria indicate that organic plus inorganic inputs sustain fertility at a higher level than the expected additive effects of either input alone (Dennison, 1961). Nutrient effects alone do not explain the benefits derived from modest organic manure inputs combined with inorganic fertilizers. High quality carbon and nitrogen provide substrate to support an active soil microbial community, which is not so much a direct nutrient supplier, but enhances the synchrony of plant nutrient demand and soil supply by reducing large pools of free nutrients (and consequent nutrient losses from the system) thus maintaining a buffered, actively cycled supply of nutrients (De Ruiter et al., 1993; Snapp, 1995). Research is now generating examples of yield gains available on-farm in southern Africa through inorganic/organic combinations. Table I gives three recent examples from our work. Often the largest gains are demonstrated on research stations where soil fertility is already high. On-farm the gains are usually lower because of inherent low soil fertility, water deficits and management compromises but still worthwhile. Many farmers already recognize
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 10

these effects and combine, where possible, small amounts of high-quality organic inputs with inorganic nutrients.
3.3 Increasing Availability And Use Of Organic Sources Of Fertility
Organic inputs available to most smallholders are rarely sufficient to maintain SOM. The rapid introduction of more organic materials is needed. Options include rotations, green manures, animal manures, intercropping, strip cropping, relay cropping, and agroforestry.
3.3.1 The Role Of Legumes
Most of the promising routes to raised SOM involve legumes. Legume-based agriculture has been the focus of well meaning efforts to transform smallholder African agriculture since the turn of the century. The efforts of Alvord in Zimbabwe are typical of the genre. Green manures were heavily researched in the 1920s-1940s (Metelerkamp, 1988) and widely used by large commercial farmers in Zimbabwe until a decline in the real price of inorganic fertilizers in the 1950s made the practice uneconomic. With the rise in real prices of inorganic fertilizers and concern over the sustainability of current cropping systems, green manures have attracted new research interest (see Hikwa and Mukurumbira, 1995). Annual legumes are used as sole crops in rotation with cereals, are intercropped, or occasionally used as green manures. Perennial legumes are sometimes retained in farmers' fields and are just beginning to be incorporated as hedgerow intercrop or alley crop systems. Giller et al. (1994) conclude that biological N fixation from legumes can sustain tropical agriculture at moderate levels of output, often double those currently achieved. Under favorable conditions green manure crops generate large amounts of organic matter and can accumulate 100-200 kg N ha-1 in 100-150 days in the tropics.
But legumes remain marginal in many of the maize-based systems of the region. Low P levels in the soils inhibit legume growth and so at a minimum P must be added. Much of the work underlying legume-based technologies has been done on research stations. Insufficient account has been taken of the need to tailor these technologies to farming circumstances where labor is short. The necessary fertilizer to "kick start" the system may be too costly or unavailable, and there are often difficulties in obtaining legume seeds (e.g., Giller, et al., 1994). Finally, the family may not be able to release land from staple food crops.
There is often a direct conflict between the short term requirement to meet today's food supply and building up the long term fertility of the soil to meet tomorrow's food needs. Farmers discount the value of a benefit that will only be achieved in several years time from investments made today. The most suitable legumes, from the soil fertility perspective, are often the hardest for the farmer to adopt. Broadly speaking, the larger the likely soil fertility benefit from a legume technology, the larger the initial investment required in labor and land, and the fewer short-term food benefits it has.
Grain Legumes
Grain legumes have the fewest adoption problems; and are widely grown by farmers, mainly for home consumption of the seed and sometimes leaves. But the more productive high harvest index grain legumes add relatively little organic matter and N to the soil since most of the above-ground dry matter and almost all the N is removed from the field in the grain (see Giller,
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems II

et al., 1994). Species that combine some grain with high root biomass and shoot-leaf biomass such as pigeonpea (Cajanus cajan) and dolichos bean (Dolichos lablab) offer a useful compromise of promoting farmer adoption and improving soil fertility (Giller pers. comm.).
Carr (1994) reports that, on severely depleted soils in Malawi, soybeans (Glycine max) will produce more calories per unit of land than unfertilized maize, in addition to fixing N from the atmosphere. Also in Malawi, Kumwenda (1995) reported average soybean yields of 2200 and 860 kg ha1, respectively, in monoculture and intercropping from a self-nodulating (promiscuous) variety 'Magoye'. Promiscuous soybeans are attractive to smallholders because they do not have to be inoculated with Rhizobium spp. bacteria to fix N and they also have good root and above-ground biomass. Magoye soybean is grown very widely in Zambia and is now grown by thousands of smallholders in Malawi, but Malawi has not yet approved recommendations for its use and Zimbabwe does little research on them. The bias in regional soybean development is towards the high grain yield types favored by large-scale producers whose interest in soil fertility improvement is secondary.
Intercropping, where two or more crops are grown mixed together on the same ground for all or most of their life cycle, is a wide-spread traditional African agricultural practice (Andrews and Kassam, 1976). Legumes in intercropping systems often contribute some residual nitrogen to the following crop (Willey, 1979). However, low growing legumes are often shaded by taller cereals (Dalal, 1974; Chang and Shibles, 1985; Manson et al., 1986) and under smallholder management in low fertility conditions, poor emergence and growth of the legume in the intercrop is common. This limits the N and organic matter contribution of the legume on-farm to levels well below the potentials found on research stations (Kumwenda et al., 1993; Kumwenda, 1995).
One of the more promising intercrops is late maturing pigeonpea. Even though early growth of the legume is reduced when intercropped with maize, pigeonpeas compensate by continuing to grow after the maize harvest and produce large quantities of biomass (Sakala, 1994). In Malawi Sakala (1994) reported a pigeonpea dry matter yield of 3 t ha' from leaf litter and flowers when intercropped with maize. Pigeonpea is easily intercropped with cereals and, even if the seed is harvested for food, the leaf fall is sufficient to make a significant contribution to N accumulation. The disadvantage is that pigeonpea is highly attractive to livestock. Thus it is rarely practical to grow where, as in many African smallholder systems, livestock are left to roam the fields uncontrolled after harvest.
The importance of intercrops in densely populated regions is widely recognized. This is due to the stabilizing effect of intercrops on food security, and enhanced efficiency of land use. However, in semi-arid areas, plant densities (and thus potential yield ha-) have to be reduced (Shumba, et al., 1990b). Those authors, working in southern Zimbabwe, found that cowpeamaize intercrops can greatly reduce the grain yield of maize in dry years. Natarajan and Shumba (1990), reviewing intercropping research in Zimbabwe, comment that cereal-legume rotations appear to offer greater prospects of raising the yield of cereals than do intercrops. While maize benefits from being grown after groundnuts (Arachis hypogaea), maize-groundnut intercrops often reduce maize yield (see Hikwa and Mukurumbira, 1995). Cowpea-maize (Vigna unguiculata) intercrops out-yield either maize or cowpea sole crops on a land equivalent basis in wetter areas of Zimbabwe (Mariga, 1990).
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 12

Legume rotations are an important practice for restoring soil fertility for farmers with larger land holdings. Rotations allow crops with different rooting patterns to use the soil sequentially, as well as reducing pests and diseases harmful to the crops, sustaining productivity of the cropping system. The amount of N returned from legume rotations depends on whether the legume is harvested for seed, forage or incorporated as green manure. In Malawi, MacColl (1989) estimated net N of 23 to 110 kg ha1 from pigeonpea, 23 to 50 kg ha"' from Dolichos bean, and 25 kg ha- from groundnuts. In Nigeria, Jones (1974) and Girt and De (1980) estimated 60 kg N ha-1 from groundnuts. The yield response of a cereal crop following a legume can be substantial. In Malawi, MacColl (1989) showed that grain yield of the first crop of maize followin. pigeonpea averaged 2.8 t ha) higher than that of continuous maize that received 35 kg N ha each year. In Zimbabwe, Mukurumbira (1985) showed large maize yields following groundnuts and bambara nuts (Voandzeia subterranea), without supplemental inorganic nitrogen. Similar studies in Tanzania by Temu (1982) showed that maize yields after sunnhemp (Crotolaria spp.) with residues removed or incorporated were 3.65 and 4.82 t ha- respectively. The latter gave a yield that was equivalent to an application of 80 kg N ha' from inorganic fertilizer. Supplementing sunnhemp green manure by incorporation with 40 kg N ha gave a yield of 6.83 t ha-I which was equivalent to an application of 160 kg haI of inorganic N fertilizer.
Natural short fallowing of overworked lands in Zimbabwe results in little or no improvement in soil fertility (Grant, 1981). Extra investment in the fallow is required for a significant improvement in soil fertility. Improved fallows, using Sesbania sesban as a way of adding significant amounts of N and organic matter to soil, have been evaluated in Eastern Province, Zambia where short duration (2-3 year) fallows are common. On station, maize yielded between 3 and 6 t ha- of grain for each of four years following a two-year improved fallow compared to 1-2 t ha' in the unfertilized control (Kwesiga and Coe, 1994). Farmers in eastern Zambia now use little fertilizer since subsidies have been removed (Place et aL., 1995). From these data Place et al., (1995) calculated that when inorganic fertilizer is not used (common farmer practice in eastern Zambia, Place et aL, (1995)) it was profitable for farmers to invest in the improved fallow. The use of improved fallows avoids the competition effect between trees and crops (Kwesiga and Coe, 1994). But where land is limiting, the feasibility of fallow systems is yet to be proved.
Agroforestry refers to land-use systems in which woody perennials (trees, shrubs, etc.) are grown in association with herbaceous plants (crops, pastures) and/or livestock in a spatial arrangement, a rotation or both, and in which there are both ecological and economic interactions between the tree and non-tree components of the system (Young, 1989). Given the widespread decline in soil fertility, the use of leguminous multi-purpose trees for soil fertility improvement has been a major thrust of agroforestry research in the region.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 13

Alley cropping, sometimes referred to as hedgerow intercropping, was developed in the late 1970s at IITA and is the best known agroforestry system developed for soil fertility improvement (Kang et al., 1990). The logic of this technology assumes that moisture and nutrients at depth in the soil, and below the roots of annual crops, are used by the deeper rooted tree crop. The tree crop is then pruned to supply nutrients to the soil and these are used by shallow rooted annuals. Jones et al. (1995a); Wendt et al. (1995); and Bunderson et al. (1991) showed that the application of Leucaena leaf prunings in an alley cropping system raised maize grain yield and increased soil pH, organic C, total N, S, and exchangeable calcium (Ca), Mg and K.
However important aspects of the technology are problematic. Inorganic P may be needed to realize increased maize yields. Leucaena leucocephala (the most widely researched hedgerow species) suffers from a number of problems including susceptibility to termite attack at the seedling stage, severe defoliation by the recently arrived Leucaena Psyllid and poor biomass production under low fertility and on soils of low pH. Saka et al. (1995) have estimated that the cost of producing N from Leuceana biomass is comparable to that from inorganic sources. The technology is labor intensive and management sensitive. Competition between the trees and associated crops can occur for moisture, and for nutrients and light (Mbekeani, 1991; Ong, 1994). Observations suggest that competition for soil moisture is particularly acute at the beginning of the rainy season before soil moisture reserves have built up. Without a better understanding of the competitive effects, especially those below ground, the potential of alley cropping is unproven at the farmer level.
There has been progress on some aspects of the above problems. Research has identified alternative species that are better adapted and produce more biomass than Leucaena leucocephala (Bunderson, 1994). Gliricidia sepium and Senna spectabilis have both shown potential to produce more biomass in a wider range of ecologies.
Other agroforestry technologies offer promise. Faidherbia albida is a vigorous leguminous tree, long used by African fanners for improving crop yields where the tree is naturally abundant. The species ha the unique characteristic among African savanna tree species of retaining its leaves in the dry season and shedding them at the onset of the rains. The resultant fine mulch of leaf litter undergoes rapid decomposition enriching the topsoil in plant nutrients and organic matter. Saka et al. (1994) have described how natural stands of the tree have been used in maize production systems in Malawi. Although the value of F. albida is recognized by farmers, the tree has not been integrated systematically into maize growing areas.
3.3.2 Cereal Residues And Animal Manure
The burning of crop residues remains common. Where it is a deliberate farmer practice (mainly in areas with few livestock), it is used to release nutrients quickly, ease land preparation and reduce pest and disease pressures. Although the nutrient content of maize stover is relatively low, stover can contribute to the productivity of the soil. Careful management of such residues is required, since N can be immobilized at the time of peak maize N requirements which can result in poor crop growth (Nandwa et al., 1995); a fact well known to farmers.
In unimodal rainfall areas where stover breakdown in the soil tends to be slow, and on sandy soils where soil N is very low, maize stover is usually fed to livestock. As well as keeping
Soil Fertility Network Results Working Paper I: Soil Fertility Research for Maize-Based Systems 14

animals alive, this is an important way of cycling the residues in a way that makes them more beneficial to the crop. Where, as in Zimbabwe, cattle are common in smallholder areas, farmers often apply cattle manure to fields that will be planted to maize. Losses of N from such systems are, however, often high. Cattle manure is applied in a dried, aerobically decomposed form, often with a high sand content, and N contents frequently less than 1.2% (Mugwira and Mukurumbira, 1984). It is broadcast-applied and ploughed into the soil before planting. Survey discussions with farmers and informal assessments show amounts applied are around 8-20 t haI with applications every 3-5 years (Mugwira and Shumba, 1986). Research shows that the most efficient use of manure is to combine it with some inorganic fertilizer (Murwira, 1994). This is common farmer practice. Station-placement or dribbling into the planting furrow, rather than broadcast application, are promising ways of increasing the crop yield benefits from cattle manure (e.g., Munguri et aL, 1995).
3.4 Maize Genotype Improvement For Low Soil Fertility
In addition to altering the soil to better support maize crops, maize genotypes are becoming available that perform better under limiting soil fertility. Slight improvement of N-use .efficiency (NUE) has been reported in tropical maize, with prospects of further gains (Lafitte and Edmeades, 1994a and b). At CIMMYT in Mexico, three cycles of full-sib recurrent selection for grain yield under low soil N, while maintaining grain yield under high soil N, was conducted in a lowland tropical population, Across 8328 (Lafitte and Edmeades, 1994a, 1994b). This resulted in a per cycle increase in grain yield under low N of 75 kg hal (2.8%), with similar yield improvements under high N (Lafitte and Edmeades, 1994b). If maintained over another 3-5 cycles of selection these improvements will be substantial. Similar work at CIMMYT-Zimbabwe in the population ZM609, adapted to mid-altitude areas of southern and eastern Africa, led to large initial gains in NUE (Short and Edmeades, 1991), but recent progress has been inconclusive (Pixley, 1995).
Also, significant progress in developing tropical maize for tolerance to high Al3+ saturation in acidic soils using full-sib recurrent selection is reported by Granados et al. (1993). These materials should be tested and modified to suit African conditions and breeding work established to develop riaize that yields well in soils that have high concentrations of H+ but
wtotA3+ M2+
without Al or Mn toxicity. Breeding for tolerance to micronutrient deficiencies may be another worthwhile long-term goal.
Agronomists and social scientists continue to have a role in encouraging breeders to persevere with work on soil fertility issues. Most breeders are more comfortable with biotic constraints and feel they can make more progress on those issues. However, the edaphic constraints found on smallholder farmers' fields need to be addressed. There is an urgent need for improved germplasm adapted to nutrient deficiencies and other edaphic stresses. Most of the gains in NUE through breeding have come from improved N utilization to produce more biomass and grain yield with no increase in total N uptake (Lafitte and Edmeades, 1994b). Nevertheless there is some concern that future genotypes bred for low soil fertility may gain their advantage from extracting more micro- and macro-nutrients from the soil. The outcome of such a strategy is uncertain and therefore it is important not to rely solely on crop improvement.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 15

3.5 Synthesis And The Way Forward
The technologies needed to manage soil fertility in southern Africa do not differ from those developed for other parts of the world. Inorganic fertilizers are a key element of fertility management as the need for added external nutrient inputs is inescapable. But, such fertilizers should not be the sole source of nutrients. There are sources of organic fertility, mainly based around the use of legumes in the cropping system, which can usefully complement inorganic fertilizers. But for both sources of fertility, important practical questions for smallholders remain unanswered. Information on the optimum use of small amounts of inorganic fertilizers, on the best combinations of organic and inorganic fertilizers, on how to produce sufficient amounts of organic manures under low fertility conditions, on the best management compromises in the use of labor between critical seasonal tasks, and on adjusting fertility management according to the seasonal rainfbdl and other external factors, is largely lacking. This is despite some attempts through adaptive research in the 1980s and early 1990s (e.g., Waddington, 1994).
Yet these are the very questions to which smallholders are seeking answers. Typically, farmers are looking for ways of combining these inputs and employing them in ways that minimize the requirements for additional cash, labor and land. There is little past, or ongoing experimentation to guide them through the choices, and little counsel from extension.
Flexible soil fertility recommendations that better address actual nutrient deficiencies, take advantage of cropping system opportunities, are efficient in the highly variable rainfall regime faced by most smallholders, and are compatible with farmer socio-economic circumstances are required. Combinations of technologies (involving organic and inorganic sources), often with each component employed sub-optimally, will be needed. Organic manures are highly variable and usually in short supply. Mixing high quality organic sources with inorganic fertilizer can make substantial improvements to nutrient use efficiency and crop productivity. The resultant practices will be practicable and profitable, and thus are candidates for wider adoption.
Transferring experience in fertility management from other regions requires not only adaptation but also a much greater understanding of the processes through which fertility can be managed under African conditions. It is to this concurrent requirement for hQJh adaptive and process research that the following sections will be devoted. We, therefore, now turn to consider improvements in the technology development and dissemination process itself to ensure more effective technology combinations are available to farmers.
4. Linking Adaptive And Process Research For Improved Technology Development And Dissemination African crop management research requires two critical thrusts to be carried out in a collaborative and highly focused manner:
1. basic research to understand better the soil fertility cycle, its inputs and its losses, on African
farms so that technology can be effective, and,
2. an integrated sequence of follow-through from basic research, through adaptive work with
farmers and other clients.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 16

The methodologies for each type of research are different and it is rare for scientists working in process research to have a full grasp of the skills needed for adaptive research (and vice versa). Fertility changes have long term effects on productivity changes in agriculture are typically incremental rather than spectacular. Understanding the changes that are taking place needs a consistent and long term effort by scientists. But it is important that they share a common vision of the problems faced by the farmers so that the research agenda remains fixed on priority productivity problems and does not get side tracked into challenging, but less relevant, avenues.
Many of the key characteristics of the applied agricultural research process are well described in literature on adaptive on-farm research (e.g., Tripp, 1992). We propose here to build on this foundation. The focus is now on adapting the techniques to embrace a long term perspective, where investigation, review, and interaction with producers and other concerned parties leads to a coherent research strategy running consistently over time. In the section following we give one example of such a process already underway. We then proceed to show how progress in applied or adaptive research requires careful underpinning with high quality, but carefully prioritized basic studies.
4.1 A Case Study Overcoming Micronutrient Deficiencies On Maize In Malawi'
Research on the relief of micronutrient deficiencies with maize in Malawi is an example of the commitment needed for "successful" soil fertility research. A study by Conroy (1993) over a three season period from 1990 to 1992, showed that smallholders following existing fertilizer recommendations (a single blanket recommendation of 92 Kg N and 40 Kg P205 per hectare was used for all farming areas of Malawi) achieved only about 3 t ha-' maize grain, using hybrids with yield potentials that exceed 10 t ha- Over 40 % of smallholders that applied the recommended rates of fertilizer failed to cover the cost of fertilizer application.
Table 2 shows the research sequence and outputs, moving from the problem of low maize yields in the presence of adequate levels of N fertilizer, through to producing fertilizers that were useful to farmers. The process embraced soil chemical analysis, through investigative trials with missing micronutrients at a relatively small number of on-farm sites, to the formulation of new fertilizer blends (including Zn and S) with a fertilizer manufacturer. This last step was combined with simple on-farm verifications of increased yield and profitability when the new formulations were used and required the active involvement of the extension service and of farmers.
In 1987, informal farmer surveys indicated that low yields were not generally due to poor planting or weeding practices or to adverse climatic conditions. But the farmers had been growing continuous maize (sometimes with a minor legume intercrop) with minimal rotation for many years. As a possible solution, researchers hypothesized that low yields might be due to deficiencies of nutrients other than N and P. Two initial goals were identified:
1. To determine yield response, in a given soil type and agro-ecozone, to fertilizers that
addressed local soil nutrient deficiencies.
The work synthesized here is the result of efforts by many persons and organizations in Malawi. In particular, credit for much of this research goes to John Wendt and Richard Jones, Rockefeller Foundation Post-Doctoral Scientists attached to the Department of Agricultural Research in Malawi from 1989 to 1993.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 17

2. To determine the relationship of soil fertility analytical values to maize yield response to
fertilizer added. This relationship is influenced by climate, soil type and profile.
From the literature and other sources of information, it was established that absence of one or more of a range of nutrients (besides N and P) could potentially depress yields significantly. It was also apparent that several of these could be discarded. Either the deficiency was unlikely, it was already addressed in the existing fertilizer formulation, or it was impractical to deal with in the Malawi context.
A series of on-farm trials were started in 1989 to evaluate deficiencies of N, P, K, S, Zn, and B at several sites throughout Malawi. A "missing nutrient, minus one" trial design was used to evaluate the relative contribution of each nutrient to maximizing yields. Because of the need to keep the clients sharply in mind, the bias was towards work conducted on smallholder farmers' fields. New methods used included tissue sampling (ear leaves were taken from all plots at silking and analyzed for N, P, Ca, Mg, K, S, Zn, Cu, Mn, Fe, and B). The soil and plant analytical data collected were correlated with yield data to determine minimum soil analytical criteria for the elements studied.
There was a regular process of review. The treatments and treatment methods evolved from year to year, based on experience from previous years. By the 199 1/92 season, a general review concluded that yield data indicated regional deficiencies of B, Zn, 5, and K. In deficient regions, average yields improved by 40% over the existing N and P application when the deficiencies were satisfied. Lime applications were necessary at some sites. Soil and plant analytical data showed P application was unnecessary at some sites, while at other sites the recommended P fertilizer rate was insufficient.
To refine these observations, the research team liaised with an existing fertilizer demonstration project run by the extension service and FAO. In 1992/93, the program saw the addition of micronutrient supplement treatments on selected extension fertilizer demonstrations throughout the country. From the analysis of soil samples from 400 smallholder sites, preliminary maps of areas with S, Zn, K, copper (Cu), P nutrient deficiencies for maize in Malawi were developed (Wendt el al., 1994).
The next step was to work with a local fertilizer company to produce practical fertilizer blends that addressed these regional deficiencies. Two types of fertilizers were jointly agreed in 1993/94 and, in 1994/95, based on further experience, another formulation was produced. Concurrently the nutrient deficiency characterization exercise was greatly expanded to better refine the nutrient deficiency maps. Over 3000 soil samples were collected, and analyzed for pH, P, Ca, K, Mg, Zn, Cu, B, soil organic matter and soil texture. The sample locations were geo-referenced (elevation, longitude and latitude) to develop a computer database that could be used with a Geographic Information System (GIS). Farmn management information was also collected. A series of verification trials, in co-operation with the extension service, were then (1993/94 onwards) put in place to verify the new fertilizer formulations as used under farmner conditions, and also to demonstrate more efficient fertilizer use methods. This work was done in conjunction with the extension service, but with guidance and assistance from DAR. In 1994/95 work was started to define new response curves for N and P with hybrid maize where the micronutrients were added. The fertilizer company has advertised the new 20:20:5+4S+0. lZn fertilizer and procedures are in place to incorporate new fertilizers into standard recommendations.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 18

Clearly blanket recommendations for fertilizer use are inadequate. They result in the inefficient use of an expensive and usually imported commodity whiich is costly for the farmer and for the nation. This is uncontroversial. But progress in dealing with this evident and crucial problem has been slow. The preceding example illustrates how progress is possible with modest resources but a clear sense of direction. It exploits the potential, through modern fertilizer blending systems, to produce small runsi" of fertilizer to given specification. It shows how, through a focused investigative and verification program, fertilizer recommendations can be developed for localized areas at a surprisingly reasonable cost. And it opens the door to refine these recommendations further through modern data base management techniques.
4.2 Basic Process Research
The applied work outlined in the preceding section must be underpinned by high quality research aimed at comprehending the basic processes underlying nutrient flows in tropical soils. Nutrient budgets and process-level research track the consequences of crop management strategies in smallholder farming systems. Synthesizing our understanding of processes involved in nutrient cycling efficiency is the goal, to be able to predict which agronomic technologies enhance yield potential over the long-term. The challenge is enormous, as detectable changes in soil organic matter and other soil fertility parameters occur very slowly. Soil organic C was not altered after 10 years of widely differing residue input rates in a Malawi hedge-row intercrop trial conducted at Bunda College of Agriculture, University of Malawi, (Materechera and Snapp, unpublished, 1995).
To develop further improved organic matter technologies the first step is to quantify nutrient losses and inputs at different scales and extrapolate among them. Typical scales include the plot level, the farm level, the watershed level and the regional level (Fresco and Kroonenberg, 1992). To evaluate the effects of technologies on soil fertility the definitive test is measurement of yields over the long term. Studies conducted over decades in a systematic, rigorous manner are clearly needed in sub-Saharan Africa and for the tropics generally. However, because farmers need improved technologies now, faster methods are required. To date, regional assessment of soil nutrient status and losses have necessarily been almost entirely based on extrapolation of plot estimates (Smaling, 1993) but with unreliable results.
New methods in soil analysis and modeling are promising tools to assess nutrient flows across the heterogeneous landscape of tropical, smallholder farms (Barrios el al., 1994; Jones et al., 1 995b; Parton, 1992; Snapp et al., 1995). Powerful tools of the trade include nutrient monitoring, new methods in soil analysis, modeling, and networked or chrono-sequence trials to allow rapid insight into soil fertility processes (Anderson and Ingram, 1993; Jackson et al. 1988; Parton, 1992; Snapp et al., 1995). Investigations must be conducted on representative sites of well-characterized agro-ecosystems. Networked trials and standardized analytical methodology, are needed to quantify the consequences of different management regimes for the soil resource base (Anderson and Ingram, 1993). Ths type of basic research holds out the promise of supporting a major shift in tropical agronomy, from empirical testing to design of crop management strategies. The work requires a continuing commitment to building bridges among fields such as agronomy, ecology, soil biology, chemistry and physics.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 19

4.2.1 Nutrient Budgeting
Nutrient budgeting can be used to develop improved ways of using available nutrients. Earlier sections underlined the importance of organic matter management on the leached and depleted soils of Africa. Using conventional methods derived from temperate zones, SOM is difficult to maintain and requires substantial amounts of organic inputs to enhance in tropical cropping systems (Jenkinson, 1981). Recent findings suggest that small additions of high quality organic material can increase nutrient cycling efficiency. High quality organic materials are high in available N and provide a source of energy (available carbon) to soil microorganisms. Enhanced activity of soil microbes can increase nutrient turnover rates, enhancing nutrient availability to crops, while minimizing nutrient losses from leaching and volatilization (De Ruiter et al. 1993; Doran et al. 1987; Granastein et al. 1987; Ladd and Amato, 1985).
This suggests an opportunity in the tropics to use active microorganism populations to buffer (or reduce nutrient losses from) low input cropping systems. Although the pool of inorganic nutrients is small in such systems, the nutrient supply may be sufficient to support crop growth, due to constant re-supply of nutrients by microbial mediated turnover rates. Nutrient losses due to leaching and volatilization are reduced (See Figure 2). Crop nutrient budget methodology (after Jackson et al., 1988), focusing on N losses (N is the nutrient most limiting growth, and most vulnerable to losses from smallholder farms in S. and E. Africa) is being developed to test this theory.
Related process studies have recently been initiated in Malawi, including 15N isotope labeling, to quantify the effects of small amounts of organic residues applied with inorganic nutrients under field conditions (Jones et al. 1995b). Researchers in Zimbabwe have undertaken a range of intensive studies of N efficiency, using 15N isotope and novel N monitoring techniques (TSBF, 1995; Tagwira, 1995). While the results from these studies remain preliminary, the initial findings are exciting.
4.2.2 Methodology Development
New techniques have recently been developed for separating SOM into fractions that may be biologically meaningful, and that appear to have predictive value for agronomists and land managers. Physically-based isolation of SOM fractions can be achieved rapidly and easily by sieving soil and separating SOM on the basis of size (Cambardella and Elliot 1994; Christensen 1992). Early indicators are that a fraction of SOM known as the light/large (LL) fraction provides a good indicator of active SOM.
Use of SOM fractionation, and measurement of LL-associated organic C and N, is being evaluated in Zimbabwe, Kenya and Malawi (Barrios et al., 1994; Okalebo et al., 1993; Snapp et al., 1995). Organic C in the LL fraction was higher in two Malawi soils amended with Gliricidia sepium or Leucaena leucocephala (a perennial legume grown as a hedge-row intercrop with maize), compared to a control (continuous maize with no fertilizer) or an intercrop with a non-N fixing legume, Senna spectabilis (Saka et al., 1995).
4.2.3 Root Studies
Roots play an important role in regulating nutrient cycling efficiency. A vigorous, effective crop rooting system is essential for efficient nutrient acquisition, particularly for mobile nutrients
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 20

such as nitrate N. Nitrogen losses from tropical cropping systems have not been well documented but are expected to be substantial, over 30-70% of N inputs. Major N losses are thought to occur early in the cropping season due to high initial mineralization rates, inorganic N accumulation during the dry season and limited root growth of young crop plants (Myers et al., 1994).
Understanding root behavior is also essential to minimize competition associated with intercropping technologies. Several important agroforestry systems are based on the presumption that the trees reduce nutrient losses by scavenging nutrients from beneath crop rooting depth (Young, 1989). Preliminary findings from two significant maize/perennial legume intercrop systems for Malawi Senna spectabilis and Gliricidia sepium alley cropped with maize (Itimu and Giller, pers. comm., 1995) suggest that ridge planting of maize (the current wide-spread practice of Malawi smallholders) reduced competition between maize and the perennial legumes. However, the data strongly pointed to intense crop/tree root competition, and modest deep rooting of the perennials. The soil environment influenced rooting patterns to as great a degree as genotype. Thus reducing negative interactions between tree roots and annual crops, and enhancing benefits from tree roots, may be fundamental to the success of the technology (Giller et al., 1995).
4.2.4 Nutrient Release Synchrony
Nutrient release synchrony with crop demand is another approach that has been advocated to increase nutrient cycling efficiency (Anderson and Ingram, 1993). Synchronizing the release of nutrients from organic materials with crop requirement for nutrients can increase nutrient cycling efficiency (Myers et al., 1994). This central hypothesis of TSBF suggests that organic inputs of varying quality can be used to manipulate nutrient supply. Laboratory incubation studies show that high quality residues (e.g., narrow C/N ratio, low percent lignin) supplied in conjunction with low quality residues (e.g., wide C/N ratio and/or high lignins) can provide a continuous nutrient supply, with N being released from high quality residues first (Itimu and Giller, pers. comm., 1995; Snapp et al., 1995; TSBF, 1995). Preliminary field data supports these findings. This was indicated for three out of four Malawi soils, in which soil inorganic N levels were enhanced in soil amended with high quality Gliricidia sepium residues, compared to controls (no fertilizer or organic inputs), or soil amended with moderately-high quality residues of Senna spectabilis (Snapp el al., 1995). Nevertheless, synchronizing nutrient release with nutrient demand using organic residues is technically demanding. Just how practicable this is for farmers on their fields every year has yet to be shown.
Quality of residues is difficult to define. The SOM fractionation methods discussed earlier may provide one method to assess residue quality after it is incorporated into soil, and most importantly, a means to quantify the consequences for soil quality. Preliminary indications are that plants with high lignins, polyphenolics and other anti-quality factors may not release nutrients for many growing seasons and thus these residues may be undesirable to use as organic soil amendments (Myers et al., 1994). Studies have been initiated in Zimbabwe and Malawi to test the synchrony hypothesis under field conditions (TSBF, 1995; Materechera and Snapp, 1995).
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 21

5. Exploiting Interactions Between Soil Fertility Technologies And Other Inputs And Management Factors
In southern Africa, the defining feature of the ecology is the 7-8 month dry season. This means that farm operations, particularly planting, weeding, and fertilizing are concentrated in the critical early weeks of the season. The options for addressing soil fertility are influenced by crop choices, by the season, the climate, and by a range of other factors some of which the farmer can control or modify, and some of which he or she cannot. Thus it is the interaction of soil fertility technology with other, possibly more readily adoptable farmer inputs and management practices that needs particular consideration. Interactions can be expected with weed control practices, with moisture availability, with fertilizer responsive and efficient maize, with labor, and with draught power.
5.1.1 Interactions With Weed Management
Kabambe and Kumwenda (1995) examined the interaction between weed growth and fertilizer use efficiency in Malawi. Their results show that farmers who weed twice at the critical periods for maize can achieve a higher yield, with half the amount of fertilizer, than farmers who only weed once (see Figure 3). Where animal draught power is available, the options for exploring interactions with weed management methods are increased (see Low and Waddington, 1990). Results from Zambia show that combining basal and top-dress fertilizer, applied when weeding the 20 cm tall maize resulted in a saving of six person days ha-' during the peak demand period for family labor. The practice also gave a 19 % yield increase compared with the standard farmer practice of a basal application just after planting, followed by late weeding and topdressings (see Low and Waddington, 1990; 1991). These results were developed into separate recommendations for ox-cultivators and for farmers who used hand hoes for weeding. Oxcultivators should weed with oxen two weeks after emergence while covering the mixed basal and topdress fertilizer (applied immediately before) in the same operation. This should be followed by hand weeding on the crop row within the subsequent two weeks. Hand-hoe weeders were recommended to combine the first hoe-weeding with a mixed basal and topdress fertilizer application, beginning 10 days after emergence. The second weeding would follow two weeks later.
Soil fertility decline may cause the build-up of weeds (and pests and diseases) and thereby may indirectly further affect crop production. A well-known example is Striga spp.. This parasitic weed tends to build up when soils are depleted (although once it is established it is hard to control just through improved soil fertility). Data from long-term trials in Kenya have shown that incorporated crop residues play an important role in reducing Striga parasitism (Odhiambo and Ransom, 1995).
5.1.2 Interactions With Moisture
The climate of southern Africa means that moisture is a frequent constraint to maize yield and to yield response to fertilizer. The efficiency (measured through grain production) of both water use and N use is raised when both are in adequate supply. But the high risk of poor response to fertilizer in dry years is a major reason why most farmers in semi-arid areas use little or no fertilizer. Over-dependence (as in Zimbabwe) on developing site specific fertilizer recommendations through soil chemical analyses alone (see Grant, 1981; Metelerkamp, 1988),
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 22

and the concept of regular maintenance dressings (for K and P) has resulted in unrealistically high recommendations for smallholder farms, especially those in semi-arid areas, that are rarely used. Uncertainty exists on appropriate NPK fertilizer recommendations for semi-arid areas. The utility and economics of fixed recommendations for such areas is increasingly questioned (e.g., Mataruka et al., 1990; Shumba et al., 1990a; Piha, 1993; Page and Chonyera, 1994).
Smallholder farmer practice in Zimbabwe reflects the need for substantially revised fertilizer recommendations. The frequency of application of inorganic fertilizer and the amounts applied by farmers is strongly influenced by seasonal rainfall. Almost all smallholder farmers apply some inorganic fertilizer to maize in the wetter areas (Natural Regions II and III). In a survey of 10 communal lands in such wetter areas Waddington et al. (1991) found that farmers applied an average of 65 kg N ha-I to their earlier plantings of maize. But with late plantings (of lower yield potential), N rates were cut to about 60% of those for earlier plantings.
In semi-arid areas, inorganic fertilizer use is low. Rohrbach (1988) found that in Chivi communal area (which is representative of these conditions) less than 10% of farmers regularly applied fertilizer. What fertilizer was used was usually low rates of topdress N. Zimbabwe data from the 1991-4 seasons in Gutu (Natural Region IV) communal area show 80% of farmers using an average of 19 kg N haI on fields that were monitored. In Chivi (Natural Region IV-V), only 23% of farmers applied fertilizer at an average rate of 8 kg N ha' (DR&SS/CIMMYT unpublished). In intermediate areas such as Shurugwi-Chiwundura and Wedza most farmers (79-91%) use topdress N but less than half applied basal fertilizer (see Huchu and Sithole, 1994).
A partial solution to excessive risk constraining fertilizer use in dry areas can be found in "response farming" techniques (Stewart and Kashasha, 1984; Stewart, 1991) that use early rainfall events to decide on the amounts of fertilizer to apply. Piha (1993) has explored the interaction between nutrient use and rainfall in Zimbabwe. His data show how, by adjusting fertilizer use to the evolving rainfall pattern in any one season, the profitability of fertilizer use can be significantly increased. Trials over 5 years on farmers' fields, show that his approach gave 25 42 % more yield and 21 41 % more profit than did existing fertilizer recommendations. He arso showed that existing recommendations were too risky for lower rainfall areas and needed to be adjusted downwards to be profitable.
Such techniques can be refined through use of crop simulation models to predict outcomes under variable water and N conditions. When coupled to GIS, outputs can be used to delineate target agro-ecological areas or groups of farmers for which a particular input level is appropriate (e.g., Dent and Thornton, 1988; Keating et al, 1992). As an example, Keating et al. (1991), using a version of the CERES-Maize simulation model modified for Kenya, was able to quantify the economic risks associated with N application for Machakos and Kitui districts in Kenya. The risk of economic loss from fertilizer application clearly rose as crop available moisture declined.
The significance of the recent work by Piha (1993) and other similar efforts (see Stewart and Kashasha, 1984; Stewart, 1991) is that productive and profitable agriculture is reliably possible on poor soils, and in semi-arid conditions, with the judicious use of inorganic fertilizers. Many of Africa's smallholders live in such environments.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 23

6. Institutional Change In Research And Development Institutions
6.1 Addressing Smallholder Diversity
To deal with the variability in the smallholder fanning community and to put in place affordable programs that can address the many and different problems involved, the population needs to be carefully and thoughtfully characterized into strata of coherent and useful size (see Blackie, 1995, for an example of stratification in this context). There are several ways in which knowledge of the characteristics of different strata might be exploited to improve the impact of new technologies. Two of these are explored as illustrations in the following sections.
6.1.1 Enhanced Farm Management Advice
There is no single route to improved maize productivity for African farmers. In the 'developed' world, farmers were guided in the use of new and expensive technologies by carefully formulated farm management advice. Such counsel is rarely available to African farmers. While clearly the cost of having highly qualified farm management advisors readily available to all African farmers is prohibitive, so too are the costs (both social and economic) of accepting the continuing decline in African farin productivity. Through joint ventures with donors and with private sector input suppliers, farm management advice can be made available to those farmers already using inputs and to those whose efficiency can be improved so as to make inputs profitable.
The emphasis moves from making inputs "affordable" (by which is typically meant in a subsidized forin, with all the inherent problems and contradictions) to making them profitable. It also moves from the diagnostic (typified by efforts in Farming Systems Research and various forms of rapid rural appraisal) and the prescriptive (from top down extension efforts to the training and visit (T+V) system favored by the World Bank) to a problem solving format in which the fanner is actively involved. This will involve recommendations that are conditional on circumstances and include IF and THEN decision making. It provides a framework for effective research/extension linkages and facilitates the evolution of a demand driven technology development process by smallholders.
6.1.2 Improving Productivity Of Those Farmers With The Fewest Resources Particular attention needs to be paid to the agricultural production problems of the resourcepoorest farmers. This group is analogous to the long term unemployed who occupy the welfare rolls of the developed world (and who remain an intractable problem for those better endowed countries). It will include the old, the work-shy, the handicapped, and many single mothers and widows. Their numbers may be substantial (as high as 40% of the smallholder population in Malawi) and are growing. These are typically farmers who do not reliably feed themselves and their families year on year. The poverty trap faced by the poorest families precludes their active participation, under present circumstances, in a market economy (except as distress sellers of labor and, sometimes, food). Both equity and reason indicate that they should not be ignored which, in fact, they are by the conventional technology development and extension process.
Cash crops can play an important role in priming the soil fertility input pump, by bringing extra income to the farmer. In Malawi, for example, burley tobacco may be the only viable crop that
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 24

can generate sufficient income for smallholders from their tiny land holdings to significantly improve cash availability (Carr, 1994). But in many situations it is unrealistic to expect sophisticated and possibly risky cash cropping to provide the first step to a better life for this group of farmers. They are desperately short of cash, cannot afford the luxury of experimentation, and often lack the confidence and the ability to deal, unaided, with many aspects of modem society. Moreover, cash cropping depends on reliable and honest markets for the inputs and outputs, and targeted production advice on a new commodity is essential for farmers.
Credit is often proposed as a solution but is only of value to individuals who are periodically short of cash to purchase inputs. For those farmers who are chronically short of cash, other alternatives will need to be sought. In some richer African countries, the informal financial sector (which commonly includes such groups as Savings Clubs, and Rotating Savings and Credit Associations) can be quite large. Clubs designed and operated by smallholders meet their special needs for financial services. In Malawi, where individual savings are minute, an innovative effort using a combination of start-up grants and savings mobilization has been remarkably successful in reaching the poorest farmers around Dowa in central Malawi (VEZA .project). It has introduced them to improved maize technologies, quickly moving them to financial viability. Savings rather than credit, therefore, provides the mechanism for introducing cash poor smallholders to improved technologies (Chimedza, 1993). The conventional wisdom holds that most households in the smallholder sector have a low marginal propensity to save, and therefore are not able to invest from savings. Experience has shown that the capacity to save is rather larger than typically assumed which gives an important leverage point when dealing with rural poverty. A sound macro-economic policy that maintains low monetary inflation is important to support such efforts.
6.2 Institutional Change To Support Smallh older Productivity Increases
The successful implementation of the technology development and transfer processes outlined in this paper requires commitment of substantial resources over extended periods. While the private sector and the non-govemnmental organization (NGO) community will be essential partners in this effort, hirgh quality public sector research and extension will remain a critical element to the process (see Blackie, 1994 for discussion on the role of the private sector). The poor funding base of most African National Agricultural Research Systems (NARSs) means that these institutions face high staff turnovers and often suffer from low morale on the part of scientists remaining in post -due to both poor financial incentives and the lack of resources with which to undertake research. Hence the NARSs, on their own, are ill equipped to take on the difficult task of maintaining institutional memory.
The agricultural technology development system in Africa is not well adapted to comprehending and responding to the long term problems of the continent. Institutional memories are poor, inadequate budgets (and the poor use of the funds that are available) lead to short lived, disparate project-orientated research which rarely is able to "follow through" to the farmer. To an extent, in plant breeding, the international agricultural research system has been able to address these deficiencies. The results are apparent in the uptake of improved germplasm. Structures for the maintenance of institutional memory for germplasm development are comparatively well developed. The International Agricultural Research Centers have a long and successful history of assisting NARSs in the development of improved varieties through
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 25

drawing on international experience of varietal development. The Centers maintain a comprehensive set of data on varietal characteristics, on breeding environments, and on pest and disease pressures. A process of collaboration and training with NARS scientists exists and is extensively used by plant breeders in developing countries.
By contrast, the effort to maintain a similar corpus of knowledge on crop management is minuscule. The role of agronomists and social scientists in the international research community is more poorly defined than that of plant breeders. Their numbers and influence are substantially less, with the research vision and agenda for crop management being largely left in the hands of the NARSs. The justification is that management issues are location specific and do not lend themselves to the kind of broad brush support that has proved so useful in crop breeding. Possibly this derives from the initial successes of the Green Revolution in Asia where improved germplasm, with broad adaptability, was taken up and incorporated into improved productivity packages by national scientists. In the African context, with a highly variable ecology and farming community, ignoring management factors avoids the heart of the problem. The continuing under-investment in long term support for high quality crop management technology development has compromised the ability of the research community to answer the real problems facing the African smallholder. Some effective ways of organizing an increased commitment follow.
6.2.1 Maintaining Institutional Memory
Crop management research is typically reported in a highly distilled format in the form of farmer recommendations. Such recommendations suffer from two important flaws. Firstly, as has been well documented in the literature, they frequently fail to take into account the diversity of farmer circumstances. Secondly, because of their highly distilled nature, they are not a good repository for long term inform-ation. The data from which the recommendations are derived may be lost long before the recommendation is discarded. Progress in soil fertility management will require that information can be retrieved and shared efficiently amongst those concerned with the promotion and development of improved technologies. Recent advances in computer databases make this task ;asier. With the small size of many African research and extension services, and their limited capacities, obvious opportunities arise from regional collaboration. Some efforts in this direction have already been made (a recent example is the Southern African Center for Cooperation in Agricultural Research and Training (SACCAR)) but have failed to realize their potential.
To get the needed impacts from soil fertility work will require a long term commitment to research and extension by interdisciplinary, inter-institutional and inter-country groups of staff that not only include technical scientists but also social scientists, extensionists, the private sector, and NGO staff. Opportunities need to be created for such staff to work closer together on common topics and trials. A climate of mutual review and criticism is essential. Financial constraints mean no one country or institution can bring sufficient resources to bear on these widespread, complex issues. There is a clear need for continuity in resources focused on soil fertility improvement, efficiencies from regional co-operation in the use of those resources, and better regional access to results. One way of developing a more integrated approach is through formal networks. A variety of networks related to soil fertility are underway at several levels.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 26

The most important process-orientated soil fertility research network in Africa is AfNet run by the Tropical Soil Biology and Fertility Program (TSBF). AfNet brings together African scientists to work on improving the synchronization of soil nutrient supply to nutrient demand by the crop, through site characterization and the use of common methods (e.g., Seward, 1995).
A good example of a more applied soil fertility research and extension network is the network established in late 1994 by the Rockefeller Foundation with national research and extension programs in Malawi and Zimbabwe, and CIMMYT. This Network emphasizes joint research priority setting, planning and integration through meetings and peer review; the conduct of priority research, including network trials across maize-based agro-ecologies; information exchange and training for network scientists, and the distribution and use of output information through enhanced interaction between the farmer, extension and research (Waddington, 1995).
The above examples overcome some of the negative attributes of previous networking efforts; they are internally driven by the participating scientists rather than by outside donors. But this is not a sufficient condition for success in addressing client needs. We need to move forward further to networks that take the lead on integrating farmers, NGOs, extension services and policy makers with research in the testing and dissemination of appropriate soil fertility technology.
We thank Paul Heisey, Greg Edmeades, Derek Byerlee and Carl Eicher for many helpful suggestions on earlier drafts of this review.
7. References
Anderson, J.M. and Ingram, J.S.I. (eds.) 1993. Tropical Soil Biology and Fertility: A Handbook of
Methods, 2nd Ed. CAB International, Wallingford, U.K.
Andrews, D.J. and Kassam, A.H. 1976. The importance of multiple cropping in increasing world food
supplies. In: Multiplg Cropping. Special Publication no. 27. American Society of Agronomy,
Madison, USA.
Araki, S. 1993. Effect on soil organic matter and soil fertility of the chitemene slash-and-bum practice
used in northern Zambia. In: Mulongoy, K. and Merckx, R. (eds.) Proceedings Soil Organic Matter Dynamics and Sustainability of Tropical Agriculture. Wiley-Sayce, Chichester, UK. pp.
Barrios, E., Buresh, R.J. and Sprent, J.1. 1994. Soil organic matter size and density fractions in
contrasting annual and perennial legume systems. Paper presented at the XV International
Congress of Soil Science, Acapulco, Mexico.
Blackie, M.J. 1994. Maize productivity for the 21st century: the African challenge. Outlook on
Agriculture 23, 189-195.
Blackie, M.J. and Jones, R.B. 1993. Agronomy and increased maize productivity in southern Africa.
Biological Agriculture and Horticulture 9, 147-160.
Blackie, M.J. 1995. Identifying smallholder farmer target groups at a local, national or regional level:
A food security based approach. In: Waddington, S.R. (ed.) Report on the First Meeting of the Network Working Group. Soil Fertility Research Network for Maize-Based Farming Systems in
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 27

Selected Countries of Southern Africa. The Rockefeller Foundation Southern Africa Agricultural Sciences Programme, Lilongwe, Malawi and CIMMYT Maize Programme,
Buddenhagen, 1. 1992. Prospects and challenges for African agricultural systems: an evolutionary
approach. Paper presented at the Carter Lecture Series on Sustainability in Africa: Integrating
Concepts. University of Florida, Gainsville, USA.
Bunderson, W.T. 1994. Comparative maize yields under alley cropping with different hedge species in
the Lilongwe Plains. Mimeo.
Bunderson, W.T., Saka, A.R., Itimu, O.A., Mbekeani, Y. and Phombeya, H.S.K. 1991. Management
of alley cropping with maize (Zea mays) and Leucaena leucocephala in Malawi. In: Saka, A.R., Bunderson, W.T. and Maghembe, J. (eds.) Proceedings, Agroforestry Research and Development in Malawi. Department of Agricultural Research, Lilongwe, Malawi. pp. 171182.
Byerlee, D., with Anandajayasekeram, P., Diallo, A., Gelaw, B., Heisey, P.W., Lopez-Pereira, M.,
Mwangi, W., Smale, M., Tripp, R. and Waddington, S. 1994. Maize Research in Sub-Saharan Africa:. An Overview of Past Impacts and Future Prospects. CIMMYT Economics Working
Paper 94-03. Mexico, D.F., Mexico.
Cambardella, C.A. and Elliot, E.T. 1994. Carbon and nitrogen dynamics of soil organic matter fractions
from cultivated grasslands. Soil Science Society ofAmerica Journal 58, 123-130.
Carr, S. 1994. The unique challenge of Malawi's smallholder agricultural sector. Mimeo. Carter, S.E. (ed.) 1993. Soil fertility management in Mutoko communal area, Zimbabwe. Report of a
Field Exercise, August 12 September 3, 1992. TSBF, Nairobi, Kenya, 75pp.
Chang, J.F. and Shibles, R.M. 1985. An analysis of competition between intercropped cowpea and
maize. II. The effect of fertilizer and plant population. Field Crops Research 12, 145-152.
Chimedza, R.M. 1993. A study of the savings behaviour of rural households using asset portfolio
analysis: The case of Zwimba Communal Area in Zimbabwe. PhD Thesis, Department of
Agricultural Economics and Extension, University of Zimbabwe.
Christensen, B.T. 1992. Physical fractionation of soil and organic matter in primary particles and
density separates. Advances in Soil Science 20, 2-90.
CIMMYT 19)2. CIMMYT 1991/92 World Maize Facts and Trends. Maize Research and Impacts in
Developing Countries. Mexico, D.F.,. CIMMYT, Mexico.
CIMMYT 1994. CIMMYT 1993/94 World Maize Facts and Trends. Maize Seed Industries
Revisited:Emerging Roles of Public and Private Sectors. Mexico, D.F., CIMMYT, Mexico.
Conroy, A.C. 1993. The economics of smallholder maize production in Malawi with reference to the
market for seed and fertiliser. PhD Thesis. University of Manchester, Manchester, UK.
Conroy, A.C. and Kumwenda, J.D.T. 1995. Risks associated with the adoption of hybrid seed and
fertilizer by smallholder farmers in Malawi. In: Jewell, D., Waddington, S.R., Ransom, J.K. and
Pixley, K.V. (eds.) Maize Research for Stress Environments, CIMMYT, Harare, Zimbabwe.
Dalal, R.C. 1974. Effects of intercropping maize with pigeonpea on grain yield and nutrient uptake.
Experimental Agriculture 10, 219-224.
De Ruiter, P.C., Moore, J.C., Zwart, K.B., Bouwmann, L.A., Hassink, J., Bloem, J., De Vos, J.A.,
Marinissen, IC.Y., Didden, W.A.M., Lebbink, G. and Brussaard, L. 1993. Simulation of nitrogen mineralization in the below ground food webs of two winter wheat fields. Journal of
Applied Ecolovy 30, 95-106.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 28

Dennison, E.B. 1961. The value of farmyard manure in maintaining fertility in northern Nigeria.
Empire Journal of Experimental Agriculture 29, 330-336.
Dent, J.B. and Thornton, P.K. 1988. The role of biological simulation models in farming systems
research. Agricultural Administration and Extension 29, 111-122.
Doran, J.W., Fraser, D.G., Culik, M.N. and Liebhardt, W.C. 1987. Influence of alternative and
conventional agricultural management on soil microbial processes and nitrogen availability.
American Journal ofAlternative Agriculture 2, 99-106. FAO 1988. FAO Year Book, FAO, Rome.
Fresco, L.O. and Kroonenberg, S.B. 1992. Time and spatial scales in ecological sustainability. Land Use
Policy. Butterworth-Heinenann. pp. 155-168.
Gilbert, E., Phillips, L.C., Roberts, W., Sarch, M-T., Smale, M. and Stroud, A. 1993. Maize Research
Impact in Africa: The Obscured Revolution. Report prepared for the Division of Food, Agriculture and Resources Analysis, Bureau for Africa, US Agency for International
Giller, K.E., McDonagh, J.F. and Cadisch, G. 1994. Can biological nitrogen fixation sustain agriculture
in the tropics? In: Syers, J.K. and Rimmer, D.L. (eds.) Soil Science and Sustainable Land
Management in the Tropics. CAB International, Wallingford, UK. pp. 173-191.
Giller, K.E., Itimu, 0. and Masamba, C. 1995. Nutrient sourcing by agroforestry species and soil
fertility maintenance in maize associations. Paper presented at the Second Meeting of the Soil Fertility Network Working Group for Maize-Based Farming Systems, 17 21 July 1995,
Kadoma, Zimbabwe.
Giri, G. and De, R. 1980. Effects of preceding grain legumes on growth and nitrogen uptake of dryland
pearl millet. Plant and Soil 56, 459-464.
Granados, G., Pandey, S. and Ceballos, H. 1993. Response to selection for tolerance to acid soils in a
tropical maize population. Crop Science 33, 936-940.
Granastein, D.M., Bezdicek, D.F., Cochran, V.L., Elliot, L.F. and Hammel, J. 1987. Long-term tillage
and rotation effects on soil microbial biomass, carbon, and nitrogen. Biology and Fertility of Soils
5, 265-270.
Grant, P.M. 1981. The fertilization of sandy soils in peasant agriculture. Zimbabwe Agricultural Journal
78, 169-175.
Heisey, P.W. and Mwangi, W. 1995. Fertilizer and maize production in sub-Saharan Africa: Use and
policy options. Paper presented at a workshop on The Emerging Maize Revolution in Africa. The Role of Technology, Institutions and Policy, 9-12 July 1995, Michigan State University, East
Lansing, USA.
HIID 1994. Fertiliser policy study:market structure, prices and fertiliser use by smallholder farmers.
Harvard Institute for International Development and Economic Planning and Development Department, Office of the President and Cabinet, Government of Malawi, Lilongwe, Malawi.
Hikwa, D. and Mukurumbira, L. 1995. Highlights of previous, current and proposed soil fertility
research by the Department of Research and Specialist Services (DRSS) in Zimbabwe. In: Waddington, S.R. (ed.) Report on the First Meeting of the Network Working Group. Soil Fertility Research Network for Maize-Based Farming Systems in Selected Countries of Southern Africa.
The Rockefeller Foundation Southern Africa Agricultural Sciences Programme, Lilongwe,
Malawi and CIMMYT Maize Programme Harare, Zimbabwe. pp. 21-33.
Huchu, P. and Sithole, P.N. 1994. Rates of adoption of new technology and climatic risk in the
communal areas of Zimbabwe. In: Craswell, E.T. and Simpson, J. (eds.) Soil Fertility and
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 29

Climatic Constraints in Dryland Agriculture. ACIAR Proceedings No.54, Canberra, Australia,
pp. 44-50.
Jackson, L.E., Strauss, R.B., Firestone, M.K. and Bartolome, J.W. 1988. Plant and soil nitrogen dynamics in California annual grassland. PlantSoil 110, 9-17.
Jenkinson, D.S. 1981. The fate of plant and animal residues in soil. In: Greenland, D.J. and Hayes, M.H.B. (eds.), The Chemistry ofSoil Processes. John Wiley and Sons, pp. 505-561.
Jones, M.J. 1974. Effects of previous crop on yield and nitrogen response of maize at Samaru, Nigeria.
Experimental Agriculture 10, 278-279.
Jones, R.B. and Wendt, J.W. 1995. Contribution of soil fertility research to improved maize production
by smallholders in southern Africa. In: Jewell, D., Waddington, S.R., Ransom, J.K. and Pixley,
K.V. (eds.) Maize Research for Stress Environments, CIMMYT, Harare, Zimbabwe.
Jones, R.B., Wendt, J.W., Bunderson, W.T. and Itimu, O.A. 1995a. The effects of N and P
application on soil properties, maize yield, and nutrient uptake in a Leucaena alley cropping
system. Mimeo.
Jones, R.B., Snapp, S.S., Giller, K.G. and Phombeya, H.S.K. 1995b. Effect of organic matter quality
on N fertiliser use efficiency as measured by 15N recovery in maize and soil. Paper to be presented at Driven by Nature Plant Litter Quality and Decomposition Conference, Wye
College, UK.
Kabambe, V.H. and Kumwenda, J.D.T. 1995. Weed management and nitrogen rate effects on maize
yield in Malawi. In: Jewell, D., Waddington, S.R., Ransom, J.K. and Pixley, K.V. (eds.) Maize
Research for Stress Environments, CIMMYT, Harare, Zimbabwe.
Kang, B.T., Reynolds, L. and Atta-Krah, A.N. 1990. Alley farming. Agronomy Abstracts 43, 315-359. Keating, B.A., Godwin, D.C. and Watiki, J.M. 1991. Optimizing nitrogen inputs in response to climatic
risk. In: Muchow, R.C. and Bellamy, J.A. (eds.) Climatic Risk in Crop Production: Models and Management for the Semiarid Tropics and Subtropics. CAB International, Wallingford, UK, pp.
Keating, B.A., Wafula, B.M. and Watiki, J.M. 1992. Development of a modeling capability for maize in
semi-arid Eastern Kenya. In: Probert, M.E. (ed.) A Search for Strategies for Sustainable Dryland
Cropping in Semi-aridEastern Kenya. ACIAR Proceedings No.4 1, Canberra, pp.26-33.
Kumwenda, J.D.T. 1995. Soybean spatial arrangement in a maize/soybean intercrop in Malawi. Paper
presented at the Second African Crop Science Conference foi Eastern and Central Africa, 19-24
February 1995, Blantyre, Malawi.
Kumwenda, J.D.T., Kabambe, V.H.K. and Sakala, W.D.M. 1993. Maize-soybean and maize-bean
intercropping experiments. Maize Agronomy Annual Report for 1992/93. Chitedze Agricultural
Research Station, Lilongwe, Malawi.
Kwesiga, F. and Coe, R. 1994. The effect of short rotation Sesbania sesban planted fallows on maize
yield. Forest Ecology and Management 64, 199-208.
Ladd, J.N. and Amato, M. 1985. Nitrogen cycling in legume-cereal rotations. In: Kang, B.T. and van der
Heide, J. (eds.), Nitrogen Management in Farming Systems in Humid and Sub-Humid Tropics.
IITA and ISF, The Netherlands, pp. 105-127.
Lafitte, H.R. and Edmeades, G.O. 1994a. Improvement for tolerance to low soil nitrogen in tropical
maize. I. Selection criteria. Field Crops Research 39, 1-14.
Lafitte, H.R. and Edmeades, G.O. 1994b. Improvement for tolerance to low soil nitrogen in tropical
maize. II. Grain yield, biomass production, and N accumulation. Field Crops Research 39, 15-25. Lele, U. 1992. Aid to African Agriculture. Johns Hopkins University Press, Baltimore, USA.
Soil Fertility Network Results Working Paper I: Soil Fertility Research for Maize-Based Systems 30

Low, A.R.C. and Waddington, S.R. 1990. On-farm research on maize production technologies for smallholder farmers in southern Africa: Current achievements and future prospects. In: Gebrekidan, B. (ed.), Maize Improvement, Production and Protection in eastern and southern
Africa. CIMMYT, Nairobi, Kenya, pp. 491-511.
Low, A., and Waddington, S.R. 1991. Farming systems adaptive research: Achievements and prospects in southern Africa. Experimental Agriculture 27, 115-125.
MacColl, D. 1989. Studies on maize (Zea mays L.) at Bunda, Malawi. II. Yield in short rotation with
legumes. Experimental Agriculture 25, 367-374.
Manson, S.C., Leighner, D.E. and Vorst, J.J. 1986. Cassava-cowpea and cassava-peanut intercropping.
III. Nutrient concentration and removal. Agronomy Journal 78, 441-444.
Mariga, I.K. 1990. Effect of cowpea planting date and density on the performance of a maize-cowpea
intercrop. Zimbabwe Journal ofAgricultural Research 28, 25-131.
Mashiringwani, N.A. 1983. The present nutrient status of the soils in the Communal Farming Areas of
Zimbabwe. Zimbabwe Agricultural Journal 80, 73-75.
Mataruka, D.F., Makombe, G. and Low, A. 1990. The contribution of economic analysis in developing
promising research agendas: A maize nitrogen x phosphorus trial in Zimbabwe. Farming Systems
Bulletin, Eastern and Southern Africa 5, 1-5.
Materechera, S.A. and Snapp, S.S. 1995. Novel assay method to quantify residue decomposition in
Malalwi cropping systems. Paper presented at Driven by Nature, Plant Litter Quality and
Decomposition Conference, Wye College, UK.
Mbekeani, Y. 1991. Effects of spacing on water relations, light interception and biomass partitioning in
a Leucaena-maize alley cropping system. MSc thesis, Washington State University, Pullman,
Metelerkamp, H.R.R. 1988. Review of crop research relevant to the semiarid areas of Zimbabwe. In:
Proceedings of a workshop on Cropping in the Semiarid Areas of Zimbabwe.
Agritex/DR&SS/GTZ, Harare, Zimbabwe, pp. 190-315.
Mugwira, L.M. and Mukurumbira, L.M. 1984. Comparative effectiveness of manures from the
communal area and commercial feedlots as plant nutrient sources. Zimbabwe Agricultural
Journal 81, 241-250.
Mugwira, L.M. and Shumba, E.M. 1986. Rate of manure applied in some communal areas and their
effect on plant growth and maize grain yields. Zimbabwe Agricultural Journal 83, 99-104.
Mukurumbira, L.M. 1985. Effects of rate of fertilizer nitrogen and previous grain legume crop on maize
yields. Zimbabwe Agricultural Journal 82, 177-179.
Munguri, M.W., Mariga, I.K. and Chivinge, O.A. 1995. The potential of optimizing cattle manure use
with maize in Chinyika Resettlement Area, Zimbabwe. Paper presented at the Second Meeting of the Soil Fertility Network Working Group for Maize-Based Farming Systems, 17-21 July 1995,
Kadoma, Zimbabwe.
Murwira, H. 1994. Nitrogen dynamics in a sandy Zimbabwean soil under manure fertilization. In: TSBF
1994 Annual Report, Tropical Soil Biology and Fertility Program, Nairobi, Kenya.
Myers, R.J.K., Palm, C.A., Cuevas, E., Gunatilleke, I.U.N. and Brossard, M. 1994. The synchronization
of nutrient mineralization and plant nutrient demand. In: Woomer, P.I. and Swift, M.J. (eds.), The Biological Management of Tropical Soil Fertility, John Wiley and Sons, Chichester, UK. pp. 81116.
Nandwa, S.M., Anderson, J.M. and Seward, P.D. 1995. The effect of placement of maize stover and N
fertilization on maize productivity and N use efficiency in the semi-arid and sub-humid agroSoil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 31

ecological zones of Kenya. In: Jewell, D., Waddington, SR., Ransom,'J.K. and Pixley, K.V.
(eds.), Maize Research for Stress Environments, CIMMYT, Harare, Zimbabwe.
Natarajan, M. and Shumba, E.M. 1990. Intercropping research in Zimbabwe: Current status and outlook
for the future. In: Waddington, S.R, Palmer, A.F.E. and Edje, O.T. (eds.) Research Methods for
Cereal/Legume Intercropping, Cl MMYT, Mexico, D.F., Mexico, pp.190-193.
Odhiambo, G. and Ransom, J.K. 1995. Long term strategies for Striga control. In: Jewell, D.,
Waddington, S.R., Ransom, J.K. and Pixley, K.V. (eds.) Maize Research for Stress Environments,
CIMMYT, Harare, Zimbabwe.
Okalebo, J.R., Gathua, K.W. and Woomer, P.L. 1993. Laboratory Methods of Soil and Plant Analysis:
A Working Manual. KARI and TSBF, Nairobi, Kenya. ''
Ong, C. 1994. Alley cropping ecological pie in the sky? Agroforesty Today 6, 8-10.
Page, S. and Chonyera, P. 1994. The promotion of maize fertiliser packages: a cause of household food
insecurity and peasant impoverishment in high rainfall areas of Zimbabwe. Development
Southern Africa 11,301-320.
Parton, K.A.. 1992. Socio-economic modelling of decision making with respect to choice technology by
resource-poor farmers. In: Probert, M.E. (ed.) A Search for Strategies for Sustainable Dryland
Cropping in Semi-aridEastern Kenya. ACIAR Proceedings No.41, Canberra, pp.110-118.
Piha, M.I. 1993. Optimizing fertilizer use and practical rainfall capture in a semi-arid environment with
variable rainfall. Experimental Agriculture 29, 405-415.
Pinstrup-Andersen, P. 1993. World food trends and how they may be modified. International Food
Policy Research Institute, Washington D.C., USA.
Pixley, K.V. 1995. CIMMYT mid-altitude maize breeding program. Report of activities during 1993/94.
In: Annual Research Report CIMMYT-Zimbabwe, November 1993 to October 1994. CIMMYT,
Harare, Zimbabwe, pp. 7-39.
Place, F., Mwanza, S. and Kwesiga, F. 1995. A cost-benefit analysis of improved fallows in Eastern
Province, Zambia. (Mimrneo).
Rohrbach, D.D. 1988. The growth of smallholder maize production in Zimbabwe: Causes and
implications for food security. PhD thesis, Michigan State University, USA.
Saka, A.R., Bunderson, W.T., Itimu, O.A., Phombeya, H.S.K. and Mbekeani, Y. 1994. The effects of
Acacia albida on soils and maize grain yields under smallholder farm conditions in Malawi.
Forest Ecology and Management 64, 217-230.
Saka, A.R., Phombeya, H.K.S., Jones, R.B., Minae, S. and Snapp, S.S. 1995. Agroforestry
Commodity Team Annual Report 1993/94 Season. Department of Agricultural Research,
Lilongwe, Malawi.
Sakala, W.D.M. 1994. Crop management interventions in traditional maize pigeonpea intercropping
systems in Malawi. MSc Thesis, Bunda College of Agriculture, University of Malawi,
Lilongwe, Malawi.
Sanchez, P.A. 1995. Tropical soil fertility research: Towards the second paradigm. Mimeo. Seward, P.D. 1995. The Tropical Soil Biology and Fertility (TSBF) Program in southern Africa. In:
Waddington, S.R. (ed.) Report on the First Meeting of the Network Working Group. Soil Fertility Research Network for Maize-Based Farming Systems in Selected Countries of Southern Africa.
The Rockefeller Foundation Southern Africa Agricultural Sciences Programme, Lilongwe,
Malawi and CIMMYT Maize Programme, Harare, Zimbabwe. pp.13-20.
Short, K.E., and Edmeades, G.O. 1991. Maize improvement for water and nitrogen deficient
environments. In: MacRobert, J.F. (ed.) Proceedings of the Crop Science Society of Zimbabwe
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 32

Twenty-First Anniversary Crop Production Congress. Crop Science Society of Zimbabwe,
Harare. pp. 63-68.
Shumba, E. 1989. Maize technology research in Mangwende, a high potential communal area
environment in Zimbabwe. Part 2: The on-farm experimental program. Farming Systems
Bulletin, Eastern and Southern Africa 1, 1-13.
Shumba, E.M., Dhliwayo, H.H., Kupfuma, B. and Gumbie, C. 1990a. Response of maize in rotation with cowpea to NPK fertilizer in a low rainfall area. Zimbabwe Journal ofAgricultural Research
28, 39-45.
Shumba, E.M., Dhliwayo, H.H. and Mukoko, O.Z. 1990b. The potential of maize-cowpea intercropping in low rainfall areas of Zimbabwe. Zimbabwe Journal ofAgricultural Research 28, 33-38.
Smaling, E. 1993. Soil nutrient depletion in sub-Saharan Africa. In: Van Reuler, H. and Prins, W. (eds.) The Role of Plant Nutrients for Sustainable Food Crop Production in sub-Saharan Africa.
Leidschendam: VKP.
Snapp, S.S. 1995. Improving fertilizer efficiency with small additions of high quality organic inputs. In: Waddington, S.R. (ed.), Report on the First Meeting of the Network Working Group. Soil Fertility Research Network for Maize-Based Farming Systems in Selected Countries of Southern Africa.
The Rockefeller Foundation Southern Africa Agricultural Sciences Programme, Lilongwe,
Malawi and CIMMYT Maize Programme, Harare, Zimbabwe. pp. 60-65.
Snapp, S.S., Phombeya, H.S.K. and Materechera, S. 1995. An agroforestry based test of new fractionation methods in soil organic matter analysis. Crop Science Society of Zimbabwe. 25th
Anniversary Symposium. June, 1995. Harare, Zimbabwe.
Stewart, J.I. 1991. Principles and performance of response farming. In: Muchow, R.C. and Bellamy, J.A. (eds.) Climatic Risk in Crop Production: Models and Management for the Semiarid Tropics
and Subtropics, CAB International, Wallingford, UK. pp.361-382.
Stewart, J.I. and Kashasha, D.A.R. 1984. Rainfall criteria to enable response fannrming through cropbased climate analysis. East African Agriculture and Forestry Journal, Special Issue, pp.58-79. Stoorvogel, J.J., Smaling, E.M.A., and Janssen, B.H. 1993. Calculating soil nutrient balances in Africa
at different scales. I. Supranational scale. Fertilizer Research 35, 227-235.
Swift, M.J. and Woomer, P. 1993. Organic matter and the sustainability of agricultural systems:
Definition and measurement. In: Mulongoy, K and Merckx, R. (eds.) Soil Organic Matter
Dynamics and Sustainability of Tropical Agriculture. Wiley-Sayce, Chichester, UK. pp.3-18.
Tagwira, F. 1995. Current and proposed soil fertility research at Africa University. In: Waddington, S.R.
(ed.). Report on the First Meeting of the Network Working Group, Soil Fertility Research Network for Maize-Based Farming Systems in Selected Countries of Southern Africa. The
Rockefeller Foundation Southern Africa Agricultural Sciences Programme, Lilongwe, Malawi
and CIMMYT Maize Programme, Harare, Zimbabwe, pp. 39-41.
Temu, A.E.M. 1982. Rotational green manuring with Crotalaria in the southern highlands of Tanzania.
Southern Highlands Maize Improvement Program, Progress Report, Tanzania.
Tripp, R. 1992. Expectations and realities in on-farm research. Farming Systems Bulletin, Eastern and
Southern Africa 11, I-13.
TSBF 1995. The biology and fertility of tropical soils. Report of the Tropical Soil Biology and Fertility
Program 1994. Tropical Soil Biology and Fertility Program, Nairobi, Kenya.
Waddington, S.R. 1994. Overview of the current situation and previous impact of adaptive agricultural
research in southern Africa In: Craswell, E.T. and Simpson, J. (eds.) Soil Fertility and Climatic
Constraints in DrylandAgriculture. ACIAR Proceedings No.54, Canberra, Australia, pp. 7-14.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 33

Waddington, S.R. 1995. (ed.) Report on the First Meeting of the Network Working Group. Soil Fertility Research Network for Maize-Based Farming Systems in Selected Countries of Southern Africa.
The Rockefeller Foundation Southern Africa Agricultural Sciences Programme, Lilongwe,
Malawi and CIMMYT Maize Programme, Harare, Zimbabwe, I I0pp.
Waddington, S.R., Mudhara, M., Hlatshwayo, M. and Kunjeku, P. 1991. Extent and causes of low yield
in maize planted late by smallholder farmers in sub-humid areas of Zimbabwe. Farming Systems
Bulletin, Eastern and Southern Africa 9, 15-31.
Weber, M., Staaz, J., Holtzman, J., Crawford, E. and Bernsten, R. 1988. Informing food security
decisions in Africa: empirical analysis and dialogue. American Journal of Agricultural
Economics 70, 1045-1054.
Wendt, J.W., Jones, R.B., and Itimu, O.A. 1994. An integrated approach to soil fertility improvement in
Malawi, including agroforestry. In: Craswell, E.T. and Simpson, J. (eds.) Soil Fertility and
Climatic Constraints in DrylandAgriculture. ACIAR Proceedings No.54, Canberra, pp. 74-79.
Wendt, J.W., Jones, R.B., Bunderson, W.T. and ltimu, O.A. 1995. Residual effects of Leucaena leaf
application, leaf removal, and P application on maize yields, maize nutrient uptake, and soil
properties in an alley cropping system. Mimeo.
Willey, R.W. 1979. Intercropping its importance and research needs. Part 1. Competition and yield
advantages. Field Crops Abstracts 32, 1-10.
Woomer, P.L., Martin, A., Albrecht, A., Resck, D.V.S., and Scharpenseel, H.W. 1994. The importance
and management of soil organic matter in the tropics. In: Woomer, P.L. and Swift, M.J. (eds.)
The Biological Management of Tropical Soil Fertility. Wiley-Sayce, Chichester, UK, pp.47-80. Young, A. 1989. Agroforestryfor Soil Conservation. CAB International, Wallingford, UK. 276pp. Zambezi, B.T. Kumwenda, J.D.T. and Jones, R.B. 1993. Closing the maize yield gap in Malawi. In:
Munthali, D.C., Kumwenda, J.D.T. and Kisyombe, F. (eds.), Proceedings of the Conference on Agricultural Research for Development. Department of Agricultural Research, Lilongwe and
Chancellor College, Zomba, Malawi.
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems 34

Figure 1. Malawi Maize Production Scenarios for 1995-2015
20 1.50
S14 0
C 0
.-12 0.00 C
C 10 -0.50
.2 V
6 -1.00
C. 6
-1.50 N
2 -2.00
0 -2.50
0) 0) a) 0) 0) 0 C0 CD 0) 0D 0) 0) 0) 0D 0CD0) 0) 0) a) 0) 0 0 0 0 0 0 0 0 0 0> 0 0 0 0 0
- -- N N N CN N N CN N N N CN N N N CN N
SI--(Scenario 2) Hybrid Maize Area Increases I(Scenario 3) As 2 + Better Management
---- Population (Millions) ---- (Scenario 1) Current Maize & Management


Table 1. Examples of gains in maize yield through inorganic/organic fertilizer combinations at
.levels practicable on-farm in Malawi and Zimbabwe. These gains are from the first cropping
season after fertilizer application. Additional gains can be expected in following seasons.
Organic Inorganic Location
Fertilizer Fertilizer and Season
No Organic Inorganic Organic + Yield of
Fertilizer Alone Alone Inorganic combination as
Combination a percent of alone
Leucaena 30 kg N ha-' Chitedze 2.24 3.32 2.72 3.60 119
leucocephala alley applied to Research cropped with maize crop Station,
maize. 1.5 2.5 t Malawi
ha"1 Leucaena leaf 1988-1990
prunings applied
to maize
Pigeonpea 48 kg N ha1 Lunyangwa 0.87 1.70 1.98 2.31 126
intercropped with applied to Research maize previous maize crop Station, season. Pigeonpea Malawi
residues 1993-1995
incorporated into
Cattle manure 112 kgN, 17 6 communal 0.79 1.11 1.30 1.93 160
13-25 t ha' kg P and 16 farms in
broadcast and kg K ha1 as a Wedza and
ploughed in before split basal and Chinyika, planting topdress Zimbabwe
1994/95, a
drought year
Soil Fertility Network Results Working Paper 1: Soil Fertility Research for Maize-Based Systems

Table 2. The relief of micronutrient deficiencies on maize in Malawi, as an example of a long term research process.
1987-89 1989-90 1990-91 1991 -92 1992-93 1993-94 1994 95 and onwards
Activities and Methods
* Review Missing nutrient trials at 10 on farm & Nutrient Map regional nutrient deficiencies
past trials station sites in selected areas supplement Formulate new basal fertilizers and distribute
and conduct Soil and tissue chemical analysis at sites treatments to Collect and analyze soil from 3000 geofarm surveys on farm demos referenced sites throughout Malawi
* Soil analysis Verification trials of new fertilizers at several
at 400 sites in hundred on-farm sites, involving farmers
central Malawi Use GIS to map sites
* Better target verifications
to sites deficient in S & Zn,
involve extension & farmers
* N response trials at sites
where micronutrient
Key Outputs deficiencies satisfied
* Micronutrient Regional deficiencies of S, Zn, B &K found Near universal Good yield Optichem to produce compdeficiency Yields increase by 40 % over standard NP S & Zn deficiency response to ound fertilizer with 1Zn
likely cause when micronutrient deficiencies satisfied in some ADDs basal ferts Include results from better
of yields < 3t P not needed at some sites while others Confirmed yield containing S targeted verifications in per ha with need more than current recommendation responses Responses at new recommendations
current NPK More verification with farmers needed Optichem to low Zn sites Results from response
recommendation produce 20:20:5 were poor, economics
+4S+0.lZn and 0.1Zn is Results from N response
15:10:5+4S+0.lZn insufficient trials where micronutrients
Organizations relieved
* Maize and Soils Commodity Research Teams in the Department of Agricultural Research, Ministry of Agriculture
* Smallholder Farmers
* FAO/MoA/UNDP project
Fertilizer Maker (Optichem) 10
* MoA Extension ServiceD-

Fertilizer Quality
Fertilizer Fics
Soil Soil
Organic Organic Subs rate
-- I: :'-;:;:'"A::'W ""--lan UpakePlant Uptake
Leaching and ILeaching and
Volatilization Losses Volatilization Losses
Figure 2. The role of high quality organic inputs in improving the efficiency of inorganic fertilizer use. Rapid cycling between inorganic N and soil microbe N allows the plant to take up N without a large pooi of inorganic N which is vulnerable to leaching.

Figure 3. Effect of weeding frequency and N application on maize grain yield
5000 j
4500 z ..
4000 ..." "
3500 -" z
3 20O
2500 2000
cc 1000
500 kg N h"
W eeding We d n at 21, 45 Weeding Weeding and 54 att28
and 45 weeding
Weeding times (days after planting)
Source: Kabambe and Kumwenda, 1995

Soil Fertility Network for Maize-Based Cropping Systems in Countries of Southern Africa
Soil Fertility Network Coordinator CIMMYT Maize Research Station The University of Zimbabwe Farm
P.O. Box MP 163 12.5 km Peg
Mount Pleasant Mazowe Road
Harare, Zimbabwe
Phone: (263) 4 301807
Fax/Phone: (263) 4 301327 E-Mail:

@ CIMMYT International Maize and Wheat Improvement Center The CIMMYT International Maize and Wheat Improvement Center holds all rights to the source text and shall be considered the copyright holder for the text and images of these publications.
The CIMMYT International Maize and Wheat Improvement Center has made this publication available to the University of Florida, for purposes of digitization and Internet distribution. The CIMMYT International Maize and Wheat Improvement Center reserves all rights to this publication. All uses, excluding those made under "fair use"~ provisions of U.S. Code, Title 17, Section 107 are restricted. Contact the CIMMYT International Maize and Wheat Improvement Center for additional information and permissions.