Developing No-Till Packages for Small-Scale Farmers

Javier Ekboir

Introduction conditions of small-scale farmers. No-till demonstrates, those differences are has been found to cut unit production extremely important to the potential Population and income growth are costs (often up to 50%) and to reduce development and adoption of no-till in inducing an intensification of agricultural risks. Small-scale farmers small-scale agriculture. agriculture, which threatens the have used the additional income and sustainability of natural resources in First, no-till represents a major shift in time to start new income-generating rural environments and the livelihoods the agricultural paradigm, with activities. No-till has also benefited of small-scale farmers who depend on consequent implications for researchers women and children in rural those resources to survive. Though no and farmers. The old paradigm was households by facilitating some of the single action will reduce poverty and based on massive soil disturbance with tasks they perform, such as weeding. increase the sustainability of plows and powerful . The new Living standards rise as families have agricultural production, technical paradigm, whose best exponent is no- more money for nutrition, education, change can play a prominent role in till, is based on three principles: and housing. In addition to its economic alleviating these problems. Among the minimizing soil disturbance, covering benefits, no-till confers significant technologies capable of achieving these the soil with plants or plant residues for environmental benefits. It substantially two goals simultaneously, no-till is one as long as possible, and rotating crops. reduces erosion and has the potential to of the most promising in most Since this technology is very sensitive to diminish pollution from agrochemicals, environments, 1 especially for small- environmental conditions, the specific because of a fall in the volume of scale farmers. components of no-till packages differ chemicals used and a shift to less toxic across locations but still share some of No-till is sometimes regarded as products. By reducing fuel consumption the principles mentioned above. Even unsuitable for small-scale farmers, but and conserving organic matter in the though packages that reflect all three as this report will show, much evidence soil, no-till reduces the emission of principles yield the highest long-term indicates that this view is generally greenhouse gas from agriculture. economic and environmental benefits, it untrue. Because it reduces the labor and No-till is characterized by two profound has been difficult to develop them. Most effort invested in food production, no- differences from other agricultural packages only minimize soil till is particularly appropriate for the technologies. As this report movements and keep the soil covered. Second, no-till technology is the outcome of a complex social process, 1 No-till has been developed basically for areas where flood irrigation is not used. Development of no-till or reduced till packages for flood irrigated areas is still in its infancy. Unless otherwise and understanding the development indicated, in the remainder of this report the term no-till refers to packages developed for areas where flood-irrigation is not used (i.e., rain fed or where other irrigation systems are used).

1 2001 CIMMYT World Wheat Overview and Outlook and adoption of no-till requires a new million hectares worldwide were Improvement Center (CIMMYT), which framework for the socioeconomic study cropped under no-till, mostly by large- has participated in developing no-till of technical change. In the traditional scale farmers (Derpsch 2001). A large technology since the 1970s. Based on an framework, technical change occurs number of small-scale farmers have extensive review of experiences with along a continuum that starts with also adopted no-till, but because these no-till, 3 including CIMMYT’s basic research, is followed by adaptive farmers crop small areas, their adoption experience, this report explores a research, and ends with adoption by of no-till should not be measured in number of issues related to the farmers. Most exchanges of numbers of hectares but in numbers of development and adoption of no-till information between the agents adopters. Although accurate estimates technology, especially in small-scale involved in technology development are lacking, in recent years adoption of agriculture. It begins with a description and adoption are represented as no-till has occurred at a rapid pace in of no-till’s special features and the unidirectional and indirect flows, several countries. It is estimated that innovation systems through which the mediated by markets or printed more than 90% of small-scale farmers technology evolves. As mentioned material. In the new framework, which in southern Brazil use no-till (Denardin earlier, a new framework for the is based on the concept of “innovation and Kochhann 1999); the number of no- socioeconomic study of technical systems,” technology development and till farmers in Ghana in 2000 was change is needed to understand the adoption are social phenomena in estimated at 100,000 (Ekboir, Boa, and development and adoption of no-till. which agents interact in several ways, Dankyi, forthcoming); and about This report describes that framework creating multiple information flows in 100,000 ha of wheat were planted with and reviews the methodological many directions. These agents (e.g., no-till in India and Pakistan in 2001 difficulties involved in measuring the public research and extension systems, (Peter Hobbs, personal communication, benefits and adoption of no-till. This innovative farmers, commercial firms, 2001). No-till has been promoted in information provides the context for a foreign research institutions) form Central America for the last 20 years series of six case studies of no-till networks that co-evolve with the with localized successes. No-till experiences among small-scale farmers technologies they create. Just as no-till programs have been established for in Latin America, Africa, and Asia. The packages are location-specific, no-till small-scale farmers in Bolivia, Burkina concluding sections of the report networks and their evolution are also Faso, Cameroon, Côte d’Ivoire, discuss key factors that condition the unique. In some cases, as in South Ethiopia, Indonesia, Kenya, Malawi, effectiveness of no-till programs for America, the main forces driving the Mali, Mozambique, Paraguay, Senegal, small-scale farmers, including the innovation system were the South Africa, Tanzania, Uganda, implications for national and commercial interests of input suppliers Zambia, and Zimbabwe, but they are international agricultural research. and commercial farmers’ need for too new to show an impact. Although sustainable technologies. In other cases, no-till is also being promoted in China What Is No-Till? as in the Indo-Gangetic Plains of South and Kazakhstan, 2 insufficient Asia, the no-till package was developed information is available to evaluate Because so many terms have been used through the interaction of researchers these experiences. to describe no-till, it is important at the in national and international Given its potential to reduce rural outset to define what we mean by no- organizations. poverty and increase the sustainability till, to describe some of its benefits, and No-till has already had a major impact of agricultural production, especially discuss the factors that encourage and on agriculture throughout the world, for farmers with few resources, no-till is restrict its adoption. Readers who are but the impact on small-scale particularly relevant to the mission of already familiar with this information agriculture is not always well the International and Wheat may wish to proceed to the next documented or understood. It is estimated that in 2000/01 about 59 2 For information on no-till in Kazakhstan, see CIMMYT and ICARDA (2000). 3 Aside from a literature review, more than 150 interviews and field visits were conducted.

2 2001 CIMMYT World Wheat Overview and Outlook section, which reviews the More specifically, no-till is defined as locations when one principle is not distinguishing features of the no-till planting crops in previously followed. research process (p. 8). unprepared soil by opening a hole, Because of technical and economic narrow slot, trench, or band of the Farmers in developing countries have constraints, packages that combine all smallest width and depth needed to practiced seeding without tillage for three principles are used in only a few obtain proper coverage of the . At centuries. Traditionally, this practice locations. The development of least 30% of the soil surface remains has been associated with slash-and- sustainable no-till packages requires a covered with residues of commercial or burn agriculture, in which farmers strong research effort, especially by cover crops 5 (Wall 1998; Derpsch 1999). slash the vegetation on a plot of land, interdisciplinary teams, which are This definition excludes systems in burn the dry residues, and then seed difficult to build. Also, short-term which residues are burnt and crops are into the uncovered soil. Slash-and-burn economic considerations may prevent planted without tillage (a practice used agriculture results in rapid soil farmers from using appropriate in ancient times and still used in parts degradation and is sustainable only rotations. For example, until the mid- of Mexico, Central America, Asia, and when it is associated with shifting 1990s, the difference between soybean Africa). It also excludes systems in cultivation, in which a plot is farmed and maize prices was more than which seed is broadcast on the surface until its natural fertility declines to the US$ 100/t. Most SouthAmerican and incorporated with complete tillage point where it is no longer worth farmers planted a continuous wheat– (practiced in the Andean Region) and farming it (usually 3-4 years). The soybean rotation, even though they systems in which a small tills the farmer then clears a new plot and knew that the practice increased the soil and plants the seed in the same leaves the degraded plot fallow for a risk of infestation by aggressive weeds. operation (practiced in China and other long period (7-10 years) to regain its The short-term price difference areas of South Asia). natural fertility. Increasing population compensated for the expected pressure has made less land available Although the name refers to only one discounted reductions in yield over for shifting cultivation, forcing farmers practice, no-till actually is a farm the long term. in many developing countries to management system that involves No-till packages differ by location and shorten the fallow period. This highly many agricultural practices, including type of farmer. The most complex no- unsustainable practice leads to severe planting, residue management, weed till package, such as the one used in soil erosion and sometimes to and pest control, harvesting, and South America, is a combination of desertification and famine. rotation. The maximum benefits of no- special machinery (planters, sprayers, till are obtained only if the package In modern no-till, the soil remains and equipment for residue 4 follows the three principles mentioned covered with plant residues (mulch). management), agrochemicals earlier: that the soil is disturbed as little Several names have been used for this (particularly herbicides), and as possible, that the soil is covered by “no-till with mulch” practice: no-till, knowledge. Variations of this package plants or plant residues, and that crops zero tillage, direct planting, direct have been developed in other locations. are rotated. In some places, such as planting on residues, and residue For example, the package used in Brazil, packages that do not follow conservation. Throughout this report, Ghana is basically a weed management these three principles have been found “no-till with mulch” is referred to system that does not use planters but to be unsustainable; it is not clear yet simply as “no-till.” relies heavily on herbicides. In the how sustainability is affected in other irrigated areas of the Indo-Gangetic Plains, where weed management 4 In flood irrigated systems, the importance of mulch in no-till is not clear yet. The opinion of techniques had already been developed most experts is that it should be helpful, but for the time being several technical problems are for several decades, the development of hindering its use (see p. 4). 5 A cover crop is not planted to be harvested but to protect soil from erosion, help control weeds, no-till depended essentially on the and/or eventually increase soil fertility. Commercial crops, in addition to performing some of design of adequate drills (planters) and these functions, are intended for sale. Cover crops are used when no commercial crops are available for a particular growing season. Some cover crops may have an economic value as water management techniques. pastures.

3 2001 CIMMYT World Wheat Overview and Outlook Permanent no-till cannot be used for cycle. Crop residues, if left in the field, one area to another. Even systems crops that require massive soil are chopped and deposited in the known to work in a given area must be disturbance for their harvest, such as furrows between the ridges. This adapted to the conditions on particular potatoes or groundnuts. However, system is used where wet, cold soils at farms in that area. For example, the Brazilian experiences show that planting delay field access for most advanced farmers in South planting these crops with no-till can conventional tillage and reduce plant America adapt their practices to reduce the impact of soil disturbances emergence. Raised, permanent beds individual plots (Ekboir and Parellada (M.F. Ribeiro, pers. comm., 2001). dry out and warm up more quickly, 2000). The components that need to be allowing earlier planting and better adapted are residue management and No-till has been adopted mainly in 6 crop establishment. The furrows also crop rotations; planters; control of nonarid areas, on soils that are not provide drainage for excess water after weeds, diseases, and other pests; and naturally flooded, or on soils where heavy rains, preventing waterlogging liming and fertilization techniques. flood irrigation is not used. In arid (Sayre and Moreno Ramos 1997). areas, where the main objective is water Residue management and crop harvesting, the soils are disturbed to Researchers working with farmers in rotations. Residue management and meet this objective. In naturally flooded irrigated areas of the Yaqui Valley in crop rotations influence soil areas, the main objective is to favor northwestern Mexico have developed a temperature, moisture conservation, drainage and/or evaporation, which system in which 2-3 rows of wheat are erosion, machinery performance, and are hampered by the mulch used in no- planted on top of permanent beds, 70- weed and pest management. Ideally till. In flood-irrigated areas, leveling 90 cm wide and 15-30 cm high. Wheat the soil should be covered with 6-10 t of disturbs the soil, some drills cannot is planted on top of the beds, and the dry matter per hectare per year, but in work properly with loose or excessive furrows between the beds conduct drier areas less than 2.5 t have been residue, and the water moves loose irrigation water. The row orientation of used successfully. Not only the residue to the tail of the plot, the crop facilitates weeding. Other quantity but the distribution and type preventing a uniform mulch cover. crops, like soybeans, have also been of residue are important. Uneven These problems are discussed later in planted on beds. Irrigation distribution results in poorer the case studies of no-till in the Indo- management with furrows is more performance of planters, herbicides, Gangetic Plains and Mexico. efficient and requires less labor than and the plants themselves: because the flood irrigation. Water savings of up to amount of residue in any particular Among the new technologies being 50% have been measured with the bed spot affects soil moisture and developed for flood-irrigated areas, the system compared to traditional flood temperature, of the same crop use of permanent raised beds is one of irrigation. variety will grow differently if residues the most promising (see p. 22). Raised are distributed unevenly. 7 The type of beds have been used in several residue also affects the performance of countries in rainfed areas to plant row- Adapting the technology to no-till technology. Some plants, like crops (e.g., maize or soybeans) but not local conditions soybeans, produce too few residues, wheat. A reduced-till system called As mentioned, no-till technology is which also decompose too fast in ridge-till has been used in the North sensitive to local conditions and subtropical environments, making it Central US since the 1950s. The ridges requires substantial adaptation from difficult to maintain adequate soil (or beds) are re-shaped for every crop cover. Other crops, like irrigated rice, produce too much straw, which hinders 6 No-till has been used successfully in Australia, however, under as little as 200 mm of annual the performance of drills. Part of this rainfall, provided that the soil was not degraded (Patrick Wall, personal communication, 2001). 7 Until recently, most combine harvesters would leave too much residue in the center of the header residue must be eliminated, either by width and too little in the extremes (Derpsch 1999). The latest models have improved straw accelerating its decomposition, burning distribution mechanisms. Strippers for small grains also leave an even cover, since most of the straw is left standing (strippers are harvesters that cut immediately below the spike, leaving most it, or removing it from the field. of the plant standing). Additional advantages of this system are the effective control of wind erosion and fewer problems with planters, because the residues are anchored to the ground.

4 2001 CIMMYT World Wheat Overview and Outlook Residue management depends on compaction mechanism to ensure soil- microorganisms, facilitating integrated environmental and social conditions. It seed contact (usually a wheel) are weed and pest management. Weed is more difficult to produce enough behind the opener. This equipment can control and disease management can residues in dry environments than in be used in areas with similar be enhanced by combining no-till with more humid ones. In many developing landscapes, such as the Bolivian other agricultural practices, such as countries, plant residues are fed to lowlands or the Great Plains of the US. crop rotations and management of livestock, reducing the amount of Even though the landscape in irrigation water. residue available for soil cover. Kazakhstan is also flat and farms are Liming and fertilization techniques. large, Kazakhstan receives half of the In environments where water Liming and fertilization techniques for rainfall received in the Pampas, so seed availability allows, a permanent cover no-till differ from those used in needs to be planted deeper to reach the with a sequence of commercial crops conventional tillage. In no-till, the moisture. Planters for Kazakhstan and short-period cover crops results in products are not incorporated into the could share many of the characteristics fewer weeds than in situations where soil, and the dynamics and interactions of the Argentine equipment, but the only mulch covers the soil between of the soil flora and fauna with the opening devices must penetrate the soil commercial crops. Derpsch (1999) mulch are different. For example, the more deeply to place the seed at the reports that in some plots there was no traditional recommendation for acidic right depth. The drills used in southern need to use herbicides for up to three soils in the tropics was to apply large Brazil have opening devices similar to years. Key to the dissemination of quantities of lime every few years and those used in Argentina but are cover crops in South America was the incorporate it into the soil. Recent trials relatively narrow because they must introduction of the knife roller, which have shown that, since many tropical operate in hilly, uneven terrain. The soil enabled cover crops to be managed soils are permeable, small quantities of openers of planters used in most easily and cheaply. Rotations also lime can be applied every year on the countries are mounted independently reduced weed infestations soil surface. The lime moves naturally to adjust to irregular terrain, but the substantially (Table 1). with rainwater into the soil (Derpsch openers on the drills developed for the 1999). Planters. Planter performance Indo-Gangetic Plains are mounted on a depends on local agricultural practices, bar to reduce the price. These drills soil conditions, and residue perform adequately on plots that have Agronomic, economic, and management practices. In the flat been laser-leveled, a technique also environmental benefits Argentine Pampas, for example, required for efficient water Compared to conventional tillage, no- planters are large and heavy in order management. till has numerous agronomic to cut residues with their weight. The Control of weeds, pests, and diseases. advantages (Beck, Miller, and Hagny planters usually have a cutting disk in No-till creates an environment 1999; Derpsch 1999; Sayre 1998). When front of the soil opening device; the favorable to beneficial insects and combined with residue management, seed and fertilizer ducts and a no-till limits soil erosion. 8 It improves Table 1. Number of weeds per square meter for two soil preparation methods, with water conservation (by increasing and without rotations, Rio Grande do Sul, Brazil infiltration and reducing evaporation) With rotation Without rotation and the condition of the soil (by No-till Conventional tillage No-till Conventional tillage increasing organic matter content, improving soil structure, and Broadleaf weeds in wheat 36 24 102 167 Gramineous weeds in wheat 17 30 41 44 preventing the formation of a plow- Broadleaf weeds in soybeans 4 20 20 71 pan). Because there is no plow-pan, Source: Derpsch (1999). root systems develop better. The mobilization of nutrients in the soil also 8 Although agronomists do not recommend eliminating contour bounds on slopes of more than improves. 15%, farmers in southern Brazil have eliminated them on slopes of up to 50%. The farmers say they have no erosion problems (Landers 1999).

5 2001 CIMMYT World Wheat Overview and Outlook No-till increases flexibility in the timing machinery and labor. In some Taken together, these benefits give a of crop operations. The turnaround instances, yields increase with no-till, good indication of how no-till time between crops is reduced, so crops and in certain areas, three harvests per contributes to the agronomic, can be planted closer to the optimal year become feasible. Production risks economic, and ecological sustainability dates. Another advantage is that crops fall. Production in marginal areas, of cropping systems. can be planted in soils with higher impeded by low moisture levels or moisture levels (farmers do not have to steep slopes, becomes possible. Factors encouraging and wait as long to plant after a rain) or in In several counties, small-scale farmers restricting adoption dry soils before the rains begin. themselves have mentioned that with Several characteristics of no-till As described earlier, no-till may also no-till they get a larger and steadier facilitate its adoption when an facilitate weed control and reduce pest income; the need for family labor falls, adequate package is available. First, the infestations when used with adequate allowing children to attend school; they economic benefits often become rotations. Integrated pest management can pay for children’s education; and evident in the first year, whereas is easier in no-till systems because more the increased income has enabled problems generally do not arise until beneficial insects are present. In flood- families to improve their quality of life. the third year (see below). Second, the irrigated areas the productivity of Finally, no-till has several important sunk costs associated with no-till are water rises as higher yields are environmental benefits. The soil cover small. 10 The specialized components of obtained with less water. Also, reduces erosion and favors water no-till are a planter and knowledge. irrigation becomes easier and faster. infiltration, diminishing the probability Because no-till planters can also be The economic advantages of no-till are of mudslides in hilly areas and used for conventional tillage, the only also numerous. It reduces costs, even reducing the pollution of waterways sunk cost would be the small cost of for small-scale farmers who buy with agrochemicals. Pollution may also converting a conventional planter into herbicides, because they replace be reduced as fewer and less toxic a no-till planter. If a no-till planter 9 expensive labor with herbicides; it agrochemicals are used. Reductions in replaces a conventional planter, there is requires fewer implements as well as agrochemicals, combined with the a cost only if the conventional planter is less tractor power, reducing investment nondisturbance of the soil, favor the replaced before it is completely in and development of beneficial insects. amortized. The other specialized inputs extending the life of tractors; and it Wildlife benefits as shifting cultivation are the farmer’s investments in reduces labor requirements and and pollution decrease. Greenhouse learning about the technology, which simplifies labor management. Reduced gas emissions are greatly reduced by include the farmer’s time, the costs of labor requirements enable small-scale lower consumption of fossil fuels, the acquiring information (mainly through farmers to undertake other income- smaller amount of organic mater that is buying specialized literature and generating activities, such as growing transformed into carbon dioxide, and participating in special events), and horticultural crops or dairying, lower methane emissions in field trials. provided that they can access the nonpuddled rice systems. Increased None of these costs is substantial markets. The physical work involved in fertilizer efficiency reduces the compared with the variable costs no-till is less demanding compared to formation of nitrous oxides, and due to invested every year by a commercial conventional tillage. Larger areas can reduced fertilizer consumption, less farmer, but it is important to note that be planted with the same amount of energy is used in their production. they can be important for small-scale farmers. As discussed later, large-scale 9 For example, by adopting no-till, small-scale farmers in Ghana reduced their maize production farmers, either individually or costs by 32% in normal years and 57% every three years, when a new plot had to be cleared under conventional tillage (Ekboir, Boa, and Dankyi 2001). collectively, often generate their own 10 A sunk cost cannot be recouped if a decision is reversed. For example, as soon as a new car leaves information about no-till, whereas the dealership it loses about 30% of its value. If the new owner finds a buyer just outside the dealership, he has to incur the 30% loss. Sunk costs are usually associated with purchases of specialized equipment and transaction costs.

6 2001 CIMMYT World Wheat Overview and Outlook small-scale farmers often must rely on after they had invested many years in demand reflects low-productivity no-till projects organized and funded working with conventional tillage. livestock enterprises. To date, farmers by other agents. have not adopted no-till with mulch in In switching from conventional tillage areas where residues are used as feed. A third characteristic of no-till that to no-till, the farmer must learn about It has been suggested that farmers may encourages adoption is that the dynamics of a system that is out of solve this problem by replacing draft investment indivisibilities are small. equilibrium and usually takes a long animals with tractors or introducing The latest generation of no-till planters time (more than five years) to reach a 11 more productive technologies for may be expensive, but many countries steady state. In the first two years of livestock (e.g., improved pastures). have active markets for second-hand no-till, the farmer obtains benefits from Because no-till involves a transition equipment. As noted, the technology the new technology. The transition to from a low biomass producing system for converting conventional planters no-till reaches a critical point in the to a high biomass system, in some cases into no-till planters is well known and third year, when factors particular to the additional biomass may sustain relatively inexpensive. Even though each farm (especially the evolution of both no-till and livestock. It is possible, the transformed planter may not be as pest and weed populations) need to be however, that farmers will prefer to use effective as a new one, it can do a addressed. Farmers can, and the additional biomass to expand their sufficiently good job. sometimes do, revert to conventional herds. tillage when they face problems with A fourth advantage is that no-till can no-till that they cannot solve. As mentioned, short-term economic be adopted partially or in stages. In considerations may induce farmers not many cases, farmers have tried no-till As mentioned, no-till requires to use adequate rotations, reducing the on a small area until they have a good substantial adaptation to local system’s long-term sustainability. The command of the package. Finally, conditions. In areas with weak research incidence of weeds, pests, and diseases although the technology is complex, systems, researchers may develop may increase, especially when rotations the immediate economic consequences packages for some regions but not for are not used. of small departures from the optimal others. In semiarid environments, it is package of recommendations are difficult to produce enough biomass to Plant varieties adapted to no-till minor. Farmers can learn about the provide a proper soil cover. No-till may agriculture may not be available; often technology over time. If a farmer does also be difficult to use in areas with seed companies and national breeding not use the best rotation, for example, sandy soils that have a tendency to programs develop plant varieties for the consequences are not felt compaction, especially in the first two traditional production systems and completely in the first year but increase years. In compacted soils, a cultivator more productive cropping gradually over time. or a subsoiler must be used before environments. Areas considered to adopting no-till. In areas covered by have low potential for agriculture What are the main restrictions to the snow during the winter, like the Great under conventional tillage may become adoption and continuous use of no-till? Plains of the US, mulch delays the high potential areas under no-till, but One of the most difficult restrictions to thawing of the soil, shortening an generally new varieties are not overcome is that no-till requires a already short season. In semiarid and developed for these areas until they complete departure from conventional windy areas like Kazakhstan, however, have consolidated as major agricultural farming practices. For years farmers the residues, if left standing, may help regions. Very few public research have been told about the advantages of to capture snow, increasing the water institutions breed plant varieties reducing the soil to a fine powder. available for agriculture in the spring. specifically for no-till conditions (one Many researchers, university exception is Brazil’s national wheat professors, and government officials In many developing countries, residues breeding program, the Empresa also found it difficult to change their are used to feed animals. Since plant Brasileira de Pesquisa Agropecuária– professional practices, particularly residues are low quality feed, this Centro Nacional de Pesquisa de Trigo),

11 There is even the question of whether a steady state exists and if a system ever reaches it.

7 2001 CIMMYT World Wheat Overview and Outlook although specially adapted varieties farmers and equipment manufacturers limitation, however, some researchers may perform better. For example, can generate their own knowledge with from these institutions have played wheat varieties perform very limited interaction with formal research important roles in the development differently when planted on beds or in institutions. Researchers themselves and adoption of no-till. the traditional manner, but breeders can develop no-till packages without The complexity of no-till processes started to develop wheat varieties for necessarily understanding all of the often results in unexpected outcomes bed planting only recently. changes occurring in the agricultural that reflect different perspectives system and the environment. Problems Another potential difficulty in the among those involved in developing that can be addressed only with more adoption of raised beds is that they the technology. For example, formal research approaches (e.g., the require an initial sunk cost for researchers in Argentina and Brazil evolution of weed populations) arise purchasing specialized equipment and found increased soil compaction under only after the no-till system is preparing the beds. Unlike no-till no-till and were reluctant to established. planters, bed planters generally cannot recommend it, but farmers found that be used for conventional tillage. Since research for traditional increased soil density did not reduce Recently, however, an innovative commercial crops has been conducted yields or hamper the operation of manufacturer in India produced a bed for many decades, private firms and planters (Ekboir and Parellada 2000a). planter that could also be used as a no- public research institutions have a There is no accepted explanation for till or conventional drill. wealth of knowledge that farmers can why soil compaction had so little effect. use. The channels for generating and The moisture conserved by the mulch Key Features of transferring information about these and the work of roots and insects may crops are well established. All of the loosen the soil enough to allow planters Research on No-Till agents involved in commercial to penetrate and encourage good crop agriculture know where to search for development (Derpsch 1999). An Before discussing the innovation information. These features allow for alternative explanation is that systems that spurred the development more distant relationships among the researchers conducted their of no-till, it is important to understand agents involved, usually through experiments on small plots where field some of the ways that research on no- markets or printed materials. No-till is traffic was intense, whereas farmers till differs from research on other a new experience, however. The produced crops on large plots, which agricultural technologies. These channels for generating and reduced the impact of machinery differences help to explain why transferring information as well as the movements (Ekboir and Parellada innovation networks are fundamental technology itself have to be created, 2000). This example highlights two to the development of no-till and why and different agents must collaborate major research issues: the importance they have evolved differently from closely (see “Patterns of Innovation,” of identifying the right indicators to networks supporting the development p. 13). evaluate an experiment (for of other kinds of technology. The researchers, it was soil density; for special features of research on no-till Not everyone can or will collaborate, farmers, yields) and choosing the also have implications for formal however. A static organizational setting of the experiment (the research programs interested in perspective, characteristic of many traditional approach is to conduct working on the technology. public research institutions in research on small plots in experiment developing countries, has prevented Unlike some agricultural technologies, stations under tightly controlled many of them from becoming catalysts the development of a basic no-till conditions, whereas no-till was or even joining innovation networks package is not science intensive—even developed through participatory devoted to no-till (see “Public Research a nonscientist can compare two plots, research under commercial conditions Institutions and Flexibility for one cultivated with conventional tillage in farmers’ fields). Innovation,” p. 9). Despite this and the other with no-till. Often

8 2001 CIMMYT World Wheat Overview and Outlook Public Research Institutions and soybeans. Weeds were controlled with Flexibility for Innovation glyphosate. After several years, new pests and more aggressive weeds that were not easily controlled with Most public research and extension institutions in developing countries were common herbicides were identified, created or consolidated in the second half of the twentieth century, when the and the development of appropriate prevailing policy was to attain self-sufficiency in food production. In this weed management practices has environment, researchers and policy makers were concerned mainly with become crucial for the sustainability of productivity rather than with competitiveness or sustainability. The focus on the system. Other issues also require productivity provided a stable framework that lasted several decades and more formal research: the evolution of prompted public agencies to think in terms of individual crops and not whole soil structure and soil compaction, and agricultural systems. Interdisciplinary work was discouraged by an the impact of soil structure on crop organizational structure divided by crop and academic discipline (e.g., yields and evolution of soil flora and breeding, soil science, entomology) and by incentives that placed great value fauna. Most of this research is science on publications in peer-reviewed journals. The financial procedures and the intensive and consequently has to be linear vision of science that prevailed among most researchers and research done by research institutions. administrators discouraged interactions with extension agents, farmers, and others. Finally, loose quality control of the work of individual researchers gave them no incentive (other than personal motivation) to generate new knowledge Innovation Systems or technology. In the mid-1980s, when competitiveness and resource and Technical Change conservation became the most important agricultural policy goals, public research institutions often could not adapt to the new economic and social Technology generation and adoption environment. Their linear vision of scientific innovation (see Figure 1, next have changed substantially in the last page) prevented them from integrating into innovation systems and quarter of the twentieth century with establishing new incentives more appropriate to the new environment. the emergence of complex technologies. 13 While mass production technologies are basically developed by isolated teams of researchers working Traditional and nontraditional research was not obtained with traditional in one institution (usually a laboratory can contribute to the development of statistical methods; sometimes this within a university or firm), complex no-till technology in different ways. perspective slows the development of technologies are developed by Traditional agricultural research is a no-till. networks of agents that co-evolve with long-term process, in which Because the environment evolves in the technologies they generate (Cohen researchers repeat experiments with a response to farming practices, more 1995; Rycroft and Kash 1999). The statistical design and few repetitions formal research may be required to networks usually involve researchers over a number of years before they ensure that no-till practices remain from different institutions along with issue recommendations. Some of this sustainable. For example, until the late users of technology, input suppliers, knowledge can be obtained faster by 1990s, the no-till package used most in government agencies, non- repeating the same experience Argentina was based on the continuous governmental organizations (NGOs), simultaneously in many farmers’ fields double cropping of wheat and and financial institutions. without an experimental design, but other knowledge can be obtained only 12 The issue here is under which conditions valid scientific information can be obtained from through formal research in well- experiments lacking a statistical design. This issue has been largely resolved in sciences that equipped and well-funded cannot perform experiments and rely on nonexperimental data (e.g., economics, astronomy, ecology). The law of large numbers guarantees valid inferences when enough repetitions 12 institutions. Many researchers are are performed. 13 reluctant to accept information that There are many definitions of complexity. In this report it is defined as a system “whose properties are not fully explained by an understanding of its component parts” (Gallacher and Appenzseller 1999).

9 2001 CIMMYT World Wheat Overview and Outlook The key concept for understanding networks, particularly an effective Because of the crucial importance of technology generation in this new interaction between a country’s path-dependence and self-organization, environment is innovation . An scientific base and its business this report carefully examines the early innovation is defined as anything new community (OECD 1999; Rycroft and stages of the development of no-till introduced into a productive or social Kash 1999). packages and the co-evolution of process. This broad definition includes innovation networks and the no-till Two important features of the NIS new technologies as well as packages they generate. framework are: institutional changes (Archibugi, • The evolution of social processes The NIS is larger than the national Howells, and Michie 1999; OECD 1999). results from the interaction of their research system (NRS). The NIS can be The basis for innovation is learning, i.e., history, trends, and random events— strong even though the NRS is weak. In the ability to gather information that is, processes are path-dependent these cases, research institutions are combined with the ability to use that and unpredictable in the long weak, but other agents may actively term. This means that limited information creatively in response to search for new technologies and adapt predictability is possible, but random market opportunities or other social events may derail these predictions. them to the local environment. Italy is needs (Lundvall 1999; OECD 1999). an example of a country with a weak • Processes are not preordained but Knowledge flows and their NRS and a strong NIS, which allowed self-organize through the interaction transformation into innovations of many agents. Even though some it to develop strong competitive depend on the idiosyncratic agents have more clout than others, clusters in the textile industry (Malerba characteristics of knowledge, formal no single agent has the power to 1993). On the other hand, the Soviet and informal regulations (including determine the development path. Union of the 1970s is an example of a laws) that regulate interactions among In the initial stages of path-dependent strong research system and a weak NIS. agents, and the history of each processes, individuals and small events The traditional analysis of technical innovation network. The agents may have a great influence in defining change is based on what has been involved in the innovation process, how the process will evolve—in other called the linear vision of science (Figure their actions and interactions, and the words, in “locking” the process into a 1). In this vision (or framework), formal and informal rules that regulate specific trajectory. After some time, knowledge flows start with basic this system constitute the national however, the trends become stronger science and continue with strategic and innovation system (NIS) (Archibugi, and the ability of individuals to applied research, followed by Howells, and Michie 1999; Nelson and influence the evolution of the process technology development and ending Rosenberg 1993; OECD 1999). 14 declines (Arthur 1994; Geroski 2000). with adoption. This framework is valid An innovation does not have to be new for the world or even for the country in Researchers conducting formal Researchers, Farmers which it is adopted, but only for the research in established institutions extension agencies agent that adopts it. The economic performance of a country or region Basic research Strategic research Applied research Technology development Adoption depends less on the agents that develop Knowledge flow the most advanced technologies than on the innovative activities of the Figure 1. The linear vision of science. majority of agents (in other words, on having many innovative agents) 14 Arthur (1994) studied this feature as an “urn” stochastic process. Assume that you have an urn (Nelson and Rosenberg 1993). The with two balls, one black and the other white. Draw one ball at random and then put it back in dynamics of the NIS depend the urn with another ball of the same color. Before the first draw, the probability of extracting a white ball was 0.5. Before the second draw, the probabilities changed to 2/3 and 1/3, a 33% increasingly on an environment that change. Assume that after 1,000 draws, the probability of extracting a white ball is 478/1,000. fosters the emergence of innovation Before the draw 1,001, the probability of extracting a white ball will be 479/1,000 or 477/1,000, a change of 0.1%. The influence of the first draw on the proportion of balls of each color after 10,000 draws is clearly greater than the influence of the draw 1,001.

10 2001 CIMMYT World Wheat Overview and Outlook only for a limited number of cases; in performance of an innovation network research system (not always present), general, technological developments also depends on its history, the innovative agents (in particular, precede scientific understanding of the complementary assets that the network farmers and equipment underlying phenomena (e.g., the steam needs to acquire, and the environment manufacturers), and an extension engine and thermodynamics) or occur in which the network operates (Rycroft system (either the public system or in production lines through the and Kash 1999). farmers’ associations). transformation of known processes (as Organizational learning is a social Five new learning routines were in many Japanese firms) (IDRC 1997; process that involves several agents introduced by the catalytic agents: Nelson and Rosenberg 1993). (e.g., producers, input suppliers, output participatory research methods, a Unlike the linear vision of science, in buyers, government agencies, research multidisciplinary approach to research, the NIS framework (Figure 2) institutes, researchers, and so forth) acceptance of information generated innovations are the result of complex interacting in evolving networks. without an experimental design, interactions among agents, which Learning routines within an institution creation of a common language that include several feedback loops. These and its ability to exploit acquired enabled communication between interactions can occur at any stage of competitive advantages depend on agents with different backgrounds, and the processes of knowledge generation internal organizational and governance active gathering and open and its application to social processes structures. These routines and dissemination of information. These (IDRC 1997; Nelson and Rosenberg institutional structures evolve jointly in routines required new types of 1993). The NIS resembles a spider web processes that are unique to each interactions among agents, basically more than a linear sequence. institution in each country (Cohen replacing the hierarchical structure 1995; Dosi 1999). arising from the linear vision of science Since agents differ in their innovative with a horizontal structure in which activities and abilities, innovation The core assets of no-till networks have farmers were partners of researchers networks arise because no one can included at least a minimal research and manufacturers. pursue all of the steps required to capability (usually individual develop and commercialize complex researchers from public institutions or The dynamics of innovation networks, innovations. Agents participate in the agrochemical firms), an institutional including the division of labor among network through formal and informal culture that valued innovation and public- and private-sector agents, have arrangements; additionally, networking (especially participatory evolved in the last 30 years because of participation in the network changes research approaches), linkages with changes in the organization of the NIS, often, reflecting changes in the agents’ international sources of information, in the appropriability of innovations, objectives and evolving technological and at least one institution (with and in the role of the government as an challenges. sufficient resources and geographic economic agent. Most research coverage) willing to play a catalytic role institutions in developing countries The performance of innovation in the emergence of the network. This have had difficulty adjusting to the networks depends on their core last factor is crucial. As will be seen, new environment because they were, capabilities , internalized complementary many individual researchers and and many still are, organized in a way assets, and organizational learning farmers experimented with no-till in that reflects a linear vision of science. routines . Core capabilities are those several countries, but widespread These institutions sought to maximize aspects of innovation in which a adoption occurred only when an academic output without regard for particular network excels. Internalized institution took the leading role. future uses; in other words, their complementary assets are the resources objective was to develop research that the network can use to innovate. The complementary assets of no-till outputs and then display them in the Organizational learning is the process networks have included agents with window for someone to take. The by which new capabilities and assets strong personalities who could consequence was the establishment of are acquired or discarded. The organize local networks, a formal weak information flows between

11 2001 CIMMYT World Wheat Overview and Outlook Access to Knowledge critical inputs supply Regulatory Economic environment Incentives environment

Capacity to Capacity to detect market detect opportunities technological opportunities

Multinational organizations

National Innovation International Open System (NIS) economy competitiveness Potential economic Available development technology Global Creativity Identify supply Innovation technology Knowledge System (GIS) trends on the market NISs Innovation

Innovation systems Development Local International National Centers International

Commodities Innovative capacity Interaction Organizations Strategic Market Social needs Companies advantages Research institutes Universities

Individual actions

Effects of Globalization Dissemination of knowledge and technology Comparative advantages

Incorporating Irregular flow of R&D concentrated Integration of technological information and in countries (United production systems Export to the developments knowledge from States, Japan, and in countries with Research Development Production rest of the country to country Europe) specific geographical world advantages Figure 2. National innovation systems. Source: Adapted from Ekboir and parellada (2000b).

12 2001 CIMMYT World Wheat Overview and Outlook researchers and users, which resulted Patterns of Innovation in the accumulation of unused technologies and lack of social support for research. The evolution of technologies can be classified as fitting three patterns (Rycroft and Kash 1999). Interactions between researchers and extension agents were also shaped by Normal (also known as incremental changes): the co-evolution of an the linear vision of science, in which established network and a technology along familiar technological standards. extension provided little input to In this pattern there is relatively low technological and economic risk. The research. Moreover, their interactions interaction mechanisms are known and relatively stable. The innovations are were hierarchical: the researcher, being incremental in the sense that they represent minor changes of standard the holder of knowledge, educated the practices. An example of this type of evolution is the replacement of a modern extension agent. This type of crop variety by another modern variety—a small improvement of an otherwise relationship was then recreated in the unchanged agricultural package. The agents who generate the changes do not interactions between the extension need to interact closely with the users or other agents because they know the agent and the farmers. rest of the package and the technological requirements. Most interactions, as well as dissemination of the innovations, are indirect, mediated by markets or Within the NIS framework, on the printed material. other hand, the emphasis rests on the multiple interactions between Transition: the co-evolutionary movement by an established network and researchers, input suppliers, extension technology to a new evolutionary path. The technological risk is greater than in agents, and users of the technology. the previous case. The innovation may be introduced into existing production These interactions lead to the processes, leaving the final product without change. Through these changes, establishment of participatory research the innovation network remains relatively unaffected but the production programs; farmer-to-farmer extension process goes through a substantial change. An example of this type of programs with strong support from evolution was the introduction of soybeans into Brazilian cropping systems in researchers, extension agents, and the 1970s. Agriculture was intensified although most of the agricultural input suppliers; and more horizontal package remained unchanged (see p. 15). flows of information within and Transformation (also known as revolutionary changes): the launching of a between institutions (see “Patterns of new development path by a new network and a new technology. The Innovation,” this page). technologies represent a major departure from conventional practices and involve major commercial and technological risks. Because the technical Methodological standards and market opportunities are not well understood, developers have Considerations to interact closely with other members of the network to reduce risk and to obtain all the knowledge required to develop the innovation. Biotechnology is This report does not follow the an example of this pattern. The creation and marketing of biotechnology traditional approaches for the analysis products required the creation of new and more complex networks whose key of technology dissemination for two players included scientists as well as input suppliers, lawyers, government reasons. First, no satisfactory indicators regulators, the media, and non-governmental organizations (NGOs). of innovation diffusion have been Eventually, the transition and transformation patterns evolve into a normal developed because innovation is a pattern. broad concept that includes tangible As will be seen later, no-till is an unusual example of a transformation pattern. and intangible outputs as well as small New networks of agents arose to develop knowledge and new products organizational changes (Nelson and (planters and sprayers) that were integrated into new technological packages Rosenberg 1993). Traditional research that, in turn, were used to produce an unchanged agricultural product (initially and development (R&D) indicators grain, but currently vegetable, pasture, and perennial crops).

13 2001 CIMMYT World Wheat Overview and Outlook measure only formal research activities. more than 150 semi-structured agronomists and mechanical They do not include resources interviews and field visits in Argentina, engineers; it also assessed the technical invested, for example, in design Bolivia, Brazil, Ghana, India, Mexico, and economic potential of the market improvement or in enhancing the Pakistan, Paraguay, the UK, and the for paraquat in several countries. effectiveness of networks. Second, US. The interviews included early Following these studies, ICI time-series data on adoption of no-till developers and adopters, farmers, established a research team in are incomplete and unreliable, researchers, agrochemical companies, Australia in the late 1960s and in Brazil preventing an econometric analysis. and planter manufacturers. in 1972. 16 ICI’s involvement in the US was limited because it had licensed This report also departs from paraquat to Chevron Chemicals and CIMMYT’s usual approach in technical Early Development of could not develop a market for itself. research publications in that it No-Till emphasizes the roles of individual In 1960, researchers in Virginia, USA, researchers and unforeseen events. The next sections of this report consist used paraquat to control bluegrass sod, Three reasons justify this approach. of six case studies of innovation atrazine for residual control, and 2,4-D First, as explained in the previous networks that developed no-till for post-planting cleanup. The section, individuals and chance packages for small-scale farmers. These experiments were soon repeated in strongly influence the evolution of a cases highlight the interactions other states. The University of process such as the development of no- between catalytic agents, institutions, Kentucky created a strong research till in its early stages. Second, catalytic chance events, and history in shaping program on no-till, led by S. Phillips. agents (i.e., institutions and often the networks. This section provides a The first use of no-till in commercial individuals that play key historical perspective on those case production was reported in Kentucky organizational roles) are an essential studies by describing the first efforts to in 1962. The first commercial no-till component of no-till networks. Third, develop modern no-till practices, planter was produced in 1966. To take in their very early stages, new which resulted in packages appropriate advantage of the fact that no-till technologies are generally not for large-scale farmers. enabled farmers to plant immediately developed by research institutions but after harvest, the double cropping of Modern development of no-till started by individuals within those wheat and soybeans was introduced in after the discovery of the herbicide institutions. Only after the technology 1966 (Ekboir and Parellada 2000a). paraquat in 1955 and its commercial has shown some potential do release in 1961 by the British company Research on no-till at Massey institutions organize research ICI. For centuries it had been assumed University in New Zealand started in programs. 15 Understanding the that tillage was necessary to improve 1967 (Ritchie, Baker, and Hamilton- interactions between the catalytic water infiltration and control weeds. Manns 2001) . This program would agents, institutions, and technologies Now that weeds could be controlled later be important in the development from the early stages of development chemically, ICI funded research to find of no-till in the Indo-Gangetic Plains to maturity is essential to the design of out if cultivation was still necessary. (p. 23). successful no-till programs for small- After the first promising results, ICI scale farmers. Two small groups of researchers from realized that a completely new the National Institute for Agricultural Information on the no-till development technological package was needed to Technology (Instituto Nacional de processes described in this report was create a market for paraquat. ICI Tecnología Agropecuaria, INTA) and a gathered through an extensive created in-house research capabilities group of large-scale farmers started literature review, complemented by on agricultural systems by hiring research on no-till in Argentina in the early 1970s. Initially each team worked 15 For example, despite of two of its centers being pioneers in the development of no-till, separately but followed the others’ EMBRAPA’s central office recognized no-till as a major agricultural technology only in the late 1990s, when the area under no-till in Brazil was about 9 million hectares. advances. Adoption of no-till 16 The development of no-till in Brazil for small-scale farmers is reviewed in one of the case studies later in this report.

14 2001 CIMMYT World Wheat Overview and Outlook proceeded slowly, because weed relatively slowly compared to South present particular problems that control with paraquat was difficult and America. In the US, a major restriction prevent them from benefiting directly development of adequate machinery on the expansion of no-till was that the from no-till packages developed for was lagging. In the late 1970s, these mulch delayed thawing of the soil in the rainfed conditions. Considerable groups devised no-till packages based spring, shortening an already short research is still needed to develop on glyphosate, a herbicide developed cropping season. Also, agricultural appropriate no-till packages for flood by Monsanto. Even though glyphosate policies promoted a range of irrigated conditions; some of this greatly simplified weed control, its conservation practices (including research will be described in the cases price made the package economically minimum till and no-till), which presented in this report. unfeasible. In the early 1980s, about 15 delayed adoption of no-till alone. Even though there are no reliable people (including researchers, The development of an adequate no-till statistics, in 2000/01 no-till was extension agents, and farmers) from the package was technically more difficult estimated to be used on 59 million three groups started to meet in Australia than in South America. hectares around the world, mainly by periodically to discuss their progress. Paraquat could not control some of the large-scale, commercial farmers In 1987 a Monsanto salesperson local weeds (e.g., rye grass). (Table 2). convinced this group to create an Agricultural practices that combined association of no-till farmers. 17 livestock with crop production caused Monsanto also progressively reduced Case 1: The Brazilian two problems: unless the pastures were the price of glyphosate; a liter cost properly managed, the animals Experience about US$ 40 in the 1970s less than US$ increased soil compaction, and it was Early research 10 in the early 1990s. By this time, difficult to control some of the pastures machine manufacturers had developed Starting in the 1960s and for the next with the available herbicides. appropriate planters and sprayers, and three decades, the Brazilian with the lower price of glyphosate, the Areas where flood irrigation is used government encouraged an expansion package became technically and of the agricultural frontier towards the Table 2. Area (ha) under no-till in different economically efficient. The association countries, 2000/01 southwest, center-west, and north. of no-till farmers was very good at Agricultural practices also became Country Area communicating with farmers; in five more intense as soybeans, as a single USA 21,120,000 years, the area under no-till surged crop or in rotation with wheat, replaced Brazil 13,470,000 from 200,000 to 7 million hectares Argentina 9,250,000 livestock and coffee production. These (Ekboir and Parellada 2000a). Canada 4,080,000 changes, combined with the heavy Australia 8,640,000 rains, hilly landscape, and conventional Development of no-till in the US and Paraguay 960,000 tillage, led to serious soil erosion, and Australia followed a more traditional Bolivia 350,000 public research and extension pattern than in Argentina, where Chile 100,000 Venezuela 150,000 institutions recommended that farmers individual researchers from public- Colombia 70,000 switch to livestock production. Some sector institutions were instrumental to Ghana 50,000 farmers, wishing to avoid the economic developing the technology. In contrast, Mexico 50,000 Uruguay 50,000 losses involved in moving to livestock the US and Australia developed no-till Other 1,000,000 production, began experimenting with packages mostly through formal Total 59,290,000 reduced tillage. research programs in established Source: Data for all countries and “other” from Derpsch (2001), institutions, with strong participation except for Ghana (Ekboir, Boa, and Dankyi 2001), India In the late 1960s, the German (Peter Hobbs, personal communication 2001), Mexico by farmers and input suppliers. (from CIMMYT estimates based on fieldwork, and cooperation agency GTZ (Deutsche Adoption of no-till grew steadily but Pakistan (Peter Hobbs, personal communication 2001). Gesellschaft für Technische Zusammenarbeit) hired Rolf Derpsch to work on a newly established 17 Monsanto paid for all startup costs and for operational costs in the first year.

15 2001 CIMMYT World Wheat Overview and Outlook research project in Londrina, Paraná to lack of a network delayed the spread of terraces and prepared a chart that increase soybean yields; the project was no-till, however. The no-till package related the distance between terraces to based in IPEAME. 18 Derpsch began was still in its infancy, especially the the slope of the terrain. When farmer testing alternative practices, including machinery and weed control Manoel Pereira’s loan request was an early version of no-till. Encouraged practices. 20 The planters that were rejected in 1976 because, according to by the results, he teamed up with available were inadequate for the chart, his land was regarded as too Herbet Bartz, a farmer who was willing Londrina’s heavy soils. Bartz continued steep for terracing, 22 he sought the to test the technology on his farm. to collaborate with Derpsch and ICI advice of an agronomist who thought but, as he was uninterested in that no-till might increase the distance When ICI transferred its no-till promoting no-till, a network to between terraces and enable the use of research team from Australia to disseminate the package did not machinery. The first no-till trial Londrina in 1972, the team established emerge. succeeded. A few neighbors also tried strong relationships with the agents no-till with help from ICI. After three working on no-till: Derpsch; a few In 1974, ICI imported a no-till planter years, these farmers felt they needed to researchers, mainly from IAPAR, the and lent it to research teams, seek specialized advice abroad. Pereira Agronomy Institute of Paraná (O equipment factories, and farmers. and Franke Dijkstra visited the Instituto Agronômico do Paraná), and Small workshops started to adapt the University of Kentucky’s no-till the Wheat Center of the Brazilian planter to local conditions with support program. Pereira and Dijkstra had very Agricultural Research Corporation from EMBRAPA-CNPT and ICI. ICI strong community ties and were very (Empresa Brasileira de Pesquisa mobilized a large sales force to promote active in their farming cooperatives. Agropecuária–Centro Nacional de no-till among pioneer farmers. Formal Upon their return, they encouraged the Pesquisa de Trigo, EMBRAPA-CNPT); research was concentrated in two 19 creation of the Earthworm Club to a planter factory (Semeato); and institutions, IAPAR and EMBRAPA- exchange no-till experiences. Many pioneer farmers. ICI became the hub of CNPT. All other research and extension members of their cooperative were a system that developed the first no-till institutionss, including the universities, small-scale farmers. Pereira and package for Brazil. The keys to its opposed the technology until adoption Dijkstra were instrumental in success were its strong development was widespread. convincing three neighboring team, its participatory and The severe erosion problems forced cooperatives to organize an extension multidisciplinary research strategy, and many farmers to default on their loans. program for them. Eventually, the a strong sales force that promoted no- In the early 1970s the manager of the Earthworm Club and this program till. The new learning routines induced Castro (Paraná) branch of the Banco do evolved into the ABC Foundation, 23 a major changes in formal research 21 Brasil convened a meeting of research and extension institution methods by fostering collaboration researchers and extension agents to funded by the three cooperatives. An among researchers, equipment find a technical solution to this active collaboration between the manufacturers, and farmers. problem. Following the guidelines of cooperatives and University of In 1972, Bartz traveled to the UK and the US Soil Conservation Service, they Kentucky emerged from the visit to the the US with the help of ICI to learn recommended the construction of US. Professors and graduate students more about no-till. Back in Brazil the next year, forced by the loss of his crop to sell all his machinery except a no-till 18 This center would later become EMBRAPA’s Soybean Research Center. planter acquired overseas (it had only 19 ICI contacted several manufacturers of agricultural machinery, but most were not interested in a scrap value), Bartz had to plant his new technology for which they did not see a market. The owner of Semeato, a small factory at that time, was also a farmer interested in conservation tillage. He immediately agreed to farm with no-till—the first large-scale participate in the development of planters for Brazil. Eventually Semeato became the largest manufacturer of no-till planters outside the US. use of the technology in South 20 In certain years Bartz manually weeded over 600 ha of soybeans. America. Technical difficulties and the 21 The Banco do Brasil was then the most important lender to the agricultural sector. 22 According to the chart, the terraces should be one meter apart. 23 In 1998 the ABC Foundation had a budget of about US$ 1 million.

16 2001 CIMMYT World Wheat Overview and Outlook from Kentucky often visited Brazil, and During these years, Derpsch interacted whole state. This success induced other many of the cooperative’s professionals actively with agrochemical companies, agents to join the program; in the third received training in the US. individual farmers, and farmers’ year, partners included seven private associations. He organized a research companies, three public research and The introduction of glyphosate and program on cover crops as well as no- teaching institutions, the extension continuous research by several agents till associations. service, local planning offices, (public-sector researchers, farmers’ cooperatives, and municipal organizations, agrochemical authorities. The project had a budget of companies, and machinery factories) No-till for smallholders about US$ 400,000 for the first three produced an efficient no-till package by Until the 1990s, small-scale farmers’ years. the end of the 1970s. The main hurdle adoption of no-till was limited. They to adoption was the high price of The project’s most relevant activities could use the same crop rotations, glyphosate. At this time, the governor were the adjustment of cover crop herbicides, and cover crops as large- of Paraná barred IAPAR from management practices to regional scale farmers but lacked technology for researching no-till on the grounds that circumstances, the validation of low- planting by hand or with animals. it was a technology for large-scale cost kits to enable small-scale farmers Three major projects (METAS, Paraná farmers, promoted by a multinational to convert conventional planters to no- Rural, and Land Management Project company. IAPAR was to concentrate on till planters, the adjustment of II) in the 1990s fostered adoption of no- technologies suitable for small-scale recommendations for liming and till by small-scale farmers by farmers. Though these directives were fertilizer management, the developing adequate planters and eventually reversed, IAPAR never improvement of chemical spraying training extension agents. regained its relevance as a major player technology, the training of extension in the development of no-till. The METAS project involved several agents, the implementation of Training public- and private-sector agents. In Demonstration Units in farmers’ fields, After the success of the Earthworm 1990, a researcher from EMBRAPA- and the establishment of a research Club, small groups called Clubes CNPT and a Monsanto technician Demonstration Farm (Denardin and Amigos da Terra (Friends of the Land investigated the causes of limited Kochhann 1999). Public- and private- Clubs) were created in the southern adoption of no-till by small-scale sector partners in METAS shared states with support from farmers and farmers in Rio Grande do Sul. They human, material, and financial agrochemical companies. These small identified three factors: the need to resources to coordinate activities and groups were a successful innovation. adapt the package to local conditions, reduce risk. The EMBRAPA centers and Periodic meetings and stable lack of adequate planters for small- the Universidade Federal de Pelotas membership fostered trust and allowed scale farmers, and insufficient were responsible for conducting farmers to discuss their technical command of the package by extension research and training extension agents. problems openly. agents. Following this diagnosis, in The state extension service and private Because weed control was easier with 1993 Monsanto promoted a project that extension agents worked with glyphosate than paraquat, Monsanto supported research as well as cooperatives to install Training eventually captured most of the extension. The project involved five Demonstration Units. These units herbicide market for no-till. Despite public and private institutions, selected served first as a means of self-training ICI’s huge investment in creating a to cover the spectrum of problems to for extension agents and later became market for paraquat, a first-class be solved. In the first year of the demonstration plots for farmers and research team, and an innovation project, the area under no-till jumped local communities. Additionally, the network, no-till was a commercial from 45,000 ha to 150,000 ha. By 1997, extension agents provided feedback to failure, and ICI cut all no-till activities the area under no-till had reached the research teams. The private in the late 1980s. 820,000 ha in the project area (90% of companies financed part of the costs, the target) and 2,200,000 ha in the

17 2001 CIMMYT World Wheat Overview and Outlook contributed with their own research, enacted by the state government was Case 2: The Bolivian and interacted with researchers and an important tool for extension agents Experience extension agents. to convince the most reluctant farmers to adopt soil conservation technologies. Crop production in Bolivia is Before the METAS project, planters The project also included credit lines concentrated in two environments: the were relatively large and expensive for the poorest farmers. A parallel eastern lowlands (150-700 masl) and (above US$ 10,000). Many of the repair project was implemented by IAPAR, the Andean Valleys (1,000-3,600 masl). shops that improved the low-cost kits the extension service, the state Most small-scale farmers are found in for adapting conventional tillage government, and the Brazilian the Andean Valleys. In 2000, an equipment to no-till became Association of No-till Farmers. In this estimated 350,000 ha in the lowlands manufacturers of cheaper seeders, project, 31 planters for small-scale were planted with no-till. Adoption of offering prices as low as US$ 7,000. farmers were bought by the state and no-till in the valleys remains minimal, Semeato was induced to reduce the placed in 31 communities for the since a basic package is still being size of its equipment and lower its farmers to experiment with the developed. prices to meet small-scale farmers’ technology. Extension agents and requirements. farmers were trained in the use No-till in the lowlands In Londrina, technical problems with of no-till. no-till had been solved by the late The lowlands were opened to In the state of Santa Caterina, the Land 1980s, but adoption was still confined agriculture in the 1980s. Large-scale Management Project II also used the to large-scale farmers. In 1988 a group and mid-size farmers predominate, micro-catchment as the planning unit. of state technicians formulated the generally growing summer soybeans In 1984 three pilot projects were set up project Paraná Rural. With financial with wheat, sunflower, or sorghum to develop soil conservation and other support from the World Bank, the sown in the winter in a double crop sustainable practices, building on project (implemented between 1989 system. Rainfall averages 1,300 mm previous research and farmer and 1996) eventually involved 210,000 and is concentrated in the summer experiences with cover crops. That farmers on 7.1 million hectares. By months; wheat is planted in May into same year, watershed commissions in combining micro-catchment planning residual moisture after the summer the project areas were created. The first with planning on individual farms, the soybean crop is harvested. experiences of no-till with animal- project marked the evolution from a A few farmers developed no-till drawn planters occurred in 1987. The purely physical approach to soil practices in the 1980s without support machinery and the rest of the no-till conservation (use of barriers) to an from formal research institutions. In package were improved in the integrated soil management approach 1993, the Association of Oilseed following years. Massive adoption by in which no-till was the key component Producers (Asociación Nacional de farmers started after 1991; by 1999 it (Landers 1999). The Project Productores de Oleaginosas, ANAPO) was estimated that over 80% of the Coordinating Unit and the Regional asked CIMMYT to post an agronomist state’s grain area was planted with no- and Municipal Soil Commissions used to eastern Bolivia at Santa Cruz de la till (Landers 1999). Technical support decentralized, participatory Sierra to improve wheat technologies for this project was provided by approaches for implementation. The and the profitability of the soybean- EPAGRI, the state research and project supported farmers’ associations wheat rotation (no-till was not extension enterprise. The project was and group purchases of equipment. considered as a specific instrument.) followed in 1998 by the Pro-Palha Training programs for implementing Patrick Wall, a CIMMYT wheat Project, a partnership between seven institutions and officials ensured that agronomist, was posted from Paraguay agribusiness firms, EPAGRI, and two consistent information was given to to Bolivia in 1994 and started to cooperatives. farmers. Soil conservation legislation promote no-till.

18 2001 CIMMYT World Wheat Overview and Outlook No-till, combined with crop rotation, agrochemical distributor, was critical to experiments had a fairly sophisticated has a strong positive impact on wheat the development of a weed- design to derive statistically valid production. Wheat yields increase management package. The inferences. Second, farmer groups, field about 25 kg/ha for each additional agrochemical company provided new days, bulletins, and events were millimeter of moisture in the soil. In information and conducted some established. Third, public-sector the lowlands, in a dry year like 1998, research with farmers. researchers were encouraged to work plots seeded with conventional or on no-till. Public research was severely vertical tillage yielded 450 kg/ha of challenged in 1998, when the Bolivian Unlike other South American wheat. Plots seeded with no-till Institute of Agricultural Technology countries, in Bolivia the medium- and without crop rotations yielded about (Instituto Boliviano de Technología small-scale farmers in the lowlands 600 kg/ha. When no-till was combined Agropecuaria, IBTA) was closed and have been more interested than large- with rotations, yields of 900 kg/ha the Center for Tropical Agricultural scale farmers in collaborating with were obtained (Patrick Wall, personal 25 Research (Centro de Investigación researchers. Small-scale farmers communication, 2001). Agrícola Tropical, CIAT), a state usually own 50 ha, of which 15-20 are Small-scale farmers did not adopt no- research center in Santa Cruz de la sown to grain crops. Owing to past till because suitable planters were not Sierra, lost most of its funding. To subsidies, most small-scale farmers are available; on areas under 150 ha, sustain wheat research, the National mechanized. Efforts to develop a no-till mechanized no-till was less profitable Wheat Program (PROTRIGO), package assumed that small-scale than traditional practices (Paz 1998). incorporating CIAT and ex-IBTA farmers were eager to use their Large-scale farmers adopted the researchers, ANAPO, three NGOs, and tractors, but when Wall invited Magin technology to save on production CIMMYT, under the Ministry of Meza 26 and Marcos Peñalva to give costs—especially diesel—and because Agriculture, was initiated with funding talks to small farmers in the lowlands, they could extend the planting period from the US and the European they showed pictures of animal-pulled from about 3 days after the first rains Community. This program financed planters that interested small-scale to about 12 days. The short seeding much no-till research and promotion farmers because they would need no period prevented the emergence of a between 1998 and 2001. Formal bank credit to use the equipment. Soon, market for custom services to serve agricultural research in the lowlands is in collaboration with ANAPO, Wall smallholders. now conducted by CIAT and CIMMYT held a series of field days on animal- working with ANAPO. CIAT soil drawn equipment (Brazilian and Over the years, two specialized no-till scientists have been more active in locally made). networks emerged; they worked promoting vertical tillage researched separately but followed each other’s with a previous British technical aid progress. CIMMYT organized a No-till in the highlands mission, whereas CIAT wheat network with public research agronomists, CIMMYT, and ANAPO In the Andean Valleys, where farmers institutions, farmers’ associations, and favor no-till. usually own 1-5 ha, farm sizes have progressive farmers to deal with been falling with successive agronomic issues in wheat production Drawing on experience gained in generations. Most farmers combine and develop appropriate machinery Paraguay (see p. 31), Wall established livestock with crop production. During for small-scale farmers. A second large demonstration plots in farmers’ the winter, the animals are heavily network, formed by Fundacruz, 24 fields with farmer collaboration; these agrochemical companies, and farmers, specialized in weed management. 24 Fundacruz is a farmer-funded research foundation that follows the model of the Brazilian ABC Carlito Los, an agronomist with the Foundation. 25 ABC Foundation (Brazil) who was Owners of large-scale farms are usually business people with little involvement in agriculture. Managers of these farms are interested in collaborating but have limited decision-making power hired by the main Bolivian about the use farm resources. 26 Magin Meza is a Paraguayan extension agent with vast experience in developing no-till packages for small-scale farmers.

19 2001 CIMMYT World Wheat Overview and Outlook dependent on crop residues (Wall et al. Mayor de San Simón in Cochabamba, involved through courses on no-till and 2000). Rainfall in the valleys varies although an FAO project which printed materials. Farmer adoption between 300 and 700 mm/yr and is terminated two years ago did establish remains minimal owing to the lack of concentrated in the summer months some no-till work using Brazilian drills. machinery and of sufficient residues (November-April). FAO has also conducted a training for no-till as well as livestock. Soil course in the Tarija Department. cover is difficult to maintain because In the more fertile regions or in areas fields are open for communal grazing where summer rainfall exceeds 500 The Universidad Mayor, together with in late June and, as wheat is harvested mm, farmers grow wheat, maize, PROTRIGO and CIMMYT, has by hand and threshed away from the potatoes, broad beans ( Vicia faba ), and developed an animal-pulled, no-till drill field, farmers do not want to incur the peas (often harvested as dried peas) in for small grains in collaboration with expense and trouble of returning the different rotations. In the less fertile or the Silsoe Research Institute (SRI) (a straw to the field. The communal drier regions (less than 500 mm British research institute) and several grazing problem could be solved by rainfall), wheat or barley monoculture, NGOs that support the diffusion of the controlling the sheep, making the broken by occasional weedy fallows, is technology. The development of the residues unpalatable, or introducing common. After each crop, plots were drill was partly funded by project pastures in fallowed plots. Given that left fallow for two years, but this managed by SRI and funded by the UK no-till increases the amount of residues practice is being abandoned as farm Department for International produced, adopters could potentially sizes decline. Most crops are hand Development, in which the Universidad have enough residues to maintain their harvested, and the small grain cereals Mayor manufactures the drills and the herds and a proper soil cover. are threshed away from the fields. After project purchases them at a cost of the threshing, the straw is stored for about US$ 330 per unit. Another The introduction of combines could animal feed. development by the CIMMYT project also increase the feasibility of no-till. was the manufacture of a human- or Recently, one NGO started offering In 1994 IBTA and CIMMYT surveyed animal-drawn sprayer, based on a custom harvesting with a combine, and farmers in the Andean Valleys and Brazilian model. Several machines, farmers immediately realized the found that the two main constraints to based on this initial model, are now savings of money, time, and effort that wheat production were drought stress being produced by the Universidad. could be achieved. 27 Since demand for and soil erosion (mainly hydric). These Before this local drill was developed, the combine is large, the NGO gives two problems had received little small-scale farmers had to rely on priority to the best fields, providing an attention because researchers assumed Brazilian equipment, which was too incentive for farmers to adopt a more that drought stress could not be solved large. As noted, they could not resort to intensive production package. One by technical means. To demonstrate custom planting for wheat because of byproduct of mechanical harvesting is that cover crops could have an impact the short planting season. that the straw remains on the soil. Even on soil moisture management, Wall if it is later grazed by sheep, enough imported a few Brazilian no-till Until 2001, few farmers participated in straw remains to provide cover. planters. researchers’ on-farm trials. With the purchase of 10 drills in 2001, small-scale Until its closure, IBTA was CIMMYT’s farmers, NGOs, and a municipal most important research partner. After government became more actively 1998, PROTRIGO funded and coordinated most research and extension activities; the latter were contracted to NGOs and farmers’ 28 With combine harvesting, farmers can remove wheat from the field in a few hours compared associations. The only other major with the three months usually required by manual harvesting, threshing, and winnowing. In research partner is the Universidad addition, the mechanically harvested wheat has higher quality because it is not mixed with stones and manure.

20 2001 CIMMYT World Wheat Overview and Outlook Case 3: The Mexican created erosion problems in the hilly the Colegio de Postgraduados (CP), a Experience Mexican landscape, and research on graduate teaching institution, also conservation practices was initiated. In supported research and extension Mexico’s research and extension 1973, CIMMYT started its own research related to no-till. Most of the research institutions have been organized along and training program on no-till for was done for masters’ and doctoral scientific disciplines or crops, reflecting maize. CIMMYT researchers, trainees, theses, in particular on soil topics. The the linear vision of science discussed and visiting scientists (including CP’s relationship with farmers has been earlier. Their structure has tended to be trainees and scientists from Mexico) top-down and the research has had relatively rigid and hierarchical. conducted research on experiment little impact in the field. Collaboration among scientists, even stations and in farmers’ fields around This specialization of tasks reflects the from the same institution, has been the world (Soza, Violic, and Haag 1998). lack of interaction between the three infrequent, researchers have had little In the late 1970s, CIMMYT organized a institutions: since no package had been say in the programs designed by their study tour to the US by technicians and developed for the different Mexican directors (Ekboir et al., forthcoming), researchers from Fideicomisos environments and with little support and they have worked in small plots on Institutidos en Relación con la from INIFAP or the CP, FIRA had to do experiment stations, seldom interacting Agricultura (FIRA) 29 and INIFAP’s its own research. At the same time, with extension agents and farmers. predecessor organization to see neither INIFAP nor the CP had “Finished technologies” were passed to conservation agriculture. Upon their extension programs. Despite their extension agents who would then return, these institutions started no-till similar programs and several transfer them to the farmers, usually as programs in their respective domains: institutional agreements to develop part of a program that included INIFAP in research and FIRA in joint projects, interaction between subsidies. The technologies to be promotion (Claveran 2000). researchers from these institutions promoted were selected by the public FIRA initiated a no-till program that remains limited. officials in the capital city who trained farmers, built specialized designed the programs in a top-down Others involved in no-till in Mexico training centers, established a large manner with little participation of have included NGOs, private network of demonstration plots, farmers or local governments (Kondo companies, and farmers themselves. A promoted conferences and farmers’ 1999, van Nieuwkoop et al. 1992). network partially financed by the organizations, and, foremost, financed Farmers began to expect monetary Rockefeller Foundation includes equipment purchases (González 1990). incentives to be attached to extension universities, research institutions, state INIFAP conducted a few isolated trials programs. governments, and other NGOs in the 1970s and 1980s. Research (Claveran 2000). A few private Trials on no-till without mulch were accelerated at the beginning of the companies organized no-till extension conducted in the 1950s by the 1990s, and in 1996 INIFAP created programs that failed because their predecessor organizations of the CENAPROS, a center devoted to traditional structure discouraged Instituto de Investigaciones Forestales, sustainable agriculture which efforts to develop an innovation Agrícolas y Pecuarias (INIFAP), coordinates all institutional activities network and a package adapted to Mexico’s national agricultural research related to no-till. In the late 1990s, 55 local conditions. For example, John organization, but this line of research researchers participated in INIFAP’s Deere promoted groups of no-till was soon abandoned due to national program on conservation farmers, but its main goal was to sell disappointing results (Claveran 2000). tillage (Claveran 2000). From the 1980s, machinery imported from the US. When the success of the Green Revolution in South Asia 28 led the

Mexican government to promote a 28 The Green Revolution package included high-yielding, input-responsive wheat and rice varieties, similar package in the 1970s, the improvements in irrigation, and increased fertilizer use. The Indian government promoted this package aggressively with important subsidies for inputs and outputs. resulting intensification of agriculture 29 FIRA is a set of trust funds administered by the Banco de México with a budget of about US$ 2 billion. FIRA combines extension and financing of agriculture. No-till is one of several programs.

21 2001 CIMMYT World Wheat Overview and Outlook Similarly, although Monsanto valuable agronomic information was of the Yaqui Valley in northwestern promoted no-till, it did not support produced, the program (organized as Mexico (Sayre and Moreno Ramos development of a package. Technical traditional on-station and on-farm 1997). 30 Trials initiated in 1961 by R.J. difficulties have now led Monsanto to research) had little impact on Mexican Laird, a soil scientist and agronomist, emphasize minimum tillage (seen as a agriculture. In 1999 it was reorganized showed that wheat yields were not precursor to no-till). to incorporate participatory affected over a large range of spacings methodologies and relocated to the (from 17 to 70 cm), indicating that it Farmers have also been exploring no- Bajío area in central Mexico; it is too was feasible to modify planting till options. In certain areas of the soon to evaluate this new stage. patterns. In 1968, O.H. Moreno Ramos southern state of Chiapas, for example, joined the research team in the Yaqui farmers used to plant in communal Declining water tables and falling prices Valley. Over the next decade, he plots in a system that included long are forcing farmers in Mexico’s irrigated conducted a series of experiments that fallows. These very steep areas cannot areas to seek alternatives. Partial laid the basis of the bed planting be planted with conventional adoption of no-till, such as its use for system, particularly planting on beds techniques. After titling of the land in only one crop in a two-cycle annual rather than flat land and using furrow the early 1990s induced an pattern or no-till without soil cover, has rather than flood irrigation. A intensification of agriculture and been documented in several areas concerted effort was made to abandonment of the fallow, farmers (Erenstein and Cadena 1997). Recent familiarize Yaqui Valley farmers with developed a no-till system based on the fieldwork in the Bajío indicated that this technology. In 1981 only 6% of use of herbicides and manual planting farmers are aware of conservation farmers were planting on beds; at (Mauricio Bellon, personal tillage technologies. More than 60% of present over 90% use the system. Bed communication, 2001). The origin of large-scale farmers had heard about the planting reduces costs, reduces water the package is not known, as no technology compared to 36% of small- use, and permits farmers to combine a institution has claimed credit for its scale farmers. All farmers, however, had pre-seeding irrigation with mechanical development. insufficient knowledge of the package cultivation as a weed control strategy, (Erenstein 1999). Adoption in the Prototype planters for small-scale reducing herbicide use (Sayre and rainfed areas has been negligible, but in farmers have been developed by Moreno Ramos 1997). The advantages irrigated areas more than half of the INIFAP and the Universidad identified by farmers include higher farmers use conservation tillage in the Autónoma de Chapingo, but they have yields, reduced costs, easier weed spring crop, followed by heavy tillage not been tested by farmers nor reached control, and reduced turnaround time for the winter crop. Obstacles to the full commercial production. A few Mexican between wheat and a summer crop. adoption of no-till in the Bajío are: 1) manufacturers and two multinational plant residues have a market as animal Most farmers practicing bed planting companies (John Deere and New feed, 2) it is difficult for individual burned the crop residues, tilled the soil Holland) produce no-till planters and farmers to keep their land covered, (destroying the beds), and formed new drills for small-scale farmers (Claveran, given the common practices of burning beds to sow wheat after the summer 2000). Demand for this equipment is residues and communal grazing, 3) crops. In the early 1990s, Kenneth small because of design problems and residue management practices are not Sayre, a CIMMYT wheat agronomist, because the rest of the package has not available, and 4) neither are adequate started working with farmers to been developed. planters. Along with its research on no- develop permanent beds, which would In the 1990s, CIMMYT and the Centre till for maize, CIMMYT worked with eliminate the cost of forming beds de Coopération Internationale en farmers to develop new planting every year and yield the benefits of no- Recherche Agronomique pour le techniques for wheat in irrigated areas till. Production costs fall as plowing is Développement (CIRAD), the French cooperation agency, had a joint no-till 30 The Yaqui Valley, in the northern state of Sonora, was the center of the wheat breeding program program in the state of Jalisco. Though initiated by Norman Borlaug and colleagues under a Government of Mexico-Rockefeller Foundation collaboration in the mid-1940s.

22 2001 CIMMYT World Wheat Overview and Outlook reduced. Permanent beds also reduce network failed to emerge, largely irrigation or only as a rainfed crop. the loss of sediment in the irrigation owing to the top-down organization of Wheat is grown under similar tail because these beds are more stable research and extension and the linear circumstances, although most wheat than beds established after vision of science that hampered inter- fields receive at least one initial conventional tillage prior to irrigation. institutional collaboration. These irrigation. Traditionally, rice is planted Crop residues in the furrows help shortcomings affected not only public in damp soils after thorough plowing, stabilize the soil and reduce soil institutions but also CIMMYT and which negatively affects soil properties erosion by water (Sayre and Moreno private firms. Bed planting, on the (see “Problems in Rice-Wheat Ramos 1997). The controlled traffic of other hand, was developed because of Systems,” p. 24). Wheat is then planted field equipment (the wheels pass only the early emergence and persistence of in these degraded soils. The rice-wheat over the furrows) reduces soil a strong localized research program, system dominates the irrigated areas, compaction. 31 The technology is still in the strong innovative attitude of whereas a larger variety of cropping the early stages, however, especially farmers, their need to cut costs, and patterns are found in the rainfed areas the development of no-till bed planters their close collaboration with CIMMYT in response to different environmental that can plant through residues and of and research institutions in the 1990s. conditions (Kataki, Hobbs, and wheat varieties adapted to bed Adhikary 2001). 32 planting. Case 4: The Indo- There is no agreement on the level of Gangetic Plains No-till wheat in the Indo- adoption of no-till in Mexico. FIRA Experience Gangetic Plains recently estimated the area under no- Only one basic no-till package for till at about 650,000 ha (Ochoa 1999), Because of its many environments and wheat is available for the region. based on the number of no-till drills rainfall patterns, the Indo-Gangetic Though substantial efforts are being delivered through several promotional Plains region in South Asia (Map 1) made to develop new planting programs. Recent CIMMYT fieldwork presents a formidable challenge for the techniques for rice, these are still in shows that those programs development of permanent no-till their infancy (see “No-Till for Rice,” p. emphasized “hardware” over packages—i.e., mutually compatible 25). Research on no-till for other crops “software”: the machines were no-till technologies for the most is conducted only by isolated delivered but farmers were not trained common crops, especially the rice- researchers and innovative farmers. to regulate them or to manage wheat sequence. Soil and hydrological Efforts to develop no-till packages for residues. Consequently, most farmers situations differ throughout the Indo- rainfed conditions have been weak. use the no-till drill for conventional Gangetic Plains. Rainfall is more tillage. limited in the west, where the most The apparently low adoption of no-till productive areas are irrigated, whereas in Mexico reflects the inability of all the lowlands of the eastern Indo- Nepal agents working on the topic to Gangetic Plains are naturally flooded Pakistan integrate into an innovation network. for periods of variable duration. Rice is Development efforts remained planted in the wet, warm season and isolated. Until recently, there was no wheat in the dry, cool season. Rice INDIA attempt (except in the Yaqui Valley) to grows in areas receiving as little as 120 Bangladesh involve users and local equipment mm of rain and in areas with more manufacturers. A catalytic agent for a than 1,800 mm, and under full

31 In another major irrigated area (El Bajío, in central Mexico) farmers successfully experimented with wide beds of 2-3 m (Bernard Triomphe, personal communication, 2001). Map 1. Indo-Gangetic Plains (dark area 32 Preliminary trials show strong variability in the response of wheat varieties to bed planting. on map).

23 2001 CIMMYT World Wheat Overview and Outlook Problems in Rice-Wheat Systems

A number of problems affect rice-wheat systems in the turnaround time is caused by excessive tillage, soil Indo-Gangetic Plains of South Asia. Support prices for rice moisture problems (too much or too little moisture), power and wheat and input subsidies (for agrochemicals, constraints for plowing, or by the need to thresh and electricity, and water) make the rice-wheat rotation the handle the rice before plowing the soil for wheat (Hobbs most profitable and least risky option for farmers while and Giri 1998; Hobbs 2001). inducing excessive input use (Pingali and Shah 2001). The repeated plowings to get a good seedbed for wheat These policies have hampered the development of markets increase costs. Currently, tillage represents about one-third for other crops; since alternative crops present more of the total production cost. In addition, excessive tillage market risk than rice or wheat, farmers are less likely to and poor soil structure create problems for wheat during grow them. the first irrigation. When water remains standing in the Another problem is that the soil is managed in completely field because of poor infiltration, oxygen stress delays different ways for rice and wheat. Traditionally rice is wheat growth (Hobbs and Giri 1998; Hobbs 2001). produced by plowing saturated soils (i.e., “puddling”) Continuous rice-wheat cultivation has led pests and (some dry-seeded, nonpuddled rice is found, however, diseases to build up. The major weed affecting wheat is mainly in rainfed areas). Puddling reduces water loss and Phalaris minor , which was normally controlled with the controls weeds in rice, but it leads to deterioration of the herbicide Isoproturon, but in some places this weed has soil’s physical structure, formation of a hard pan at a developed resistance to the herbicide (Malik, Gill, and shallow depth, poor water infiltration, waterlogging, and Hobbs 1998). poor root development. Physically degraded soils create problems for the wheat crop, because farmers have to Water is also becoming a serious problem. In some areas of make many passes with a plow to prepare an acceptable northwestern India and Pakistan, water tables are seedbed. Tilling with animal traction is more prevalent in declining rapidly as water is pumped faster than it is the eastern areas where yields are lower. Commonly, replenished. In other areas, water tables are rising and farmers in these areas make six to ten passes with a local leading to waterlogging and salinity. plow (a wooden stick with a metal point), followed by the Transplanting rice is a demanding, labor-intensive same number of plankings (a heavy wooden log that technique. Increasing labor shortages during major compacts and breaks clods). Tractor power is rapidly agricultural operations (planting, harvesting, and replacing animal power and is the major power source in threshing) have reduced the profitability of puddled rice the higher production areas of the western Plains. and fostered mechanization. The use of tractors and power Although tractors enable faster and more thorough land tillers in Thailand increased by a factor of 6-7 between 1978 preparation, farmers still make six to eight passes with a and 1993, and similar trends have occurred in India, nine-tine cultivator or disk harrow, usually followed by Vietnam, Malaysia, Indonesia, and Bangladesh (Bakker et planking, to prepare the land for wheat (Hobbs and Giri al. 2000). Throughout the eastern Indo-Gangetic Plains, 1998). small-scale mechanization is increasing, and in the Late planting is another problem; yield potential declines regional as a whole custom operations are more common by 1-1.5% per day when planting occurs after 20 (Bakker et al. 2000; Hobbs 2001). November. The three major causes of late wheat planting are delays in planting rice due to drought, late harvest of * Basmati rice is a long-duration crop that commands a high price because of its particular aromatic characteristics. In some Indian the previous rice crop, or long turnaround time from rice states, like Haryana, more than 40% of the area under the rice-wheat harvest to wheat planting. Late harvest results from the system is planted to basmati varieties (Mehla et al. 2000). After * liberalization of the Indian rice market in the late 1990s, basmati rice use of long-duration rice varieties, late rice planting, or has expanded into the states of Ghaziabad and western Uttar Pradesh the production of a short-duration crop after rice. The long at the expense of sugarcane. Basmati rice is also grown in the Punjab of Pakistan.

24 2001 CIMMYT World Wheat Overview and Outlook Surface seeding is the simplest no-till tractor, constituting a reduced tillage India and Pakistan, where generally system. This common practice for system rather than a no-till package. In four-wheel tractors are used to prepare wheat establishment in parts of India the lower, flood-prone areas of the land. The equipment used is a tractor- and Bangladesh has been introduced eastern Indo-Gangetic Plains, this mounted seed and fertilizer drill with into the Terai (lowlands) of Nepal. system results in the highest wheat inverted-T openers that place seed Wheat seed is broadcast either before yields because it enables farmers to directly into the standing rice residues or after rice is harvested. The key to seed at least one week before no-till without any land preparation (Hobbs success is the correct management of and almost one month before and Giri 1998). Adaptation of this soil moisture. Too little moisture results conventional tillage. equipment for use with two-wheel in poor germination and too much hand tractors has started recently (Peter Mechanized no-till has been tested in moisture causes seed to rot (Hobbs and Hobbs, personal communication 2001). Pakistan, India, and Nepal (Hobbs and Giri 1998; Hobbs 2001). In some areas The smaller equipment is better suited Giri 1998). This package is more (e.g., Bangladesh), seed is incorporated to eastern areas where fields are small relevant to the higher yielding, more with a rotavator pulled by a two-wheel and fragmented. mechanized areas of northwestern

No-Till for Rice

Research on no-till for rice started in Australia in the address the lack of information on the technology, refine 1950s. Hood (1961) and Boerema (1965) showed that rice technologies developed in Australia and the US, analyze could be sown directly into a pasture and yield differing results and lack of follow-up from earlier comparably to rice sown after conventional cultivation. research at IRRI, and identify systems that could save Weeds were controlled and pasture was prevented from water and labor (Piggin et al. 2000). re-growing through flooding. No herbicides were used. Preliminary results at IRRI suggest that a satisfactory In the US, methods of crop establishment, weed control, lowland rice crop can be established using no-till seeders and water and cover crop management for no-till mounted on hand tractors. Careful crop establishment irrigated rice were investigated in the 1980s and 1990s in and management are needed to avoid problems with rice Louisiana (Bollich 1991, 1997) and California (Williams et seedling emergence and survival. Effective weed control al. 1992). Yields of the best treatments were consistently can be obtained with glyphosate and adequate water above 9 t/ha. Farmers in Australia and Florida establish management. Post-emergence weed control (either about 10% of the rice area using no-till. manual or with a combination of herbicides) is critical. The economic advantage of no-till rice relative to Research on no-till for irrigated rice in Brazil started in the conventionally tilled rice depends on the relative prices of 1970s. By the 1990s, irrigated no-till rice had expanded herbicides, labor, and fuel. Wet seeding requires 4-5 tillage strongly in the southern states. In 1993, an estimated operations, seedling propagation, and transplanting. No- 270,000 ha (33% of the rice area) in Brazil’s largest rice- till requires one spraying and one sowing operation. At producing state, Rio Grande do Sul, was planted with no- 2000 values, costs were comparable (Piggin et al. 2000). till irrigated technologies thanks to satisfactory weed control, an improvement of the cost/benefit relationship, As with most early research on no-till, the main and a better integration of crop and animal farm knowledge gaps in the package for irrigated rice are production (Sousa et al. 1994). related to the development of appropriate machinery, weed control, crop establishment, and crop management. Despite these advances, little has been done for small- A related line of research is the development of practices scale rice growers such as those in South Asia. In the late for growing nonpuddled rice in plots where puddling is 1990s, the International Rice Research Institute (IRRI) not needed (i.e., heavy, poorly drained, lowland, and salt- established a long-term study on no-till irrigated rice to affected soils) (Bakker et al. 2000).

25 2001 CIMMYT World Wheat Overview and Outlook Nonmechanized no-till has become applied 4-5 days earlier and took only to maintain bullocks all year on the possible since animal-pulled planters 8-10 hours, costing about Rs 475/ha farm. Often farmers replace bullocks have been produced locally. These (Hobbs and Giri 1998). Savings per with water buffaloes to obtain implements are still too new to show hectare with no-till reach up to 1 additional income from selling their any impact. million liters of irrigation water and milk. In Haryana in 2001, 70% of about 60 liters of diesel (Melha et al. farmers that adopted no-till did not Bed planting, described earlier, is 2000). No-till has the potential to save own a tractor and used custom being tested in the region. A local 6-10 plowing operations, reducing costs services; additionally, 40% of the manufacturer in the Punjab of India from Rs 825 with conventional tillage to adopters were small landholders with has built a prototype low-cost planter Rs 125 (Hobbs and Giri 1998). By farms smaller than 2 ha (R.K. Malik, that plants three rows of wheat on 70 substantially reducing turnaround time personal communication 2001). cm beds at the same time that the beds between rice harvest and wheat are formed (Hobbs and Giri 1998; planting, no-till increases wheat yields. Hobbs 2001). Constraints to the adoption No-till has proven very effective in of no-till controlling weeds in wheat because Benefits of no-till in the Indo- The main constraints to farmers’ germination of most weeds is triggered Gangetic Plains adoption of no-till for wheat are by light or by lower temperatures. imperfect drill performance and some The problems affecting rice-wheat Since the soil is disturbed less with no- shortcomings in the no-till package. systems in the Indo-Gangetic Plains till, less weed seed is exposed and Additional factors are lack of have no single solution. Appropriate germinates (Hobbs and Giri 1998). knowledge and information about no- economic and agricultural policies can Also, through timely planting, wheat till and skepticism among extension foster diversification of crops and can emerge early and shade out weeds. and scientists about no-till. techniques (Kataki, Hobbs, and Farmers can use the savings obtained Adhikary 2001; Pingali and Shah 2001). by reducing the number of tillage Issues related to the no-till drill. Cost New technologies that reduce input operations to buy the new, more does not appear to be a constraint to use (at current prices), such as no-till, effective but expensive herbicides and the adoption of the drill. 33 More nonpuddled rice, and planting on improve weed control (Malik, Gill, and important factors behind the limited permanent beds, can also help in the Hobbs 1998). Recent data suggest that demand seem to be technical transition to a more sustainable no-till reduces weed infestations over limitations and current extension agriculture. time, and eventually no herbicides are policies. Cost savings are substantial with no- required in some seasons (Peter Hobbs, The drills have several technical till. In a trial with wheat, five personal communication 2001). limitations. First, the openers are fixed irrigations, each at a depth of 8-10 cm, Thanks to the availability of custom to a bar and cannot follow the contour were applied under conventional machinery services, small-scale farmers of the soil. If the field has not been tillage, whereas only 4 were needed have been able to use no-till. By leveled, some of the seed is not placed under no-till. Under conventional adopting no-till, small-scale farmers at the proper depth. Second, the drill tillage, on average, between 26 and 34 obtain two additional benefits: their lacks a compacting mechanism behind hours are required to apply two operating capital requirements fall the openers, so in heavier or damp soil irrigations at pre-sowing and another because they need to contract fewer some seeds do not make good contact at the crown root stage (21 days after tractor hours, and they no longer need with the soil. Some farmers overcome seeding), costing 1,800 Indian rupees (Rs) per hectare. With no-till, farmers 33 did not apply the pre-sowing Drills currently cost about US$ 300 in India and US$ 600 in Pakistan. (The difference results from different input prices in both countries and slight design dissimilarities.) During the fieldwork for irrigation; the first irrigation was this report, farmers who already owned equipment mentioned that they would buy a better but more expensive drill. Nor does the price of the drill seem to be a deterrent for small-scale farmers who do not invest in machinery and rely on custom services.

26 2001 CIMMYT World Wheat Overview and Outlook the problem by planking or by subsidies are needed to promote new active in the region but do not irrigating after seeding. 34 Third, the technologies, and farmers expect to participate actively in the development inverted-T openers do not perform receive them. The state governments of of no-till packages, a point that will be well if there is loose straw because they Haryana and Punjab, India, subsidize discussed in greater detail later. Weed work as a rake. 35 Fourth, if the drill is up to 30% of the price of the drill; management could be improved not properly calibrated, up to 15% of farmers in other states are waiting for through integrated weed management the seed is broken. Finally, most similar subsidies. strategies that include rotations with farmers find it difficult to regulate other crops; researchers at the Regional In recent years, demand for no-till fertilizer and seed flows in the drills. Research Station of Haryana drills has increased faster than supply, Agricultural University in Karnal The last two problems have been to the point that manufacturers cannot found that introducing sugarcane after solved in the latest version of the drill cope with demand. Several new the rice-wheat sequence greatly developed by a manufacturer in factories have started producing drills, reduced weeds in the subsequent rice Amristar, India, but other but the quality of these new entrants is crop. manufacturers have not yet adopted uneven. In the medium-term this the changes. In a recent survey should not be a problem since the Three problems prevent good residue conducted in Punjab by the National market will eventually weed out the management. First, livestock compete Agricultural Research Center (NARC) lowest quality manufacturers. with no-till for the use of residues. of the Pakistan Agricultural Research Second, the drill is less efficient with Issues related to the package. The main Council (PARC), 70% of farmers the loose residues left on the soil after problems of the package currently in pointed out operational problems with combine harvesters have been used. 37 use are weed management (especially the drills. Even though some of the The regional importance of this for rice), residue management, a lack of problems could result from problem will become more acute as profitable rotations, insufficient inappropriate training, the field combine harvesting increases in command of the package by local research for this report confirmed the response to labor shortages. Currently extension agents and researchers, and design shortcomings. Farmers mention farmers remove or burn all loose insufficient training of farmers. that obtaining customer support from residues. Although extension agents manufacturers and finding local The lack of an effective weed discourage this practice, there are no technicians to repair the drills are also management package, especially for alternatives except to incorporate the major problems. As the equipment rice, is a major barrier to the adoption residue into the soil. Third, as 36 industry develops, these issues should of permanent no-till. Weed mentioned previously, irrigated rice be resolved. management should integrate produces large amounts of residue that, alternative herbicides with mechanical, when left in the field, hinder Extension policies in India and cultural, and agronomic practices. The germination of the subsequent crop. Pakistan make the drill available herbicides currently available are not Since wheat has to be planted as soon without cost to farmers in targeted very effective against several weed as possible after rice, farmers cannot villages for limited periods. Although species, especially those that compete wait for the rice straw to decompose. the drills are lent for three years at with rice. Agrochemical companies are most to each village, farmers expect to use them for the foreseeable future. Some farmers indicated that they 34 Apparently, in heavy soil closing the slit opened by the drill reduces germination (Peter Hobbs, would buy a drill if they could not get personal communication 2001). However, during the field visits farmers complained about the one for free. Also, for almost four open slits. 35 Loose straw results from combine harvesting, land preparation, or grazing of the standing rice decades, government officials, stubble after harvest. 36 The problem is more relevant for direct seeded, nonpuddled rice than for transplanted rice on researchers, and extension agents in beds. both countries have assumed that 37 Researchers are developing equipment that can operate with loose residues, including chopping and spreading the straw as it leaves the combine or after, and a drill that can plant into residues (Peter Hobbs, personal communication 2001).

27 2001 CIMMYT World Wheat Overview and Outlook Economic policies that make the rice- were not trained to operate the drill. In establishment. The International wheat sequence the most profitable the field interviews conducted for this Center for Research in the Semi-Arid option have hindered the introduction report, many farmers complained about Tropics (ICRISAT) has also joined the of alternative rotations (Pingali and difficulty in calibrating the drill, network by providing improved seed Shah 2001). These policies started to whereas more experienced users found of several leguminous crops, mainly change in India in 2000, and farmers no problem. Similar training chickpea. Although researchers and are actively trying alternatives. Most deficiencies were observed with respect manufacturers from both countries researchers, on the other hand, are still to herbicide use. interact with colleagues and farmers focused on the rice-wheat system. One within the country, knowledge of of the major hurdles for diversification Development of no-till developments occurring in is the lack of markets for alternative for wheat in the neighboring countries is limited. The crops. Even though markets could Indo-Gangetic Plains RWC has organized several traveling develop on their own, adequate workshops to address this deficiency, policies could ease and accelerate the The development of no-till packages in but their impact has been limited. transition. India and Pakistan followed different Knowledge of developments outside paths. In India, a few researchers at the Indo-Gangetic Plains is restricted to Another difficulty is that many state universities conducted research, a handful of researchers, mainly from researchers and extension agents in the but except in Haryana and Punjab, international organizations. region still do not interact effectively. extension did not join these efforts. In Extension agents learn how to operate Another issue is that most researchers most states, the Rice-Wheat Consortium the drill but not about other have accepted no-till only recently. In a (RWC) for the Indo-Gangetic Plains components of the package, such as 1996 workshop in Haryana on long- coordinated research and extension for weed control. For example, at a state term soil fertility issues in the region, no-till. 38 In Pakistan, research was university a farm machinery professor several soil and crop management conducted initially by NARC in who demonstrates no-till to farmers techniques were discussed, but no-till partnership with CIMMYT, whereas mentioned that, even though he knew did not feature among the extension was conducted almost weeds could be controlled chemically, recommendations (Abrol et al. 1997). exclusively by the On-Farm Water that was the expertise of weed Adoption of no-till by farmers in the Management Wing (OFWM) of the specialists. For fields with severe weed Indo-Gangetic Plains is also a recent Department of Agriculture, infestations he recommended and localized phenomenon. Even in Government of Punjab. In both mechanical control with a rotavator 1998, Hobbs and Giri (1998) reported countries, interactions between research instead of no-till. that adoption was scant. It is estimated and extension agencies are limited. that the wheat area planted with no-till There are indications that farmers do First CIMMYT and later the RWC in 2001 will reach 50,000 ha in not receive proper training in using the provided important research support. Haryana, 10,000 in Uttar Pradesh, and drill. In the survey conducted by In 1999 the RWC started an important 5,000 ha in the Punjab of India. NARC, about 20% of farmers who had program in India and Pakistan to study Adoption in other Indian states has bought a no-till drill still plowed the alternative methods for rice been more limited. Adoption in the field before sowing, and more than half Punjab of Pakistan was estimated at 30,000 ha for 2000/01. 38 The Rice-Wheat Consortium (RWC) was established in 1994 as an Ecoregional Initiative of the Consultative Group on International Agricultural Research (CGIAR), involving the national Pakistan. In the 1970s, Massey agricultural research systems of South Asia, international agricultural research centers, and University (New Zealand) established advanced research institutions. Funding for the RWC comes from many sources, including the participating countries; the Governments of Australia, the Netherlands, UK, and the US; the a collaborative program with Pakistan, International Fund for Agricultural Development, the Asian Development Bank, and the World Bank. The activities of the RWC are decided on and approved by the steering committee chaired through which several Pakistani by the heads of the four national agricultural research systems on a rotating basis. CIMMYT agronomists pursued doctoral studies currently acts as the convening center of the Consortium on behalf of the steering committee and the CGIAR.

28 2001 CIMMYT World Wheat Overview and Outlook related to no-till. Among them was but the extension service lacked funds though NARC supports the research Ashraf Choudhary, who visited to keep the drills in working order, and done by OFWM, the research effort is Pakistan in 1982 to promote no-till they deteriorated rapidly. The limited. Extension agents conduct on- under a UNDP-funded assignment. He demonstrations failed and the farm trials with active support from the brought literature and two inverted-T technology was not recommended. RWC and more limited interactions openers developed in New Zealand by After Hobbs moved to Nepal in 1988 with NARC. Aitchison Industries, which were (where he assumed regional The Government of New Zealand passed on to Peter Hobbs, a CIMMYT responsibilities, including work with continued to support the diffusion of wheat agronomist based in Pakistan. the RWC) and Choudhary shifted the no-till in Pakistan in the 1980s and The openers were mounted on a frame focus of his work to Iran in 1991, all 1990s, financing several graduate and tested at the Agricultural promotion of no-till in Pakistan came to students and, more recently, research a halt. Machinery Division of NARC. and dissemination programs. Two Interactions with NARC were good During the mid-1990s, the OFWM in other important sources of funds for because the center had just been Punjab had also tried no-till, but the research on no-till were DFID and created and the researchers were eager results were not convincing. In 1997, CIMMYT. to cooperate. Choudhary invited Mushtaq Gill, The OFWM has supported the creation Hobbs used funding from the US OFWM Director General, to visit of an association of no-till farmers, but Agency for International Development Massey University and see no-till in its presence in the field is small. The (USAID) to circumvent traditional operation. In 1998 Gill participated in a Pakistani association has not organized public-sector regulations and make the traveling seminar organized by the a network like those in South America RWC that made him confident about network more agile. In 1984 he to generate and disseminate imported an Aitchison drill into the potential of no-till. Gill realized that information. Pakistan with financial support from no-till could save substantial amounts USAID and tried it with NARC. of water and instructed his staff to India. Research on no-till for wheat in Assuming that farmers would not promote it. OFWM had a large network India started almost three decades ago abandon their late-maturing rice of extension agents that interacted with (Aslam et al. 1993; Brar et al. 1983; varieties, researchers focused on irrigation associations at the village Verma and Srivastava 1989, 1994; reducing turnaround time so that level; this contact with farmers enabled Verma, Srivastava, and Verma 1988). wheat could be planted earlier. The OFWM to establish a large number of Several state agricultural universities first no-till work occurred in farmers’ demonstrations. In 1998, Hobbs was tried no-till in the 1970s but their efforts fields. Even though farmers were introduced to Gill, and they started were marred by technical problems. convinced by the results, the New collaborating on a package for Lacking adequate planting equipment, Zealand drill was still too big and too Pakistani conditions. OFWM collected they planted seed manually; chemical expensive for Pakistan. the old no-till drills and started to weed control was also difficult. This promote no-till aggressively. The line of research was soon abandoned From his base in New Zealand, OFWM in collaboration with the RWC by all except a handful of researchers Choudhary continued to promote no- contacted local drill manufacturers and working in isolation. till in Pakistan with NARC and PARC convinced them to start production of between 1985 and 1991. In 1988, Hobbs In 1990, Hobbs introduced the the drill that had been developed in obtained additional funds from USAID inverted-T openers to Indian India. As noted, collaboration with to make 20 copies of the Aitchison drill, researchers. In 1991, engineers at G.B. local extension and research agencies which were used by NARC used until Pant University of Agriculture and remained weak (although Choudhary the early 1990s. NARC researchers Technology, Pantnagar, modified the continued to provide assistance). Even convinced the extension services to traditional rabi (cool season) seed drill conduct a few no-till demonstrations, by replacing the old seed coulters with

29 2001 CIMMYT World Wheat Overview and Outlook inverted-T openers (Hobbs and Giri Recently, the University of Adelaide forthcoming). Because small-scale 1998; Melha et al. 2000). The (Australia) started a program with farmers in Ghana traditionally researchers contacted several local Haryana Agricultural University to practiced bare no-till, the no-till shops to produce the drill control Phalaris minor ,further develop package developed for Ghana commercially, but none were interested no-till, and foster adoption. consisted only of weed and stubble in starting a new line of production for management practices. The similarity In Punjab, three factors have delayed which they saw no market. Eventually of the traditional and no-till packages adoption. First, researchers and they found a manufacturer in Punjab favored adoption while the lack of extension agents often gave conflicting who was willing to start production. appropriate planters has prevented advice. Second, farmers planted rice mechanized farmers from adopting In 1993, S.S. Dhillon of Punjab that matured more rapidly than the this technology. Agricultural University participated in varieties grown in Haryana, so the a wheat agronomy course managed by benefits of no-till were smaller. Third, The earliest research on no-till in Ken Sayre at CIMMYT headquarters in in Punjab, recommendations have to be Ghana started in the late 1960s (Ofori Mexico, where he became acquainted approved by the state university before 1973) and was continued in the 1970s with bed planting. Upon his return to they can be taken to farmers. This by local researchers who interacted India, he tried this technology on meant a delay in getting farmers to little with other agents, including wheat with the objective of reducing experiment and become convinced agrochemical companies and foreign production costs. To his surprise, about the technology. In 2001 the state researchers. Their findings showed the infestations by Phalaris minor were government introduced a US$ 70 beneficial effects of no-till on soil and greatly reduced, and the incidence of subsidy for the purchase of no-till water conservation (Mensah-Bonsu this weed decreased over the years as drills. and Obeng 1977). no-till was used. Agrochemical companies collaborate In the 1990s, no-till research was In 1994, R.K. Malik attended the same with the state universities, but not with concentrated in the Crops Research course in Mexico. Malik, a weed the RWC. In the late 1990s Monsanto Institute (CRI) in Kumasi and the scientist at Haryana Agricultural supported no-till research at the Ghana Grains Development Project University, was also an advisor to the universities but soon reduced its (GGDP). 39 Roberto Soza, a CIMMYT university’s Vice Chancellor. Malik was activities: the potential market for agronomist working with the GGDP, able to explain the benefits of no-till glyphosate was small because it could organized no-till research from 1990 and, through his position, managed to not control Phalaris minor efficiently. until 1996. In 1991, the GGDP adopted get the university to promote it. To Even though other herbicides are used the no-till system for planting maize overcome bureaucratic barriers, the in the region, agrochemical companies and grain legumes on five research RWC bought several drills and have not played a major role in stations. The system was tested donated them to the university, which developing no-till. extensively in farmers’ fields across the contributed extension agents and their country. expenses. The rapid adoption of no-till Case 5: The Ghanaian In 1993, Sasakawa-Global 2000 (SG in Haryana was the consequence of a 40 2000) and Monsanto joined the concerted effort by state researchers Experience GGDP and the extension service from and extension agents, the state subsidy In 2000, an estimated 100,000 Ghanaian the Ministry of Food and Agriculture for purchasing drills, and farmers’ farmers used no-till on 45,000 ha to promote no-till. The program agricultural practices. Since a large (Ekboir, Boa, and Dankyi, received strong political support from proportion of farmers grew long- duration Basmati rice, wheat was usually planted late. The benefits of no- 39 The Ghana Grains Development Project involved the Crop Research Institute, the Grains till came not only from cost reductions and Legumes Development Board, the Ministry of Food and Agriculture, CIMMYT, and the International Institute of Tropical Agriculture (IITA). It was funded by the Government of but also from higher wheat yields. Ghana and the Canadian International Development Agency. 40 Sasakawa-Global 2000 is an NGO financed by the Government of Japan.

30 2001 CIMMYT World Wheat Overview and Outlook the government (Findlay and demonstrations, from site selection developing planters and mechanized Hutchinson 1999). The results obtained through herbicide application to sprayers because of the weak research by the researchers were used to train harvesting. Demonstration plots were effort and because most farmers in the extension agents, who carried out situated at strategic points—along central Ghana did not use planters. As their own demonstrations in farmers’ major footpaths and roads linking a result, adoption of no-till in Ghana plots. Monsanto helped CRI evaluate villages/towns as well as on highways has been restricted to small-scale the efficacy of powder glyphosate in (Findlay and Hutchinson 1999). farmers, while large-scale mechanized small packages for no-till on farmers’ farmers continue with conventional The diffusion program established fields with maize and beans. Trial tillage. links with the Sasakawa Center at the protocols and guidelines were University of Cape Coast, making it established by Monsanto and discussed possible for agricultural students in Case 6: The with CRI researchers, who their final year to view farmers’ no-till implemented the trials with financial Paraguayan Experience fields at the end of their soil support from SG 2000. Even though the conservation course. The Paraguayan agricultural sector powder presentation was more traditionally has been characterized by expensive per unit of active ingredient From an institutional point of view, the the prevalence of small-scale, than the liquid formulation, farmers research effort (measured by the subsistence farmers, who occupy only initially preferred it because it was the number of researchers and public 6% of the land but produce 35% of the exact amount they needed for one resources committed) was quite agricultural output (Sorenson et al. backpack sprayer. After farmers limited. Even in 2000, public research 1998). In the late 1970s, the government became familiar with the no-till institutions had no formal programs on opened forested areas in eastern package, they started to demand no-till. In contrast, the few researchers Paraguay to agriculture and promoted glyphosate in larger, cheaper actually doing the research were the settlement of immigrant farmers presentations. extremely motivated and innovative. The overall research program may and small-scale farmers from This program established have been weak, but a strong extension Paraguay’s Central Region. Most of the demonstration plots large enough to program was implemented. Extension immigrants originated in southern show the advantages of no-till under agents and researchers at CRI work Brazil and came with a great deal of no- farmers’ conditions; other activities closely together with a participatory till experience. Since the newly cleared included pre-season farmer training, approach, to the point that farmers areas of Paraguay were ecologically on-farm demonstrations, field days, cannot distinguish the activities of each similar to the Brazilian state of Paraná, field tours, workshops and seminars, group. Some rural banks and district many Brazilian technologies could be and distribution of fact sheets and assemblies also promote no-till by adopted without modification. The production guides. The program did providing credit to selected farmers. immigrant farmers interacted closely not promote weed control alone but with farmers and researchers in Brazil. rather an entire farm-management After a hand-operated no-till maize In the 1980s, the Japanese International system that encompassed the use of planter was imported from Brazil by Cooperation Agency (JICA) mounted a certified seed, fertilizing at planting SG 2000 and proved a success, no-till research program in Paraguay to and as a top dressing, pre-planting researchers approached engineers at support the country’s important weed control with herbicide, manual or CRI to have it copied, but they showed community of Japanese farmers. One of chemical in-crop weed control, and no interest. Unlike other countries JICA’s activities was to import no-till new harvesting techniques that left (Brazil, India, and Pakistan), in Ghana planters. Even though the farmers used crop residues in the field. Farmers were local craftsmen were not approached to the planters, they still burned the straw involved in implementing copy the planter. Research in Ghana did not dedicate a substantial effort to because residue management practices had not been developed.

31 2001 CIMMYT World Wheat Overview and Outlook In the early 1980s, local cooperatives reduced income variability. Costs fell Today, a Paraguayan Association of started no-till validation programs. due to lower labor and machinery No-till Farmers and local cooperatives After moving to Paraguay in 1988, Pat requirements, which also reduced the interact with the GTZ program and Wall fostered interactions between need to take expensive short-term Brazilian institutions to adapt no-till Paraguayan farmers and farmers’ credit. Net farm income increases of 35- technologies to local conditions. There associations and Argentine and 236% have been reported among no-till is an active exchange of information Brazilian associations of no-till farmers. adopters (Sorenson et al. 1998). These between farmers’ associations from Wall also worked with researchers from changes resulted in higher living Brazil, Argentina, and Paraguay. The a public institution, the Centro Regional standards, including improved housing, members of the Association of No-till de Investigación Agropecuaria (CRIA), vehicle ownership, and university Farmers are large- and medium-scale to create a no-till program. education for children. farmers; small-scale farmers do not participate because they face different One precursor of no-till programs for The MAG-GTZ program supported the problems and consequently need small-scale farmers was the formation of a no-till association among different solutions (for instance, unlike introduction of green manure crops in small-scale farmers in Edelira, but its small-scale farmers in Brazil, those in the Edelira region by the public impact has been limited. Committees Paraguay use different cover crops). extension service in 1989. In 1990, GTZ, from the association received no-till in association with the Ministry of equipment from the GTZ project or It is estimated that in 1999/2000, no-till Agriculture (MAG), started to promote from the local government, but demand was used on one million hectares. green manure crops among small-scale for the equipment is too high and Despite widespread use of no-till on farmers, but the level of adoption is not sometimes farmers cannot use it when medium- and large-scale farms, known. needed. At other times, the shared adoption by small-scale farmers is just equipment is not maintained properly starting. Wall (1998) and Sorensen et al. Since 1993 the Soil Conservation (Sorenson et al. 1998). One of the main (1998) estimated that the area under Program, involving MAG and GTZ, has benefits that large-scale farmers some type of no-till on small farms in encouraged no-till among small-scale obtained from their no-till associations 1997 was about 4,500 ha, but the area farmers in five pilot areas, especially was the gathering and generation of under permanent no-till was less than Edelira and San Pedro (Sorenson et al. information. Small-scale farmers, 2,000 ha and involved less than 150 1998). The program supported however, do not have resources to farmers scattered around the towns of extension agents and provided some search for information by themselves. Edelira and San Pedro, located in no-till machinery, but the key factors in These resources include access to eastern and southeastern Paraguay, the diffusion of no-till among small- technical literature (through respectively. scale farmers were two highly publications or the Internet), travel to motivated extension agents, Magin distant sites, and setting up a trial Meza and Elba López de Meza, who Key Factors for network. A market for no-till custom built an effective partnership with seeding operation for small-scale Developing No-Till farmers. After two years of working on farmers has not developed because the Packages for Small- green manures, they convinced a group large-scale farmers who own the of small-scale farmers on the brink of Scale Farmers machinery are located in different bankruptcy to try no-till. In the regions (Rolf Derpsch, personal As the case studies have shown, no-till following years they jointly developed communication, 2001). Research was is a complex technology in which an appropriate no-till package. Six greatly strengthened when GTZ opened several components must be adapted to years after starting with no-till, the a program in Paraguay in late 1988, local conditions. Because of this farmers had completely turned around transferring Derpsch from Brazil. Work complexity, a single agent cannot their financial situation, reduced their initially focused on large- and medium- develop a package; what is needed is a dependence on soybeans and cotton, scale farmers, but in 1999 a program for network that may include researchers, increased their total income, and small-scale farmers was implemented.

32 2001 CIMMYT World Wheat Overview and Outlook input suppliers, equipment and later by farmers’ associations. In organization of extension. The early manufacturers, NGOs, extension Ghana, the catalytic agents were development of no-till machinery is agents, and farmers, among others. Monsanto and SG 2000; in India, potentially hampered by the lack of Networks are also necessary for CIMMYT and the RWC; and in demand for specialized equipment. diffusion. Individual agents (e.g., Pakistan, CIMMYT, the RWC, and the Since farmers are not sure about the researchers or input suppliers) have OFWM. CIMMYT is trying to organize benefits of the technology and the researched no-till in many countries, similar networks in Bolivia and Mexico, quality of the initial technical but substantial adoption (by large- and while GTZ is active in Paraguay. standards, they are reluctant to buy the small-scale farmers) has occurred only equipment. At the same time, The core assets contributed by the when an innovation network emerged. manufacturers are unwilling to invest catalytic agents described in this report in developing the equipment because The composition and evolution of were participatory research demand is lacking and it is difficult for innovation networks have varied methodologies, information exchanges them to protect their discoveries. In across countries and times in response among participating agents, linkages Brazil, this hurdle was overcome by the to local technological needs, the with international information sources, collaboration between ICI, EMBRAPA- different agents active in each region, and funds. The funds served three CNPT, and Semeato. ICI and and the socioeconomic environment. purposes: they allowed other agents (in EMBRAPA-CNPT lowered What are the key factors and problems particular, public researchers and development costs by importing that have influenced the performance extension agents) to participate in the foreign prototypes and by contributing of no-till networks? They include those network; in some countries they were formal research capabilities, while identified in traditional analyses of lack used to purchase the first no-till Semeato added their production lines. of adoption of modern technologies: equipments, (creating the demand); market failures or missing markets. 41 and they allowed farmers to In Bolivia, planters were developed In addition to these problems, the experiment with the technology. through collaboration between literature on innovation highlights CIMMYT, PROTRIGO, the local Only the Brazilian network has evolved system failures—that is, failures in the university, and the SILSOE Research into a diverse network in which several innovation networks caused by barriers Institute; in the Indo-Gangetic Plains, agents play catalytic roles in their to collaboration among agents. the agents were CIMMYT, universities, regions, and local and foreign agents and local manufacturers. In both cases, generate multiple information flows. In the networks reduced the development Key factors the other networks reviewed in this cost and created the initial demand, report, information flows converge on The resources that determine the reducing uncertainty for the catalytic agent, who distributes the performance of innovation networks manufacturers. In Brazil, ICI lent information to other agents (sometimes are the particular agents active in imported machinery to interested in the form of regional workshops or different environments, their core and farmers. In Bolivia and the Indo- study tours) and also serves as the complementary assets, and learning Gangetic Plains, the no-till programs main link with foreign sources of routines. bought a few drills and lent them to information. The most important role in no-till farmers. In all cases, interest was In addition to supporting research, the networks is that of the catalytic agent. created among farmers by allowing catalytic agent often provided funds to In the case studies reviewed in this them to experiment with the planters accomplish two important objectives: report, this role was played in Brazil after the demonstrations. the development of planters and the initially by agrochemical companies The contribution of extension infrastructure took different forms. In Brazil and Ghana, agrochemical 41 Market failures arise when the market cannot value the consequences of the actions of one agent over other agents (externalities), when one participant in a transaction has more information than companies provided the initial the other (asymmetric information, also known as moral hazard), or when certain goods do not have a market value that reflects their production cost (public goods).

33 2001 CIMMYT World Wheat Overview and Outlook resources for the extension effort, either centers) were the sources of involved at different stages and in using their sales force or funding public information. The extension programs different ways in generating no-till agencies. In other cases, resources for for small-scale farmers had a more packages. They contributed formal extension came from the public sector; traditional structure than those for research capacity and scientific in India, for example, state universities large-scale farmers. information from their fields of provided the extension infrastructure; specialization. Despite this Even smaller and more focused in Pakistan, it was the OFWM. involvement, they did not become the associations, such as those for hub of the innovation network because Farmers played an important role in purchasing and sharing machinery, they lacked financial resources and, in organizing no-till networks in South have seldom worked for two reasons. many cases, a participatory approach America, but in most cases, farmers First, the seeding period has to be long to research. have remained nonleading partners in enough to allow all members to plant research and/or extension activities. on or very close to the optimal date. Agrochemical companies provided Farmers’ associations played a major Second, there is a free rider problem: 43 important support for the development role in the diffusion of no-till for not all association members maintain of no-till packages in countries and medium- and large-scale farmers in or calibrate the planter properly. When regions with potential markets for their Brazil, Paraguay, and Argentina; only these associations have worked, a herbicides. Their contributions in southern Brazil were these catalytic agent set and enforced the involved conducting in-house research, associations important for small-scale rules of use. funding research and extension by farmers. Attempts to create similar third parties, and exchanging It is important to make a distinction associations in other countries have not information. Even in their high-priority between public research institutions been successful. The South American countries, however, these companies and the individual researchers who associations succeeded because they have not had a uniform impact. In work within them. No institution were created by small groups of highly some countries, such as Brazil and (except for IAPAR and EMBRAPA- motivated farmers with resources to Ghana, private companies were closely CNPT) has ever played a major role in search for, produce, and disseminate involved in organizing the network the early development of no-till or in information. 42 The associations were and contributed key assets. In other the organization of a no-till network. very focused and efficient in these countries, such as Mexico and India, Research institutions have been slow to activities, and they enabled farmers to agrochemical companies shared the recognize new areas of research that benefit from the economies of scale that public institutions’ linear vision of cannot be described along disciplinary characterize these processes. science and could not catalyze no-till lines or directed toward specific crops; Associations of small-scale farmers networks despite important this difficulty has become more were not successful because such investments. prominent as institutions have been farmers lack resources to generate and/ forced to generate an increasing share Equipment manufacturers have been or gather information. When small- of their funding and have introduced less innovative than agrochemical scale farmers have successfully more formal priority setting companies in generating information adopted no-till, other institutions (e.g., procedures. On the other hand, many and machinery because they have agrochemical companies, NGOs, researchers from public institutions, fewer resources for research, their farmers’ associations, international working individually, have been ability to capture research benefits is limited, 44 and the potential market is smaller. Manufacturers receive 42 They could pay for travel in their countries and abroad, hire consultants, and experiment on their own farms. information from a number of sources, 43 The free rider problem arises in a network because each member can increase his individual including local farmers, public benefits if he contributes to the common effort less than is expected from him. Since this is true for every member, in the absence of a mechanism that forces everybody to contribute his institutions, agrochemical companies, expected share, the collective effort and benefits are smaller than when the rules can be enforced. and international research institutions. 44 Patents have never been important for agricultural equipment manufacturers and are almost impossible to enforce in most developing countries.

34 2001 CIMMYT World Wheat Overview and Outlook The most sophisticated manufacturers, Most local universities, as institutions, justification for the public sector’s like those in Argentina or Brazil, have been absent from no-till networks involvement in research. The recent actively follow developments in in developing countries. Professors are literature on NIS highlights a new area industrialized countries by visiting not required to interact with farmers, for public policy: system failures that farm shows in the US to copy advanced and there is no effective quality control arise from insufficient interaction designs. Conversely, US manufacturers of their activities. These features, among agents. In the case of the copy innovations developed in South combined with insufficient resources networks that generate no-till, such America. Manufacturers in most other for research and interaction with system failures include difficulties in developing countries rely on foreign institutions, have prevented broadening the focus of existing information provided by agents with professors from staying up to date in networks, poor understanding of the international links, such as their disciplines. Most professors have particularities of no-till research, the international research centers or been reluctant to change research and linear vision of science that still prevails agrochemical companies. teaching approaches developed over in most public research institutions, many years. lack of a systemic approach to the Numerous NGOs and international analysis of production systems, cooperation agencies (especially GTZ) Foreign research institutions and contradictory reforms of public have played important roles in the farmers, on the other hand, have research institutions, and restrictive evolution of no-till in many countries, generated no-till information that is frameworks of funding agencies. providing funds, research staff, and useful to local agents. The activities of information. Their impact has been advanced research institutions are Difficulties in broadening the focus of limited, however, except for a few particularly important, because they existing networks. In all cases, except cases, as in Ghana. perform science-intensive research that in South America, no-till research has in many cases can be transferred to not focused on developing a complete With the exception of CIMMYT, most developing countries. These package (keeping the soil covered, international agricultural research institutions have mostly interacted minimizing soil disturbance, and centers of the Consultative Group on with researchers linked to international practicing crop rotations). Usually only International Agricultural Research research centers or multinational some of the required components were (CGIAR) have had limited companies, although some have developed, such as weed management participation in no-till networks. The recently established collaborations with strategies or machinery. To develop the International Institute for Tropical farmers’ associations or public missing components, the network Africa (IITA) has maintained a no-till institutions in regions where no-till must be enlarged to acquire new program in Africa since the 1980s, and networks have emerged. complementary assets. In many cases, the International Center for Research in local agents can contribute those assets, the Dry Areas (ICARDA) has studied Several international networks have but their participation is restricted by tillage methods for several years (Pala been created in the last decade to foster inter-institutional rivalries or by 2000). These programs have produced the development of sustainable working in an organizational structure a wealth of information, but their agriculture. Even though no-till is one that reflects a linear vision of science. In impact has been circumscribed because of the most important technologies other cases, foreign agents hold the they were based on traditional recommended, other packages are also needed complementary assets, and experiment station or on-farm research. promoted (see “Networks for their participation is expensive. Recently IRRI initiated a program on Sustainable Agriculture,” p. 36). Because of weak interactions with no-till and began to participate in the foreign agents, many local researchers Indo-Gangetic Plains no-till network. System failures do not fully understand the deficiencies ICRISAT has had limited participation in the packages being promoted. in no-till research in the RWC and in The existence of market failures (in Consequently, they fail to recognize the Africa. particular, the prevalence of public goods) has been the traditional

35 2001 CIMMYT World Wheat Overview and Outlook need to incorporate new agricultural technologies requires new provide incentives for researchers and complementary assets into the awareness on the part of the catalytic technology users to interact. network or to identify the owners of agents and continuous assessment of Researchers plan their activities with those assets. the network. Agents must interact little formal (and in many cases Poor understanding of the more closely and informally than in ineffective) interaction with extension particularities of no-till research. networks aimed at developing better agents and farmers. known technologies. The novelty of Because no-till is a relatively new The linear vision of science does not this institutional framework makes it technology, the mechanisms for emphasize a systemic approach to the more difficult to obtain institutional developing and diffusing information analysis of production systems. Given support and funding for many of these are not well established. One of the that many public research institutions activities compared to more traditional main roles of the catalytic agents is to are still organized on the basis of research programs. establish those mechanisms by disciplines or crops, researchers are identifying the weaknesses of their Public research institutions were (and accustomed to working in projects networks and the agents that can help most still are) organized on the basis specific to their areas of expertise. No- to overcome them. This nontraditional of a linear vision of science. This till research requires interdisciplinary approach to the development of organizational structure does not collaboration and a systemic (versus commodity) focus, but incentives generally discourage (or at least, do Networks for Sustainable Agriculture not encourage) participation in interdisciplinary programs. CIMMYT recently created a global program on conservation tillage to exchange information on agronomic issues and exchange machinery for small- Public research institutions in many scale farmers. The bed planter developed in Mexico, as well as the drill countries are going through a developed in Bolivia, were sent to India and Pakistan, where they were further transformation, but the objectives refined and returned to their countries of origin. A CIMMYT-CIRAD project is and instruments are often organizing the exchange of information on planters between Mexico and contradictory. The main features of Brazil. CIMMYT scientists participate in the networks in the countries where this transformation in research they work, visit countries where no-till programs are being implemented, and institutions are 1) new priority setting sustain an active dialog with each other on technical issues. mechanisms (usually relying on more formal procedures and greater At Cornell University, the Management of Organic Inputs in Soils of the emphasis on identification of Tropics (MOIST) group collaborates with partners in Africa (West Africa and technology demands), 2) an emphasis Madagascar), Asia (Philippines, Indonesia, and China), South America (Brazil on diversifying funding sources, which and Paraguay), and Central America (Honduras and Costa Rica). Following in all cases included a substantial MOIST-organized workshop on cover crops and organic inputs in tropical reduction in direct budgetary soils, a consortium was formed to address information-access issues allocations, 3) greater pressure to synergistically. MOIST also maintains an electronic network, several electronic generate resources by selling goods discussion groups, and three databases. and services, 4) reductions in research The Indigenous Fallow Management Network, housed at the International staff and support personnel, and Centre for Research in Agroforestry (ICRAF) in Bogor, Indonesia, comprises 5) serious reductions in investments, Southeast Asian organizations. which resulted in a general decay of The World Bank and FAO have promoted and financed several networking research equipment and outdated activities, including workshops and study tours by farmers, researchers, and libraries. There is evidence that manufacturers from Africa and Asia. FAO also supports international research capability has diminished meetings on no-till. The participation of these institutions in networks has through the lack of operational funds been, for the time being, restricted to isolated activities.

36 2001 CIMMYT World Wheat Overview and Outlook and investment. Greater pressure on transformed into complex technologies • The importance of commercial inputs — public research institutions to generate by foreign competition in domestic When a commercial input is a key component of a technology with a their own resources has forced them to markets. The challenge is far greater for large market potential, its producer concentrate on producing goods that small-scale farmers, who lack resources has a strong incentive to invest in the have a market value—with a reduction to search for new economic and development of the whole package, in the production of “public goods”— technological opportunities. even though a large part of it could or on research that responds to political be a public good. This expanding universe of complex needs with very short-term objectives technologies demands more (IDRC 1997; Rozelle, Pray, and Huang sophisticated agricultural and scientific Although the development of no-till 1997; Ekboir and Parellada 2000a). policies. A discussion of some of the packages and their adoption by small- The linear vision of science restricts economic and innovation policy scale farmers followed different paths activities of funding institutions. Most changes required in this new than for large-scale farmers, the paths funding institutions do not understand environment may be found in OECD shared one important common feature: the importance of fostering interactions (1999). Less understood are the new all successful programs resulted from among innovative agents and give policies required to accelerate technical networks that worked with priority to more traditional research change among small-scale farmers. participatory research approaches (for activities. This approach restricts the These policies should be based on the a partial list of some recently availability of funds for networking. recognition that complex technologies established networks, see p. 36). Large- and their adoption are social processes scale farmers had the resources to Conclusions that result from the co-evolution of create novel institutions and to search innovation networks and the for information; small-scale farmers, In areas where adequate no-till technological packages they develop, lacking those resources, relied on other packages have been developed, no-till restricted by technical, economic, and agents to organize no-till programs. has had a major, positive impact on the social factors. Several other factors Other characteristics of the successful lives of small-scale farmers and on the condition policy design: development and adoption of no-till management of natural resources. • The complexity of the package—Simpler for small-scale farmers are listed below. Developing such packages has not been packages are easier to develop and/ • A catalytic agent was willing to or transfer than more complex easy, because novel approaches to mobilize others to organize packages, because adaptations of research and extension are required. programs for small-scale farmers. complex packages require the These lessons are important, not only The catalytic agent took different involvement of more agents and, forms: private companies, for no-till but for other complex potentially, more complex research. technologies. The range of complex associations of large-scale farmers, • The scientific base of each component in state governments, agronomists technologies, and the need for them, is the package —Modification of science- international research institutes, in or expanding as farmers are pushed into intensive technologies requires international cooperation agencies. globalized markets and increasingly strong research institutions (either • Although the most active agents in public or private). Conversely, lay required to integrate production and the network were not small-scale users can modify technologies that marketing strategies. Since farmers farmers, they were willing to do not depend heavily on science. now have to learn to produce and sell develop or adapt existing techniques in ever-changing markets where • The sensitivity of the package to local to the local needs of those farmers. conditions—Drip irrigation can be competitors are permanently searching • Small-scale farmers may not have immediately transferred because its for new strategies to increase their taken a leading role in developing dependence on the local the technology, but they were competitiveness, formerly simple environment is very small. The important participants in the process. technologies (e.g., designed for crop success of a crop rotation, on the production for local markets) are being other hand, depends to a great extent on local soils and climates.

37 2001 CIMMYT World Wheat Overview and Outlook • The information was presented in a • Farmer-to-farmer communication is a Despite the common belief that farmers way that enabled small-scale farmers crucial instrument in convincing are reluctant to accept change, no-till to assimilate it. For instance, the small-scale farmers that they can experiences in several countries show dissemination program in Ghana adopt the technology even with their that both large- and small-scale farmers established accessible demonstration limited resources. plots that were large enough to show adopt new technologies very fast when farmers the advantages of no-till the benefits are clear and the package under farmers’ conditions (at least Although agrochemical companies fits their needs and constraints. A slow 2 1,000 m for the farmers’ “standard realized that no-till could open rate of adoption indicates deficiencies practice” plot and for the no-till plot). important markets for their products (either technical or economic) in the The size was also convenient for and were willing to invest in package. More than three decades of calibrating the 15-liter sprayer (Findlay and Hutchinson 1999). The developing the appropriate research on no-till have yielded a demonstration program included technological packages, their wealth of technical information; the key pre-season farmer training, on-farm participation was not a determining issue faced by no-till networks in demonstrations, field days, field factor in the success of the technology. nonirrigated areas is how to use the tours, workshops and seminars, and For example, Monsanto invested in available information to develop a distribution of fact sheets and developing no-till packages for small- locally appropriate package. When the production guides. scale farmers in many countries on right network emerges, research and • Since small-scale farmers lack three continents with limited success. adoption proceed quickly: resources to travel, a large network of development of the first packages in demonstration plots must be established. Farmers should not have South America took about two decades, to travel more than 50 km to attend but in Ghana it took about five years. field days.

38 2001 CIMMYT World Wheat Overview and Outlook