Advice from the Field

Agroforestry: learning as we go in Africa
by Peter Cooper

Introduction

Peoples’ awareness of major global environmental issues such as deforestation, loss of plant and animal biodiversity, global warming and carbon sequestration and land degradation - and their links with poverty - has increased enormously in recent decades. The seriousness of these issues is not contested in sub-Saharan Africa. It is not surprising therefore that interest in peoples’ age old reliance and traditionally sustainable interaction and interdependence on trees has been re-awakened. This is logical. Trees can provide a wide range of useful and valuable products, together with the services which help mitigate some of the negative impacts people have on their own environment (Cooper et al. 1996).

There are many who justifiably believe that the future of trees, and hence the role they play in sustaining people and their environment, will rely not so much on maintaining the remnants of our forests, but rather in encouraging the deliberate re-introduction of trees as an integral part of the land we farm (Sanchez et al. 1998). It is this conviction that underpins the emerging science and practice of agroforestry. I stress this because, in his editorial to this Handbook, Roger Leakey points to the fact that as agroforesters attempt to define the underlying principles of the science of agroforestry (Sanchez 1996) the definition of agroforestry has become increasingly complex. I have personally enjoyed, but frequently been frustrated by, semantic debates on this issue. However, I remain to be convinced that the scientific complexity of studying agroforestry needs to be matched by even more complex definitions! Running the risk of being accused of confusing simplicity with being simplistic, I prefer to think of agroforestry as “managing trees on farm”. Those four words, to me at least, embody the essence of what agroforestry is all about.

I alluded earlier to the complexity of agroforestry as a science, not only from a conceptual perspective, but also from the practical point of doing rigorous and relevant agroforestry research that leads to widespread adoption of agroforestry innovations by farmers. Much has been written about:
• ways to evaluate farmers’ concerns and priority for different agroforestry tree species (Franzel et al. 1996);
• issues which bedevil the design of valid agroforestry research field trials (Coe 1994, Rao et al. 1990);
• approaches to the participatory evaluation of promising agroforestry practices with farmers and how to assess their adoption potential (Franzel and Scherr in preparation) and
• the complexity of assessing the impact of natural resource management interventions, such as agroforestry, at different spatial scales (Dumanski et al. 1998).

It is no coincidence that the references cited above are all fairly recent. Agroforestry is an emerging science and the research and development community is making mistakes and learning as it goes. We know today a great deal more than we knew ten years ago, and I do not doubt that in ten years time we will marvel at our naiveté in 1999! But in spite of this agroforestry is on the move and research is yielding results, not only in increasingly perceptive publications, but on the ground and in farmers’ fields. To illustrate these points, I present a case study from an ICRAF (International Centre for Research in Agroforestry) and Government of Zambia collaborative project, and highlight some important factors which appear to have contributed to its success and suggest some lessons we have learned.

Mrs. Zulu, a participant in the ICRAF/Government of Zambia Chipata project, shows her bumper yield of maize following a two-year Sesbania fallow. Photo © Anthony Njenga

Chipata, Zambia: a case study

Chipata district is situated in Zambia’s Eastern Province, close to the Malawian border. It was given priority as the study area in this collaborative project by the Government of Zambia as landuse is dominated by a mixed maize and livestock farming system. This system is common throughout the Eastern, Central and Southern provinces of Zambia. The district is well served by Msekera Research Station, which is situated near to Chipata township (13° 39’ S, 32° 34’ E). Franzel et al. (in preparation) have summarized the principal features of the study area as shown in Table 1.

Farmer surveys undertaken in 1987 identified declining soil fertility as one of farmers’ main perceived problems. Nitrogen deficiency was judged to be the most important problem responsible for low maize yields. Increased pressure on land has reduced fallow periods - farmers’ main method for maintaining soil fertility - to one to three years. Some farmers even practice continuous cropping because such short natural fallows do not result in greatly increased yields. Fertilizer use was common during the 1980s but the removal of subsidies caused the ratio between the price of nitrogen and the price of maize to increase from 3.1 in 1986/87 to 11.3 in 1995/96. Fertilizer use in Zambia declined by 70% between 1987/88 and 1995/96 and the decline in the smallholder sector was even greater.

Table 1 Principal environmental, landuse and farm features
at Chipata, Zambia (Franzel et al. in preparation).

Agroforestry interventions for nutrient replenishment

Research on soil nutrient replenishment started at Msekera Research Station in the 1987/88 season. Given that nitrogen deficiency was severe and widespread and that almost all farmers were still practicing natural grass/bush fallows, emphasis was given to assessing the potential of improved fallows. In our on-station research, we studied a range of leguminous shrubs and trees and their interaction with management strategies such as fallow length, planting density, method of establishment, intensity of fallow weeding and other factors. The effects of fallow species and management strategies were evaluated by assessing maize yield responses compared with controls which represented a range of common farmer practices. On-station responses were dramatic. Yields following two year improved fallows greatly out-yielded those following grass fallows and often exceeded those obtained by fully fertilized maize. Farmers’ reaction at field days was overwhelmingly enthusiastic.

In 1995, we witnessed the start of a substantive on-farm participatory evaluation of two-year improved fallows by the ICRAF/Government of Zambia Team in formalized partnership with Farmer Training Centres, several NGOs and farmer organizations. These partnerships have become known as Adaptive Research and Dissemination Teams (ARDTs). Farmers were invited to choose one of six possible improved fallow strategies comprised of a choice of three species (Sesbania sesban, Tephrosia vogelii, Cajanus cajan) either established as pure stand or intercropped with maize in the first year.

What happened thereafter has been well-documented (Kwesiga et al. 1998). In summary, experimenting farmers’ own experiences mirrored on-station results and their feedback has confirmed that improved fallows were feasible, acceptable and profitable. Farmers appreciate the greatly increased maize yields of up to 300-400% (ICRAF 1997), which result from improved fallows. In addition to these high maize yields, village level workshops held in 1996 indicated that experimenting farmers have a broad range of expectations of improved fallows at the plot, farm, household and community scale (ICRAF 1997; see Box 1). The numbers of experimenting farmers rose from 204 in 1995 to 958 in 1996, and to 3023 in 1997.

Such is the proven potential of improved fallows. So great is the enthusiasm of farmers that in 1997 the project and ARDT partners started preparations to move into the wider dissemination of improved fallows, both within Zambia and in similar ecozones in Malawi, Zimbabwe and Tanzania. In 1999, it is estimated that over over 6000 farmers in the four countries have planted improved fallows of Sesbania and Tephrosia.

Possible constraints to widespread adoption and lessons learned.

Optimism for real change is running high in Zambia, but our partners in research and development are aware that certain issues remain to be addressed. The constraints and potential solutions being evaluated are given in Table 2. Although somewhat specific to this project, these constraints illustrate the need for an integrated and multi-disciplinary research planning and implementation approach which, right from the outset, includes consideration of the following:
• the need to effectively monitor wider dissemination to identify issues such as the possibility of pest and disease outbreaks as “scaling up” occurs;
• the need to have a range of choices for farmers and to train them well in the management of those choices;
• the usefulness of an ex ante analysis of labor demand profiles of potential interventions;
• the value of examining the impact of existing and alternative policies on adoption potential; and lastly
• the desirability of anticipating germplasm demand and supply issues well in advance.

Factors contributing to success

When the project started in 1988 there was no clear long-term strategy of evolution from research through to development nor was there a clearly spelt out conceptual framework within which to operate. These have developed over time and in response to evidence of emerging success. Kwesiga et al. (1999) recognize that the driving force behind development is people’s enthusiasm for change. In hindsight they attribute the apparent success of this project to the following factors:
• correct initial diagnosis of problems;
• the early involvement of farmers, extension staff and NGOs in the research process;
• starting small and building on local knowledge and success as the project evolved;
• demonstrating clear and recognizable results with substantive impact;
• offering a range of options to farmers and encouraging their innovation and adaptation through farmer designed and managed trials;
• improved fallows appear gender and wealth neutral with regard to adoption;
• the formalization and function of the ARDT that has resulted in several outcomes, the most important of which are:

’ the cost of on-farm research is spread among the various research and development partners;
’ the breadth of input into and the relevance of the research have been greatly enhanced;
’ as research progresses towards dissemination, all partners have become increasingly well informed on key aspects of technology options and their management;
’ both research and development partners, as well as farmers, have developed a sense of involvement, enthusiasm and ownership of promising innovations;
’ as the project moves into the wider dissemination of successful innovations our development partners, through whom the vast majority of farmers will be reached, are better placed to undertake informed and targeted dissemination and to critically analyze and respond to farmer feedback.

Box 1. Farmers’ expectations of improved fallows as identified by farmers in village workshops (Zambia).

In the plot
• Soil fertility, soil structure, rainfall infiltration will improve
• Less erosion will occur in plots
• Better maize seedlings will emerge and maize yields
will be higher
• Sesbania will control striga, act as a windbreak,
and provide shade for rest periods

In the farm or in the household
• Increased harvest will mean increased food supply
• More firewood will be available
• More time will be available for other activities because of reduced time spent collecting fuelwood*
• More cash will be available (earnings from maize
and savings from buying no or less fertilizer)
• Standard of living and nutrition will improve
• Sesbania poles will be available for building storage bins
and fencing
• More maize stover will be available for cattle fodder
• More labor will be needed – for work in nurseries, for transplanting and weeding – but less for cropping
• More time will be needed for herding in dry season
to ensure animals do not damage fallows*

In the village
• Indigenous trees, forests, and wildlife will be saved
as fuelwood collection declines*
• Food security will increase
• Grazing area will be reduced*
• Community spirit of working together will be enhanced
through group nurseries
• Need will be greater for community regulation
of grazing and fires

*Indicates there was lack of agreement among farmers

Table 2. Current constraints to adoption and impact of improved fallows and potential solutions, Chipata,Zambia.

Conclusions

The case study from Zambia has illustrated a range of issues related to successes and possible failures in the research-development continuum. It is, however, an initiative centered around addressing a focused and single, albeit important, problem, that of declining maize yields. Landuse systems, their problems and opportunities are complex and are not static. They evolve with time. Cooper et al. (1996) have concluded that “Agroforestry Systems which provide solutions for today’s land degradation problems will need to evolve in both diversity and intensity if they are to remain relevant and effective for tomorrow’s Africa”. I take this opportunity to end by re-iterating and highlighting this conclusion as it is vitally important. It reinforces the need to avoid possible complacency that may come with success. We must continue to ensure that we learn as we go.

Acknowledgements

The author acknowledges the dedication and drive of all the research and development personnel involved in the Zambia project and invaluable involvement and hard work of Zambia’s farmers as genuine research partners in this initiative. They have taught us much and will continue to do so.
Peter J.M. Cooper, a systems agronomist and agroforester, has 30 years of research and development experience in sub-Saharan Africa and the Mediterranean region. He is currently Leader of ICRAF’s ‘System Evaluation and Dissemination Programme’.

Contact: Dr. P.J.M. Cooper, ICRAF,
P.O. Box 30677, Nairobi, Kenya; Tel. +254.2.521450, Fax +254.2.521001,
E-mail P.Cooper@cgiar.org.

References Cited
Coe, R. 1994. Through the looking glass: 10 common problems in alley-cropping research. Agroforestry Today 6:9-11.
Cooper, P.J.M., R.R.B. Leakey, M.R. Rao and L. Reynolds. 1996. Agroforestry and the mitigation of land degradation in the humid and sub-humid tropics of Africa. Experimental Agriculture 32:235-290.
Dumanski, J., W.W. Pettapiece and R.J. McGregor. 1998. Relevance of scale dependent approaches for integrating biophysical and socio-economic information of agroecological indicators. Nutrient Cycling in Agroecosystems 50:13-22.
Franzel, S. and S. Scherr, editors. In preparation. Trees and Farmers: Assessing the Adoption Potential of Agroforestry Practices in Africa.
Franzel, S., D. Phiri and F. Kwesiga. In preparation. Assessing the adoption potential of improved fallows in Eastern Zambia. Book chapter to appear in S. Franzel and S. Scherr, editors. Trees and Farmers: Assessing the Adoption Potential of Agroforestry Practices in Africa.
Franzel, S., H. Jaenicke and W. Jansen. 1996. Choosing the Right Trees: Setting Priorities for Multi-purpose Tree Improvement. The Hague, ISNAR Research Report No. 10.
ICRAF. 1997. Annual Report 1996, pp 209-211. Nairobi, International Centre for Research in Agroforestry.
Kwesiga, F., S. Franzel, F. Place, D. Phiri and C.P. Simwanza. 1999. Sesbania sesban improved fallows in Eastern Zambia: their inception, development and farmer enthusiasm. Agroforestry Systems. In press.
Rao, M.R., C.S. Kamara, F. Kwesiga and D. Duguma. 1990. Agroforestry field experiments: methodological issues for research on improved fallow. Agroforestry Today 2:8-12.
Sanchez, P.A. 1996. Science in agroforestry. Agroforestry Systems 30:5-55.
Sanchez, P.A., A.J. Simons and F.J. Place. 1998. The Future of Trees is On-farm in Africa. Presented at the ASA Meeting (Division 6), Baltimore, USA. October 1998.

On-farm evaluation of Tephrosia vogelii (foreground) and Sesbania sesban (background), key elements in the two-year fallows that have enhanced maize yields and spurred widespread farmer interest. Photo © Joan Baxter.

Playing the mbao game: evaluating local perspectives on the value of trees
by Martina Backes

The problem of obtaining reliable information

Most researchers, ethnobotanists, forest technicians and agricultural extensionists face a common problem when gathering data during field studies. Direct questions usually encourage people to give responses that may not entirely reflect their knowledge or true beliefs. Being aware of this, interviewers are keen to avoid direct or dichotomous questions.

There are many reasons to believe that information given by the informant will be represented differently by the interviewer after being screened through his or her culturally biased way of perceiving things, understanding terms and interpreting language. It is probable that some erroneous information will be given by informants, either intentionally (and the intentions are manifold), or because people are naturally cautious. They may be restrained in their comments about an issue or a plant for fear of not knowing the “right” answer, that is, the one that will satisfy the interviewer.

When the interviewer is able to spend more time on a certain topic, he or she will come to understand the range of possible answers and the complexity of meanings. Social and cultural attributes of plants may complicate the phenomenon, especially when particular use or management strategies are restricted to certain members of the community. In this case, different if not contradictory answers from various people may all be accurate. Only the full range of possible answers will reflect the complete truth, especially when accessibility, resource allocation or user rights are gender dependent or age related. Let me illustrate these reflections with a few examples from my fieldwork in Africa.

In Bungoma, Kenya all female informants mentioned the tree Rhus vulgaris (Anacardiaceae) as a high quality firewood species. Yet just a few women used it for this purpose, although the tree is very common, being found in most farms and off-farm areas. The explanation for its limited use by women is found in a popular belief:

“The tree is not allowed to be used as firewood because it may then cause chaos in the home especially between husband and wife. It is also said that young women who are still able to give birth are not allowed to use the firewood of this tree but there is no ban for elder women and widows. Furthermore, the tree is used as a drug: when burnt the roots are taken as a love charm. The man will prepare the drug secretly. The wife will never run away nor divorce while taking the drug.”

Because Rhus vulgaris is common, multipurpose and well accepted within crop fields (as it seems to be non-competitive with most crops), the tree was identified as an agroforestry species of high potential. Although regeneration after pruning is strong, its cultural attributes show that it cannot be conveniently promoted as a firewood crop.

Misunderstandings in ethnobotanical data analysis may arise because of the size or age of trees. In Bungoma, the juvenile plant of Steganotaenia araliacaea (Apiaceae) is called kumapepenembusi, meaning “the tree whose leaves are chewed raw by goats”. Once the tree has grown up the name changes to kumutomolo, referring to the bark that can be removed easily even when uncut. The resulting tube is a preferred source of pop-guns used by small boys. Both vernacular names refer to other tree species as well, further complicating the case. Talking about one of these trees without a living specimen increases the risk of mistaking the identification threefold.

Playing customary games - An alternative data collection method

The mbao game is widely known in many African countries although the rules vary. It is played with a wooden board that has several holes (16 holes in Kenya, 32 holes in Rwanda where it is known as ikisoro) in which tree seeds, stones or shells are placed.

In western Kenya, we used the game as a participatory rural appraisal tool in social forestry surveys. The aim was to identify tree species that farmers want to grow in their farms and to specify the reasons behind their decisions. We found the game very helpful for creating a congenial atmosphere that facilitated information sharing about trees, their uses and their values among extension staff, ethnobotanists and indigenous people, including farmers, healers, herbalists, midwives and elders. Each participant was able to learn more about specific knowledge and decision-making strategies concerning these trees.

At times, the discussion between local people about the correct meaning, history, folklore, usage and other details was contradictory. Many details seemed to be unclear. Participants expressed the need to encourage further reflections and internal information sharing on the characteristics and roles of trees in order to safeguard cultural knowledge about them.

The Play Sequence and Results

Using an mbao board with only six holes, we placed a branch with leaves of each selected tree next to the hole meant to represent the species. Alternatively, you may use drawings as pictured in this article, but leaves are easier to recognize and varieties may be identified more reliably. The list of species used may vary according to the questions to be asked.

The informants were asked to give a score for each tree on a given criterion by putting from one to six seeds, beans or shells in its hole. The visual results were then reviewed and modified by the informants. The ranking allows participants to obtain quantitative data by evaluating the relative use or value of different species for a particular purpose.

At the same time, the game provides qualitative data. When informants themselves are motivated to ask questions of other members of the community or even of the interviewer, the ensuing discussions reveal information that had not been requested originally. The interviewer observes the dialogue between local people from a distance.

These innovations are valuable since they facilitate process-oriented learning on both sides. The game is a participatory ethnobotanical tool that encourages each participant to be highly creative and innovative in terms of the selection of species, questions, comparisons, potential informants, objectives and other elements.

Keeping records facilitates analysis of the data and comparison of results drawn from different informant groups. Trees may be compared in terms of quality for certain purposes (for example, firewood quality as measured by energy, smoke, and time needed to dry) and cultural attributes (for instance by finding out who is allowed to plant, prune, harvest, consume and fell a specific tree).

Re-Evaluating the Findings with Pairwise Comparisons

Another participatory technique, pairwise comparison, may be used to review, evaluate and verify the results. This technique allows participants to prioritize species for particular use categories, environmental roles (including ecologically important functions such as mulching, soil improvement, windbreaks and erosion control), or cultural and social attributes such as spiritual characteristics, gender-based knowledge, access to or allocation of resources, taboos or management strategies.

In pairwise comparison, trees are ranked in order of preference with respect to one specific use or function. Each plant (drawn from a limited number of species identified in the mbao game) is compared with all others in pairs. One of the two within each pair is selected as better for that particular use or function. Ranking is achieved by counting the number of times a species was chosen as the better of the pair. Visualization of the results motivates the participants to explain why or under which conditions one option is better or worse than another. This increases the completeness and quality of the information.

The mbao game provides an opportunity to demonstrate the use and value of indigenous as opposed to exotic species, a relevant topic for agricultural, forestry or agroforestry extension projects. This participatory evaluation process allows participants to compare intrinsic emic values with external economic valuation of species or modern scientific ways of evaluating ecological functions. When local people are encouraged to ask questions, reverse the roles by interviewing the interviewer, or evaluate the results of their own questions, the relationship between all players – locals and outsiders – is egalitarian, avoiding a top-down approach but encouraging mutual learning. The mbao game acts as a ritual that creates a more equally balanced exchange of experience.

When I used these methods in my fieldwork, I learned a lot from innovative and unusual questions asked by local people. At the end of each workshop, participants found themselves more sensitive to the wealth and value of their indigenous knowledge. From the whole process of mbao there has emerged an awareness that indigenous knowledge of plant resources is increasingly disappearing. Playing the mbao game further strengthens the link between cultural identity and the use of indigenous trees.

Martina Backes is an ethnobiologist who has conducted research on agroforestry systems in East Africa, and who now works with the Third World Information Centre (iz3w) in Freiburg, Germany.

Contact: Dr. Martina Backes, Frankenweg 14, D-79117 Freiburg, Germany; Tel.+49.761.7075125, Fax +49.761.7075123,
E-mail fernweh-iz3w@t-online.de.