INTRODUCTION

Participation has been on the lips of development professionals for many years now. People's participation in development requires that the top-down approach be replaced by a process of helping people to articulate their problems, form self-help groups and formulate plans to use their own resources to achieve their objectives. Such methods have been used in projects to rebuild infrastructure in the wake of cyclones, to organize villagers to run their own night schools for adult education, and to form groups to guarantee jointly the repayment of business loans.

'Participation' aims to empower people through joint activities; the alternative style of intervention – imposing projects on passive recipients, or offering them project 'handouts' – is disempowering, and creates dependency on the external agency. In an era when aid is increasingly rationed across ever-expanding areas of need, helping people make the best of their own resources rather than hope for bounty from outside seems to be the only realistic policy.

For poor farmers in the marginal lands around the world the drive for participation in agricultural projects comes with an additional impetus. The goods from outside – in this case the 'green revolution' technologies of high-yielding seed varieties, packaged with fertilizers, herbicides and pesticides – are not only unaffordable, but they are inappropriate. It is now well recognized that introducing high-input agriculture to areas of the world where poor people cultivate difficult soils in adverse climatic conditions is not successful, not because the farmers themselves are ignorant or are unable to apply the new technologies according to the book, but because the technologies are often not suitable for the variable conditions under which the farmers work.

Technologies developed on research stations and promoted by agricultural extension services may yield well in the prosperous lands of, for example, the Punjab or the commercial farming regions of Zimbabwe, but perform erratically in the semi-arid, dryland farming regions of southern India, and the communal lands of Zimbabwe. The result is that farmers are barely able to cover the investment costs of these seed packages and feed their families as well.

Those areas where modern varieties were successful can be termed 'green revolution' regions: where rainfall is reliable, or irrigation possible. Marginal lands which are distributed all over the Third World, and cover much of sub-Saharan Africa, have been characterized as complex, diverse and risk-prone (Chambers 1989): here agriculture is rain-fed, and droughts are a regular occurrence.

This book describes the Chivi project, a pilot study which explored alternative ways of working with smallholder farmers to develop technological options appropriate for such conditions. The first part of this chapter considers farmer innovation and technology development, and contrasts this with the failure of agricultural extension services to deliver relevant technologies for marginal conditions. Two themes are then explored: participation – but by whom and to what extent? – and the sustainability of livelihoods and institutions.

Farmers as experimenters

Poor farmers in marginal lands traditionally plant a range of crops and vegetables, often intercropping, as well as preserving trees in their fields for their fruit and firewood. To the eyes of most commercial farmers, their farms appear an untidy mixture of crops, with tiny plots and low yields. Agricultural scientists and social scientists increasingly recognize, however, that what poor farmers have always used – their own varieties and traditional farming techniques – are well adapted to their environment, and are more appropriate than the monocultures that research stations and extension services tend to promote. From all over the world, practices previously dismissed as unscientific and backward are now recognized to have a rational basis, and to be the result of informal innovation and experimentation on the part of often uneducated farmers.

There is now plenty of evidence that farmers are skilled at selecting seed varieties to suit the varying agronomic conditions on their farms, and plant a range of varieties to suit different soils and water availability within their landholding. Mende farmers in Sierra Leone, for example, select shorter- or longer-season rice varieties to plant in marshy areas, in water-retentive soils in valley floors, in free-draining upland areas, and in water courses – and thus avoid the need to irrigate (Almekinders 1999). This adaptation of the germ plasm to suit the conditions contrasts with the conventional approach of adapting the conditions (for example, applying irrigation) to suit the requirements of modern varieties. Farmers identify favourable characteristics in other farmers' varieties and combine these with their own varieties, selecting the plants with the most promising characteristics and multiplying up the seed. While formal research tends to concentrate on yield alone, farmers' selection criteria are more complex and include taste, texture and cooking properties, suitability for intercropping, pest resistance, ability to grow in the shade of trees, length of duration, reliability of yield during drought years, and so on.

All this does not mean that farmers are producing enough to feed their families: regular food relief in these regions is evidence to the contrary. Nor does it suggest that farmers have all that they need in terms of varieties and technologies at their disposal, and that agricultural research stations or extension services have nothing to offer them. Farmers are interested in formal variety trials, as was demonstrated by the case of the Mahsuri rice, a variety that 'escaped' from a research station in India via a farm labourer, was multiplied and spread informally among farmers with whom it was popular, and was only officially released by the government some time later as a result of popular demand (Maurya 1989). It is also a mistake to label modern varieties as 'bad' and traditional varieties 'good' for marginal farmers – they themselves are usually interested in trying out the new alongside their native varieties, incorporating a new variety that has performed well in their fields within their own germplasm banks along with several of their favoured native varieties; indeed they probably do not see a great dichotomy between new and native varieties (Rhoades 1989).

What is becoming clearer is that, when farmers adopt new technologies, they do...