Africa's image in the West is grim: a continent hovering on the brink of famine, with eroding soils and shrinking forests, at the mercy of a climate of fearful unpredictability. Good rains over the past two years have brought respite, but they have not changed the realities that largely justify the image.

Aid projects and government programs in Africa have an appalling failure rate. One in every two World Bank agriculture projects in East Africa flops -- against one in 20 in South Asia. High imports, high costs and high dependence on government experts or bureaucrats make such projects prey to Africa's cash and manpower shortages.

Projects often ignore Africa's unpredictable climate, poor and easily eroded soils and diversity of cultures. They often involve costs that the peasant can't afford or risky practices, rather than tried, traditional ones. Prevailing low prices for farmers' produce kill many attempts to boost output.

Yet success is possible. Africa could feed its growing populations and save its threatened environment. Low-cost self-help methods could bring results by 1990, without unduly straining African budgets or Western aid donors' funds.

I visited 20 projects across Africa that have broken through where so many have failed, to find out how they pulled it off.

New high-yielding crops are often unpopular in Africa because they need fertilizer and careful management. But planting of a new breed of the root crop cassava -- developed by the International Institute for Tropical Agriculture -- is spreading spontaneously among Nigerian farmers. It doubles yields, with no input and less labor.

New tree plantations in the Sahel rarely flourish. Yet in Niger's Majjia valley, farmers have planted and nurtured to maturity 330 kilometers of windbreak trees, which boost crop yields by 20 to 25 percent and cut wind erosion.

Dozens of stoves that save firewood have been designed, but no more than a few thousand of each type are in use. Yet within three months of its nationwide launch in Burkina Faso, one simple improved mud stove spread to 86,000 households, cutting their fuel-wood use by half. The stove costs nothing and can be individually fashioned in a day to fit any pot.

Northern Burkina Faso has been devastated by desertification. Land that once yielded crops has been so overfarmed and overgrazed that it has crusted over. Plants and trees wither and die while rain sheets off uselessly.

Oxfam's Peter Wright, working closely with local farmers there, found that placing lines of stones across the region's gentle slopes could solve the problem if they were exactly aligned with the land's contours. The stones hold the water back long enough for it to sink into the ground, depositing soil and leaf litter. They have raised crop yields by up to 50 percent and have literally pushed back the desert.

To find the contour levels, Wright designed a cheap ($6) device made from lengths of clear, water-filled hose tied at each end to a notched stick. He trained peasants from each village to use them. Hundreds of villages have built their own lines and watched their land turn green and their granaries fill again.

From ventures like these, a rough blueprint for success can be drawn. Africa's farmers are poor; its governments are also poor and always short of foreign exchange. Import restrictions and budget crises are a daily reality. It follows that development efforts should rely as little as possible on imports of any kind and should be ultra-low in cost to farmers and governments.

They must promise a good return with no extra risk -- farmers on the bread line will invest scarce cash or labor only in sure-fire successes. Luckily, for most African environments, low-cost techniques now exist that can boost food production by 20 percent or more and conserve soil and trees at the same time.

Efforts in Africa must be based on self-help, catalyzed by simple training spread by networks of village-level workers or volunteers. If an activity cannot be kept going by the local people, it will collapse just as soon as the government jeeps break down.

Aid donors need to absorb the lessons of success as well. They must no longer encourage imports of machinery -- or pull out from high-cost projects after five years, leaving Africans to foot the continuing bills. They must stand ready to support low-cost programs for as long as it takes them to break through to the rural majority.

The most crucial requirement is incentives. Wherever farmers have been given attractive prices for their produce, as in Zimbabwe or Malawi, they have responded with handsome increases in output. State marketing boards must pay more for food and cash crops -- or, better still, allow a free market, with freely floating currencies.

If these approaches are followed, all but a handful of African countries could become self-sufficient in food production within three or four years and be on a secure footing to move on gradually to higher cost approaches. At the same time, African nations would conserve their soils and forests, and be less vulnerable to drought and famine.

Otherwise, Africa will remain in a chronic state of near crisis -- interrupted from time to time by apocalyptic catastrophe.

Paul Harrison is a British journalist who writes about Third World development problems.