Many wonder if plant breeding can achieve much in the African drylands, because the growing conditions are so harsh. Most major historical breeding successes occurred where water was ample. But too many lives are at stake to shun the challenge.

In 1972, the CGIAR created the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) which ever since has worked closely with national counterparts across the developing world to improve and disseminate two major dryland cereal grain crops: sorghum and millet; and three legume crops: chickpea, groundnut, and pigeonpea.

Despite the difficulties of the drylands, farmers now grow improved varieties on about a million hectares across Africa. The adoption rates for improved millet (34%) and sorghum (23%) across southern Africa are particularly remarkable. Other pockets of high adoption include sorghum in southern Chad and adjacent parts of Cameroon (30%), millet in Namibia and Zimbabwe (50% or more), and pigeonpea in the Babati district of Tanzania (35% or more).

Surrounding these pockets of success, however lie vast areas that are yet to benefit. Across West and Central Africa as a whole, where the bulk (about 80%) of sub-Saharan Africa's dryland cereals are produced, adoption rates are below 2% for sorghum and 1% for millet. On an Africa-wide basis, the adoption of improved legume varieties appears equally low, although large parts of eastern and southern Africa have not been assessed yet.

What traits do farmers want most in new varieties? National agricultural research system (NARS) and ICRISAT breeders frequently ask them. Farmers are invited to field trials to score different varieties for their preferred traits, and breeders use this data when deciding which new varieties to release to the public.

Four characteristics of the improved varieties are especially valued by African farmers:

Every dryland African knows the meaning of the phrase ‘hungry season' in the month or two before harvest, the previous year's grain stores have run nearly empty. Without food to eat or money to buy, the poor simply go hungry as they anxiously anticipate the next harvest. Women, children and the elderly tend to suffer the most. All to often, some starve — as in Niger in 2005.

Plant breeding has helped ease this seasonal hunger by developing varieties that mature weeks or even months sooner than traditional varieties. Not only does this put food in hungry bellies; the farmer benefits by getting the year's highest prices. Another crucial advantage of early maturity is it reduces farmers' risk in years when the rains end too soon.

For example the millet variety ‘Okashana 1' in Namibia, selected largely by farmers themselves, matures 4-6 weeks earlier than previous varieties; it was so popular it spread in just a few years in the mid-1990s to cover half of the country's millet-growing area. The sorghum variety ‘S 35' quickly spread in southern Chad around the same time, giving farm families a way to protect themselves from drought and the hungry period. Another early sorghum variety, ‘Macia' that was previously released across southern Africa is now spreading in eastern Africa as well.

Early-maturing varieties can fit into important niche opportunities such as growing an extra crop per year. ICRISAT helped Sudan identify a early-maturing pigeonpea called ‘Tayba' for 50,000 hectares during summertime in the Gezira irrigation area (about 250 kilometers south of Khartoum), complementing the wintertime crop of irrigated wheat. Enthused by this success, Sudan asked ICRISAT's help in finding pigeonpea varieties suited to dryland conditions in the northeast of the country around Gedaref - a vast potential of over a million hectares.

The drylands are diverse, with rainfall varying across locations and years from 300 to 800 or more millimeters per annum. Different soil types and field histories lead to differing levels of soil fertility. Often the soil close to the house compound is more fertile due to ash, wastes and animal manure. Low-lying areas such as valley bottoms often contain rich soils. And farmers who can afford fertilizer and manure can use them to convert a poor soil into a good one.

Breeders have developed varieties that respond to such favorable conditions by yielding 30-100% more than traditional varieties. To capitalize on the opportunity, ICRISAT and partners are researching practical ways for farmers to obtain soil fertility-boosting fertilizer and organic matter. If the combination of soil fertility amendments and higher-yielding varieties can become widely available to smallholders, dryland Africa could attain the food security that has eluded her for so long.

Large additional yield gains lie ahead through ‘hybrid' varieties, such as hybrid millet, sorghum and (in the not-too distant future) pigeonpea. Sophisticated seed production and distribution systems are needed to disseminate hybrids. ICRISAT is breeding hybrid sorghum and millet varieties so that Africa can bring them on-stream as soon as conditions can support a hybrid industry.

Some of the largest gains from plant breeding in Africa have been achieved by developing resistance to devastating diseases. Insect resistance is harder to achieve, but large gains are possible if pests such as legume pod borers and sorghum midge can be repelled.

Pigeonpea is a nutritious foodstuff and also a vital cash earner for the poor in parts of Tanzania. The wilt fungal disease was beginning to cause despair in the 1990s, but the resistant variety ‘Mali' is now saving the livelihoods of these dryland farmers, who have few other options.

Downy mildew infects 30-50% of the millet area in Eritrea, a sorghum and millet-dependent country. The resistant variety ‘Hagaz' was just released in 2004 following intensive partnership with ICRISAT to help build the new national breeding program. Rapid adoption is expected in the coming years.

Groundnut used to be West Africa's number one agricultural export in colonial times, but a disease called rosette virus, which the Hausa people of Nigeria tellingly call ‘groundnut leprosy', brought the industry to its knees in recent decades, including a wipeout of the entire crop in 1975. Several resistant varieties have been released in recent years, raising hopes for a revival of this huge income-earner. So far these varieties have spread over about 18,000 hectares in four Nigerian states.

The new tools of biotechnology have been particularly powerful in battling diseases and pests. The parasitic weed ‘striga' (also known as witchweed, because it appears to intoxicate the plant) is an interesting example. No high-level resistance has yet been found; different sorghum varieties have moderate levels of resistance that is difficult to observe consistently in the field. ICRISAT and partners used molecular markers to map some of these genes, and the markers are being used to combine them into varieties that will hopefully have a higher and more stable resistance than any of their predecessors.

Pod borer insects are extremely damaging to pigeonpea, and resistance not been found despite an intense search in the pigeonpea gene bank (see below). In recent years, ICRISAT scientists have inserted a Bt ( Bacillus thuringiensis ) gene into the plant that is toxic to the borers, but harmless to humans. Bio-safe greenhouse testing now underway will demonstrate whether this results in a breakthrough against this pest.

Unlike environmental stresses, diseases and pests can themselves evolve to defeat a new resistance in the plant. A constant watch has to be kept for new disease/pest strains and the ‘breakdown' (actually the bypassing) of plant resistance. Breeders continuously develop multiple genetic sources of resistance so that they will be ready to quickly deploy a new type when necessary.

Like everywhere, African consumers value grain qualities related to taste, texture, and ease of milling and cooking, and durability in storage for all five ICRISAT crops. In addition, urban and export markets demand specific, uniform quality characteristics. As both consumers and sellers of grain, farmers are very attentive to the quality characteristics of new varieties. Improved grain quality can substantially raise the market value of their harvests, typically on the order of 20-30%.

New chickpea varieties are spreading rapidly in Ethiopia because the market prizes their large ‘kabuli' (garbanzo) type grains. Large, creamy-white seeded improved pigeonpea varieties are preferred in Mozambique and Tanzania because they are readily sold for export to India. Groundnuts for the snack market must have large seeds, while those for the processing industry must be high in oil.

The preferred white grains of sorghum S 35 fetch 25% higher prices than red-grained varieties in Chad. But different tastes prevail in different locations. Some East African consumers prefer red sorghum grain because it imparts a dark color when mixed with cassava to form their staple starch dish ‘ugali'; while others in the same area dislike any darkness in their grain, having gotten used to maize, which imparts a white color to ugali. Two improved varieties are spreading in Nigeria because of their suitability for malting, earning hard cash for farmers and helping the country save millions annually in barley imports.

Millets in West Africa are prized as a weaning food for babies; new varieties like ‘SOSAT-C88' are preferred for this characteristic. An important recent finding is that a few pearl millet types from the gene bank have yellow endosperm and high beta-carotene levels. Beta-carotene, also known as provitamin A, is a molecule that the human body can use to make vitamin A. Deficiency of this vitamin in the diet is a major health issue across Africa, particularly in dry areas where crop options are few. If the vitamin A genes from these ‘golden millets' are confirmed and can be shifted into adapted varieties, the poor, who can least afford a diverse and nutrient-rich diet, will benefit most.

Despite all these exciting traits, the limited availability of seed slows the adoption of improved crop varieties across Africa. Western models of highly structured and regulated national seed systems have been too costly and difficult to manage in Africa. For example, over the past three decades, a bevy of donor agencies and the Government of Mali pumped in over $16 million to strengthen the public seed sector in the country. Yet, the sector supplies less than 1% of the total sorghum seed sown every year by farmers.

A study of the problem in Mozambique concluded that there needs to be a simplification and streamlining of procedures and conditions for variety registration and release. Varieties long-tested and released in one country ought to be considered for expedited release in similar areas of neighboring countries rather than repeating the whole process again.

Instead of government controlling the whole seed process, it can focus on selling the first generation of seed (breeder seed) to the private sector and to non-governmental projects. They are well suited to multiply it and get it to farmers in appropriately small packs through rural retail shops, as long as they adhere to agreed quality standards.

Affordable seed of acceptable quality is more useful for the poor than top-quality seed they cannot pay for. In outlying areas where expensive commercial seed is not available or affordable, communities can themselves produce seed by following practical quality control standards, such as the Quality Declared Seed (QDS) approach recently instituted by a number of countries in eastern and southern Africa.

On both sides of the continent legume crops present an especially difficult challenge, because their seeds are readily damaged by insects in storage. And since groundnut produces relatively few seeds per plant, it takes years to multiply enough seed to cover large areas.

This discourages the private seed sector, but smallholder farmers are taking up the challenge. In Malawi ICRISAT and national partners have passed out hundreds of small (1 kilogram) packs of pure seed of a new high-yielding, preferred-quality variety called CG 7 to women (who largely cultivate this crop) on the condition that they would pass some of their harvest on to neighbors. This chain-reaction scheme has helped CG 7 spread widely across the country.

The concept of ‘foreign aid', which implies repeated handouts of emergency assistance, is being supplanted by the more hopeful vision of ‘development investment.' But are development investors getting a good return on their money? Plant breeding in wetter environments has been shown time and again to deliver value that far exceeds its costs; but what about the dry areas, where gains are harder to come by?

Return-on-investment studies cannot be made on every variety released because they are painstaking and costly. But the few that have been carried out suggest that the effort has been well worth it.

The three-million dollar effort to create and disseminate the early-maturing millet variety Okashana 1, released in 1989 in Namibia, was estimated in 1998 to be returning net benefits worth 50% of the investment, year after year — a rate of return far outstripping that which can be earned in banks or in the stock market, while directly helping society's poorest. Sorghum varieties released since 1975 for the Nigerian food industry were evaluated in 2002 and estimated to be earning a 62% annual return on investment for the period. Sorghum variety S 35 in Chad generated an even more spectacular result: a 95% return on investment per annum, generating about $4 million per year in net benefits to the poor of this impoverished region due to its 50% yield advantage.

Cash value, of course is not the only measure of success; the alleviation of human suffering is a priceless good in its own right. Nor do these studies assess all the indirect benefits of economic growth and employment stimulated by the success of a new crop variety, for example the training of national scientists, and the new employment and profits associated with agricultural input supply crop processing and marketing.

Close engagement with national partners, farmers, NGOs and extension agents is at the heart of the ICRISAT approach, and has been essential for the successes so far. But much remains to be accomplished.

Large areas in Africa have still not benefited from improved varieties. Both soils and crops need to be simultaneously improved in order to feed Africa's growing population. Quality as well as quantity needs to be enhanced in order to maximize farm incomes and improve human nutrition. Seed systems need to be revolutionized so that farmers have ready access to a diverse set of choices and the information on how best to use them. A number of disease and pest constraints remain to be conquered. And training of African scientists must enable them to take over more and more of these challenges. Clearly, there is much to keep everyone busy for many years to come.