Hybrids require a sufficient capacity for plant breeding and a well­organized seed multiplication and distribution system. In those (parts of) developing countries where these requirement are basely met, it may be doubted whether strong emphasis on hybrids will contribute to increased household food security, as is illustrated by the case of Tanzania. Supplying a wider range of varieties, including hybrids, improved open­pollinated varieties and landraces, could enable especially poor farmers to shift their little resources around according to changing circumstances.
 
Discovery of hybrid vigour has turned out to be of importance for private involvement in the seed sector. Hybrid varieties can potentially out­yield the conventional cross­ or open pollinated varieties. This yield potential combined with farmers' need to buy seeds annually, secures a market for the industry. Since the individual seed company is normally able to protect the parent inbred lines of its commercial hybrid seed from competitors, the industry has a strong incentive to invest in research and development of hybrid varieties as opposed to R&D of non­hybrids.

For a number of technical reasons, hybrids have only been possible for cross­pollinated crops, such as maize, sorghum, millet, sunflower, faba beans, pigeon peas. Recently, the self­pollinated crop rice was added to this list. Hybridization in Africa has been especially important for maize. Parastatal seed companies and national plant breeders have followed the same trend in maize improvement as took place in Western countries and concentrated research and multiplication efforts on hybrids rather than on improving open­pollinated varieties.

Conditions for hybrid seed
Hybrid seed may be an appropriate type of seed, even for resource­poor African peasants, if a number of conditions are satisfied:
(1) The national seed industry should be able to ensure a stable and sufficient supply of hybrid seeds to the farmers. Many African national seed industries suffer from limited capacity due to inefficient management, and shortages of resources, spare parts, etc. Consequently, peasants often experience an erratic supply of hybrid seeds, and peasants unable to buy hybrid seed have to rely on second generation hybrids or on local landraces. Household food security may be threatened if dependency on hybrids has been established and the market fails to supply new seed each year.
(2) The seed industry should be able to supply hybrid seed before the optimal planting time. In the highland areas in Africa, for example, the time between the harvest of the hybrid seed and the optimal time of planting these hybrids is limited to only two to three months. This implies that the national seed companies have to be very efficient in processing the seed and distributing it to farmers. As this is seldom the case in Africa, many peasants experience crop losses due to late delivery of hybrid seeds.
(3) Hybrid seeds need to be sold at a competitive price, for maize typically double the grain price. Because of the high degree of centralization of national seed industries in Africa, and the resulting exorbitant and ever increasing transport prices, hybrid seeds are often unnecessarily expensive for peasant communities.
(4) Hybrid seeds need to be offered in competition with non­hybrid improved varieties and well­established landraces. The plant breeders and seed industries have strongly emphasized hybridization, often at the expense of developing improved open­pollinated varieties. Because of the absence of improved open­pollinated varieties the existing local landraces have been the only alternative to hybrids. The availability and quality of local landraces varies greatly, both geographically and between crops. A wider range of alternative seeds, including improved varieties and established landraces, increases farmers' options to harmonize their seed choice with their limited resources and changing circumstances, in order to ensure household food security. Besides, with respect to biodiversity and sustainable agriculture, the existence of a range of local varieties is important too.

Seed industry
Seed industries in the developing countries are commonly expected to perform two quite different economic functions simultaneously: an equity function of delivering the types and quantities of seed required by different categories of users in a timely manner to appropriate locations at an acceptable price; and an efficiency function of fully recovering the fixed and variable costs of seed production and delivery. They concluded that for the small­farmer seed market it is often impossible to fulfil both these functions simultaneously. Peasant societies are only commercialized to a certain extent, implying that their buying power is subsequently low. A large proportion of the peasants live scattered in areas with a poor infrastructure and long distances separate them from urban and industrial centres.
Private seed companies are commonly found to be reasonably efficient, but they are seldom engaged in operations which are not profitable, such as multiplication of improved varieties of open­pollinated subsistence crops or the seed delivery to marginal areas with low density of demand. Equity goals are often embodied in public seed industries, but poor efficiency often makes these organizations not fulfil their equity goals, despite substantial subsidies.
In Africa, both private and parastatal seed companies have in the past been strongly biased towards hybrid seed production as this guarantees a stable annual demand. The effective demand of a new improved variety of open­pollinated, self­pollinated or vegetatively propagated crops is likely to decline within a limited number of years after release. Peasants become as good as self­sufficient with the variety in question through on­farm propagation.

Hybrids in practice: Tanzania
From the late 1970s to the late 1980s, the Tanzanian Southern Highlands experienced a strong growth in marketed agricultural production. With regard to maize, the Southern Highlands turned from a subsistence economy into a surplus area, with a total marketed production of approximate 250,000 tonnes in 1983.
Three factors that contributed to this are:

• High and rising producer prices for maize, partly realized on the black market.
• The combined use of hybrid maize seed, mineral fertilizers and pesticides supplied at subsidized prices. The cheap access to these inputs made it economicly viable and technically possible to increase yields.
• Peasants were provided with a strong incentive to intensify cultivation due to villagization, the governmental resettlement of all farmers into some 8000 villages in the mid 1970s. This resulted in a break­down of the traditional systems of cultivation based on fallow­rotation.

After the break up of the East African Community in 1977, Tanzania had to buy hybrid seeds from Kenya each year. In the late seventies, samples of breeding material (inbred lines) were brought to Tanzania and became the basis for its own production of hybrid seeds. Three hybrid long­season varieties of dent maize were produced and sold in the Southern Highlands during the 1980s: H614, H632 and H6302. The Tanzanian research establishment and seed industry exclusively focused on hybrid maize seed, and consequently no improved open­pollinated maize variety was bred or made available during this period. Hybrid maize seed, in combination with fertilizer and pesticides was thus the only alternative to continuation of subsistence agriculture using local open­pollinated landraces.
Hybrid maize seed sales in Tanzania fluctuated greatly during the eighties. Since the estimated demand for hybrid maize seeds exceeded supply by more than three times, this fluctuation was closely linked to low production capacity of the parastatal national seed company TANSEED. Detailed research carried out in Njombe District in 1987, elucidates how this deficit supply of hybrid maize seed was distributed and what the consequences were for peasant production.
The generally weak and underdeveloped marketing infrastructure in Njombe District for agricultural inputs, including hybrid seed, underwent successive, turbulent changes in this period. The village­based primary co­operatives were only able to handle one third of the hybrid maize seed sales, resulting in an unequal distribution: a politically well­connected village could receive more than it demanded, while other villages received only a fragment of their requirement. The remaining two­thirds of the seed supply was sold directly from the warehouse of TANSEED's processing plant or by private shops in Njombe town. All hybrid maize seed was sold within a few days after release, and mainly went to the villages closest to Njombe town or with access to transport and a progressive chairman.

Impact on peasant production
What were the consequences for peasant production of this erratic supply of hybrid seed, in a situation where no improved open­pollinated maize seed was available and in which farmers were under strong pressure to intensify agricultural production? Two points could be highlighted:
(1) A common strategy for those farmers who lack annual access to the limited supply of hybrid seed, is to cultivate second or later generation of hybrids combined with landraces. The yield of hybrids will decrease considerably, although the yield of a second generation hybrid may still be higher than of local landraces. More important is, however, that also other genetic characteristics will deteriorate, such as its 'original' resistance to pests. Dependency on hybrid maize when the seed supply and marketing are irregular results in increased vulnerability of food production.
(2) Diffusion of hybrid maize in the Southern Highlands took place gradually. Most households using hybrid maize seed also sow local landraces and had separate fields with different varieties. Where fields of local landraces of maize are cultivated close to fields of hybrid maize, cross­pollination occurs. The outer rows in the field with landraces will be pollinated from the hybrid crop when flowering. Since mass selection of seed of landraces for the coming season is done from the stored maize, part of the selected cobs are likely to be those which are cross­pollinated with hybrids. Cross­pollination from hybrid to local maize varieties changes the characteristics of the landraces.
One of the important characteristics of landraces in Njombe is the time to maturity. The common practice in Njombe was to plant the long­season hybrid maize immediately after the first rains and plant the short­season landraces thereafter. When seeds of landraces that were no longer pure after they had been cross­pollinated by hybrid varieties, were planted late, part of the field failed to mature before the end of the season. It was common in the late 1980s to see fields of 'mixed' landraces, of which a quarter of the field had not matured by the time of harvest. Some farmers were conscious of cultivating local landraces isolated in time and space to ensure their purity. Most farmers, however, were unaware of cross­pollination from hybrids to landraces, and were simply worried about what they called the declining quality of landraces.

Many things have changed in Tanzania since the economic liberalization and structural adjustment policies were introduced in the late 1980s. The seed industry has been partially privatized and efficiency has been improved considerably. Meanwhile the removal of subsidies on fertilizer and transport costs has made it economically unviable to produce maize in the more distant parts of the Southern Highlands for the market in Dar es Salaam. Whether the private seed industry will be able to supply farmers in Tanzania timely with the types and quantities of seed required, including hybrids and open­pollinated varieties, and short and long­season seeds, to appropriate locations and at acceptable prices, is yet to be seen.
Esbern Friis­Hansen

Sources
E. Friis­Hansen (1988), Seeds of Wealth ­ Seeds of Risk? The vulnerability of Hybrid Maize Production in the Southern Highlands of Tanzania. CDR Project Paper 88.2. Copenhagen: Centre for Development Research.

E. Cromwell, E. Friis­Hansen, and M. Turner (1992), The Seed Sector in Developing Countries: A framework for performance analysis. ODI Working paper 64. London: Overseas Development Institute.