4.4 Agricultural research and modern varieties

Contents - Previous - Next

The development and diffusion of improved cultivars constituted the mainstay of increasing yields and production in the recent past and, subject to the considerations discussed in the preceding section, will probably continue to do so over the next two decades. Research efforts related to improved cultivars can be categorized in three classes (FAO/TAC Secretariat, 1993): research focusing on raising the absolute yield ceilings, research related to closing the gap between average farm yields and the yield ceiling, and productivity maintenance research. As discussed in the preceding section, current average yields in most countries and land classes are well below potential yields and in most cases this also holds for projected year 2010 yields. After initial quantum jumps in yield ceilings in the 1950s and 1960s following research breakthroughs, further research led to small yearly increments in yield ceiling. The accumulated effect of such yearly changes over time resulted in many cases in yield gains that were a multiple of the initial one-time increases (Byerlee, 1994; Duvick, 1994). With the possible exception of rice (Box 4.1), the prospects are that production growth based on quantum jumps in cereal yield ceilings will be rarer than in the past. However, the slow evolutionary progress

Box 4.1 Hybrid rice

Hybrid rice has shown a quantum leap in yield in China and hybrid rice seed technology is progressing in other major rice producers such as India, Indonesia and Vietnam. Hybrid rice was introduced to farmers in China in the mid-1970s, with adoption accelerating in the 1980s. In 1992-93 it was grown on 19 million ha in China (65 percent of the country's total harvested area of rice), 10000ha in Korea DPR, and 20 000 ha in Vietnam. It is expected that India's hybrid rice area will leap from 10000ha in 1993 to over 500000 ha in 1996.

Hybrids in irrigated conditions generally yield I tonne/ha (paddy) more than the semi-dwarf modern high-yielding varieties with the same amount of or sometimes even fewer inputs. Many Indian and Vietnamese hybrids have shown a yield advantage of 1.5 to 2.5 tonnes/ha. Evidence suggests that the yield advantage of heterosis is relatively greater in lower productivity areas. The International Rice Research Institute (IRRI) hybrid variety IR64616H has been recommended for on-farm evaluation in the Philippines after three seasons as the highest-yielding early-maturing variety in national trials. At 10.7 tonnes/ha it has the highest yield ever recorded at the IRRI farm.

Progress on hybrid rice seed production technology has dramatically lowered the price of seed in China. Yields of hybrid seed were only 1-1.5 tonnes/ha in 1970s but increased to 2.3 tonnes/ha in the 1980s and 4.5 tonnes/ha in 1991-92 (with a peak of 6.35 tonnes/ha). Because of the high cost of hybrid rice seed, hybrids were only profitable in transplanted rice and in irrigated and favourable lowland rice ecologies during the 1980s. With seed costs having fallen, it is now estimated that hybrids would be profitable on 70 million ha worldwide (nearly half of the 145 million ha total rice harvested area). If the cost of hybrid rice seed, as is expected, falls even further, hybrid cultivation could become profitable in direct seeding (which requires 10 times more seeds than in transplanting) in irrigated and favourable lowland ecologies.

Hybrid rice programmes are underway in Colombia, Brazil and Guyana. A programme to develop less labour-intensive ways to produce hybrid rice has been formulated in Indonesia and Vietnam based on the discovery of photo- and thermo-sensitive genetic male sterility. Different types of hybrid rice are under development and results have been impressive. The future of hybrid rice seed technology could be based on apomixis-asexual reproduction, requiring the cooperation of upstream research (genetic engineering and protoplast fusion) in China and developed countries. With apomictic hybrids, farmers can retain seeds from their crops for many seasons whereas with conventional hybrids farmers must renew seeds for each planting cycle. will likely continue and can underpin the growth of average yields on the basis of research for adapting existing high-yielding varieties to local conditions by overcoming site-specific constraints, e.g. by breeding for increased tolerance to biotic stresses (diseases, insects, weeds) and abiotic ones (soil toxicity, temperature, water).

Seeds produced by the formal seed sector (public agricultural research and the private sector) are referred to as "modern varieties" (MVs), though they are often also described as "high-yielding" or "improved" seeds. The genetic material produced by the system formed the basis of the green revolution.

Figure 4.2 Harvest areas

"Traditional", "local" or "farmers"' varieties (TVs) are indigenous types but include those with some foreign material. Between TVs and MVs there exists a wide range of intermediate varieties embodying improvements in various degrees. The share of modern varieties in total seed use has been increasing rapidly (see below). Nevertheless, traditional varieties remain important because of their particular characteristics as well as a source of genetic diversity. They sometimes have superior performance in particular locations, especially in marginal environments. On-farm selection of traditional varieties has led to substantial improvements in their yields. The loss of varieties and individual genes has in certain cases been accelerated by the application of the products of the modern plant breeding methods and a number of analysts fear that this could be a threat to long-term sustainability of agriculture. On the other hand, a number of authors hold that the development and diffusion of modern varieties have had less impact than commonly claimed on diversity and in the post-Green Revolution period they contributed to increase it (Byerlee, 1994). It is argued that in many cases, in particular in the more productive agroecological areas, genetic diversity in the pre-green revolution period was already low since farmers concentrated on a limited number of varieties which were subsequently replaced by a similar number of modern varieties. In the later stages of the green revolution period and afterwards, the number of varieties that has been and is still being developed has increased rapidly and so has genetic diversity as a result of efforts to breed varieties with desirable traits.

Earlier MVs often lacked resistance to many diseases and insects. Since then, research has concentrated on coping with this problem and the more recent MVs incorporate resistance to multiple pests. In this respect, it is held that they are superior to both traditional and the earlier modern varieties and the effort has generally improved yield stability (Byerlee, 1994). As discussed in the section on plant protection, this trend is expected to continue. Of crucial importance in this respect is the success of maintenance research to continually renew pest resistance in the face of evolving new pest biotypes. Naturally, of equal importance are the economic factors and policies that would induce farmers to replace periodically the varieties in favour of the ones with new sources of resistance.

In the last few decades, MVs of cereals have been rapidly replacing the more traditional varieties (Table 4.12). Nevertheless there are a number of situations where the adoption of modern varieties has been slow or non-existent, e.g. the unfavourable agroecological environments. Despite the progress made in adapting MVs to some of these environments, particularly the success in developing MVs resistant to soil stresses, MVs have not been widely adopted in areas with frequent drought stress or, in the case of rice, in areas with poor water control (FAO/TAC Secretariat, 1993). For the future, little progress in raising yield potentials in these environments can be expected from new breeding techniques given the low rates of return to the relevant research effort. Likewise, MVs have not been widely adopted in areas with poor infrastructure

Table 4.12 Share of area (%) planted to modern varieties of rice, wheat and maize in developing countries

  Rice: Wheat: Maize
  1970 1983 1991 1970 1977 1983 1990 1990
All developing countries 30 59 74 20* 41* 59* 70 57
Africa (sub-Sahara) 4 15 n.a. 5 22 32 52 43
West Asia/North Africa 0 11 n.a. 5 18 31 42 53
Asia (excluding China) 12 48 67 42 69 79 88 45
China 77 95 100 n.a. n.a. n.a. 70 90
Latin America 4 28 58 11 24 68 82 46

Source: Reproduced by permission from Byerlee (1994).
*Excluding China.

"Excludes tall varieties released since 1965. If these varieties are included, the area under MVs increases, especially for rice in Latin America. and lack of market access. These constraints are particularly pronounced for hybrid maize (in sub-Saharan Africa and the hillside systems of Latin America and Asia), the diffusion of which depends heavily on an efficient system for producing and marketing seed.

Sometimes yield advantages of MVs are more than offset by quality advantages of traditional varieties, e.g. Basmati rice in Asia, white flint maize in Southern Africa. The slow spread of MVs, and sometimes return to TVs, are partly attributed to their quality advantages. However, the breeding of desirable quality traits into the MVs is gradually contributing to their wider diffusion. For example, a high-quality MV of Basmati rice was almost completely adopted within a three-year period in Pakistan, after it became available in the mid-1980s. Another feature limiting the adoption of MVs in marginal areas has been the higher straw yields and fodder values of TVs. In dry areas in India, for example, the high value of fodder of the traditional varieties is often considered to have inhibited the adoption of modern sorghum varieties.

Finally, there are areas where traditional varieties still have a yield advantage over MVs. Such areas are often characterized by insufficient local research capacities to do the necessary adaptation work to the local production conditions as, for example, in the eastern and southern savannas of Africa where rainfall is mainly bimodal and drought is a serious problem. For these drought-prone areas, only limited appropriate material is available (Oram and Hojjati, 1994). On the other hand, there are situations where farmers may adopt a lower-yielding modern variety, if it confers other advantages.

Genetic potentials of many staple crops (other than wheat, rice and maize) have increased little in developing countries. The formal seed sector generally has not been responsive to the needs of small farmers and marginal environments. Catering to their needs is often a difficult and expensive undertaking for national seed systems. Recognizing this, local level seed supply is receiving increased support from governments and development agencies. Likewise, in the last decade, national and international breeders have been paying more attention to genetic improvement of the crops that benefited little from the green revolution, the so-called "orphan crops". This shift of emphasis reflects the increased awareness of the importance of these staples to food security in many developing countries, particularly in marginal areas. Efforts to raise yields of "orphan crops", notably sorghum, millet, pulses, roots and tubers, have been frustrated in part because of the unfavourable environments (socioeconomic as well as physical) in which these crops are typically grown. Recent research has offered improved varieties for some environments, usually those with adequate moisture. However, under conditions of low and uncertain rainfall the risk in using such seeds is high and this often discourages farmers from doing so, especially when other inputs are needed to benefit from the improved varieties. Efforts to increase yield stability are therefore often of higher priority than raising potential yield.

Summary renew by major food crops


As discussed earlier, in many countries rice yields realized by farmers are still relatively low. Table 4.12 shows that the adoption of MVs of rice is not yet complete. MVs developed so far have been mostly suitable for irrigated and favourable lowland areas. Differences in adoption rates of MVs are strongly influenced by the degree of water control. For example, in the Philippines and Indonesia production environments are generally favourable and adoption rates are among the highest in Asia, while in India adoption rates are lower than in the Philippines because a greater proportion of its rainfed areas are drought-prone, flood-prone or both. In some countries, including the major exporters, yield advantages of modern varieties are more than offset by the higher prices of traditional varieties (such as Basmati rice) due to their superior palatability. Recent progress in breeding higher-yielding aromatic types may therefore have a substantial impact on rice production in South Asia. In Pakistan and Thailand, MV adoption is relatively low; even in irrigated areas MVs have not been completely adopted because the quality of available MVs has been poor. Semi-dwarf Basmati varieties have only recently been released in Pakistan and India.

The gap in rice yields between irrigated and non-irrigated areas has widened, and is expected to widen further (Table 4.10). Some progress in raising yield potentials in more marginal environments is expected but the main impact on production will be from adaptation of high-yielding semi-dwarf and hybrid types of rice to local conditions in non-marginal environments. Hybrid rice will make possible further yield increases in East Asia and it is also being introduced in India, Vietnam and the Philippines. It is expected that hybrid rice will largely be in substitution for, rather than incremental to, existing high yielding varieties. The extension of hybrid rice to other areas can have a profound influence on production before 2010 (Box 4.1).


Semi-dwarf varieties of wheat with their high responsiveness to fertilizer were key elements of the green revolution and are now estimated to account for some 70 percent of the wheat area in developing countries (Table 4.12). Virtually all irrigated wheat is now based on semi-dwarfs. The spread of these types has been slower in rainfed areas, though their diffusion was facilitated by breeding varieties with resistance to major fungal diseases in the 1970s. About half the area sown to semi-dwarfs in developing countries is now rainfed but the use of these varieties is generally correlated with moisture: they cover around 60 percent of the area in good rainfall environments but only 20 percent in drier areas.

The possibility to lift the current yield ceiling much further appears limited and no hybrid wheat is foreseen. Following the advantages conferred by the first widely successful varieties, more modest but still significant increases in yield have occurred. Since semi-dwarf wheats were released in the 1960s, plant breeders have maintained a long-term average yield gain estimated at 0.7 percent per year. Much current breeding effort is maintenance research, i.e. maintaining crop resistance to changes in insect pests and diseases.


Hybrid maize has been established for at least three decades in developing countries and no other breakthrough is foreseen in the short term to shift the yield ceiling of this crop. Open-pollinated varieties are less expensive to develop and received more attention in the 1970s but improved germplasm for tropical hybrids recently shifted the emphasis in national breeding programmes back towards hybrids. MVs have not yet been adopted everywhere, partly because breeders still have to find improved varieties which meet the needs of particular groups of farmers. More location-specific adaptive breeding of improved varieties and hybrids is still needed. Some progress in reducing the vulnerability to drought during the crucial silking period is however probable in the near future, which would be of major importance in much of the semiarid tropics.

The MVs of maize have been adopted mostly in the more favourable environments and among more commercially oriented farmers. The pattern of adoption is typically a progression from local varieties to improved open pollinated varieties and then to hybrids. The highest rates of adoption are in southern Latin America and the East Asian countries, using genetic material derived mainly from the United States and which is suited to temperate climates. High rates of adoption are also found in East and Southern Africa. The lowest rates of adoption are in Western Africa and the Andean countries of South America, where maize is grown under low fertility conditions and primarily by poor farmers for home consumption. Intermediate levels of adoption are reported in Mexico and in most countries of South Asia.

Although maize benefited from the green revolution, it has shown pronounced geographic variability in yield gains in developing countries which is indicative of the difficulties in raising yields of the so-called "orphan" crops. While wheat and rice are grown under relatively homogeneous agroclimatic conditions where new technologies can be disseminated more easily, maize is grown under a wide range of agroclimatic conditions and improved varieties cannot always be diffused rapidly. Much of the developing world's maize is grown in marginal environments characterized by unreliable rainfall or low soil fertility. Hybrid maize seed generally needs specialized production and distribution facilities which are lacking in many developing countries.

Recent progress in breeding, however, promises to deliver the technical knowledge required to cope better with the difficulties of adapting modern maize varieties to the extraordinary variety of potential growing environments for maize (Byerlee and Lopez-Pereira, 1994). Examples include CIMMYT's work in Mexico's highlands to produce traits suited to highland areas of Asia and Africa (Oram and Hojjati, 1994), and success in adapting cultivars to more intensive farming systems as in the case of rice maize rotations with transplanted or directly seeded maize as a dry season crop. Progress has also been made in the development of cultivars resistant to multiple species of insects. In general, the technical knowledge to tackle the adaptation problem of maize varieties is already available or forthcoming.

Sorghum and Millet

Sorghum hybrids are widely used in South Asia and Latin America. Hybrids have shown good yield potential also in Africa. Hybrid yields are up to 50 percent higher than yields of control varieties, but seed production is still in the experimental stage (ICRISAT, 1991). Improved higher-yielding cultivars are available but they do not seem to express themselves in unfavourable environments. It is widely held that improved soil moisture and fertility status are indispensable for higher-yielding varieties. Breeding in these grains is therefore often aimed at stabilizing rather than maximizing yields. Sorghum varieties with resistance to the parasitic weed striga, which is a severe problem in Africa, have recently been released, and lines are now available with tolerance to acidity and aluminium toxicity, which is especially important on some Latin American and African soils. Identification of the mechanisms of drought tolerance are producing improved millets and some new varieties are less vulnerable to downy mildew, which is a severe disease of pearl millet.

Roots and tubers

The value of roots and tubers lies in their ability to produce large quantities of dietary energy and in the stability of their production under conditions where other crops may fail. They can be grown under a wide range of adverse conditions such as acid soils, fluctuating soil moisture content, high rainfall and infertile marginal lands. Dramatically higher yields can be achieved even with non-improved cultivars through better soil management and crop husbandry. Yields have been shown to double or triple by introducing a single factor such as fertilizer, though fertilizers are not commonly used on such crops despite demonstrated gains with moderate doses of nutrients. Stagnant or even declining yields in developing countries have resulted not only from the lack of high-yielding cultivars and poor cultural practices but also from rapid degeneration due to diseases, especially viruses. With healthy planting material, large yield increases can be realized.

The potential to produce improved varieties is considered high and so far these crops have remained underexploited in genetic terms. In the past two decades, high-yielding cultivars have been bred for cassava, sweet potato and potatoes with resistance to major diseases and insect pests. The production of planting materials from those varieties with stable resistance is relatively easy. However, certain root and tuber crops (yams, aeroides, plantains and cooking bananas) remain vulnerable to insects and diseases. The problem of disease transmission and multiplication through vegetative propagation is perhaps the most important constraint to production increases of starchy staples. The dissemination of healthy planting materials requires establishment of nurseries with efficient distribution systems, which has proven to be difficult to implement. A more realistic approach would be to develop cultivars with inherent resistance to diseases and insect pests which could be easily multiplied and used by growers with traditional methods. True seed, rather than cuttings, generally does not transmit viruses and diseases and can be used where quality considerations, particularly regularity in size and shape, are of less importance.

Contents - Previous - Next