PROSEA, Introduction to Cereals

Plant Resources of South-East Asia Introduction

List of species

Introduction

Definition and species diversity

Choice of species

Strictly speaking the definition of cereals is limited to Gramineae species with edible starchy grains. In this volume the definition is widened to include those grain crops cultivated for their starchy seed which is used as a basic food for humans, as a fodder for domesticated livestock or as a raw material for industrial purposes. Most starchy grain crops are "true" cereals belonging to the Gramineae family. By far the most important one for South-East Asia is rice (Oryza sativa L.), followed at some distance by maize (Zea mays L.). The other true cereals described in full articles in this volume are grown very locally and in much smaller quantities compared with rice or maize. Cereals like sorghum (Sorghum bicolor (L.) Moench), proso millet (Panicum miliaceum L. cv. group Proso Millet) and wheat (Triticum spp.) have some economic importance in South-East Asia, whereas the others like pearl millet (Pennisetum glaucum (L.) R. Br.), finger millet (Eleusine coracana (L.) Gaertner cv. group Finger Millet), foxtail millet (Setaria italica (L.) P. Beauvois cv. group Foxtail Millet), barnyard millet (Echinochloa spp.), Job's tears (Coix lacryma-jobi L.), rye (Secale cereale L.), and barley (Hordeum vulgare L.) have been included because they are produced locally on a minor scale, or are being investigated at research institutes as potential new crops. Three non-graminaceous crops, the so-called "pseudo-cereals", have been added to this group of major species, namely grain amaranth (Amaranthus spp.), grain chenopod (Chenopodium spp.) and buckwheat (Fagopyrum esculentum Moench), because they have the potential to become economically important in South-East Asia. These pseudo-cereals are dicotyledons, which differ from the Gramineae in many aspects. They are mentioned only occasionally in this introductory chapter, but are briefly described in general terms in section 1.10.

In addition to the major species, Chapter 3 briefly describes 9 minor species (8 graminaceous and 1 non-graminaceous), which could be or are actually used as emergency food in times of scarcity. Chapter 4 lists those cereals whose primary use is not as a grain crop, and gives the appropriate references to other volumes of the Prosea handbook.

Domestication

There is archaeological evidence that in the Palaeolithic, 30 000 years ago, primitive people were already collecting dry seeds of grasses. Barley and wheat have been found in the delta of the Nile and the Euphrates, in the Fertile Crescent, in excavations of settlements dating from around 8000-7000 BC. During the Neolithic period (7000-3000 BC) wild types of wheat, millets, barley, sorghum, maize and rice were collected and used as energy food in many places. Domestication and cultivation started gradually at several locations in the world. In the fourth millennium BC, cereal cultivation was already widespread in the Mediterranean area, Western Asia and Western Europe (Evans, 1993; Leakey & Lewin, 1977).

How domestication started can be deduced from the gathering and cropping practices of traditional people in remote areas. Grains of certain grass types with big kernels of good eating quality, the ancestors of modern cereals, would have been collected by sweeping a basket through the dry grass vegetation. Small-seeded types such as millets and amaranths were also harvested, because of their ability to yield large quantities of seed from a dense vegetation. Harvesting in this way, one person could collect over 10 kg of wild cereal seed per day, and the surplus could easily be stored to bridge the period to the next harvest. The next logical step was to protect this grass vegetation and to weed selectively by removing competing broad-leaved plants. Spilled seed will have germinated in compounds or on land near dwellings where the vegetation was burned to stimulate grass growth for livestock.

The distinctive step marking the beginning of agriculture was the intentional saving and actual sowing of part of the collected grain. Aboriginals in Australia, while harvesting wild millet, sometimes scatter some seed to stimulate regrowth. Only the haulms of plants with non-shedding seed will have been picked, and therefore this characteristic plant type will have been selected unconsciously for all cereals and pseudo-cereals. Deliberate roguing and the discarding of plants or seed with undesirable characters marked the beginning of plant breeding. This went hand-in-hand with the refinement of cultural practices. However, this made the incipient landraces more vulnerable to predators and more dependent on the farmer's protection (Brown & Eckholm, 1975; Evans, 1993). Over hundreds of generations humans learned how to grow a good grain crop and that it was possible to improve yield and quality by keeping the most productive plants with desirable traits for the next sowing. In this way, wild grass species were gradually developed into new crop species. They differed from their wild ancestors in aspects such as the size of the grains, the non-shattering seed, the free-threshing glumes, reduced dormancy and other desired traits (de Wet, 1986; Riley, 1975).

Geographic distribution

Irrigated rice cultivation probably developed in China during the Neolithic period 5000 years BC. Already in antiquity it spread from South China to all over South and South-East Asia, and in recent centuries it has spread westwards to central Asia, tropical and subtropical Africa, the Mediterranean area, and to South, Central and North America. Maize originated in Central America, spread to the north and the south, and went eastwards all over the world up to the east coast of Asia.

Although rice, maize, sorghum and millets are thermophile crops and wheat and barley are more at home in cooler climates, the overall adaptation to other environments is noteworthy (Figure 1). Rice is grown from the equator to northern China at 50 °N. Wheat is produced from Alaska, Norway and Siberia at 65 °N, almost at the Arctic Circle, to the tropical areas of 20 °N in Asia, while in the highlands of Africa and Latin America it touches the equator.

Role of cereals

Social and cultural aspects

Cereals, food crops of eminent importance for humanity and the backbone of human nutrition derive their English name from Ceres, the Roman goddess of the most basic human need: daily bread. They are interwoven with all aspects of human civilization. When around 2600 years BC pharaoh Zoser in ancient Egypt engaged thousands of workers to build his tomb, the Step Pyramid of Saqqara, he was only able to do so because he had a well-organized supply of emmer wheat (Triticum turgidum L.) from the farms in the Nile delta. In ancient times, the most advanced civilizations could arise only in those areas where suitable growing conditions, paired with inventive cropping practices, guaranteed a high and sustainable cereal production. Examples are the wheat-based civilizations of Ancient Egypt, Mesopotamia and the Roman Empire, and the rice-based civilizations of South-East and East Asia. For thousands of years the Incas and Aztecs in Latin America cultivated maize as a staple food together with the pseudo-cereals amaranth and quinoa (Chenopodium quinoa Willdenow), using them in religious ceremonies too.

Even today the planting and harvesting of cereals is accompanied by feasts and ceremonies. Everywhere in South-East Asia the rice harvest is a happy occasion and the people are thankful to the gods because in the coming period their staple food is guaranteed (White, 1994). The Javanese rice goddess Dewi Sri who guards life and fertility has much in common with the Roman goddess Ceres. And every year citizens of the United States celebrate Thanksgiving Day, a tradition started by a group of European settlers in gratitude for their first wheat harvest.

Why are cereals so popular as a staple food? They have attractive traits: they are compact, energy-rich, easy to grow, store, prepare, carry, trade and ship. Most governments attach great importance to being self-sufficient in cereals, because dependence on imports is considered politically and economically undesirable. Not surprisingly, cereals have always been a trump card in international politics.

Uses

The many uses of cereal grains and by-products can be grouped into four major categories: human consumption, livestock feed, raw material, and subsidiary uses.

Human consumption

Cereals are consumed after the husked kernels have been processed in some way. The whole grain of rice, maize, sorghum, millets and sometimes also of wheat is boiled or steamed and eaten as such or used in soups or to make porridge. The flour obtained by grinding or milling the kernels is used to prepare bread, porridge, pancakes, pasta (macaroni, spaghetti, noodles, couscous) or pastries. Wheat flour, the main ingredient for baking bread, is often mixed with other cereals for economic reasons. Maize, sorghum and millet flours are usually boiled when used as a main dish. Amaranth and quinoa grains are toasted or ground into flour. Various kinds of fermented dishes are prepared from whole kernels or flour.

Livestock feed

Grain and derivatives of grain from all major cereals (called concentrates) are widely used as feed for livestock. Not only the whole kernel but also the waste from the grain-processing industry is widely used as feed. Waste products from rice-polishing or wheat-grinding (bran or meal) are rich sources of protein and macro-nutrients. The straw of many cereals is a popular feed for ruminants. It is used for hay and fodder. In temperate regions it is common to cultivate cereals (mainly maize) and harvest them when immature for use as forage or silage, but this practice is scarcely known in the tropics. There are special forage cultivars of maize and sorghum with a high protein content (2-3%) and excellent digestibility (70%).

The conversion rate of cereal protein into animal protein is about 23% for eggs or milk, 18% for chicken, 12% for pork and 4% for beef. Under conditions of scarcity, animal products from animals fed with cereals are considered a luxury food, and direct consumption of cereals as human food should be encouraged.

Raw material

Cereal grain is an important raw material for many products. It is used to prepare beer and liquors and for diverse industrial products: starch, alcohol, glue, dextrin. In the United States, maize starch is used for the large-scale production of a sugar substitute called "High Fructose Corn Syrup", obtained by enzymatic transformation. The introduction of this technology in the beverage industry (aerated drinks, fruit juices) was a major cause of the collapse of sugar cane production in the Philippines in the 1980s. Oil extracted from maize and sometimes also from rice bran is used for cooking or industrial purposes. Rice husk is a raw material for the preparation of building board. Straw of all cereals is used for making cardboard, for weaving mats and for decorative items. Pulp from straw is becoming increasingly important as a substitute for wood pulp in paper manufacture. The manufacture of degradable bioplastics is a new and promising use of starch.

Subsidiary uses

Cereal straw is commonly used for bedding and mulching in vegetable production. Maize and sorghum plants are used as living supports for climbing beans. Straw is used to make compost, which in turn is used as a substrate for mushroom cultivation. Straw is also applied as litter or bedding in livestock sheds. The straw of wheat and rice is a popular material for thatching. Maize and sorghum stalks are harvested for fuel, as are rice husks. As wood becomes scarcer, the use of straw for fuel will increase. Rice straw rope is used in religious ceremonies. Cereals are sown as green manure and as cover crop. Green inflorescences are mixed in bouquets with cut flowers, and dry inflorescences are popular in dry bouquets.

Nutritional aspects

Relative importance

Cereals constitute the main source of energy in human nutrition. On average, direct consumption of cereals provides 56% of the world food energy (rice 21%, wheat 20%, maize 5%, other cereals 10%), compared with livestock products and fish 11%, vegetables, fruits and nuts 10%, fats and oils 9%, sugar 7%, roots and tubers 7% (Brown & Eckholm, 1975). Moreover, a large part of the feed needed for livestock is derived from cereals. The income elasticity for cereal consumption is negative; direct consumption of cereals decreases with increasing wealth, whereas the consumption of livestock products and other more expensive "luxury" foodstuffs increases. Most people perceive meat and dairy products as being more palatable than rice or bread. About two-thirds of the cereal production of western countries is used for the production of meat, eggs and dairy products. In South-East Asia, the proportion of cereal used for livestock feed is still rather low - possibly about 10% - but it is increasing rapidly, in line with earlier developments in industrialized countries.

Composition

The chemical composition of cereals and cereal products is presented in Table 1, in which foodstuffs from other commodity groups have been included for comparison. Cereals are an important source of carbohydrates, proteins, B vitamins, minerals and dietary fibre. The low fat content is an advantage for the keeping quality of the flour. The consumption of non-refined cereals (high-extraction flour) is a particularly healthy practice. Yet a cereal-based diet with insufficient other foodstuffs will lead to nutrient deficiencies because of the lack of certain essential amino acids and vitamins. In a well-balanced diet the energy-supplying staple food of cereals is complemented with pulses, livestock products, vegetables, fruits and other foodstuffs which compensate the nutrients deficient in cereals.

On a fresh weight basis, the differences in energy value between the commodity groups are rather large. However, if compared in terms of dry matter, all cereals, pseudo-cereals, pulses, roots and tubers have about the same energy value. The popular method of fine milling, followed by fractionation and removal of coarse parts, or, for rice, the practice of polishing, means that a certain percentage of the grain is lost for human consumption. Moreover, the composition of the flour is less nutritious than the composition of the whole kernel or the remaining coarse part, which is destined for feed. In industrialized countries, many diets are based upon refined cereal products supplemented with much sugar, fat, meat, eggs and dairy products. This type of diet supplies large amounts of energy and protein, but is deficient in micro-nutrients and fibre and contains too much fat. It is associated with a high incidence of constipation, cancer of the bowel, heart attacks and other diseases.

Carbohydrates

Cereals contain about the same content of carbohydrates as roots and tubers. Carbohydrates constitute the bulk of the grain, about 80% of the dry weight. They are mainly stored as starch in the endosperm cells, and only a minor part is present as free sugar. Starch is a polysaccharide (C₆H₁₀O₅)n. Cereal starch is roughly composed of 25% amylose (15-30%) and 75% amylopectin (70-85%). Its digestibility is improved by heating. The ratio of amylose to amylopectin strongly affects the palatibility and the industrial use. Amylose is water-soluble, and the industrial separation of amylose and amylopectin is based on this property. Amylopectin gelatinizes in hot water (60-80 °C); the higher the amylopectin content, the higher the glutinosity. The endosperm of waxy cultivars of rice, barley, sorghum and maize lacks amylose (Southgate, 1988).

Dietary fibre is mainly present as non-starch polysaccharides from cell-wall structures (indigestible cellulose, hemicellulose and lignine). The highest concentrations are found in the outer bran layers of the grain. The lower the extraction rate, the lower the fibre content of the flour. Dietary fibre stimulates peristalsis.

Protein

Protein is mainly present in the endosperm (about 70% of the total) but the highest concentration is found in the aleurone layer and the germ. Even within the same species the protein content of different samples may show large variation, depending on cultivar, growing conditions and cultural practices (mainly on N fertilizer). The variation is largest in maize (6-15%). The biological value of cereal protein is rather low compared with egg protein or meat protein because the content of some essential amino acids (mainly lysine, methionine, cystine; and in maize also tryptophan) is limiting. Lysine is present in high concentrations in the germ. High-extraction flour (for which almost the whole grain is used) has a higher protein content and a better biological value than low-extraction flour. The biological value of protein from the pseudo-cereals amaranth, buckwheat and quinoa is higher than that from the true cereals. Some of the cereal proteins (gliadin, glutenin) are insoluble in water, producing the "gluten", a swelling, sticky, elastic compound, when the meal is wetted and kneaded. Carbon dioxide bubbles formed in the dough during fermentation are retained by this gluten, thus bringing about the porous crumb structure of bread. Only wheat and, to a lesser degree, rye contain enough gluten to enable this type of bread to be baked.

Vitamins and micro-nutrients

Cereals are an important source of B vitamins, mainly thiamine (B1), riboflavine (B2) and nicotinic acid or niacin (PP). Polished rice contains considerably less thiamine than home-pounded or parboiled rice. The thiamine is present in the outer layers of the grain. Beri-beri disease, caused by a deficiency of thiamine, claimed many victims in Java in the early 20th Century when the use of polished rice became widespread. The symptoms of this disease are nausea, neuritis, muscular convulsions and cardiac disturbances. The riboflavine concentration is highest in the germ. Riboflavine deficiency causes irritation of the mucocutaneous junctions of eyes and lips and desquamation of the skin. Nicotinic acid is present in wholemeal cereals, but in maize the content is low. It is associated with the amino acid tryptophan. In a normal diet, part of the tryptophan is converted into nicotinic acid. Deficiency of tryptophan, and thus of nicotinic acid, causes pellagra, a dangerous skin disease which occurs in areas with a monotonous maize diet. Vitamin A and vitamin C are generally lacking in cereals. Yellow maize is the only cereal with a significant vitamin A content. Durum wheat and proso millet contain a small quantity of carotenoids.

Cereals contain many inorganic constituents, mostly in the germ and the outer layers of the grain. Only zinc and iron are considered as important contributions to the diet. In low-extraction meal (white wheat flour) a large part of these valuable micro-nutrients has been lost.

Economic aspects

In the South-East Asian countries the first priority is to safeguard food production for the nutrition of the rapidly growing population. Because cereals are the most important staple food, governments follow a strategy to become or to remain self-sufficient - especially in rice.

Tables 2 and 3 present some figures derived from national statistics collected by FAO (1994; 1995). These figures are reasonably accurate for the larger cropping areas of main cereals, but are less reliable for the minor cereals, especially those used for subsistence.

The global average yield of cereals has increased tremendously during recent decades, from 2.19 to 2.83 t/ha. This is an increase of 29% in the period from 1979-1981 to 1994, and is mainly attributable to higher yields of paddy rice (which has risen from 2.74 to 3.65 t/ha), wheat (from 1.86 to 2.45 t/ha), and maize (from 3.34 to 4.33 t/ha). Yields of barley have increased considerably (from 1.90 to 2.19 t/ha), whereas sorghum (1.46 versus 1.39 t/ha) and millets (0.68 versus 0.69 t/ha) have remained at the same level. The increase in yield undoubtedly reflects the research efforts and the increased area planted with improved cultivars.

Table 2 shows that there was a large increase in production in South-East Asian countries between 1979-1981 and 1994.

The total population of the 9 South-East Asian countries increased from 360 million in 1980 to 476 million in 1994 (a 32% increase). During this period, annual cereal production rose from 95 million t to 140 million t (an increase of 48%), the harvested cereal area increased from 43.1 million ha to 47.2 million ha (an increase of 9.2%), and the yield rose from 2.20 to 2.96 t/ha (an increase of 35%). In spite of the extremely rapid population growth in South-East Asia, the overall effect is a significant improvement of the general food situation: the annual per capita availability of cereals increased from 263 kg in 1980 to 293 kg in 1994. These figures are not solely related to direct human consumption in South-East Asia; probably around 10% are used for other purposes. Part of the production is exported outside the region, an important portion is used for feed and industrial use, and a minor part is saved for seed.

Rice is by far the leading cereal in South-East Asia. Other cereals are mostly planted as secondary food crops or used for feed or industrial purposes. The average yield of paddy rice for all South-East Asian countries is 3.06 t/ha, which is 15% below the world average. The highest score is for Indonesia (4.34 t/ha), the lowest is for Cambodia (1.06 t/ha).

Maize is the second most important cereal, covering only about 20% of the area under rice. Indonesia, the Philippines and Thailand are major producers, whereas Vietnam and Burma (Myanmar) also produce a considerable amount. Maize is often planted as a secondary crop after rice, or as the main cereal crop where irrigated rice cannot be grown. About 20% of the production is used for animal feed, and the demand for feed is still increasing. The average yield of maize for all South-East Asian countries is 2.10 t/ha, which is 51% below the world average. The highest score is for Thailand (3.17 t/ha), the lowest for Cambodia (1.32 t/ha).

Sorghum is an important crop in Thailand, but is rather insignificant in the rest of South-East Asia. It is eaten mixed with rice or replaces rice in times of food shortage, but its more important uses are as feed and raw material.

Wheat, globally the leading cereal, is a minor crop in South-East Asia, although it is rather important in Burma (Myanmar). It is grown on small areas in the north of Thailand and in North Vietnam. The major constraints to wheat production in the region are climatic conditions and disease problems resulting from high relative humidity.

Burma (Myanmar) is the only South-East Asian country producing millets to a reasonable extent. Barley is grown on a very small scale in Thailand. These cereals have little economic importance in other countries in the region. Minor cereals and pseudo-cereals are generally not recorded in the statistics. They may be cultivated solely by certain ethnic groups, mainly for subsistence.

Cereal production in Burma (Myanmar) increased from about 13 million t in 1979-1981 to 19.6 million t in 1994. Rice is the dominant crop, but maize, wheat, millets and other cereals are becoming increasingly important. The only cereals Burma (Myanmar) imports are wheat and flour (in relatively small amounts). Rice is an important export commodity.

The harvested areas of rice in Cambodia and Laos in 1994 were respectively about 1.7 and 0.6 million ha, with production being 1.8 and 1.6 million t respectively. Import of cereals in Laos (mainly rice) had fallen to about 8000 t by 1993, while in Cambodia it had increased to 106 000 t by that year. Indonesia became self-sufficient in rice in 1984; however, in some years there is a relatively modest shortage. The country imports an increasing amount of wheat and wheat-flour mixtures, totalling 2.6 million t in 1993. With a yield of 2.6 t/ha, a cropping area of about 1 million ha would be needed to meet the demand! Recent Indonesian studies have shown that East Timor has potential for wheat cultivation. The import of maize for feed is increasing. There is a minor import of barley for breweries.

Total cereal production in Malaysia, which is completely dominated by rice, remained stagnant at about 2.1 million t from 1979-1981 to 1994. However, imports of wheat, maize and other grains are steadily increasing.

Papua New Guinea hardly produces any cereals for domestic consumption; it relies on imports.

Cereal production in the Philippines increased from about 10.9 million t in 1979-1981 to 15.6 million t in 1994. Small amounts of cereals were exported, but wheat and wheat-flour mixtures were imported at an increasing rate.

Rice has traditionally been an export commodity of Thailand. Its production increased slightly from 17.0 million t in 1979-1981 to about 18.4 million t in 1994. The cultivated area has decreased but yields have increased. Cereal production as a whole increased from 20.3 million t to 22.6 million t in the same period. Thailand's total rice exports in 1993 were about 5 million t, valued at US$ 1300 million. Thailand also exports a considerable quantity of maize. Similar to other South-East Asian countries, Thailand's import of wheat and wheat-flour mixtures has sharply increased.

In Vietnam, rice production increased sharply from about 12 million t in 1979-1981 to 22.5 million t in 1994. Vietnam has become a rice exporter in recent years, but its wheat imports are increasing.

Botany

Taxonomy

All true cereals are monocotyledons belonging to the Gramineae, the grasses, one of the largest plant families with about 650 genera and 10 000 species. Gramineae are distinguishable from other families by their typical long and narrow leaves alternating in two opposite vertical rows, their cylindrical stems (culms) with conspicuous nodes and their internodes which are hollow or filled with soft tissue. The inflorescence is usually terminal on the culm and its basic units are termed spikelets (functionally comparable to the flowers of other families).

The classification of the grasses has not yet been agreed and no wholly satisfactory overall account is available. At present the family is subdivided into 6 subfamilies with 40 tribes. Important characters for classification are the spikelet structure, the internal leaf anatomy, the photosynthetic metabolism, the basic chromosome number and the embryo structure. The currently widest accepted classification of the cereals described here is summarized in Table 4.

Morphology (see Figures 2 and 3)

Cereals are tufted, annual herbs with erect stems (culms). Usually several erect branches (tillers) arise from the axils of leaves at the base of the stem and behave ultimately like true stems.

Seminal roots arise from the germinating embryo but are very soon replaced by an adventitious nodal root system that arises from the culms and tillers. The roots are fibrous and usually penetrate 1-2 m into the soil. Maize also has stout prop roots arising from the lower nodes above the soil surface.

The stem is cylindrical, with elongated, usually hollow internodes (but pith-filled in maize and sorghum), connected by short, harder, disc-shaped solid nodes that look very different from the internodes and from which the buds and leaves originate. The lower internodes usually remain very short. The internodes have a soft meristematic zone immediately above the nodes which has the important function of bending the stem upright again by differential growth when the plant has lodged because of rain or trampling.

The leaves are arranged in two rows along the stem and consist of a sheath, a ligule and a blade. The sheath clasps the stem tightly and gives mechanical support to the meristematic zone of the internode. At its upper end the sheath passes into a parallel-veined, typically long and narrow blade, which also has a meristematic zone at its base permitting it to continue its growth despite the removal of its distal parts by e.g. grazing or cutting. A short membranous or ciliate rim, the ligule, is present at the junction of sheath and blade. Its function might be to prevent rain entering the sheath. The base of the blade or the top of the sheath often bears auricles (ear- or teeth-like appendages).

The inflorescence is a specialized leafless branch system, which usually terminates the stem.

Morphologically, its basic units, called spikelets, are partial inflorescences, but functionally they can be compared with the flowers of petaloid plants. The spikelets are arranged in various ways, ranging from a single spike or raceme through an intermediate stage of several spikes arranged digitately or along an axis to a many-branched panicle. The spikelet structure is essential for the identification of grasses; on the outside two opposite rows of scales are visible, arranged alternately along an axis (rachilla); the two lower scales, the glumes, are empty, but the remainder form part of a floret, whose floral parts are enclosed by the lemma on the outside and a delicate membranous scale, the palea, on the inside; glumes and lemmas often terminate in one or more long stiff bristles termed awns; the floral parts consist of two or three tiny scales, the lodicules (they have a function in opening the floret by their ability to swell), three stamens (6 in rice) each with a delicate filament and a 2-celled versatile anther and a pistil consisting of a superior single-loculed ovary with a single ovule and 2 styles, each ending in a feathery stigma. This basic pattern of a spikelet structure may be modified by reduction, suppression or elaboration. Bisexual spikelets are the rule, although some of their florets are often unisexual or barren. Separate male and female spikelets are occasionally borne on the same plant, rarely on separate plants. The florets open for only a few hours to expose the sexual organs to wind pollination. Cross pollination is usually ensured by protandry; the pollen is viable for less than a day.

The fruit (grain) is usually a caryopsis, with a thin pericarp adhering firmly to the seed; sometimes it is a utricle (with free soft pericarp), e.g. in finger millet, or an achene (with free hard pericarp). The fruit is surrounded by the husk, which consists of the hardened palea and lemma. The husk of wheat, maize, millet and some sorghum types is easily removed by threshing, but the husk of barley and rice is tenacious and is traditionally removed by pounding and winnowing (nowadays by machine). The outer layers of the caryopsis are (beginning at the outside): the outer and inner pericarps (derived from the ovary), the true seed coat or testa and a protein-rich aleurone layer.

The seed consists of an embryo at the base of the abaxial face and a starchy endosperm. The embryo is peculiar, with no homologue among the angiosperms; it has a flat haustorial cotyledon (the scutellum) and a special outer sheath (the coleoptile) protecting the plumule during soil penetration.

Growth and development

Stages of growth and development

Unlike the cultivation of root and tuber crops or vegetables, cereal production implies completing the whole life cycle from seed to seed. The following is a practical subdivision of the life cycle in externally visible growth stages:

• vegetative phase: leaf initiation, tillering, flower initiation;
• reproductive phase: stem elongation, spikelet and floral development, anthesis;
• ripening phase: grain filling, grain ripening.

Standardized development stages of the leading cereal crops have been defined on the basis of visual observations. For rice a scale is used in which the development stages of photoperiod-sensitive rice are defined, including a time range adaptable to cultivar and location (Stansel, 1975). The Decimal Code (Zadoks et al., 1974) is widely applied to most cereals, including transplanted rice, but not yet to maize and sorghum. Crop development is divided into 10 stages, each of which is subdivided into a maximum of 10 steps (Table 5). When observing a plant population, the criterion for assigning it to a certain stage is that more than 50% of the individual plants show the characteristic in question.

During the vegetative phase, tillering is of paramount importance, as it is the natural mechanism enabling individual plants to correct for suboptimal planting density, in order to attain an optimal crop canopy. Tillers produce fewer leaves than the main stem, and this tends to synchronize their development with that of the main shoot. The number of tillers further depends on species and cultivar, and increases with applications of nitrogen-rich fertilizer and shallow sowing. Modern maize cultivars have no or very few tillers; therefore, the crop has no buffer for correc¬ting poor emergence and thus correct sowing density is crucial.

The real start of the reproductive phase (inflorescence initiation in the apex which is still at ground level) can only be observed by studying the apex, split lengthwise, under the microscope (Kirby & Appleyard, 1981). In most cereals meiosis takes place during booting (Decimal Code 41), but in rice meiosis occurs earlier (Decimal Code 39). Environmental stress during inflorescence initiation or during meiosis easily leads to a serious reduction of yield. Anthesis may last one to three weeks for the whole crop of a self- or cross-pollinating cereal.

The transport of assimilates to the grain ceases when the dry matter content reaches about 60%. This is the start of the ripening phase. Harvest-ripe seed (Decimal Code 92) has a dry matter content of 80-90%. Mature seed, well dried in the field to a moisture content below 14%, will germinate easily when wetted by rain. Some cereals have a short dormancy period: indica rice 1-3 months, some millets a few weeks, sorghum 1-4 weeks.

Source-sink relationships

The leaf canopy is the driving force for interception of solar energy (the source), while a sufficient number of panicles must provide the storage capacity for the assimilates (the sink). Variation in rate and extent of canopy development, expressed as Leaf Area Index (LAI), is one of the main causes of yield variation. Canopy development and decline during the life cycle of a crop are strongly influenced by environmental factors (temperature) and the availability of water and mineral nutrients. The green leaf area including the green inflorescence are the "source" of the flow of assimilates to the "sinks". Substantial amounts of assimilates have already been stored in the stem and leaf-sheaths before and during anthesis. Grain filling can continue to some extent under adverse conditions because of the translocation of these reserves. Stems and leaf-sheaths are an additional source of reserves for the main sink, the grain. Under normal conditions, these stored reserves also gradually flow to the maturing grain and continue to do so after the leaves senesce. Nitrogen, an important nutrient for achieving a high protein content in the grain, is mainly taken up before anthesis.

Understanding the source-sink relationships is helpful to overcome imbalance in yield-determining factors, and consequently to ensure good grain yield. The grain yield may be either source-limited or sink-limited. When it is source-limited, increased availability of assimilates results in higher yields, indirectly by an improved build-up of plant parts, directly by a better grain filling. Source-sink relationships between all plant parts exist throughout the plant's lifetime, starting with the germinating seed which functions as source for the young shoot and radicle. Higher yields are obtained by reducing all possible constraints to source and sink functions.

Yield-determining factors

Agronomists use the following simple formula to analyse grain yield:

grain yield (g/m²) = plant density (plants/m²) × inflorescences per plant
× grains per inflorescence × mean grain weight (g).

Plant density and number of well-developed inflorescences per plant are inversely related. The inflorescence may be divided further into the number of spikelets per inflorescence and florets per spikelet. The number of florets per spikelet is usually rather stable (in wheat: up to 9), but only a small proportion of the florets are fertile (in wheat: 2-5). The average weight of grain varies greatly, and is a function of the total quantity of assimilates transported to the grains and the number of grains/m².

The potential yield of a grain crop depends on its capacity to use incident solar radiation for the production of grain. The potential grain yield (Y) is determined by the following parameters (Hay & Walker, 1989):

Y = Q × I × ε × H

• Q = quantity of incident solar energy over the crop period. Depends on daylength and weather, varies with latitude, season and region.
• I = intercepted fraction of Q. Depends on leaf area and morphological characteristics (canopy structure) that determine light extinction. In practice, this factor is a major cause of yield variation.
• ε = efficiency of conversion of radiant into chemical energy, expressed as dry matter produced per unit of intercepted radiation. A high ε is an interesting target for breeders.
• H = harvest index, i.e. harvested (useful) product as a fraction of total (aboveground) dry matter.

Modern cereal cultivars have medium-short, stiff straw to withstand lodging and to tolerate high nitrogen gifts, and an adequate canopy structure to maximize light interception. They have a high yield potential, translated into increased total dry matter production, more inflorescences per unit land, larger inflorescences, improved floret survival, and heavier grains. An example is the yield of winter wheat cultivars in western Europe. The long-straw wheat cultivars dating from before 1900 had a harvest index of only 35% and yields of about 2 t/ha. At present, the best farmers achieve grain yields of about 10.5 t/ha with a harvest index slightly above 50%.

How far can the genetic yield potential be raised? It seems possible to further increase the harvest index (H), thus the potential yield, to some extent, but not by a spectacular amount. Consequently, there seems little scope for further genetic improvement of the yield capacity of cereals. However, yield can be increased dramatically by improved cultural practices.

Another question is whether root and tuber crops surpass cereals in energy capture. The total dry matter production of these two types of food crops is very similar, but the root and tuber crops have the potential to produce a much larger amount of edible dry matter per ha and per day than cereals because of their favourable harvest index (up to 80%). However, the gap between this yield potential and the yield realized on the farm is larger than for cereals, the main reason being a lack of dissemination of research results. At present, average world yields (edible portion) of rice and maize in terms of energy per ha per day are superior to cassava (Manihot esculenta Crantz) and yams (Dioscorea spp.). Only sweet potato (Ipomoea batatas (L.) Lamk) appears superior to cereals (de Vries et al., 1967). The advantage of cereals over non-cereal energy crops is not their yield of edible energy per crop, but the fact that a moderate energy yield may be secured in a relatively short cropping period, and that cereal grain can easily be stored due to its compactness. Moreover, several ecological and socio-economic factors such as climate, soil, irrigation water, need for labour and inputs (fossil energy), suitability for mechanization, food habits and possibilities for marketing influence the choice between a cereal crop and a root or tuber crop.

Ecology

Cereals are grown in all regions of South-East Asia, under conditions varying greatly in climate and soil. These ecological conditions have been studied extensively, mainly for the three leading crops rice, maize and wheat. Temperature, incident radiation and daylength are important site-specific factors determining the potential yield, i.e. the yield realized when growth-limiting factors (water, nutrients) are optimal and growth-reducing factors (diseases, pests) are absent. The potential yield is reduced to the attainable yield under the prevailing suboptimal supply of growth-limiting resources (water, nutrients), but in the absence of growth-reducing factors. The attainable yield is reduced to the actual yield by biotic (diseases, pests) and abiotic (soil pollutants, extreme weather) growth-reducing factors.

Climatic factors

Temperature

Seasonal temperatures tip the scale in demarcating which cereal can be grown in a certain area or season, and are the main constraint to the further expansion of cereals to lower or higher latitudes and altitudes. Air and soil temperature, and for rice the temperature of the irrigation water, affect the total cropping period as well as all the growth processes from germination to grain filling. High temperatures generally promote flowering and low temperatures delay flowering. Critical and optimal temperature values for all main cereals have been thoroughly studied, especially for rice, wheat and maize. These values vary with the cultivar and the plant development stage. The optimum day temperature range for maize is 21-30 °C, for rice it is 20-35 °C. These values are much lower for cereals grown in the temperate areas: 10-24 °C for wheat, 10-25 °C for rye. It is well known that in temperate areas the highest cereal yields are obtained when ripening coincides with declining temperatures. This effect is attributed to the extension of the grain-filling period. Also, at least for wheat, cool temperatures are favourable for the total quantity of carbohydrates transported to the grain and - vice versa - extreme high temperatures will shorten the grain-filling period and thus diminish the mean kernel weight. The harvest index is negatively influenced by higher temperatures. Low soil temperatures retard germination and emergence, giving pathogens more chance to cause seedling death.

The symptoms of cold injury are similar for all cereals: loss of vigour, slow and stunted growth, leaf discolouration, reduced tillering, leaf senescence, irregular heading, increased susceptibility to diseases, incomplete panicles, sterility, degeneration of spikelets, late maturation, low yield, shrivelled seed and inferior grain quality. Cold injury is a common cause of low yields. Sterility is mainly due to cold injury during booting and anthesis. In rice it has frequently been observed at higher elevations in South-East Asian countries. Local landraces in Java, however, seem to be adapted to highland conditions.

High temperatures during anthesis also cause injury resulting in sterility of the spikelet. The range of minimum-optimum-maximum temperatures during meiosis is 11-33-41 °C for rice and 9-33-42 °C for maize (Goldsworthy & Fisher, 1984).

A common index to quantify the influence of temperature on cereals is 'growing degree-days' (GDD), i.e. the cumulated daily mean temperatures above a certain zero-growth base temperature. This base temperature is low for cereals with a C3-cycle photosynthetic pathway (e.g. 2.6 °C for wheat) and high for C4-cycle cereals (e.g. 9.8 °C for maize, 11.8 °C for pearl millet) (Jones, 1985).

Radiation

Although the light is more intense in the tropical zone, the daily useful photosynthetic period is longer during summer months in temperate areas and in the subtropics. Hence potential and attainable yields during summer in temperate and subtropical areas are 10-25% higher than in the tropics. Large regional and seasonal fluctuations exist due to cloudiness. This explains why yields of irrigated crops during the dry season tend to be 10-20% higher than yields during the wet season.

In the early stages of leaf area expansion, both leaf area index (LAI) and crop dry matter increase exponentially. Once a young crop reaches a leaf area index of about 5 and maximum light interception is attained, crop photosynthesis does not increase any further, hence the growth rate is constant and linear. In this phase the biomass of a healthy and vigorous cereal crop will increase by 150 to 200 kg/ha/day (Lövenstein et al., 1992).

The maximum energy utilization of a full-grown canopy of wheat and rice, both with a C3-cycle photosynthetic pathway, is estimated to be 6% (Hay & Walker, 1989). The photosynthesis of maize, sorghum, and millets, and of the pseudo-cereal amaranth follows a C4-cycle pathway, characterized by a higher light saturation point and by lower respiration losses than C3-cycle plants. Maximum utilization of energy measured on C4-cycle crops reaches values up to 8.9%. In contrast to C4-cycle plants, C3-cycle plants (rice, wheat, barley, quinoa, buckwheat) suffer from high carbohydrate losses due to high photorespiration at high temperatures. In general, C4-cycle plants are more adapted to hot climates with high light intensities than C3-cycle plants (Jones, 1985).

The utilization of photosynthetic active radiation (PAR) during the crop cycle is subject to many limiting factors such as below-maximum leaf area, extreme low or high temperatures, water deficit, poor soils and poor crop management. This may lead to values between 2-4%, while the average daily radiation in itself is limited due to cloudy weather and/or short daylength. A value of LAI 4-5 is needed for maximum interception of PAR, but this varies greatly between species and cultivars. LAI often increases to values well over 4 during crop growth but then declines below 3 due to leaf senescence and diseases.

Daylength

Photoperiodic daylength is the interval between civil twilight before sunrise and civil twilight after sunset. At the equator, daylength hardly varies during the year. At 10 °N (southern parts of the Philippines, Thailand and Vietnam) it varies from about 11.30-12.40 hours and at 20 °N (northern parts of the Philippines, Thailand and Vietnam) from about 10.50-13.20 hours. All cereal crops exhibit a more or less pronounced photoperiodic response, i.e. the daylength influences the time span of the vegetative period from germination to anthesis. Local cereal cultivars show a photoperiodic response that fits the latitude and growing season. Near the equator only short-day cultivars are usually found. At higher latitudes the local cultivars cultivated during the summer require longer days, while those used for winter cultivation may be either long or short-day. Traditional rice cultivars in South-East Asia are short-day crops; long days delay flowering and prolong the total cropping period. In some cases a long-day reaction may be an advantage, since late-maturing cultivars often outyield early-maturing ones. Many modern cereal cultivars for tropical and subtropical areas are almost daylength insensitive, which makes them in principle suitable for a wider geographical area. Another advantage is that the farmer can plant them at any moment of the year when temperature and water availability are suitable, without much variation in crop duration.

Under natural conditions the photoperiodic sensitivity serves as a mechanism to achieve crop growth during a period of peak water availability, with grain maturation during a dry period. In the northern part of South-East Asia, long-day maize cultivars are sown at long intervals at the beginning of the rainy season. All these fields will flower and mature in a short time span at the end of the rainy season.

Water

Under the climatic conditions of South-East Asia, the potential evapo-transpiration (ETP) is in the range of 4-8 mm/day (4-8 l/m²/day). Important factors determining ETP are soil structure, texture and moisture content, soil cover, crop type and stage, and canopy develop¬ment. The total water requirement mainly depends on environmental conditions, crop type and growth stage. Because of the difference in photosynthetically active leaf area, a full-grown cereal crop needs more water than during the juvenile stage or at maturation. Soil moisture may be supplied exclusively by rain, by irrigation alone or by rain water supplemented with irrigation. Cereals can cope with a range of soil moisture conditions. Their roots penetrating to a depth of 1-2 m are capable of extracting water from saturated soils to soils near wilting point. Pearl millet is especially noted for its dense root system and high root suction potential. Although moisture stress at any time may reduce yields, cereals do have certain critical periods in which they are extremely sensitive to drought. Early moisture stress hampers germination and tillering. Maize is very sensitive during anthesis and in the period of grain filling up to the milky stage. Rice is most sensitive during anthesis and the milk development stage; sorghum is very sensitive during ear emergence and during flowering and seed formation (Doorenbos & Pruitt, 1984). Rainfed cereals often fail to establish after a small shower which is sufficient to initiate germination but not enough to sustain the seedlings.

The water use of a crop is also strongly associated with its photosynthetic pattern. The water-use efficiency (water used per g dry matter produced) is around 500 g for barley and 700 g for rice (both C3-cycle cereals) but is only 280 g for pearl millet and 350 g for maize and sorghum (all C4-cycle cereals). The lower water use of C4-cycle plants makes them more suitable for drier areas. The ability to withstand drought differs among cultivars. Drought resistance is based on plant characteristics such as earliness and the development of a deep rooting system. Drought-resistant cultivars are extremely important for regions with unreliable rainfall and lack of irrigation. On the other hand, soil drainage should be adequate. All cereals except rice are rather sensitive to waterlogging. Flooding for some days or even some hours may cause wilting, root rot or other irreversible damage.

Ideally, a dry period is needed to facilitate the harvest of a cereal crop. Yet rice and maize have become important food crops even in the year-round per-humid equatorial zone of Sumatra and Kalimantan.

Soil factors

Cereal crops are traditionally grown on a wide range of soils varying from heavy clay to light sandy loams, as long as the nutrient and water requirements of the crops can be met. Cereals differ in their adaptation to specific soil types. Maize and wheat are rather sensitive to salinity, whereas rice and barley are more tolerant. Millets are grown on less fertile soils with low precipitation because of their ability to give a reasonable yield when other cereals fail. Rice is a marshy plant for which the physical soil conditions are not so important.

Cereals thrive at pH values from 5.5-7.0. Many soils in South-East Asia are very acid, with a pH-H₂O below 5.0, inducing toxicity of Al, Fe and Mn. These soils are often low in available P or liable to P fixation, and low in K, Mg and S. Their nutrient storage capacity is low, necessitating fertilizer application to be split between several dressings. Liming is recommended to raise the pH to at least 5.5.

The soil fertility of land submerged for rice production is strongly influenced by the anaerobic conditions. Major chemical changes take place which have a large influence on soil nutrient transformations and availability. The pH tends to be neutral, the supply of N, P, Si and Mo is improved, but Zn and Cu availability decreases, whereas reduction products like methane and hydrogen sulphide may reach toxic levels.

Erosion is a main cause of declining cereal yields. Cereal fields in the heavy rainfall areas of South-East Asia are very erosion-prone.

Agronomy

Production systems

From shifting cultivation to permanent cropping systems

Shifting cultivation is still practised in some provinces of Indonesia (Kalimantan, Irian Jaya), Vietnam and the Philippines, mainly with upland rice and maize. Cereals and pseudo-cereals of minor economic importance but vital for subsistence, such as sorghum, millet, Job's tears and buckwheat are also produced in this traditional system. Although the area under shifting cultivation still amounts to several million ha, it only accounts for a minor percentage of the total cereal production.

Permanent cropping instead of shifting cultivation has been practised for thousands of years. The planting of irrigated rice on terraces was a milestone in agricultural development. Better tools, the introduction of animal traction, the use of farmyard manure or chemical fertilizers, improved seed, disease and pest control measures, timing of planting, improved grain storage, crop rotation, all elements of crop husbandry in optimal combination help make the system more productive. Increased demand for food from the steadily increasing population drives farmers to intensify their use of external inputs. However, the increased use of agro-chemicals, in South-East Asia mainly insecticides on rice and N fertilizers on rice and maize, is jeopardizing the sustainability of the system and bringing about soil and water pollution.

Rice-based cropping systems

Rice-based cropping systems are often classified according to the water supply of the rice crop as:

• lowland rice grown on irrigated land,
• lowland rice grown on flooded land (rainfed),
• upland rice grown as a rainfed crop,
• deep-water rice (floating rice) grown in areas of deep flooding (up to 5 m or more).

In literature the terms "lowland rice" versus "upland rice" are commonly used to indicate irrigated or flooded versus rainfed, non-irrigated and non-flooded cultivation. Some authors use the terms "wetland rice" and "dryland rice".

The most common cropping system in South-East Asia is the cultivation of irrigated rice during the rainy season, followed by secondary crops, e.g. maize, soya bean, groundnut, tobacco, mung bean and vegetables, which are planted at the end of the rainy season and the beginning of the dry season. Double cropping - in some areas even triple cropping - of lowland rice is practised when the supply of irrigation water is sufficient. In areas with rainfed land only, upland rice or maize is grown during the wet season, followed by a secondary crop if rainfall is still adequate. In Indonesia, in areas where water cannot be controlled properly, a system called "surjan" is promoted. It entails planting lowland rice on sunken beds 3-4 m large. These are alternated with raised beds used for growing vegetables, maize or other upland crops. In the northern parts of Thailand, the Philippines and Vietnam, cereals such as wheat, maize and millets are grown during the cool, dry season, whereas during the hot rainy season the land is used for lowland rice. The rice-wheat cropping system is found in a very large area of the zone between 25N and 40N.

Planting

Information of importance for the cultivation of the main cereals and pseudo-cereals is summarized in Table 6.

Planting material

Traditionally, the grower saves a portion of the grain at harvest as seed for the next season. During cultivation little if any extra care will be given to the part of the crop destined for seed, but some farmers select particular plants or parts of the crop for seed. If the own seed is insufficient, the farmer has to procure seed from other farmers or from the local market. Seed of improved cultivars for replacement of landraces is often available from the public or private seed industry or from farmers who had adopted a new cultivar earlier. Numerous on-farm comparison trials with farm-saved seed and commercial seed of local and improved cultivars have shown that a good genetic, physiological, physical and sanitary quality of the seed is the most important means to assure sustainable high yields (see section 1.9). Most farmers are well aware of the advantages of good seed, but will only purchase certified seed on a regular basis when there is a marked advantage over the use of farm-saved seed.

Tillage and sowing

The surface of a seed-bed for cereals should be regular, to enable sowing at a uniform depth. Normal tillage for seed-bed preparation starts with hoeing or ploughing, to be followed, if needed, by harrowing. Ridges may be needed for irrigation, or during the rainy season to improve drainage; they are made after ploughing or during hoeing.

Tillage for irrigated rice is different: after ploughing and harrowing, the land is soaked and puddled. Cereal crops are direct-sown, with the exception of irrigated rice, which is usually sown in a nursery and transplanted later. The most common method in temperate regions is sowing in rows by machine drilling or, for maize, precision sowing at a uniform depth of 2-4 cm. Machine drilling uses less seed and promotes a uniform stand. In South-East Asia this method is rather uncommon because cereal farms are small, sowing machines expensive, and the soil surface irregular. The normal practice is to dibble the seed in pockets. Broadcasting followed by harrowing is practised for small grains (wheat, barley, rye, millets), for the nursery seed-bed of lowland rice, or for direct sowing of lowland rice. Dry-seeded lowland rice known as "gogo rancah" occurs in some areas of Java.

The optimal sowing density in terms of number of kernels per m2 or kg seed per ha depends on the species and the cultivar (weak or strong tillering) and on expected field emergence and crop establishment. Shallow sowing, high N fertilizer dose and a low sowing rate stimulate tillering. Too dense sowing not only wastes seed but also results in excess tillers dying, which weakens the plants. High plant densities cause stem etiolation and increased susceptibility to lodging. Although farmers tend to use higher seed rates than strictly needed, very dense sowing or planting hardly occur in practice. Slow and uneven germination, often observed on dry, hard soil, or caused by inferior seed quality, results in uneven stands with abundant weed growth and low yield.

The data on sowing rate and plant density presented in Table 6 are average values. The percentage of kernels, sown out per m2, that results in harvestable plants, is highest for species with large seeds. It may reach 90% in maize and 70% in wheat and barley. Small-seeded crops show a much lower percentage of plant establishment. In areas with limited rainfall, the plant density practised for crops such as maize, sorghum and millets is strongly reduced. Thinning-out after field emergence is sometimes practised. Pseudo-cereals show self-thinning when sown very densely.

Intercropping

Subsistence farmers in tropical lowlands lacking irrigation facilities for rice cultivation sometimes intercrop their rainfed cereals (maize, upland rice, sorghum) with crops such as cassava, sweet potato, groundnut and other pulses, and vegetables. Stalks of growing maize and sorghum are used as supports for climbing beans. Intercropping with pulses may add some nitrogen to the soil, to the benefit of the cereal. But the main advantages of intercropping are a better use of available space, reduction of risk for complete crop failure, better soil protection against erosion, reduction of diseases, pests and noxious weeds. A disadvantage of intercropping is that mechanization of sowing, weeding and harvesting operations becomes difficult. In lowland rice, the borders and bunds are often used for planting vegetables and root and tuber crops.

Unlike several other crops (taro, groundnut, pineapple, Capsicum pepper), which tolerate some shade, cereals produce best under unshaded conditions, hence they are less suited for agroforestry systems. Research reports on intercropping of tree crops with cereals attribute the low yields to reduced radiation and, to a lesser degree, competition with the trees for nutrients and water. In trials with cereals and other crops under coconut in India, pearl millet and kodo millet (Paspalum scrobiculatum L.) yielded remarkably well, almost as high as in the open, but the yield of maize (1.2 t/ha) cultivated under coconut was only 37% compared to the yield in the open (Ohler, 1984).

Crop rotation

To obtain a maximum grain yield it is recommended to rotate cereals with non-cereal crops as a general control measure to reduce the build-up of diseases and pests. The important exception is irrigated rice, which is grown continuously on the same land, often without noticeable yield depression. However, very intensive cropping does stimulate the build-up of populations of rice diseases and pests such as leafhoppers transmitting the tungro virus, and brown planthoppers transmitting grassy stunt and ragged stunt virus. Monoculture also favours the build-up of populations of certain weed species such as Cyperus in irrigated rice. Farmers sometimes rotate cereals with non-cereals in an attempt to control these weeds. However, most farmers practise crop rotation not because of the beneficial effect of other crops on the cereal crop, but because of the beneficial effect of the cereal on the succeeding dicotyledonous crops such as tobacco, cotton, vegetables, pulses, roots and tubers. This positive effect may be attributed mainly to the reduction of plant parasitic nematode populations and reduced weed competition.

Husbandry

Weeding

In young cereal crops, some weeds of Gramineae are difficult to distinguish from the cereal plants. In South-East Asia manual weeding by pulling out or hoeing is most common, though mechanical weeding using animal traction or small tractors is becoming more popular. On well-prepared and puddled land, irrigated transplanted rice is very competitive because when planted it has a head start over the weeds. If water control is complete in lowland rice crops, most weeds can be eradicated by alternating between draining the fields for a few days followed by flooding to the top of the young rice plants.

Weed control with herbicides is gaining popularity. Of the pre-emergence or pre-planting herbicides, roundup (glyphosate) is often used against grasses and sedges (e.g. nutgrass) and paraquat against broad-leaved weeds. Other popular pre-emergence herbicides are atrazine, butachlor and propachlor. Popular post-emergence herbicides used in maize, sorghum, millets and wheat include dinoseb and 2,4-D. Weeds are no longer harmful once cereal crops have attained a full-grown leaf canopy. The application of herbicides, when carried out incorrectly or with the wrong products, may be hazardous for the environment. There are alternatives to chemical weed control, and farmers need to be informed about them.

Mechanization

Animal traction with oxen and buffaloes is commonly used by farmers in South-East Asia for soil tillage, but 2-wheel tractors are gradually becoming popular. Small motor-powered threshing machines are also used widely. Yet cereal cultivation is still very labour-intensive. This can be illustrated by comparing the labour requirement for rice in South-East Asia (about 1200 hours/ha, yield 3.6 t/ha, output 3 kg/hour), with labour for mechanized rice farming in the United States (20 hours/ha, yield 6 t/ha, output 300 kg/hour). A century ago wheat growers in western Europe needed 200 hours/ha using horse power for tillage and harvest, with a yield of 2 t/ha, hence an output of 10 kg/hour. Nowadays, large-scale farmers in western countries obtain quadruple yields (8 t/ha) by using herbicides for weed control and tractor-driven machinery for ploughing, sowing, weed control, spraying, harvesting and threshing, reducing labour requirements to 15 hours/ha, hence an output of about 500 kg/hour. The greatest progress in the reduction of manual labour was achieved by the introduction of herbicides for weed control and by the invention of the combine, a machine which harvests and threshes the grain in one single operation (Leonard & Martin, 1963). The downside of this is the exorbitant use of fossil energy and of herbicides. As costs of labour rise, the use of motor-driven machinery in South-East Asia will increase, but it is unlikely that the millions of small cereal farmers will be able to attain the same degree of mechanization as big farmers on the great plains in Europe, America and Australia. The development of simple, relatively cheap and well-designed machinery (2-wheel tractors, spraying equipment, threshing machines) has brought a certain degree of mechanization within reach of the small farmer.

Irrigation

Cereals other than lowland rice are usually grown as rainfed crops, but in some regions, maize, sorghum and wheat are cultivated under irrigation. Various factors determine the total quantity of irrigation water and the frequency of application required for undisturbed crop growth and a high yield: evapotranspiration, rainfall, topography, soil type, seepage and percolation losses, crop duration, cultural practices. Dry hard soil must be made tillable by one or more pre-irrigations. In a trial to measure the influence of water manage¬ment practices on rice, the water-use efficiency (grain weight in g per litre of water) ranged from 0.6 (deep continuous flooding with 15 cm water, yield 8.9 t/ha) to 1.4 (continuous soil saturation with 1 cm water, yield 9.0 t/ha) (De Datta, 1981).

Furrow irrigation is the most common method used for cereals other than lowland rice. Irrigating implies a certain risk of waterlogging and causing soil salinity by salts accumulating near the surface. These salts are either brought in with the irrigation water, or by capillary rise from a high water table. In the first case, the remedy is to ensure that the salts are leached beyond the root zone, either by natural rainfall or by irrigation. In the second case, the water table must be lowered, by avoiding excess irrigation or by installing a drainage system.

Plant nutrition

Nutrient removal The uptake of minerals from the soil solution by cereals (rice, maize, wheat) has been extensively investigated. A good insight into nutrient uptake and distribution can be obtained by analysing data on nutrient removal (Table 7).

The data illustrate that the grain takes up much N and P, while the straw is richer in K, Mg and Ca and moderately rich in N. If the straw is returned and incorporated into the soil, the removal of K, Mg and Ca is greatly reduced. Straw and other crop residues remaining in the field can be ploughed in. Commonly, for practical reasons, the straw which is not taken away for other uses will be burned on heaps in the field. Although the remaining ash enriches the soil with minerals, the other advantages of organic material are lost. Moreover, burning means considerable losses of nitrogen, and the remaining soluble minerals are easily leached. However, one advantage of burning the straw is that it may destroy certain diseases and pests which could contaminate a following crop.

Nitrogen

Nitrogen is the most important element for cereal cultivation. It is an essential element of proteins in all plant tissues, including the kernel. An optimal N supply increases the leaf area, stimulates growth, delays leaf senescence, improves tillering and tiller survival, increases the number of grains per ear and raises the protein content of the grain. Too high doses of N stimulate excessive vegetative growth, especially of leaf tissues with low dry matter content. Luxury consumption of N makes the plant susceptible to diseases and pests. Also, plants may form longer and weaker stems that are susceptible to lodging. By experience, farmers learn to apply doses of N fertilizer to obtain high grain yields without too much risk of crop loss by lodging. In South-East Asian countries the use of fertilizers, especially of urea in lowland rice, is often extremely high, far beyond the recommended doses. However, as Integrated Pest Management (IPM) becomes more widespread, the application of N will become more balanced. The N efficiency in upland cereal cultivation is 50-60%, but in irrigated rice it is generally only 25-30% as a consequence of ammonia volatilization, denitrification in the anaerobic zone, leaching losses and runoff. Split application with one or more top dressings is recommended in order to guarantee a more regular supply of this very soluble element and to reduce losses. In addition to correctly timing the N fertilizer gift, losses are minimized by incorporating the fertilizer in the soil, by using ammonium (NH4) instead of nitrate (NO3) fertilizer, and possibly by applying slow-release fertilizer and nitrification inhibitors. In upland cereal cultivation, 50-60% of the N is often applied as basal dressing and the rest in one or two split applications.

Fertilizer recommendations

With the exception of farmers practising shifting cultivation, most subsistence farmers and practically all cash crop cereal farmers of South-East Asia apply fertilizer. Farmyard manure is an excellent fertilizer to improve both the chemical fertility and the physical properties of the soil. However, supplies of farmyard manure or other organic fertilizers are far from sufficient, and therefore cereal farmers depend on inorganic fertilizer for a reasonable yield.

It is difficult to give general recommendations, because appropriate fertilizer recommendations are tailored to cultivar, season, soil type, water supply, and the fertilizing of preceding crops. In South-East Asia, recommendations for rice range as follows: 75-140 kg/ha for N, 10-20 kg/ha for P, 0-35 kg/ha for K (Halliday & Trenkel, 1992). Official recommendations for hybrid maize in South-East Asia are 90-180 kg N, 20-25 kg P and 0-50 kg K, and for local maize cultivars are 30-90 kg N, 15-20 kg P and 0-25 kg K. Many farmers do not follow official recommendations but rely on own experience. Most tropical soils are very acid (pH < 5.5). The acidity is aggravated by heavy applications of acid-forming N fertilizer, urea being the worst. Liming is still not very common. The lime or calcium carbonate equivalent needed to raise the pH by 0.5 is about 1.5 t/ha on light, sandy soils, about 2.5 t/ha on loam and 4.0 t/ha on heavy clay or organic soil. The best lime source is magnesium limestone (dolomite). Fertilizer application, which is a necessity in primary crop production, is practically environmentally harmless if carried out in accordance with official (and correct) recommendations. Any pollution which may occur beyond an acceptable low level is mainly the result of faulty fertilizer practices (Halliday & Trenkel, 1992).

Application methods

For most cereal crops the fertilizer is applied in granular form by manual broadcasting. Mechanical distribution is not common in South-East Asia. Plant or row application is generally more effective than broadcasting because the fertilizer is placed close to the stem base where the roots are most concentrated. Plant or row application is rather common for maize and sorghum, which are cereal crops with a large inter-row distance and a low planting density. The fertilizer is worked into the topsoil or left on the soil surface to penetrate slowly when dissolved in rainwater.

A common practice for lowland rice is to apply the total amount of P and K as basal dressing at planting (because these elements do not leach out easily), and N in two or three split applications, as basal dressing and in one or two top dressings, one during tillering (about 3 weeks after planting) and the next facultatively at inflorescence emergence.

Foliar application is seldom applied because it is expensive and the amount of macro-nutrients per dressing is limited. In the case of nutrient deficiencies it may be useful to spray with the required element, usually Mn, Zn or Cu.

Crop protection

Cereals are affected by many diseases and pests. The relative importance of cereal diseases and pests is indicated in Table 8.

Most diseases of cereals are caused by fungi. Some bacteria from the genera Pseudomonas and Xanthomonas cause losses, especially under high relative humidity. Several viral diseases are devastating in rice, maize, sorghum and other cereals. Contrary to other food crops, cereals are not very susceptible to nematodes.

Borers are among the most harmful insect pests of rice and maize. Hoppers damage the plants by sucking and they transmit viral diseases. They have become a major pest since the early 1960s as the result of the widespread application of synthetic insecticides for the control of stem borers (Brader, 1982; Heinrichs & Mochida, 1984). Birds, such as weaver birds, parakeets and sparrows, may cause considerable damage to all grain crops. Rats and other rodents may be very harmful to grain crops at all stages.

Crop losses

Yield losses of food grains before harvest are mostly below 5%, but may reach up to 30% (Thurston, 1984). Exceptionally the whole crop may be destroyed by a disease or pest. Losses cannot be completely avoided, even with adequate crop protection measures. Crop loss assessment is supremely important because it correlates crop damage with yield reduction, enabling farmers to improve crop husbandry, especially their planning and timing of crop protection measures, and to establish economic thresholds in relation to the costs of control equipment and materials. Yet crop losses are rarely assessed adequately (Jago, 1993). For example, in Malaysia rice stem borers were considered to be the most serious pest of rice. Chemical control as a preventive method was recommended. However, surveys showed only a low incidence of rice stem borers, with less than 5% of bored tillers. Apparently, rice stem borers were under natural control. Simulation models can be used to analyse the development of diseases and pests with the help of collected data. EPIPRE is such a model. It uses linear regression to analyse disease or pest effects on crop yield. It is a field-oriented, computer-based, interactive system, which stores specific data from every registered field in a data bank. At intervals, the farmer observes disease and pest incidence in his field and communicates his observations to the EPIPRE database. Combining various data, EPIPRE calculates the probability of disease and pest development and of yield loss in this field. If the expected loss outweighs the expected costs, a recommendation to spray the field with a biocide is issued to the farmer. The strong involvement of the farmers in this Integrated Pest Management (IPM) project taught them to distinguish symptoms of all kinds of diseases and pests and made them conscious of environmental effects of chemical control measures (Zadoks, 1980).

Control methods

Although a large part of the global cereal area is cultivated without any crop protection, the number of farmers practising certain control measures is increasing. The following control methods may be distinguished: control by cultural practices, control by resistant cultivars, biological control, chemical control, and integrated pest management.

Cultural practices

With appropriate cultural practices, damage by diseases and pests can usually be kept at a low level. One general control measure is crop rotation to reduce the build-up of harmful populations of diseases and pests. Other methods are based on removing potential sources of infestation. The use of clean seed is a very important preventive measure. Rice fields can be left inundated to control soilborne diseases. Infection by fungi is reduced by a lower plant density. The supply of nitrogen fertilizer should be minimized; the higher the N content of the plants, the more susceptible they are to many diseases and pests. Another control method is proper timing of the cropping period. For example, maize is only vulnerable to downy mildew in the first weeks after field emergence; sowing early when the infection level is low can avoid severe damage. Army worm, a pest of maize and sorghum, can be controlled by drawing the plough furrows across the paths of migrating larvae, with the steep side of the furrows facing the oncoming "army"; pits about 30 cm deep are sunk at intervals in the furrow; the larvae move along the furrows and gather in the pits where they can easily be destroyed (Häfliger, 1979).

Resistant cultivars

About twenty years ago, resistance as a control method against diseases and pests gained scientific attention through the work of the international agricultural research centres. The International Rice Research Institute (IRRI) developed several rice cultivars with resistance to leafhoppers and planthoppers. IR 26, a cultivar resistant to brown planthopper, was successfully cultivated over extensive areas in the Philippines, Vietnam and Indonesia. However, it was only effective for 2-3 years, because a new virulent biotype of the insect became abundant (van Vreden & Ahmadzabidi, 1986). Cultivar IR 36, released in 1976, was resistant to brown planthopper, green leafhopper, gall midge, stem borers, blast, bacterial blight, grassy stunt, and tungro; it became the most successful cultivar of any improved food crop (Litsinger, 1989). Resistant cultivars of other cereals have also been successfully bred and released (Hagan, 1991; Hunger, 1991; Peters & Starks, 1991; Wellso et al., 1991). The first resistant cultivars were monogenic by nature, but modern ones have multiple genes for resistance. Combining two or more major genes in a single cultivar results in better and more sustainable (vertical) resistance. Horizontal resistance, a type of durable resistance governed by many minor genes, is becoming important in breeding programmes. Resistance to stem borers in various rice cultivars is polygenic but of a low level. Many rice cultivars with resistance to blast have been bred. Breeding for resistance to bacterial and viral diseases has been successful for other cereals. However, it will never be possible to combine resistance to all diseases in one cultivar, and the occurrence of new races or biotypes in the field is unpredictable.

Biological control

Two types of biological control can be distinguished. Natural biological control is the control of pests by their naturally occurring enemies (predators, parasites or pathogens) in the agro-ecosystem. This occurs everywhere and is of prime importance. The other type is classical biological control, which is the introduction of natural enemies to an area where they originally did not occur. In rice, classical biological control has been success¬ful in only a few cases, using exotic parasites to control exotic pests, e.g. the control of Chilo suppressalis in Hawaii and of Marasmia exigua in Fiji. In these relatively small and well-isolated regions, the parasites could easily spread and maintain themselves, meeting little competition from other natural enemies.

The use of exotic parasites to control indigenous pests has never been successful. Therefore, most research on biological control is now focused on natural biological control. Most rice insect pests in South-East Asia are indigenous and have their own indigenous natural enemies, but little research has been carried out on methods to maximize their impact. To maximize the benefit of the action of predators and parasites, an inventory should be made and regularly updated, and outbreaks investigated to determine their causes (Ooi & Shepard, 1994). This may lead to the development of practical pest surveillance and forecasting techniques and to recommendations for chemical intervention that take into consideration the pest/predator ratio, instead of the pest population alone.

Chemical control

Many farmers treat their rice crop routinely 1-5 times with insecticides. The application of fungicides is rather limited. Many farmers and policy makers believe that insecticide is a necessary input for ensuring high yields. Strong promotion of insecticides by the government and by private companies in many areas has led to over-use, provoking outbreak of insect pests that normally do not cause problems. Usually, the farmers' practice of pesticide application is not very effective, because of inadequate application technology, less than optimal timing and choice of pesticides, and the habit of mixing different products. In mixed spraying, the concentration of each component is often lower than recommended. This results in pest resurgence and secondary pest outbreaks, and leads to much waste of pesticides and to environmental pollution.

Some fungi can be controlled with chemicals, seed treatment being most effective (with 3 g of a mixture of thiram and carbofuran per kg of seed, the quantity of pesticide per ha is very minor), but chemical control of bacteria is difficult. Some viruses can be controlled by eliminating their insect vectors with chemicals.

The use of pesticides has reached levels at which they pose great health risks to the farmer and cause considerable damage to the environment. The risk of residues for the consumers is present, although less with cereals than with vegetables. Another negative effect of pesticides is the destruction of predators, parasites and other natural enemies of the pest. The disturbance of the natural balance has led to a vicious circle of an ever-increasing use of pesticides without reaching the desired 100% control. Many pests have developed resistance to pesticides, thus forcing farmers to spray more frequently and at higher doses. For lowland rice it has been shown that indiscriminate pesticide use leads to larger pest-related yield losses than not applying pesticides at all (Rola & Pingali, 1993).

Chemical control should only be applied if the economic threshold for damage is surpassed, if no other control methods are available or effective and if adequate precautions are taken for safe use. The International Code of Conduct on the Distribution and Use of Pesticides (FAO, 1986) gives guidelines for the safe use of pesticides.

Integrated pest management

Integrated Pest Management (IPM) is identical to Integrated Pest Control (IPC), which was defined by the FAO as: "... a pest management system that, in the context of the associated environment and the population dynamics of the pest species, utilizes all suitable techniques and methods in as compatible a manner as possible and maintains the pest populations at levels below those causing economic injury." Utilization of natural biological control is the cornerstone of IPM (Brader, 1979).

The success of IPM training programmes for rice farmers in several countries proved that high yields can be obtained without chemical pesticides. It is based on the following principles:

• communities of indigenous natural enemies keep insect pest populations in check most of the time; this naturally occurring control is the cornerstone of pest management, and therefore the conservation of these natural enemies, by reducing insecticide use as much as possible, receives full attention in IPM farmer training;
• resistance is an important component of IPM; cultivars with resistance against several diseases and insect pests are widely cultivated;
• good crop management practices (water, fertilizer) result in a healthy crop, able to tolerate fairly high infestation levels without suffering economically important yield loss.

As a result of these three factors, diseases and pests in non-treated rice fields do not usually cause important yield losses. In Indonesia, "IPM farmer field schools" have been set up. The philosophy behind them is that by learning in theory and practice farmers will become IPM experts and trainers (van de Fliert, 1993).

Many of the farmers trained in IPM see chemical control as a last resort. Biocides should not be used before the economic damage threshold has been reached, and then should be applied at the lowest effective dose, at the moment when diseases or pests are most vulnerable, using formulations that are the least toxic to beneficial organisms. Spot treatment, if adequate, is preferable to blanket spraying.

Harvesting and post-harvest handling

Harvesting and threshing

In the maturation stage of the kernel, starch turns from being watery, milky and doughy to being hard and white. Once the hard-dough stage has been reached the total dry matter stored in the kernel does not increase any more, and the moisture content has fallen to about 35%. The whole plant starts yellowing and the moisture content of the grain decreases further to 13-15%. From this moment, the grain can be harvested, but the drier the grain at the harvest stage, the better its storage quality. Premature harvesting reduces yield and quality because the kernels are light and shrivelled, whereas delayed harvesting causes losses due to lodging, sprouting, shattering and damage by rain, rats and birds. Cereal stems collapse and break soon after the right harvest stage has passed.

The traditional way of harvesting is either to cut off the panicles or spikes with a knife, leaving the rest of the plant (the straw) behind, or, more commonly, to cut the whole plant with a scythe or sickle. Maize is harvested manually by picking the cobs. Mechanical harvesting and threshing with combine machines, the common method in large-scale cereal production areas in temperate regions, is rare in South-East Asia.

Threshing to separate the grain from the husk is carried out by beating, treading, or mechanically, using small threshing machines. Maize is threshed manually or by a maize-sheller. In rice, the husks are removed traditionally by pounding the grain to break off the husk, or mechanically in hullers, producing brown rice in which the bran remains attached to the grain.

Drying and cleaning

Harvested grain with a moisture content above 14% should be dried as soon as possible to avoid deterioration caused by fungi and insects. Dry seed is hygroscopic, in equilibrium with the relative humidity of the surroun¬ding atmosphere. Often, harvested cereal is left in the field for a few days to dry before threshing, or shelling in the case of maize. Small farmers in South-East Asia usually dry their grain in the sun. Spreading out the grain for 3-5 days on a drying floor in a layer of 5-10 cm is sufficient to reduce the moisture content to 11-13%. Grain is often artificially dried with heated air before storage. It is mechanically loaded and circulated in a batch dryer or a bin dryer. Multipass drying in a continuous flow of hot air, for rice at 38-54 °C, lowers the moisture content to about 14% (Grist, 1986).

The grain needs to be cleaned to remove all particles which might affect the milling quality. Small farmers mostly clean the grain by hand winnowing, taking advantage of the wind. In a hand or machine-operated winnower all the light material such as chaff, straw and dust is blown away. To a certain extent, cleaning prevents storage losses. Dirty grain is more vulnerable to fungal and insect infestation because dirt contaminates the grain and attracts moisture.

Storage conditions

Stored grain must be protected against diseases and pests and any increase in air humidity. The lower the temperature and relative humidity, the longer the storage time without noticeable loss of consumption quality. For example, rice grain can be held in good condition for two years when stored at 18 °C and a moisture content of 13%, which is in equilibrium with a relative humidity of 65%. A high air humidity causes heat production by increased respiration and, consequently, the germination capacity will decrease. These conditions are also favourable for fungi, mainly Aspergillus and Penicillium species. They produce mycotoxins which frequently cause food and feed poisoning. Wheat with lower than 14% moisture, stored at a temperature below 20°C remains fungus-free and can be kept for many years, while with 17% moisture and at 15 °C the fungus-free period is at most 30 weeks and at 20 °C it is only 6 weeks.

Seed grain is usually treated differently to food grain. Seed must be properly selected, threshed, cleaned and dried, and preferably kept in airtight containers. Small farmers often keep their seed grain unthreshed under the roof of the house, or above the cooking place where it is well protected and dried by smoke and heat.

If the stored grain is not dry enough (water content > 14%), ventilation will be needed to reduce fungus development and fermentation. On the other hand, if the grain has been properly dried (water content < 13%), airtight storage will be the best method, because in that condition the grain respiration will cause the oxygen content to decrease, whereas the carbon dioxide content will increase, leading to an atmosphere preventing insect damage. In a humid climate, well-dried grain in an open store will absorb water from the air when ventilated, with increased risk of fermentation and infection with fungal diseases. In a dry climate, ventilation may further reduce the water content of the stored grain.

Post-harvest diseases and pests

Diseases and pests are the main cause of losses during storage. As infestation often starts in the field, timely harvesting is recommended. While it is being dried in the open the grain is exposed to rodents and birds, and to infestation by insects, fungi and bacteria.

Many insect pests of stored grains, such as the rice weevil (Sitophilus oryzae) and the lesser grain borer (Rhyzopertha dominica) require a high moisture level. Well-dried grain provides a poor environment for these pests. Although seed damaged by insects may still be capable of germinating, it will be very susceptible to fungal attack both during storage and after sowing. Many farmers do not use chemical insecticides to control insect pests because of the dual destination of the stored grain, for food and for seed. Insect control by natural compounds of plant origin, such as the leaves of neem (Azadirachta indica Juss.) or neem oil, is not widespread, although it is reasonably effective. Some farmers mix sand and ashes with the grain; the effect is based on scratching the insect's cuticle, causing it to die from dehydration. Sun drying is a cheap and effective form of insect control, since most insects will leave the grain when temperatures reach 40-44 °C (Clements, 1988).

Rodents cause damage to stored grain in two ways. The obvious effect is the eating of the grain. The less obvious, but probably more important damage is caused by biting holes in plastic storage bags, which results in increased seed moisture content and subsequently insect and fungal infestation. Rodents may be controlled by storing the grain in rodent-proof containers, by application of rodenticides or by traps. The micro-organisms of stored grain are essentially the same throughout the world, consisting chiefly of airborne species of Alternaria, Aspergillus, Cladosporium, Fusarium, Penicillium and Rhizopus. Fungi carried over from the field to the store include Epicoccus spp., Helminthosporium oryzae and Magnaporthe grisea (syn. Piricularia grisea) (Grist, 1986). Grain contaminated by fungi often contains dangerous levels of mycotoxins, such as aflavotoxin (from Aspergillus spp.), ochratoxin (from Aspergillus spp. and Penicillium spp.), ergotoxins (from Claviceps purpurea) and trichocenes (from Fusarium spp.) (Sharma & Salunkhe, 1991). Mycotoxins are very stable and persist in foods and feeds for a long time. Since decontamination is costly and impractical, prevention of contamination is very important.

Utilization and processing

Dry milling

Most cereals are subjected to some kind of processing of the whole grain before they can be used as food. The most common processing is dry milling, i.e. mechanical milling between rollers or grinding between two stones. Two steps can be distinguished in the milling process, i.e. reduction of the size of the kernels, followed by separation of the fractions.

The type of milling depends on the type of cereal and the purpose for which the flour will be used. Milling wheat separates the bran and embryo from the endosperm. The wheat grain contains about 82% of white starchy endosperm. It is not possible to separate this completely from the 18% of bran, aleurone and embryo, and to obtain a white flour with 82% extraction rate. White flour with about 70% extraction is often used for bread. Wholemeal is light brown and contains all the fractions of the milling of cleaned wheat, while intermediate flours vary in colour from white to light brown.

Rice grain is subjected to a different milling process than most of the other cereals. The bran layers are removed by polishing the grain in a rice mill. Abrading the pericarp, testa and some germ leaves a white grain. Rice can be milled into flour, but this is not common practice. In India, Bangladesh and Pakistan the parboiling of rice is an ancient tradition. Unhusked rice is soaked in water and then steamed or boiled. This process softens the husk but produces other changes, because the water-soluble components, especially the vitamins, migrate into the grain together with some of the oil. The grain structure is modified so that the aleurone layer is retained with the grain when the bran layers are milled off. Parboiling thus improves the nutritional value of the polished grain.

Maize is milled to produce maize meal, which may be sieved to remove the pericarp and germ. Sorghum and millet grains are traditionally pounded to produce flour.

Bread baking

The most important cereal product is bread made from wheat. Bread is made by baking a dough whose main ingredients are wheat flour, water, yeast and salt. The flour is wetted and the protein begins to hydrate and forms gluten, a cohesive protein which binds the flour particles together into a dough. Air bubbles are kneaded into the dough and eventually the dough becomes spongy, with the bubbles trapped in the gluten network. Enzymes in the yeast start to ferment the sugars present in the flour, which are broken down to alcohol and carbon dioxide. The carbon dioxide gas mixes with the air in the bubbles and causes the dough to expand. The extent of leavening depends on the amount of gluten and on the amount of carbon dioxide additives. Baking powder and eggs can be added instead of or as well as yeast. Rye contains only little gluten, but in combination with wheat flour it may give an acceptable bread dough. Wheat flour is often mixed with flour of other cereals (maize, rice, sorghum) or cassava starch for economic reasons, or with pseudo-cereals, e.g. of grain amaranth or buckwheat, to improve the nutritive value of the bread.

Pasta and whole-grain foods

Pasta is the collective term for macaroni, spaghetti, vermicelli and noodles. It can be prepared from flour of wheat, maize, sorghum and rice, and from mixtures, e.g. with buckwheat. Hard durum wheat (Triticum turgidum L.) is especially suited for this purpose. It is milled to produce coarsely ground endosperm particles known as semolina. Semolina is used in the manufacture of pasta and is also cooked unprocessed as couscous. To prepare pasta, dough is first made by mixing semolina and water. After kneading, the dough is heated to 49°C and extruded through a press to form a thin sheet which is cut into strips. A modification of this process, in which the dough is extruded through special dies to make shaped products, was first introduced about 50 years ago (Pomeranz, 1987). The pasta is cooked in boiling water until it is soft and is then ready to be eaten.

Breakfast cereals and porridge

Apart from bread baking, cereal starch can be made digestible and acceptable for humans by cooking. If the cereal is cooked with an excess of water and only moderately heated, as in boiling, the starch gelatinizes and is easily hydrolyzed by enzymes of the digestive system. Breakfast cereals are products that are consumed after cooking. They fall into two categories: those made by a process that does not include cooking and which therefore have to be cooked domestically (hot cereals) and those which are cooked during processing and which require no domestic cooking.

Many cereals and pseudo-cereals are used to make porridge. In Africa, maize grits are boiled with water. Barley meal is used for making a type of porridge in many countries in the Far East, the Middle East and North Africa. Buckwheat groats are used for porridge in Europe and North America.

Maize, wheat and rice are the cereals generally used for flaking. To make corn flakes, a blend of maize grits plus flavouring materials, e.g. sugar, malt syrup or salt, is pressure-cooked to a moisture content of about 28%. When the colour of the grits has changed from chalky-white to light golden brown, the grits have become soft and translucent, no raw starch remains, and the cooking is complete. The dried grits are then flaked on counter-rotating rollers, and the flakes thus formed are toasted (Greenfield & Southgate, 1992; Kent & Evers, 1994).

Malting, brewing and distilling

The essential process involved in brewing is the conversion of cereal starch into alcohol to make a palatable, intoxicating beverage. Fermentation is mediated by yeasts, usually belonging to Saccharomyces cerevisiae. Two processes are involved. First, the starch has to be converted to soluble sugars by amylolytic enzymes. If the grain's own enzymes are employed this is called malting. Second, the sugars have to ferment into alcohol by the enzymes of yeasts. All cereals are capable of undergoing malting, but barley is particularly suitable. Malting barley has a low content of husk, a high starch and a low protein content. In Africa, many malts are produced from sorghum and millets.

Malt is sprouted barley or another cereal crushed in water to hydrolyze starch by enzymatic action into soluble sugars. The solution extracted from the malt is called "wort". It forms the feedstock for fermentation for brewing beer or distillation of spirits (Briggs, 1978). The separation of the liquid wort from the solid remains of the malt is carried out by a process called lautering. The liquid is allowed to pass through the spent grains while retaining the fine solids, thus giving sweet wort. At this stage syrups may be added if the amount of fermentable sugars needs to be increased. Hops are also added at this stage. Now the wort is boiled, producing the bitter taste from the hops. After filtering it is transferred to tanks where yeasts are added together with air. Fermentation takes place for 7-9 days. Carbon dioxide may be added back when the beer is bottled or casked. The "green" or young beer is run off from the aggregated yeast cells and cooled and aged before filtering and carbonating. Depending on the type of beer, sugar may be added to the casks to allow a second fermentation.

An essential difference between beer and saké (rice beer) is that for saké the natural enzymes present in the grain are expressly inactivated before the saccharifying phase of saké brewing. Enzymes are derived from the fungus Aspergillus oryzae. The alcohol content of beer depends on the tolerance of the yeast strain used and is usually 5-12%. However, the alcohol content of saké is up to 19%.

Distillation is used to produce potable spirits with an alcohol content above that of fermented drinks. Spirits produced from grains are of two major categories: whisky (or whiskey) and neutral spirits. The difference is that in whiskies care is taken to retain flavours and colour carefully introduced during production, while in neutral spirits the introduction of flavours and colour during production is avoided. The various types of whisky differ in their origin, their carbohydrate source and the manner of their production. Beer for distillation of neutral spirits is produced as economically as possible, flavour being undesirable. The cheapest available cereal can be used and it is more economical to use enzymes derived from micro-organisms than those from malt.

Wet milling: starch and gluten

Wet milling is a maceration process in which the dry flour is soaked in water or an alkaline solution to extract starch and gluten. The objective is to bring about a complete dissociation of the content of the endosperm cells, and the release of the starch granules from the protein network in which they are enclosed. The starch may further be separated in amylose and amylopectin. Although the grains of all cereals contain starch, those most widely processed by wet milling are wheat and maize.

All wet processes for the manufacture of starch and gluten from wheat comprise the steps of extracting the crude starch and crude protein, purifying, concentrating, and drying the two products. To obtain pure gluten it needs to be separated from the germ and bran by first dry milling with conventional methods. Consequently, white flour is used as starting material for wet processing to separate starch and gluten. Vital gluten is separated from wheat by processes which permit the retention of the characteristics of natural gluten, i.e. the ability to absorb water and form an extendable, elastic mass. Vital gluten is used to improve the texture and raise the protein content of bread and to fortify weak flours, composed of mixtures poor in gluten. It is also used as a binder and to raise the protein level in meat products, breakfast foods, pet foods, dietary foods and textured vegetable products (Kent & Evers, 1994).

Maize is wet-milled to obtain starch, oil, cattle feed and the products of hydrolyzed starch (i.e. liquid and solid glucose and syrup).

Animal feed and industrial uses

Apart from human food, animal feed is by far the largest use for cereals, both as whole grain and as milling by-products. The treatments applied to cereals by animal feed processors are both expensive and time-consuming, and obviously would not be undertaken unless such treatments offer considerable advantages over the feeding of untreated whole grain. Both cold and hot, dry and wet, mechanical and chemical methods of treatment may be used, with the objective of improving palatability, avoiding wastage, improving digestibility and nutritive value, and encouraging consumption, thus leading to more efficient use of feed and faster growth. The actual treatment used will depend on the kind of cereal involved and the proportion of that cereal in the feed. It also depends on the type of animal for which the feed is intended, particularly whether it is for ruminants or for monogastric animals, and probably also for what stage in the animal's life cycle (Kent & Evers, 1994).

Rice bran and other milling by-products have traditionally been used for livestock feed rather than for human use because the removal of bran from the grain mixes an enzyme with oil in the bran, which eventually gives rise to rancid odours and flavours. The growing interest in rice bran as a foodstuff is because of its benefit in cholesterol reduction. This health benefit is attributed to the oil component of the bran, which constitutes about 20%. The technology to process rice-bran oil is similar to that used for other vegetable oils. In the near future the use of edible rice-bran oil may gain popularity (Luh, 1991; United Nations, 1985; Young et al., 1991).

Ethanol is produced by the enzymic action of yeast on sugars, which are themselves produced by the hydrolysis of starch. It can be regarded as a modification of the brewing process, in which starch separated from the grains is the starting material, and pure (96%) ethanol is produced by distillation of the aqueous solution. Ethanol made from cereals can be used as a partial replacement of gasoline for fuel of internal combustion engines. The process is a useful way of dealing with surplus grain whenever it arises.

In the United States, maize starch is important as raw material for the sweetener industry. Enzyme technology is used to produce dextrose and high-fructose syrups, mainly for sweetening soft drinks (Inglett, 1984).

Maize cobs, hulls of oats and rice and the fibrous parts of other cereals are rich in pentosans, condensation products of pentose sugars. Pentosans are the starting material for the manufacture of furfural. The commercial process for manufacturing furfural involves boiling pentosan-containing material with strong acid and steam for several hours. The pentosans are dissociated from the cellulose, then hydrolyzed to pentose sugars, and finally the pentose sugars undergo cyclo-hydration to form furfural, a heterocyclic aldehyde (Kent & Evers, 1994; Pomeranz, 1987). The most important use for furfural is in the manufacture of nylon.

Wheat and maize starches are applied in paper coating and as adhesives in the manufacture of paper, boards and plywood. They can also be used in paints, in plastics and to produce crude latex in the manufacture of rubber. Rice hulls can be a source of high-grade silica. Rice hull ash is used as a constituent of cement, since it is more acid-resistant than Portland cement. Furthermore it is a silica source in glass and ceramics industries. Rice hulls heated to temperatures up to 700 °C yield amorphous silica which is suitable for making solar-grade silicon for solar cells (Kent & Evers, 1994).

Genetic resources and breeding

Genetic resources

Thanks to Vavilov the importance of the centres of origin and/or diversity of cultivated plants for plant breeding is now recognized. Central America is the centre of origin of maize, whereas rice originates in South-East Asia/China, sorghum in North-East Africa, proso millet in China, barnyard millet in India, finger millet in East Africa, pearl millet in West Africa, and wheat, barley and rye in the so-called Fertile Crescent of West Asia. The process of domestication and diffusion of these crops has taken thousands of years and has given rise to numerous landraces in all production areas. Modern plant breeders rely on a continuous renewal of genitors for desirable characteristics, which means that they need wild relatives of crops, landraces and improved cultivars from their own region as well as from other production areas. In turn, tremendous genetic erosion is provoked by the adoption of improved cultivars, especially in areas with high production potential.

The International Plant Genetic Resources Institute (IPGRI, Rome, Italy) has the mandate to promote and coordinate conservation of genetic variation. Loss of genetic variation is partly counteracted by two complementary approaches: ex situ conservation in genebanks, and in situ conservation on-farm or in the natural habitat.

Genebanks

Ex situ conservation is effected through genebanks, which store samples of seed or planting material under controlled conditions of temperature and humidity. The aim is to conserve as much as possible of the existing genetic diversity, ensuring its availability as resource material for plant breeding and research in the future. Table 9 reviews collections of cereal accessions in the genebanks of the International Agricultural Research Centres (IARCs), national genebanks of global importance, and genebanks in South-East Asia.

In situ conservation

In situ conservation means that valuable genetic diversity is maintained in its natural habitat, allowing adaptation and evolution to continue. This method is particularly appropriate, at least theoretically, for habitats that are under threat and for areas that are still farmed in a traditional way, where landraces are often enriched by gene exchange with wild relatives. The problem with in situ conservation as conceived at present is that unless crops are permanently guarded, security is low. Natural habitats may dry out or disappear as a result of the intensification of agriculture. As a result local materials may be replaced by foreign genetic material. Paying farmers cash to conserve their landraces in situ appears to be financially unsustainable.

It has been argued that crop conservation and development at the farmer or community level should be recognized as complementary to institutional efforts, and the formal crop improvement and conservation efforts should be integrated more at the local level (Altieri & Merrick, 1987; Brush, 1991; Hardon & de Boef, 1993; Vaughan & Chang, 1992). These efforts seek to support and enhance, rather than to replace the community management of plant genetic resources. Various national and international institutions are making efforts to integrate the conservation of biodiversity of traditional crops in rural development. Germplasm in the form of landraces, breeding lines or mixtures of modern cultivars is put directly at the disposal of farmers, to be used and developed according to their own needs and practices; this is called "participatory breeding" (Eyzaguirre & Iwanaga, 1996).

Breeding

In the past the overall main objective was to breed for yield and broad ecological adaptability; resistance to diseases and pests was considered as part of the yield-determining factors. At present, the main objectives are tending to shift to breeding for the maintenance of high yields by durable resistance and by introducing genes for resistance to new biotypes of diseases and for tolerance of adverse conditions. In the case of self-pollinating (self-fertilizing) crops such as rice, wheat, finger millet and barley, the breeding process leads to very homozygous and uniform pure-line cultivars. The selection of cross-pollinating cereals like maize, sorghum, pearl millet and rye, requires a different approach, the result being a "selection" with a narrow composition compared with the original population, but still rather heterogenous with a high degree of heterozygosity. Composite and synthetic cultivars have been developed for maize and sorghum. Composites are a deliberate mixture of selections, multiplied as open-pollinated cultivars. Synthetic cultivars are composed of a mixture of some well-defined populations or inbred lines, which continue to show the effects of heterosis in later generations.

Hybrid cultivars offer the advantage of easy combination of alleles for disease resistance or other desired characters. A hybrid cultivar is the F1 offspring of two more or less homozygous lines, showing heterosis for yield and a high uniformity as most attractive characteristics. In cross-pollinating crops, well-combining inbred lines are used as the parents. The classic example is hybrid maize, which has become very successful all over the world, but is not suitable for marginal growing conditions. Hybrid sorghum is also widely grown, and hybrid pearl millet is becoming popular. Hybrid wheat remains largely in the research phase; apparently its advantages do not outweigh the high cost of the seed. In contrast to seed production of self-pollinating crops, the cultivar maintenance and seed production of hybrids are difficult and expensive.

Hybrid rice is widely cultivated in China; about 18 million ha or 55% of the total area. The seed production is based on cytoplasmic male sterility. Parent lines are selected on strong heterosis effect. The seed yield may attain 2.5 t/ha. It is difficult to synchronize the flowering of the parental lines. The yield of these hybrids on the farm is about 5 t/ha, i.e. 15-20% higher than that of convential seed. IRRI and the National Agricultural Research Systems (NARSs) are jointly developing parental lines and cultivars suitable for South-East Asian countries. Photoperiod-sensitive and temperature-sensitive genic male sterility systems are being investigated, as well as chemical emasculators. So far, farmers in South-East Asia have been reluctant to adopt hybrid rice. This is because the hybrids may show similar characteristics to open-pollinated cultivars regarding growth and resistance to diseases and pests, and the grain quality may vary from poor to acceptable. A major disadvantage is that the seed is 4-10 times more expensive than open-pollinated seed, because of the complicated system of seed production and the low seed yield from the crosses of the parental lines. Furthermore, the farmers are obliged to buy F1 seed for each planting.

Landraces versus improved cultivars

The traditional local seed supply systems that have evolved over hundreds of generations of farmers, have led to a multitude of landraces. A landrace may be defined as a mixture of genotypes resulting from biotic and abiotic selection pressures in specific agro-ecological and socio-economic conditions (Louwaars & van Marrewijk, 1996). On the large-scale, highly mechanized farms in the world-leading cereal production areas of the temperate zone, these landraces have been replaced by genetically uniform cultivars that react favourably to high fertilizer applications, in particular to high nitrogen fertilization. In South-East Asia, the area under high-yielding cultivars/varieties (HYVs) - especially of rice and maize - is increasing, but landraces still remain widely cultivated, both for subsistence farming and as cash crops. Possibly 75% of the lowland rice area and 45% of the maize area are planted with HYVs, whereas almost all upland rice still consists of landraces.

Landraces usually compare favourably to HYVs in adaptation to specific agro-ecological conditions, production systems and people's taste. The high degree of heterogeneity of a landrace implies a lower yield potential but a more stable yield compared with the very uniform HYV. The individual farmer may find that his traditional local cultivars with their high yield stability under adverse conditions and low demand for external inputs represent better value than the higher costs and risks involved in maximizing yield and profit of HYVs.

In practice, the distinction between landraces and improved cultivars is not always clear, since selection in landraces may lead to improved local cultivars. Also, simultaneous growing of landraces and HYVs leads to introgression of new genes into the landraces. Plant breeders attempt to combine the superior characteristics of introduced HYVs with lines from landraces to create improved cultivars well adapted to local conditions.

The Green Revolution

In this century, modern wheat cultivars with a high yield potential have gradually become widely cultivated in industrialized countries in the temperate regions. These cultivars have a more favourable harvest index than traditional landraces. Their main feature is short straw, which implies a favourable response to high nitrogen fertilizer gifts without lodging problems. Another important feature of these new wheat cultivars is their superior resistance to fungal diseases. However, often the resistance is broken down by the rise of new fungal biotypes within a few years, and the cultivar must be replaced. Every year new cultivars with higher yield potential are released.

In the sixties, similar high-yielding cultivars for tropical and subtropical regions were very successfully created for rice (IRRI) and wheat (CIMMYT). In the period 1965-1975, the wide introduction of these cultivars combined with increased fertilizer application and improved cultural practices (irrigation, use of pesticides) caused an enormous boost in rice and wheat production, known as the Green Revolution. This Green Revolution enabled many developing countries to meet the steadily increasing demand for food resulting from population growth and urbanization. Next to their merits, HYVs in combination with their technology package proved to have some negative properties. Firstly, the HYVs of rice were either resistant or highly susceptible to diseases, and sometimes resistance broke down by mutation of the pathogen, leading to crop failure. These HYVs of rice were genetically very uniform, lacked multigenic resistance to diseases and pests, and lacked tolerance of adverse conditions. At present, rice breeding at IRRI and other institutes is putting much emphasis on multigenic, durable resistance. Secondly, HYVs were found to have some negative social effects. Resource-poor farmers who could not afford the high costs of extra inputs were confronted with decreasing prices and could not compete with richer farmers. In some cases, small farmers who risked planting HYVs had a crop failure and lost their investments, and in the worst case they became landless farm labourers or migrated to the cities (Lipton & Longhurst, 1985).

Seed supply

Seed sources

Seed-supply systems

Traditionally, at harvest the grower saves a portion of the grain to use as seed for the next season. If the own seed lot is insufficient, the farmer has to procure seed from other farmers or from the local market. In South-East Asian countries, the supply of genetically improved non-hybrid cereal seed is primarily an activity of the public sector. The low net profit from seed production of self-pollinating cereal crops is not appealing to private enterprises. The seed market is more profitable for cross-pollinating crops and for hybrid cultivars. The production and handling of cereal seed require high investments because of the bulky nature of transport, storage and processing. Yet the international trend for market liberalization is pushing governments to strive for increased involvement of the private sector, including in the cereal seed industry. In western countries, the private sector has been very active in spreading improved germplasm developed by public breeding programmes. The Asia and Pacific Seed Association (APSA, Bangkok, Thailand) stimulates cooperation between both sectors and among countries in the region.

Seed price

The price of cereal seed primarily depends on the multiplication factor, which may vary from very low (e.g. for wheat 8-30) to very high (e.g. for sorghum up to 400). The price of the seed must compensate the producer for all extra expenses: cost of breeder seed or foundation seed, extra labour for processing, losses by roguing and cleaning, cost of seed inspection, storage, and distribution.

For the farmer, seed costs represent only a small part of the total production costs, particularly for small-seeded cereals such as sorghum which have a low sowing rate. Farmers will be reluctant to buy seed if they are not convinced that the expected superior physical, physiological and sanitary quality will increase their profit. Cash-constrained small farmers often perceive little benefit in purchased seed; they may compensate for uncertain quality (germination) of their own farm-saved seed by using higher sowing rates (Cromwell et al., 1992).

Seed replacement rate

The seed replacement rate is the percentage of the total seed requirement (in amount or area) which is annually renewed by seed from the formal sector. It is an important parameter for evaluating the seed situation for a certain crop in a given area. For rice, for instance, a replacement rate of 20% is judged appropriate. In most countries in South-East Asia the rate hardly reaches 10%, Indonesia being an exception with 25%. Clearly, the seed replacement rate for crops with a high degree of cross-breeding (sorghum, maize) should be high, e.g. 30-50%, since these cultivars easily interbreed with landraces or other cultivars. For hybrids, commonly used for maize and sorghum, the seed replacement rate should be 100% because farm-saved seed shows so much genetic decline that the farmer should purchase new seed for each new planting.

Accurate data on areas planted with landraces or improved cultivars and the replacement rates are scarce. Also, a low seed replacement rate does not mean that farmers do not accept new cultivars. New cultivars also disseminate through unregistered local seed-supply systems. They often become mixed with landraces during seed handling or by genetic introgression.

Seed production planning

Commercial seed production is a concern for specialist enterprises. The multiplication factor and the replacement rate are important parameters for planning a seed production project. It is important to plan adequately for the amounts of breeder seed, basic seed (foundation seed), and market seed (certified seed) required.

While planning seed supplies, the government has to take into consideration the risk of crop failures due to natural disasters. In time of shortage, farmers' families may be forced to consume the seed lot. Some national seed policy authorities take measures to secure seed stocks for emergency situations.

Seed crop husbandry

Crop husbandry for cereal seed production is principally the same as that for grain for consumption. The differences are related to the objective of a seed crop: extra care is needed because it is not yield maximization but the supply of first quality seed that is important. For good genetic seed quality, contamination with alien pollen or with other seed lots must be avoided. Isolation distances depend on the nature of pollen transport and the degree of cross pollination. For the production of certified seed of self-pollinating cereals (rice, wheat, barley, finger millet) an isolation distance of 10 m between the fields is normally judged to be sufficient, whereas for maize, sorghum and pearl millet at least 200 m is required. Careful roguing (the removal of off-type plants during crop growth) is most important, especially during the first multiplication stages.

To ensure good physical quality of the seed, much attention should be given to weeding out species such as red rice or wild sorghum which are difficult to remove later by seed cleaning. The physiological quality of the seed is improved by proper cultural practices and by harvesting in a dry period. The amount of N fertilizer applied must be carefully controlled, because too much nitrogen results in weak plants, prone to fungal diseases; however, potassium is beneficial. The sanitary quality is upgraded by respecting the prescribed crop rotation period, stringently controlling diseases and pests, and removing and destroying diseased plants.

Seed conditioning

Seed conditioning starts with threshing or shelling. Seed drying and cleaning is an important part of post-harvest handling because it provides an opportunity to upgrade the physical, physiological and sanitary quality by mechanically removing impurities. Seed cleaning may be combined with grading by size. The most common machines used for cleaning are the air-screen cleaner, the indented cylinder and the gravity table. Fumigation with an insecticide is often applied to kill all insects and insect eggs. Cereal seed has to be stored at least until the next sowing season. Some seed lots must be stored for two seasons, as an insurance against harvest failure. Deterioration by diseases and pests or physiological loss of viability during storage must be avoided. The principle is to store seed in a clean, dry and cool place. A prerequisite for good storage of cereal seed is that it should be dried to a moisture content of 13% maximum, in equilibrium with a relative humidity of about 65%. It is better to dry to 10-12%, which means an equilibrium with 45-60% of relative humidity. Storage in airtight containers is ideal. This is relatively easy for small quantities, e.g. farm-saved seed or breeder seed, and is sometimes also done in modern seed storage.

Nucleus seed and breeder seed have to be stored for long periods, but since relatively small amounts are involved, more ideal storage conditions can be created. Harrington's rule of thumb is worth quoting here (Harrington, 1972): for each 5 °C decrease in temperature and/or for each 1% decrease in seed moisture content (about 10% decrease in relative humidity) seed life is doubled.

Cereal seed is often dressed with fungicides and insecticides to protect against diseases and pests. The insecticide lindane and the fungicides thiram and captan are widely used.

Seed legislation

Governmental control of seed quality of cereal crops is more advanced than for other crops because of the significance of cereals to the national economy and food supply, and because the government itself is the only large seed producer.

All South-East Asian countries have a Seed Act intended to regulate and encourage the release of cultivars, the production and dissemination of high-quality market seed and the control of seed quality. The Philippines and Thailand are among the South-East Asian countries which are member of the International Seed Testing Association (ISTA, Zurich, Switzerland). This organization standardizes testing procedures for germination capacity, purity and health. Yet all countries refer to ISTA rules to some extent when defining seed quality.

In most countries an independently operating national seed control system has been established, mandated to control that seed legislation is being implemented properly. The objective of seed legislation is to guarantee that farmers have a constant supply of superior seed material. A national cultivar release committee is responsible for the release of new cultivars, whereas a national seed certification agency controls the cultivar identity and purity of the seed lots for certification. Full seed certification means official control of cultivar identity and purity through field and store inspection and labelling, and post-control plots (seed samples sown out to check the cultivar identity and the purity). Physical, physiological and sanitary quality control are carried out through laboratory testing.

Full certification of all the seed produced by the formal sector, meaning that only certified seed lots may enter the market, is very expensive. The modern concept of seed quality control is based on the breeder being responsible for the continuous supply of true-to-type breeder seed, and on the seed producer being responsible for the quality of the basic, foundation and market seed he produces. Seed producers should perform their own inspection and all seed entering the market should be "truthfully labelled" (TL seed) with a label indicating cultivar name, lot number, date of production, date of testing, germination capacity, purity, moisture content. This enables the government's role to be restricted more to guidance than to control.

Pseudo-cereals

Four non-graminaceous grain crops have been included in this volume. The three most important ones are grain amaranth (Amaranthus L.), grain chenopod (Chenopodium L.) and buckwheat (Fagopyrum esculentum Moench). At present, these three pseudo-cereals are only sporadically found in South-East Asia, but in view of their success in East Asia it might be worthwhile introducing them in locations with suitable ecological conditions. They were domesticated in a similar way as the true cereals. In the past, these pseudo-cereals were important as staple food crops in large areas, where they currently survive only as secondary food crops for subsistence or cash. Other uses, e.g. as forage, are of minor importance. In most industrialized countries they have been ousted by higher-yielding true cereals. For some decades, however, there has been renewed interest in these crops because of their excellent nutritional value (high content of essential amino acids).

The botany of pseudo-cereals is quite different from the true cereals. The seedling forms a taproot and a main stem. The plant does not tiller but exhibits branching as a mechanism to compensate for inadequate plant density. A clear disadvantage compared with true cereals is the large spread in time of anthesis and seed maturity. The seed is a less strong sink of assimilates than in true cereals, and the harvest index is much lower (around 30%). The climatic conditions needed for the satisfactory performance of pseudo-cereals are optimal in warm temperate areas during the summer months. Buckwheat (with a C3-cycle photosynthetic pathway) is cultivated during spring or autumn in subtropical areas, whereas the grain chenopod quinoa (with a C3-cycle photosynthetic pathway) is much cultivated in subtropical areas and in the highland tropics. Grain amaranth (with a C4-cycle photosynthetic pathway) is cropped from warm temperate to subtropical areas and also in the highland tropics.

Pseudo-cereals are cultivated similarly to millets. Generally, manual weeding is practised. Chenopod and amaranth are very competitive to weeds. They respond well to nitrogen fertilizer, whereas buckwheat does not tolerate high doses of nitrogen. Pseudo-cereals are more susceptible to insect pests than cereals. Crop losses reported as a consequence of diseases and pests are relatively moderate. Ripening per plant is rather uneven, and seed shattering is a problem. Hence, the choice of harvest time is a compromise between a high percentage of unripe seed and large seed losses by shattering. After harvesting, field drying is more difficult than for true cereals because of uneven ripening and the great leaf mass. Well-cleaned and dried seed can be kept for a long time, as with true cereals.

The small-seeded pseudo-cereals are processed by milling, comparable to methods used for small-grain cereals. The whole grain is also toasted to produce a nut-like flavour.

Although research and breeding of pseudo-cereals have received much less attention than in the leading cereals, quite a lot of work has been done in countries where their economic importance is obvious. Substantial research has been done on buckwheat in Japan, Korea, China, India, Nepal and Russia, and much research has been done on grain amaranth in Latin America, India and Nepal. Quinoa research has mainly been executed in the Andean countries of South America. Renewed interest in pseudo-cereals in western countries like the United States has led to research and breeding activities with the emphasis on large-scale cultivation with combine harvesting for bulk production.

Organic farming practices for the production of health food are gaining ground.

Cereal research in South-East Asia

The research attention on cereals in South-East Asia is primarily focused on rice, maize and wheat. Most of the cereal research in the region is conducted by the International Agricultural Research Centres (IARCs) in collaboration with the National Agricultural Research Systems (NARSs). The rice cultivars produced by IRRI and the wheat cultivars produced by CIMMYT have been especially widely adopted by farmers. Also, but to a lesser extent, cultivars and breeding material of maize from CIMMYT, cultivars of millet and sorghum from ICRISAT, and barley cultivars from ICARDA have diffused to farmers' fields, as well as cultivars developed by the NARSs. The rise in average yield over the last 50 years is partially attributable to agricultural research, although it is difficult to quantify these effects.

International and national institutions

At the International Rice Research Institute (IRRI) in Los Baños, the Philippines, the following research issues are currently receiving much attention (IRRI, 1995):

• development of a new plant type of semi-dwarf indica rice with high yield potential (up to 15 t/ha) and multiple resistance to diseases and pests;
• hybrid rice development with effective seed production (male sterile and restorer lines with synchronized flowering);
• causes of long-term productivity decline in intensive cropping systems;
• optimization of technology for rice-based cropping systems;
• development of sustainable production technology for upland rice;
• development of environmentally sustainable and economically viable pest management technology;
• biotechnology, mainly to develop DNA-marker-assisted breeding and genetic transformation protocols.

The International Maize and Wheat Improvement Center (CIMMYT) in Mexico, with a branch in Bangkok, Thailand, reports the following research priorities (CIMMYT, 1994):

• improvement of wheat and maize yield potential and resistance to biotic and abiotic stress;
• development of drought-tolerant maize cultivars;
• improvement of the nitrogen recovery in wheat from 50% to 65%;
• gene mapping and marker-assisted selection in maize and wheat;
• optimization of technology for the rice-wheat cropping system (collaborative programme with IRRI).

The International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) based in India performs research on sorghum and pearl millet (ICRISAT, 1994, 1995; Ramakrishna, 1993). Research priorities are:

• development of improved genetic material, with special attention for drought tolerance;
• integrated management strategies for control of Striga, insects and abiotic stress in sorghum;
• control of downy mildew, Striga and insect pests in pearl millet;
• development of forage cultivars of sorghum with related technology.

The research on cereals done at the NARSs in all South-East Asian countries has mainly been applied research, directed to the development of adequate technology packages for extension officers and farmers. Soil fertility and crop protection have received particular attention. Many cultivars suitable for regional ecological conditions have been selected on the basis of breeding material from the IARCs. Now that the NARSs are becoming stronger in breeding and applied research, the IARCs are shifting to "upstream" research.

Cereal research in all South-East Asian countries has much in common. Everywhere rice is the priority crop, followed by maize. Little attention is given to sorghum, and even less to millets, other cereals of secondary importance and to pseudo-cereals. Only Burma (Myanmar), Vietnam and Thailand contribute some research on wheat.

Special topics

Crop modelling

Computerized simulation models are used in advanced cereal research to quantify the influence of ecological and biological yield-determining factors on crop yield. Crop modelling is a useful tool for assessing the effect of prevailing weather conditions (temperature, radiation, rainfall) on crop performance to optimize resource use efficiency (fertilizers, irrigation water, pesticides), to characterize the ecological zones of production areas (quantification of yield potential and risks), to analyse biotic stress caused by epidemic diseases and pests for the benefit of IPM programmes, and to evaluate plant ideotypes for specific environments (Elings & van Keulen, 1994; Goudriaan & van Laar, 1994; Lövenstein et al., 1992). EPIPRE is an example of a simple epidemic and yield loss model implemented in an IPM application (see 1.5.4). Another example of the application of dynamic simulation models in crop research is Systems Analysis and Simulation for Rice Production (SARP), a project coordinated by IRRI with NARSs in Indonesia, Malaysia, the Philippines, Thailand and other Asian countries. This project was initiated by Wageningen Agricultural University and the Research Institute for Agrobiology and Soil Fertility (AB-DLO) in the Netherlands (Aggarwal et al., 1995; ten Berge et al., 1994). The project covers research on potential production (crop responses to light and temperature, crop development and morphogenesis), crop and soil management (water and nitrogen management for different soils and cultivars), biotic stress, and crop protection (damage mechanisms of diseases and pests), and studies the optimization of rice agro-ecosys¬tems (agro-ecological zonation, timing of crops and crop sequences, the rice-wheat system, and water use).

Biotechnology

Biotechnology has become a major tool in international plant-breeding research (Komen et al., 1995). Anther culture is used to create doubled haploid lines. This technology is especially useful for highly heterozygous crops such as maize, sorghum and pearl millet. Embryo rescue by tissue culture is used to cross parents whose embryos normally abort (often in interspecific crosses). Diagnostic techniques with isozyme electrophoresis are applied for detection of resistance genes and identification of genotypes. At present, much emphasis is put on molecular markers, many of which have already been placed on the genetic maps of rice (by IRRI), wheat and maize (by CIMMYT) and sorghum and pearl millet (by ICRISAT), using a technique called RFLP (Restriction Fragment Length Polymorphism) or PCR (Polymerase Chain Reaction)-based methods, e.g. RAPD (Random Amplified Polymorphic DNA). Linkages of molecular markers to QTL (Quantitative Trait Loci), resistances and quality features have been determined. The association between marker alleles and genes controlling agronomically important traits enables breeders to shorten and facilitate the selection process. This has proved to be especially useful in selection for desirable alleles in backcrossing programmes. Efforts with transgenic maize plants containing crystal protein genes from Bacillus thuringiensis have resulted in cultivars with resistance to insects.

Apomixis

Apomixis is the phenomenon of natural asexual seed production, the seed producing a clone of the mother plant. The creation of apomictic cereal cultivars would mean that farmers could rely on cheap farm-saved seed with a superior genetic composition (Jefferson, 1994). Apomixis makes it possible to clone any superior plant with hybrid vigour resulting from a crossing or with useful genes introduced from genetic engineering. Genes which control apomixis are found in some wild plant species, such as in the genus Tripsacum L., an ancestor of maize, and in the genus Pennisetum Rich. In these genera, apomixis is being transferred to the cultivated species by wide hybridization and backcrossing (Ozias-Akins et al., 1993; Savidan & Berthaud, 1994). In both programmes of apomixis transfer, molecular markers are used to assist in the selection of the apomictic mode of reproduction. Research is in progress to introduce apomixis through molecular biology into diverse cereal crops. Since the introgression of new genes requires sexuality, apomictic plants would spell the end of plant breeding, unless the apomixis is facultative and may be switched off temporarily for crossing. Farmers' applications are still a long way off.

Nitrogen fixation

Sorghum and millets are grown largely on rainfed, marginal soils by resource-poor farmers who cannot afford nitrogen fertilizer that can increase yields. In recent years, research has indicated that certain bacteria living in the rhizosphere of sorghum and pearl millet can fix atmospheric nitrogen symbiotically and partially satisfy the nitrogen requirement of these crops. A maximum nitrogen gain of 33 kg/ha per crop was obtained. Statistically significant increases in grain yield have been obtained in maize, pearl millet, sorghum and rice, inoculated with nitrogen-fixing bacteria such as Azospirolla, Azobacter chroococcum and Azospirillum brasilense (Wani, 1986). Seed inoculation might become a practical application.

Prospects

Consumption

Cereals are the dietary mainstay of mankind, providing about 60% of the required energy (Johnson, 1984). The simulation models for food strategies being made to predict the future food situation (Luyten, 1995; Meadows et al., 1991; Penning de Vries et al., 1995) take into account that the world food supply is largely determined by the cereal situation. The following assumptions regarding cereal consumption in South-East Asia seem to be justified:

• In the next century, in an optimistic scenario, the annual population increase is expected to slow down gradually from about 2% at present to about 1.6% in the year 2020. The world population might increase from 5.4 billion in 1996 to about 8.1 billion in 2020, and the population in South-East Asia from 0.46 to about 0.71 billion. The annual world demand for cereals might double during that period, from the present 2 billion to 4 billion t; in South-East Asia from 0.16 to 0.32 billion t.
• In South-East Asia, rising family income will mean that the direct consumption of cereals will decline, largely in favour of more luxury alternatives, mainly vegetables and animal products. However, indirect consumption will increase as a result of greater demand for animal feed. The outcome will be a slightly increased cereal consumption per capita. Cereal products considered as "health food" will become more popular.
• Only a minor part of the consumption of bread and other wheat products in South-East Asia will be met by locally produced wheat or by wheat substitutes, thus wheat imports will rise. The import of maize, mainly for chicken and pig feed will also increase. Only with strong support of the public sector will most countries be able to remain self-sufficient in rice.

Production and trade

Rising labour costs may compel production systems to change to larger units using less labour-intensive technology. Direct sowing of rice will become common practice, which implies the use of more herbicides. The use of fossil energy to power machinery will increase sharply. The worlwide trend towards the bulk production of cereals implies a farming system characterized by farms that are large in size but small in number. However, many scientists point to the negative aspects of such a system: great dependence on purchased inputs, high risk of farm failure, poor system diversity, loss of biodiversity, high risks to environmental quality, high risks to human health, low reliance on rural communities. In the 1990s there is a growing movement in industrialized countries to arrive at an integrated, sustainable production system: use of solar energy, prevention of overconsumption, promotion of biodiversity, and organic farming without chemical pesticides.

Because bulk transport is easy and relatively cheap, cereals have become the main food commodity traded nationally and internationally, of utmost importance to balance national food shortages or surpluses. The spectacular increase in the world cereal production since the Green Revolution in the 1960s, however, has gradually levelled off. Since 1985, population growth has outpaced the increase in grain production, hence production per capita has declined. There are two main reasons for this. First, new land for cropping is becoming scarce. Even in those regions with an expanding economy the area planted with cereals is tending to decrease because farmers are switching to more remunerative crops (vegetables, fruits) or activities (meat, dairy products). Secondly, the increase in yield per ha is tending to level off. It seems to be becoming more difficult to create new cultivars yielding significantly more than existing ones, or to improve cropping techniques bridging the large gap between actual and potential yield. Another reason is the decline of cereal yield in the former Soviet Union.

It is of the utmost importance for global food security and thus for economical and political stability in the world, that sufficient reserves of cereals are maintained. Many countries have sufficient cereals in store to meet several months of their national annual need. However, an alarming development is the shrinking cereal stocks in the world, mainly caused by the rapidly increasing demand for feed cereals in China and by yields below long-term prospects. Between 1990 and 1995 annual world cereal production stabilized at around 1.9 billion t, whereas the surplus quantity in store decreased from 0.38 billion t (21% of the annual consumption) in 1992 to 0.26 billion t (14%) in 1995.

World trade regulations as agreed in the GATT (General Agreement on Tariffs and Trade) strive for increased liberalization of cereal trade as a function of demand and supply. This is in contrast to the tendency of governments to subsidize national cereal production and to prohibit export in periods of national shortage. Food experts do consider free global trade as an important step to food security. Rising cereal prices automatically stimulate farmers to increase their production, while consumers are stimulated to buy alternative staple foods, livestock farmers to use non-cereal feed, and the industry to search for other raw materials. Consequently, consumer prices will be lowered. Yet, in the short term, the replacement of a notable part of the cereal consumption by any other energy food is not feasible. In periods of worldwide cereal shortage, the poorest cereal-importing countries will face big problems in supplying this staple food to low-income classes. So, national policies are directed to guaranteeing an adequate supply of this staple.

Research priorities

Future research priorities reflect the expected demand for cereals and market trends. At present, policy makers point to the urgent need to match the basic food production with the population explosion. In the coming 25 years, governments will probably be willing to strengthen their NARSs out of necessity. The expected increase in cereal demand can only be matched by raising land productivity to almost double the present level. This increase in yield might partly be achieved by genetic improvement, but will also largely be brought about by improved cultural practices, including the use of external inputs. With increased effectiveness of research at the NARSs, the research programmes of the IARCs will move to more fundamental or complex research. More research will be devoted to crop diversification, away from rice. For the main cereals, the use of advanced technology will become normal practice (e.g. molecular markers in breeding programmes), although it may be a decade or more before some impact will appear. Genetic transformation will also gradually be incorporated in breeding programmes. The perpetual fight against newly arising biotypes of diseases and pests will remain an important objective for breeders. The NARSs may restrict their breeding work more to the development of interesting lines for the private seed sector, which is likely to become much stronger than at present. Researchers will be obliged to meet strong demands from society, e.g. relating to the environment and to food quality. And increasing urbanization and industrialization will cause labour costs to rise, forcing farmers to mechanize.

Authors

G.J.H. Grubben, Soetjipto Partohardjono & H.N. van der Hoek

Table 1

Table 2

Table 3

Table 4

Table 5

Table 6

Table 7

Table 8

Table 9

Figure 1