Introduction

Millet is a collective term referring to a number of small-seeded annual grasses that are cultivated as grain crops, primarily on marginal lands in dry areas in temperate, subtropical and tropical regions. The most important species are pearl millet, finger millet, proso millet and foxtail millet (see Annex I for types of millet). Pearl millet accounts for almost half of global millet production. It is the most important species of millet both in terms of cropped area and contributions to food security in regions of Africa and Asia that can produce little else. Finger millet is widely produced in the cooler, higher-altitude regions of Africa and Asia both as a food crop and as a preferred input for traditional beer. Proso millet is important for bird seed in the developed countries and for food in parts of Asia. Foxtail millet is important in parts of Asia (mainly China) and Europe. The other species (barnyard, kodo and little millets, the fonios and teff) are locally important food grains restricted to smaller regions or individual countries. The various species differ in their physical characteristics, quality attributes, soil and climatic requirements and growth duration.

Developing countries, mainly in Asia and Africa, account for about 94 percent of global output, estimated at some 28 million tons (1992-94 average, Table 1). Of this, pearl millet accounts for about 15 million tons, foxtail millet for 5 million tons, proso millet for 4 million tons and finger millet for over 3 million tons (Annex II). Almost all millet is produced by small-scale farmers for household consumption and localized trade. Pearl millet, in particular, is critically important for food security in some of the world's hottest, driest cultivated areas.

Very limited quantities of millet are produced in the developed countries, primarily for a high-value specialty market as bird seed. Correspondingly, only limited quantities of millet are recorded in international trade.

Statistical documentation for millet is generally poor and fragmentary. Few national statistics distinguish between the various botanical species. Some countries combine millet figures with those of sorghum and other cereals, and include millet under the general category "other coarse grains". Many of the statistics are only rough estimates; analyses derived from these data should, therefore, be treated with caution.

Millets are better adapted to dry, infertile soils than most other crops, and are therefore often cultivated under extremely harsh conditions - for example, high temperatures, low and erratic precipitation, short growing seasons and acidic and infertile soils with poor water-holding capacity. Most millets have strong, deep rooting systems and short life cycles, and can grow rapidly when moisture is available. As a result, they can survive and reliably produce small quantities of grain in areas where mean annual precipitation is as low as 300 mm. This compares with a minimum water requirement of 400 mm for sorghum and 500-600 mm for maize. Some species (pearl and proso millets) also appear to tolerate higher temperatures than sorghum and maize, although they do not tolerate long drought periods as well as sorghum.

Millet production systems

In most parts of the world, millet is grown as a subsistence crop for local consumption. Commercial millet production is risky, especially in Africa, because the absence of large market outlets means that fluctuations in output cause significant price fluctuations, particularly in areas where millet is the main food crop. Apart from grain production, millet is also cultivated for grazing, green fodder or silage. Livestock are an important component of most millet production systems, and millet crop residues contribute significantly to fodder supplies. Some popular landrace varieties in India, for example, are over 3-meter tall, and are valued for the large amount of fodder they provide, even though grain yields are relatively low¹.

[¹. Dry millet stover often has a lower total fodder value than stover from sorghum or other grains because the stalks are lignified and have a lower digestible energy content. However, it is often the only fodder available in areas where millet is grown.]

In developing countries, millet cropping systems tend to be extensive, with limited application of improved technologies, except in some of the more commercialized farming regions in India. These crops are usually grown without irrigation or chemical fertilizer, on light, well-drained soils that are poor in organic matter content. When supplementary or full irrigation is available, farmers prefer to cultivate more remunerative crops, although exceptions occur in some regions (such as Gujarat in India) where there is seasonally high demand for pearl millet crop residues as fodder for milch animals. Short-duration millet cultivars are also grown under irrigation, before or after higher-value crops, in areas where the season is long enough to permit double cropping.

For these reasons, and others discussed in the section on Production Trends, millet yields are usually much lower than yields of other cereals (which are grown under more favourable conditions). Although millet occupies about 5 percent of the world's cereal area, it accounts for only 1.5 percent of world cereal production. Furthermore, yields are highly variable from one season to another. In Niger, for example, average pearl millet yields fell from 510 kg/ha in 1988 to 240 kg/ha in 1990, then increased to 360 kg/ha in 1992.

Table 1. Millet area, yield and production by region.¹

		Area (million ha)			Yield (t/ha)			Production (million tons)
		1979-81	1989-91	1992-94	1979-81	1989-91	1992-94	1979-81	1989-91	1992-94
Developing countries		34.70	34.40	35.60	0.68	0.73	0.75	23.67	25.00	26.60
Africa		11.50	15.80	18.50	0.67	0.66	0.61	7.68	10.46	11.36
	Northern Africa	1.10	1.05	1.96	0.40	0.18	0.28	0.44	0.19	0.55
	Sudan	1.10	1.05	1.95	0.40	0.18	0.28	0.44	0.19	0.55
Western Africa		8.30	12.60	14.00	0.67	0.68	0.64	5.52	8.55	9.00
	Burkina Faso	0.80	1.21	1.24	0.49	0.54	0.64	0.39	0.65	0.79
	Ghana	0.18	0.19	0.20	0.64	0.64	0.82	0.12	0.12	0.17
	Cote d'Ivoire	0.06	0.08	0.08	0.58	0.61	0.84	0.04	0.05	0.07
	Mali	0.64	1.19	1.20	0.72	0.69	0.61	0.46	0.82	0.73
	Niger	3.01	4.19	4.87	0.44	0.34	0.38	1.31	1.43	1.86
	Nigeria	2.40	4.50	5.20	1.04	1.04	0.89	2.50	4.67	4.62
	Senegal	0.93	0.90	0.89	0.60	0.64	0.61	0.56	0.58	0.55
	Togo	0.12	0.13	0.13	0.36	0.51	0.50	0.04	0.07	0.06
Central Africa		0.63	0.79	0.93	0.59	0.51	0.48	0.37	0.40	0.45
	Cameroon	0.13	0.06	0.05	0.75	1.06	1.01	0.10	0.06	0.06
	Chad	0.36	0.54	0.59	0.50	0.40	0.47	0.18	0.22	0.28
Eastern Africa		1.46	1.33	1.46	0.89	0.97	0.91	1.31	1.29	1.33
	Ethiopia	0.23	0.25	0.25	0.90	0.95	1.05	0.20	0.24	0.27
	Kenya	0.08	0.10	0.09	1.05	0.67	0.65	0.08	0.07	0.06
	Tanzania	0.45	0.23	0.32	0.80	0.94	0.71	0.36	0.22	0.23
	Uganda	0.30	0.38	0.41	1.59	1.53	1.57	0.47	0.58	0.63
	Zimbabwe	0.35	0.27	0.25	0.43	0.50	0.27	0.15	0.14	0.07
Southern Africa		0.09	0.11	0.21	0.41	0.49	0.18	0.04	0.06	0.04
Asia		22.98	18.29	16.99	0.69	0.79	0.89	15.75	14.45	15.17
Near East		0.19	0.18	0.15	1.02	0.58	0.78	0.19	0.10	0.12
Far East		22.79	18.41	16.84	0.68	0.78	0.89	15.56	14.35	15.05
	China	3.98	2.25	1.90	1.45	1.74	1.93	5.79	3.92	3.67
	India	17.84	15.19	13.95	0.51	0.64	0.77	9.19	9.76	10.70
	Myanmar	0.18	0.17	0.20	0.45	0.69	0.66	0.08	0.12	0.13
	Nepal	0.12	0.20	0.21	0.99	1.16	1.14	0.12	0.23	0.24
	Pakistan	0.51	0.44	0.43	0.50	0.41	0.44	0.25	0.18	0.19
Central America and the Carribean		0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00
South America		0.20	0.04	0.04	1.21	1.49	1.53	0.25	0.06	0.06
	Argentina	0.20	0.04	0.04	1.21	1.49	1.53	0.25	0.06	0.06
Developed countries		2.94	4.13	2.49	0.65	0.88	0.72	1.93	3.64	1.79
	Australia	0.03	0.03	0.03	1.00	0.88	1.05	0.03	0.03	0.03
	United States	0.09	0.15	0.15	1.20	1.20	1.20	0.11	0.18	0.18
	CIS2	2.79	3.92	2.27	0.63	0.87	0.68	1.76	3.40	1.54
World		37.60	38.60	38.10	0.68	0.74	0.74	25.70	28.65	28.38

1. Each figure is a 3-year average for the respective period, e.g., 1979-81.
2. Until 1991, area of the former USSR.
Source: FAO

Crop distribution

In Asia, millet is restricted almost exclusively to two countries, India and China, although Myanmar, Nepal and Pakistan also produce small quantities. India is the world's largest producer, harvesting about 11 million tons per year, nearly 40 percent of the world's output (Fig. 1). Pearl millet, which accounts for about two-thirds of India's millet production, is grown in the drier areas of the country, mainly in the states of Rajasthan, Maharashtra, Gujarat, Uttar Pradesh and Haryana. Finger millet is produced mainly in the state of Karnataka, but also in Orissa, Uttar Pradesh and Tamil Nadu. It is also the most important millet in Nepal and Bhutan. China produces about 3.7 million tons of millet (mainly foxtail) per year, largely in the provinces of Hebei, Shanxi and Shandong.

Millet production in Africa (Fig. 2) is distributed among a much larger number of countries, notably Nigeria (over 40 percent of the regional output), Niger, Burkina Faso, Mali, Senegal and Sudan (Table 1). Pearl millet is grown along the southern peripheries of the Sahara (i.e., the Sahelian countries and the northern parts of the coastal countries in Western Africa) and in the drier areas of Eastern and Southern Africa. Finger millet production is concentrated in Eastern and Southern Africa, where the leading producers are Uganda and Tanzania. As a grain crop, tef is largely confined to Ethiopia. Small quantities of white fonio are grown throughout sub-Sahelian Western Africa, most importantly in Mali. Black fonio is grown in isolated pockets in Nigeria, Togo and Benin. Guinea millet is cultivated only on the Fouta-Djallon plateau of northwestern Guinea and adjacent Sierra Leone. Foxtail and proso millets are very minor crops in Africa, but are cultivated to a limited extent in Kenya and other upland areas in Eastern Africa. Kodo millet is commonly harvested from wild forms in Western Africa, but cultivated forms of this "ditch millet" are only found in Asia. In Latin America, millet production is confined to a small area in Argentina.

Figure 1. The world's major millet producers.

Figure 2. Relative importance of millet worldwide.

Among the developed countries, millet cultivation (almost entirely proso millet) is concentrated in the CIS, particularly in the Russian Federation, Kazhakastan and the Ukraine. Production in North America, Australia and Europe is extremely limited. In some countries, millet is sown as a catch crop when sowing conditions for the main crop are unfavourable. However, even in such situations the grain is sometimes left unharvested and the area simply grazed by livestock.

Production trends

Worldwide, the area sown to millet has remained relatively stable at around 38 million hectares for the past two decades (Table 1). Both production and yield increased by a little over 10 percent through the 1980s, but have remained unchanged since then. Current global production is about 28 million tons, and average yields are 0.75 t/ha. At a regional level, however, there are sharp differences in trends, especially between the two main producers, Asia and Africa (Fig. 3, 4 and 5).

Developing countries

In Asia, millet area declined by 2.4 percent per annum between 1979 and 1994, falling from 23 million to 17 million hectares (Tables 1 and 2). However, part of this decline was compensated by yield increases (1.5 percent per annum). During the past three decades, yields have roughly doubled in China - where they are now among the highest in the world - and increased by more than half in India. This progress is essentially a result of successful breeding research and the widespread dissemination of pearl millet hybrids in India and improved open-pollinated foxtail millet varieties in China.

Output trends in Asia have been heavily influenced by policy changes in China. The millet economy, which was earlier subjected to production quotas and farm and consumer prices set by the government, was virtually fully liberalized by 1987. This allowed farmers to shift to more remunerative crops, and respond to changing consumer preferences. Consequently, the millet area declined from 2.3 million hectares in 1989-91 to 1.9 million hectares in 1992-94, and current production is about half the peak levels reached in the mid 1980s. In fact, the government has discontinued millet procurement as a result of this decline.

Figure 3. Global trends in millet production, 1979-94.

Figure 4. Global trends in millet area, 1979-94.

Figure 5. Global trends in millet yield, 1979-94 (3-year moving average).

Table 2. Millet annual growth rates, 1979-94.

		Area (%/yr)	Yield (%/yr)	Production (%/yr)	Per caput production (%/yr)
Developing countries		0.3	0.4	0.6	-1.4
Africa		4.1	-0.6	3.4	0.6
Northern Africa		2.7	-2.7	-0.1	-2.6
	Sudan	2.7	-2.7	-0.2	-2.5
Western Africa		4.7	-0.4	4.2	1.2
	Burkina Faso	3.8	2.0	5.9	3.0
	Ghana	0.7	3.0	3.7	0.4
	Côte d'Ivoire	2.6	2.5	5.2	1.3
	Mali	5.1	-1.0	4.0	2.3
	Niger	3.9	-1.0	2.8	-1.3
	Nigeria	7.7	-2.3	5.2	2.2
	Senegal	0.1	1.4	1.5	-1.4
	Togo	3.5	-1.5	1.9	-1.2
Central Africa		3.6	-1.3	2.3	0.0
	Cameroon	-6.1	3.3	-3.0	-5.8
	Chad	5.4	-0.5	4.8	2.0
Eastern Africa		0.5	-0.1	0.4	-2.4
	Ethiopia	1.1	0.8	1.8	-1.0
	Kenya	3.7	-2.4	1.2	-2.3
	Tanzania	-2.0	-2.4	-4.4	-7.3
	Uganda	2.3	0.5	2.8	-0.5
	Zimbabwe	-1.6	-2.6	-4.2	-7.2
Southern Africa		5.9	-4.5	1.1	-1.8
Asia		-2.4	1.5	-0.9	-2.8
Near East		-2.0	-3.1	-5.0	-7.7
Far East		-2.4	1.5	-0.9	-2.7
	China	-6.1	1.8	-4.5	-5.8
	India	-1.8	2.7	0.9	-1.2
	Myanmar	0.5	1.5	2.0	-0.2
	Nepal	4.6	1.5	6.2	3.5
	Pakistan	-1.8	-1.2	-3.0	-6.3
Central America and the Caribbean		0.0	0.0	0.0	0.0
South America		-12.5	2.3	-10.5	-12.2
	Argentina	-12.5	2.3	-10.5	-11.8
Developed countries		-0.3	0.4	0.1	-0.9
	Australia	0.6	-0.5	0.2	-1.3
	United States	4.3	-0.1	4.3	3.3
	CIS1	-0.6	0.2	-0.3	-2.0
World		0.3	0.4	0.7	-1.1

1. Until 1991, area of the former USSR.
Source: FAO

Africa is the only region where millet production is growing, having risen from 8 million to over 11 million tons between 1979-81 and 1992-94 (Table 1). Most of the increase in production, however, occurred during the first half of the 1980s and has since been sustained by area expansion, mainly in the Sahel and to a smaller extent in other countries.

For many African countries, millet yields have remained stagnant or fallen (Tables 1 and 2), partly because much of the expansion has been into areas with poor soils and low, erratic rainfall. Overall, millet production has grown slightly faster than population with 'per caput production increasing by 0.6 percent per annum between 1979 and 1994. However, this situation is likely to be reversed in the near future.

In a number of countries, consumption levels of millet have been maintained only through area expansion. Even so, most millet production areas remain food-insecure. As land constraints become more severe, it is imperative that productivity of this key staple is increased to ensure at least minimal food security.

Developed countries

The only millet producer of any significance in the developed countries is the CIS (Table 1). Millet production rose sharply during the 1980s as a result of large increases in productivity (from 0.63 to 0.86 t/ha between 1979-81 and 1989-91). However, subsequent developments were similar to those in China.

In the Russian Federation, for example, production levels were supported by prices unrelated to production costs and determined by production quotas set by the government. When price setting was discontinued and quotas abolished, the market collapsed. Land was shifted from millet to wheat and other grains. Millet output dropped from 2.3 million tons in 1989 to 482,000 tons in 1994.

Kazhakastan traditionally produced millet mainly for export to other parts of the former USSR. When this trade disappeared during the early 1990s, following declining demand, production fell from 1.0 million tons to 300,000 tons per annum.

Production constraints

Millet production in the developing world, particularly in Africa, suffers from a number of constraints - poor soil fertility, low and erratic rainfall, high temperatures, widespread Striga infestation, downy mildew disease and loss of grain to birds.

Population pressures have led to a shortening of fallow periods, which in turn has accelerated the decline in soil fertility. These processes have also prompted the expansion of millet into more marginal lands. The impact is evident in the declining yields of millet in the major producing countries in Africa (Niger, Mali, Nigeria) over the past 15 years. Similar trends are also evident in the harsher millet production environments in Asia (e.g., western Rajasthan in India).

In addition, millets are cultivated on small, fragmented production units and are often intercropped (usually with legumes and sometimes with sorghum or maize).

Unreliable precipitation tends to keep the use of inputs such as chemical fertilizer, pesticides and hired labour to a minimum; and limited commercial demand depresses the incentive to use purchased inputs.

Crop improvement is generally more difficult in millet than in most other crops, largely because of the nature of the environment in which they are grown (see section on Technological Change). National millet improvement programmes began much later and remained weaker than those for many other crops.

Budgets for millet breeding research are low in most countries. Moreover, almost no experience has been acquired on millet breeding in developed countries that could be transferred to developing countries, as has been done for wheat and maize. Among the different types of millet only pearl millet, and to a small extent finger millet, has so far been researched at the international level. Where new technologies for crop and resource management have been developed, adoption has been poor, partly because of inadequate extension, but, equally, because farmers in harsh environments are generally more risk-averse than their counterparts in more favourable environments. The returns from investing labour and capital in millet production may be lower than the gains derivable from such investments in other farm and non-farm enterprises. Further, many new technologies may not be properly tailored to farmers' severely resource-constrained circumstances.

Varietal improvement

Hybrid breeding programmes have traditionally targeted the relatively better environments, although even these environments are harsher than those for most other crops. Hybrid grain cultivars have been developed for pearl millet in India and the United

States, but perform best in areas where rainfall is reliable. In drier areas with more erratic rainfall, it is far more difficult and time-consuming for breeders to identify dual-purpose grain/stover combinations that are superior across a range of growing conditions. For these areas, crop breeders have concentrated on developing open-pollinated varieties that give stable grain and straw yields and suit the prevailing rainfall pattern, rather than on attempting to maximize yield potential under more favourable conditions.

Grain yields of improved cultivars grown with low to moderate inputs can exceed those of local landraces by about 20 percent; an even more important advantage is that they often mature earlier, and thus perform better under terminal drought stress. However, the adoption of improved varieties remains poor outside a few countries, such as China (foxtail millet), the CIS (proso millet), India (pearl millet) and the United States (proso, foxtail and pearl millets). For example, in Niger, the world's fourth largest producer, improved varieties account for only 5 percent of the millet area, in part because the seed multiplication and distribution system is inadequate.

Utilization

Although millet represents less than 2 percent of world cereal utilization, it is an important staple in a large number of countries in the semi-arid tropics, where low precipitation and poor soils limit the cultivation of other major food crops.

Millet utilization is mostly confined to the developing countries, even more so after production and utilization fell sharply in the CIS, the largest producer in the developed world. Accurate data are not available for most countries, but it is estimated that about 80 percent of the world's millet (and over 95 percent in Asia and Africa) is used as food, the remainder being divided between feed (7 percent), other uses (seed, beer, etc.,) and waste (Table 3).

Table 3. Millet utilization by type, region and selected countries, 1992-94 average.

		Direct food ('000 tons)	Feed ('000 tons)	Other uses1 ('000 tons)	Total ('000 tons)	Per caput food use (kg/yr)
Developing countries		21776	966	3767	26509	5.08
Africa		8673	187	2328	11188	13.40
	Burkina Faso	683	2	126	811	68.52
	Chad	217	0	41	258	33.73
	Ethiopia	108	0	153	260	1.97
	Mali	658	3	119	781	74.63
	Niger	1440	17	259	1716	162.45
	Nigeria	3315	100	1155	4570	31.50
	Senegal	505	5	83	593	61.61
	Sudan	364	20	76	460	14.14
	Tanzania	177	2	53	233	6.41
	Uganda	517	20	95	633	25.93
Asia		13103	748	1433	15284	4.17
	China	3277	327	257	3861	2.74
	India	9216	283	1100	10599	10.23
Central America and the Caribbean		0	0	0	0	0.00
	South America	0	31	6	37	0.00
Developed countries		513	970	323	1805	0.40
	North America	0	180	0	180	0.00
	Europe	0	4	1	5	0.00
	CIS	504	736	316	1555	1.73
	Oceania	0	1	0	1	0.00
World		22289	1936	4090	28314	4.00

1. For seed, manufacturing purposes and waste.
Source: FAO

Food use

Per caput food consumption of millet varies greatly between countries; it is highest in Africa, where millet is a key food staple in the drier regions. Millet represents about 75 percent of total cereal food consumption in Niger and over 30 percent in most other countries in the Sahel. It is also important in Namibia (25 percent of total cereal food consumption) and Uganda (20 percent).

Outside Africa, millet food consumption is important in parts of India, China and Myanmar. Utilization is negligible in Latin America, the Caribbean and all the developed countries. The exception was the former USSR until the late 1980s. However, under the economic transition process and the removal of price subsidies, utilization in the USSR/CIS fell sharply.

Millet is a high-energy, nutritious food, especially recommended for children, convalescents and the elderly. Several food preparations are made from millet, which differ between countries and even between different parts of a country. These consist primarily of porridge or pancake-like flat bread. However, because wholemeal quickly goes rancid, millet flour (prepared by pounding or milling) can be stored only for short periods.

Millet is traditionally pounded in a mortar, but mechanical dehulling and milling are increasingly used since they eliminate a considerable amount of hard labour and generally improve the quality of the flour.

Worldwide, millet food consumption has grown only marginally over the past 30 years, while total food use of all cereals has almost doubled. Millet is nutritionally equivalent or superior to other cereals². However, consumer demand has fallen because of a number of factors, including changing preferences in favour of wheat and rice (cheap imports are available in several countries), irregular supplies of millet, rising incomes and rapid urbanization. Particularly in urban environments, the opportunity cost of women's time has encouraged the shift from millet to readily available processed foods (milled rice, wheat flour, etc.,) that are far quicker and more convenient to prepare.

[². Protein contents in pearl, proso and foxtail millets are comparable with those in wheat, barley and maize. Finger millet has a slightly lower protein content, but is in fact nutrition-ally superior because the protein quality is generally as good or better than in other cereals. Finger millet is also high in calcium and iron, and contains fairly high levels of methionine, a major limiting amino acid in many tropical cereals.]

Animal feed

Utilization of millet grain as animal feed is not significant. It is estimated that less than 2 million tons, (about 7 percent of total utilization), is fed to animals, compared with about 30 million tons of sorghum (almost half of total output). In the developing countries, use of millet grain for animal feed is concentrated in Asia; very little is fed in Africa. However, millet fodder and stover are a valuable and critical resource in the crop/livestock systems where millet is grown.

Feed use estimates are heavily influenced by assumptions made for China, the world's third largest producer. In fact, little reliable information is available on feed use in this country. Based on very rough calculations of feed use in the CIS, it is estimated that about 1.0 million tons per annum are currently used as animal feed in the developed countries (Table 3). Western Europe, North America and Japan together use slightly over 200,000 tons, almost exclusively as bird seed. Recent increases in the use of pearl millet as a lower-cost substitute for maize feed in aquaculture and dairy and poultry farming in India and the southeastern United States are not well documented, and, in any case, at the moment represent only a small fraction of overall feed grain utilization.

Feeding trials have shown that pearl millet grain compares favourably with maize and sorghum as a high-energy, high-protein ingredient in feed for poultry, pigs, cattle and sheep. Nevertheless, very little millet is used as feed. First, as millet is grown mostly on marginal lands and production is barely sufficient to satisfy food requirements, little surplus is left for animal feed. Second, millet production fluctuates widely from year to year because of rainfall variability and drought in the main production areas. This deters a closer integration of millet production with intensive livestock operations. Third, millet yields are generally lower than those of other crops produced commercially in more favourable environments. Thus, production and transport costs are often prohibitive compared to alternative ingredients of compound feeds.

Other uses

There are few other uses of millet. Small quantities of finger millet are used in Zimbabwe for commercial brewing and opaque beer. Food technologists have experimented with the incorporation of pearl millet into composite flour, but the commercial application of this technology is limited.

International trade, market prices and stocks

Global trade in millet is estimated to range between 200,000 and 300,000 tons (Table 4a and b), representing roughly 0.1 percent of world trade in cereals or 1.0 percent of world millet production. The major exporters are India, the United States, Argentina and China, which together supply about two-thirds of recorded exports. A sizeable proportion (about 100,000 tons) of the recorded international trade is in proso millet, exported by the United States, Argentina and Australia to other developed countries. Another 60,000 tons are pearl millet exports by India. In recent years, China has started to export some quantities of foxtail millet.

The European Community³, accounts for more than 50 percent of global imports. During the 1992-94 period, the EC purchased an average of 145,000 tons per year (Table 4b). Other major, regular importers are Japan, Switzerland and Canada. In contrast, countries like Kenya, Mauritania, Nigeria, Tanzania, and Uganda no longer import millet, at least officially.

[3. Under the Lomé IV Agreement, ACP countries (Africa, Caribbean, Pacific) can export up to 60,000 tons per year to the EC without import levies.]

Besides this official trade, a substantial unrecorded quantity of millet is traded within subregions in Africa, with grain moving from surplus to deficit areas. In Western Africa, for example, there is movement of millet during good years from surplus producing areas along the southern boundary of the Sahara both southward to higher-rainfall but millet-deficient areas and northward to supply nomadic populations.

The magnitude of officially recorded trade has marginally declined over the past 20-30 years, and there has been a slight change in direction. Imports by developed countries have tended to decrease over the past two decades, while those by the developing countries remained steady through the 1960s, rose during the 1970s, but fell thereafter, having been replaced by rising imports of wheat and rice. Developed countries now account for an estimated 70 percent of recorded world imports, compared with about 50 percent during the early 1960s.

Table 4a. Millet recorded international trade: exports¹.

Exports		1979-81 ('000 tons)	1989-91 ('000 tons)	1992-94 ('000 tons)
Africa		57.9	26.4	20.2
	Mali	0.02	15.0	18.0
	Niger	36.7	0.1	0.0
	Sudan	2.1	1.3	0.0
Asia		12.0	16.8	84.6
	China	8.7	4.6	21.6
	India	0.0	7.0	58.5
North, Central and South America, and the Carribean		145.9	119.1	90.3
	Argentina	112.9	41.0	42.9
	United States	33.0	75.5	45.5
Europe		20.5	33.0	43.7
	EC (12 countries)3	15.6	22.6	28.3
	Hungary	4.4	6.9	13.0
Oceania		14.6	13.6	16.3
	Australia	14.6	13.6	16.3
World		250.9	208.7	255.0
Developing countries		181.9	84.0	147.4
Developed countries		69.0	124.7	107.6

1. Each figure is a 3-year average for the respective period, e.g., 1979-81.
2. Shown as zero for trade less than 50 tons.
3. Including intra-trade among member countries.
Source: FAO

Table 4b. Millet recorded international trade: imports¹.

Imports		1979-81 ('000 tons)	1989-91 ('000 tons)	1992-94 ('000 tons)
Africa		82.0	7.9	40.9
	Angola	0.02	0.0	21.7
	Côte d'Ivoire	0.0	2.6	1.2
	Gabon	0.0	0.0	0.1
	Mauritania	1.0	0.0	0.0
	Mali	40.0	0.0	0.5
	Niger	8.0	2.4	0.5
	Nigeria	26.7	0.5	0.0
	Senegal	0.0	2.1	15.0
	Sudan	0.0	0.0	0.3
	Zimbabwe	0.2	0.0	0.3
Asia		58.9	40.3	44.3
	Japan	53.1	23.7	20.3
	Kuwait	1.1	0.3	0.5
	Malaysia	0.7	2.1	2.4
	Saudi Arabia	1.4	1.5	2.8
	Singapore	0.4	1.1	0.7
	Thailand	0.8	1.7	1.8
North, Central and South America and the Caribbean		4.0	26.2	18.1
	Brazil	3.8	3.7	5.8
	Canada	0.0	5.9	8.2
	Europe	145.7	145.5	155.4
	Austria	2.3	1.2	0.8
	EC (12 countries)3	114.9	131.5	145.2
	Switzerland	26.0	9.2	8.2
Oceania		0.8	0.8	5.7
World		291.4	220.8	264.3
Developing countries		90.5	43.8	75.4
Developed countries		201.0	177.0	188.8

1. Each figure is a 3-year average for the respective period, e.g., 1979-81.
2. Shown as zero for trade less than 50 tons.
3. Including intra-trade among member countries.
Source: FAO

Note: The discrepancies between imports and exports are largely because some exporting countries do not report millet sales at all, or include them under "other cereals".

International trade in millet is controlled by a few specialized trading companies and generally conducted on a sample basis. Only Argentina is reported to have established official export quality standards. International prices are highly volatile, determined largely by supply volumes, and are usually unrelated to those of other major coarse grains such as maize, sorghum or barley. Quotations⁴ are not regularly published or recorded according to official statistics. Table 5 therefore compares export prices for millet in Argentina, Australia and the United States. The high degree of price variability among suppliers, even in the same year, is due to the "thin" market, with small trade volumes and very few buyers and sellers.

[4. International quotations are published in the Public Ledger, London, for UK imports and in STAT (Stat Publishing, Blaine, WA 98230, BC, Canada) for Argentine and US exports.]

Since millet yields in the major exporting countries are substantially lower than for other cereals, and taking into consideration the opportunity cost of growing millet rather than other crops, prices have to be considerably higher to make cultivation remunerative. As a result, prices are generally higher than those of other grains, except in India, where millet and sorghum prices are roughly equal. These prices discourage the use of millet in compound feed. Only on rare occasions when prices are extremely low is millet used as a substitute for sorghum or maize in feed formulations.

Table 5. Average annual export prices for millet.

Year	Argentina (US$/ton)	United States (US$/ton)	Australia (US$/ton)
1979-81 average	129	186	224
1983	147	175	251
1984	166	176	254
1985	107	171	210
1986	139	151	195
1987	108	154	162
1988	123	173	110
1989	190	177	249
1990	143	188	318
1991	107	156	249
1992	114	170	249
1993	156	223	245
1994	228	254	325

Source: FAO

World millet stocks, currently estimated at 3.0 million tons, are relatively unimportant in the global cereal context. They represent only 1.0 percent of world cereal carryover stocks. Most non-commercial stocks are held by farmers (but not officially recorded) in developing countries for household consumption, seed and limited trading on local markets. Such stocks, while critical in terms of food security at household or local level, are not significant in terms of global trade.

Most millets have excellent storage properties and can be kept for up to 4-5 years even in simple storage facilities, such as traditional granaries. This is because the seeds are protected from insect attack by the hard hull covering the endosperm, and because grain is usually harvested and stored in dry weather conditions. Thus, although there may be large year-to-year variations in production, stocks can easily be built up after favourable years.

Internal marketing and domestic policies

Millet marketing channels in many developing countries are not well developed. There are three main reasons: scattered and irregular supplies, large distances between producing areas and the main urban centres and limited demand in urban areas. Only 15-20 percent of the pearl millet produced in India, and perhaps 5-10 percent in Africa, enters the commercial marketing system. Moreover, although a number of developing countries have market intervention regulations to stabilize domestic millet prices, these regulations are effectively enforced in only a few countries. In several cases, the large year-to-year variations in the size of the harvest make it difficult for governments to provide adequate farm income support and simultaneously maintain adequate stocks in anticipation of lean years. Also, many governments do not include millet in their farm price-support programmes.

Technological change, environmental issues and focus of research

Research has generally focused on pearl millet, the most important species. Adaptation is a more serious problem in millet than in many other crops. Pearl millet originated in Western Africa, evolving in a harsh environment in association with a number of diseases and insect pests. Local landraces developed through natural and human selection gave poor yields, but showed reasonable tolerance to many of these hazards. Improved varieties introduced from Asia, Eastern Africa or the United States, where these problems are less prevalent, have generally failed to show any superiority over local varieties in Western Africa. Indeed, these introduced materials generally have serious adaptation problems, because of lack of tolerance to high soil temperatures and sandstorms as seedlings, and greater susceptibility than local landraces to diseases and insect pests.

Breeding for drought tolerance, a major problem in all millet environments, is also difficult because it is hard to accurately simulate drought conditions and because manipulating a plant trait to improve tolerance to severe drought stress at one growth stage may result in increased sensitivity to drought at another stage.

These difficulties notwithstanding, national and international research programmes have made significant advances. Improvements in pearl millet yields in the developing countries, mainly in India, have occurred largely due to the development, release and widespread multiplication of improved open-pollinated and hybrid cultivars. Because of their low sowing rates (3-4 kg/ha of seed) and high multiplication rates (200 to 500-fold per generation), these improved cultivars have been adopted fairly widely even by subsistence farmers in specific, relatively favourable, millet environments, e.g., Gujarat, Haryana and Maharashtra in India.

In Africa, hybrids have yet to make a significant impact because extension is inadequate and the seed industry poorly developed. However, there has been some adoption of improved open-pollinated varieties in Southern Africa. These offer marginal gains in grain yield over traditional landraces, and are less prone to end-of-season drought because they mature earlier, thus reducing the risk of crop failure.

The problem of environmental degradation is common for many crops, but particularly serious in millet. Population growth has forced farmers to shorten fallow periods (which in turn has resulted in declining soil fertility) and to expand millet cultivation into more marginal lands. This is most evident in Africa, where millet area has increased and yields have declined over the past 15 years. This problem is less serious in Asia; however, in the harsher environments in Rajasthan, the expansion into more marginal areas will make future productivity increases harder to achieve.

Improving the reliability of grain and stover yields continues to be the major focus of millet research. Current millet research is moving in two broad directions:

· shorter crop life cycles of 70-80 days (as against the more usual 90 days), so that the plant can escape end-of-season drought;

· better tolerance to mid-season drought.

Crop improvement programmes are now integrating farmers more closely than before into the breeding and diffusion process in order to develop technologies that are more accurately targeted at farmers' constraints, and, therefore, more likely to be adopted. The emphasis is on breeding cultivars with durable resistance to downy mildew and foliar diseases. Other important research objectives are to:

· identify Striga resistance sources and develop better crop management technologies to reduce losses due to Striga, particularly in Western Africa;

· develop integrated pest management strategies to reduce losses due to the millet stem borer.

Medium-term outlook5

[⁵. Covers the period from 1992-94 to 2005. The supply outlook is based on estimates of future area and yields projected from recent trends, with some adjustments based on judgement of how individual countries are likely to perform, assuming no major policy changes. Demand projections are based on United Nations population projections and World Bank income growth rates.]

World millet production is projected to increase from 28 million tons (1992-94 average) to about 33 million tons in the year 2005 (Tables 6 and 7). Most of the growth will be in the developing countries, where production is projected to grow at 1.4 percent per annum from 27 million tons in 1992-94 to 31 million tons in the year 2005. Africa is expected to show the highest growth rates (2.4 percent per annum) and the largest absolute increase in production.

At the global level, growth will come mainly from yield increases (Table 7). In Africa, both area expansion (1.1 percent per annum) and yield increases (1.4 percent) will contribute. However, growth in output will remain slower than population growth, and per caput consumption in Africa will decline. In Asia, production is projected to increase marginally from 15 million tons in 1992-94 to 16 million tons in 2005. The increase will come mainly from higher productivity - yields are expected to grow from 0.9 t/ha in 1992-94 to about 1.1 t/ha in 2005. Most of the production growth is expected to occur in India. In China, millet yields are already among the highest in the world and will increase still further, especially if millet hybrids, which are still not widely used, are developed and disseminated. However, overall output is likely to fall because land-use patterns are changing in favour of other agricultural products and economic activities.

Table 6. Projected millet production, demand and trade ('000 tons), 1992-94 to 2005.

		Actual (1992-94 average)					Projected (2005)
		Production	Total use	Food use	Feed use	Trade gap1	Production	Total use	Food use	Feed use	Trade gap1
Developing countries		26,592	26,509	21,776	966	83	31,394	31,421	25,510	1,542	-27
Africa		11,358	11,188	8,673	187	170	15,072	15,138	11,705	438	-66
	Northern Africa	554	467	365	26	87	765	738	499	136	27
	Western Africa	8,986	8,921	6,987	129	65	12,024	12,051	9,430	243	-27
	Central Africa	447	435	356	4	12	505	502	437	10	3
	Eastern Africa	1,332	1,326	929	29	6	1,740	1,792	1,289	49	-52
	Southern Africa	39	39	35	0	0	39	55	50	0	-16
Asia		15,171	15,284	13,103	748	-113	16,229	16,235	13,805	1,104	-6
	Near East	117	118	75	30	-1	148	145	79	49	3
	Far East	15,054	15,166	13,028	718	-112	16,081	16,090	13,726	1,055	-9
South America		63	37	0	31	26	92	46	0	39	46
Developed countries		1,786	1,806	513	970	-20	1,662	1,595	488	1,009	67
World		28,378	28,314	22,289	1,936	64	33,056	33,016	25,998	2,591	40

1. Production minus utilization.
Source: FAO/ICRISAT

Food demand

Millet will continue to be used primarily for human food, and will remain a major source of calories and a vital component of food security in semi-arid areas in the developing world. With the exception of the CIS, food use will remain confined to the developing countries, which currently account for 98 percent of total food use. Food demand for millet in these countries is expected to grow at 1.3 percent per annum between 1992-94 and 2005, with important differences in growth patterns between Asia and Africa. Asia now accounts for 59 percent of total millet food use and Africa for 39 percent. By the year 2005, however, Asia's share will fall to 53 percent while Africa's will rise to 45 percent (Table 6).

In Asia, food use is projected to grow by only 0.4 percent per annum (Table 7), as consumers shift to other foods. Growth will be stronger in Africa (2.6 percent per annum), but will be constrained by supply rather than demand factors. As per caput consumption falls, calories for African households will have to be provided increasingly by imports or by food grains produced in higher potential areas within each country.

One major concern is the likelihood of growing millet deficits, particularly in Africa. By 2005, the projected millet deficits for Africa will be 66,000 tons per year, as against a current "surplus" of 170,000 tons (Table 6). Some of this deficit could be covered by imports or food aid. However, falling per caput production could have serious consequences for food security and nutrition in a region that already experiences frequent food shortfalls. It should be noted that even this projection is based on a somewhat optimistic production growth rate of 2.4 percent per annum between now and 2005.

Feed demand

Global demand for millet as feed in 1992-94 was 1.9 million tons, projected to grow to about 2.6 million tons in 2005 (Table 6). Again, whether such growth will be achieved depends largely on the developing countries. Their feed use of millet is projected to increase by 60 percent by the year 2005, depending mainly on developments in India, Nigeria and Sudan. Feed use in China is expected to decline. The use of millet for bird seed in developed countries is unlikely to change significantly from the current level. These countries are expected to continue to use millet mainly as bird seed because it is too expensive to be competitive as an ingredient in livestock feed, except as a locally produced feed grain on light soils in parts of the Unites States.

Trade

Future world trade in millet is very difficult to project because of its small size, the unknown volume of unrecorded trade and uncertainties regarding both supply and demand. If larger surpluses of millet become available in some countries (for example, in Western Africa), trading opportunities in those regions would increase. However, in view of the huge distances and the high transport costs, and the large variability of tradeable volumes, any significant trade expansion is unlikely. Most international trade in millet up to the year 2005, therefore, is envisaged to remain largely restricted to border transactions among developing countries and limited but regular purchases by the developed countries as in the past.

Table 7. Millet projected growth rates, 1992-94 to 2005.

		Area (%/yr)	Yield (%/yr)	Production (%/yr)	Per caput production (%/yr)	Utilization
						Total (%/yr)	Food (%/yr)	Feed (%/yr)
Developing countries		0.2	1.2	1.4	-0.5	1.4	1.3	4.2
Africa		1.1	1.2	2.4	-0.7	2.6	2.5	7.4
	Northern Africa	0.9	1.8	2.7	0.4	3.9	2.6	14.8
	Western Africa	1.1	1.4	2.5	-0.9	2.5	2.5	5.4
	Central Africa	-0.1	1.1	1.0	-2.1	1.2	1.7	7.9
	Eastern Africa	1.8	0.4	2.3	-1.0	2.5	2.8	4.5
	Southern Africa	6.6	-6.1	0.5	-3.0	2.8	3.0	0.0
Asia		-0.9	1.5	0.6	-1,1	0.5	0.4	3.3
	Near East	0.0	1.9	1.9	-0.7	1.7	0.4	4.2
	Far East	-0.9	1.5	0.6	-1.0	0.5	0.4	3.3
South America		2.5	0.8	3.3	1.6	1.8	0.0	1.9
Developed countries		-2.2	1.6	-0.6	-1.1	-1.0	-0.4	0.3
World		0.1	1.2	1.3	-0.3	1.3	1.3	2.5

Source: FAO/ICRISAT

Summary and Conclusions

Pearl millet is grown largely for its ability to produce grain under hot, dry conditions on infertile soils of low water-holding capacity, where other crops generally fail completely. Correspondingly, it is produced mainly in outlying areas peripheral to the major production and population centres of the developing world. Yields are low, averaging only three-quarters of sorghum yields in Africa and Asia. Most farmers who rely on this crop are quite poor and frequently experience food shortfalls. Little of the millet production enters the commercial market; most never leaves the farm on which it is grown. Rather, many millet farmers are more likely to be food buyers than sellers.

The combination of poverty and severe environmental conditions makes it difficult to improve productivity in pearl millet. While yields are growing in Asia, many African producers are unable to raise yields because of the continuing expansion into even drier and harsher agroecologies and poor adoption of "improved" technologies in these environments. A major reason for poor adoption is that some of these technologies are expensive or otherwise inappropriate for these harsh environments.

The growth of pearl millet yields in Asia is due to the adoption of improved cultivars (both hybrid and open-pollinated) and at least limited investments in fertility maintenance. Farmers are also expanding investments in water conservation technologies as land constraints become more severe. Yield improvements would be greater if the move to more remunerative oilseed crops (e.g., groundnut, sesame and castor) were not so prevalent in the more favourable pearl millet production areas in Asia.

In Africa, by contrast, most farmers continue to plant traditional landrace cultivars. While there are signs of interest in new open-pollinated cultivars, private seed companies do not believe this area is profitable, and public sector investments in seed production are limited. The widespread promotion of hybrids in Asia has encouraged private investment in seed production, but the prospects for hybrid adoption in Africa remain unknown. The costs of distributing hybrid seed are higher than in Asia (because population densities are lower), and the willingness of the often poorer African farmer to purchase hybrid seed remains untested. However, given the low seed requirement and the low production costs (because of high multiplication rates), even poor pearl millet producers in Africa, similar to their counterparts in Asia, may find it worthwhile to invest in improved seed, either hybrid or open-pollinated.

There are strong justifications for more government investment in millet seed production and distribution as a means of reducing the costs of relief during droughts. Care must be taken to ensure that certification regulations for improved seed are reasonable and enforceable. Previous experience has shown that unreasonably stringent regulations serve only to restrict competition in the seed industry, resulting in seed shortages and unnecessarily high seed prices.

Prospects for the adoption of improved management technologies in both Africa and Asia are limited, for several reasons. Firstly, the high variability in annual rainfall, especially in Africa, makes it difficult for farmers to judge potential investment returns. Secondly, labour constraints restrict the adoption of improved soil and water conservation systems as households send children to school and adults to urban areas in search of employment. And thirdly, farmers judge the returns to cash investment in inputs, such as fertilizer, against the gains obtained by saving to buy food or livestock or education for their children.

Such factors require scientists and extension workers to be more imaginative in developing technologies suited to these difficult production environments. Breeders need to consider more carefully the trade-offs that farmers calculate between grain and fodder, between yield and yield stability, and between input responsiveness and productivity under low-input conditions. Resource management scientists must assume that farmers' decisions will change depending on rainfall levels over the course of the season, and target narrow opportunities for even marginal improvements in water-use efficiency and soil fertility. These may include aiming for a small investment in chemical fertilizer to complement the use of manure, or a legume rotation rather than a short-run profit-maximizing investment entailing higher production risk.

The prospects for expanding commercial trade in Africa are limited. The biggest opportunity lies in the expansion of trade between surplus and deficit rural households. This is made difficult by the variability of year-to-year production and the long distances between households in areas of relatively low population density. Traders face difficulties in identifying surplus and deficit areas, and the costs of grain collection and transport are high. However, there may be scope for improvements in market information systems and investment incentives to encourage private investment in grain trade. These have proven beneficial in India where there is greater, but still limited, commercial trade in millet grain. These investments can be justified as a component of national and regional drought relief strategies. In areas where inter-seasonal and inter-annual millet prices are highly variable, drought relief programmes could also seek to strengthen household and village grain stocks, for which millet is well adapted.

Small quantities of millet grain are traded for use as flour and beer malt in both Africa and Asia. In Africa, low productivity and high transport costs will restrict this trade to a high-priced premium market. In Asia, higher productivity and lower marketing costs (associated with higher population densities and better market infrastructure) offer better prospects for expanding millet sales. However, it will still be difficult for millet to compete with other cereals grown on substantially more productive land in regions with higher rainfall. In areas where millet is competitive in terms of price and feed value, demand for millet grain for fish and poultry feed may grow. Pearl millet has the advantage of superior adaptation to high temperatures and infertile soils with low water-holding capacity. In specific areas where these constraints are important, millet grain will compete effectively as a livestock feed against other cereals that must be transported across long distances at considerable expense. Further, there will remain a market niche for millet trade as bird seed.

There are indications that pearl millet is becoming increasingly important as a forage crop and as a cover crop or mulch for intensive legumes production on tropical acid soils. Further, it appears that pearl millet will soon become a regionally important alternative feed grain in subtropical areas in several countries. However, these new uses are relatively minor compared to the importance of millets as food crops of the rural poor, primarily in the tropics.

In sum, millet will remain largely associated with the food security of drought-prone human populations. Productivity has lagged, particularly in Africa, because of the severity of this environment and the pressure of human population growth on traditional land-extensive fallow systems. Correspondingly, productivity improvements will contribute most directly to the alleviation of poverty and food insecurity. The prospects for the expansion of market flows are reasonable if targeted within food-deficit areas. However, the prospects for commercial trade are limited, except in small specialty markets for flour, malt, feed grain and bird seed.

Annex I: Types of millet

The various millet species can be divided into two broad categories: pearl millet and "small" millets. The latter group, with the exception of proso millet, have smaller grains than pearl millet.

Pearl millet

Pearl millet(Pennisetum glaucum, P. typhoides, P. tyhpideum, P. americanum) is the most widely grown of all millets. It is also known as bulrush millet, babala, bajra, cumbu, dukhn, gero, sajje, sanio or souna.

Pearl millet is a traditional crop in Western Africa, particularly in the Sahel; in Central, Eastern and Southern Africa; and in Asia, in India and Pakistan and along the southern coast of the Arabian peninsula.

Pearl millet has been recently introduced as a grain crop in the southeastern coastal plain of the United States, where it has been used as a summer forage. Pearl millet can be grown on poor, sandy soils in dry areas that are unsuitable for maize, sorghum or finger millet. It is a summer cereal grass with large stems, leaves and heads. It is more efficient in its utilization of moisture than sorghum or maize.

The grain grows on condensed panicles (spiked) 10 to 150 cm in length. Pearl millet has the highest yield potential of all millets under drought and heat stress.

Finger millet

Finger millet (Eleusine coracana), known as ragi in India, is another important staple food in Eastern Africa and in Asia (India, Nepal). It has a slightly higher water requirement than most other millets and is found in cooler, elevated regions up to 2000 metres above sea level. The plant carries several spikes or "fingers" at the top of the stem. The grain is small (1-2 mm in diameter).

Proso or Common millet

Proso or Common millet (Panicum miliaceum) is grown in temperate climates. It is widely cultivated in the Russian Federation, the Ukraine, Kazhakastan, the United States, Argentina and Australia. The plant has open, branching, drooping panicles and is tolerant of a wide range in temperature.

Foxtail millet

Foxtail millet (Setaria italica) is also adapted to moderate climates. It produces long, cylindrical or lobed, bristly, condensed panicles. China ranks first in the production of foxtail millet in the world. It is grown there for both food and feed. The crop is also grown in India, Indonesia, the Korean peninsula, and some parts of southern Europe. It is not grown to any extent in Africa outside the eastern highlands. Prior to the availability of sorghum-sudangrass forage hybrids, foxtail millet was an important temporary pasture species.

Teff (Eragrostis tef) is a very small-seeded grass that is cultivated for grain in the Ethiopian highlands, where its production exceeds that of most other cereals. It tolerates heavy soils with poor drainage characteristics. Several of its relatives are highly valued forage grasses in the world's arid zones.

White fonio (Digitaria exilis), Black fonio (Digitaria iburua), and Guinea millet (Brachiaria deflexa) are minor cereals of dry areas in sub-Sahelian Western Africa. White fonio is cultivated throughout much of this region, except Liberia. It is a very important crop in southern Mali, northeastern Nigeria, extreme southern Niger, western Burkina Faso, eastern Senegal and northern Guinea. Black fonio is found in isolated pockets in the Jos-Bauchi plateau of Nigeria and the northern parts of Togo and Benin. Guinea millet cultivation is confined to the Fouta-Djallon plateau of Guinea and Sierra Leone.

There are several other "minor" millets, some of which are of regional importance.

Barnyard millet

Barnyard millet (Echinochloa crusgalli, E. colona) is important in the tropics and subtropics of India.

Little millet

Little millet (Panicum sumatrense) is widely grown in India, Nepal, Pakistan, Sri Lanka, eastern Indonesia and western Myanmar.

Kodo millet

Kodo millet (Paspalum scrobiculatum) is harvested as a wild cereal in Western Africa and India, where it grows abundantly along paths, ditches and low spots. The species was domesticated in India about 3000 years ago.

Job's tears (Coix lachryma-jobi) is a minor cereal even among the small millets, with production confined largely to Southeast Asia.

Annex II: Relative importance of millet species, 1992-94.

		Total millets ('000 tons)	Pearl millet (%)	Finger millet (%)	Proso millet (%)	Foxtail millet (%)	Teff millet (%)	Fonio millet (%)	Other millets (%)
Developing countries		26591	55	12	9	20	1	0	3
Africa		11358	87	7	0	0	1	0	5
Northern Africa		554	98	2	0	0	0	0	0
	Libya	2	100	0	0	0	0	0	0
	Morocco	5	100	0	0	0	0	0	0
	Sudan	547	82	18	0	0	0	0	0
	Western Africa	8986	95	0	0	0	0	0	5
	Benin	26	100	0	0	0	0	0	0
	Burkina Faso	793	99	0	0	0	0	0	1
	Côte d'Ivoire	71	85	0	0	0	0	0	15
	Gambia	39	97	0	0	0	0	0	3
	Ghana	166	100	0	0	0	0	0	0
	Guinea	13	95	0	0	0	0	0	5
	Guinea-Bissau	26	100	0	0	0	0	0	0
	Mali	732	95	0	0	0	0	5	0
	Mauritania	4	100	0	0	0	0	0	0
	Niger	1858	100	0	0	0	0	0	0
	Nigeria	4620	98	0	0	0	0	0	2
	Senegal	549	100	0	0	0	0	0	0
	Sierra Leone	24	100	0	0	0	0	0	0
	Togo	64	100	0	0	0	0	0	0
Central Africa		447	87	13	0	0	0	0	0
	Cameroon	55	100	0	0	0	0	0	0
	Central African Republic	11	87	13	0	0	0	0	0
	Chad	282	100	0	0	0	0	0	0
	Zaire	32	67	33	0	0	0	0	0
Eastern Africa		1332	35	50	0	0	9	0	6
	Burundi	12	0	100	0	0	0	0	0
	Eritrea	14	66	17	0	0	17	0	0
	Ethiopia	265	0	14	0	0	86	0	0
	Kenya	57	55	45	0	0	0	0	0
	Malawi	9	40	60	0	0	0	0	0
	Mozambique	19	80	20	0	0	0	0	0
	Rwanda	1	0	100	0	0	0	0	0
	Tanzania	230	70	30	0	0	0	0	0
	Uganda	634	6	94	0	0	0	0	0
	Zambia	49	40	60	0	0	0	0	0
	Zimbabwe	67	70	30	0	0	0	0	0
Southern Africa		39	100	0	0	0	0	0	0
	Angola1		80	20	0	0	0	0	0
	Botswana	2	100	0	0	0	0	0	0
	Namibia	37	100	0	0	0	0	0	0
Asia		15171	34	16	14	33	0	0	3
Near East		117	60	31	2	5	0	0	2
	Afghanistan	22	0	100	0	0	0	0	0
	Iran, Islamic Republic of	11	0	100	0	0	0	0	0
	Iraq	1	0	100	0	0	0	0	0
	Jordan	2	0	100	0	0	0	0	0
	Saudi Arabia	11	100	0	0	0	0	0	0
	Syria	7	100	0	0	0	0	0	0
	Turkey	4	0	0	33	67	0	0	0
	Yemen	60	100	0	0	0	0	0	0
Far East		15054	34	16	14	33	0	0	3
	Bangladesh	64	90	10	0	0	0	0	0
	China	3671	10	30	10	50	0	0	0
	India	10703	58	27	5	5	0	0	4
	Korea Rep.	2	0	0	0	100	0	0	0
	Korea DPR	50	0	0	31	69	0	0	0
	Myanmar	135	85	15	0	0	0	0	0
	Nepal	238	0	98	1	1	0	0	0
	Pakistan	190	97	1	1	1	0	0	0
	Sri Lanka	6	0	100	0	0	0	0	0
South America		63	0	0	100	0	0	0	0
	Argentina	63	0	0	98	0	0	0	2
Developed countries		1786	1	0	98	1	0	0	0
	Australia	32	0	0	100	0	0	0	0
	Greece	1	0	0	100	0	0	0	0
	Hungary	7	0	0	0	100	0	0	0
	Japan	1	0	0	100	0	0	0	0
	Portugal	10	0	0	0	100	0	0	0
	South Africa	10	100	0	0	0	0	0	0
	Spain	1	0	0	0	100	0	0	0
	United States	180	0	0	100	0	0	0	0
	Yugoslavia	1	0	0	100	0	0	0	0
	CIS	1540	0	0	100	0	0	0	0
World		28377	52	12	14	18	1	0	3

1. Data for Angola not included as the millet figures are combined with those of sorghum.

Source: Accurate figures on millet production and particularly on species-wise composition, are difficult to obtain. These figures are based on official country statistics and FAO estimates for 1981-85 production, revised to obtain the 1992-94 average. The revision was based on relative proportions of the different millets in the 1981-85 production. Updated information on species-wise composition of the 1992-94 production was provided by respondents (usually millet scientists and extension personnel) to ICRISAT questionnaires.