Bulletin 1454 ' December 1983 The Qrigin 0nd Early Cultivetio of Sorg The Texos Agricultural Experiment Station / Neville P. Clarke, Director/ The Texos AGM University System / College Stotion, Texos Contents lntroduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Trade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Climatic Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 The Origins of African Agriculture . . . . . . . . . . . . . . . . . . . . 6 The Origins and Domestication of Sorghum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Race Bicolor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Race Guinea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Race Caudatum . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Race Kafir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Race Durra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Alternate Hypotheses for the Timing and Placement of the Origins of Sorghum Cultivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Literature Cited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Additional Bibliography Dealing with the Origins of Agriculture in Africa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Archeological Evidence . . . . . . . . . . . . . . . . . . . . . 19 Domestication/Origins . . . . . . . . . . . . . . . . . . . . . . 2O Climatalogical/Environmental . . . . . . . . . . . . . . . . 2O Cultural . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2O Botanical/Paleobotanical . . . . . . . . . . . . . . . . . . . . 21 "* The Origin and Early Cultivation of Sorghums in Africa J. A. Mann, C. T. Kimber and F. R. Miller* 1 Q *Slick Fellow, Professor of Geography, Texas AEBM University, and Professor/Sorghum Breeder, The Texas Agricultural Experiment Station 30° 20° I A T L A N TI C O C E A N 10° \ 20° 30° 0 500 1000 miles | 1 . | I I , j prepared by: 0 750 1500 kilometers I C""‘°9'°P"‘“ 0° 10° 20° 30° 40° 50° T°"°s A8“ Major physical features of Africa identified in the text. TUNISIA f, f" MOROCCO" z fm/ t?‘ '- !...._..- -.~.. ._ .. H,’ . . A i. Tropic of Cancer | : _.._.= I MAURITANIA i I J. * " UPPER s u J VOLTA "y; . iAuxeu. (,-'&.--,.-'.r-J : I . v i \ IN," ‘"1 ' 3 ‘- " .- ’ /‘ . \____ IVORY i ._ _. -- -. _ ' ETHIOPIA J - -' , -' / ..~-- - ' - - .- z. A GHANA. = _. (Ab ) \ COAST ,- f‘; i‘ CENTRAL AFRICAN \-- 9'8"“ x - - ' = REPUBLIC cAMsnooNl__ ,--\,____ ,../-- ‘a ‘r1 . k" : ‘J BURUNDI TANZANIA ,'.__/.I'>§'" ; o __ A \_ ZIMBABWE= a?‘ ,. '\..,_ .-’ ! aoTswANA f‘ i f ‘= r» ../~~~-/ .4 ' “d (1 j soum AFRICA I.~.. ./QO\: National boundries and locations of archiological sites mentioned _in the text. ii Foreword Sorghum is a major food and feed crop throughout the world. lt is a species of tropical origin, but in recent history it has been adapted, through selection, to the temperate regions of the world. Sorghum remains the staple food of many countries in Africa and V; Asia. It is now a major feed grain crop in the United States, Mexico, Argentina, Australia, and South Africa. Because of drought resistance found in the species, sorghum is important in marginal rainfall areas of the world. The ability of sorghum to perform well under limited moisture and high temper- ature conditions where other crops often fail makes it an extremely important commodity in providing the necessary food and feed for millions of people in both developing and developed countries. The improvement of any species depends largely upon an accumulated knowledge of the species and the efficient utilization of its genetic diversity. Africa is either the center of origin, or a major site of diversification of much of the sorghum germplasm available in the world. Knowledge of where types originated and how they might have moved within Africa is very important to our efficient utilization of the germplasm in our breeding programs. lf we know more about the history of Africa, and can draw from that knowledge sound inferences about how, where, and when sor- ghum was domesticated and diversified, it will help explain the adaptation of sorghum to specific climatic zones, and what traits various types are likely to possess. Thus, a better knowledge of sorghum history will allow breeders to more accurately forecast what traits might exist in materials presently used, and also to forecast where we might collect materials with particular traits needed in germplasm collections. This publication presents several alternate hypotheses on origin and domestication of sorghum in Africa, in consideration of archeological findings. A new look at ancient events should lead us to new or expanded concepts. This treatise primarily expands on previous theories of the origin of sorghum and suggests that domestication of sorghum may well have occurred much earlier than previously thought. lt is not the purpose of the authors to dispute or discredit previous concepts on the origin of sorghum, but to present in a concise manner the various theories of domes- tication of sorghum. D. T. Rosenow Sorghum Breeder The Texas Agricultural Experiment Station The Origin and Early Cultivation of Sorghums in Africa J. A. Mann, C. T. Kimber and F. R. Miller Introduction fi The origin, domestication and early distribution of sorghum (Sorghum bicolor (L.) Moench) in Africa is of interest to those concerned with the origins of agricul- ture, but it has a special importance to those who deal directly with the crop today as agriculturists. The histori- cal geography of sorghum’s origins gives clues as to why the crop exists in its present form, and also provides valuable insights into where exploitable germplasm of a given type might exist, and what characteristics relate to sorghums adaptation to different ecological zones within the general area of origin. The interaction of early sorghums and man may well have come about far earlier than current evidence sug- gests (56,57). Alternatives will be explored from a human ecology viewpoint based on changing physical environ- ments. Climate The climate of central Africa from the Tropic of Cancer in the north to the South African plateau in the south is largely controlled by the position of the lntertropical Convergence Zone, associated with the shifting of the high sun throughout the year (Figure 1) and the equato- rial trough (35). The tropical marine air mass originating over the Atlantic contributes the moisture for most of Africa as far east as the Rift Valley. The monsoon, another system from the lndian Ocean, influences the areas from the Rift eastwards (79). The relationship of temperature to precipitation in this tropical rain belt is opposite to that in the extreme north and south of the continent. At these extremes, rainfall is inversely cor- related to temperature, so that more rain falls in the cooler seasons. Within the central tropical belt, rain is ‘more likely in the warmer seasons. Because this central belt across Africa is the main area of concern for this publication, a knowledge of long term temperature and rainfall patterns in the area will be useful in looking for "Nimes which may have been propitious for the cultural/ biological events of domestication to occur. The general pattern in northern Africa (south of l5-20 degrees N) has been one of dessication and revegetation over a long period of years (5,17,56). The period from about 20,000 Before Present (B.P.) to 12,400 B.P. was cool and arid. A warming trend about l2,000 years B.P. expanded the tropical rainbelt into higher latitudes to include the Air, Tebesti, Ahaggar and other mountain and highland areas of the north African ‘Fertile Crescent’. This wetter, warmer interpluvial affected all of central Africa from Mauritania to Ethiopia and south to Tanzania (4) for much of the period from 12,000 to 4,500 B.P. The period 8,000 to 7,000 B.P. was drier, however. From 4,500 B.P. until present much of central Africa has been in a cooling and drying phase (35,56). The cycles of cool, dry to warm, wet weather which have characterized this area for so many thousands of years may have played a role in the domestication event(s) of sorghum and other crops (12). During the warm, wet periods, the Sahara, Sahel and Sudanic belts were mostly savanna occupied by plains game, and containing numerous streams and waterfilled depres- sions (4,53). lt is likely that among the evolving flora of these areas were the wild ancestors of today’s African cereal grains. As the weather cycle turned dryer, the true desert and its sahelian fringe would have expanded southward. Within this area, changes in altitude are equivalent to changes in latitude, such that the moun- tains would have dried from the bottom upwards at the same time and in the same way the lowlands dried from the southern desert fringes southward. As the true desert moved both southward and upward, the semi-desert steppe and savanna contracted (11,44,56). This meant a gradual drying of permanent lakes, streams and other features likely to have been food source foci for early hunters and gatherers. The two major rivers, the Nile and the Niger, would have varied in both level and volume as the climate changed. The shrinking of habitable areas might have de- creased food supplies to the point that the inhabitants turned to agriculture and domestication of crops. It is also likely that during some of the following wet periods they abandoned agriculture as it became unnecessary for their sustenance. The kind of variation possible in vegetation due to the aforementioned wet and dry cycles is shown in Figures 2 and 3. Of special interest is the dramatic contraction of Nonh Latitude South 4O 25 2O 15 10 4O l - 10:35 — 10:54 N ~ -11:11 —11:28 11244 l T l l l 12:00 12:16 12:32 12:49 13:06 13:25 Precipitation Total Position of the sun on January 15. Figure 1. The climates of central Africa from the Tropic of Cancer in the north to the South African plateau in the south are controlled by the high sun the savanna and semi-desert steppe regions from the wet situation of 5,000 B.P. (Figure 2) to the dry situation today (Figure 3), and the enormous expansion of true desert. lt should be noted that as livestock herding became more common, environmental degradation due to overstocking may have confounded the effects of climatic change. What in the record looks like climatic change may have been desertification caused by over- grazing. As the expansion and contraction of the savanna and semi-desert steppe regions north of the equator took place, the most advantageous habitats would have also moved north and south, and thus been in slightly differ- ent day length patterns over time. These critical few minutes in day length would have affected the adaptation and distribution of wild plants as they came under selec- tion and domestication pressure. Wild and early domes- ticated sorghums were photoperiod sensitive, and re- quired short days to flower. Migration Other events or processes which may have been important in the origin, dispersal, and domestication of sorghum are movements of people around the continent (migration) and trade routes which might have carried sorghums or ideas about agriculture from one place to another (Figure 4). The archeological records to docu- ment such movements are rare for Africa, especially south of the Sahara. Modern man seems to have replaced Neanderthal Day Length North Latitude South Monthly Precipitation Total '“ .0 - Position of the sun on April 15. position and the location of the Equatorial Trough. Day length, moisture, and temperature provided the framework for cultural adaptation of crops. man some forty to thirty thousand years ago (11,31). Probably well before 9,000 B.P., the present races in Africa had differentiated, with Koisan or Bushmanoid peoples in the south and east, Negroid peoples in the west, north and central parts, and Afro-Mediterranean/ Afro-Asiatic (Caucasoid) peoples along the Mediterra- nean coast and in northeastern Africa (11,63). During the Makalian Wet phase (about 7,000 B.P. to 4,500 B.P.) there was likely considerable movement of peoples north and south across the Sahara and Sahel (9,10) as well as a number of at least semi-permanent habitations in that region (11). Further east there is evidence of Caucasoid pastoralists moving into East Africa from Ethiopia/Sudan as early as 12,000 B.P. (63). More importantly for this discussion, the Cushites, after being driven out of Egypt into the Sudan, may have filtered down into Kenya and Tanzania as early as 5,000 B.P. At some point they possibly introduced eleusine millets (Eleusine spp.) and sorghums into the area, which were later adopted by the Bantu (45,62). Another migration of great importance is that of the Bantu (or proto-Bantu) from the area of eastern Nigeria and northern Cameroon through the forests to Katanga, and later on to East Africa. The proto-Bantu were a Negroid stock closely related to the Niger-Congo peoplew (if not actually of that stock). While originally thought to have been due to the movement south of a war-like lron Age people bent on conquest (78), the accepted view now is that the ‘Bantu Expansion’ was a slow movemeng of an expanding populace into the forest as their technol- ogy permitted and their population size demanded ( 1y Length 13:26 13:07 12:49 12:32 12:16 12:00 -)atitude 11:44 11:28 11:11 10:53 10:34 Monthly Precipitation Total South Position of the sun on July 15. (61,62). The movement of the proto-Bantu may have begun after the introduction of iron tools and Southeast Asian food crops such as banana (Musa sapientum L.), taro (Colosasia esculentum (L.) Shott), yam (Dioscoria spp.), etc. into West Africa about the turn of the Christian era (61). Newer evidence indicates that the forest was penetrated by people using indigenous yams and stone tools much earlier (57). McIntosh argues for acquisition of Southeast Asian crops via the Zambezi valley after about 1,400 A.D. (55). The reason for the Bantu move is difficult to guess, although increased population pres- sures due to a contracting forest/savanna may be one reason (c.f. 17, 85). The route of the migration is dis- puted, but a ‘cultivation belt’ cuts through the humid forest from present Kinshasha to Katanga, indicating a slow migration with concurrent agriculture (34) south through the forest, then southeast along its margin (66). Others speculate the course may have been east along the northern margin, then south just west of the Rift Valley (30). ln any case, the migration seems to have stopped temporarily in the light woodlands of Katanga, a region of similar ecological nature to the homeland they had left (55,61). By 700 to 800 A.D. a Bantu Kingdom had been established there (62), and linguistic evidence indicates that Katanga was the center of origin of the Bantu languages. The food production complex of sorghum and millets (Eleusine, Panicum, Pennisetum and Setaria spp.), the banana and Asian yam, and iron tools set the stage for the rapid increase in population. This resulted in the ”\ expansion of the Bantu to the Atlantic and lndian Ocean coasts, then into the higher rainfall areas of East and central Africa (i.e. the highlands and coastal belt) and finally the drier areas of the savanna (55,61). By the Q 2Q Day Length 11:26 11:33 11:40 11:47 11:54 12:00 12:06 12:13 12:20 12:27 12:34 s i - '5 z 40- — F _ P _. 25- - 20- - <1) ‘O _ B ‘<6 ...J 3 jlf§f§f§f§f§f Precipitation <3 40— 11FI11¢I1I1I= rota| * Position of the sun on October 15. middle of the present millennium, the expansion of the Bantu into eastern, central, southern and southwestern Africa was complete (55). The last migration of concern was the movement of several Caucasoid peoples out of the central Sudan after the Muslem invasions and collapse of the Christian King- dom of Nubia. ln the subsequent unrest, about 1,300 to 1,500 A.D., the Lwoo (now Luo) moved south and occupied areas north and east of Lake Victoria, evidently carrying their sorghums with them (9,16,18,25,61, 62,86). Trade The other process which may have played a role in the dispersal of sorghum is trade, both intercontinental and intra-African (Figure 5). There is no firm archaeological evidence for trade across the Sahara before Roman times. Law, however, describes earlier trade routes from Egypt to the Fezzan then southwest via Tassili and Ahaggar to the Niger (51). A more complete network developed in the first millennium A.D. (58), which in- cluded routes from Sudan/Ethiopia across the Sudanic belt to Lake Chad and on into Niger, Upper Volta and Mali. By 500 A.D. there was trade between areas of central Sudan and East Africa down the Western Rift (52). This may have connected the proto-Bantus with the Nilotes of Sudan. ln the first millenniun A.D., most of the trade in East and northeast Africa was with Arabic, lndian, and lndone- sian traders. They brought in a variety of manufactured goods and other products (including foods and food Day Length TROPICAL RAIN FORESTS SAVANNA MOUNTAINS ‘ DESERT DRY GRASSLAND © THE H. M. GOUSHA COMPANY 5m JOSE Reprmted by perm|ss|on Figure 2. Five thousand years ago the temperature regimes, rainfall patterns, and resultant vegetation were markedly different from today. These different environmental conditions made much of northern Africa a more hospitable environment for sorghum. 4 DPICAL RAIN FORESTS SAVANNA MOUNTAINS DESERT DRY GRASSLAND © THE H. M. GOUSHA COMPANY sm JOSE Reonnted by perrmssnon Figure 3. Drier climate resulted in an expansion of the deserts and contraction of savanna and tropical vegetation (3,000 B.P. to present). Figure 4. Human migrations over a period of 6,500 years had a marked effect on develop- ment of agriculture and utili- plants) and took out gold, ivory, slaves and probably certain indigenous African food crops (2,40). Permanent Arabic settlements were in place in Zanzibar by 107 A.D. (2) and there was extensive trade from Pate south to Kilwa. ln the Red Sea, Egypt dominated trade from the zation of sorghum in Africa. first century A.D. for about 500 years. They were replaced by Persians, who were in turn displaced by Muslem Arabs and lndians by 700 A.D. By this time, then, Arabic influence extended from the Red Sea as far south as Sofala (19), and as far inland as Lake Victoria, Lake Tanganyika, and southwest into Zambia (8). With this trading, sorghum could have and probably did move in and out of the African interior to and from all Indian Ocean states of that era. The Origins of African Agriculture Ideas about the origins of agriculture in Africa have continued to evolve as new archeological evidence has accumulated. Until very recently, no evidence existed to cast doubt on the generally held view that Africa, al- though the oldest peopled continent, had no agriculture until it was introduced from Asia Minor about 7,000 years ago. The accepted view is that the wheat/barley (Triticum/ Hordeum spp.) complex, together with animal husban- dry of sheep and goats, was introduced about this time up the Nile, and that all agricultural developments in Africa arose from that base (1,42,58,71). lt is presumed that this new technology then spread southwards and westwards during the period from 7,000 to 4,500 B.P., first south up the Nile and then into the Sahara during the Makalian wet phase (or Saharan subpluvial) when the desert was hospitable to movement and colonization. During the latter part of this phase, many streamside, mountain and valley sites were occupied by agriculturists (9,42). lt is also usually accepted that during this 2,500- year period cultural contacts, trade, and migration, spread agricultural knowledge westward throughout the Sudanic belt. ¢ ln I959 Murdock advanced the hypothesis of an early center of domestication and agriculture in western Africa near the headwaters of the Niger River (58). He con eluded that Mande-speaking people of this area indepen-¢" dently discovered agriculture, and domesticated some thirty African crops. While not proving domestication, the W Figure 5. Trading ac- tivities utilized between 1000 B.C. and 1500 A.D. are believed to have influ- enced the movement of sorghums across Africa. presence of agricultural/cooking implements lends cre- dence to the idea that a permanent change in diet could have taken place about 7,000 B.P. Other workers (1,42) have given the theory botanical credence, but the latest botanical and anthropological evidence suggests only a few of these thirty such as yam (Dioscorea cayenenis) and oil palm (Elaeis guineensis Jacq) were of West African origin (42,66). Recent evidence from the Nile region of Egypt and northern Sudan suggests that the origin of agriculture and the timing of its movement across Africa may have been during an earlier climatic period than is currently thought (75,81,83,84,85). Wendorf and his associates have, over the past 12 years, reported a series of archeo- logical discoveries on the western side of the Nile which, if they have been interpreted correctly, would fit with the hypothesis of agriculture dating from one of the earlier favorable periods. Their discoveries also bring into question the accepted model of agricultural origins and dispersals. Agriculture, the planting of selected seeds and the harvesting of whole inflorescences of grain, may in fact be originally African, but agriculture as the permanent resource base of established village life may have developed only when the new technology reached Asia Minor (71). The in- troduction of the ‘Asia Minor Agricultural Complex’ into Egypt 7,000 years ago may have been an introduction of revised technology and germplasm rather than the in- troduction of a totally new system. ln addition, Wendorf i and his colleagues found no evidence that the cultivation of wheat and barley was due to the kind of population pressure or declining environment often thought neces- sary for domestication of plants and the agricultural exploitation of the land (82,84). The people of Wadi ‘Kubbaniya planted their crops in the still damp pond beds left by the seasonal flooding of the Nile, much like the ‘decrue’ system used in West Africa today. They then left the ponds, to return only when the crops were ready “for harvest. These small groups of people were appar- ently not pressed for food, as wild game, fowl, and fish appeared to have been plentiful along the fringes of the Nile. Thus, a kind of transient agriculture seems to have developed there without the kind of crisis often deemed necessary for such change (85). On the other hand, because of the hospitable environ- ment in the Sahara and the Sahel during the wet interplu- vial (12,000 to 8,000 B.P.) it seems that this would have been an auspicious time and place for the beginnings of agriculture. These incipient agriculturists would have found wild cereals adapted to the region, and agriculture might have sprung up in the Sahara as much as 10,000 or more years ago. As people moved south to north and east to west in the Saharan region, the basic technology of agriculture and domestication could have spread across sub-Saharan Africa. lt would seem possible and even probable that while Murdock was wrong in his conception about a center of domestication at the head- waters of the Niger, he may have been right that 7,000 years ago there developed an agricultural center which grew a number of crops (including sorghums and mil- lets) currently thought domesticated long after 7,000 B.P. The expansion of true desert southward would have forced the ‘casual’ agriculturist in the African fertile cres- cent southward around 8,000 B.P., and put them into contact with an established hunting and gathering popu- lation along the Niger River. The continued drying of the Sahara might have provided the push for the adoption of agriculture by more sedentary peoples such as the Mande. By the time the Sahara and Sahel became moist again about 7,000 B.P., agriculture (as described by Murdock) may have been well underway. Other hypotheses exist to explain the movement of agriculture across northern Africa. Some have argued for Ethiopia as the center of origin and diversity for several African crops (27,68,80), and many of these authors have concluded these domesticates and agriculture spread directly westward across the savanna to Lake Chad and on to the headwaters of the Niger. The relative lateness of sorghum at Daima (550 A.D.) and Jabel et Tomat (250 A.D.) (14) lead Munson (57) to think this direct movement unlikely or at least too late to account for early west African agriculture. Another possibility is that agriculture and possibly some domesticates moved across North Africa in the so called ‘Saharan Fertile Crescent’. This is a chain of highland plateaus and mountains which stretches from southwestern Egypt to the central Sahara then to within a few hundred kilometers of the great bend of the Niger (44,48,66). The earliest evidence for cereal cultivation in the Sahara comes from the Ahaggar (8,100 to 6,800 B.P.), and other evidence also exists for more recent times (6,49). Because even in today’s dry Sahara, wheat, barley, sorghum, and bulrush millet are grown in these highland areas (60), the possibility exists that the move- ment of agriculture from the Nile to the Niger could have occurred even in the dry periods of prehistory. The agriculture practiced in this area is similar to the ‘decrue’ system (39), and might well have been practiced all over the Sahara during the wet interpluvial times. The spread of agriculture south of the equator is generally considered to be a later development. From the Great Lakes (Albert, Victoria, Tanganyika, and Malawi) westward the movement of agriculture was impeded by the tropical rain forest. During wet periods the forest extended north as far as 13 degrees N, and receeded to 7 or 8 degrees N during the dryer periods (56). ln any case , without crops suited to such areas, penetration of the forest was difficult. Some authorities think that agricul- ture did not penetrate the forest belt until the acquisition of the banana/yam complex from Malaysia (58,66). Oth- ers think that indigenous yam agriculture is much older, perhaps 4,000 to 5,000 years (58,66), but that deep penetration of the forest was possible only after the introduction of iron implements into the agricultural complex about 2,500 B.P. (15). The expansion of the Bantu, about 2,000 B.P., finally accomplished the penet- ration of the forest, and moved an established agricultur- al system from the Cameroon highlands to Katanga, via whatever route they followed. ln East Africa the situation was different because of the lack of a barrier such as the tropical forest. While it would still appear that agriculture was late developing there, as compared to Abyssinia (Ethiopia), Nubia (Sudan), and Egypt, there is evidence of agricultural peoples from Nubia moving into the Rift Valley of Kenya and Tanzania more than 5,000 years ago (27,28). Doggett notes the great similarity in agricultural systems found in the Cen- tral Highlands of Ethiopia and the area near Lake Victoria in Tanzania. A second wave of Caucasoid peoples left Ethiopia before 3,000 B.P. and with them came both animal and plant domesticates. Apparent vestiges of these people can still be found in the Rift Valley of northern Tanzania (58). These Cushitic peoples probably did not move much further south than northern Tanzania, and the spread of agriculture southwards probably waited until the end of the first millennium after Christ. As the Bantu expanded their territory, they spread agriculture through much of Africa from south of the tropical forest to Cape Province. The Origins and Domestication of Sorghum Domestication of plants is a process which has drawn a considerable amount of attention from a number of geographers, historians, archeologists and botanists (1,9,12,42,43,50,70,71,80). ln sorghum, domestication is initiated by allelic changes at only two loci resulting from different selection pressures when harvest tech- niques change. This initial simple change has resulted in dramatic changes in the plants themselves and how they are used. The basic distinction between a wild and domesticated cereal is the presence or absence of an abscission zone at the rachis, panicle, or spikelet nodes (65). The wild type progenitors of all cereals disperse their seeds by the breakage of these nodes and subsequent shattering of the seeds. As long as these wild grasses were collected by threshing the inflorescences while still on the plant, this character was maintained (even reinforced) by this type of selection. Even if these seeds were later sown, no progress was made towards domestication in this narrow sense. Selection for other characters such as seed size or quality or inflorescence size might well have been suc- cessful without completing domestication. The essential” step in the domestication was the harvest of whole inflorescences, and the utilization of this grain for seed. This change in harvest technique makes shattering z ,y strongly negative character, and those mutant types irtJ’ which the rachis, panicle branch, or spikelet node re- mained intact suddenly had the selective advantage. K- Figure 6. Mapping sorghum collections illustrates the distributions of the wild varieties of Sorghum bicolor: var. aethiopicum 9 ; var. arundinaceum O ; var. verticilliflorum I ; var. uirgatum Q . Each dot represents one or more collections. The approximate limit of present day tropical forest is delineated by the dashed line. (From de Wet and Harlan [22]). 6 ounnA ¢- 0000 0.9. > g ' PUNJAB * emu BICOLOR \\" nooomw. > . g i K W c. zooo n. Figure 7. Presumed souce region of domes- ticated sorghum and early diffusion to secondary centers. (From Harlan and Stemler [40]). / .\. ti? l V Figure 8. Later movement of sorghum: C and 1/zC represents Caudatum and half Caudatum types respectively; K, G, and GK represents Kafirs, Guineas and Guinea-Kafir types respectively; DB represents Durra-Bicolor types. (From Harlan and Stemler [40]). Within a very few generations, this recessive character would have beenfixed, and domestication complete. The domestication process is obviously somewhat different for tree crops, root and tuber crops, and bananas. Once domestication has occurred, the next important step may be the inclusion of the domesticate in a stable ‘village agriculture’ setting, so that the results of selection and domestication are not undone by such practices as the use of wild seeds as planting stock (54,71). lt may be that domesticates were not always maintained in the years between domestication and inclusion in a perma- nent agricultural system. However, even under hunting and gathering conditions, it is likely that continuous selection increased genetic diversity. The actual course of diversification depended on the crop, how it was used, the pollination system of the crop, and the amount of diversity present (42,50,71). The hows of domestication are considerably easier to define than where and when the events occurred. Vavilov (80) was perhaps the first scientist to systematically examine, on a worldwide basis, the diversity in crop plants. From the evidence gathered, he proposed that 1O eight centers of origin existed in the world, including the African center, Ethiopia. Sorghum was one of the crops he considered to have its origin there. The Vavilovian concept of centers was modified by his associate Zhukovsky so that these centers included whole conti- nents, with the ‘African center’ including most of sub- saharan Africa. Murdock lists sorghum as one of the crops originating in the Sudanic complex, enobled by Negroid peoples of the Nigritic language group. As previously mentioned, the West African center of origin has been largely discounted. Harlan (36) has given a definitive look at agricultural origins, using archeological, paleobotanical, anthropological as well as botanical evi- dence. He delineates two general patterns of origin, centers and noncenters. Centers of origin apply to such crops (on an individual basis) as Brachiaria deflexa andv@ Oryza glaberrina. These crops have not spread much, and are still grown in their centers of origin. More crops, including sorghum and millets, arose from noncenters a Harlan states that any crop unlimited in time, space an genetic variability is likely to exhibit a broad area of origin where it was probably enobled many times over many years (37). Sorghum fits this definition nicely, and Harlan envisions a noncenter for sorghum stretching from near "\ the Ethiopian border west through Sudan and Chad to near Lake Chad. The diversity and abundance of wild and weedy species, as well as the presence of primitive race bicolor lead Harlan to this conclusion (38). Snowden considered sorghum to have separate cen- ters of origins for different types, with the wild race aethiopicum giving rise to races durra and bicolor, arun- dinaceum to guinea, and verticilliflorum to kafirs, respec- tively (68). De Wet and Huckaby (26) proposed a similar scheme, with durras arising from the race kafir. Doggett (27) says, given the distribution of wild sorghum races in Africa (Figure 6) and the known amount of introgression of genes from wild to cultivated species (and visa versa), the similarities observed by Snowden (between aethiopicum and durras, for example) should be expect- ed because of mutual modification of wild and cultivated forms. ln other words, the diversity seen in the wild form may, to some degree, reflect man’s intervention in the selection of domesticated types. The most widely published view of the origins of sorghum comes from the work done by Harlan, de Wet, Stemler, and others at the University of lllinois. ln a long series of publications (20,21,22,23,24,25,36,37,38,40, 41,70,71,73,74,75,76) the general outline is detailed of what they find to be the most logical sequence of events from the time of sorghum’s first enoblement to its wide Figure 9. Bicolor panicle and three positions of the spikelet and kernel; front, back, and side. Note the long clasping glumes and round but slightly elon- gated seed. distribution over the continent (Figures 7 and 8). The classification scheme used in this publication is that of Harlan and de Wet, (38), in which their cultivated races of sorghum correspond closely with the working groups of Murty et al. (59). Race Bicolor From the distribution of races in the subspecies S. bicolor arundinaceum, de Wet et al. concluded that race vigatum, a desert grass, and race arundinaceum, a forest grass, both exist outside the boundries of primary sor- ghum areas, and as such are unlikely candidates for the progenitors of modern sorghum (24). The race aethiopicum/verticilliflorum complex was domesticated first, giving rise to primitive race bicolor, probably 3,000 to 5,000 years ago. Race bicolor is the most primitive of the five races of subspecies bicolor as deliniated by Harlan and de Wet (38), and now occurs everywhere in Africa that sorghum is grown, and is widely distributed and apparently ancient (24) in Asia from lndia to ln- donesia. The kaoliangs of China were also apparently derived from race bicolor introduced from northwestern lndia. Race bicolor sorghums are characterized by: panicles open and medium sized; long slightly stiff rachis 11 branches terminating with long clasping glumes; glumes thick and coriaceous with obscure nerves; tips of lower glumes depressed and hairy; pedicellate spikelets persis- tent; grain elliptic to subglabose but elongated and gen- erally small in size; grain enclosed or almost enclosed and tightly clasped by the glumes; pedicels short; glumes, seed and plant parts generally highly pigmented; grain persistently attached to panicle; fairly low yielding, medium height but tending to tiller profusely (Figure 9). Race Guinea Harlan and Stemler (40) conclude that with the possi- ble exception of the race kafir, the other cultivated races of sorghum were derived from race bicolor, probably through selection without introgression of wild races. They hypothesize early bicolor ‘moved’ west (presumably through trade or migration) into the savanna west of Lake Chad to the Senegalese coast sometime prior to 3,000 B.P. From the early bicolor the first guineas were selected for adaptation to wet tropical habitats (23). Harlan and Stemler consider the guineas to be the oldest of the 12 Figure 1 0. Guinea panicle and three positions of the spikelet and kernel. Note the long loose and pendu- lous panicles as well as exposed grain in the invo- lute glumes. specialized races because of its wide distribution and diversity (40). Because, they claim, the ecological niches of race arundinaceum and race guinea do not overlap, they discount arundinaceum as a direct ancestor. Snow- den (68) has suggested guineas gained adaptation to wet climates by introgression with arundinaceum, and de Wet, et al. (24) leave it as a possibility. The guineas are now the dominant sorghums of western Africa, grown in areas which receive up to 5,000mm of rainfall. Guinea sorghums were moved into east and southeastern Africa before 2,000 B.P. and were then transported, probably from East African ports, to the Malabar Coast of lndia (23,40). Race guinea sorghums are characterized by: panicles long, loose, glabrous and pendulous at maturity; sessile spikelets opening when mature exposing the grain; glumes involute, opening widely, hairy and awns con- spicuous; pedicellate spikelets both persistent and de- ciduous; grains small to medium, biconvex and nearly¢ ovate but some flattened; grain color light or only slightly pigmented; plants medium to tall and low yielding (Fig- ure 10). w u, Figure 1 1. Caudatum panicle and three posi- tions of the spikelet and kernel. Note the shorter glumes and longer grain which is turtle backed (flat on one side and bulging on the back). Race Caudatum Race caudatum was, according to Stemler et al., prob- ably also selected directly from early race bicolor, be- cause caudatums are for the most part restricted to the area of early bicolor domestication. The distribution of caudatums is closely associated with speakers of the Chari-Nile language group. Caudatums are grown to some extent outside of the areas associated with Chari- Nile speakers, but usually these areas have had contact with traders of Cheri-Nile origin or are occupied by immigrants from the core area. This is the case in Cameroon and Ethiopia (74). The caudatums grown in the Ethiopian lowlands are grown by people whose traditions and farming systems very much resemble their Sudanese neighbors. Stemler et al. (76) conclude that the caudatums of the Ethiopian highlands are intrusive, probably acquired by trade or other such means. They are most common in areas receiving from 250 to 1300mm of rainfall, and are described as having great adaptation to harsh conditions (74). Race caudatum sorghums are characterized by: pani- cles medium large oblong, dense to slightly open, hairy with stoutpeduncle and rigid primary branches; sessile spikelets obovate to elliptical; pedicellate spikelets de- ciduous; glumes coriaceous, shorter than the large grain and all are pubescent; grains are turtle backed (flat on one side and round or bulging on back); grain chalky white or pigmented, most with a recessive spreader gene; plants medium to tall and generally high yielding (Figure 11). Race Kafir According to the construct of Harlan et al., (21,40) race kafir was derived from early bicolor carried south from the north African savanna. Electrophoretic data (67) indicate kafirs to be more closely associated with wild race verticilliflorum than the other four races of cultivated sorghum, so the possibility of independent domestication exists. The kafir sorghums are closely associated with the Bantu speakers of eastern and south- eastern Africa; indeed the word ‘kafir’ comes from the Arabic word for ‘non-believer’. lt is thought that kafirs are a relatively recent race because they did not move to India via the same channels as guinea types. Harlan and Stemler (40) also theorize that more northerly ports in East Africa were used for this trade, and so kafirs were not included. Race kafir sorghums are characterized by: panicles erect, elongate, mostly semicompact and cylindrical; branches and sessile spikelets hairy but glumes at 13 "‘~‘\ u‘: I \\\\‘\“I= "ii/ ' v", ' \\\\“. v‘! (w/ l;l/:/ / \\‘\\\‘t \‘ \ \ §\ “ff: v , "I ~ " w w l \ ‘Y1 | ' ////1.,:,/ \ , .' '2" / \ maturity almost glossy; glumes moderately coriaceous and much shorter than grain; grain is broadly elliptical, sometimes compressed, flattened or biconvex (almost round); plants medium height and high yielding (Figure 12). Race Durra Because of its compact panicle and predominantly white grain, race durra is adapted to areas of low rainfall and low grain mold risk. lt is the most important sor- ghum type in Ethiopia and is restricted in distribution to areas of Africa north of the equator. lt is also an impor- tant type in lndia. According to Harlan and Stemler (40) and Stemler et al. (76), durra sorghums were selected from early bicolors which had been carried to lndia before 3,000 B.P. They were acquired by Arabs from lndia by trade or invasion before the sixth century A.D., and established in Arabia and parts of Africa that the Arabs invaded. During the migration of Arabic Muslems into Ethiopia around 615 A.D., durra sorghums were introduced and these sorghums were the economic base of those Muslem states. Harlan et al. base their hypothe- sis on the fact that durras are primarily grown by Muslem Africans and Arabic peoples in Ethiopia, while they are grown in both lslamized and Hindu areas of lndia. They 14 Figure 12. Kafir pani- cle and three positions of the spikelet and ker- nel. Note the erect mostly semi-compact cylindrical panicle con- taining elliptical and sometimes compressed but almost round grain. indicate that the main growers of durra sorghum in Ethiopia today are the Muslem Oromo (Galla), who settled the fertile warm highlands about 450 years ago, and adopted race durra sorghums as the basis of their agricultural system. Durras are grown today in the mid-altitude highlands of Ethiopia, the Nile Valley of Sudan and Egypt, and in a belt across Africa between 10 and 15 degrees N. latitude from Ethiopia to Mauritania. Race durra sorghums are characterized by: panicles stiff, dense and compact (often very compact), ovate to oblong shaped or rhomboidal and covered with dense velvety pubescence; peduncle often recurved but occa- sionally erect; panicle branches short, semi-erect and hairy; rachis varies from elongate to hidden; pedicellate spikelets large and persistent; sessile spikelets obovate elliptic to rhomboidal with glumes coriaceous on lower half and slightly to strongly depressed with a central transverse wrinkle, the lower half of the glume strongly nerved with tip papery; glumes lightly pigmented; grain generally medium to large, biconvex, broad top and wedge shaped base; plants medium to tall and good” quality (Figure 13). Figure 13. Durra pani- cle and three positions of the spikelet and ker- nel. Note the often re- curved peduncle and mostly rhombidal pani- cle as well as the cen- tral transverse wrinkle on the lower half of the u‘! .,-' ~ . ...\ m‘. , . QIUITIG. Iflfifillfitéi‘ it I“ ' H")! l; “A “MN n": l f”! ||'Ifl‘l‘ ‘W4 - \‘ ‘i’ '-;' flflinfl‘; l 1 l “l ' 4]] -,/ "l", in h 559i} ikflalilllsaigyglwuii;,his , \ l‘==‘s.';,.,=laim|>!{l; I, — ‘- =\\:~ -._:~'_".Ajy~ é 3L“ , N». Q; ‘ u’! ‘we [Blank Page in Orwaal F‘ '1'" ‘,- . ua-v 2.5M—1 2-83