L., I IDRC-080e ARCHIV MACINTedings of the Fourth Symposium of the 23101 Lational Society for Tropical Root Crops

Held at CIAT, Cali, Colombia, 1-7 August 1976

Edited by James Cock, Reginald Maclntyre, and Michael Graham ('

N

The International Society for Tropical Root Crops in collaboration with Centro Internacional de Agricultura Tropical International Development Research Centre United States Agency for International Development IDRC-080e

PROCEEDINGS

of the FOURTH SYMPOSIUM

of the INTERNATIONAL SOCIETY FOR TROPICAL ROOT CROPS

held at CIAT, Cali, Colombia, 1-7 August 1976

Edited by James Cock, Reginald MacIntyre, and Michael Graham

'041

20 IV1977

The International Society for Tropical Root Crop in collaboration with Centro Internacional de Agricultura Tropical International Development Research Centre United States Agency for International Development @ 1977 International Development Research Centre Postal Address: Box 8500, Ottawa, Canada K1G 31-19 Head Office: 60 Queen Street, Ottawa

Cock, J. MacIntyre, R. Graham, M. International Society for Tropical Root Crops CIAT IDRC USAID IDRC-080e Proceedings of the Fourth Symposium of the International Society for Tropical Root Crops held at CIAT, Cali, Colombia, 1-7 August 1976, Ottawa, IDRC, 1977. 277 pp. / IDRC pub CRDI /. Proceedings of a symposium on / root crop / / production / in the / tropical zone /includes / list of participants /, / bibliography/s, and / statistical data /. UDC: 633.4(213) ISBN: 0-88936-115-0

Microfiche Edition $1 CONTENTS

Foreword5 Society Council, 1976-79 6

Welcoming addresses 7

Participants 11

Section 1:Origin, dispersal, and evolution 19 Papers by:Léon 20; Plucknett 36; Sadik 40; Martin 44; Mendoza 50; Kobayashi and Miyazaki 53; Degras 58; and Warid et al. 62 Summary of discussions65

Section 2:Basic productivity69 Papers by:Loomis and Rapoport 70; Holmes and Wilson 84; Ferguson and Gumbs 89; Dharmaputra and de Bruijn 94; Nitis and Suarna 98; Obigbesan et al. 104; Ngongi et al. 107; Howeler et al. 113; Rendle and Kang 117; Mohan Kumar et al. 122; Edwards et al. 124; Wahab 131; Umanah 137; Montaldo and Montilla 142; Montilla et al. 143; Wilson et al. 146; Tanaka and Sekioka 150; and Sykes 151 Summary of discussions152

Section 3:Preharvest and postharvest losses155 Papers by:Lozano and Terry 156; Bock et al. 160; Mukiibi 163; Mukiibi 169; Terry 170; Ninan et al. 173; Leu 175; Terry 179; Obigbesan and Matuluko 185; Bellotti and van Schoonhoven 188; Nyiira 193; Yaseen and Bennett 197; Pillai 202; Thompson et al. 203; and Albuquerque 207 Summary of discussions208

Section 4:Utilization211 Papers by:Christiansen and Thompson 212; McCann 215; Chandra and De Boer 221; Valdes Sanchez 226; Phillips 228; Oke 232; Delange et al. 237; Hew and Hutagalung 242; Khajarern and Khajarern 246; Varghese et al. 250; Hutagalung and Tan 255; Gomez et al. 262; Gregory et al. 267; Nartey 270; Nakayama et al. 274; and Jeffers 275 Summary of discussions277

3 Origin, Evolution, and Early Dispersal of Root and Tuber Crops Jorge Léonl

Tropical root and tuber crops have been domesticated in SoutheastAsia, west-central Africa, and tropical Latin America (including the high ).The crops belong to different families. of the one ()were domesticated independently in each of the three regions. Species of Colocasia and Xanthosoma(family Araceae) and Pachyrrhizus and Pueraria (family Leguminosae)were domesticated in separate regions. Many of these crops have restricted areas of distribution dueto physiological requirements and are becoming relict crops. Roots and tubers are ancient crops, and even todaysupport groups of people who gather them from wild . Poisonous, acrid,or bitter qualities were found in the most important crops by early man, who learned how toremove or destroy these undesirable qualities. Most of the root and tuber crops are polyploids, andmost of them are vegetatively propagated. Fertility traits are therefore of special importance in their evolution under cultivation, but there is no evidence that clonal propagation has ledto sterility. Information on their evolution is extremely scarce as cultural sources, archaeological, linguistic, and historic information is scanty and unevenly distributed. On biologicalsources, compara- tive , metaphase cytology, and hybridization have givensome important clues, but there is still very little information available on the evolution of these species. Root and tuber crops dispersed slowly between Southeast Asia andAfrica. After the seventeenth century a very active interchange occurred especially withthe American crops. Since then, there has been a continuous replacement of crop species, especially in Africa. The sweet , of American origin, was found in Oceania whenthe Europeans arrived, but no satisfactory explanation of how it got to Polynesia hasever been made. Root and tuber crops are thought to be oftion" with cereals. In vegetative propagation ancient origin, and are often regardedas relicsprimitive farmers apply simple husbandry, but of primitive agriculture. This concept is basedthe same is true when they grow seed crops. On on the important role these crops play in exist-the other hand, some vegetative crops, such as ing primitive societies, and on the rudimentarypotatoes, have reached an advanced stage of husbandry they require, particularly vegetativeproduction and technology, comparable to propagation. These crops are easily adapted inmany other crops. The view persists, however, the less-advanced agricultural systems becausethat vegetative propagation represents a low of their high yields, resistance and earliness,stage of progress, and a distinguished cyto- and in the dietary pattern by their bulk andgeneticistinhisinterpretationofhistory taste qualities. (Darlington 1969), points to "the fatal abun- Since the last century, geographers andan-dance of tropical root crops imported from thropologists have contrasted root and tuberAsia and America" as one of the main factors production, which includesotherclonallyin the decline of Africa. propagated crops such as bananas, breadfruit, The contrastingof agriculturalsystems sugarcane, with seed agriculture. Vegetativebased on the differences between seed and propagation, developed in tropical regions, isclonal propagation is a simplification of a assumed to be a static system, whereas seedproblem that is too complex to be reduced to agriculture is associated with the developmentthe duality of planting materials. of more advanced societies. Geographers and historians are tempted to associate "civiliza- The Basic Materials Roots and tubers are storage organs that are 1Programa de Germoplasma, CATIE, Turrial-developed in many families of plants, probably ba, Costa Rica. as a result of selective pressures in environ- 20 LION: ORIGIN, EVOLUTION, AND DISPERSAL 21 ments with varying amounts of rain. Theon crop evolution comparative taxonomy, storage organs permit the accumulation of nu-cytological analysis especially at metaphase, trients elaborated by the aerial parts of the and experimental hybridization have been plant. By growing underground, they maintain applied to some of the root and tuber crops. the nutrients with minimal loss. Once the tem-However, as a whole, the information avail- porary branches or foilage have dried, newable, with the possible exception of potatoes, shoots develop from the storage organs. Byis very poor and scattered. harvesting roots and tubers before the plants Comparative taxonomy aims to establish the have flowered, man has interrupted this pro-relationships among existing taxa with the cul- cess, and has thus kept the plants in a kind oftivated species. The definition of their taxo- permanent juvenile stage. nomic position permits the identification of Although storage organs may vary in theirputative parents and the delimitation, within phylogeny and structure, the nature of theirthe genus, of the cluster of species more storage tissues is common. These organs areclosely associated with the cultivated taxa. The large masses of parenchyma thatcontaintraditional methods of taxonomy do not work mainly water and starch grains. The paren-well with root and tuber species. Often foliage chyma is intimately connected with the vascu- isdifficultto accommodate in herbarium lar system, which permits easy transport fromsheets; flowers are bulky, fleshy, and frequently and to the storage organs. Frequently, there are absent; root and tubers too difficult to preserve. poisonous, bitter, or acrid substances in theIn some genera, like Xanthosoma, the taxo- storage tissues, which present an obstacle to the nomic status is extremely unsatisfactory; the utilization of roots and tubers. However, theseabundant synonymy in the aroids and yams is, substances are a deterrent to animals and there-in part, a result of studies based on herbarium fore play an important role in the survival ofmaterials. As in other crops, the identification plants growing in natural conditions. Theof a wild population closely related to the cul- quantity of these materials (e.g. raphides ortivated species, raises the question of whether it glucosides)variesconsiderably within theis an ancestor of the cultivated type or a feral or same species, a trait that is mainly determinedweedy form. In Ullucus, an aboriginal species by inherent factors. has been described (Brucher 1962), but this The storage organs may be roots or stems. could be only a wild variety. A hexaploid popu- In roots, such as , the storage tissueslation of Ipomoea from Mexico, has been con- may derive from a normal cambium, or as insidered as an ancestor species (Nishiyama the from tertiary cambiums that1971) or a weedy variety of the sweet potato develop around vascular elements. Storage(Martin et al. 1974). stems are of different kinds (e.g. rhizomes, Cytological studiesare interrelated with tubers, and corms), and as in roots, the storagetaxonomy and both have helped to solve some tissues may derive from different types ofproblems. But itis surprising how littleis cambium. In most of the root and tuber crops,known in this field. Of some of the cultivated storage organs are more than carbohydratespecies there is no information even on chro- sinks. They are also reproductive organs, andmosome number. On the other hand, some of this double function has been of foremost im-the root tuber crops offer technical problems portance in their propagation by man. such as chromosome size and high number, The common classification of crops by thewhich make it difficult to detect possible link- utilization of certain organs results in artificialages or identification points. groupings. Thus, temperate species utilized for There is a trend to attribute to interspecific their roots or stems (radishes, beets) are con-hybridization the origin of most cultivated sidered as vegetables, whereas tuber cropsplants, especially if they are polyploids. In yielding essential oils (ginger) are includedroots and tubers, sweet potatoes (Nishiyama among the spices. The present discussion iset al. 1975), African yams (Ayensu and Cour- limited to tropical and subtropical species usedsey 1972), common potatoes (Dodds 1965), mainly for their starch content as energy foodsand others have been assigned such origin, (Table 1). and putative parents have been suggested for some of them. Genome identification lends Factors in Evolution some support to this contention, but it is very The usual sources of biological informationdifficult to obtain definitive evidence from 22 TROPICAL ROOT CROPS SYMPOSIUM

Table 1. Tropical and subtropical species used mainly for their starch content as energy foods.

Monocotyledoneae Agavaceae Cordyline terminalis Ti, palm lily Araceae Alocasia Ape, biga, birah Amorphophallus Suran, elephant Colocasia Taro Cyrtosperma Xanthosoma Cannaceae Canna edulis Achira Cyperaceae Cyperus esculentus Eleocharis tuberosus Water chestnut Dioscorea (I 2 species) Yam Marantaceae Maranta arundinacea Arrowroot Calathea alluia Lairen Taccacaeae Pia, Polynesian arrowroot Zingiberaceae Curcuma angustifolia Curcuma zeodoaria Dicotyledoneae Ullucus tuberosum Ulluco Compositae Polymnia sonchifolia Yacon Convolvulaceae lpomoea batatas Sweet potato Cruciferae Lepidium meyenii Maca Euphorbiaceae Manihot esculenta Cassava Labiatae Plectranthus esculentus Kafir potato, dazo Solenostemon rotundifolius Hausa potato Stachys sieboldii Leguminosae Paclzyrrhizus (spp.) Ahipa Psyplzocarpus tetragonolobus Sigarilla Pueraria lobata Sphenostylis stenocarpa Oxiladaceae Oca Solanum tuberosum Potato Tropaeolaceae Tropaeoleum tuberosum Mashua Umbelliferae Arracacia xanthorrhiza Arracacha

hybridization work. Often the terms "nobiliza-of somatic mutation in the evolution of these tion" or "ennoblement" are applied to root andcrops. However, hybridization in these crops is tuber crops, meaning improvement by primi-a potential force of considerable value for tive agriculturists. These terms imply a plannedfurther improvement (Abraham et al.1964). hybridization,as insugar cane breeding, which is not the case in root and tuber crops. On the other hand, information obtained from Polyploidy hybridization aimed at crop improvement is As expected in cultivated crops, a good num- rather incidental to crop evolution. ber of root and tuber crops have different As a main force in evolution, the impact ofploidy levels. The early cytological studies on recombination is possibly less evident now thanpotatoes led to the separation of the numerous in the past. Under original conditions, root andcultivars included in Solanum tuberosum into tuber species were closer to their primitive al-several species (Juzepczuk1937),or into one lies and hybridization may have been morespecies formed by five groups and two hybrids frequent. But as man moved them to new en-with other species (Dodds1962).The culti- vironments, not only did the possibility forvated diploids have been derived either from further crossing and segregation decrease, butwild diploids, again with no total agreement the mechanisms of seed setting and sexual re-on the identification of the ancestors, or from a production were affected. The practice of vege-primitive complex in which many diploid tative propagation helped to reduce the im-species may have taken part (Ugent1970). portance of segregation and increased the roleThe tetraploids may have arisen through allo- IION: ORIGIN, EVOLUTION, AND DISPERSAL 23 polyploid or doubling of diploids (Simmondsthe second process is more likely to have 1976) ; the triploids as hybrids of tetraploids occurred. Other species showing different levels and diploids; and a pentaploid group fromof ploidy are: Maranta arundinacea, tetra- crosses between a triploid (unreduced gametes) ploid; Canna edulis, triploid; Ullucus tuber- and a tetraploid (Ugent 1970). osus,triploid;Oxalistuberosa,hexaploid; Perhaps the group of root and tuber crops inTropaeoleum tuberosum, hexaploid. On the which polyploid is most complex is Dioscorea,other hand, the cultivated Xanthosoma (X = but here again the studies have been scarce and13) are diploids, 2n = 36 (Plucknett 1976). isolated. In D. alata, a survey of chromosomeCassava, 2n = 36, has been considered as a numbers in Indian cultivars has established adiploid (Magoon 1967) and as a tetraploid series of 2n = 30, 40, 50, 70 for different(Jennings 1976). In the Euphorbiaceae, X = clones that do not show phenotypical differ-9 is found in Manihot and allied genera, and ences (Sharma and Deepesh 1956). Also poly-the presence of three nucleolar chromosomes somaty has been detected in this species, andand some chromosome duplication at pachy- this could be a possible source of new typestene, suggest that cassava is possibly a seg- through vegetative propagation. In D. bulb-mental allotetraploid (Magoon et al. 1969; if era, there are morphological differences inJennings 1976). the cultivated types of Africa and Asia, the There are several questions to consider in former being considered by some authorities as relation to polyploidy. For example, does it a distinct species. Cytological counts tend tohave any significance, as in other crops, on the support such differences. The African cultigens size and quality of the useful parts of the have 2n = 36, 40, 54, 60 with a basic numberplants. The information is meagre, except on lenta, 2n = 40, 90, 100; D. cayenensis 2n = potatoes, and in this crop it is masked by long 40, 60, 80, 100 with a basic number of 20selection, environmental conditions, and crop (Martin 1974). Other polyploids are D. escu-protection. There is no information, for in- lenta, 2n = 40, 90, 100; D. cayennensis 2n -= stance, of any correlation in size of tubers and 140; D. opposita 2n = 40; D. penthaplylla 2nploidy in the series mentioned = 140. However, Coursey (1976) points outabove. that in Dioscorea polyploidy is not restricted to In triploids of Canna edulis, the starch con- cultivated species only. tent is almost three times higher than in the In taro, a general survey of its variability isdiploid, but there is no information on yield badly needed, as it is one of the most ancient(Mukerjee and Khoshoo 1971). On the other of the root and tuber crops and one of the mosthand, induced polyploids in cassava, assuming widely spread. Two basic chromosome num-that it is a diploid, are not superior in yield to bers have been recorded in taro, X = 12 and diploids. In I pomoea, wild hexaploids do not X = 14. Clones with 2n = 24 and 2n = 48show any root thickening, and therefore, the are reported from India, while clones with 2ndomestication character may result from pro- = 28 and 2n = 42 are found from India tocesses other than polyploidy. Japan in one direction, and to Timor, New It is well known that most polyploids have a Caledonia, and in the other. But wide adaptability to new conditions. In culti- east of 180°, that is in most of Polynesia, allvated polyploids, this characteristic aided by clones show 2n = 28 (Yen and Wheeler 1968).cultural practices, especially crop protection, The most interesting example of polyploidyis a key factor in their success and expansion. in root and tuber crops is the sweet potato, a Finally, an important aspect is to consider hexaploid, 2n = 90, which could be derived,the possible relationships between polyploidy as has been proved experimentally, in twoand vegetative propagation. Through the latter, ways: (1) by the multiplication of a diploida not fully balanced allopolyploid, for instance, (I. leucantha), or (2) by the duplication of acould be multiplied and meiotic irregularities triploid resulting from the crossing of a diploidor gene-determined sterility bypassed. Such (I. leucantha) and a tetraploid (I. littoralis).polyploids may spread, through clonal propa- The resulting hexaploid, I. trifida, is consideredgation, in a way that would be difficult through a primitive form of I.batatas (Nishiyamaseed reproduction. 1971). As two of the three genomes in the sweet potato show more homology with each Fertility and Vegetative Propagation other than with the third (Magoon et al. 1970), Vegetative propagation is necessary in most 24 TROPICAL ROOT CROPS SYMPOSIUM root and tuber crops because of their ineffi-seed production in taro (reports from Papua ciency in producing seed, a result of natural andNew Guinea, Raratonga, Philippines) could be cultural factors. Among the natural factors aredue to harvesting practices. In yams, seed set- incompatibility,dichogamy, abnormal seed ting is often limited by harvesting practices, but and seedling development, seed dormancy, andwith some African species the constraints for pests and diseases attacking flowers and seeds.sexual reproduction include a large number of Cultural factors are equally important. First,male clones, imperfect seed, and dormancy man has taken clonal crops to regions whereperiods. In potatoes, seed reproduction and fac- environmental conditions do not favour seedtors conditioning low setting are well known, setting. Second, plants are harvested when thebut even hybrids of induced haploids can set storage organs have reached maturity, whichseeds under special conditions of temperature occurs in most cases well before flower ini-and air moisture (Subramanyan et al. 1972). tiation. Third, man, by copying the same typeAll the Andean tubers produce abundant seeds. (clone) in millions of individuals may haveIn oca (Oxalis tuberosa), however, seeds are increased the possibilities, if a natural traitextremely scarce in field conditions, but if pro- limiting fertility is present, to extend it intotected from wind and frost, which produce large populations. abcission of the inflorescences, seed setting is It has been stated often, especially by non-normal (Alandia 1967). In arracacia, seeds are biologists, that continuous vegetative reproduc-produced the second year but as the rootma- tion leads to sterility. Sauer (1952) stated thattures in 8-14 months, flowers and seeds are cassava reproduction by cuttings has been car-rarely seen. ried out for so long that it has lost completely Vegetative propagation as a cultural practice its ability to set seeds. Vegetative propagationis very important in the evolution of root and in itself cannot lead to sterility although, astuber crops. In modern agriculture, it permits mentioned above,it may increase the fre-the multiplication of superior and uniform ma- quency in a population of a trait favouringterialsin large monoclonal plantings. This sterility, or has an indirect effect as it permitsleads to a continuous replacement of cultivars the cultivation in environments unfavourablewhen superior clones become available, with for seed setting. However, some recent work inconsequent losses in germ plasm. In addition, Nigeria shows that seed production is higher monoclonal plantings may be wiped out in yams coming from seed-producing plantsquickly by diseases or pests. Vegetative propa- than those obtained from tubers of continuousgation materials may become sinks of viruses clonal propagation (Sadik and Obereke 1975). and mycoplasma, and are subject to physio- This could be due to the nature of the sampleslogical degeneration also requiring clonalre- studied, but is worth further study. Someex-placement. Under systems of primitive agricul- perimental work in cassava (Jennings 1963) ture the effect of changes in planting materials showing different degrees of male and femaleis of less importance, as the standard practice sterility in hybrids and backcrosses may beis to plant several clones in thesame plot, partly due to the different materials used and tomixed or separated. Frequently the planting environmental effects. material is the edible part, and in time of As was said before, all root and tuber cropsscarcity or famine the "seed" has to be eaten. set seed. Reluctant clones are found, but ex-This double utilization made it possible for the perimental work shows that it may be possiblePolynesians and possibly Incas to propagateas to promote the formation of viable seed bycrops what they had brought as food for their changing environmental conditions or manipu- long sea journeys. lating the flowering process. In cassava, flower As a whole, vegetative propagation, espe- induction has been obtained by moving plantscially in monoclonal plantings, is an important to higher altitudes (above 1000 m in Java,restriction in increasing variability. In species Costa Rica, etc.) or in areas with specialof imperfect evolution, sexual reproduction is climatic conditions, such as the coast of central a second best choice in their reproductive , where most of the clones set seed regu-system (Martin 1967). larly. In sweet potato, seed setting is reported from many countries (Yen 1974), often out- side its natural area (e.g. the Philippines and The Role of Mutation Papua New Guinea). The scarcity of records of In root and tuber crops, evolution through LÉON: ORIGIN, EVOLUTION, AND DISPERSAL 25 chromosome or gene recombinations is limitedof the leaf up to insertion of the petiole, often by the predominance of vegetative propaga- blotched), or the branching in the corms of tion. Therefore, somatic mutation plays a veryXanthosoma and Colocosia, could also be at- important role. However, any new type hastributed to somatic mutations. Colour changes to be carefully evaluated before determiningdue to mutation were mentioned above in roots if it is a bud mutation or a chance seedling.of the sweet potato and a characteristic feature There is no survey of somatic mutations inof somatic mutation,sector colouring,is root and tuber crops to determine mutationfound in the petioles of many taro and Alocasia rates. In primitive populations of sweet potato,clones. Quite common are chimeras affecting changes in the skin colour of the root fromcolour or structure of the outer layers in tubers white or cream to orange or purple haveand corms. In potatoes and Ullucus, buds been recorded roughly at 1:1000 (Yen 1974).coming out of areas of different colour give In more reduced samples of improved clones,rise to different clones. In potatoes, it is com- mutation rates vary from 0 to 2.9% (Hern-mon to find purple areas in white tubers. In andez et al. 1964). The frequency of mutationUllucus, there is a group of clones with tubers may give an indication of the age of the crop,either completely yellow or magenta or with although it is mainly determined by the num-large areas of both colours. In this crop, as in ber of individuals. Because mutations changepotatoes, russeting is common and types with only certain characters, the original popula-this kind of skin are recognized as different tion could be recognized. The mutants formclones. It is very likely that the ornamental "groups of varieties" (e.g. in taro and othercassava has a similar origin. As mutation rate tuber crops). Within a mutant population notis a function of population size, it is quite likely all the individuals are alike (e.g. potatoes).that in the relict crops so common among root Some may produce subclones that differ in im-and tubers, this force is decreasing its effects portant characters such as yield and resistanceand therefore reducing crop diversity. to disease. Somatic mutation is probably the most im- The same mutation may appear in differentportant factor in the evolution of root and places, and this is one of the reasons for thetuber crops under cultivation. It is probably high number of repetitions in collections. How-easier to find a bud mutation than a chance ever, two mutations phenotypically alike mayseedling. However, gene and chromosome be different in their physiological responses.mutation, although less well-documented and Most mutations do not have any agriculturalwith far less chances of occurring, may have value and in an advanced system of agricul-contributed considerably to the diversity of ture, where uniformity is highly desirable, theythese crops, especially when they were grown are immediately eliminated. But in primitiveunder more natural conditions and seed setting systems of agriculture, farmers like to main-was more frequent. tain as many types as possible, as a kind of agricultural asset. It is likely that the "magic Domestication gardens" of the taro growers of Polynesia are collections of aberrant clones. A similar situa- The process of domestication in root and tion occurs in the mixed agricultural system intuber crops, including potatoes, has not been the Andes. Primitive farmers have learned bywidely studied. Domestication of grain crops experience that some of these new types mayin arid lands has been studied, but the results show special resistance and become an im-are of little value in explaining root and tuber portant resource when the common clones aredomestication. The archeological evidence has wiped out by disease and pests. considerable limitations. Evidence depends on Somatic mutations affecting the whole plantpreserved plant materials, from which we can are known in few cases. Brachytic types aredetermine the structure of the organs and en- found in cassava and sweet potato and the mainvironmental conditions. However, because of difference between wild and cultivated Ullucusthe fleshy tissues of roots and tubers, they are is internode length. easily destroyed by fungi, bacteria, and in- Several instances are known in sweet potatosects. As well, most grow in wet regions and in which a normal branch produces othersthe materials preserved are very scarce and with leaves of different shape (Yen 1974).irregularly distributed. The presence of plant Characters like the "piko" in taro (a deep cutremains in dry areas, like the coast of Peru, 26 TROPICAL ROOT CROPS SYMPOSIUM provides secondary clues often of great im-Polynesians also learned to prepare a taro portance, as in the case of the sweet potato,mash ("poi"), which was more nutritious and but not the fundamental information on orig-easier to keep than the fresh product. The inal places and processes. Even less helpful is extraction of glucosides in cassava required the the indirect evidence derived from tools anddevelopment of special techniques and tools. pottery. Historical evidence is most important The planting and cultivation practices and but, like archeology, it provides an irregulartools are less complicated than those used in picture. Africa has been the foremost meet-preparing the crop for meals. The digging stick ing place of root and tubers, however its earlyof the yam gatherers can be easily transformed written history is very fragmentary. A similarinto a planting tool. In a very simple situation occurs in southeast Asia and tropicalinstrument was developed to cut the upper part America. of the taro to obtain propagation material Plant domestication was initiated to answer (Buck 1964). Evidently early man learned the needs of primitive man for food, clothing,simple cultural practices, such as providing body painting, medicines, and poisons. It couldsupport for yams and piling earth on the base have been carried out in areas where materialsof the plants to supply better and looser soil for domestication were abundant but also tothe growing tubers. The processes of under the pressure of scarcity (Harlan anddomestication of roots and tubers, particularly Zohary 1966). The common concept that plantthe development of practices and instruments domestication led to the establishment of sed-for their preparation as food, were far more entary human communities is open to somedifficult than for fruits or grains. Also in select- questioning.Burkill(1960)suggested thating less poisonous or acrid clones and develop- fishing people, who were rather sedentary,ing cultural practices, primitiveman showed after a period of gathering may have startedsuch ingenuity that Burkill (1951) said they domestication and cultivation of yams. How-must have "already graduated in horticulture." ever it may have been a different process forOut of simple management practices,man each species, and it is extremely importantnotevolved more complex systems of agriculture. to generalize. In southeast Asia, terracing was developed for Some domestication characteristics commontaro cultivation, which eventually was adapted in root and tuber crops are: (1) large size offor rice production. In the Andes, the most edible parts; (2) earliness; (3) low content ofcomplex system was developed to cultivate poisonous or acrid substances; (4) attractivetubers and other crops, including terracing, shape and colour; (5) shallow undergroundfertilization, irrigation, and food storage. This growth; (6) sugar content. Other characteris-development was far superior to any agricul- tics, such as resistance to disease, may come intural system in western Europeor elsewhere in more advanced stages of agriculture. The sizethe world in the fifteenthcentury. of the useful parts, like the corms of aroids, and the shape and colour (e.g. Andean tubers) may have attracted the attention of gatherers. Geographical Origin In the transition from gathering to cultivation, and Early Dispersal the main obstacle was the presence of bitteror Root and tuber crops were domesticated in- poisonous substances in the edible parts. Thedependently in three regions: (1) southeast discovery of techniques to eliminate these sub-Asia and its geographic continuation the stances played a decisive role in domestica-Sunda Islands, Papua New Guinea, Oceania; (2) tion. The processesdifferedaccording toAfrica ; (3) Tropical America. species. One special technique was washing theA few species were domesticated in southern roots or tubers for many hours to removeChina and Japan. The three regionswere active poisonous principles. In the Andes this methodcentres of domestication of animals as well. permitted the use of bitter types of potatoes, Agricultural systems were developed independ- mashuas (Tropaeolum), and oca (Oxalis).ently but they have many common features This also led to the preparation of "chuño,"abecause of similar environments. Prior to the dry mass easier to store and transport than the 1500s there were no exchanges of materials and whole tubers. Roasting or washing methodstechniques between the Old and the New were developed in Polynesia to remove theWorlds. The sweet potato was the only known acrid substances from the stems of aroids. Theroot crop in tropical America and Oceania. LÉON: ORIGIN, EVOLUTION, AND DISPERSAL 27

Southeast Asia Mongoloid, came in successive waves from dif- ferent regions. They mixed together and the The Indo-Malayan region, the land betweenresulting population, to which the names of the Deccan peninsula and the South China Sea, Indo-Malays or Malays have been applied, was could be considered as a primary agriculturalpossibly the group that started plant domestica- hearth because of its domestication of plantstion. Burkill (1953) says it is likely that root and animals, the development of systems ofand tubers were the first domesticated crops in agriculture typical to the region, and the in-this area. vention or changes in the utilization of plant As was stated earlier, the development of materials for food and other uses. The regionmore efficient techniques in food preparation is limited by natural barriers to the north (themay have been as important in domestication Himalayas but in its northeast corner is openas the development of crop husbandry. Yams towards China) and to the west (the Indiangathered in India and Malaya are roasted or Desert). The rest of the region is surroundedcooked. The removal of poisonous substances by water. The larger Sunda Islands and Papuaor acrid materials in yams and aroids is done New Guinea could be considered as a naturalby pounding, washing, and heating. Thus, continuation of the region. The Indian sectionthrough a combination of simple practices and comprises mainly the coastal areas of thetools, it was possible in this region to start an Deccan peninsula. In Indochina, the relief isagricultural system based on the production of quite complex, and is determined mainly bytuber crops. The region is extremely rich in three major mountain ranges, which run from otherfoods,especiallyfruits mangoes, northwest to southeast with narrow valleysdurian, rambutan and others which grow and alluvial plains in between. Its western side,wild in the forests. Fish and wild animals sup- towards the Bay of Bengal, from the foot ofplied the necessary protein. The later domesti- the Himalayas to the tip of Malaysia andcation of rice in the region started a new pat- farther on to Sumatra, Java, and Borneo, is atern of agriculture. "tropical rainy" area (over 2000 mm per year). The most important expansion of root and Towards the centre, there are large areas withtuber crops from the Southeast Asian region "tropical rainy" or "wet and dry" climates withwas eastward, carried out by a mixed group, savannas and open forests. The vegetation isthe proto-Malays, which moved from the con- one of the richest in the world and decreasestinent first to the Great Sunda Islands and then in number of species from west to east. Theto Papua New Guinea, about 3000 years ago. conditions in the region are not favourable forThe first expansion occurred about 2000 years the preservation of plant remains, and arche-ago with the settlement of Polynesia when the ological surveys have been sporadic. Malays from Samoa and Tonga reached the There were five racial groups in this regionMarquesas to the east, Tokelau to the north, belonging to the Negrito stock who, up to thisand the Ellice islands to the northeast. Pre- century, lived only by gathering roots andviously, Micronesia had been settled by other tubers. These were the Andamanese in theimmigrants from the Sunda and Philippine Andaman Islands;the Semang inNorthislands. The second expansion occurred before Malaya; the Kadar and the Chenchu in the500 AD when the Polynesians starting from Western Ghats in India, and the Veddas in Srithe Marquesas reached the extremes of the Lanka (Burkill 1953; Coon 1974). Polynesian triangle(i.e. Hawaii, Easter Is- The gathering practices of these groups giveland, New Zealand). There are no written a picture of how man lived before agriculture.records before 1500, and linguistic and archeo- Even today, small groups of food gatherers inlogical evidence is not strong enough to sup- this region collect yams and aroids which areport the view that in Oceania there was a pre- abundant in the rain forest. Because of their re-and post-Polynesian stage in agriculture. duced numbers and the availability of other Towards the west, the expansion of crops food sources, they have survived in the samewas prevented by the dry and desert areas in areas for many centuries. northeast India. Only one species (Colocasia The Negrito stock settled in the region some esculenta) may have followed a land route, 25 000 years ago, but was pushed eastwards byeither through Syria or by the Sabean Lane to other immigrants and is represented today onlyEgypt. by enclaves. The new arrivals, Australoid and To the northeast, taro and some yams 28 TROPICAL ROOT CROPS SYMPOSIUM

moved into the subtropical areas of China andseems to have originated in India. It has spread from there to Japan. eastward in recent times, but is not found in From the background of roots and tubers,Polynesia (Yen and Wheeler 1968). new crops and techniques were developed in It reached China and the Lower Yangtze Southeast Asia and Malaysia, thus creatinga valley and is mentioned in literature towards large agricultural complex. Ricewas the main100 BC. From China it moved into Japan. The crop on the continent and the large islands,introduction into the Philippines came possibly alone with bananas, sugar cane, breadfruitsthrough the Sunda Islands. and many other minor crops. The oldest arche- The spread of C. esculenta to the west is ological date for rice in India is 4300 BC. Bypoorly documented. It reached Egypt around that time other cereals such as wheat and100 AD, either through Syria (and there is sorghum had been introduced andwere al-some linguistic support for this, Tackholm and ready in cultivation (Rao 1974). Dar 1950), or through the Sabean Lane, since Aroids it is found in Yemen from where it may have originated. From Egypt, it went through North Alocasia macrorrhiza (A. indica). This is aAfrica to Morocco and then to Spain and very primitive crop, possibly domesticated in Portugal. It spread also from Egypt to Italy India (Assam, Bengal), or in Indochina; inand to Cyprus, where it is an old and im- India other species (e.g. A. cucullata)are cul-portant crop. tivated and wild Alocasia is used as food. Its When, where and by whom Southeast Asian large trunk contains a fine starch, but becausecrops were transported to Africa is still open of the high oxalate content, it must be cut andto question. Indians or Indonesians settled baked on hot stones, or boiled. In Java andsouth of Ethiopia around 500 AD, leaving in- Tonga some cultivars are used only for theirstruments and practices, like certain types of leaves. Alocasia spread only to the east to-boats along the coast of Zanj and the lakes, wards Melanesia and Polynesia. It is ofsomeand they brought in their crop plants from importance in Tonga and Samoa and to all ofMalaysia. Madagascar is culturally linked with Micronesia, especially the Marshall Islands,Indonesia, and many of the words for crops and was introduced into Brazil in the lastlike Tacca, coconut, taro, are the same in the century as cattle feed. two areas. The Malay sailors may have reached Amorphophallus campanulatus. This plantthe coast of Africa with the favourable winds is found from India to Polynesia but withnoduringthemonsoon season.Propagation clues as to the area of domestication. It isanmaterial of roots and tuber crops, brought in ancient plant, low-yielding, and difficulttothese trips, may have remained viable for prepare for eating, with the result that it isweeks and very likely were established in being grown less and less. It is cultivatedfromAfrica after many failures. Taro, bananas, India to Malaysia, and in Javaas a backyardgreater and lesser yams, and sugar cane were crop (Sastrapradja 1970). In Polynesia,itadopted by the Bantu people and other tribes grows wild and is occasionally gathered, but ison the continent. Either by the geographic unknown in Micronesia (Barrau 1962). spread of the former ethnic group, or through Colocasia esculenta. This species is founddiffusion intodifferenttribes,these crops wild from India to Southeast Asia, and has reached central Africa and later on west Africa. spread throughout the tropical world and toTaro was already in cultivation in Gambia and the fringes of the temperate regions. San Thomé around 1500 (Mauny 1953). Towards the east, the plant was spread by Taro was taken from west Africa to tropical the Malayans and Polynesians to all the islands America, probably in the early 1500s. How- of Oceania, including Hawaii, Easter Island, ever, it is difficult to establish its arrival because and New Zealand. In this vast area, someearly descriptions confuse it with Xanthosoma. hundreds of clones are known, but there is noBy the end of the 18th century it had spread complete survey of its diversity. From chromo-from the Caribbean to Brazil, and early in this some counts, it has been established that therecentury to the southern coast of the United are two types, 2n = 28 and 2n = 42, with theStates. Again, very little is known of its diver- former the predominant type from India to sityinthisarea.Superiorclones,called Japan and Polynesia. Type 2n = 42 occurs in"dasheen," are recent introductions but the India, New Zealand, and the Philippines andnative Xanthosoma, being more productive LÉON: ORIGIN, EVOLUTION, AND DISPERSAL 29 and resistant, has prevented the expansion ofIndia and . Before the arrival of the Colocasia esculenta in the American tropics.Europeans, it had spread from Southeast Asia Cyrtosperma chamissonis. This aroid wasto the Philippines and into Oceania but not not domesticated on the continent since it is notbeyond Tahiti, and north to China, where it is cultivated in India and Malaysia. Its range ex-mentioned in the literature around 200 and 300 tends from Indonesia to the north side of NewAD. It was taken by the Portuguese, along with Guinea; in Melanesia, the Solomons and ,D. alata, around the Cape to west Africa. By but not in ; in Polynesia, in theselection, superior clones with larger and fewer central part as far as the Marquesas, but nottubers and less thorny stems have been ob- in Hawaii or Southeast Polynesia; throughouttained in Southeast Asia and Oceania. Recent Micronesia, as it grows well in the low atollscollections suggest that thisspecies has a (Barrau 1962). greaterpotentialthanpreviouslyrealized (Martin 1974b). Yams (Dioscorea) Dioscorea bulbifera was independently do- Six species of Dioscorea were domesticatedmesticated in the region, and the Asiatic clones in this region: D. alata, the "greater yam,"show morphological and cytological differ- originated in the area occupied by Burma andences from the African cultivars. It is found China where the rivers Irrawaddy, Salween,from India to north Australia (only wild types and Mekong ran closely and parallel to eachon this continent) and all over Oceania. Asiatic other (Burkill 1951). This is a mountainousclones have recently been introduced to tropi- area with alternate seasons. Two wildspecies,cal America (Martin 1974a). closely allied to D. alata, and many primitive Dioscorea hispida extends from west India, cultivars are found here. These yams growwhere it is sporadically cultivated, to Malaysia large rhizomes deep in the soil to survive theand Papua New Guinea. In Java it is planted as dry season, and this characteristic may havea minor crop. The cultivated types are often as attracted the attention of man since early times.poisonous as the wild plants. The greater yam was taken first to the Sunda Dioscorea nummularia is a relict crop, found Islands, to the east, quite possibly only asfrom the Philippines to Borneo, Celebes and clones with shallow-growing tubers. ManyPapua New Guinea and Tahiti, but is not culti- mutants were concentrated or appearedinvated in Java, Sumatra, or New Caledonia. these islands, differing in tuber shape and size, Dioscorea pentaphylla grows wild in India, and in other characteristics, and this area hassouthern China (to 22°N), Philippines, In- the highest diversity of the species. It spreaddonesia, and all over Oceania; cultivated types also to the Philippines and to all parts ofhave been selected in separate localities in Oceania, including New Zealand. Toward theMalaysia and Oceania. west, it extended to west India, stopped by the Great Indian Desert. It spread toward Africa, Other minor root and tuber crops maybe taking the same route as taro, banana, originating in this region and other Southeast Asian crops. It spread to East Africa and Madagascar, and later to Cordyline terminalis is found wild in South- central and west Africa. In the latter region,east Asia, Australia, and most of Oceania; a however, it did not become important becauseclone with green foliage extensively planted there were already native yams under cultiva-for the fleshy roots that contain levulose tion. An historical expansion took place after(Ezumah1970). Curcuma augustifoliais 1500, when the Portuguese brought it to theplanted in south India as a source of starch, west coast of Africa. It became the main food"East Indian arrowroot," and C. zeodaria, used in the slave ships and was marketed widely asfor the same purpose, is cultivated mainly in "Lisbon yams," especially from San Thomé.north India and Sri Lanka (Kundu 1967). With the slaves, the greater yam arrived quiteTacca leontopetalcides is planted sporadically early, around 1530, in the Caribbean andfrom Southeast Asia, Philippines to eastern Brazil, but in the New World its expansion wasPolynesia as a source of starch, the "Tahiti checked again by the African yams. arrowroot." It was possibly domesticated by Dioscorea esculenta, the "lesser yam," wasPolynesians, who developed several ways of possibly domesticated in the same area as D.preparing it for food or starch, but it is now alata. Wild types have been reported fromlosing importance. It was taken by Malays to 30 TROPICAL ROOT CROPS SYMPOSIUM Madagascar but it is likely that the Africanthe Malays or Indians. Among the root and cultivars found from west Africa to Ethiopiatubers, Plectranthus tuberosus was introduced may have an independent domestication. It hasto India and Indonesia. Perhaps the two cul- not improved as has the Malaysian crop.tivated species of Psophocarpus, if theyare Pueraria thunbergiana (P. lobata) is cultivatedAfrican, followed the same route. in the highlands of Papua New Guinea for its From tropical America the introductions are fleshy roots. Psophocarpus tetragonalobus ismore recent and important: cassava, sweet cultivated from India to Polynesia, especiallypotato Xanthosoma, and potatoes have dras- in Burma, for the fleshy, sweetish roots. Itstically changed the agricultural systems and origin is unknown, though it could be tracedtofood habits in west and east Africa. The Africa. The recent interest in this crop is due to Guinean yams were taken very early to the the protein value of the seeds. It is widelyAntilles and the coast of Brazil, during the planted as a vegetable for its green pods. slave trade, and are now the most important Root and tuber crops in the Far East yams in the region. Several species have been domesticated in Yams China or Japan but they have not spread much The main contribution of Africa in root and outside this region. In the last decades intuber crops is the domestication of the Guinean China, both native and introduced root andyams Dioscorea cayenensis and D. rotundata. tuber crops have had a large expansion in areaThese have been considered one species with and production. These far eastern species in-D. rotundata as a subspecies of D. cayenensis, clude: Dioscorea opposita, the "China yam,"but the recent trend is to keep them apart, is possibly derived from D. japonica; A mor-partiallybasedonanatomicalcharacters phophallus rivieri, or "bonjac," is found in (Coursey 1967).The Guinean yams are Japan, China, and possibly Vietnam. Its originespecially important in West Africa, and were is possibly south China, with quite a complexprobably domesticated in this area 5000 BP. utilizationin Japan; Eleocharis dulcis,the(Coursey1976;Ayensu and Coursey1972). "water chestnut," is assumed to be the culti- Dioscorea cayenensis, the "yellow yam," vated form of E. tuberosa, a wild species widely which is less important than the white yam, isa distributed in the Asiatic tropics; Sagittariapolyploid of unknown origin, although it may sagittifolia is cultivated mainly in China for itsderive from D. minutiflora or other closelyre- tubers, and was introduced by the Chinese tolated species. It grows wild throughout West Polynesia and Stachys sieboldii, which isre-Africa, and was probably domesticated in the ferred to in the Chinese literature of the four-Guinea coastal area, spreading through the teenth century, is also grown in Japan. It wasGuinean region in areas of high rainfall. Thorn- introduced into Europe at the end of the nine-less clones have been introduced into tropical teenth century and became quite popular inAmerica. France. Dioscorea rotundata, the "white yam," is supposed to be a hybrid between D. cayenensis Africa and D. praehensilis (Ayensu and Coursey 1972).It is more widely adapted to moisture In Africa, more than in any other continent,conditions, and its cultivation is most intense we see a full range in the gradual transitionin Nigeria. It has spread from Senegal to south in utilization of root and tuber crops, fromEthiopia, including parts of the savanna area, gathering of wild materials to well-establishedand to Uganda, Angola, and Northern Rho- cropping systems. The African root and tuberdesia. It is also cultivated in the Comores and crops were domesticated south of the Sahara,Madagascar. The primitive types have thorny some in the savanna region, others in theroots which provide good protection,but Guinean forest. Africa is also the meeting pointthrough selection of mutants, thornless clones for Asian and American roots and tubers, andhave been established and were taken to tropi- nowhere else has there been such drastic re-cal America during the slave trade where it placements in these crops. Among the Asianhas become the most importantyam. roots and tubers, taro (Colocasia esculenta) The following yams are of lesser importance. was the first. Dioscorea bulbifera, which was mentioned be- Several African crops were taken to Asia byfore, is found in Southeast Asia and Polynesia. LÉON: ORIGIN, EVOLUTION, AND DISPERSAL 31 There are differences between the Asian andfound from tropical Africa to Papua New the African clones, the latter being less ad-Guinea. vanced in their domestication. Some clones do Sphenostylis stenocarpa, one of the "yam not produce underground tubers.It growsbeans,"ofAfricanoriginpossiblyfrom roughly between 10°N and to 10°S lat through-Ethiopia, is cultivated in east Africa and the out west Africa and from the Nile valley nearGuinea area for its spindle-shaped tubers and the Ethiopian border to Southern Rhodesia.dry seeds. The clones in tropical America seem to belong to the African group. Dioscorea dumetorum is Tacca leontopetaloides cultivated particularly in the border of the yel- (T. involucrata, T. pinnatifida) low yam plots, and is found wild through This species is found wild from Senegal to Africa from 15°N to 15°S (Ayensu and Cour-East Africa and also in Southeast Asia and sey 1972). Other species cultivated are D.Oceania. In Africa it is seldom cultivated. The abyssinica, D. colocasiifolia, D. hirtiflora, D.tubers require careful preparation to remove praehensilis, D. quartininiana, and D. san-the toxic principles; in some places they are sibarensis. In Madagascar a number of localused as a source of starch. species (D. antaly, D. ovinala, D. soso, the latter with very sweet tubers) have been domesticatedTropical America but their cultivation is being reduced by cas- sava and sweet potato. The three most important root and tuber crops potatoes,cassava,sweet potato The African tuberous Labiatae come from Tropical America. Other crops,like Xanthosoma, are of high potential value and In the mint family (Labiatae) the formationseveral minor crops offer limited possibilities of tubers is not uncommon, as was mentioneddue to their physiological requirements. under Stachys sieboldii. Two African species The American root and tuber crops have have been domesticated. One was taken sometwo main areas of domestication the high centuries ago to India and Indonesia, and hasAndes and the lowlands in northern South become a regular crop in these countriesAmerica, and a secondary area, Middle Amer- (Chevalier 1905). ica. The Andean crops include potatoes, oca Plectranthus esculentus (Coleus dazo, C.(Oxalis tuberosa), mashua (Tropaeolum tu- esculentus, C. langouassiensis, C. floribundus),berosum), ulluco (Ullucus tuberosum), and the "Kafir potato" is native to west and centralmaca (Le pidium meyeni) that weredomesti- Africa, though its origin and variability are un-cated in the PeruvianBolivian altiplano, above known. The plant produces many elongated300 m, and two crops native to the northern tubers, arising from the central stem. section of the Andes (i.e. Colombia), at lower Solenostemon rotundifolius (Coleus dysin-elevations (1000-2500 m), arracacha (Arra- tericus, C. rotundifolius, C. cop pini, Plectran-cacia xanthorrhiza)and yacon (Polymnia thus tuberosus, P. ternatus), the "Hausa po-sonchifolia). In the Andes, other root and tato" is cultivated in west and central Africa totuber crops were early introduced from lower Transvaal and Madagascar. The plant pro-areas and became an important part of the duces spheric to ovoid tubers, dark red andagricultural complex: sweet potato,achira white.Itisan old introduction to India(Canna .edulis), ahipa (Pachyrrhizus sp). ("koorkan"), Malaysia, Indonesia, and recently The second source of root and tuber crops to the Philippines. is the lowlands of northern in- cluding the Antilles, an area of very imprecise Tuberous Leguminosae limits, in which Xanthosoma spp., arrowroot Two species of possible African origin,(Maranta arundinacea), lairen (Calathea al- Psophocarpus palustris and P. tetragonolobus,louia), Pachyrrhizus tuberosus, and possibly are cultivated widely in tropical Africa and cassava and sweet potato were domesticated. Asia for their fleshy, sweet roots and for theMiddle America, which is so rich in native green pods. The second species, by far thecrops, has contributed few and unimportant most important,isintensively cultivated incrops: jicama (Pachyrrizus erosus) is the most Burma for its tuberous roots. The origin ofoutstanding and has spread to Asia and these speciesis unknown. Psophocarpus isOceania. In pre-Colombian times and even to- 32 TROPICAL ROOT CROPS SYMPOSIUM day in remote localities, some plants areMayan agriculture is still open to discussion gathered for their tubers or corms: Bomarea(Bronson 1966; Cowgill 1971), but it is un- edulis, Dalechampia spp., some beans (Phase-likely that cassava could have been a very olus), Tigridia pavonia, and Dahlia spp. Theimportant food source in the conditions of last two species are now grown as ornamentalsYucatan and Petén. All this may point to a but were once used for their corms and fleshySouth American domestication and the fact roots respectively. Sechium edule is planted forthat its spread towards the north was restricted its fruit though it also yields edible roots. Theto the sweet varieties. On the other hand, early introduction of cassava and sweet potatoHumboldt suggested that the sweet types may to Middle America probably prevented thehave been domesticated first and that later on domestication and expansion of the local tuberman learned how to utilize the bitter varieties. crops. Cassava was introduced early to Africa by the Portuguese. The first published report by Cassava Barré and Thevet is dated 1558 (Mauny 1953). Cassava (Manihot esculenta), known onlyFurther spread inside Africa was determined under cultivation, is a complex of clones show-by its adoption first as a vegetable and later as ing the widest morphological diversity in the a flour source in the Kingdom of Congo, which ParaguaySouthBrazilarea.Clustersofwas an advanced state that influenced the rest closely-related species to cassava are located inof tropical Africa. The spread apparently was both North and South America (Rogers andrather slow, but was favoured by the resistance Fleming 1973), but no wild species have beenof cassava to locusts (Jones 1959). suggested as a possible ancestor. The time and Cassava was introduced to India and South- place of domestication are unknown. The mosteast Asia late in the nineteenth century. important trait for the use of cassava as a food is the HCN content in the roots, which has a Sweet potato wide range from high (bitter ) to very Sweet potato (Ipomoea batatas) was the low (sweet cassavas). There is a clear correla-only food crop common to Tropical America tion in the geographic distribution of the twoand Polynesia before the Discovery. As such, it kinds: sweet cassavas occur in the western sidehas raised a long discussion on which of the of South America, Central AmericaMexico,two regions is its place of origin and on how while bitter clones are planted mainly in theits early dispersal occurred (Yen 1974). The eastern side of South America and the Antilles,recent discovery in coastal Peru of sweet potato with an overlapping area in between (Renvoizetubers dating from 10 000 BP (Engel 1970) 1972). Archeological evidence is very scarce.settles the question of the origin, as this date by The remains of cassava leaves have been iden-far antedates any agricultural development in tified in caves in Mexico dated 2500 BP, andPolynesia. However, it should be considered tubers in coastal Peru from about the same age.that, like all other plants cultivated in the Indirect evidence, such as the presence in earlycoastal region of Peru, sweet potato was in- times of grinding stones in Colombia andtroduced from elsewhere, possibly from the assumed to be used for grindingnorth, the coastal area of Ecuador and Colom- cassava roots, is not very convincing. It is alsobia, where close wild types have been found assumed that cassava flour was an important(Martin et al. 1974), or from across the Andes, article of commerce in northern South Americalike Canna edulis and othercrops. in the second and third millenia BC (Jennings At the arrival of the Europeans, the sweet 1976). What is clear is that cassava was morepotato was known in all Tropical America, intensively used in South America than inwith an important area of diversity around the Middle America. In the former area,theCaribbean. Oviedo, writing in 1530, reports artefacts for the preparation of flour were farthat several varieties he had seen in the early more developed, and other uses, such as thedays of the Conquest were already disappear- utilization of leaves as vegetables or for the ing. preparation of sauces, are typical of South The spread of the sweet potato to the Old America. Archeological information, such asWorld was quite rapid; it was introduced in representations in ceramics and early historicalSpain, after several failures, as living plants information, gives additional support to a morebefore 1550. It is not known how it reached intensive use in South America. Its role in theAfrica, whether from Spain or from tropical LÉON: ORIGIN, EVOLUTION, AND DISPERSAL 33

America. A report that sweet potato was grown In spite of its importance, very littleis in San Thomé in 1520 seems doubtful (Maunyknown on the domestication and early dis- 1953). More reliable information shows thatpersal of the cultivated potato, but the com- it was widely cultivated by the end of the seven-plexity of its structure as a species shows that it teenth century in West Africa, and a centuryhas a long history. The oldest tubers are dated later all over the tropical areas of the con-200 BP (Ochoa, personal communication) and tinent. potatoes are represented in ceramics of the The introduction to Polynesia, as discussedthird century BC. Although very little is known above, has not been properly explained. Itof the domestication process, the early spread could have been accidentally transported into Europe and other continents is fairly well one of the Peruvian rafts lost in the Centraldocumented (Dodds 1966; Hawkes 1967). Pacific, which reached Polynesia where the Other root and tuber crops of the highlands crop was established by Indo-Americans andare: Oxalis tuberosa, or "Oca", of which no developed later on by Polynesians. It has beenwild ancestors are known. It has a large num- proposed also that the sweet potato may haveber of clones differing in size, colour, and shape been taken to Polynesia by one of the Spanishof the tuber, plant size, foliage colour, and expeditions that visited the area starting fromheterostyly. Clones with bitter tubers are used Peru in the sixteenth century. It was taken toto prepare "chuño." The oca was introduced China in 1594 and after a famine in Fukien, itinto Mexico during colonial times ("papa ex- later became an important crop. Sweet potatotranjera"),into southern France, and last was introduced early to Japan from Okinawacentury to New Zealand where itis called and cultivated and adopted in the southern"yam." Ullucus tuberosus has slimy tubers region up to 35°N. which are not as attractive as ocas, but they are consumed even in the large towns. Wild or Xanthosoma ancestor types grow in the highlands of Peru The identification of the species of Xantho-and . Two main groups of clones are soma, cultivated for the corms, is still not clear. known: in the northern extreme of the range The "species" described vary between them-(Colombia), with trailing branches and large, selves like the clones of taro which is now con-red tubers; and in Peru and Bolivia, erect, short sidered to be one species. The genus is foundbranches,with multi-colouredtubers. The from Mexico toBrazil, but the cultivatedulluco was introduced into southern Europe "species" are centred around the Caribbean.but it is not planted. There is no information on the evolution ofgrows in the same area as oca. No wild rela- this crop. It is superior to taro in yield, re-tives are known, although some other South sistance to disease, adaptability, and taste, andAmerican species are reported to form large therefore it is not surprising that this species istubers. Two main groups of clones are known: replacing taro all over the tropics. Xanthosomain Colombia, tubers are slim, white, with deep was introduced to Africa by the middle of theeyes from which emerge fine rootlets; and in last century, where the replacement is mostPeru and Bolivia, the tubers do not have root- active. lets and the predominant colours are yellow with purple lines or fine points. The mashua Tubers of the high Andes grows often ataltitudes where only bitter A group of tuber crops was domesticatedpotatoes are produced. Lepidium meyeni pro- in the high Andes, above 3000 m, where theyduces a radish-like root, sweet, yellow or dark are now intensely cultivated (Léon 1964). Thepurple in the highlands of Peru, above 400 m. most important are the potatoes, which at pres-It is a relict crop that is rapidly disappearing. ent are considered as one species, Solanum tuberosum, including: (i) two tetraploid groups, Other Andean root and tuber crops Tuberosum and Andigena;(ii)a triploid, At lower elevations in the Andes, between Chaucha;(iii)two diploids, Phureja and0 and 3000 m, several root and tuber crops Stenotomum; (iv) two intersperific hybrids, xare grown: A rracacia xanthorrhiza. "Arra- juzepozukii, triploid, a cross between S. tuber-cacha" is especially important in Venezuela and osum X S. acaule, and x curtilobum, resultingColombia; no wild types are known and it from S. tuberosum x juzepozukii (Ugentseems to be of ancient cultivation. Several 1970). clones are known to differ in shape and size of 34 TROPICAL ROOT CROPS SYMPOSIUM the roots, foliage, colour, etc. It has been in-tuber crops, the roles of polyploidy, mutation, troduced into Middle America, Brazil, Eastand vegetative propagation give some clues to Africa, India; Canna edulis is possibly nativethe general process. Some outstanding research to the eastern part of the Andes, and wasin sweet potatoes has established the possible brought to the coast of Peru where it has beentransition stage of the wild diploids to a com- cultivated since 4000 BP; diploid and triploid plex hexaploid, and has recognized wild popu- types are known (Mukherjee and Khoshoolations which may have had a role in the de- 1971);cultivated in Australia ("Queenslandvelopment of the crop. But in sweet potato, and arrowroot"), Hawaii, India, Polynesia; Po- even in the common potato, there is an appall- lymnia sonchifolia is cultivated from Vene-ing ignorance as to how domestication traits zuela to Argentina, and according to Bukasovoccurred. Today there are increasing doubts (1930)grows wild in Colombia. The tuberous among anthropologists and botanists about the roots are fleshy, contain sugar (10% ), andcapacity of primitive man to have carried out were used in colonial times on long sea trips;crop selection. Primitive farmers profited from introduced to southern Europe as a foragethe presence of edible organs in certain species. crop; Pachyrrizus sp. are ancient crops in theHowever, it is quite difficult to explain how he Andes, probably introduced from the Ama-could carry on the improvement of inherent zonian lowlands. traits without knowledge of the rules of ge- netics and the help of permanent records. On Minor tropical American the other hand, primitive man exchanged root and tuber crops planting materials with his neighbours, and in The West Indian yam (Dioscorea trifida) isvegetative crops he introduced superior clones the only species of the genus domesticated inand thus eventually contributed to their hybrid- the American tropics, although other species ization. At the same time, by moving crops to are gathered, particularly in Brazil. The speciesnew areas, he restricted the possibilities of has the highest diversity in the area between thefurther crossing with the species of its native Guianas and Brazil and selected types arehabitat. Man-made isolation has been, there- planted in the Antilles. Its domestication was,fore, an important factor in the evolution of of course, independent of the Asiatic andcrop species. Man has also executed an im- African yams (Alexander and Coursey1969). portant action in taking plants to new habitats. Arrowroot(Maranta arundinacea)was, This is seen more clearly in relation to thern around the middle of the eighteenth century,selective impact of diseases and pests, such as used mainly to cure the wounds from poisonedthe attack of cassava viruses in Africa. arrows, and also started to be used in the Roots and tubers are ancient crops and Antilles as a source of starch (the "St. Vincentprimitive man independently domesticated arrowroot" Stutervant1969). species of the same genus (Dioscorea) or of Lairen (Calathea allouia) was intensivelyallied species(Colocasia and Xanthosoma, cultivated at the time of the Discovery in thePachyrrhyzus and Pueraria) in different parts Great Antilles and the Continent. The ovoidof the world and at different times. Many of tubers are now used for food mainly in Vene-these crops, due to physiological constraints or zuela. lack of acceptability, have not spread farther Several species of jicamas(Pachyrrhizusspp.) than their native habitat, and in fact many are cultivated in South America, and are par-are now becoming relicts in old agricultural ticularly important in Mexico and Centralsystems. Their spread, particularly after the America. From Mexico the plants went, by theDiscovery, has led to competition among them- AcapulcoManila connection, to the Philip-selves and to the eventual replacement of some pines and subsequently spread to Southeastspecies, a process that continues today. Asia and Oceania. Plants of the same "species" Root and tuber crops are associated with offer such a variability in size, shape of leaves,primitive systems of agriculture. A duality has and tuberous roots that the specific limits arebeen established between seedagriculture, difficult to recognize. which is supposed to be a dynamic process characteristic of advanced communities, and vegetative propagation which is supposedly Conclusion maintained by more primitive communities. In considering the evolution of root andThe fact is that in the tropics there is no such LtON: ORIGIN, EVOLUTION, AND DISPERSAL 35 difference, and root and tuber crops as the onlyCowgill, U. M. Some comments on Manihot source of energy food are found only in a few subsistence and the ancient Maya. Southwest isolated communities. Grains and tubers are J. Anthr. 27, 1971, 51-63. integral parts of most agricultural systems inDarlington, C. D. The evolution of man and society. Allen and Unwin. London, 1969. which they have different not antagonistic Dodds, S. K. Classification of cultivated potatoes. roles. Perhaps the best answer to the academic In Correl, D. S., ed., The potato and its wild problem of seed versus vegetative culture lies relatives. Renner, Texas, 1962. in the ceramics of the Trujillo valley in North- The history and relationships of cultivated ern Peru. On one of these ceramics dating from potatoes. In Hutchinson, Sir J., ed., Essays on around 1500 BP an Indian farmer is shown crop evolution. Cambridge, 123-141, 1965. holding, at the same level, a corn plant in one La evoluciofi de la patata cultivada. Endea- hand and a cassava plant in the other. vour 25, 1966, 83-88. Engel, F. Exploration of the Chilca canyon. Curr. Anthr. 11, 1970, 55-58. Abraham, A., Panicker, P. K. S., and Mathew, Ezumah, H. Miscellaneous tuberous crops of P. M. Polyploidy in relation to breeding in tuber Hawaii. Tropical root and tuber crops tomor- crops. J. Ind. Bot. Soc. 43, 1964, 278-282. row. 1, 1970, 166-171. Alandia, S. Producciofi de semilla sexual en oca. Harlan, J. R., and Zohary, D. Distribution of wild Sayafia, 5, 1967, 12-15. wheats and barley. Science, 153, 1966, 1074 Alexander, J., and Coursey, D. G. The origins of 1080. yam cultivation. In Ucko, P. J., and Dibley, Hawkes, J. G. The history of the potato. J. Roy. G. W., eds., The domestication and exploitation Hort. Soc. 42, 1967, 207-224; 249-365. of plants and animals. 405-425, 1969. Hernández, T. P. et al. Frequency of somatic Ayensu, E. S., and Coursey, D. G. Guinea yams. mutations in several sweet potatovarieties. Econ. Bot. 26, 1972, 301-318. Proc. Amer. Soc. Hort. Sc. 85, 1964, 430-433. Barrau, J. Les plantes alimentaires de l'Oceanie. Jennings, D. L. Variation in pollen and ovule Ann. Mus. Com. Marseille Sér. F. 1955-61. fertility of cassava and the eflect of interspecific 1962. crossing on fertility. Euphytica, 12, 1963, 69 Bronson, B. Roots and the subsistence of ancient 76. Mayas. Southwest. J. Anthr. 22, 1966, 251-279. Cassava, Manihot esculenta (Euphorbiaceae). Brucher, H. Ullucus aborigen eous, spec. nov. die In Simmonds, N. W., ed., Evolution of crop wildformeiner andinenkulturpflanze.Ber. plants. Longman, London, 81-84, 1976. Deutsch. Bot. Ges. 80, 1967, 376-381. Jones, W. O. Manioc in Africa. Stanford, Calif., Buck, P. H. (Te Rangi Hiroa). Arts and crafts 1959. of Hawaii. I. Food. B. P. Bishop Museum Spec. Juzepczuk, S. M. New species of the genus Publ. 45. Honolulu, 1964. Solanum in the group Tuberarium. Bull. Acc. Bukasov, S. M. The cultivated plants of Mexico, Sc. USSR, 2, 1937, 295-331. Guatemala and Colombia. Bull. Appl. Bot. Gen. Kundu, B. C. Some edible rhyzomatous and tuber- Pl. Br. Supl. 47, 1930. ous crops of India. Proc. Intern. Symp. Trop. Burkill, I. H. The rise and decline of the greater Root Crops, 1, 1967, 124-130. yam in the service of man. Adv. Sci. Lond. 7,Leon, J. Plantas alimenticias andinas. IICA Bol. 1951, 443-448. Tec. 6. Lima, 1964. Habits of man and the origins of the culti- Magoon, M. L. Recent trends in cassava breeding vated plants in the Old World. Proc. Linn. Soc. in India. Proc. Intern. Symp. Root Crops. 1, Lond. 164, 1953, 12-42. 1967, 100-113. Organography and evolution of the Diosco-Magoon, M. L. et al. Cytomorphological studies reaceae, the family of yams. J. Linn. Soc. (Bot.) on induced polyploids of cassava. Gen. Iber. 21, 56, 1960, 319-412. 1969, 1-47. Chevalier, A. Les vegetaux utiles de l'AfriqueMagoon, M. L., Krishnan, R., and Vijaya, K. tropicale.I. Coleus. Depot des Publications, Cytological evidence on the origin of the sweet Paris, 1905. potato. Theor. Appl. Gen. 40, 1970, 360-366. Clark, J. D. The spread of food production inMartin, F. W. The sterility-incompatibility com- sub-Saharan Africa. J. Afric. Hist.3,1962, plex of the sweet potato. Proc. Intern. Symp. 211-228. Trop. Root Crops, 1, 1967, 3-15. Coon, C. S. The hunting peoples. Penguin, London, Tropical yams and their potential. Part I. 1974. Dioscorea esculenta. USDA. Agrie. Handb. No. Coursey, D. G. Yams. Longman, London, 1967. 457, 1974a. Yams, Dioscoreaspp.(Dioscoreaceae).In Tropical yams and their potential. Part 2. Simmonds, N. W., ed., Evolution of crop plants. Dioscorea bulbif era. USDA. Agric. Handb. No. 70-74, 1976. 466, 1974b. 36 TROPICAL ROOT CROPS SYMPOSIUM Martin, F. W., Jones, A., and Ruben& R. M. A improvement of yams, Dioscorea rotundata. wild Ipomoea species closely related to the Nature, 254, 1975, 134-135. sweet potato. Ec. Bot. 28, 1974, 287-292. Sastrapradja, S. Inventory, evaluation and mainte- Mauny, R. Notes historiques autour des princi- nance of the genetic stocks at Bogor. Trop. pales plantes cultivées d'Afrique occidentale. Root and Tuber Crops Tomorrow, 2,1970, Bull. Inst. Franc. Afrique Noir 15, 1953, 684 87-89. 730. Sauer, C. O. Agricultural origins and dispersals. Mukerjee, I., and Khoshoo, T. N. V. Genetic The American Geogr. Society, New York, 1952. evolutionary studies in starch yielding Canna Sharma, A. K., and de Deepesh, N. Polyploidy in edulis. Gen. Iber. 23, 1971, 35-42. Dioscorea. Genetica, 28, 1956, 112-120. Nishiyama, I. Evolution and domestication of the Simmonds, N. W. Potatoes, Solanum tuberosum sweet potato. Bot. Mag. Tokyo, 84, 1971, 377 (Solanaceae). In Simmonds, N. W., ed., Evolu- 387. tion of crop plants. Longmans, London, 279 283, 1976. Nishiyama, I., Miyazaki, T., and Sakamoto, S.Stutervant, W. C. History and ethnography of Evolutionary auto ploidyinthe sweetpotato some West Indian starches. In Ucko, J. J., and (Ipomea batatas (L). Lam.) and its preogenitors. Dimsley, G. W., eds., The domestication of Euphytica 24, 1975, 197-208. plants and animals. Duckworth, London, 177 Plucknett, D. L. Edible aroids, Alocasia, Colo- 199, 1969. casia, Cyrtosperma, Xanthosoma (Araceae). InSubramanyan, K. N., Kishore, H., and Misra, P. Simmonds, N. W., ed., Evolution of crop plants. Hybridization of haploids of potato in the plains London, 10-12, 1976. of India. Curr. Sci. 41, 1972, 580. Rao, M. V. Crops of West Asia. Cereals. InTackholm, V., and Dar, M. Flora of Egypt. Vol. Hutchinson, Sir J., ed., Evolutionary studies in 2. Cairo, 1950. world crops. Cambridge, 1974. Ugent, D. The potato. Science, 170, 1970, 1161 Renvoize, B. S. The area of origin of Manihot 1166. esculenta as a crop planta review of theYen, D. E. The sweet potato and Oceania. An evidence. Ec. Bot. 26, 1972, 352-360. essay in ethnobotany. B. P. Bishop Museum Rogers, D. J., and Fleming, H. S. A monograph Bull, 236, Honolulu, 1974. of Manihot esculenta with an explanation ofYen, D. E., and Wheeler, J. M., The introduction the taxotnetric methods used. Ec. Bot. 27, 1973, of taro into the Pacific: the indications of the 3-113. chromosome numbers. Ethnology,7,1968, Sadik, S., and Obereke, O. U. A new approach to 259-267.

Giant Swamp Taro, a Little-Known Asian-Pacific Food Crop Donald L. Plucknettl Cyrtosperma chamissonis (Schott) Merr., a member of the Araceae and probably native to Indonesia, has now spread eastward to become a minor crop in the Philippines, Papua New Guinea, and some Pacific Islands, but a major crop on coral atolls and low islands of the Pacific. A hydrophyte, and extremely hardy perennial, it grows in coastal marshes, natural swamps, and man-made swamp pits in conditions unsuitable for other staple crops. Yields vary, but 10-15 metric tons of the large edible corms per hectare per year have been produced. There is a great need to collect and evaluate cultivars now in use for salinity and flooding limits, short crop duration, superior food value and acceptance, and other factors, before they are lost through neglect.

Recently, interestin subsistence tropicalbenefited from this new emphasis on in- food crops has increased. Root crops havedigenous staple foods. Some roots crops, however, are so poorly known or understood that they continue to be 1Soil and Water Management Division, Officeneglected. Cyrtosperma chamissonis (Schott) Agency for Internation.al Development, Washing- College of Tropical Agriculture, University of Merr., commonly called giant swamp taro, is of Agriculture,TechnicalAssistanceBureau, such a crop. The only cultivated member (if ton,D.C., USA 20523. (Permanent address: C. merkusii is a separate species, then there are Hawaii, Honolulu, Hawaii, USA 96822). two cultivated species) of a pantropic genus of