58 TROPICAL ROOT CROPS SYMPOSIUM Vegetative and Sexual Management in Food Improvement L. Degras1 Yam models are described, giving several required characteristics for clonal cultiva­ tion of alata, D. cayenensis, and D. trifida. Intraclonal selection is possible because of tuber internal heterogeneity, clonal popUlation diversity, and time dispersed production and germination of individual bulbi!. A kind of "somatic drift" is observed. Several aspects of determinism of tuberization and flowering are cited, and day Inight ratio, light intensity, soil mineral balance, internal vegetative phases, and genetic factors are recognized in flowering and in sex-ratio determinism. Basic management using axillary structure growth-substance treatment for vegetative, tuberization, flowering, and sex control is discussed. In addition, the genetic analysis of characteristics, and different cultivation systems are examined.

Yam improvement has normally been based existing widely in the species are: good on the introduction of clonal cultivars selected dormancy, resistance to Penicillium oxalicum, from traditional popUlations. Use of the sexual germination in dry conditions, and good yield system for food yams has been attempted; how­ without staking); D. cayenensis ssp. rotundata ever, it is among pharmaceutic yams that the - ( 1) fair tuber maturity long before foliage sexual system has been extensively used and decay, (2) good tuber regrowth after com­ studied. mercial harvest, (3) high cooking quality Improvement of the food yam through its (whitish flesh), (4) year-round tuber develop­ sexual system began in 1966 with D. trifida in ment, (5) medium storage duration of com­ Guadeloupe (Degras 1969), and it is now mercial harvest, (6) prickless roots, (7) high spreading through an inter-Caribbean selection early yield; D. trifida - (1) tubers available on behalf of the ISTRC-CFCS yam study year-round because of: (a) fresh production all group. The largest food yam breeding poten­ season, (b) food storage duration, (2) high tial now comes from lITA where, since 1970, cooking quality with sweet taste, (3) tuber Sadik and Okereke (1975) have developed grouping near soil surface, (4) high yield with the sexual utilization of D. cayenensis ssp. 10% seed sized tubers, (5) resistant to viruses, rotundata, and tested a large amount of seeds Penicillium oxalicum, nematodes, and mealy throughout the tropics. bugs, (6) spheroidal tuber shape, (7) resistant This does not, however, mean that clonal or tolerant to drought. selection should be stopped, rather it still may have an important role to play. Intraclonal Selection Variation within a clone exists. Heads, Yam Models in Current Clonal middles, and tails from the same tuber differ Cultivation in earliness of germination, yield, and number of stems or tubers produced. In addition, At the present time, all cultivated yams are normal bud regulation is suppressed when the clonally propagated. The characteristics of D. slice size is greater than 5 g (Degras and alata, D. cayenensis, and D. trifida in the Mathurin 1975). It is important that differ­ French West Indies are as follows: D alata­ ences have been repeated over first and second (1) high cooking quality (white flesh), (2) generations of two D. trifida clones obtained long storage without loss of weight, (3) re­ from different parts of a tuber of cuItivar sistant to anthracnosis and viruses, (4) resistant INRA 25: flowering time differences at the to water stress, (5) high yield with medium second generation were in accordance with be­ and regular size tubers (other characteristics haviour in the first generation, differences in time to maturity were up to 1 month. 1 Plant Breeder, Institut national de la recherche agronomique (INRA), Centre des Antilles et de We do not know the level of genetic homo­ la Guyane, Domaine Duclos, Prise d'eau, Petit geneity of traditional cultivars, but some may Bourg, France. include mutational variations. Off-types are DEGRAS: YAM MANAGEMENT 59

Table 1. Agronomic variability from D. cayenenis cultivar Krengle intraclonal tuber selection.

Tuber shape Plant size Growth Tubers Mean tuber (seed tuber) (200 days) durationa Yieldb harvestedb weightb

Cy Iindrical e 100 100 100 100 100 Cylindro-conical 110 111 91 103 89 Conical 100 104 88 108 82 Ovoidal 113 104 82 88 94 Spheroidal 103 100 73 102 67

aStatistical significance has been found for shape X seed tuber weight interaction in cycle duration. bSignificative differences p = 0.05. "Cylindrical data are taken as basis = 100 for each trait. normally easily recognizable when cultivated the sexual phase is very depressed in annual with cultivars: cultivar individuality is a fact. cultivation. A kind of competition exists be­ Nevertheless, in the well known cultivar D. tween vegetative accumulation and sexual de­ cayenensis ssp. rotundata "Krengle," in central velopment for the following reasons. Ivory Coast, a number of tuber forms could be Tuberization and flowering seem to be separated. When cultivated separately they favoured by short days for a number of species. have distinct performances (Table 1). This has been shown for bulbils and for tuber Another case of intraclonal variation that growth. D. alata and D. trifida normally flower may be explored arises from different times to in the autumn, but Henry (1967) obtained maturation of the bulbils of the plant. In D. earlier flowering for a number of seedlings bulbi/era and D. alata, the difference may under short days. reach 1 month. High light intensity is necessary for flowering These examples show that a form of and tuberization: Henry (1967) observed a "somatic drift" of clonal expression of a yam lower level of flowering of D. trifida seedlings genotype could proceed from conscious suc­ when shaded. For most species, staking, which cessive selection of extreme phenotypic varia­ permits more light to enter the canopy, gives tions. It is obvious that in current cultivation, higher tuber yields. The difference is striking such variations do not cause marked changes for a number of dry-forest climax species like of the clonal population. But we think that D. cayenensis, whereas it is less important for through controlled cultivation (planting time rain-forest climax species like D. alata and D. and density) and controlled dormancy dura­ trifida. tion we could expect to alter the phenotypic The physiological response to fertilization balance of clonal properties. The selection of suggests a common process for initiation of the type of planting material from tubers of a tuberization and flowering. We observed that clone could at the same time determine the given balances of NPK in D. cayenensis in­ best rate of mUltiplication and the highest rate creased the tuber yield and the percentage of of phenotypic variation that could be explored that flower. In D. alata, stem fasciations through environment X growing condition in­ which may be considered in some cultivar~ teractions (Mathurin and Degras 1974, Degras from central Ivory Coast as a substitute for and Mathurin 1975). In this respect, an ex­ flowering, occurred when tuber yields were periment is now in progress using D. trifida greatest. cultivar INRA 25. Generally speaking, flowering seems to be triggered by a certain level of development of Flowering Determinism the vegetative organs. This could account for the low level of flowering observed in the first Flowering Stage and Vegetative growing cycle of D. trifida seedlings when com­ Development Relation pared to the second cycle. This level is not merely a matter of biomass: plants with heavy Wild and domestic yams seem to give vegetative organs do not necessarily flower. preference to the vegetative system over the It seems that the relation is both qualitative and sexual system (Burkill 1960). In many cases quantitative. 60 TROPICAL ROOT CROPS SYMPOSIUM Aspects of the control of the vegetative and enensis spp. rotundata, Sadik and Okereke flowering stages are also seen in the balance (1975) observed no female inflorescence in between bulbil and inflorescence development the first seedling cycle, but in the following in some species. Along one plant of D. bulbi­ growth cycle, female plants appeared in a fera (Martin 1974) there are a succession of higher proportion than in traditionally vege­ foliar axils, first with bulbils alone, then with tatively reproduced populations. both bulbils and inflorescences, then with Another aspect of Dioscorea flowering seems inflorescences alone. Among populations of constant: the time of female flowering with re­ primitive cultivars of D. alata some clones bear spect to that of males. This may be associated only bulbils and others only inflorescences. with the tendency toward higher tuber yields Bulbil setting begins somewhat before the from female plants. flowering time. To obtain high levels of yam crossing the Genetic factors also control flowering following are recommended: (1) a high level formation. Apart from the wide differences of vegetative growth, which is necessary for known between cultivars in a D. cayenensis full female expression; (2) several plantations for instance, it is known that sexual reproduc­ of both parents for simultaneous flowering; tion and selection can increase the percentage and (3) common staking of male and female of flowering (Sadik and Okereke 1975). inflorescences so that flowers are close (pollen So, ecological factors like day/night ratio, is sticky and wind dispersal poor). light intensity, soil mineral balance, internal New research could lead to a better manage­ vegetative phases, and genetic makeup control ment of yam flowering. The very specific nature yam flowering. of the nodal and axillary yam structure is obvious from the special vascular organization Sex-Ratio Determinism (glomeruli) ascribed to this genus (Ayensu Though dioecism is the general case in yams, 1972). These glomeruli are capable of develop­ a number of cases have been observed where ing into flowers, bulbils, or chiots. A systematic some level of monoecism and even of herma­ screening of growth substance, and the nutri­ phroditism is obvious. In central Ivory Coast tional and physical effects on the yam nodal the occurrence of monoecious plants in D. axillary system, as applied to the axillary com­ cayenensis is affected by mineral fertilization plex of Euphorbiaceae (Champault 1973), and season, and its occurrence is higher in may be successful in increasing flowering. sexual progenies than in clonal material. Sadik and Okereke (1975) have recently con­ First Data from Genetic Analyses firmed these findings. Martin (1966) attempted an interpretation In a number of species a prevalence of male of sex-ratio heredity in yam, and we have flowers is reported. In some wild species, the limited data on D. trifida (Degras 1969, De­ female form is unknown, for example in Mada­ gras et aI. 1973). We now have more complete gascan flora (Burkill et al. 1950) or in Mexican data for a number of crosses in this species. flora (Matuda 1954). Martin (1966) who studied the behaviour of steroid species, sug­ Anthocyan in Tuber Flesh gested that the variations of sex-ratio in prog­ Table 2 gives the distribution of anthocyan enies proceed from the different heterosomic in progeny. lnterpretation is difficult because status of the male parents, which in tetraploid of: the limited seed germination; the death of a species could be XXXY, XXYY, or XYYY. number of plants in the field; and the possible The latter is lethal or only gives males. interaction of anthocyan expression with de­ In D. trifida, Henry (1967) observed a high gree of maturity of harvested seedlings. Never­ prevalence of males in common clonal prop­ theless we observed: (1) a dominance of purple agation, a lower prevalence in the first seed­ over white; (2) a transgression beyond the ling growth cycle, and a still lower prevalence purple; and (3) an occurrence of intermediate in clonal mUltiplication of these seedlings. De­ levels; suggesting a number of modifying gras et aI. (1973) observed an increase of factors. female plants when flowering was increased. In D. trifida seedlings a 3/1 male/female ratio, Length of Nontuberous Parts of Stolon which in the case of predominant tetraploidy, The length distribution of the stolons, for 40 could result from a XXYY male. In D. cay- progeny of D. trifida (C.C.V. X INRA), was DEGRAS: YAM MANAGEMENT 61 Table 2. Distribution of anthocyan in tubers of 40 sex type and yield has also been found by progeny of D. trifida (e.e. v. X INRA). Sadik and Okereke (1975). Seed germination and many aspects of the Skin colours two first progeny growing cycles of D. trifida Deep Light were studied by Henry (1967). Since 1965 in Aesh colours purple Purple purple White Guadeloupe, about 5000 seedlings from 20 crosses have been observed and tested, and Deep purple 3.3 1.6 improved cultivars have been obtained. Among Purple 22.9 18.oa them INRA 25 and INRA 5-20 associate a Light purple 3.3 18.0 number of the required characteristics. Resist­ Very light purple 1.6 3.3 ance to diseases (Penicillium rot, viruses) and White 1.6 6.6 19.6b pests (mealy bugs, nematodes) are still lacking ac.c. Violette type. in our hybrids. Special genetic searches are bINRA 40 type. projected in the Guyanas. as follows: less than 10 cm 32.2%; 10-15 cm New Horizons 26.2 %; 15-20 cm 20.8 %; and more than 20 Now that the feasibility of genetic improve­ cm 20.8 %. Long stolons appear dominant over ment of yams is well established, we can ex­ short ones. Here also a transgression beyond amine some new cultivation systems. parent limits is observed. It seems that for most First, more intensive cropping will proceed characteristics, in accordance with the general from the following genetic modifications: (1) high level of polyploidy in cultivated yams, greater efficiency in translocation of assimilates multiple level factor determinism is the rule. to the tuber, including limitation of stem de­ The analysis of yam genetics will benefit velopment, shortening of the prebulking tuber from the utilization of auto fertile monoic or phase, and faster tuber maturation; (2) lower hermaphroditic plants. If androgenesis could interplant competition, permitting higher be applied to yam pollen to give haploid clones, plantation density; (3) more rapid drying of this would open new opportunities of genetic the tops at maturity to aid mechanical harvest­ analysis. Recently attempts have been suc­ ing; (4) underground structure superficially cessful (Arnolin 1976, personal communica­ compacted for better mechanical harvesting; tion). and (5) higher nutritive value of tuber. Known variability permits a reasonable expectation of Sexual Progeny Selection obtaining clones with most of these traits. The success of selection from free and con­ Other quite new genetic modifications can trolled pollinated crossings of D. cayenensis in be envisaged if sexual seeds become the basic West Africa, and D. trifida in the West Indies, material for plantations. Preliminary require­ shows that theoretical knowledge of the genetic ments are: highly prolific female parents; mechanisms are not needed to make advances. good seed germination; and knowledge of the The first conscious agronomic utilization of combining ability of the parents. hybridization in D. cayenensis occurred in Two levels of these crop modifications can 1955 at Bouake (Ivory Coast) where Franck be projected: (1) the tubers from the seedling harvested 3020 seeds from cultivar "Assaoua" may be used as propagating material for com­ and obtained 48 seedlings. Only 13 clones were mercial production; and (2) the seedling cycle retained. We studied them from 1956-58, and must give good tuber yields and hence breeding selected from them Assaoua B9 (Van de Venne selection will change from clonal to seedling 1973, personal communication). performance. In both cases a relative genetic Among many selection criteria we noticed, homogeneity is wanted in the progeny. Both at the first clonal cycle, the value of flowering, parents, or at least one of them, should be built sex expression, and phyllotaxy. Nonflowering through successive brother X sister crossings plants produced 668 g, monoic 770 g, and fe­ in order to reach some level of homozigosity male 1135 g (mean tuber weight). Leaf and for main characteristics. branch balance on successive nodes gave a yield Ayensu, E. S. Anatomy of the . of 1000 g (worst balanced system) to 3000 g VI. . Metcalfe, e. R. (ed.), Claren­ (best balanced systems). A relation between don Press, Oxford, 1972, 182 p. 62 TROPICAL ROOT CROPS SYMPOSIUM

Burkill, I. H. The organography and the evolution Ibandan, Nigeria, 1973 (in press) . of Dioscoreace, the family of the yams. J. Linn. Degras, L., and Mathurin, P. New results in yam Soc. Bot. 56,1960,319-412. multiplication. Proc. 13th Ann. Meet. Carib. Burkill, I. H., and Perrier de la Bathie, H. Dio­ Food Crops Soc., Trinidad, 1975. scoreacees (). Flore de Madagas­ Henry, V. C. R. Studies on botanical and agro­ car et des Comores, 44e famille. Firmin-Didot, nomic characeristics in cush-cush (Dioscorea Paris, 1950,73 p. trifida L.). Thesis, McGill Univ., Montreal, Champault, A. E/Jets de quelques regulateurs de Canada, 1967. la croissance sur des noeuds isoles de Mercuri­ Martin, F. W. Sex ratio and sex determination in aUs annua L. (2n = 16) cultives in vitro. Bull. Dioscorea. Heredity, 57(3), 1966,95-99. Soc. Bot. Fr. 120, 1973,87-100. Mathurin, P., and Degras, L. E/Jects of division Degras, L. Quelques donnees sur la variabilite des levels of seed tubers on yams (D. alata, D. descendances d']gname Cousse-couche (D. trifida) germination and yield. Proc. 12th Ann. trifida). Proc. VlIth Ann. Meet. Carib. Food Meet. Carib. Food Crops Soc., Jamaica, 1974. Crops Soc., Guadeloupe-Martinique, 1969, 59- Matuda, E. Las Dioscoreas de Mexico. Ann. Inst. 65. BioI. Univ. Mex. 24, 1954,279-390. Degras, L., ArnoHn, R., and Poitout, R. Selection Sadik, S., and Okereke, O. U. A new approach to de l'Igname Cousse-couche (Dioscorea trifida L.) improvement of yam Dioscorea rotundata. Proc. 2nd Int. Symp. Trop. Root Crops, Soc., Nature, 254, 1975, 134-135.

Sweet Clones Adapted for Libyan Agriculture Warid A. Warid, Boleid W. Dahmani, and Mosbah M. Kushad1 Five experiments were conducted from 1971 to 1973, on the adaptability of some introduced clones of sweet potato. An average yield of marketable enlarged roots of 7.4- 21.2 tlha was produced by clone American from USA, 24.3-26.6 t/ha by clone Kahera Hybrid, and 14.5-29.6 t/ha by clone Mabrooka. The last two clones were introduced from Egypt. The most variable clone in weight of , number of roots, or yield was Mabrooka, and the least variable was Kahera Hybrid. Within a given clone, the yield showed the highest degree of variability. A nonsignificant correlation existed between vine weight and number of roots. A positive correlation was found between vine weight and yield in two clones, and between num ber of roots and yield in all clones.

The sweet potato has always been an im­ marketable roots. Perennial plots were estab­ portant food crop. However, the sweet potato lished for promising clones to obtain stem crop is not popular among Libyan farmers. cuttings for use as planting material in clonal The enlarged roots offered for sale are of in­ evaluation tests. ferior quality. They are elongated and not Little work has been done on sweet potatoes uniform in shape. The skin surface is ridged in Libya, although recommended cultivars and tan coloured. The flesh is whitish, almost were reported by Mazzocchi and Thrower tasteless when boiled, and has a high fibre con­ (1962). The aim of this study was to screen tent. The vegetable breeding program at the introduced clones, and possibly recommend University of Tripoli has included the improve­ one or more clones for food and industry. ment of this crop since 1969. A few enlarged roots of some clones bred in Egypt, Japan, and Materials and Methods the United States were introduced and asex­ ually propagated for the first 2 years. Some of A total of five experiments were conducted these clones were discarded early because of from 1971 to 1973 at the University of Tripoli. poor growth of , low yield of enlarged The soil was a sandy loam (pH 7.5-8.5). Stem roots, or the production of only thick and non- cuttings, 20-25 cm long, were planted 30 cm apart in rows 60 cm wide. The experimental lDepartment of Plant Production, College of plot unit varied from 2 to 8 m 2 (excluding Agriculture, University of Tripoli, Tripoli, Libya. guard rows). The crop was grown as a summer crop under irrigation. Mean air temperature