Bulletin de Ressources Phytogénétiques...and anti-evolutionist (whose work Darwin nevertheless...

Food and Agriculture Organization of the United Nations and theInternational Plant Genetic Resources InstituteOrganisation des Nations Unies pour l'alimentation et l'agriculture etl'institut international des ressources phytogénétiquesOrganización de las Naciones Unidas para la Agricultura y la Alimentación yel Instituto Internacional de Recursos Fitogenéticos

Noticiario de Recursos Fitogenéticos

Bulletin de Ressources Phytogénétiques

Plant Genetic Resources Newsletter

No. 127, 2001

Bulletin de Ressources Phytogénétiques


Noticiario de Recursos Fitogenéticos

ISSN 1020-3362

Managing EditorPlant Genetic Resources NewsletterIPGRIVia dei Tre Denari, 472/a00057 MaccareseRome, ItalyTel.: (+39)0661181Email: [email protected]: (+39)0661979661Web: http://www.ipgri.cgiar.org

EditorialOffice

© IPGRI/FAO 2001

Bureau derédaction

Oficina deRedacción

The designations employed, and thepresentation of material in the peri-odical, and in maps which appearherein, do not imply the expressionof any opinion whatsoever on the partof IPGRI or FAO concerning the legalstatus of any country, territory, cityor area or its authorities, or concern-ing the delimitation of its frontiers orboundaries. Similarly, the views ex-pressed are those of the authors anddo not necessarily reflect the viewsof IPGRI or FAO.

Les appellations employées danscette publication et la présentationdes données et cartes qui y figurentn’impliquent de la part de l’IPGRI etde la FAO aucune prise de positionquant au statut juridique des pays,territoires, villes ou zones, ou deleurs autorités, ni quant au tracé deleurs frontières ou limites. Les opin-ions exprimées sont celles desauteurs et ne reflètent pasnécessairement celles de l’IPGRI oude la FAO.

Las denominaciones empleadas, yla forma en que aparecenpresentados los datos en estapublicación, no implican, de partedel IPGRI o la FAO, juicio algunosobre la condición jurídica de países,territorios, ciudades o zonas, o desus autoridades, ni respecto de ladelimitación de sus fronteras olímites. Asimismo, las opinionesexpresadas son las de sus autoresy no reflejan necesariamente laopinión del IPGRI o la FAO.

Cover: Diversity of shape, size andcolour of coconut germplasm in In-dia. Discussed by Arunachalam,Jerard, Elangovan, Ratnambal,Dhanapal, Rizal and Damodaran onpp. 39–43.

Couverture : Diversité de forme, tailleet couleur du germoplasme de la noixde coco en Inde. Une discussion parArunachalam, Jerard, Elangovan,Ratnambal, Dhanapal, Rizal etDamodaran aux pages 39–43.

Portada: Diversidad de forma,tamaño, y color de la germoplasmade Coco in la India. Debate deArunachalam, Jerard, Elangovan,Ratnambal, Dhanapal, Rizal yDamodaran en pp. 39–43

Plant Genetic Resources Newsletter, 2001, No. 127 1Plant Genetic Resources Newsletter, 2001, No. 127: 1 - 10

Benefits from giving and receiving geneticresources: the case of wheat1

Kelly Cassaday¹, Melinda Smale²? , Cary Fowler³ and Paul W. Heisey4

¹ International Maize and Wheat Improvement Center (CIMMYT), Apartado Postal 6-641, 06600 Mexico, D.F. Mexico² IPGRI, Rome, Italy and the International Food Policy Research Institute (IFPRI), 2033 K Street N.W., Washington, DC 20006, USA. Tel: +1 202 8625600; Fax: +1 202 4674439; Email: [email protected]³ Agricultural University of Norway, PO Box 5001, N-1432 AAS, Norway4 Economic Research Service, US Department of Agriculture

SummaryBenefits from giving andreceiving genetic resources:the case of wheatThis paper examines the origins of mod-ern wheat varieties, especially springbread wheat varieties produced by thepublic research systems in developingcountries, of which CIMMYT (the Inter-national Maize and Wheat ImprovementCenter) is a part. Centuries of selectionand evolution produced wheat landracesalmost everywhere in the world. Thecomplex pedigrees of CIMMYT’s springbread wheats contain vast arrays of theselandraces from developed and develop-ing countries. CIMMYT’s wheat breed-ing lines are distributed mainly to re-searchers in developing countries, andcontribute significantly to the pedigreesof most wheats released in these coun-tries. These wheats exhibit a broader ge-netic base, more stable yields, and betteradaptation to conditions faced by poorfarmers than the ‘modern’ varieties of 30years ago or contemporary varieties lack-ing CIMMYT germplasm. These wheatsalso provide substantial economic bene-fits, particularly to developing countries,largely because of a policy environmentthat encouraged the flow of genetic ma-terials and supported public sector re-search with those materials. The case ofwheat illustrates the benefits of promot-ing flows of germplasm of all crops, es-pecially through laws and policies thatallow institutes such as CIMMYT to ac-quire genetic materials and freely distrib-ute research products.

Key words: Developing countries,genetic diversity, genetic resources,germplasm flow, modern varieties,wheat

ResumenBeneficios de dar y recibirrecursos genéticos: El casodel trigoEn este documento se examinan los orí-genes de las variedades modernas de tri-go, especialmente las de trigo harinero deprimavera generadas por los sistemaspúblicos de investigación en los países endesarrollo, de los cuales forma parte elCIMMYT (Centro Internacional de Mejo-ramiento de Maíz y Trigo). Tras siglos deselección y evolución se generaron var-iedades criollas de trigo en prácticamentetodo el mundo. Las complejas genealogíasde los trigos harineros de primavera delCIMMYT incluyen un gran número deesas variedades, que provienen tanto delos países en vías de desarrollo como losdesarrollados. Las líneas de trigo del CIM-MYT se distribuyen principalmente a in-vestigadores de los países en desarrollo ycontribuyen de manera significativa a lasgenealogías de la mayoría de los trigoslanzados en esos países. En consecuencia,esos trigos poseen una base genética másamplia, rendimientos más estables y me-jor adaptación a las condiciones de losagricultores de escasos recursos que lasvariedades “modernas” de hace 30 años ylas que no contienen germoplasma delCIMMYT. Estos trigos generan beneficioseconómicos sustanciales, sobre todo enlos países en desarrollo, en gran partegracias a una política que ha fomentado ladistribución de materiales genéticos yapoyado la investigación del sector públi-co con esos materiales. En conclusión, elcaso del trigo ilustra los beneficios de pro-mover la distribución de germoplasmade todos los cultivos, especialmente medi-ante leyes y políticas que permitan a insti-tutos como el CIMMYT adquirir materi-ales genéticos y distribuir gratuitamentelos productos de su investigación.

ARTICLE

RésuméLes advantages des échangesde resources genetiques : lecas du bléCe document examine les origines desvariétés modernes de blé, notamment lesvariétés de blé tendre de printemps pro-duites par les systèmes de recherche pub-lique des pays en voie de développement,dont le CIMMYT (Centre Internationalpour l’Amélioration du Maïs et du blé) faitpartie. Des siècles de sélection etd’évolution ont généré des variétéstypiques de blé dans pratiquement lemonde entier. La généalogie complexedu blé tendre de printemps du CIMMYTinclut un grand nombre de ces variétésmodernes qui proviennent aussi bien depays développés que de pays en voie dedéveloppement. Les lignées de blé tendrede printemps sont distribuées principale-ment aux chercheurs des pays en voie dedéveloppement, et contribuent, demanière significative, à la généalogie desblés les plus distribués dans ces pays. Enconséquence, ces blés offrent une basegénétique plus large, des rendements plusstables, et une meilleure adaptation auxconditions des petits fermiers que lesvariétés « modernes » d’il y a 30 ans oucelles composant actuellement le matérielgénétique du CYMMIT. Ces blés génèrentdes bénéfices économiques substantiels,surtout dans les pays en voie de dével-oppement, en grande partie grâce à unepolitique qui a encouragé la distributionde matériel génétique et soutenu la re-cherche du secteur publique à l’aide de cematériel. Pour conclure, le cas du blé illus-tre l’avantage de promouvoir la distribu-tion de matériel génétique de toutes lescultures, spécialement au travers de lois etde politiques qui permettent aux institutscomme le CYMMIT d’acquérir du matéri-el génétique et de distribuer gratuitementle produit de leurs recherches.

IntroductionJean Henri Fabre, a celebrated 19th century French entomologistand anti-evolutionist (whose work Darwin nevertheless ad-mired), once astutely observed that “History celebrates the battle-fields whereon we meet our death, but scorns to speak of theplowed fields whereby we thrive; it knows the names of the

King’s bastards, but cannot tell us the origin of wheat. That is theway of human folly.”

In recent years, a great deal has been uncovered about theorigins of wheat as a crop (Heun et al. 1997; Kawahara andTaketa 2000). Even so, knowledge of and curiosity about the

1 At the time the research described in this paper was conducted, M. Smale and P.W. Heisey were Economists with CIMMYT. The viewsexpressed in this paper are the authors’ and do not necessarily represent views or policies of their respective organizations.

2 Plant Genetic Resources Newsletter, 2001, No. 127

origins of modern wheat varieties, who has fashioned them, how,and with what genetic materials, remains low both in scientificcircles and among the general public. Were he here today, Fabre,a keen observer of insect behaviour, might remark that our think-ing has scarcely evolved at all.

In this brief paper, we look at the origins of modern wheatvarieties, especially spring bread wheat varieties produced bythe public sector and sown in developing countries. In particu-lar, we examine how these varieties are constituted, what mate-rials are used in the breeding process, and where these materialswere acquired. We demonstrate the effects of the internationalwheat breeding system on spring bread wheat genetic diversityand varietal performance, the wide use of scientifically bredvarieties in developing country wheat production, and the eco-nomic impacts of this international exchange of genetic re-sources.

Today, the origins of wheat are important for reasons Fabrecould never have imagined. In 1993, the Convention on BiologicalDiversity came into force. This historic and legally binding treatylaid down rules for the conservation, management and move-ment of genetic resources. Access to genetic resources, accordingto the Convention, would henceforth be granted by the country oforigin on the basis of prior informed consent and mutually agreedterms. Negotiators, however, recognized the peculiar case ofagricultural crops and invited the Food and Agriculture Organi-zation (FAO) of the UN to initiate negotiations specifically onplant genetic resources for food and agriculture. These negotia-tions are still underway2.

To a large extent, the negotiations are influenced by whatdelegates know or suspect about the origins of crops and cropvarieties, ancient and modern. Understandably, countries wish tobenefit from the transfer of genetic resources. All want access tothe best and most appropriate new crop varieties, but somequestion whether the contribution of landraces and wild rela-tives—the ‘raw material’ for new varieties—has been worth it.Perhaps more has been lost than gained through systems offacilitated access to agricultural genetic resources.

The rules that countries establish through the FAO negotia-tions could dramatically affect the origins, i.e. the genetic content,of future wheat varieties, as well as varieties of all other cropsvital to food security and economic development. Whether posi-tive or negative, the impact of the decisions taken at FAO is likelyto be far more than trivial. Indeed, the impact could be earth-shaking, particularly for the poorest and most vulnerable inhab-itants of this world. Simply stated, it is important, therefore, thatone understand wheat and its ancestry.

The history and genetic background ofmodern wheatsTo make any sensible statement about the place of origin ofwheat, one would have to specify ‘what’ wheat and ‘when’(Harlan 1971). Zohary (1970) has referred to wheat’s origins as‘confused’, and Harlan (1992) has described them as ‘diffuse’(neither centric nor non-centric). Wheat’s history is a long one, thefirst domesticated forms appearing some 9000 years ago (Harlan

1977). The great diversity of the species can be attributed par-tially to its extended cultivation and conservation in so manyparts of the world over such a long period. Perhaps as early as6000 BC, wheat started its travels from West to East Asia. By4000 BC, Neolithic farmers were growing and improving wheatacross large areas of today’s developing and developed world,from North Africa and the Near East to South Asia and onwardto China. Wheat was being grown in Europe and as far north asScandinavia, as well as in present-day Russia. Part of the famous‘Colombian Exchange’ wheat arrived in the New World whenSpaniards took it to Mexico in 1529, and then spread to southernAfrica, Australia and beyond (Feldman 1976). The dispersal ofwheat was marked by continual flows. Just as wheat moved outof the Near East to Europe and Africa, ‘new’ wheats—trans-formed by evolution and selection in different environments overmany generations—also flowed back to the Near East and weretaken to other regions. Viewed over time, these flows ofgermplasm were continuous, large and multidirectional.

Common wheat (also known as hexaploid or bread wheat) isan example of ‘genome building’ that occurred through interspe-cific hybridization of three diploid grasses, each of which do-nated a genome (Zohary et al. 1969). As far as we know, breadwheat began to appear relatively recently (7000 years ago) some-where near the southwest corner of the Caspian Sea (Zohary et al.1969), when plants of emmer wheat growing on a Neolithic farmwere fertilized by pollen blown from a nearby plant of a weedydiploid called goatgrass. Progeny of this spontaneous hybridsurvived and were then domesticated by humans. Zohary et al.(1969) contend, however, that hexaploid wheats did not originatefrom one single hybridization event, but from numerous occa-sions with a variety of races and forms.

Bread wheat has since become one of the most commonlygrown, commercially valuable food crops in the world. Thoughover half of the world’s wheat crop is produced on the large,mechanized farms of industrialized countries, most of the world’swheat farmers live in the developing world. About two-thirds ofthe total area sown to wheat in the developing world is sown tospring bread wheat, with the remaining third sown to a combina-tion of durum wheats and bread wheats with winter and faculta-tive growth habit. Heisey et al. (in press) have estimated that thetotal area sown to spring bread wheat in the developing worldtoday is about 72 million ha.

Indicators of diversity in modern wheatsThe modern varieties of spring bread wheat grown in the develop-ing world today are the products of wheat’s complicated history,a fact that is reflected in their pedigrees. Genealogies and thecomplexity of breeding history provide a glimpse of the ‘latentdiversity’ present among modern varieties of highly bred crops.Latent diversity is the extent of genetic variability in varieties thatis not observed in the field until challenged by some environmen-tal stress (Souza et al. 1994).

Although molecular data can be used to depict the diversityof both landraces or populations of wild relatives and modernvarieties of highly bred crops, genealogical information exists

2 This article was accepted forpublication before treaty negotiation was completed in November 2001.

Plant Genetic Resources Newsletter, 2001, No. 127 3

only when breeding history is recorded. In the following sections,most of the indicators we present are based on genealogies, asexplained below. Some molecular evidence is also reported in thefinal section.

The numbers of unique landrace ancestors and unique paren-tal combinations shown in the tables later in this paper indicatepedigree complexity and the number of new materials that arebred into the background of each variety. The total numbers oflandrace ancestors and parental combinations indicate the scopeof the breeding effort—the number of crosses made and theresearch time dedicated to producing a new variety.

Modern professional wheat breeders using conventional meth-ods only rarely cross landraces with advanced breeding lines be-cause many breeding cycles are required to eliminate deleteriouscharacteristics while incorporating heritable, useful traits. Lan-draces are usually incorporated into modern wheats through cross-ing breeding lines with different pedigrees. Therefore, the number oflandraces and the frequency of use are by no means a measure oftheir genetic contribution to yield or other economically importanttraits. Nor are they indicators of diversity in and of themselves.

The coefficient of parentage (COP) provides a theoreticalestimate of the genetic relationship between two cultivars based

on genealogical information. The COP estimates the probabil-ity that a random allele taken from a random locus in onevariety is identical, by descent, to a random allele taken fromthe same locus in another variety (Falconer 1981). Calculatedwith detailed pedigrees and Mendelian rules of inheritance, thepairwise COP has been employed in the crop science literatureas a measure of similarity between two modern varieties due toinheritance. Genetic diversity or distance is then calculated asunity minus the coefficient of parentage. The matrix of pairwisecoefficients of diversity for a set of modern varieties is used toconstruct indices of latent genetic diversity (Souza et al. 1994).

In the molecular analysis cited here, DNA was extractedand molecular fingerprinting analyses conducted using SSR(simple sequence repeat) and AFLP (amplified fragment lengthpolymorphism). Simple matching coefficients of similarity werecalculated for all pairs of individuals. Genetic distance wasmeasured as unity minus the coefficient of similarity.

The international wheat improvementsystem and pedigree complexityThe international wheat improvement system, formed byCIMMYT (the International Maize and Wheat Improvement

Fig. 1. The CIMMYT open recurrent selection program makes crosses in Mexico; the resulting germplasm is distributedinternationally, and information and reselected germplasm are returned to CIMMYT. Source: Skovmand and DeLacy(1999).

Fig. 2. Percent distribution of wheat seed samples sent byCIMMYT International Nurseries to developed and developingcountries, 1994–1999. Source: Calculated from data providedby CIMMYT International Nurseries. Note: ‘Developed’ or‘developing’ nation status follows FAO classification. Datarepresent the total number of seed samples sent for eachspecies in all nursery sets.


Fig

. 3. S

egm

ent o

f whe

at v

arie

ty S

onal

ika.

Sou

rce:

Sm

ale

(199

6).

NO

TE

:1.

BO

XE

S W

ITH

TW

O L

INE

S A

RE

LA

ND

RA

CE

S,

CR

OS

SE

S,

OR

SE

LEC

TIO

NS

AB

OU

T W

HIC

HN

O F

UR

TH

ER

IN

FO

RM

AT

ION

IS

KN

OW

N.

2. N

OT

ALL

SE

GM

EN

TS

OF

TH

E P

ED

IGR

EE

AR

E S

HO

WN

.3.

DA

TE

S M

AY

RE

FE

R T

O R

ELE

AS

E D

AT

E O

RC

RO

SS

DA

TE

, A

ND

AR

E A

PP

RO

XIM

AT

E.


Center) and public research programmes in developing countries,has facilitated large flows of germplasm (Fig. 1). CIMMYT’swheat breeding programme makes about 19 000 crosses everyyear. Products of these crosses, as well as seed from CIMMYT’sgenetic resources centre, are made available to breeders world-wide. From 1994 to 2000, CIMMYT distributed 1.2 millionsamples of bread wheat, durum wheat, triticale, barley and otherseed. In each year, 65–77% of samples were sent to developingcountries, with 71.3% of the total distributed to developing coun-tries over the 6-year period (Fig. 2). All but a very minor share ofthese materials was transferred to public agencies.

Sonalika and Veery wheats, two of the best known (and mostwidely grown) wheats ever released in developing countries 3, areCIMMYT-related spring bread wheats that are themselves de-scended from numerous varieties and landraces that have beentraced to many parts of the world. Figure 3 shows a segment ofSonalika’s pedigree. Both wide and long, the pedigree revealsthat, for this portion alone, breeders used landraces from 17 coun-tries and breeding lines from 14 countries. In all, six continents‘contributed’ to this small section of a much longer pedigree. Thepedigree of Veery is considerably more complex than Sonalika,with 10 more distinct landraces and eight times as many parentalcombinations.

Table 1 shows the percentage distribution by source regionof distinct landrace ancestors in the pedigrees of bread wheats

3 Sixty-four cultivars were released from the Veery family of wheats.

Table 1. Source and destination of landraces in bread wheat crosses grown in the developing world in 1990

Region of destination in developing world

Source region Sub-Saharan North Africa West Africa South Asia Mexico/ Andean SouthernAfrica Guatemala Region Cone

Percentage distribution of numbers of landraces in pedigree

Sub-Saharan Africa 12 9 7 9 10 12 7North Africa 2 4 2 3 2 1 1West Asia 2 1 7 2 1 1 1South Asia 10 8 7 21 6 10 6Mexico and Guatemala 4 3 7 6 9 7 5Andean Region 0 0 0 0 1 1 1Southern Cone, 14 16 8 11 16 17 31South AmericaChina 1 1 1 1 1 1 1Developing world 45 42 39 53 46 50 53

North America 8 6 9 4 9 9 10Northern Europe 10 8 5 6 9 8 6Southern Europe 7 10 15 8 8 9 8Poland, Germany, 15 21 16 18 21 19 14Former Soviet UnionJapan and Korea 1 3 2 2 2 2 2Australia 1 0 2 0 1 0 1Developed world 42 48 49 38 50 47 41Unknown 13 10 12 9 4 3 6All 100 100 100 100 100 100 100

Source: Smale (1996).Note: All countries in the ‘developing world’ category are low- or middle-income countries. All countries in the ‘developedworld’ category are high-income countries except for Poland and the countries of the former Soviet Union.

Fig. 4. Frequency of landrace use in pedigrees of breadwheats grown in the developing world in 1990.Source: Smale (1996).


grown in developing countries in 1990—including both thosewith spring and winter habit, though winter wheats are a minorproportion of those grown. Comparable data were not availablefor the bread wheats grown in developed countries. In thisinstance, each landrace ancestor is counted only the first time itappears in a pedigree. (See Fig. 4 for data that reflect thefrequency with which landraces are used.) Averaged over re-gions, the developing world contributed 47% and the industri-alized world 45% of the distinct landrace ancestors found in thepedigrees of modern bread wheats. (The source of 8% of thelandrace ancestors in those pedigrees is unknown.)

Landraces can enter the pedigree of a given wheat varietymore than once through various breeding lines (Fig. 4). Lan-draces with important traits, such as disease resistance, mayappear with high frequency. Measured by the frequency oflandrace use, the proportional contributions of the industrial-ized world (Poland, Germany, the Former Soviet Union andEast Asia) to modern wheats grown in the developing worldappear higher than the contributions of the developing worlditself. Frequently used landrace progenitors of both spring and

winter wheats include some spring wheat ancestors such as RedFife (originally from Poland), as well as major winter wheatancestors, such as Turkey, from the Caucasus (Zeven andZeven-Hissink 1976). The high frequency with which breedersused East Asian landraces is explained largely by the use ofonly two landraces, Akagomughi and Daruma. These landracesprovided the dwarfing genes that made wheat plants shorter.Daruma contributed the Rht1 and Rht2 genes in the semidwarfwheats that have had such a dramatic effect on wheat yieldsover the past 30 years.

Tables 2 and 3 give some idea of the scope of the interna-tional wheat improvement effort and the resulting complexityof modern wheat pedigrees. The tables present pedigree char-acteristics of all major and a sample of minor wheat varietiesgrown in developing countries in 1997. (Each major varietywas grown on more than 0.25 million ha in the developingworld in 1997; each minor variety occupied less than0.25 million ha.)

The international exchange of breeding materials makes itmore likely that a breeding programme will introduce ‘new’

Table 2. Pedigree characteristics of major spring bread wheats (varieties planted on more than 0.25 million hain the developing world in 1997)

Variety Country Year No. of parental Percentage No. of No. of Area (ha)released combinations of parental ancestors different

in pedigree combinations in pedigree landraceused once ancestors

Sonalika India 1966 419 21 419 39 1 073 000UP-262 India 1977 696 16 693 45 1 261 000WH-147 India 1977 293 29 292 48 1 288 000HD-2189 India 1979 1347 10 1264 49 609 000Sakha-69 Egypt 1980 1155 10 1134 45 471 000Lok-1 India 1981 653 16 649 39 2 094 000HD-2285 India 1983 2899 6 2694 59 1 137 000Kanchan Bangladesh 1983 418 22 416 46 613 000Palmiet South Africa 1983 1842 6 1789 49 398 000HUW-234 India 1984 2004 7 1829 62 3 387 000Marchouch Morocco 1984 656 12 642 41 365 000WH-283 India 1984 675 12 662 40 414 000HD-2329 India 1985 1804 8 1736 58 4 248 000HUW-206 India 1985 2684 5 2435 49 479 000Chakwal-86 Pakistan 1986 904 12 889 47 310 000Gonen Turkey 1986 786 13 777 45 491 000Granero INTA Argentina 1987 3005 7 2791 68 744 000Achtar Morocco 1988 1413 10 1344 57 512 000ProINTA Federal Argentina 1989 2539 6 2385 52 1 398 000Tilila Morocco 1990 2684 5 2435 49 365 000Falat Iran 1991 2684 5 2435 49 707 000Inqalab-91 Pakistan 1991 1644 10 1528 52 4 219 000Klein Cacique Argentina 1991 2423 6 2075 56 1 313 000ProINTA Imperial Argentina 1992 2690 6 2510 55 266 000ProINTA Quintal Argentina 1992 2245 5 2176 42 375 000Sonali India 1992 2292 4 2153 40 527 000WH-542 India 1992 3970 4 3506 52 651 000Klein Dragon Argentina 1993 4712 3 4119 52 357 000PBW-343/ India 1995 4502 4 3881 66 923 000PBW-373 (Attila)Average 1932 10 1781 50 1 062 000

Source: Smale et al. (2001). Calculated from data in Heisey et al. (in press) and CIMMYT’s Wheat Pedigree ManagementSystem.


landraces into the genetic background of itsmaterials. Figure 5 shows that for the morethan 1000 spring bread wheat varieties re-leased in developing countries from 1966to 1996, the number of distinct landraceancestors per pedigree has increased sig-nificantly.

This positive trend can be attributedpartly to efforts made by the CIMMYTwheat breeding programme and partly tothe use of these materials by breeders innational programmes. In developingcountries, CIMMYT-related wheats re-leased by national breeding programmeshave a larger number of differentlandraces in their pedigrees than wheatvarieties that do not incorporate CIMMYT-produced materials. Between 1965 and1990, developing countries (includingChina) released 151 wheats whose pedi-grees are known, but which have no iden-tifiable CIMMYT ancestry. The averagenumber of distinct landrace ancestors per

Table 3. Pedigree characteristics of minor spring bread wheats (a sample of varieties planted on less than0.25 million ha in developing world in 1997)

Variety Country Year No. of parental Percentage No. of No. of Areareleased combinations of parental ancestors different (ha)

in pedigree combinations in pedigree landraceused once ancestors

Ahgaf Yemen 1983 503 17 501 39 6000Alborz Iran 1978 1605 8 1534 61 <1000Annapurna-1 Nepal 1988 2684 5 2435 49 71 000Buck Guarani Uruguay 1993 6500 3 5366 59 7000Cham-6 Syria 1991 872 11 857 43 149 000Chane-CIAT Bolivia 1986 2684 5 2435 49 24 000CPAN-3004 India 1992 3536 5 3149 57 41 000EMBRAPA-21 Brazil 1993 3098 5 2714 54 <1000Faisalabad-85 Pakistan 1985 2838 6 2656 58 8000HP-1209 Egypt 1982 209 32 210 39 10 000HPW-42 India 1979 5202 3 4549 56 88 000IAPAR-28 Brazil 1988 2684 5 2435 49 35 000K.PAKA Kenya 1974 463 16 461 42 24 000Kozi Tanzania 1976 843 8 829 40 <1000Marmara-86 Turkey 1986 2539 6 2385 52 48 000Mbega Kenya 1993 1674 6 1607 43 3000Mbuni Kenya 1987 851 14 813 56 14 000Mbuni† Tanzania 1975 851 14 813 56 1000Mexivano-1481 Algeria 1979 85 52 86 29 1000Nata Zimbabwe 1989 2684 5 2435 49 2000Pasa Kenya 1989 4191 5 3865 72 8000Pavon Pakistan 1978 1668 7 1629 47 84 000ProINTA Oasis Argentina 1989 1026 11 1000 45 194 000ProINTA Puntal Argentina 1994 232 25 208 30 61 000Tota-63 Colombia 1963 96 41 97 24 1000WH-416 India 1990 685 15 678 51 6000Average 1935 13 1760 48 34 000

Source: Smale et al. (2001). Calculated from data in Heisey et al. (in press) and CIMMYT’s Wheat Pedigree ManagementSystem.† Mbuni appears twice because it is grown in two countries.

Fig. 5. Landrace ancestors in spring bread wheats released by developingcountries, 1965–1997. Source: Smale et al. (2001). Note: Calculations basedon data in Heisey et al. (in press) and pedigree information in CIMMYT WheatPedigree Management System. Data were available for 1162 (included here) ofthe 1749 spring bread wheats recorded (in the survey data) as released duringthese years. Coverage is less complete for China and for wheats released in thelast few years.


Breeding successes and diversityThe data presented in the tables and figures above demonstratethat germplasm with a diverse genetic background is continu-ally brought into national breeding programmes through theinternational wheat improvement system. One indicator of thesystem’s success is that its products now cover an estimated90% of the spring bread wheat area in developing countries (Fig.6). The remaining 10%, lumped together in the ‘traditional’category in Fig. 6, is sown to older tall varieties that were alsoproduced by breeding programmes, landraces and varieties ofunknown heritage. More than 80% of the spring bread wheats

released by national programmes inthe developing world between 1966and 1997 resulted from the interactionbetween CIMMYT and national wheatscientists.

In one sense it is true, therefore,that the ancient patterns of diversity—the diversity contained in the heteroge-neous landraces grown in the fields oftraditional farmers throughout theworld—have been replaced in largepart by the products of modern cropbreeding programmes. However,much potentially valuable diversitystill remains in the hands of farmersgrowing landraces across western Asiaand parts of Africa (for example, seeBrush and Meng 1998). These “patchesand islands of farming systems”(Brush 1995) still represent at least2.4 million ha. Considerable diversityis also stored in the wild and weedygrasses that contributed to the wheatgenome (Zohary et al. 1969).

Table 4. Coefficients of parentage and pedigree characteristics for spring bread wheat varieties grown in the developingworld, 1990 and 1997

All varieties grown on All varieties grown onmore than 0.25 million ha less than 0.25 million ha

1990 1997 1990 1997

Coefficient of parentageAverage 0.229 0.211† 0.255 0.252†

Area-weighted 0.175 0.172 0.194 0.201Minimum 0.019 0.028 0.042 0.042Maximum‡ 0.563 0.563 0.563 1Pedigree characteristicsAverage total number of parental 960 1932 1326 1906combinations per pedigreeAverage percent of parental 16 10 15 13combinations used once in pedigreeAverage number of different landrace 45 50 44 48ancestors per pedigree

Source: Smale et al. (2001).†Distributions of pairwise coefficients of diversity are different for varieties grown on more than 0.25 m ha and less than 0.25 m ha, at 1% level of statistical significance, with Wilcoxon test.‡Maximum in 1990 is 0.563, the coefficient of parentage for selections from the same cross.

Fig. 6. Area planted to modern and traditional spring bread wheat varieties, by region ofthe developing world, 1997. Source: Heisey et al. (in press). Note: ‘Modern’ varietiesare defined by their semidwarf character. ‘Traditional’ varieties include tail varietieswith pedigrees as well as landraces.

pedigree for these varieties was 19. Over the same period,developing countries released 999 CIMMYT-related wheats.The average number of distinct landrace ancestors per pedigreefor these varieties was 45, over twice as high.

Table 4 shows the average, area-weighted average and rangeof pairwise COPs among all spring bread wheats sown on morethan 0.25 million ha in 1990 and 1997, as well as for the variet-ies sown on less land in each year. The level of diversity appearshigh at this scale of analysis, since all average COPs are farbelow the level commonly used for sisters from the same cross(0.5625).


Furthermore, to say that 90% of the spring bread wheat areain developing countries is sown to CIMMYT-related modernvarieties does not imply that as a group they are uniform or thattheir uniformity has increased over time. Evidence assembled bya number of scientists in CIMMYT’s Wheat Program has demon-strated that as molecular genetic diversity and genealogical di-versity has increased among CIMMYT spring bread wheats overthe past 30 years, their performance with respect to yield poten-tial and yield stability also has improved (see results reported inSmale et al. 2001). In each decade, leading CIMMYT varietiesrequired smaller amounts of land and lower rates of nitrogenapplication to produce the same amount of grain. Resistance todisease and tolerance to heat and drought have improved. Care-ful statistical analysis of International Spring Wheat Yield Nurs-ery (ISWYN) trials confirms the performance of the internationalwheat breeding system in developing varieties that combine highyield potential and wide adaptation. Varieties developed by theinternational system “perform better than or on a par with theNARS [national agricultural research system] varieties in most ofthe major spring wheat environments” (Maredia et al. 1999).Because national programmes have crossed CIMMYT breedinglines with their own breeding lines, the varieties they releasedbased on these crosses are at least as diverse as the originalCIMMYT lines.

Economic impacts of international wheatbreedingBy any measure, public sector wheat breeding research, whichhas been based on free exchanges of genetic resources, hasgenerated great benefits for farmers and consumers through-out the world. For example, Byerlee and Traxler (1995) esti-mated that the annual benefits from wheat breeding for springbread wheat alone in the developing world were aroundUS$2.5 billion by the late 1980s. As noted, spring bread wheatis the most widely grown type of wheat in developing coun-tries. The precise economic benefits from international wheatbreeding efforts are difficult to ascertain, however, for manyreasons. For example, yield gains observed in farmers’ fieldsresult partly from advances in breeding and partly from im-provements in crop management practices, and it is oftendifficult to separate the two. Depending on how the credit foryield gains is distributed between these two sources of pro-ductivity growth, and depending on numerous other eco-nomic and technical assumptions, our preliminary rough esti-mates of the value of the current additional grain productionattributable to international wheat breeding range fromUS$2 billion to US$4 billion per year.

However the benefits of international public sector wheatbreeding programmes might be calculated, they can be balancedagainst annual costs. Annual wheat breeding investment byCIMMYT may be around US$10 million (1990 dollars) today,and has never been more than about US$14 million (1990 dol-lars). The International Center for Agricultural Research in theDry Areas (ICARDA) may invest about US$1 million in wheatbreeding annually. NARS investments in wheat breeding mayrange from about US$100 million to US$150 million (1990 dol-lars) today. Much but not all of the CIMMYT investment is in

spring bread wheat; about 70% of NARS breeding investment isin spring bread wheat (Heisey et al., in press).

Taking into account the time lag between breeding investmentand the benefits from that estimate, Byerlee and Traxler (1995)estimated a high rate of economic return to international wheatbreeding investment, and they forecast that future rates of returnwould be in the range of 35% under conservative assumptions.Evenson’s (2001) review of returns to research studies includes15 studies of returns to wheat research in general in individualdeveloping countries. Thirteen of these studies found signifi-cantly positive social profitability.

DiscussionWheat’s history as a domesticated crop spans more than 9000years. The spread of the crop began early and as a consequence,landraces that are the result of hundreds, even thousands ofyears of selection and evolution, arose in virtually all wheat-growing regions around the globe.

CIMMYT’s spring bread wheats have complex pedigrees anddraw upon this vast array of materials from both developed anddeveloping countries. CIMMYT’s breeding lines are distributed inlarge numbers, mainly to developing country national researchprogrammes, and now contribute significantly to the pedigrees ofthe majority of wheat varieties released in these countries. Thesewheats exhibit a broader genetic base, more stable yields andbetter adaptation to the conditions faced by resource-poor farm-ers than the ‘modern’ varieties of 20 or 30 years ago or thanpresent-day varieties that do not carry CIMMYT-developedgermplasm. The economic benefits provided by these wheats aresubstantial, particularly to developing countries.

While it is impossible to imagine a world without wheat, it isalmost as difficult to imagine wheat remaining a dominant cropin world agriculture without recourse to the vast genetic diversitynow found both in farmers’ fields and in genebanks. The historictriumph of wheat was facilitated and made possible by a condu-cive policy environment that encouraged the flow of geneticmaterials and supported public sector research with those mate-rials. The future of wheat—and of other crops—is inextricablylinked to that policy environment. Though the case of wheatdiffers from that of other crops in its evolutionary and breedinghistory, it is representative in some important respects of otherhighly bred crops of self-pollinating species.

The history of wheat’s development provides a chilling per-spective on current calls for restrictions on the exchange ofgermplasm of agricultural crops. One may ask how many peoplemight have been benefited, and how many would have lost orsuffered, had the restrictions being advocated today been in place10 years, 100 years, 1000 or 10 000 years ago—or had Daruma,the source of the two most important dwarfing genes, beenpatented! Have the many countries that have furnished wheatgenetic resources to breeding efforts benefited from those contri-butions? A cursory examination of modern bread wheat pedi-grees, the wide-scale use of these wheats, and the dependence ofnations and peoples on this crop, would seem to make the answerabundantly clear. Interdependence dictates cooperation. The caseof wheat shows us that enlightened self-interest should leadcountries to promote rather than restrict flows of germplasm.


Certainly, if policy-makers expect further progress to be madein international plant breeding, they will have to fashion laws andpolicies that allow institutes such as CIMMYT to acquire neededgenetic materials and freely distribute the products of their re-search. If such efforts are to be replaced by a plethora of nationalprogrammes in all crops, then governments will have to be pre-pared to devote considerably greater financial and human re-sources to plant breeding and acquisition of materials than theyseem prepared to provide today.

AcknowledgementsThe calculations and summaries reported here are based on thedata and work of numerous CIMMYT and national wheatprogramme scientists, drawing in particular on data from theCIMMYT Wheat International Nurseries, CIMMYT Wheat Pedi-gree Management System, and CIMMYT Global Surveys of WheatImpacts, 1990 and 1997.

ReferencesBrush, S.B. 1995. In situ conservation of landraces in centers of

crop diversity. Crop Sci. 35:346-354.Brush, S.B. and E. Meng. 1998. Farmers’ valuation and conserva-

tion of crop genetic resources. Genet. Resour. Crop Evol.45:139-150.

Byerlee, D. and G. Traxler. 1995. National and internationalwheat improvement research in the post-Green Revolutionperiod: Evolution and impacts. Am. J. Agric. Econ. 77(2):268-278.

Evenson, R.E. 2001. Economic impacts of agricultural researchand extension. In Handbook of Agricultural Economics (B.L.Gardner and G.C. Rausser, eds.). Elsevier, Amsterdam.

Falconer, D.S. 1981. Introduction to Quantitative Genetics.Longman, Essex, UK.

Feldman, M. 1976. Wheats. In Evolution of Crop Plants (N.W.Simmonds, ed.). Longman, London.

Harlan, J.R. 1971. Agricultural origins: Centers and noncenters.Science 174:469-474.

Harlan, J. 1977. The origins of cereal agriculture in the Old World.In Origins of Agriculture (C. Reed, ed.). Mouton, The Hague.

Harlan, J.R. 1992. Crops and Man. American Society ofAgronomy and Crop Science Society of America, Madison,WI.

Heisey, P.W., M.A. Lantican and H.J. Dubin. Forthcoming. As-sessing the Benefits of International Wheat Breeding Researchin the Developing World: The Global Wheat Impacts Study,1966-1997. International Maize and Wheat ImprovementCenter (CIMMYT), Mexico, D.F.

Heun, H., R. Schäfer-Pregi, D. Klawan, R. Castagna, M. Accerbi,B. Borghi and F. Salamini. 1997. Site of einkorn wheat domes-tication identified by DNA fingerprinting. Science 278:1312-1314.

Kawahara, T. and S. Taketa. 2000. Fixation of translocation 2A4B infers the monophyletic origin of Ethiopian tetraploidwheat. Theor. Appl. Genet. 101:705-710.

Maredia, M.K., R. Ward and D. Byerlee. 1999. Assessing poten-tial international transferability of wheat varieties. In TheGlobal Wheat Improvement System: Prospects for EnhancingEfficiency in the Presence of Spillovers (M.K. Maredia and D.Byerlee, eds.). CIMMYT Research Report No. 5. InternationalMaize and Wheat Improvement Center (CIMMYT), Mexico,D.F.

Skovmand, B. and I.H. DeLacy. 1999. Parentage of a historicalset of CIMMYT wheats. 1999 Annual Meeting Abstracts.American Society of Agronomy, Madison, Wisconsin.

Smale, M. 1996. Understanding Global Trends in the Use ofWheat Diversity and International Flows of Wheat GeneticResources. Economics Working Paper 96-02. InternationalMaize and Wheat Improvement Center (CIMMYT), Mexico,D.F.

Smale, M., M. Reynolds, M. Warburton, B. Skovmand, R.Trethowan, R.P. Singh, I. Ortiz-Monasterio, J. Crossa, M.Khairallah and I. Almanza-Pinzon. 2001. Dimensions of Di-versity in CIMMYT Bread Wheat from 1965 to 2000. Interna-tional Maize and Wheat Improvement Center (CIMMYT),Mexico, DF.

Souza, E., P.N. Fox, D. Byerlee and B. Skovmand. 1994. Springwheat diversity in irrigated areas of two developing coun-tries. Crop Sci. 29:595-601.

Zeven, A.C. and N. Zeven-Hissink. 1976. Genealogies of 14,000Wheat Varieties. Institute of Plant Breeding, AgriculturalUniversity, Wageningen.

Zohary, D. 1970. Centers of diversity and centers of origin. InGenetic Resources in Plants: Their Exploration and Conserva-tion (O.H. Frankel and E. Bennett, eds.). Blackwell, Oxford.

Zohary, D., J.R. Harlan and A. Vardi. 1969. The wild diploidprogenitors of wheat and their breeding value. Euphytica18:58-64.

Other sourcesCIMMYT Wheat International Nurseries dataCIMMYT Wheat Pedigree Management SystemCIMMYT Global Surveys of Wheat Impacts, 1990 and 1997


Variabilidad genética en el géneroXanthosoma en CubaMarilys Milián? , I. Sánchez, Magaly García, D. Guerra y Amparo CorralesInstituto de Investigaciones en Viandas Tropicales (INIVIT), Apdo. 6, Santo Domingo, Villa Clara, Cuba

ResumenVariabilidad genética en elgénero Xanthosoma en CubaEste trabajo se desarrolló en la colecciónde germoplasma del género Xanthosomaubicada en el INIVIT donde se conservan73 clones. De estos, 64 se caracterizaronmorfológicamente, con la utilización dela Lista de Descriptores del género encuestión (IBPGR, 1989; García, 1990; Mil-ián et al. 1993a) lo que permitió, además,definir que cuando se analizan todos losdescriptores no existen clones duplica-dos morfológicamente; pero si se anali-zan sólo los cualitativos más importantes,los clones Morada (1421) y Morada 3(1450) muestran duplicidad. La infor-mación procedente de la caracterizaciónque incluye todos los descriptores se pro-cesó utilizando métodos de estadísticamultivariada (análisis de conglomeradoso “cluster analysis”) para hacer una clas-ificación múltiple de los cultivares basadaen estos descriptores. Se pudo definir quela variabilidad genética en X. violaceum esmayor que en X. atrovirens y, a su vez,mayor en éstas que en X. caracu y X.sagittifolium, aunque las cuatro especiesno se diferencian claramente. El color dela carne es el carácter que más refleja lasdiferencias entre las especies, pero él solono es representativo de la variabilidad enla totalidad de los clones. También seconcluyó que la clasificación basada enuno o muy pocos caracteres morfológi-cos no refleja la verdadera variabilidadgenética dentro del género. Se recomien-da profundizar con análisis de marcado-res del DNA para realizar una clasificaciónmás acertada.

SummaryGenetic variability inXanthosoma genus in CubaThe Xanthosoma germplasm collection atINIVIT includes 73 clones, of which 64were characterized morphologically us-ing the Descriptors List for the genus(IBPGR 1989, García 1990, Milián et al.1993a). There were no duplicates in thecollection morphologically when all de-scriptors were analyzed, but based onqualitative descriptors the Morada (1421)and Morada 3 (1450) clones emerged asmorphologically similar for such descrip-tors. The data on characterization includ-ing all descriptors were processed usingmultivariate statistical methods (ClusterAnalysis) to carry out a descriptor-basedmultiple classification of cultivars. Thegenetic variability in X. violaceum is higherthan in X. atrovirens and in the latter it ishigher than in X. caracu and X.sagittifolium; although differences withinspecies were not evident. Flesh colour isthe most distinct and identifiable charac-ter but is not representative of the vari-ability in all clones. Classification basedon one or a few morphologic charactersdoes not show the true genetic variabil-ity within the genus. A recommendationis made to deepen the analysis usingDNA markers to produce a more de-tailed classification.

Key words: Cuba, cluster analysis,descriptors, Xanthosoma

ARTICLE

RésuméVariabilité génétique au seindu genre Xanthosoma à CubaLa collection de matériel génétique deXanthosoma de l’Institut de recherche surles denrées tropicales (INIVIT) com-prend 73 clones, dont 64 ont été car-actérisés morphologiquement en util-isant la liste des descripteurs du genre(IBPGR 1989, García 1990, Milián et al.1993a). L’analyse morphologique de lacollection, réalisée en prenant en comptel’ensemble des descripteurs, n’a pas misde doublons en évidence. Cependant, ense basant sur des descripteurs qualitatifs,les clones Morada (1421) et Morada 3(1450) sont apparus comme mor-phologiquement similaires. Les donnéessur la caractérisation comprenantl’ensemble des descripteurs ont étésoumises à une analyse statistique multi-variée (Cluster Analysis) afin de réaliserune classification des cultivars fondée surun grand nombre de descripteurs. Lavariabilité génétique de X. violaceum estplus importante que celle de X. atrovirens,elle-même plus grande que celle deX. caracu et X. sagittifolium. Toutefois, lesdifférences intraspécifiques ne sont pasnettes. La couleur de la chair est le critèredistinctif le plus aisément identifiable,mais il n’est pas représentatif de la vari-abilité inter-clonale. Une classificationbasée sur un seul ou un petit nombre decaractères morphologiques ne reflète pasla véritable variabilité au sein du genre. Ilest suggéré d’affiner l’analyse en utilisantdes séquences d’ADN comme mar-queurs afin d’obtenir une classificationplus détaillée.

IntroducciónLos recursos genéticos proporcionan al hombre la reserva másimportante para la obtención de productos alimenticios, indus-triales, medicinales, y otros. Por tanto constituyen la base de suexistencia.

En Cuba el género Xanthosoma es el de mayor importanciaentre las aráceas comestibles siendo de amplia distribución,adaptación y consumo. El Instituto de Investigaciones enViandas Tropicales (INIVIT) es el responsable a nivel nacionaldel manejo del germoplasma de este género cuya clasificaciónpor especies aún no alcanza la claridad suficiente.

H. León (1946) prefieren limitar la identificación taxonómicaal género de cada clon, teniendo en cuenta la confusión existentepara las posibles especies de cada género. Sin embargo, Cordero(1975) clasificó las variedades de yautía en cuatro especies: X.sagittifolium, X. atrovirens, X. violaceum y X. caracu, basado en la

forma y el tamaño de la lámina, la forma y el tamaño de loscormelos y el color de la pulpa.

Cordero (1986) añade que hay un entendido general de que laespecie más cultivada es la X. sagittifolium de pulpa o masablanca y que existen otras especies económicamente importantes.Por su parte García (1990) asegura la presencia en Cuba comocultivo comercial de clones de cuatro especies: X. sagittifolium(masa blanca), X. violaceum (masa rosada o morada), X. caracu(masa crema) y X. atrovirens (masa amarilla). Sobre la base deesta observación, Milián et al. (1993b) clasificaron 64 clones de lacolección cubana en las cuatro especies mencionadas. Estosautores coincidieron con los criterios de León (1987) quien sostieneque la clasificación de las especies y su descripción se hizo enejemplares de herbario. El propio autor reafirma que hasta que nose haga una comparación de estas entidades en vivo es más


conveniente considerarlas como una sola especie, talcomo se ha hecho con Colocasia esculenta. Al respecto,Giacometti y León (1994) también reconocen que noestá clara la posición taxonómica de las especies deXanthosoma cultivadas y que las características quelas distinguen no están bien definidas.

Analizando la caracterización morfológica de lacolección cubana de Xanthosoma spp., Milián et al.(1993b) observaron que existen grandes diferenciasentre clones de una misma especie y similitudesentre clones de especies diferentes, lo que confirma loplanteado en la literatura cuando se refiere que lataxonomía de Xanthosoma a este nivel es todavíaconfusa (Cordero 1986). Por estas razones, elpresente trabajo tiene como objetivo fundamentalutilizar métodos de análisis multivariado paraanalizar los agrupamientos que se forman cuando seprocesa la información procedente de dichacaracterización utilizando todos los descriptores.

Materiales y métodosEl trabajo se desarrolló en el Instituto deInvestigaciones en Viandas Tropicales (INIVIT).Luego de caracterizar morfológicamente 64 de losclones que forman parte de la colección nacional delgénero Xanthosoma (Milián et al. 1993b) con lautilización de la Lista de Descriptores del género(IBPGR 1989; García 1990; Milián et al.1993a) paradefinir la ausencia de clones duplicadosmorfológicamente cuando se analizan todos losdescriptores; y la presencia de duplicidad de losclones Morada (1421) y Morada 3 (1450) cuando serealizó el análisis teniendo en cuenta los caracterescualitativos más importantes, se procedió aprofundizar en el procesamiento de la información.Se utilizaron los clones clasificados, según Milián etal. (1993b), en las cuatro especies (X. sagittifolium, X.caracu, X. atrovirens y X. violaceum ) (Tabla 1).

Se emplearon métodos de estadísticamultivariada (análisis de conglomerados o “clusteranalysis”) para hacer una clasificación múltiple delos cultivares basada en sus descriptoresmorfológicos. A cada descriptor se le dio un rango enuna escala basado en el valor que toma cada unopara cada clon, se construyó una matriz de similitudentre los clones aplicando una distancia euclidiana.A partir de esta matriz se calculó cuáles son losindividuos que se agrupan por sus características.

Resultados y discusiónEl color de la carne está muy bien diferenciado parala mayoría de los clones. Es decir, estos clones noforman una gama continua, sino que toman valoresdiscretos. Basado en eso, fundamentalmente, se hanclasificado las malangas del género Xanthosoma encuatro especies. Se podría tomar otras característicaspara hacer esa clasificación, como el color del

Tabla 1. Clones de la colección cubana de Xanthosoma spp.utilizados y su clasificación por especie (Milián et al. 1993b)

No. de introducc. Clon Especie

1229 Amarilla Trinidad X. atrovirens1230 Riza X. caracu1259 Blanca Selección X. caracu1471 Blanca - 6 X. caracu1242 Amarilla Criolla -1 X. atrovirens1246 Viequera X. caracu1482 Blanca - 7 X. caracu1486 Blanca - 8 X. sagittifolium1262 Blanca Baracoa X. caracu1460 Morada Ceniza X. caracu1445 Blanca - 4 X. caracu1231 Blanca X. sagittifolium1427 Tricolor X. caracu1240 Macal Sport X. sagittifolium1267 Batabala Blanca X. caracu1457 Blanca - 5 X. caracu1422 Blanca - 1 X. caracu1436 Blanca - 3 X. caracu1241 Blanca Venegas X. sagittifolium1480 Chopo Amarillo X. atrovirens1475 De Seda X. atrovirens1232 Amarilla Criolla X. atrovirens1255 Amarilla Ceniza X. atrovirens1261 Amarilla Especial - 4 X. atrovirens1464 Amarilla - 2 X. atrovirens1238 Amarilla Especial X. atrovirens1235 Rosada X. violaceum1248 Blanca Pinar del Río X. caracu1237 Macal X. violaceum1258 Morada - 1726 X. violaceum1249 Morada - 1727 X. violaceum1256 Blanqui morada X. caracu1263 México - 8 X. violaceum1266 Cuarentena - 1 X. violaceum1247 Blanca Mutación X. caracu1251 México - 27 X. violaceum1253 México - 2 X. violaceum1234 Japonesa X. violaceum1250 México - 1 X. violaceum1445 Blanca - 4 X. atrovirens1478 Americana X. caracu1474 Blanca Morada X. caracu1490 Blanca - 9 X. caracu1485 Morada - 7 X. violaceum1233 Morada X. violaceum1428 Ceniza X. violaceum1243 Amarilla X. atrovirens1244 Stoupan X. sagittifolium1466 Encintada X. atrovirens1254 México - 3 X. violaceum1265 Cuarentena X. violaceum1257 Morada - 1 X. violaceum1254 México - 3 X. violaceum1421 Morada - 2 X. violaceum1450 Morada - 4 X. violaceum1435 Morada - 3 X. violaceum1446 Sergio Cuarentena X. violaceum1470 Morada - 5 X. violaceum1481 Morada - 6 X. violaceum1239 Javánica X. violaceum1264 Morada de México X. violaceum1236 Jardín X. violaceum1245 Belembe X. atrovirens1437 Picante X. caracu


pseudotallo. Este procedimiento genera cierta ambigüedadcuando la clasificación se basa en uno o muy pocos caracteres, loque se sustenta en el siguiente argumento que parecieraperfectamente razonable, a saber que aun cuando un caracterpuede ser poligénico, en él no van a estar involucrados más decuatro o cinco genes. Si se tiene en cuenta que en el genoma de unaplanta hay cientos de genes, una clasificación basada en unos

pocos de ellos no sería representativa aun cuando estos genesestén muy conservados en la especie.

Según los resultados que se muestran en la Figura 1, no existenagrupamientos bien definidos. Los clones que están agrupados enX. violaceum, donde existe la mayor variabilidad, presentan unagran distancia entre si y en muchos casos se solapan con X. caracuy ésta a su vez, con X. atrovirens y X. sagittifolium. Existen clones

Figura 1. Dendrograma que muestra la similitud genética entre los clones de las diferentes especies generado por elanálisis de conglomerados (“cluster analysis”) con la utilización de caracteres morfológicos cualitativos y cuantitativos.


que premanecen muy distanciados, fuera de grupo: el clonAmarilla Trinidad posee carne amarilla, pero sus peciolos sonmorados y no verdes como el resto de los clones de la especie X.atrovirens; el clon Belembe presenta hojas trifoliadas, carne concentro amarillo y borde blanco, a diferencia de los demás clones dela especie donde está clasificado (X. atrovirens); el clon Picantepresenta carne de color crema, sin embargo por el resto de suscaracterísticas (ausencia de cormelos, hojas con el ápice apuntandohacia arriba, nervaduras prominentes, etc.) que son muy diferentesde las del resto de los clones pero con una distancia apreciable delclon Belembe, se lo puede ubicar en la especie Alocasia macrorrhiza yno como X. caracu. Los clones Morada (1421) y Morada 3 (1450)muestran también similitud como resultado de este análisis.

Los resultados obtenidos indican que la taxonomía del géneroXanthosoma no está totalmente definida a nivel de especie, lo quecoincide con lo planteado por Cordero (1986). También León(1968) considera conveniente hablar de clones del género, antesque presentar aproximaciones a las especies de cada uno de losgéneros en razón de las deficiencias de la clasificación existente.

Conclusiones y recomendaciones● La variabilidad genética en Xanthosoma violaceum es mayorque en Xanthosoma atrovirens y, a su vez, mayor en ésta que enXanthosoma caracu y Xanthosoma sagittifolium.● El análisis realizado no diferencia claramente las especies.● El color de la carne es el carácter que más refleja las diferenciasentre las especies pero por si solo no es representativo de lavariabilidad en la totalidad de los clones.● La clasificación basada en uno o muy pocos caracteresmorfológicos no refleja la verdadera variabilidad genética dentrodel género.● Se recomienda profundizar en el estudio utilizando análisis demarcadores del DNA para realizar una clasificación más acertada.

ReferenciasCordero García, M. 1975. Origen, distribución y clasificación

botánica de la yautía. En: Curso de adiestramiento en elcultivo de la yautía. CATIE-CENDA.

Cordero García, M. 1986. Curso Nacional de Yautía. 90 p. FAO.García, M. 1990. Lista de descriptores de malanga Xanthosoma

spp. Conferencia mimeografiada. INIVIT Santo DomingoVilla Clara.

Giacometti, D. C. y J. León. 1994. Tannia, yautía (Xanthosomasagittifolium). En: J. E. Hernando Bermejo and J. León. Ne-glected crops: 1492 from a Different Perspective. Plant Pro-duction and Protection Series No. 26 FAO, Rome, Italy pp.253-258.

IBPGR 1989. Descriptors for Xanthosoma. International Board forPlant Genetic Resources, Rome.

Hermanos León. 1946. Flora de Cuba: 270-275.León, J. 1968. Fundamentos botánicos de los cultivos tropicales.

IICA. San José, Costa Rica, 445p. Colección Libros ymateriales educativos. IICA; No. 84.

León, J. 1987. Botánica de los cultivos tropicales. IICA.Milián J. M., I. Sánchez y M. García. 1993a. Observaciones sobre

la Lista de Descriptores para la caracterización de losrecursos genéticos en Xanthosoma spp. Instituto deInvestigaciones en Viandas Tropicales (INIVIT), SantoDomingo, Villa Clara, Cuba.

Milián J. M., I. Sánchez y M. García. 1993b. Caracterizaciónmorfológica de la colección cubana de malanga (Xanthosomaspp). Instituto de Investigaciones en Viandas Tropicales(INIVIT), Santo Domingo, Villa Clara, Cuba.


Computer tools for spatial analysis ofplant genetic resources data: 1. DIVA-GISR.J. Hijmans1? , L. Guarino2, M. Cruz1 and E. Rojas1

1 International Potato Center, Lima, Peru. Email: [email protected] International Plant Genetic Resources Institute, Americas Office, Cali, Colombia

SummaryComputer tools for spatialanalysis of plant geneticresources data: 1. DIVA-GISThe DIVA-GIS software allows analysisof genebank and herbarium databasesto elucidate genetic, ecological and geo-graphic patterns in the distribution ofcrops and wild species. It useful for scien-tists who cannot afford generic commer-cial GIS software, or do not have the timeto learn how to use it, and for others whorequire a GIS that is specifically designedfor genetic resources work. Coordinatedata are often absent from genebankdatabases, or if present are sometimesinaccurate. DIVA-GIS helps improvedata quality by assigning coordinates,using a large digital gazetteer. DIVA-GIScan also be used to check existing coordi-nates using overlays of the collection-siteand administrative boundary databases.Maps can then be made of the collectionsites. Analytical functions implementedin DIVA include mapping of richness anddiversity, distribution of useful traits andlocation of areas with complementarydiversity. DIVA can also extract climatedata for all terrestrial locations, whichcan be used to describe the environmentof collection sites.

Key words: Diversity,documentation, GIS, geographicdistribution, spatial analysis

ResumenSistemas de InformaciónGeográficas parainvestigación en RecursosFitogenéticos: 1. DIVA-GISEl programa DIVA-GIS apoya al análisisde bases de datos de bancos de genes yde herbario para hallar patrones genéti-cos, ecológicos y geográficos en la dis-tribución de especies silvestres y cultiva-das. Está dirigido a científicos que no ten-gan recursos para un Sistema de Infor-mación Geográfica (SIG) comercialgenérico o no tengan tiempo para apren-der a usar éstos y para todos los quequieren un SIG hecho específicamentepara trabajar en recursos genéticos. Paramuchas accesiones de bancos de genes,faltan las coordenadas geográficas, y aveces son inexactos. DIVA ayuda a mejo-rar la calidad de los datos por la asig-nación automática de coordenadas, us-ando un diccionario geográfico digital.DIVA también puede ser utilizado paraverificar coordenadas existentes hacien-do sobreposiciones de sitios de colecta ybases de datos de límites administrati-vos. Después de depurar los datos, sepueden hacer mapas de los sitios dondelas accesiones fueron colectadas y anali-zar los mismos. Las funciones de análisisimplementadas en DIVA incluyen el ma-peo de número de especies y otros índi-ces de diversidad; de la distribución decaracteres útiles; y de áreas con diver-sidad complementaria. DIVA tambiénpuede extraer datos del clima para cual-quier localidad en la tierra; estos datospueden ser usados para describir el me-dio ambiente de los sitios de colección.

ARTICLE

RésuméOutils informatiques pourl’analyse des données deressources génétiquesvegetales: 1. DIVA-GISLe logiciel DIVA-GIS permet l’analyse debases de données des banques de gènes etdes herbiers pour étudier les patronsgénétiques, écologiques et géographiquesdans la distribution des cultures et espècessauvages. Cet outil est dirigé a des cher-cheurs qui ne peuvent faire l’acquisitiond’un Système de InformationGéographique (SIG) générique commer-cial ou qui n’ont pas le temps d’apprendreá utiliser ces logiciels, et pour ceux quisouhaitent disposer d’un SIG adapté á larecherche de ressources génétiques. Lescoordonnées géographiques des bases dedonnées des banques de gènes sont sou-vent absentes, et parfois imprécises.DIVA-GIS aide à améliorer la qualité desdonnées par l’assignation avec un dic-tionnaire géographique digital. On peutaussi utiliser DIVA-GIS pour vérifier lescoordonnées avec des juxtapositions deslieux de collecte avec des cartes digitalesdes limites administratives. Après cettecorrection de données on peut réaliserdes cartes des localités de collection etanalyser ces données. Les fonctions ana-lytiques de DIVA contiennent des cartesde numéro d’espèces et d’autres indicesde diversité, de la distribution de car-actères importants, et des zones avec di-versité complémentaire. DIVA peut aussiextraire des données climatiques pourn’importe quel endroit sur terre; ces don-nées peuvent être utilisées pour décrirel’environnement des sites de collection.

IntroductionN.I. Vavilov developed his theory of centers of diversity and originof crops on the basis of numerous expeditions to collect cropgenetic resources and subsequent evaluation and characterizationwork in the Soviet Union. Vavilov’s work is a classic example of thedual role of collecting expeditions: to make genetic variation avail-able for use and also help elucidate genetic, ecological and geo-graphic patterns in the distribution of species (Bennett 1970).Analysis of such eco-geographic patterns can make considerablecontributions to several plant genetic resources research activities,including planning collecting programs, targeting genetic resourcesfor breeding programs, developing core collections, selecting anddesigning sites for in situ conservation and assessing the potentialimpact of the products arising from the use of plant geneticresources. For detailed overviews, see Guarino et al. (1999, 2001).

Mapping and spatial analysis of genebank data can be carriedout with off-the-shelf geographic information system (GIS) software.However, some of these software packages are too expensive forsmall programs or institutes and they do not provide specific optionsthat enable rapid and uncomplicated analysis of biological diversitydata. This paper is the first in a series describing GIS softwarespecifically designed to be used for spatial analysis of data associ-ated with genetic resources collections. The computer program de-scribed here, DIVA-GIS version 1.4, hereinafter called DIVA, wasdeveloped at the International Potato Center (CIP) in collaborationwith the International Plant Genetic Resources Institute (IPGRI), andwith support from the System-wide Genetic Resources Program(SGRP). It is available free of charge from the CIP website (http://gis.cip.cgiar.org). Base-map data (e.g. administrative boundaries,


no letters, with the sign indicating the hemisphere (+ = N or W, –= S or E) (e.g. –12.5700), hence only two variables are needed.Decimal degrees should be stored with 4 or 5 decimals. At theequator, one unit of the fourth decimal (0.0001 degrees) equalsabout 10 metres (less at other latitudes). That should be suffi-ciently precise unless a differential GPS (Global Positioning Sys-tem) is used during the fieldwork and precision at the meter levelis available. In those cases five digits would be better. To allowthe user to assess the accuracy of the coordinate data it would begood practice to document how these were obtained (e.g. whetherwith a GPS or read from maps).

Assign coordinatesCoordinate data are often absent from genebank databases, par-ticularly in older collections. For example, only 9% the accessionsof six major genebanks of the US Department of Agriculture havecoordinates (Steiner and Greene 1996). However, 50% of theaccessions have a locality description. That means that there isscope for assigning coordinates to at least another 41% of theaccessions (most accessions with coordinate data will also have a

locality description). This can be done by searching for the localitynames on maps or in gazetteers. A gazetteer is a list of names ofgeographic features and the coordinates of their locations. Fortu-nately, digital gazetteers are available.

DIVA can search for the coordinates of localities. The userspecifies an input file that should ideally have the following fieldsindicating where the accession was found: country, first andsecond administrative subdivision and up to two locality names.For both locality names the distance and direction from thecollecting site to that locality can be indicated. A digital gazetteer(the database of foreign geographic feature names from the USNational Imagery and Mapping Agency1) is then used to searchfor the locality and assign its coordinates to the accession.

Create shapefileHaving assigned coordinates to all records, the data can be

altitude) for use with DIVA are also provided for use with DIVA viathe Internet. Other software that can be used specifically with geneticresources (or biodiversity) data include Floramap (Jones et al. 2001)and Worldmap (Williams 1994).

DIVA can import genebank databases containing passport,characterization and evaluation data, using the latitude andlongitude fields. If latitude and longitude are not known, butlocality information is available (such as department, province,and place name), DIVA can help in assigning the most likelycoordinates. DIVA can also automatically check the accuracy ofcoordinate data.

When the data have been imported, completed, and checked forerrors, DIVA can map the locations where genebank samples werecollected. More interestingly, DIVA can also create analytical mapsfor use in developing plans and strategies for future collecting and insitu conservation activities. These include maps indicating the num-ber of observations, the number of distinct classes of observations,and the value of diversity indices for an array of grid cells. DIVA canalso provide estimates of the climate in the locations wheregermplasm was collected (or of any other terrestrial location).

The DIVA desktopMost of the DIVA screen is made up of a map and its associatedlegend (Fig. 1). A map is drawn using geo-referenced databasescalled themes. For example, a map of the world can be made up withthe following themes: altitude, national boundaries, main rivers andcapital cities. Each theme on the map is also listed in the legend.

To manage the content of the map and they way themes aredisplayed, three menus are available: File, Theme and View. TheFile menu has functions for file and project management, export-ing data and maps and printing. The Theme menu has functionsfor inspecting and managing individual themes. The View menuhas functions that allow managing the map (e.g. zooming in andout). These menus are described in more detail in the DIVAmanual (Hijmans et al. 2001).

There are three additional menus: Analysis, Tools and Help.The functions in the Analysis and Tools menus are discussed inmore detail below.

Using DIVA: toolsCoordinate data in genebank databases are frequently scarce andoccasionally inaccurate. This seriously complicates spatial analy-sis of genebank data and makes the results unreliable. However,there is much that can be done to improve the quality of the dataand DIVA can make this task easier. DIVA can assign coordi-nates to accessions that have a locality description but no coordi-nates, and can help verify the accuracy of accessions that do havecoordinates. These processes are described below.

Coordinate notationThe best system for the digital notation of geographic coordinatesis decimal degrees. The commonly used sexagemal system con-tains numbers, symbols and letters (e.g., 12°34’12’’S) and needsto be stored as text (which is prone to error), or as six separatenumerical variables. The decimal system only has a number, and

Fig. 1. The DIVA main window with a map of four themes.

1 http://164.214.2.59/gns/html/index.html


imported to create a ‘shapefile’—a file format that is used toshow data on a map. The user has only to specify a DBF file andthe two numerical fields in the file that contain the latitude andlongitude data. A DBF file is a database format that can be madeby many programs including Microsoft Excel® and Access®.

The advantage of using the shapefile format is that it is acommonly used format. It is the native format of the popular GISsoftware ArcView, and most other GIS programs can importshapefiles or export their data towards this format.

Check coordinatesThe first time a shapefile is made from a genebank database it islikely to have some gross errors, such as points falling in the sea oron the wrong continent. Such unlikely locations are easy to iden-tify, and often also easy to correct. However, there may also beother errors than cannot be so easily recognized.

DIVA’s Check Coordinates tool helps to identify such errors, usinga method described by Hijmans et al. (1999). By simultaneouslyquerying the accessions database and an administrative boundariesdatabase, a new (temporary) database is created. For each accession,this new database contains the location names according to thegenebank database and according to the administrative boundariesdatabase. These names should be the same, and any mismatchesprobably reflect errors (or changes in names or boundaries).

ExtractThe Extract tool assigns environmental data to points. This allowsfor so-called retro-classification: environmental characterization ofcollecting site after, rather than during, collecting (Greene et al.1999). This can be useful because traits can be related to ecologicalconditions at the places where the collections were made. Cur-rently, monthly mean data for mini-mum and maximum temperatureand precipitation are included. Thedata are extracted from a globalinterpolated climate database at a10-minute resolution.

Using DIVA: Dataanalysis with gridsOnce the accession database hasbeen mapped, DIVA can carry outvarious analyses. Most of theseanalyses are based on grids. Agrid divides the world into equal-sized cells, the size of which canbe changed by the user. A calcula-tion is then performed on each ofthe cells. For instance, the numberof observations (points) in eachcell can be calculated. The advan-tage of using grids rather than ar-eas such as countries or adminis-trative regions is that equal-sizedgrid cells can be compared moreobjectively. DIVA’s analytical fa-cilities are described below.

Number of ObservationsThe Number of Observations function allows the user to determine thenumber of observations in each grid cell. Points in the shapefile thatare not relevant for a given analysis may be excluded by (de)selectingthem on the basis of their value for a specific field in the database thatdescribes the points. The number of observations in each grid cell canbe determined by three methods: the ‘Simple’, ‘Inverse Distance-Weighted’ and ‘Circular Neighborhood’ methods.

In the ‘Simple’ method, points are simply assigned to the gridcell they fall in. Shortcomings of this method are that: a point thatis on a border between grid cells is arbitrarily assigned to one gridcell; the value of a point that falls within a grid cell is assigned tothat grid cell only, irrespective of the nearness of the point to othergrid cells; the results are sensitive to the arbitrary origin of the grid;and, uncertainty about the location of the point is not taken intoaccount. These shortcomings can largely be overcome with thecircular neighbourhood and inverse distance-weighted (IDW) tech-niques, as implemented in DIVA. When the circular neighborhoodoption is chosen, calculations are made for a circle with its center inthe middle of a grid cell and a specified radius. In the IDW method,inverse distance-weighted values are assigned to the four nearestgrid cells. For more details see Hijmans et al. (2001).

Number of Different ValuesThe Number of Different Values function calculates the numberof distinct classes of a certain variable that occurs in each gridcell. For example, if the input database consists of the loca-tions where different wild species were observed, the databasefield that indicates the species names can be selected and thenumber of different species per grid cell will be produced.Figure 2 shows an example of such an analysis.

Fig. 2. The Create Grid and Output Options windows, together with a main map windowshowing wild potato species richness in Peru.


Diversity IndicesA number of diversity indices can also be calculated for each gridcell. A variable (field) from the input database is selected (e.g.,species) for which a diversity index is calculated. The formulasfor all indices were taken from Magurran (1988), who provided adetailed description of their properties. The mathematical de-scriptions of the different diversity indices are given in Table 1.

Complementary Site SelectionThe Complementarity Site Selection (abbreviated to Complementarity)procedure aims at identifying sets of grid cells that are complemen-tary to each other, i.e. that capture a maximum amount of diver-sity in as few cells as possible. Instead of using richness, anadjustment can be made in which rare observations are given moreweight.

The procedure is based on the algorithm described by Rebelo(1994). The discussion below covers species, but any multi-statevariable could be used for the analysis. The procedure used is lessstraightforward than it might seem. Whereas the selection of thefirst cell is easy – the cell with highest species richness (or arandom choice between ties if there are any) – the choice of thenext cell(s) depends on the previously selected cells. This isbecause the species in the cell with the second highest number ofspecies may also be present in the first cell. This is a non-linearoptimization problem. Rebelo (1994) developed a procedure thatcalculates an approximate optimal solution, and this has beenimplemented in DIVA.

An iterative procedure is used. In each iteration the ‘value’ ofeach grid cell is calculated, based on the observations in that cell,and in relation to the observations in the cells already selected. Ifthere are two or more cells with the same value, one is selected atrandom. Hence, this procedure can lead to slightly differentresults each time it is run.

StatisticsIf a numerical variable is selected from the accession database,statistics can be calculated for that variable, for each grid cell. Thestatistics included are listed in Table 2.

Using DIVA: point-based data analysisAn alternative to the use of grids are ‘point-based’ approaches,such as used by the Spatial Intra-specific Diversity (SID) soft-ware described by Nelson et al. (1997) and now also implementedin DIVA. Diversity indices are calculated based on all observa-

tions lying within a user-defined circle about each point. Theresults are assigned to the location of the central observation andoutput to a database. The results can then be mapped again inDIVA.

With the Distribution Statistics function, statistics for eachunique value (class) of a multi-state variable can be calculated,for example, for each species in a database of wild relatives in agiven genepool. Currently there are two statistics (others arebeing added), number of observations and MaxD.

Freely available GIS databasesCountry-level GIS databases can be downloaded from http://gis.cip.cgiar.org. These databases can be used together with thegenetic resources data that are being mapped and analyzed.There are shapefiles with data on administrative boundaries,country boundaries and first and second level administrativesubdivisions for most countries. For all countries there are gridsavailable for altitude, land cover and population density. Thesedatabases are all taken from existing public domain databases.In most cases, however, the data in these databases are difficultto obtain, being available in huge (global) files stored in difficult-to-use formats and are therefore not available to the non-special-ist. Data from other sources can also be used in DIVA.

ConclusionsDIVA is easy to use stand-alone software and is available cost-free. It is thus a good starting point for people who work on plantgenetic resources but who do not have access to commercial GISsoftware. A particularly useful feature of the DIVA project is thatwe also provide many country-level GIS databases. The lack ofaccess to base map data is often perceived as another constraint toGIS adoption.

However, people who do have access to commercial GISsoftware programs may still want to use DIVA as it has func-tions specific to plant genetic resources that are not available, orare difficult to carry out, in other programs. This is facilitated bythe use of standard GIS data formats (the ESRI shapefile) inDIVA. There is also a function to ‘export’ the gridfiles to IDRISIformat and in forthcoming releases more formats for import andexport of data will be included.

DIVA is probably most useful for analysis of distributiondata covering larger areas, such as can be typically obtained fromgenebanks. An example in which DIVA was used extensively is astudy by Hijmans and Spooner (2001), who describe the geo-graphic distribution of wild potato species in North, Central andSouth America.

Table 1. Diversity indices

Index Formula

Margalef DMg=(S–1)/ln(N)Menhinick DMn=S/??? NShannon H´=–? pi ln piSimpson D=? (ni(ni–1)/N/(N–1))Brillouin HB=(ln N!–? ln ni!)/N

Where S is the number of unique classes per cell; N is thenumber of observations per cell; ni is the number of individualsin the i th class; and pi is the proportional abundance of the i thclass (=ni /N).

Table 2. Statistics

Min Minimum valueMax Maximum valueMean Mean valueSTD Standard deviationCV Coefficient of variationRange Difference between Max and MinRange/Mean Range divided by the meanMedian Median valueMode Mode value


The problems that may occur with the quality of the coordi-nate data of genebanks and how one can deal with these problemshave been discussed. However, it is equally important that atten-tion is given to the information quality of genebank databases.Genebank databases do not necessarily provide an unbiasedsample of existing diversity due to the way collections are made(Hijmans et al., 2000). The extent to which this influences theresults is partly dependent on the size of the grid cells chosen. Forexample, an area (grid cell) with high species richness may beassociated with low species richness due to a small number ofobservations in that area. However, this problem may diminish ifthe size of the grid cells is increased.

The next release of DIVA (Version 2), which is planned forlate 2001, will have more analytical functionality, particularly forgeographic analysis of molecular data, but also improved datahandling functions.

AcknowledgementsWe thank the Environmental Systems Research Institute, Inc.(ESRI), the International Potato Center (CIP), the InternationalPlant Genetic Resources Institute (IPGRI) and the System-widePlant Genetic Resources Program (SGRP) for support. AndrewJarvis and Karen Williams provided constructive feedback dur-ing the development of DIVA-GIS.

ReferencesBennett, E. 1970. Tactics of plant exploration. Pp. 157-179 in

Genetic resources in plants: Their exploration and conserva-tion (O.H. Frankel and E. Bennett, eds.). F.A. Davis Com-pany, Philadelphia.

Greene, S.L., T. Hart and A. Afonin. 1999. Using geographicinformation to acquire wild crop germplasm: II. Post collec-tion analysis. Crop Sci. 39:843-849.

Guarino, L., N. Maxted and M. Sawkins. 1999. Analysis of geo-referenced data and the conservation and use of plant geneticresources. Pp. 1-24 in Linking genetic resources and geogra-phy: emerging strategies for conserving and using cropbiodiversity (S.L. Greene and L. Guarino, eds.). AmericanSociety for Agronomy Special Publication 27. ASA, CSSA,and SSSA, Madison, Wisconsin.

Guarino, L., A. Jarvis, R.J. Hijmans and N. Maxted. 2001. Geo-graphic information systems (GIS) and the conservation anduse of plant genetic resources. Proceedings of the Interna-tional Conference on Science and Technology for ManagingPlant Genetic Diversity in the 21st Century, Kuala Lumpur,Malaysia (in press).

Hijmans, R.J. and D.M. Spooner. (2001).Geography of wild po-tato species. Am. J. Botany/88:2101-21120.

Hijmans, R.J., M. Schreuder, J. De la Cruz and L. Guarino. 1999.Using GIS to check coordinates of genebank accessions. Genet.Resour. Crop Evol. 46:291-296.

Hijmans, R.J., K.A. Garrett, Z. Huamán, D.P. Zhang, M.Schreuder and M. Bonierbale. 2000. Assessing the geo-graphic representativeness of genebank collections: thecase of Bolivian wild potatoes. Conserv. Biol. 14(6):1755-1765.

Hijmans, R.J., M. Cruz, E. Rojas and L. Guarino. 2001. DIVA-GIS, Version 1.4. A geographic information system for themanagement and analysis of genetic resources data. Manual.International Potato Center, Lima, Peru.

Jones P.G., L. Guarino and A. Jarvis. 2001. Computer tools forspatial analysis of plant genetic resources data: 2. FloraMap.Plant Genet. Resour. Newsletter (in press).

Magurran, A.E. 1988. Ecological diversity and its measurement.Princeton University Press, USA.

Nelson, A., G. LeClerc and M. Grum. 1997. The development ofan integrated Tcl/Tk and C interface to determine, visualise

and interrogate infraspecific bio-diversity. Internal document.CIAT, Cali, Colombia.

Rebelo, A.G. 1994. Iterative selection procedures: centres of ende-mism and optimal placement of reserves. Strelitzia 1:231-257

Steiner, J.J. and S.L. Greene. 1996. Proposed ecological descrip-tors and their utility for plant germplasm collections. CropSci. 36:439-451.

Williams, P.H., 1994. WORLDMAP priority areas for biodiversity.Version 3. Privately distributed, London.


Variability and its characterization in Indiancollections of blackgram [Vigna mungo (L.) Hepper]Sanjeev Gupta? , S.R. Gupta, H.K. Dikshit and R.A. SinghIndian Institute of Pulses Research, Kanpur 2080 24, India. Tel: 011-0512-572464; Fax: 011-0512- 572582,E mail: [email protected]

SummaryVariability and itscharacterization in Indiancollections of blackgram[Vigna mungo (L.) Hepper]Blackgram (Vigna mungo) is an impor-tant grain legume for South and SouthEast Asia. It is grown in various agroeco-logical conditions and cropping systemswith diverse cultural practices. No singleplant type could be appropriate for allthe production systems. A total of 670accessions of Indian national collectionswere grouped yield-wise and character-ized for several agromorphological de-scriptors. Considerable variation was re-corded for all descriptors except for seedsper pod and 100-seed weight. The high-est variability was recorded for the yieldfollowed by pods per plant and plantheight. Semi-determinate to indetermi-nate plant types of 50–60 cm height withlonger maturity are desirable for the wetseason. No single accession was foundhaving these attributes. Potential donorsfor early maturity trait were IPU 96-3,IPU 31-5, PLU 710, L25-7 and STY 2848;for YMV resistance, IC 27026, IC 106088,UL 2, HPU 4, HPU 188, STY 2848, UH 80-26, IP 99-127, PLU 62, PLU 158 and PLU227; for more number of pods per plant,KL 1, UH 81-44 and IPU 99-79 and formore number of seeds per pod, IC106088, HPU 193, HPU 2 and PLU 257were identified. Correlation was signifi-cant between yield and pods per plantbut non-significant between yield andseeds per pod. The number of pods perplant could be increased by increasingbranches per plant because a strong pos-itive correlation existed between them.The green-seeded types were more sus-ceptible to YMV than black-seeded ones.None of the green-seeded types was freefrom YMV incidence. Such germplasmcharacterization will provide valuable in-formation for further breeding work.

Key words: Agromorphologicaldescriptors, black gram, diseaseresistance, grain legume, India,Vigna mungo

ResumenLa variabilidad y sucaracterización en lascolecciones indias del frijolurd [Vigna mungo (L.) Hepper]El frijol urd (Vigna mungo) es una impor-tante legumbre de grano en Asia merid-ional y sudoriental. Crece en diversascondiciones agroecológicas y en sistemasagrícolas con diferentes prácticas de culti-vo. No hay un tipo único de planta quesea adecuado para todos los sistemas deproducción. Se clasificaron según su ren-dimiento un total de 670 accesiones decolecciones nacionales indias, que se car-acterizaron por varios descriptores agro-morfológicos. Se registró una consider-able variación para todos los descriptoresexcepto para las semillas por vaina y elpeso de 100 semillas. La mayor variabil-idad se encontró en el rendimiento,seguido por vainas por planta y altura dela planta. Tipos de planta semidetermi-nados o indeterminados de 50 a 60 cm dealtura con madurez más prolongada sondeseables para la estación húmeda. Nose encontró una sola accesión que reúnaestos atributos. Donantes potencialespara la característica de madurez tem-prana eran IPU 96-3, IPU 31-5, PLU 710,L25-7 y STY 2848; para la resistencia alvirus mosaico amarillo, IC 27026, IC106088, UL 2, HPU 4, HPU 188, STY 2848,UH 80-26, IP 99-127, PLU 62, PLU 158 yPLU 227; para mayor número de vainaspor planta, KL 1, UH 81-44 y IPU 99-79 ypara mayor número de semillas por vai-na, IC 106088, HPU 193, HPU 2 y PLU257. La correlación era apreciable entrerendimiento y vainas por planta, pero nosignificativa entre rendimiento y semill-as por vaina. El número de vainas porplanta podría elevarse aumentando lasramas por planta, ya que hay una fuertecorrelación positiva entre estas caracterís-ticas. Los tipos de semillas verdes eranmás susceptibles al virus mosaico amaril-lo que los de semillas negras. Ningunode los tipos de semilla verde estaban li-bres del ataque de ese virus. Esta carac-terización del germoplasma proporcio-nará una valiosa información para pro-seguir el trabajo de mejora genética.

RésuméEtude de la variabilité duharicot mungo (Vigna mungo (L.)Hepper) dans des collectionsindiennesLe haricot mungo (Vigna mungo) est unprotéagineux important pour l’Asie dusud et du sud-est. Cette espèce est cul-tivée dans différentes conditions agro-écologiques et selon diverses pratiquesculturales. Aucune variété ne pourrait àelle seule convenir à l’ensemble dessystèmes de production. Au total, 670 ac-cessions de collections nationales indi-ennes ont été regroupées en fonction deleur rendement et caractérisées selonplusieurs descripteurs agro-mor-phologiques. Des variations con-sidérables ont été observées pour tousles descripteurs, sauf en ce qui concernele nombre de graines par gousse et lepoids de 100 graines. Les paramètres lesplus variables sont le rendement, suividu nombre de gousses par plante et de lahauteur de la plante. Les types de plantessemi-déterminés à indéterminés, de 50 à60 cm de hauteur, à maturité plus tardiveconviennent pour la saison humide. Onn’a trouvé aucune accession unique pos-sédant l’ensemble de ces caractères. Lesdonneurs potentiels conférant une ma-turité précoce sont IPU 96-3, IPU 31-5,PLU 710, L25-7 et STY 2848 ; ceux con-férant la résistance au virus de la mo-saïque jaune (YMV), IC 27026, IC 106088,UL 2, HPU 4, HPU 188, STY 2848, UH 80-26, IP 99-127, PLU 62, PLU 158 et PLU227 ; ceux permettant d’obtenir un plusgrand nombre de gousses par plante, KL1, UH 81-44 et IPU 99-79 et, pour obtenirun plus grand nombre de graines pargousses, IC 106088, HPU 193, HPU 2 etPLU 257. Il existe une corrélation signifi-cative entre le rendement et le nombrede gousses par plante, mais pas entre lerendement et le nombre de graines pargousse. Le nombre de gousses par plan-te peut être amélioré si le nombre debranches par plante augmente, car il ex-iste une forte corrélation entre ces deuxparamètres. Les types à graines vertessont plus sensibles au YMV que ceux àgraines noires. On n’a observé aucuntype à graines vertes sans tracesd’infestation par YMV. Une telle car-actérisation du matériel génétiquefournira des informations utiles pour lapoursuite des schémas de sélection.

ARTICLE Plant Genetic Resources Newsletter, 2001, No. 127: 20 - 24


IntroductionBlackgram is an important grain legume for South and South EastAsia. Presently, it is cultivated in India, Pakistan, Myanmar,Bhutan, Bangladesh, Thailand, Malaysia, Philippines, Afghani-stan, Iran, Kenya, Malawi and the United States. India is the mostimportant producer of all. The average production in India is 1.28Mt annually from an area of 2.96 million ha (Anonymous 1998).The production in India and Bangladesh is almost entirely fordomestic use as food. In Thailand the production ranges between80 000 and 99 000 t annually, and is mainly exported to Japan forbean sprouts. In Japan blackgram is preferred to green gram(Vigna radiata) for bean sprouts for its longer shelf life. Blackgramis a promising legume crop of South and South East Asia.

Blackgram is basically a tropical crop but it is grown in bothwinter and summer in India. It is primarily intercropped withsugarcane, cotton, groundnut, sorghum and pigeonpea during thewet season, monocropped on residual moisture in winter (in ricefallow) and in spring/summer season in between the two maincrops. Since the crop is grown in various agroecological conditionsand cropping systems with diverse cultural practices, no singleplant type is appropriate for all production systems (Singh 1997).This calls for an extensive survey of existing germplasm collectionsfor potential utilization in development of appropriate plant typefor the various cropping systems in tropical Asia.

The Asian Vegetable Research and Development Centre(AVRDC), Taiwan, maintains a collection of nearly 200 acces-sions, while the National Bureau of Plant Genetic Resources(NBPGR), New Delhi, India maintains about 2100 accessions ofwhich the Indian Institute of Pulses Research holds 829 activecollections. The Indian gene centre has been considered an impor-tant gene conservation centre for Vigna species (Arora 1988).Considerable diversity is available in Indian collections (Singh etal. 1991; Singh and Shukla 1994; Nautiyal and Shukla 1999) butits exploitation is not appropriate owing to lack of characteriza-tion and classification. The detailed evaluation of germplasmwill help identify accessions of potential relevance for improve-ment programmes and encourage and improve the usefulnessand utilization of germplasm collections.

Materials and methodsThe blackgram active germplasm collection comprises 829 acces-sions of which 24 are cultivars, 576 advanced breeding lines and 229are landraces. Of these, 670 accessions were evaluated for 2 years(1997 and 1999) at the Indian Institute of Pulses Research, Kanpur(26.28°N and 80.21°E), India, in a single-replicate augmented designwith three checks, namely T 9, DPU 88-31 and ‘Barabanki Local’after every 10th row. Each plot consisted of a single row 4 m long.Rows were spaced 30 cm apart and interplant distance was 10 cm.The ‘Barabanki Local’, a highly susceptible cultivar for Yellow Mo-saic Virus (YMV), was used as indicator cum infector and was alsosown all around the field to increase the YMV incidence. T 9, anational check, and DPU 88-31 an improved cultivar, were used tocompare the results. Screening for YMV disease reaction at thepresent test location was appropriate as it has been considered thehot spot for YMV incidence (Singh and Gurha 1994).

Accessions were evaluated for 18 descriptors, including plantcharacteristics, seed characteristics and disease reaction, following

the IBPGR (1992) Vigna descriptor guidelines. These 18 descriptorsbelonged to three data types: binary (5), ordinal (6) and numerical(7). The data on numerical descriptors were analyzed in aug-mented design to obtain adjusted values of means as suggested byFedrer and Raghava Rao (1975) and further elaborated by Paterson(1985). These data were used to create a database in standardformat with the help of ‘DIPVIEW ‘ software package. Correlationcoefficients between yield and related parameters were determinedusing M’STAT C statistical programme (Anonymous 1988).

Results and discussionIn the existing collection, considerable variation was present forall the important descriptors except for seeds per pod and 100-seed weight. This confirms the earlier results from the work ofAcharya et al. (1993) and Singh and Satyanarayana (1994). Thehighest variability, as indicated by coefficient of variation, wasrecorded for the yield (59.61%) followed by pods per plant(52.47%) and plant height (39.43%). The data are not in agree-ment with previous work of Kasundra et al. (1995) who observedlow variability in pods per plant and plant height in 42 advancedbreeding lines. This could be because the present study includeda relatively large number of accessions in which the pods perplant and plant height showed considerable variation. Verma andKatna (1998) reported a large variability in yield tested in bothmonocrop and intercrop conditions.

All the 670 germplasm lines were evaluated for 18 descrip-tors related to plant characteristics, seed characteristics and dis-ease, insect and pest reaction. Each descriptor was also assessedwith regard to yield potentials (Table 1). Following is a detaileddiscussion of these results.

Plant characteristicsThe crop is basically a tropical one, but it is grown in both winter andsummer in India. It is also grown in the dry season in areas of highrainfall. Semi-determinate to indeterminate plant types of 50-60 cmheight with longer maturity (90–120 days) are desirable for the wetseason (Saxena and Yadav 1975). No single accession showed theseattributes. Therefore genetic enhancement is needed by combiningthese traits. A plant type combining determinate growth habit, 30 cmplant height and early maturity (60–90 days) will be appropriate forspring, summer and winter seasons. Only one accession, UP 83-2,showed all these traits but it was characterized by low seed yield. Onthe basis of evaluation for 2 years, the utilization of IPU 96-3, IPU31-5, PLU 710, L25-7 and STY 2848—the potential donors for earlymaturity—was suggested to develop suitable cultivars for sum-mer/spring cultivation. Early maturing types were shorter becauseof fewer nodes, had a smaller pod-bearing length and tend to matureafter the first flush of flowering. Satyanarayana et al. (1989) sug-gested that the fruiting zone could be increased by attemptingcrosses between determinate and indeterminate types and selectingplants with greater pod-bearing length.

Seed characteristicsOn the basis of seed colour and other characteristics, blackgramhas been grouped under two main types: (1) var. mungo, withlarge black seeds and early maturity; and (2) var. viridis, withsmaller greenish seeds with late maturity (Rao and Jana 1976).


The green-seeded types are locally known as ‘Katikaihia’ urd andgenerally grown as mixed crop with sorghum, pigeonpea andcotton. This type has been increasingly popular among consum-ers of certain areas of the country. It could also be exploited forbean sprouts but the delayed maturity and susceptibility to YMVhas often restricted its production (Shanmugam and Rangaswami1984). Therefore, there is a need to develop green-seededblackgram with early maturity, YMV resistance and determinategrowth habit. The existing collections contain 97 green-seededtypes of which all are medium to late maturing types, 64% aresusceptible to YMV and only 16 accessions are of erect growth

habit (Fig. 1). All these 16 erect types had low productive poten-tial and nine of them were highly susceptible to YMV.

Plants that produce more pods per plant along with moreseeds per pod would be desirable. In the present collection, 12accessions had more pods (?100 pods per plant) containing moreseeds (6–9 seeds per pod). Correlation coefficients were com-puted between these characters and yield (Table 2). Significantcorrelation (r=0.169**) was recorded between yield and pods perplant while it was non-significant between yield and seeds perpod. Singh et al. (1994), Sarkar et al. (1984), Wanjari (1988) andSood and Garton (1994) also discussed similar relationships.

Table 1. Descriptors and yield distribution of 670 germplasm accessions of blackgram

Plant descriptor Range in expression No. accessions of different yield potential

Low Medium High

Growth habit Erect 188 14 2Semi erect 242 27 3Prostrate 157 35 2

Growth pattern Determinate 299 24 2Indeterminate 288 52 5

Leaf shape Deltoid 220 39 3Ovate 363 37 4Others 4 0 0

Petiole pubescence Glabrous 585 76 7Pubescent 2 0 0

Days to flowering <39 days 10 2 040–50 days 477 59 7>50 days 100 15 0

Pods per plant < 60 474 46 461–120 100 27 2>120 7 3 0

Pod pubescence Glabrous 471 70 6Pubescent 116 6 1

Mature pod colour Brown 85 5 0Black 502 71 7

Days to pod maturity Early 9 8 0Medium 430 45 7Late 149 23 0

No. of seeds per pod 3–5 9 1 16–8 577 75 79–10 2 0 0

Seed colour Black 465 64 5Green 87 8 2Brown 35 4 0

100-seed weight (g) 2–3 141 13 14–6 447 63 6

No. of primary branches 1–2 106 4 03–5 462 70 76–8 19 1 0

Plant height (cm) 34–95 251 20 296–157 218 31 3158–219 105 21 2220–300 14 4 0

Peduncle colour Green 265 68 4Purple 151 4 1Others 171 4 2

Peduncle length Short 232 62 3Intermediate 240 9 2Long 115 5 2

Yellow Mosaic Virus susceptibility Low 205 42 3Moderate 366 32 4High 16 2 0


3). Plants with pods of 6–8 seeds and an average 100-seed weight?4 g are generally preferred. But these traits were not correlated.However, in the existing collection 499 accessions were found tohave the 6–8 seeds per pod with an average 100-seed weight of?4 g. Out of these, five accessions (IC 106088, UH 81-7, UH 82-42, UH 82-29 and IPU 99-102) were highly productive also.These can be exploited in further improvement programmes.

Disease, insect and pest reactionsYellow Mosaic Virus (YMV) is the most serious and widespreaddisease of blackgram and has posed a threat to its cultivation intropical Asia. Disease reaction after 75 days of sowing wasrecorded among all the accessions using a mean score on a 1–9scale. The susceptible check developed >7 grade of yellow mosaicdisease. Out of 670 lines, 250 lines showed a low susceptibility toYMV. During 1999, 25 lines (IC 106088, UL 2, HPU 4, HPU 188,HPU 227, DUS 34, K 116-86, PDU 110, JU 2, CN 28-1, CN 33-1,UH 80-26, UH 82-21, UH 82-42, UH82-50, UH 86-54, PLU 24,PLU 79 A, PLU 98, PLU 431, PLU 707, PLU 1049, PLU 1146 andPLU 1174) exhibited resistant reaction against YMV. After 2years of evaluation, IC 27026, IC 106088, UL 2, HPU 4, HPU188,STY 2848, UH 80-26, IP 99-127, PLU 62, PLU 158 and PLU227 were identified as potential donors for YMV resistance. Itwas also observed that the green-seeded types were generallymore susceptible to YMV than black-seeded types. In the existing

Table 3. Potential donors for different traits identifiedin 670 accessions of blackgram

Trait Accessions

Earliness IPU 96–3,IPU 31–5, PLU 710,(<90 days to maturity) L25–7, STY 2848Large seed size PLU 1829, HPU 2, IPU 96–10,(>4 g/100-seed weight) PLU 63Seeds/pod (>8 per pod) IC 106088,HPU 193, HPU 2,

PLU 257Pods per plant KL 1, UH 81–44, IPU 99–79(>130 pods)YMV resistance IC27026, IC 106088, UL 2, HPU(Mean score on 1–9 4, HPU 188, STY 2848, UH 80–scale) 26,IP 99–127, PLU 62,PLU 158,

PLU 227Thrips tolerance PLU 59, UH 86–40, Sym 2115,(Mean minimum H 11population comparedwith check)White fly tolerance PLU 707, PLU 158, UH 86–40(Mean minimumpopulation comparedwith check)

Fig. 1. Distribution of blackgram accessions according toYMV susceptibility, growth habit and maturity period with inblack, green and brown seeded types.

Table 2. Correlation coefficients between some yield-contributing traits in blackgram

Character 1 2 3 4 5 6 7

Days to flowering 1.00Plant height 0.249** 1.00Primary branches 0.127** 0.015 1.00Seeds/pod 0.036 –0.097* 0.016 1.00Pods/plant 0.116** 0.039 0.379** 0.012 1.00Seed weight –0.106* –0.043 –0.029 0.011 –0.145** 1.00Yield –0.052 0.046 0.124** 0.032 0.169** –0.045 1.00

** = Significant at 1 % level of probability

Seeds per pod and pods per plant were also not correlated witheach other. The number of pods per plant can be increased byincreasing branches per plant because a strong positive correla-tion (r=0.379**) existed between them. On the basis of detailedevaluation for 2 years, KL 1, UH 81-44 and IPU 99-79 wereidentified as potential donors for greater number of pods perplant (?150 pods) and IC 106088, HPU 193, HPU 2 and PLU 257were found to have more seeds per pod (?8 seeds per pod) (Table


stock nearly 64% green-seeded accessions were moderately tohighly susceptible, while only 47% black-seeded accessions weresusceptible to YMV (Fig. 1). None of the green-seeded types wasfree from YMV incidence.

Blackgram crops are attacked by more than two dozeninsect pests (Nayar et al. 1976). Among them, white fly andthrips cause the most serious damage. Mean number of white-flies were minimum in entries PLU 707 (1.3/m), DPU 88-31(1.5), PLU 158 (2.3), UH 86- 40 (2.5), Sympodial 2115 (3.3)and PLU 1146 (3.5) compared with the mean population oncheck variety ‘Barabanki Local’ (24.3/m). Entries with fewerthrips in flowers were PLU 59 (6.9), UH 86-40 (10.0), Sympo-dial 2115 (10.6) and H11 (11.7/10 flowers) compared withthe check (29.0/10 flowers). Entries Sympodial and UH 86-40showed less infestation of both these insects.

AcknowledgementsThe authors wish to acknowledge Dr P.N. Mathur, IPGRI’s Associ-ate Coordinator for South Asia, based in New Delhi, for his technicalguidance in database management and information retrieval using‘DIPVIEW ‘software package.

ReferencesAcharya, S., S.K. Gupta and B.S. Jamwal. 1993. Evaluation of

local germplasm of Urdbean. Indian J. Pulses Res. 6:198-199.Anonymous. 1988. Microcomputer Statistical Programme. Michi-

gan State Univ.Anonymous. 1998. Area and Production of Principal Crops.

Directorate of Economics and statistics, Department of Agri-culture & Cooperation, Ministry of Agriculture, Govt. ofIndia, New Delhi.

Arora, R.K. 1988. The Indian gene centre – Priorities and Pros-pects for collection. Pp. 66-75 in Plant Genetic Resources –Indian Perspectives (R.S. Paroda, R.K. Arora and K.P.S.Chandel, eds.). NBPGR, New Delhi.

Fedrer, W.T. and D. Raghava Rao. 1975. On augmented designs.Biometrics. 31:29-35.

IBPGR. 1992. Descriptor for Vigna sp. International Plant GeneticResources Institute, Rome, Italy.

Kasundra, J.K., K.V. Pethani and K.B. Kathiria. 1995. Studies ongenetic variability, correlation and path analysis in Urdbean(V. mungo L.). Indian J. Pulses Res. 8:113-118.

Nautiyal, M.K. and Arvind Shukla. 1999. Classification ofUrdbean germplasm from diversity zones of U.P. Indian J.Plant Genet. Resour. 12:26-41.

Nayar, K.K., T.N. Ananthakrishanan and B.V. David. 1976. Gen-eral and Applied Entomology. Tata McGraw Hill Pub. Com-pany Ltd., New Delhi, India, pp. 498-499.

Paterson, R.G. 1985. Augmented design for preliminary yieldtraits. Arachis, Barley & Triticale Newsletter (ICARDA) 4:27-32.

Rao, S.A. and M.K. Jana. 1976. Alteration of seed charactersticsin Blackgram. Indian J. Agric. Sci. 44:657-660.

Sarkar, A., M.A. Wahab, M.A. Daza, S. Mohamad and K. Begum.1984. Genetic variability and interrelation in Blackgram.Bangladesh J. Agric. Res. 9:108-111.

Satyanarayana, A., P. Sreenath, Y.K. Rao and N.V. Naidu. 1989.Inheritance of main stem bearing habit of Blackgram. J. Res.Andhra Pradesh Agric. Univ. 17:185-187.

Saxena, M.C. and D.S. Yadav. 1975. Multiple cropping with shortduration pulses. Indian J. Genet. 35:194-208.

Shanmugam, A.S. and S.R. Sree Rangaswami. 1984. Geneticdifferentiation between green and blackgram. Genetic Agraria38:149-160.

Singh, C. and A. Shukla. 1994. Classification of germplasm inurdbean. Indian J. Pulses Res. 7:112-115.

Singh, D.P. 1997. Tailoring the plant type in pulse crops. PlantBreed. Abstr. 67:1213-1220.

Singh, D.P. and A. Satyanarayana. 1994. Creating higher ge-netic yield potential in Urdbean. P. 6 in International Sym-posium on Pulses Research, 2-6 April 1994, New Delhi,India.

Singh, D.P., H.P. Singh and B.L. Sharma. 1991. Evaluation ofBlackgram (Vigna mungo (L.) Hepper) germplasm. NarendraDeva. J. Agric. Res. 6:132-137.

Singh, R.A. and S.N. Gurha. 1994. Influence of cropping seasonon the incidence of YMV in Mungbean genotypes. Indian J.Pulses Res. 7:206-208.

Singh, N., H.S. Yadav and R.P.S. Dhakarey. 1994. Factors influ-encing seed yield in early interspecific generation of Vigna.Pakistan J. Sci. Indust. Res. 37:54-570

Sood, B.C. and S.L. Garton. 1994. Variability and character asso-ciation in Urdbean. Indian J. Pulses Res. 7:56-58.

Verma, S. and G. Katna. 1998. Variability studies in Urdbean forintercrop and monocrop conditionns. Indian J. Pulses Res.11:33-37.

Wanjari, K.R. 1988. Variability and character association inBlackgram. Indian J. Agric. Sci. 58:48-51.


The appearance of chlorophyll defects in cerealsduring regeneration of genebank accessionsW. Schliephake, M. Grau and A. Börner?Institut für Pflanzengenetik und Kulturpflanzenforschung (IPK), Corrensstrasse 3, D-06466 Gatersleben, Germany.Tel: +49 39482 5229; Fax: +49 39482 5155; Email: [email protected]

SummaryThe appearance of chlorophylldefects in cereals duringregeneration of genebankaccessionsOver recent years occasional chlorophylldefects have been observed during seedregeneration of cereals in the Gatersle-ben genebank. The defects appeared onthe leaves as white-green stripes. Thereason for this is unknown. The analysisof all available data recorded since 1993did not point to a specific category ofaccessions concerned and it indicated amore or less coincidental appearance ofthis phenomenon. Field and growthchamber experiments were performedin order to define reasons for the emer-gence of chlorophyll-damaged plants. Itwas shown, that the chlorophyll defectsare reproducible testing seeds from thesame storage container in different ex-periments (early spring sowing, summersowing, growth chamber). This indicatesthat environmental conditions like tem-perature or soil are not responsible forthe appearance of the symptoms. Muta-tions after long-term storage as a possi-ble main reason could also be excluded,because, with one exception, the freshlyharvested progenies of damaged plotswere not affected. Although the reasonfor the chlorophyll defects cannot be fi-nally clarified, it is shown that the suscep-tibility of the genebank accessions de-pends on the genotype interacting withthe storage time.

Key words: Chlorophyll damage,cereals, long-term storage, seedregeneration

ResumenAparición de defectosclorofílicos en los cerealesdurante la regeneración deaccesiones en bancos de genesEn los últimos años se han observadodefectos clorofílicos ocasionales durantela regeneración de semillas de cerealesen el banco de genes de Gatersleben. Losdefectos aparecían en las hojas como ra-yas blancas-verdes. Se desconoce larazón. El análisis de todos los datos regis-trados desde 1993 no ha singularizadouna categoría específica de accesionesafectadas y ha revelado una apariciónmás o menos coincidente de este fenó-meno. Se han hecho experimentos en elcampo y en cámara de cultivo para pre-cisar por qué aparecen plantas con dañosclorofílicos. Se constató que los defectosclorofílicos son reproducibles. Con semi-llas de prueba del mismo contenedor dealmacén en diferentes experimentos (si-embra temprana de primavera, siembrade verano, cámara de cultivo) se vio quelas condiciones ambientales como tem-peratura o suelo no influían en la apar-ición de los síntomas. También podíanexcluirse como posible razón principallas mutaciones tras un prolongado al-macenamiento, porque, con una ex-cepción, las descendencias nuevas deparcelas dañadas no estaban afectadas.Aunque finalmente no pueda aclararse larazón de los defectos clorofílicos, resultaque la susceptibilidad de las accesiones enbancos de genes depende de la interac-ción de genotipos con el tiempo de al-macenamiento.

RésuméApparition d’anomalieschlorophylliennes dans descéréales durant larégénération d’accessionsd’une banque de gènesAu cours des dernières années, desanomalies chlorophylliennes occasion-nelles ont été observées lors de larégénération de semences de céréales dela banque de gènes de Gatersleben. Cesanomalies se manifestent au niveau desfeuilles par des stries blanc - vert. Leurorigine est, pour l’heure, inconnue.L’analyse de l’ensemble des données dis-ponibles collectées depuis 1993 ne per-met pas de l’attribuer à une catégoriespécifique d’accessions et fait penser quele phénomène se manifeste de manièrealéatoire. Des essais en champ et enchambre de culture ont été réalisés afinde déterminer les causes de l’apparitionde plantes présentant des anomalies chlo-rophylliennes. On a montré que cesanomalies étaient reproductibles. Dif-férents essais réalisés sur les semencesissues d’un même lot de stockage (semisen début du printemps, semis d’été,chambre de culture) indiquent quel’apparition de ces symptômes n’est pasattribuable aux conditions de milieu tellesque la température ou le type de sol. Il estégalement exclu que des mutationssoient apparues après une longue péri-ode de stockage car, à une exception près,la descendance des plantes présentantdes anomalies n’est pas affectée. Bien quel’origine des anomalies chlorophylli-ennes n’ait pu être élucidée, on a montréque la sensibilité des accessions de labanque de gènes dépendait du génotypeet de la durée de stockage.


IntroductionCurrently more than 40 000 accessions of cereals are maintained inthe Gatersleben genebank (IPK Annual Report, 1999). The regen-eration of genebank material in the field is labour intensive andgenetic integrity may be endangered by different types of contami-nation, e.g. foreign pollen during fertilization, gene mutation, ge-netic drift, genetic shift or imperfect handling during harvest.Therefore, the intervals between regeneration cycles should bemaximized. This becomes possible by storing the seeds underdefined low temperature and humidity. Surveys of seed ger-minability after long-term cold storage showed that high ger-minability is maintained for a period of more than 15 years underthe Gatersleben storage conditions (Specht et al. 1997, 1998).

About 2500 accessions of grain are grown in the field toregenerate the seeds every year. During regeneration chlorophylldefects have been observed on the leaves of cereal plants in severalseasons. Different numbers of plants per plot of various acces-sions showed typical symptoms. The first leaves of the plantsshow more or less white-green stripes (Fig. 1). The leaf areawithout chlorophyll varies considerable and sometimes plantsdie. The defects become partly overgrown in further development.

Chlorophyll defects are generally known to appear if plantsare grown under conditions of nutrient deficiency. Chlorotic leavesmay be the result of low amounts of nitrogen, copper or iron in thesoil. The plants usually regrow and lose the symptoms if the


lacking nutrients are added. Chlorophyll damage may be causedgenetically as well as by environmental stress. Many differentmutants were described in cereals like wheat (McIntosh et al.1998), barley (Lundqvist et al. 1997) or rye (Melz 1989). Thegenetics of most of these mutants is well understood.

A high number of spring barley accessions was affected in theGatersleben seed regeneration field plots in 1999; this was thereason for the investigations described here. The following fourquestions needed to be addressed:(i) Which accessions show chlorophyll defects and are they re-lated to each other?(ii) Is there a relationship between long-term storage and chloro-phyll damage?(iii) Are environmental conditions or pest and disease agents apossible cause?(iv) Will the defects be reproducible by testing the progenies ofdamaged samples?

Material and methodsAnalysis of documented chlorophyll defects ofcereals in the Gatersleben genebankThe first step was to analyse the documented examples of chloro-

phyll defects recorded during the last 6 years with the aim ofidentifying communities. The examination covered the followingpoints:● which accessions were affected and how often;● the peculiarities (soil and growth conditions, weather) duringthe seasons of regeneration;● the origin of affected accessions;● whether the time of storage had any influence.

Field experiment to test the reproducibility of thechlorophyll defectsBecause spring barley (Hordeum vulgare L.) was most frequentlydamaged, this species was chosen for the field experiments carriedout in July/Aug 1999. The objective was to find out whether thechlorophyll defects detected earlier could be reproduced. In totalnine accessions were tested. Three accessions (HOR7766, HOR7884,HOR7892) were known to show the defect after sowing in the earlyspring of 1999. Seeds were taken from the same storage containerfrom the cold store as for the spring sowing. A further threesamples including (HOR1735, HOR1743, HOR1766) had exhib-ited the symptoms in spring 1993. Because of a low yield in thatparticular year the accessions were grown again in 1994. However,on that occasion, they showed no symptoms. Only seeds from the1994 harvest were available and tested. Three accessions(HOR7869, HOR7903, HOR8946) were additionally selected ascontrols and these have been recorded as exhibiting chlorophylldefects in seed regeneration. In contrast to the usual practice, seedswere not treated with any fungicide. In the field the seeds weresown in 3 m2 blocks with three replications (seeds from 1994 withsix replications). Per accession and replication, 40 seeds were used.Four weeks after sowing (three-leaf stage) the percentage of plantsshowing chlorophyll damage was recorded.

Growth chamber experiment to test the inheritanceof the chlorophyll defectsThe aim of this test was to compare the extent of chlorophylldamage of long-term stored with that of fresh harvested seeds. Forthis purpose six accessions were selected because they showedobvious chlorophyll defects during field cultivation in spring 1999.Additionally, three samples without any symptoms in the fieldwere chosen as controls (see Table 5). For this experiment seedsfrom the same storage container already employed for the springsowing were grown together with freshly harvested progenies of 15single ears, selected randomly from the plots showing chlorophylldefects in spring. For each variant, 55 seeds were sown in threereplicates in seed trays, containing vermiculite as substratum. Thecontainers were watered frequently and well supplied with nutri-ents. The plants were grown in a climate chamber. The germinationtook place under a 12 h light/12 h dark regime at 18 and 14°C,respectively. With the appearance of the seedlings a 16 hlight/8 hdark regime at 14 and 8°C, respectively, was used. The percentageof plants with chlorophyll defects was scored at the three-leafstage.

Results and discussionDatabase analysisThe first well-documented chlorophyll defects were recorded in

Fig. 1. Chlorophyll damage of spring barley seedlings ap-peared during regeneration in the experimental field of thegenebank Gatersleben in spring 1999.


1993. The symptoms may have appeared earlier but no recordsare available. The species and number of cereal accessions show-ing chlorophyll damage during the seed regeneration between1993 and 1999 are shown in Table 1. Two facts are remarkable;firstly, the defects were not observed every year and, secondly,spring barley was most affected, whereas symptoms were notdetected in rye or winter wheat. It is shown that the years 1993,1995, 1997 and 1999 exhibited comparably high numbers ofdamaged accessions. However, if the data for temperature orrainfall during the seedlings stage of these particular years arecompared with the conditions of seasons with no or very lowdamage, no significant difference was detected. Furthermore,periodically performed analyses of the nutrient contents of thesoil did not indicate any deficiency. The following more detailedinvestigations were concentrated on spring barley only, becausethis cereal was most affected.

The extent of defected plants per barley accession is drawn upin Table 2. In most cases only some plants per plot were affected(<10%). However, there were also accessions where more than 50per cent of the plantlets showed chlorophyll damage. Analysis ofthe origin of the accessions showing chlorophyll defects in 1999showed no preference for a certain region. The symptoms weredetected in materials from Africa, Asia, Europe and South America(Table 3). No defects were found in 102 accessions from 31 addi-tional countries including Japan (21 accessions) USA (11 acces-sions), Libya (9 accessions) and 28 further regions where less thanfive accessions per country were grown. The number of accessionsshowing the defects was correlated with the number of cultivatedsamples in the particular countries. Taking into consideration thetime of cold storage, no clear relationship was found. A differenceof about 2 years on average between accessions with (9.8 years)and without (7.7 years) chlorophyll defects was discovered, al-though, in some cases the defects were already observed aftersowing the seeds in 2 consecutive years. It is concluded that thetime of storage may have some effect but this, however, is stronglydependent on the genotype.

Field experimentThe results of the field experiment are given in Table 4. The threeaccessions sown in early spring and in summer 1999, where seedsfrom the same storage containers were used, showed high per-centages of defects in both investigations. Consequently, seedtreatment with fungicides, or unfavourable climatic conditions inearly spring, such as low soil temperature during germination, orlate frosts, are not reasons for the chlorophyll damage. Theresults obtained from the three samples grown from seeds of the

Table 1. Number of cereal accessions with chlorophyll defects recorded between 1993 and 1999 during seedregeneration in the experimental fields of the Gatersleben genebank

Species Year of multiplication

(Growth habit) 1993 1994 1995 1996 1997 1998 1999 Total

Hordeum vulgare (spring) 29 0 14 0 14 3 81 141Avena sativa (spring) 1 0 23 0 0 0 0 24Triticum aestivum (spring) 0 0 21 0 1 0 0 22Hordeum vulgare (winter) 0 0 0 0 0 1 0 1Total 30 0 58 0 15 4 81 188

Table 3. Geographical origin of spring barley acces-sions regenerated/damaged in 1999

Country Number of Number ofof origin cultivated accessions with

accessions chlorophyll defects

Ethiopia 184 42China 38 17Unknown 38 6Turkey 34 4India 34 1Nepal 20 2Pakistan 12 1Afghanistan 11 2Iran 7 1Spain 3 2Chile 2 1Russia 2 1GDR 1 1Other 102 0Total 488 81

Table 2. Extent of chlorophyll defects in spring barley

Percentage of plants Number of accessions per yearper plot showingchlorophyll defects 1993 1995 1997 1999

H < 0 13 0 11 2410–19 8 2 2 1920–29 4 6 1 1330–49 0 1 0 19>50 4 5 0 6

1994 harvest are somewhat surprising. Whereas in 1993 theaccessions were damaged, in 1994 they were not. In this experi-ment the plants showed the symptoms again, suggesting that thedamage is somehow related to the time of storage. Anothersurprising fact was that one of the control accessions showeddamage of 16%.

Growth chamber experimentThe reason for performing this experiment under highly con-trolled conditions was to make sure that environmental influ-ences, such as soil or weather conditions, could not cause thechlorophyll defects. In Table 5 it is clearly demonstrated that allaccessions that showed the symptoms in the field in spring 1999were also affected in the growth chamber experiment, providedthat seeds from the same storage container were tested. Analysisof the freshly harvested seeds of five of the six accessions showed


no, or only a very low, percentage of plants without chlorophyll(albinos), but no plants with typical white-green striped leavesoccurred. An exception was accession number HOR2738. For thisbarley accession of unknown origin, 10% of the plants obtainedfrom fresh seeds were damaged. It should be noted that thisaccession also showed the symptoms during two consecutiveregeneration cycles in 1998 and 1999. This genotype should beclassified as highly susceptible. As in the field experiment, one ofthe controls showed damage as well.

ConclusionsAnalysis of the data recorded during regeneration in formeryears, and obtained from the experiments described here, doesnot completely clarify the reason for the appearance of chloro-phyll damaged plants during the multiplication of cereals. Dur-ing 7 years of observation in Gatersleben it is shown that springcereals, mainly spring barley, were most frequently damaged.The explanation for that, however, seems neither to be the usualpractice of treating of seeds with fungicides, nor unfavourable

conditions during the time of sowing in early spring, as could beconcluded from the experiments performed in the field in sum-mer and in the growth chamber.

With one exception, the freshly harvested progenies of dam-aged accessions did not show the symptoms again; it can, there-fore, also be concluded that the chlorophyll defects were notinherited, i.e. mutations during long-term storage can be ex-cluded. Whether the period between harvest and transfer to thelong-term storage facilities may have had some influence canonly be conjecture. Badly matured or moist harvested grainsstored for a short time under unfavourable conditions may causesome damage. To prove this, further investigations are necessary.The time of storage has some effect on the susceptibility of thematerial; however, this is strongly dependent on the genotype.Some accessions (HOR2738) seem to be highly susceptible aftershort-term storage. For further analysis, crosses were initiatedbetween highly susceptible and non-susceptible genotypes.

ReferencesIPK Annual Report. 1999. Report of the genebank department.Lundqvist, U., J.D. Franckowiak and T. Konishi. 1997. New and

revised descriptions of barley genes. Barley Genet. Newsl.26:22-516.

McIntosh, R.A., G.E. Hart, K.M. Devos, M.D. Gale and W.J.Rogers. 1998. Catalogue of gene symbols for wheat in Proc.9th Int. Wheat Genet. Symp., vol. 5: 1-236 (Slinkard A.E.,ed.). University Extension Press, University of Saskatchewan,Canada.

Melz, G. 1989. Beiträge zur Genetik des Roggens (Secale cereale L.).Dsc. Thesis, Berlin.

Specht, C.-E., E.R.J. Keller, U. Freytag, K. Hammer and A. Börner.1997. Survey of seed germinability after long term storage inthe Gatersleben genebank. Plant Genet. Resour. Newsl. 111:64-68.

Specht, C.-E., U. Freytag, K. Hammer and A. Börner. 1998.Survey of seed germinability after long term storage in theGatersleben genebank (part 2). Plant Genet. Resour. Newsl.115:39-43.

Table 4. Percentage of chlorophyll damage in spring barley grown both in early spring (regeneration) andsummer (field experiment)

Accession number Year of harvest of Percentage of plants with chlorophyll defectsanalysed seeds

Early spring sowing in 1999 Summer sowing in 1999

Hor7766 1985 70 54Hor7884 1990 70 91Hor7892 1990 80 78Hor1735 1994 70‡ 39Hor1743 1994 70‡ 29Hor1766 1994 80‡ 18Hor7869† 1990 – 16Hor7903† 1990 – –Hor8946† 1988 – –

†Control. ‡In 1993.

Table 5. Chlorophyll defects of stored and newlyharvested seeds in container trials in comparison withthe damage recorded during last cultivation in field

Accession Percentage of plants with chlorophyll defectsnumber

Container experimentEarly springsowing with Stored Freshstored seeds seeds harvestedin 1999 seeds

Hor1576 35 84 3‡

Hor2738 40 37 10Hor4046 20 34 1‡

Hor7764 70 58 0Hor7875 60 27 0Hor8092 65 98 0Hor5417† 0 0 0Hor7191† 0 0 0Hor7701† 0 27 0

†Control.‡Albino plants not showing the typical white–green stripedsymptoms.


Système de reproduction et variabilitémorpho-phénologique chez Allium roseum L.R. Jendoubi1, M. Neffati1? , B. Henchi2 et A. Yobi1

1 Institut des Régions Arides, 4119 Médenine, Tunisie2 Faculté des Sciences de Tunis, Tunisie

RésuméSystème de reproduction etvariabilité morpho-phénologiquechez Allium roseum L.Le présent travail a porté sur l’étude dusystème de reproduction et de lavariabilité intra-spécifique au niveau dedifférentes accessions autochtonesd’Allium roseum—géophyte spontanéeappartenant à la famille des liliaceaelargement utilisée par la population lo-cale du sud tunisien pour ses nombreusesvertus thérapeutiques et comme condi-ment—en provenance de différenteszones bioclimatiques de la Tunisieméridionale. Cette étude a démontréque cette espèce est allogame et qu’elleprésente une importante variabilité mor-pho-phénologique entre les différentesaccessions. Cette variabilité se traduit pardes différences de vigueur et de précocité.Les analyses statistiques (analyse de lavariance, des corrélations et encomposantes principales) ont montréque l’accession Sfax appartient auxgroupes des accessions les moinsvigoureuses et les plus précoces. Parcontre, celle de Zarzis s’est avérée la plusvigoureuse et la plus tardive. Les acces-sions de Jerba et de Bengardane sont lesplus polymorphes pour les caractèresétudiés.

SummaryReproductive system andmorpho-phenologicalvariability in Allium roseum L.Allium roseum is a spontaneous geophyticspecies belonging to the family ofliliaceae, extensively used by local popu-lation of southern Tunisia for its numer-ous therapeutic virtues and as condi-ment. The goal of this study is to deter-mine the reproductive system and theinfra-specific variability of different au-tochthonous accessions of Allium roseumfrom different bioclimatic areas of south-ern Tunisia. It has been demonstratedthrough this study that this species isallogamous and that an important mor-pho-phenological variability exists withinthe different accessions studied. Thisvariability is expressed in differences ofvigour and precocity. The statisticalanalysis (analysis of variance, correla-tions and principal component analysis)showed that the accession from Sfax be-longs to the least vigorous and most pre-cocious group of accessions. On the otherhand, the one from Zarzis proved to bethe most vigorous and the most late inmaturity. Accessions from Jerba andBengardane are the most polymorphicsfor the studied characters.

Key words: Allium roseum,reproductive system, morpho-phenological variability

ResumenSistema de reproducción y dela variabilidad morfo-fenológica de Allium roseum L.El presente trabajo trata del sistema dereproducción y de la variabilidad infra-específica a nivel de diferentes accesionesautóctonas de Allium roseum (geófitaespontánea perteneciente a la familia delas liliáceas y largamente utilizada por lapoblación del sur tunecino por susnumerosas virtudes terapéuticas y comocondimento), provenientes de diferenteszonas bioclimáticas de Túnez meridional.Se ha demostrado mediante este estudioque esta especie es alógama y que existeuna importante variabilidad morfo-fenológica entre las diferentes accesiones.Esta variabilidad se traduce en diferenciasde vigor y de precocidad. Los análisisestadísticos (análisis de varianza, decomponentes principales y correlaciones)han mostrado que la accesion de Sfaxpertenecía a los grupos de procedenciamenos vigorosos y más precoces. Encambio, las accesion de Zarzis son lasmás vigorosas y la más tradías. Lasaccesiones de Jerba y de Bengardane sonlas más polimorfas por los carácteresestudiados.


IntroductionAppartenant à la famille des liliacées, Allium roseum est une espècegéophyte spontanée. Elle est constituée d’une tige réduite à unplateau, de feuilles composées d’une gaine incolore à la base et d’unlimbe lisse et étroit et d’une inflorescence constituée d’une hampeflorale supportant une ombelle (Figure 1). La fleur de cette espèce esthermaphrodite et comporte normalement trois sépales, trois pétales,six étamines et trois carpelles. Les bulbes sont plus ou moinsarrondis et les graines sont de couleur noire et de forme irrégulière.

Cette espèce largement répandue en Tunisie méridionale, oùelle est connue pour ses nombreuses vertus thérapeutiques contrele rhumatisme (Trotter 1915, cité par Le Floc’h 1983), est utiliséepar la population locale comme condiment, remplaçant parfoisl’oignon (Gobert 1940, cité par Le Floc’h 1983). Au cours desannées pluvieuses, cette espèce est vendue aux bords des routeset dans les marchés de fruits et légumes à des prix intéressants.Cependant, les pratiques de sa cueillette s’avèrent trèsdestructives puisqu’au cours de sa collecte, même les bulbessont arrachés. Etant donné l’intérêt économique croissant

d’Allium roseum et les risques d’érosion génétique qui pèsent surelle, la sauvegarde et la domestication de cette espèce sontdevenues aujourd’hui une nécessité impérieuse.

Au niveau mondial, certaines études ont été consacrées à lataxinomie du genre Allium, à sa répartition géographique et à la

Fig. 1. Plante entière d’Allium roseum : Parties aériennes etparties souterraines.


caractérisation de ses bulbes (Camefort 1977, Purseglove 1979,Hanelt et al. 1992, Ferchichi 1997). En Tunisie, comme dans toutle reste de l’Afrique du Nord où Allium roseum est présente, voirelargement répandue sous forme endémique (var. odoratissimum)(Cuenod 1954), rares sont les études qui s’y sont intéressées àl’exception, toutefois, de celle réalisée par Jendoubi (1999) sur lacaractérisation morpho-phénologique de certaines accessions decette espèce.

La connaissance précise des caractéristiques biologiques etdes exigences écologiques d’une espèce végétale constitue,pourtant, la pierre angulaire de tout programme visant à saconservation, son amélioration et sa domestication.

Dans cette perspective, le présent travail a été consacré àl’étude d’Allium roseum afin de mieux caractériser cette espèce etd’évaluer ses aptitudes productives en rapport avec sa diversitébiologique. Il a porté précisément sur le mode de pollinisation etsur la variabilité morpho-phénologique d’accessions en prov-enance de différents sites bioclimatiques de la Tunisie aride etdésertique.

Matériels et méthodesEtude du développement des parties aérienne etracinaire chez Allium roseumAfin d’étudier le développement des systèmes aérien et racinairechez Allium roseum, des bulbes de diamètres compris entre 6,5 et10,4 mm ont été semés le 16 octobre 1998 dans des rhizotrons(pots rectangulaires présentant une face vitrée qui permet devisualiser une partie du système racinaire) remplis d’un substratsablo-limoneux. Après émergence, les descripteurs de lacroissance des systèmes aérien et racinaire telles que la longueurmoyenne des racines, la longueur moyenne des deux premièresfeuilles et la longueur moyenne de la hampe florale ont fait l’objetde différents suivis.

Etude du mode de pollinisation chez AlliumroseumAfin de déterminer le mode de pollinisation chez cette espèce, 15pieds d’accessions différentes maintenues en collection aupastoretum de l’Institut des Régions Arides à El Fjè, Médenine,ont été ensachés. De type méditerranéen, le climat de ce site estsemi-aride à hivers tempérés à doux (Emberger 1945).L’ensachage a été fait avec des sacs à mailles fines empêchant lapénétration du pollen exogène. Quinze autres pieds de mêmevigueur non ensachés ont servi de témoins. Après la maturationdes semences, il a été procédé à la récolte de toutes les inflores-cences et au comptage des graines formées aussi bien au niveaudes pieds ensachés qu’au niveau des témoins.

Etude de la variabilité morpho-phénologique chezcertaines accessions d’Allium roseumCette étude a porté sur 20 accessions originaires de quatre régionsdifférentes de la partie littorale de la Tunisie aride et désertique :Bengardane (11), Zarzis (1), Jerba (5) et Sfax (3) (Tableau 1 etFigure 2).

Après une phase d’installation qui a duré une année, 15plants de chaque accession ont été choisis pour faire l’objet demesure des descripteurs suivants :

Tableau 1. Lieux de collecte des différentesaccessions d’Allium roseum maintenues encollection au pastoretum de l’IRA en 1998

N° N° de Site de Régiond’ordre l’accession collecte

1 1 Jamila-Sebkha Bengardane2 2 Jamila-Hamada3 3 Kourkata-Sebkha4 4 Saidane5 6 Oum Chrakat6 13 Oued Fessi7 15 Choucha- Sebkha8 16 Elktef-Port9 17 Elktef- Themed10 18 Elktef - Elfeskiya11 19 El bibane

côté Elktef12 20 Grabat Zarzis13 5 Ejim Jerba14 7 Chabbar15 8 Basim16 9 Kantra17 10 Mahboubine18 11 Bir Ali Sfax19 12 74 route de

Gabès20 14 Mahrès

Fig. 2. Localisation géographique, en Tunisie, des sites decollecte des différentes accessions d’Allium roseummaintenues en collection au pastoretum de l’IRA à Médenine.


• Nombre de feuilles (NMF) ;• Longueur moyenne des deux feuilles les plus longues (LMF)

en cm ;• Longueur de la feuille la plus longue (LM) en cm ;• Largeur moyenne des deux feuilles les plus larges (LRF) en

cm ;• Largeur de la feuille la plus large (LRM) en cm ;• Taux moyen d’apparition des feuilles pendant les deux pre

miers mois (TAF) ;• Longueur maximale de la hampe florale (LMH) en cm ;• Nombre de fleurs au niveau de l’inflorescence la plus longue

(NFI) ;• Nombre d’inflorescences par pied (NI) ;• Durée de la floraison (DFL) en jour ;• Vitesse de croissance de la hampe florale (VCH) en cm /jour ;• Diamètre maximum des boutons floraux (DMB) en cm ;• Précocité de la floraison (nombre de jours après le 01/01/99)

(PRC).Il y a lieu de signaler que la précocité des différentes acces-

sions étudiées a été évaluée en déterminant la longueur del’intervalle de temps compris entre la date de l’ouverture desboutons floraux et le 01/01/1999. Cette dernière date a étéretenue d’une façon arbitraire, mais raisonnée, puisqu’il étaitdifficile de considérer la date précise de l’émergence de chaqueplant.

Les observations ont été effectuées une fois par semaine surdes plants se développant dans des conditions naturelles (sansirrigation). Totalisant 306 mm de pluie, l’année d’observations’est révélée relativement pluvieuse (Tableau 2), puisque lamoyenne inter-annuelle des précipitations au niveau de la stationexpérimentale est de l’ordre de 180 mm.

Les données relatives aux différents paramètres mesurés ontfait l’objet d’analyses de la variance, des corrélations et d’uneanalyse en composantes principales au moyen du logiciel SAS(Statistical Analysis System, 1998).

Résultats et discussionEtude du développement de la partie aérienne chezAllium roseumL’examen de la figure 3, représentant l’évolution du nombremoyen de feuilles et de la longueur moyenne des deux premièresfeuilles (les plus longues) chez des plants d’Allium roseum issusdes bulbes de 6,5 à 10,4 mm de diamètre et semés en rhizotron,montre que l’émission des feuilles s’est poursuivie jusqu’à la findu mois de janvier, soit trois mois après l’émergence. Cette figuremontre également que le nombre de feuilles par plant n’a pasdépassé les cinq feuilles. A partir de ce stade, l’émission denouvelles feuilles s’est souvent accompagnée de la chute de cellesqui les ont précédées. La longueur moyenne des deux premièresfeuilles a évolué graduellement au cours des deux premiers moisaprès émergence (novembre et décembre). La longueur maximale

Tableau 2. Pluviométrie enregistrée au niveau du site d’El Fjè au cours de la campagne agricole 1998–1999

Mois S O N D J F M A M J J A TotalPluviométrie (mm) 9,2 122,5 5,5 28 78,5 53 9,5 306,2Nbre de jours de pluie 3 2 1 1 4 4 2 17

Fig. 3. Evolution du nombre moyen de feuilles (-o-) et de lalongueur moyenne des deux premières feuilles (-o-) chezdes plants d’Allium roseum.

Fig. 5. Elongation moyenne des racines chez des plantsd’Allium roseum.

Fig. 4. Elongation de la hampe florale en fonction du tempschez des plants d’Allium roseum.

moyenne atteinte par ces feuilles a été de 36,5 cm et leurdessèchement a commencé au cours de la deuxième semaine dumois de mars.


La figure 4, représentant la croissance de la hampe florale enfonction du temps chez des plants d’Allium roseum issus desbulbes et cultivés en rhizotron, montre que l’élongation de lahampe florale s’est poursuivie progressivement jusqu’à atteindreune longueur maximale moyenne de 35,5 cm au début du moisde mars (129 jours après émergence). A partir de cette date,l’épanouissement des fleurs (anthèse) a commencé.

Une durée d’élongation de la hampe florale de 120 jours a étéégalement enregistrée chez des plants d’Allium roseum cultivés enpépinière au cours de la même période (Jendoubi 1999). Cesrésultats nous permettent de dire que si le semis est précoce(début d’automne), la durée d’élongation de la hampe floralepeut atteindre quatre mois.

Les résultats concernant la longueur moyenne des racineschez cette espèce sont représentés à la figure 5. Il s’est avéré quel’élongation des racines s’est poursuivie jusqu’à atteindre unelongueur maximale moyenne de 33,5 cm au début du 3ème moisaprès émergence.

L’arrêt de croissance des racines a coïncidé avec celui del’émission de nouvelles feuilles, le dessèchement des extrémités desfeuilles et l’apparition de la hampe florale. Il semble par conséquentque le passage du stade végétatif au stade reproductif dépendplutôt des bulbes que des racines. A ce stade de développement,nous avons aussi remarqué que les racines gonflées formées àproximité des bulbes se sont atrophiées en raison probablement dela mobilisation de leurs réserves qui seraient utilisées pourpermettre à la plante d’accomplir son cycle de vie.

Etude du mode de pollinisation chez Allium roseumLes quantités de graines produites par les deux types de plantes(ensachées et non ensachées) sont consignées dans le tableau 3.

Ce tableau montre qu’une inflorescence ensachée n’a produit enmoyenne qu’une seule graine, alors que chez le témoin (plantessoumises à une pollinisation libre), ce nombre est de 179 ± 23 graines.Un tel résultat montre que la pollinisation de cette espèce, commec’est le cas de la majorité des espèces du même genre (FAO 1961), estde type allogame. D’après Richards (1997), l’allogamie peut concerneraussi bien des fleurs d’une même plante (cléistogamie)que des fleurs de plants différents (xénogamie). ChezAllium cepa, plus de 93% des cas de pollinisation sontallogames par entomophilie (FAO 1961). Chez cetteespèce, le transfert du pollen entre les fleurs d’uneombelle ou d’une plante s’effectue par l’action d’unagent extérieur, mais l’auto-pollinisation à l’intérieur dela fleur est difficile (Pouvreau 1984). Le même auteursignale également que les fleurs sont protandres aussibien chez cette espèce que chez Allium porrum.

Etude de la variabilité intra-spécifiquechez certaines accessions autochtonesd’Allium roseumAnalyse de la diversité génétique chezAllium roseumL’analyse de la variance à deux dimensions a permisde mettre en évidence un effet accession hautementsignificatif pour tous les caractères étudiés (Tableau4). Pour chaque descripteur étudié, il existe donc, au

moins, deux accessions qui sont différentes entre elles. Lesdifférences enregistrées entre certaines accessions seraient de na-ture génétique puisque ces dernières ont été cultivées dans lesmêmes conditions édapho-climatiques.

Dans le tableau 5, sont consignées les corrélations existantentre les différents descripteurs mesurés chez l’espèce étudiéesans distinction entre les accessions. Ce tableau montre que lescorrélations entre ces descripteurs sont, pour la plupart, posi-tives et très hautement significatives. Les corrélations positives(quel que soit le niveau de leur signification) entre les descripteursrelatifs à la vigueur végétative (NMF, LMF, LM, LRF, LMH) etceux relatifs aux caractéristiques de la reproduction (NFI, NI,DFL et DMB) permettent de dire que ce sont les plants les plusvigoureux qui produisent le nombre de boutons floraux le plusimportant. Ce tableau montre également que la précocité de lafloraison est négativement corrélée, et d’une façon très hautementsignificative, avec tous les autres paramètres étudiés à l’exceptionde la vitesse de croissance de la hampe florale (VCH).

Quant à la durée de la floraison (DFL), qui est positivementcorrélée avec tous les descripteurs étudiés, elle ne semble avoiraucun rapport avec la longueur de la feuille la plus longue (LM),la corrélation entre ces deux paramètres n’étant pas significative.

Signification des 3 premiers axes de l’ACPLes composantes principales de valeur propre supérieure à 1, quisont représentées dans le Tableau 6, donnent une estimation dupourcentage de la variabilité représenté par chaque axe. En général,on ne considère que les composantes principales de valeur propre

Tableau 4. Analyse de la variance de 13 descripteurs morpho-phénologiques chez 20 accessions d’Allium roseum

Variable (1) Degré de Carré moyen Test F Niveau deliberté accession signification

NMF 19 002.84 05.68 ***LMF 19 169.87 10.13 ***LM 19 125.34 08.38 ***LRF 19 000.04 07.71 ***LRM 19 000.04 08.05 ***TAF 19 000.05 05.83 ***LMH 19 169.51 05.01 ***NFI 19 313.76 08.96 ***NI 19 001.16 05.53 ***DFL 19 674.39 12.23 ***VCH 19 000.27 03.68 ***DMB 19 010.27 05.14 ***PRC 19 517.07 10.38 ***

(1) Même légende que celle utilisée dans la partie méthode*** hautement significative.

Tableau 3. Etude de l’effet d’ensachage sur le modede pollinisation chez Allium roseum

Traitement Nbre de Nbre Nbreplants fleurs/ plant graines/ plant

Pieds 15 37 ± 4 1 ± 1ensachésTémoin 15 37 ± 4 179 ± 23


supérieure à 1 (Jeffer 1967). Ainsi, les trois premières composantes(CP1, CP2 et CP3) ont été retenues pour décrire la variabilité totaledes accessions car elles représentent à elles seules 70 % de cettevariabilité. Seuls les descripteurs ayant une corrélation supérieureà 0,70 (en valeur absolue) avec l’un des trois axes ont été considéréscomme étant significatifs pour celui-ci.

Le premier axe (CP1), qui décrit 50 % de la variation, estdéfini par les 6 descripteurs de la croissance végétative (NMF,LMF, LM, LRF, LRM, TAF). Le deuxième axe (CP2), qui décrit 11% de la variation, est représenté par les descripteurs de la repro-duction (LMH, NFI, NI, DFL, PRC). Le troisième axe (CP3), quiabsorbe 9 % de la variation, n’est défini que par un seul paramètrequi est la vitesse de croissance de la hampe florale (VCH).

Tableau 5. Matrice de corrélation entre les différents descripteurs de croissance mesurés chez 20 accessions localesd’Allium roseum

Dscpt NMF LMF LM LRF LRM TAF LMH NFI NI DFL VCH DMB PRC

NMFLMF 0,44***LM 0,56*** 0,81***LRF 0,56*** 0,56*** 0,74***LRM 0,56*** 0,55*** 0,73*** 0,99***TAF 0,67*** 0,50*** 0,63*** 0,68*** 0,69***LMH 0,44*** 0,36*** 0,47*** 0,61*** 0,63*** 0,57***NFI 0,51*** 0,41*** 0,52*** 0,68*** 0,70*** 0,55*** 0,70***NI 0,36*** 0,37*** 0,42*** 0,60*** 0,60*** 0,42*** 0,41*** 0,50***DFL 0,16 * 0,16 * 0,10 NS 0,32*** 0,32*** 0,22 ** 0,60*** 0,50*** 0,496***VCH 0,22** -0,09 NS 0,03 NS -0,03 NS -0,02 NS -0,07 NS 0,18** 0,09 NS -0,14* -0,08 NSDMB 0,22** 0,41*** 0,44*** 0,49*** 0,48*** 0,44*** 0,45*** 0,41*** 0,39*** 0,32*** -0,05 NSPRC -0,32*** -0,42*** -0,41*** -0,53*** -0,541*** -0,52*** -0,64*** -0,514*** -0,51*** -0,65*** 0,28 *** -0,50***

Même légende que celle utilisée dans la partie méthode.*** Corrélation très hautement significative.** Corrélation hautement significative. * Corrélation significative.NS Corrélation non significative.

Fig. 7. Répartition des 20 accessions locales d’Alliumroseum dans un plan défini par les axes 1 et 3.

Tableau 6. Valeurs propres et représentation de lavariation par les trois premières composantesprincipales (ACP) définies par 13 paramètresmorpho- phénologiques chez 20 accessionsd’Allium roseum

Composante CP1 CP2 CP3principaleValeur propre 6.55 1.42 1.16% d’inertie 50 11 09

Variables† NMF 0.53 0.28 0.45LMF 0.80 0.10 -0.14LM 0.91 0.12 0.07LRF 0.75 0.47 0.11LRM 0.73 0.49 0.12TAF 0.65 0.39 0.14LMH 0.26 0.73 0.39NFI 0.42 0.64 0.31NI 0.36 0.61 -0.17DFL 0.08 0.88 -0.01TCH -0.05 -0.11 0.91DMB 0.49 0.39 -0.04PRC -0.34 -0.77 0.19

†Même légende que celle utilisée dans la partie méthode.

Fig. 6. Répartition des 20 accessions locales d’Alliumroseum dans un plan défini par les axes 1 et 2.


La projection des nuages de points (accessions) sur desplans définis par les principaux axes pris deux à deux estreprésentée par les figures 6 et 7.

Répartition des différentes accessions dans un plandéfini par les axes 1 et 2La figure 6, représentant la répartition des différentes accessionsdans un plan défini par les axes 1 et 2, montre un grandpolymorphisme des accessions pour les descripteurs étudiés. Cepolymorphisme se manifeste plus au niveau des descripteurs dela croissance végétative (CP1). Les accessions 7, 13, 20 et 3 sontles plus vigoureuses, alors les accessions 11, 9, 16, 18 et 19présentent les vigueurs végétatives les plus faibles.

Quant aux descripteurs de la vigueur reproductive (CP2), ilsopposent les accessions 7 et 12, considérées les plus vigoureuseset les moins précoces, aux accessions 15, 4, 3, 20, 9 et 16,considérées comme étant les moins vigoureuses et les plusprécoces. Les deux axes 1 et 2 n’ont pas permis de discriminer lesaccessions 5, 6, 8, 14, 10, 17 et 2.

Ces résultats indiquent que ces deux axes n’ont pas permisde discriminer les accessions en provenance de Jerba, àl’exception de celle de Chabbar (7). Ils n’ont pas égalementpermis de discriminer les accessions en provenance de Sfax, àl’exception de celle de Bir Ali (11). Alors que l’accession deZarzis (20) est considérée parmi les plus vigoureuses, celles deBengardane se caractérisent par des vigueurs variant des plusfaibles aux plus fortes.

Répartition des différentes provenances dans un plandéfini par les axes 1 et 3La figure 7, représentant la répartition des différentes accessionsdans un plan défini par les axes 1 et 3, montre une grandevariabilité au niveau du taux de croissance de la hampe florale(TCH) entre les accessions étudiées.

L’axe 3 oppose en effet les accessions 18, 17, 7 et 20,caractérisées par une croissance plus rapide de la hampe florale,aux accessions 6 et 11 dont le taux de croissance de la hampeflorale est plus faible. Toutefois, cet axe n’a pas permis dediscriminer les accessions 15, 12, 8, 4 et 2 qui, sur la base de cesdescripteurs, forment un groupe relativement homogène.

ConclusionAyant porté sur l’étude du système de reproduction et l’analysede la variabilité morpho-phénologique chez plusieurs accessionsd’Allium roseum, la présente étude a permis de tirer les conclu-sions suivantes :● Le stade 5ème feuille constitue chez cette espèce le dernierstade de la phase végétative à partir duquel l’émission denouvelles feuilles correspond souvent au dessèchement de cellesqui les précédent;● L’arrêt de l’émission des nouvelles feuilles s’est produit aumême moment que celui de la croissance racinaire et del’apparition de la hampe florale qui marque la transition de laphase végétative à la phase reproductive. Il apparaît donc que ledéveloppement de la plante au cours de cette deuxième phase estassuré beaucoup plus par la mobilisation des réserves déjàstockées au niveau des bulbes et des racines gonflées;

● A l’instar de la majorité des autres espèces du même genre,Allium roseum s’est avérée une espèce allogame;● La variabilité intra-spécifique mise en évidence entre lesdifférentes accessions de cette espèce se traduit par desdifférences de vigueur et de précocité. L’analyse en composantesprincipales a montré que les accessions les plus septentrionales(Sfax) appartiennent aux groupes des accessions les moinsvigoureuses et les plus précoces. Par contre, celle de Zarzis s’estavérée la plus vigoureuse et la plus tardive. Les accessions deJerba et de Bengardane sont les plus polymorphes pour lescaractères étudiés.

RéférencesCamefort, H. 1977. Morphologie des végétaux vasculaires.

Cytologie, anatomie et adaptations. Doin. Edit. Paris. 432p.Cuenod, A. 1954. Flore de la Tunisie. Office de l’Expérimentation

et de la Vulgarisation Agricole de Tunisie. p.287.Enberger, L. 1945. Une classification bio-géographique des

climats. Recueil des travaux des laboratoires de Botanique,de Géologie et de la Faculté des Sciences de l’Université deMontpellier. Bot. 7:3-43.

FAO. 1961. Semences agricoles et horticoles. Etudes agricoles del’Organisation des Nations Unies pour l’Alimentation etl’Agriculture. No.55, 607 p.

Ferchichi, A. 1997. Contribution à l’étude caryologique,caryosystématique, morpho-biologique et écologique de laflore de la Tunisie pré-saharienne. Thèse d’Etat es Sciencesbiologiques. Univ. Tunis II. 214p.

Hanelt, P. 1992. The genus Allium . Taxinomic Problems andGenetic Resources. Proceedings of an International Sympo-sium held at Gaterleben, Germany, June 11-13, 1991. Ham-mer K. et Knupffer H. (eds.) 257-263.

Jeffer, J. N. R. 1967. Two case studies in the application ofprincipal component analysis. Appl. Stat. 16:225-236.

Jendoubi, R. 1999. Etude de la diversité biologique chez Alliumroseum L. : Ecologie-Phénologie. DEA en EcophysiologieVégétale à la Faculté des sciences de Tunis, Tunisie. 56p.

Le Floc’h, E. 1983. Contribution à une étude ethnobotanique de laflore tunisienne. Imprimerie Officielle de la RépubliqueTunisienne. 401p.

Pouvreau, A. 1984. Production de semences potagères. InPollinisation et production végétales. Ouvrage collectif del’INRA, Paris. 472-493.

Purseglove, J.W. 1979. Alliaceae. In Purseglove Edit : Tropicalcrops monocotyledons. Longman Group, London. 37-57.

Richards, A. 1997. Plant breeding systems. Depart. of PlantBiology. Univ. of New Castle upon Tyne, London, USA.


Assemblage of sesame germplasm for conservationand genetic improvement in NigeriaO.A. Falusi¹? and E.A. Salako²¹ Department of Science Technology, Federal Polytechnic, P.M.B. 55, Bida, Niger State, Nigeria, West Africa² Crop Production Department, School of Agric. and Agric. Technology, Federal University of Tech., Minna, Niger State, Nigeria, West Africa

SummaryAssemblage of sesamegermplasm for conservationand genetic improvement inNigeriaSurvey missions were undertaken to allthe major sesame-producing areas ofNigeria toward the end of the croppingseason in September 1998, when thefarmers were expected to be harvestingthe crop. The survey covered 21 townsand villages in 18 local government areasof 11 states and the Federal Capital Terri-tory. These areas include parts of thecentral, northwestern and northeasternzones of the country. A total of 36 farm-ers were interviewed and 27 accessionswere collected from 20 farmers. Resultsfrom this study showed that the highestnumber of sesame accessions collectedwas from Benue and Kaduna states.These were closely followed by Nigerand Nassarawa States, an indication thatthese states have the greatest diversity ofthe crop genetic resource in Nigeria. Themajor constraint pointed out by thefarmers was the non-availability of im-proved varieties of the crop. Scientificcharacterization of these accessions istherefore necessary to group them intoknown Sesamum species which will pro-vide the resources required for sesamecrop improvement programmes.

Key words: Germplasm collection,improved varieties, Nigeria, sesame,Sesamum, survey missions

ResumenRecogida de germoplasma desésamo para su conservacióny mejora genética en NigeriaSe emprendieron misiones de encuestaen todas las grandes zonas productorasde sésamo de Nigeria hacia el final de laépoca de recolección en septiembre de1998, cuando se suponía que losagricultores recogían la cosecha. Laencuesta abarcó 21 poblaciones y aldeasen 18 circunscripciones locales de 11estados y del territorio de la capital fed-eral, correspondientes a las zonas cen-tral, noroccidental y nororiental del país.Se realizaron en total 36 entrevistas conagricultores y se recogieron 27 accesionesde 20 agricultores. La mayoría de lasaccesiones recogidas procedían de losestados de Benue y Kaduna, seguidos decerca por Níger y Nassarawa, lo que in-dica que estos estados poseen la mayordiversidad de recursos genéticos de laplanta en Nigeria. La principal dificultadseñalada por los agricultores era la nodisponibilidad de variedades mejoradas.Es por lo tanto necesaria unacaracterización científica de estasaccesiones para agruparlas en especiesconocidas de Sesamum que ofrezcan losrecursos requeridos para un programade mejora del cultivo del sésamo.

RésuméCollecte de matériel génétiquede sésame à des fins deconservation et d’améliorationgénétique au NigeriaDes missions d’étude ont été réaliséesdans les principales régions de produc-tion de sésame au Nigeria vers la fin de lasaison de végétation en septembre 1998,période habituelle de la récolte. L’étudeconcerne 21 villes et villages et est réal-isée dans 18 parcelles du gouvernementlocal réparties sur 11 états et la capitalefédérale. Ces parcelles sont situées dansles zones centrales, nord-ouest et nord-est du pays. Au total, 36 agriculteurs ontété interrogés et 27 accessions ont étérécoltées auprès de 20 d’entre eux. Leplus grand nombre d’accessions desésame récoltées durant cette étude pro-vient des états de Benue et Kaduna, suivisde près par les états du Niger et du Nas-sarawa, ce qui indique que ces états pos-sèdent la plus grande diversité de res-sources phytogénétiques au Nigeria. Lesagriculteurs éprouvent principalementdes difficultés à se procurer des variétésaméliorées. Il est donc nécessaire de car-actériser scientifiquement ces accessionspour les grouper selon les espèces con-nues de Sesamum, ce qui fournira le maté-riel nécessaire aux programmesd’amélioration du sésame.


IntroductionSesame (Sesamum indicum L.) belongs to division Spermatophyta,subdivision Angiospermae, class Dicotyledoneae, orderTubiflorae, family Pedaliaceae and genus Sesamum (Bruce 1953;Joshi 1961; Hutchinson and Dalziel 1963; Kumar et al. 1967;Weiss 1971; Purseglove 1974). The genus consists of about 36species of which the most commonly recognized is Sesamumindicum (Purseglove 1974). Other related species are Sangustifolium, S. angolense, S. radiatum, S. indicatum, S. capense, S.alatum, S. schekiti,, S. laciniatum, S. prostratum and S. occidentatemost of which have been recorded in Africa (Salunke and Desai1986; Irvine 1969; Van Rheenen 1970). The plant Sesamum indicumis an important edible oil seed crop. It is commonly referred to as‘the queen of the oil seeds’ by virtue of the excellent quality of oilit produces.

In Nigeria, the crop (often referred to as Beniseed) is widelyused and very popular in parts of the Central, North Western

and North Eastern zones where it is usually grown. The localnames are ‘Ridi’ Hausa, ‘Ishwa’: Tiv, ‘Yamati’or ‘Eeku’: Yoruba,‘Igorigo’ Igbira and ‘Doo’: Jukun. The seeds, which yield half oftheir weight in oil, are most commonly used in soups while theyoung leaves are used as a soup vegetable. Various parts of theplants are also used in native medicines. The stems are usuallyburned as fuel where firewood is scarce and the ash is commonlyused for local soap production. The pressed cake remaining afterthe oil is removed is a rich source of protein for farm animals.

Nigeria is endowed with favourable ecologies for sesamecultivation, but the low yield potential of the crop coupled withproblems encountered during its establishment and harvestinghave tended to discourage growers, thus leading to a decline inthe total area devoted to its cultivation. Considering the impor-tance of the crop, it is anticipated that varieties which are moreproductive than those currently grown by farmers can be devel-


oped. This background has made the collection of sesamegermplasm a priority in Nigeria. A survey of producers wastherefore conducted in the producing areas of the country andsamples of sesame accessions and useful species were collectedwhere necessary.

Materials and methodsA survey of producers of sesame was conducted in 11 states ofCentral, North Western and North Eastern Nigeria in September1998, when the farmers were expected to be harvesting the crop.The exercise was carried out in collaboration with the NationalCereals Research Institute (NCRI) Badeggi, Nigeria. This insti-tute has the national mandate to conduct research work on thegenetic improvement of sesame. The states visited were Benue,Borno, Gombe, Jigawa, Kaduna, Kano, Kebbi, Kogi, Nassarawa,Niger, Plateau and the Federal Capital Territory (Fig. 1). Ques-tionnaires (see Appendix) were administered through an inter-preter in some cases and samples of available sesame accessionsunder husbandry were collected either directly on the farms orfrom barns at farmer’s homes. The seeds were collected, packedand sealed in thick paper envelopes, each of which was given acollecting code, local name and locality. The seeds were stored ina refrigerator before they were handed over to the National Cere-als Research Institute Badeggi, Nigeria, for conservation andfuture utilization. Identities and collection localities of the mate-rials are summarized in Table 1.

Results and discussionThe survey covered 21 towns and villages in 18 local governmentareas of 11 states and the Federal Capital Territory in Nigeria.Thirty-six farmers identified through the state Agricultural De-velopment Project (ADP) agencies were interviewed and 27 ac-cessions were collected from 20 farmers (Table I). Some of theaccessions were found to be duplications. Collections from Niger,Kano, Nassarawa, Kebbi, Plateau and Gombe States revealedthat many of the accession entries were replicated over states,local government and villages. This was the first extensive collec-tion of sesame germplasm ever carried out in Nigeria.

The highest number of accessions was collected from Benueand Kaduna States. This was closely followed by Nassarawa andNiger States (Table 1). These states had the greatest diversity of thecrop genetic resource. In Benue State the crop was produced incommercial quantities in Guma, Gwer, Katsina-ala, Makurdi andVendeikeja local government areas. This confirms the report ofAgboola (1979) that more than 90% of the crop in Nigeria comesfrom Benue State alone (part of the central zone of the country).

Sesame therefore might have developed from being a crop ofnegligible importance to being one of the major cash crops in theareas of production in Nigeria. More than 60% of the farmerspreferred the white-seeded sesame to other types because theybelieve it is more productive, more useful and has a higher oilcontent. The seed contributes to the diet of the people in theseareas. It is use mainly as a soup-thickening condiment. Tradi-

Fig. 1. Map of Nigeria showing collecting areas.


tionally the seed is roasted and ground together with roastedgroundnut to a pasty consistency. The product is analogous topeanut butter in appearance and partly in flavour.

To avert food shortages and eventual famine, there is a needto pay more attention to our food crops. Mengesha and AppaRao (1991) reported the gradual dwindling of the enormousdiversity which exists within crops. In view of the popularity ofsesame as a food crop in Nigeria, collecting of its germplasmmay save the crop from genetic extinction and provide forfuture utilization. The exercise provided an opportunity todiscover the existing potentials for sesame production in somestates in the country. Some local government areas such asZuru in Kebbi State, Jamaa in Kaduna State and Doma area ofNassarawa State have great potential (in terms of large landarea with deep-draining sandy loam soils) for large-scale pro-duction of the crop. The major constraint pointed out by 50% ofthe farmers was the non-availability of improved varieties ofsesame to farmers in their states. This demonstrates the need todevelop improved varieties of sesame seeds and distributethem to farmers in the producing areas so as to increase theproductivity of the crop. Considering the importance ofgermplasm with broad genetic variability in crop improvementprogrammes, it is suggested that collecting efforts should con-tinue at NCRI Badeggi, Nigeria.

The wide variety of accessions collected in this study togetherwith their great adaptability, as demonstrated by the consider-able diversity of environments under which they are grown, indi-cate that there is great scope for both improvement of existingtypes and production of new varieties of the crop. Scientific

characterization of these collections is necessary to ascertain thegenetic diversity existing within the species in Nigerian. Thepresent assemblage of sesame accessions may therefore consti-tute a reasonable contribution to the backbone information andmaterials required in the improvement of the crop. Results of thisstudy also provide additional challenge for the national andinternational scientific communities to collaborate in the effort toconserve the germplasm of neglected crops.

AcknowledgementThe authors are grateful to the National Cereals Research Insti-tute Badeggi, Nigeria for support and permission to publish thispaper.

ReferencesAgboola, S.A. 1979. An Agricultural Atlas of Nigeria. Oxford

University Press, London.Bruce, E.A. 1953. Notes on African Pedaliaceae. Kew Bull. 69:417-

429.Hutchinson, J. and J.H. Dalziel. 1963. Floral of West Tropical

Africa II. Crown Agents, London.Irvine, F.R. 1969. West Africa Crops. Oxford University Press,

London.Joshi, A.B. 1961. Sesame. Pp. 103-107 in Crops of the West Africa

Semi-Arid Tropics. International Crop Research Institute forthe Semi–Arid Tropics (ICRISAT), India.

Kumar, K., P.D. Bhargava and S.K. Upadhyaya. 1967. Classifi-cation of Rajasthan Sesame. India J. Agric. Sci. 37:193-199.

Mengesha, M.H. and S. Appa Rao. 1991. Status and diversity ofAfrican germplasm collections maintained at the Interna-tional Crops Research Institute for the Semi-Arid Tropics. Pp.45-52 in Crop Genetic Resources of Africa, Vol. I, Proc. Int.Conference on Crop Genetic Resources of Africa (F. Attere, H.Zedan, N.Q. Ng and P. Perrino, eds.). IBPGR, IITA, UNEP.

Table 1. Description and sources of sesame germplasm materials

Accession Local name Source Seed colour Seed length (mm)number

BE-01 Ishwa Otukpo Benue State Creamy White 2–3BE-02 Ishwa Yandev Benue State White 2–3BE-03 Ishwa Markurdi Benue State White 2.5–3BE-04 Ishwa Otukpo Benue State Black 2–3BO-01 Fari Ridi Maiduguri Borno State White 2–3BO-02 Fari Ridi Gozo Borno State White 2–2.5GO-01 Fari Ridi Billiri Gombe State White 2–3GO-02 Beki Ridi Billiri Gombe State Black 2–3.5JG-01 Fari Ridi Gumel Jigawa State White 2–2KD-01 Beki Ridi Kafanchan Kaduna State Black 2–3KD-02 Jai Ridi Kafanchan Kaduna State Light Brown 2.5–.5KD-03 Ijai Ridi Kafanchan Kaduna State Light Brown 1.5–2.5KD-04 Fari Ridi Kafanchan Kaduna State White 2–3KN-01 Fari Ridi Yankaba Kano State White 2–3KN-02 Fari Ridi Janguza Kano State Light Brown 2.5–3KB-01 Fari Ridi Zuru Kebbi State Cream White 2–3KG-01 Igorigo Okene Kogi State White 2–3KG-02 Igorigo Ogaminana Kogi State Cream White 2.5–3NA-01 Fari Ridi Doma Nassarawa State White 2–3NA-02 Fari Ridi Doma Nassarawa State Creamy White 2.5–3NA-03 Fari Ridi Lafia Nassarawa State White 2–2.5NG-01 Esso Bida Niger State Black 2–3NG-02 Fari Ridi Gawu-Babangida Niger State White 2.5–3NG-03 Fari Ridi Minna Niger State White 2–3PL-01 Fari Ridi Pomshere Plateau State Creamy White 2–3PL-02 Fari Ridi Panshin plateau State Creamy White 2–3FCT-01 Fari Ridi Gwagwalada FCT Creamy White 2–3


Purseglove, J.W. 1974. Tropical Crops, Dicotyledons, pp. 430-435. Longmans, London.

Salunke, D.K. and B.B. Desai. 1986. Post harvest Biotechnologyof Oil Seeds, pp. 105-116. CRC Press, Boca Raton, FL.

Van Rheenen, H.A. 1970. Intergeneric hybridization betweenCeratotheca sesamoides End 1 and Sesamum indicum. Nigerian J.Sci. 4(2):251-254.

Weiss, E.A. 1971. Castor Sesame and Safflower. Hill, London.

APPENDIXQUESTIONNAIRES FOR SESAME GERMPLASMCOLLECTION1. State of collection____________________________________________2. Local Government Area______________________________________3. Place of collection____________________________________________4. Date_______________________________________________________5. Local name of the accession___________________________________6. How long have you been growing the crop? ___________________7. How did you get your first supply of the seeds? ________________

___________________________________________________________8. What yields are you getting?_________________________________

___________________________________________________________9. What is the placement of Beniseed in your farming system? _____

___________________________________________________________10. Which type of seed do you prefer? ____________________________11. What constraints are you facing in the cultivation of the crop? ____

_________________________________________________________________________________________________________________________________________________________________________________

12. What are the economic importance of the crop________________________________________________________________________________________________________________________________________

Other comments / observation________________________________________________________________________________________________________________________________________________________________________________________________________________________________


Unexploited diversity in coconut palm(Cocos nucifera L.)V. Arunachalam²? , B. Augustin Jerard¹, M. Elangovan, M.J. Ratnambal², R.Dhanapal², Saran Kumar Rizal and V. DamodaranWorld Coconut Germplasm Centre (CPCRI RC), P.B. 172, Sipighat, Port Blair 744 101 Andaman & Nicobar Islands, India.Email: [email protected]¹ Present address: PhD programme (Horticulture), T.N.A.U. Coimbatore, Tamil Nadu, India² Present address: Central Plantation Crops Research Institute, Kasaragod 671 124 Kerala, India

SummaryUnexploited diversity in coconutpalm (Cocos nucifera L.)The morphological characters of four co-conut palms possessing mutant traits aredescribed: plicata (fused leaflets) bi-spatheate (double spadix cover), persis-tent petioles and inflorescence, and ro-sette seedling with 12 leaves. Of 71 palmsin a Niu Leka Dwarf population, onepalm (no. 5/36) has a high degree offused (plicata) leaflets (>70%) and twopalms are bi-spatheate (nos. 5/18 and 5/16). The plicata palm had a long juvenileperiod (13 years), which agrees withother reports. The bi-spathe palm hasreduced leaf length, also in agreementwith earlier reports. A palm in Hut Bay,Little Andaman having a persistent peti-ole, and a rosette seedling in a farmer’sgarden are also described. Reports onplants with such novel traits supportedby morphological, genetic, biochemical,cytological and ecological data are dis-cussed. These traits result mainly frominbreeding or recessive mutations. Afterdetailed investigation on these mutanttraits, they could serve as markers inidentifying inbred palms or palms withrecessive alleles in a population. Thiscould help to exploit the hybrid vigour incoconut palm.

Key words: Bi-spatheate, coconut,Cocos nucifera, plicata, WorldCoconut Germplasm Centre (WCGC)

ResumenDiversidad no explotada delcocotero (Cocos nucifera L.)Se describen las característicasmorfológicas de cuatro cocoteros conrasgos mutantes: plicata (hojasplegadas), bi-spatheate (doble capa deespata), peciolos e inflorescenciapersistentes, y plántula en roseta con 12hojas. De 71 cocoteros en un rodal de NiuLekha Dwarf, uno (n° 5/36) tiene unaalta proporción de hojas plegadas(plicata) y dos cocoteros son bi-spatheate(n° 5/18 y n° 5/16). El cocotero plicatatuvo un largo periodo juvenil (13 años),lo que concuerda con otros informes. Elcocotero bi-spathe tiene hojas cortas,también de acuerdo con informesanteriores. Se describen asimismo uncocotero en Hut bay, Little Andaman,con un peciolo persistente, y una plántulaen forma de roseta en el jardín de ungranjero. Se analizan Informes sobreplantas con rasgos novedosos apoyadosen datos morfológicos, genéticos,bioquímicos, citológicos y ecológicos.Estos rasgos resultan principalmente dela endogamia o de mutaciones recesivas.Una vez investigados detenidamenteestos rasgos mutantes, podrían servircomo marcadores para identificarcocoteros endogámicos o cocoteros conalelos recesivos en un palmeral, lo quepodría ayudar a explotar el vigor híbridodel cocotero.

RésuméUne diversité non exploitéechez le coconut palm (Cocosnucifera L.)Les charactéristiques morphologiques dequatre cocotiers possédant des traits mu-tants sont décrits: plicata (feuilles en fuse-au) bi-spatheate (couverture par un dou-ble spadice), petiole persistant et inflores-cence, et rosette faisant des graines avec12 feuilles. De 71 palmiers d’une popula-tion de Niu Lekha, un palmier (no. 5/36)atteint un haut degree de feuilles en fuse-au (plicata) (>70%) et deux palmiers sontbi-spatheate (nos. 5/18 et 5/16). Le palm-ier plicata a eu une longue prériode juve-nile (13 ans), ce qui est en accord avecd’autres rapports. Le palmier bi-spathe ades feuilles de longueur réduite ce qui estégalement en accord avec des rapportsprécédents. La description est égalementfaite d’un palmier à Hut Bay, Little Anda-man ayant un petiole persistant et unerosette plantule chez un fermier. Desrapports contenant les données mor-phologiques, génétiques, biochimiques,cytologiques et écologiques sur des culti-vars ayant ces traits recherchés sont dis-cutés. Ces traits sont principalement leproduit d’une reproduction intraspéci-fique ou de mutations recessives. Aprèsune investigation détaillée de cescharatères mutants, ils pourraient êtreutilizes comme marqueurs dansl’identification de palmier issu de repro-duction intraspécifique ou des palmierspossédant des alleles recessives dans unepopulation donnée. Ceci pourrait aiderl’exploitation de la vigueur des hybrideschez le cocotier.


IntroductionCoconut palm is one of the multipurpose perennial crops of thehumid tropics. It is grown in more than 80 countries in an area ofnearly 11 million hectares. The crop supports small and marginalfarm families living in the Developing World, especially in and ispart of the fragile ecosystems of small islands. Genepool enrich-ment is one of the proven methods of coconut improvement. Acompilation by Batugal and Ramanatha Rao (1998) givesprogress on germplasm collections of different countries.Bourdeix et al. (1989) suggest the use of individual combiningability tests to obtain planting materials of 20–30% superiorityover currently available hybrids. The efforts made on coconut

improvement so far have concentrated on utilizing variation overa geographical range. However, the quarantine risk of germplasmexchange poses a serious limitation in exploiting that variation.Therefore it is important to use morphological diversity within aregion to its maximum potential. The use of Macapuno (the jelly-like endosperm) in the food industry is a good example of suc-cessful utilization of morphological diversity. There are othermorphological variants (Menon and Pandalai 1960) such as albi-nism (lack of chlorophyll), aromatica (fragrant endosperm),change in sex expression and plicata (fused leaflets). These traitshave not been used to their potential. This report discusses such


variants and the relevance of these rare traits to the plant breeder.Arunachalam (1999) reviewed the range of variability of morpho-logical traits in coconut (Table 1).

The variations in such traits could be either genetic or physiologi-cal in nature. Tall and Dwarf are the two morphological contrastingforms of coconut which have given sufficient scope in hybrid produc-tion to lead to a reduction in the pre-bearing age and improvement innut yield. Rare traits such as plicata, late flowering / bearing, bi-spatheate (spadix covered by two spathes instead of one) andsecondary spikelets (further branching of spikelets) have been re-ported in Talls. Dwarfs possess other rare traits such as poly-embryony, vivipary (general observation) and pigment variation inleaf / nuts (de Lamothe and Rognon 1977; Ratnambal et al. 1995).

Materials and methodsIn the World Coconut Germplasm Centre (WCGC), SipighatFarm located in Sipighat village of Port Blair, Andaman &Nicobar Islands, India, 20 Tall and 4 Dwarf exotic accessions ofcoconut collected from six Pacific Ocean territories and 6 Tallindigenous accessions from Nicobar islands are being maintained.Out of the 4 Dwarf accessions, Niu Leka (Fig. 1) Dwarf (FijiIslands), is an important accession because it combines the traitsof Dwarfs—short internodal distance and early flowering/bear-ing—with that of Tall variety traits such as allogamy and largefruits. There are 71 palms in this population, which are 17 yearsold. One palm (no. 5/36) has shown a high degree of fused(plicata) leaflets (>70%), two palms have shown bi-spathes (nos.5/18 and 5/16). Apart from these, another two palms and twoseedlings were studied for the following traits of interest:

● A palm of local Tall at Hut Bay, Little Andaman, Indiapossessing persistent petioles and inflorescences (Rao andSampathkumar 1998)

● A palm showing bispatheate and secondary spikeletswith leaves of plicata nature (Fig. 2) at CPCRI field trials,Kasaragod, Kerala, India (TT 50)

● Two seedlings with rosette leaves (Fig. 3).

Results and discussionMorphological details of the palms having mutant traits are

Table 1. Reported variants for morphologicaltraits in coconut palm

Variants

Habit Tall, dwarf and intermediate formsLeaf Normal, albinism, plicata, small or long

leaf/leafletFloral Vegetative transformation (bulbils),

midget/hapxanthic, very early to verylate flowering/bearing, spicata,secondary spikelets, apocarpy,persistent inflorescences on stem

Sex expression (Dioecy) Male/female, male sterility,hermaphrodite

Fruit Jelly like endosperm, fragrantendosperm, edible huskVariable size of fruit (<50 g to 3 kg) andhusk proportion (20-80%)Beak, ring or horn-like structures on fruit

Fig. 1. Niu Leka dwarf.

Fig.2. TT 50 coconut palm.

Fig. 3. Rosette seedlings.

given in Table 2. The details of palms having such mutant traitsreported by earlier workers along with the genetic, biochemical,ecological, cytological basis are discussed below.

PlicataPlicata is a situation in which the leaflets remain fused, as seen inseedling leaves (Sugimura et al. 1994b). Palm no. 5/36 of WCGChas the characteristic plicata leaves (recapitulation/embryonicsimilarity). It remained in the vegetative phase for 13 years. As ithas been reported that boron deficiency can cause similar symp-toms in coconut, palm 5/36 was given a regular application ofborax (50 g/plant annually). In contrast to the palms sufferingwith boron deficiency, plicata palms applied with boron did notshow any response. In spite of the treatment, it produced onlyone inflorescence without fruits. However, species and varietal


differences in uptake and translocation of boron have been ob-served in tomato (Bellaloui and Brown 1998). In Oil Palm, this iscontrolled by a single recessive gene (Zeven 1964). The genetics ofthis trait in coconut have not been studied yet. In future we couldanalyze this trait to study its heritability, so that the trait could beutilized better.

The degree of fusion of leaflets in any leaf of this palm wasmore than 70%. Poor people use plaited coconut leaves as hous-ing materials in coastal areas. Leaves of these variant palmscould be used in those areas to save work, as plaiting (knitting theleaflets by hand) the leaves is not necessary.

Amphiblastic species produce two types of leaves. One form ofleaf structure persists only during the juvenile stage. The otherform persists throughout the rest of the plant’s life. In species fromdry regions, seedling type of leaf production stops at a relatively

early age (Stebbins1964). Plicata palmsare known to beginflowering late com-pared with otherpalms. A delay of 29months in floweringwas observed in aplicata palm bySugimura et al.(1994b). The cytologi-cal nature of an abnor-mal palm, which re-mained without fruitseven after 14 years, wasanalyzed and found tohave meiotic abnor-malities leading to ste-rility (ThankammaPillai and Vijayakumar1972). Plicata palm is

Table 2. Morphological characterization of palms with mutant traits

Character 5/18 regular 5/16 recent 5/36 plicata TT 50 Persistentbi-spathe bi-spathe (bi-spatheate inflorescence

+ plicata + palm at Littlesecondary Andamanspikelets)

Plant height (cm) 265 260 453 973 1200Girth of collar (cm) 106 107 110 108 213No. leaves on crown 20 21 14 23 15Petiole length (cm) 115 126 116 135 145Length of leaf-bearing portion (cm) 266 275 244 370 310No. leaflets 196 215 183 86 98Breadth of leaflet (cm) 5.4 5.7 5.6 7.6 3.8Length of leaflet (cm) 96 100 85 130 100No. leaf scars in 1 m 32 45 24 17 12Internodal length (cm) 3.12 2.224.16 4.16 5.88 8.12Length of inflorescence (cm) 98 95 NA 90 102Length of spikelet-bearing portion (cm) 50 52 NA 50 33Length of spikelet (cm) 23 23 NA 40 35No. spikelets/inflorescence 66 58 NA (204)† 12 (122) †51No. female flowers/inflorescence 34 58 NA 14 90

NA—Not available †Figures in parentheses are number of secondary spikelets.

reported to have sterility (Mao and Lai 1993), delayed flowering(Sugimura et al. 1994b) and no female flower/fruit formation. Apalm having similar traits at Batlagundu village (Dindigul Dt.Tamil Nadu, India) in a farmer’ s garden (Fig. 4) also had pro-duced no fruits 15 years after planting.

Multi-spatheateCoconut palm produces a single spathe to cover an individualinflorescence. In some rare palms, two to five spathes are seencovering a single inflorescence. These palms are known as multi-spatheate palms. Multispatheate palms have been reported bymany workers to occur in West Coast Tall populations:bispatheate by Davis and Menon (1952), tri-spatheate by Tho-mas and Mathew (1960) and penta-spatheate by Michael (1963).Multi-spatheate palms have been reported at Hainan Island ofChina (Mao and Lai 1993). These multi-spatheate palms providesufficient mechanical strength to the inflorescence to protect itfrom insect attack in the early stage of development (Michael1963). Presence of this trait may offer protection against windand reduce buckling (instead of being erect) of bunches. In PortBlair, the wind velocity ranges from 4.5 to 19.0 km/h, reaching apeak in June of 17–19 km/h. During cyclones, the wind velocityis very high.

At WCGC, palm 5/18 produced bispathes in every inflores-cence. This palm is a very poor yielder. The other palm (5/16),which was a good regular bearer, has recently started the habit ofproducing bi-spatheate inflorescences (Fig. 5). This palm is grow-ing on a portion of undulated land and there is a serious problemof wind damage causing buckling of bunches.

It triggers the interest of a breeder when the trait is seen in anew palm of the same population at a later stage. The implica-tion of this trait offering evolutionary fitness and the nature ofdevelopmental genetics of the trait needs to be understood. Aregular bi-spatheate palm at WCGC had brown petioles, whichindicates its hybrid origin (possibly involving green and othercoloured types).Fig. 4. Plicata palm.


Change in sex expressionCoconut palm is monoecious in nature, producing male andfemale flowers on spikelets in the same inflorescence. Female-dominant palms have no spikelets and are known as spicata.Male-dominant palms show branching of spikelets and areknown as androgena. Spicata palms were reported in the Philip-pines by Sugimura et al. (1994c). Two of the five inbred lines ofMarkham Valley Tall had mutants (Sugimura et al. 1994a)showing secondary spikelets but with a normal pre-bearingperiod. This indicates the trait to be due to either recessivealleles or inbreeding depression. Secondary spikelets were al-ways known to occur with an increased number of male flowers.Open-pollinated progenies of spicata palm segregated in 1:1ratio of spicata and normal palms. Thus, earlier workers as-sumed the spicata trait to be in the (Ss) heterozygous state anddominant over normal (ss) inflorescence. Spicata palm also hasshown many meiotic irregularities and a high degree (12%) ofpollen sterility (Ninan and Satyabalan 1963). Androgena palmshave shown meiotic abnormalities such as aneuploidy (2n–1) ina few cells (Ninan et al. 1960).

Palms with persistent petiole bases /inflorescencesA palm in Hut Bay of Little Andaman showed persistent petiolebases and inflorescences. This character is common in other palmssuch as Palmyrah (Borassus flabellifer). When we characterized thispalm for morphological traits, the collar girth was very large(213 cm). Trunk diameter is an important character in classifyingcoconut forms. Self-pollinated Dwarfs are known to have slendertrunks. Selfing up to three generations in families of MapangetTall (Indonesia) has reduced trunk diameter (Novarianto et al.1991). This palm started bearing very late and the inflorescenceshad secondary spikelets (Fig. 6).

Rosette seedlingsTwo-rosette seedlings found in a farmer’s garden at Batlagunduvillage (Dindigul Dt. Tamil Nadu, India) had an unusually highnumber of leaves and large collar girth (Fig. 3). Seedlings withhigh collar girth and large number of leaves are known to havegood yield potential (Satyabalan and Mathew 1984). As bulkseed nuts were used, the parental palms which resulted in theseseedlings could not be traced. However the performance of theseseedlings could be studied in the future.

Biochemical / molecular markersIn makapuno mutant endosperm, the activity of alpha galactosi-dase enzyme was found to be 8300-fold lower than the normalendosperm. Makapuno and normal endosperms also differed inthe activity of peroxidase and tryptophan aminotransferase en-zymes (Mujer et al. 1984).

Molecular markers were helpful in analyzing the genetic di-vergence and understanding the origin and domestication pro-cess (Perera et al. 2000). Furthermore Teluat et al. (2000) noticedthe heterogeneity of Niu Leka Dwarf (Fiji) using (AFLP andSSRs) molecular markers. They attributed it to the cross-pollinat-ing nature of the accession.

Relationship with other traitsThe measurements of morphological characters of these rarepalms improve the understanding of relationship between traits.The palms with Horned fruits (Jerard et al. 1999) had the widestleaflets (7.5 cm) and a harder shell (0.45 cm). Generally, theleaflet breadth in coconut ranges from 4.0 to 6.0 cm and shellthickness varies from 0.2 to 0.3 cm (Ratnambal et al. 1995). Thepalm TT 50 (Fig. 2) having plicata, bispatheate and secondaryspikelets also had the widest leaflets (7.6 cm) and as hard a shell(0.4 cm) as the Horned fruit type. Reduced leaf length was alsoobserved in dual and multi-spatheate plams by Mao and Lai(1993). During the germplasm survey in the Seychelles, Kumaranet al. (1998) observed that the mutant Macapuno type havingjelly-like endosperm (‘Coco Gra Tall’) had reduced leaflet length.This macapuno trait results in abortion of embryo and is known

Fig. 5. Bi-spatheate inflorescence.

Fig. 6. Secondary spikelets.


to be due to a lethal recessive gene (Cedo et al. 1984). Embryoculture could rescue plants with this trait. Reduced leaflet lengthalso was observed in spicata palms in CPCRI field trials. Nor-mally self-pollinated Dwarfs also have short leaf / leaflet. Selfingup to three generations in family 55 of the Mapanget Tall (Indo-nesia) form showed reduction in leaf and petiole length(Novarianto et al. 1991). Of the four pure F2 Macapuno bearingpalms (progenies of Dwarf x Macapuno Tall), two showed verygood vigour and early (36–37 months) flowering (Nunez and De-paz 1994). This indicates the importance of the mutant traits inexploiting the hybrid vigour / transgressive segregation.

Generally Dwarfs are early bearing. Hybrid progeny betweenTall and Dwarf accessions generally resemble the Dwarf parent inearly bearing trait. Dominance of long leaf, early bearing andsingle branching over short leaf, late bearing and secondary spike-lets respectively is generally seen in the hybrid progenies. How-ever, a detailed investigation on the genetics of traits associatedwith rare palms such as late bearing, short leaves and secondaryspikelets is required to confirm this idea. Selfing of the palms 5/18 and 5/16 at Port Blair and a palm (TT-50) with secondaryspikelets is in progress, which is expected to yield interestingresults. Molecular biological investigations of these traits couldprovide additional information.

Markers for inbred palmsThis information hints at the possibility of linkage between noveltraits and recessive alleles or of those of an inbred nature. Devel-opment of inbred lines in a perennial crop such as coconut wouldtake decades as there is no viable vegetative propagation methodavailable in this crop. The inferences from the study will aid inlocating palms of use in hybrid development using morphologi-cal markers. The results after confirmation have practical impli-cations in aiding the coconut breeder.

AcknowledgementsThe authors are grateful to Dr K.U.K. Nampoothiri, Director,CPCRI and Dr V.A. Parthasarathy, Head, Division of CropImprovement, CPCRI for the facilities. The first author (V.A.) isgrateful to IPGRI for providing financial aid from the CanadianInternational Development Agency (CIDA) to present theseresults at the SAT 21 Conference at Kuala Lumpur, Malaysiaduring June 2000.

ReferencesArunachalam, V. 1999. Prospection and collection of coconut

diversity. Pp. 12-19 in Improvement of Plantation Crops(M.J. Ratnambal, P.M. Kumaran, K. Muralidharan, V. Niraland V. Arunachalam, eds.). CPCRI, Kasaragod.

Batugal, P.A. and V. Ramanatha Rao (eds.). 1998. Coconutbreeding. Papers presented at a workshop on standardiza-tion of coconut breeding research techniques, 20–25 June1994, Port Bouet, Côte d’Ivoire. IPGRI, APO, Serdang, Ma-laysia.

Bellaloui, N. and P.H. Brown. 1998. Cultivar differences in boronuptake and distribution in celery (Apium graveolens), tomato(Lycopersicon esculentum) and wheat (Triticum aestivum ). Plantand Soil 198(2):153-158.

Bourdeix, R., A. Sangare et J.P. Le Saint. 1989. Efficacité des testshybrids d’aptitude individuelle à la combination chez lecocotier: Premiere resultants. Oleagineux 44:209-214.

Cedo, MA, E.V. De Guzman and T.J. Rimando. 1984. Controlled

pollination of embryo cultured makapuno coconut (Cocosnucifera L.). Phil. Agric. 67:100-104.

Davis, T.A. and K.P.V. Menon. 1952. A bi-spatheate coconutpalm. Indian Coconut J. 6:30-34.

Jerard, B.A., R. Dhanapal, V. Damodharan and V. Arunachalam.1999. Horned Cocos nucifera L. Plant Genet. Resour. Newsletter(in press).

Kumaran, P.M., P.K. Koshy, V. Arunachalam, V. Niral and V.A.Parthasarathy. 2000. Biometric clustering of coconut popula-tions from three Indian Ocean islands. Pp. 73–81 in Proc.PLACROSYM XIII. Recent Advances in Plantation CropsResearch (N. Muraleedharan and R. Rajkumar, eds.). AlliedPublishers, India.

de Lamothe, Nuce De M. and F. Rognon. 1977. The dwarf coco-nuts at Port Bouet. Oleagineux 32(8-9):373-375.

Mao, Z. and Y. Lai. 1993. The coconut germplasm of HainanIsland, China. Plant Genet. Resour. Newsletter 91/92:53-57.

Menon, K.P.V. and K.M. Pandalai. 1960. Coconut palm—a mono-graph. Indian Central Coconut Committee, Ernakulam.

Michael, KJ. 1963. A multi-spatheate coconut palm. Indian Co-conut J. XVI (2):78-80.

Mujer, C.V., D.A. Ramirez and E.M.T. Mendoza. 1984. alpha-D-galactosidase deficiency in coconut endosperm: its possiblepleiotropic effects in makapuno. Phytochemistry 23(4):893-894.

Ninan, C.A. and K. Satyabalan. 1963. Cytogenetic studies inCocos. II. Some observations on the spicata character in coco-nuts (Cocos nucifera L.). Indian Coconut J. XVI (3):109-114.

Ninan, C.A., R.V. Pillai and Josy Joseph. 1960. Cytogenetic studiesin the genus Cocos. I. Chromosome number in C. nucifera L. vars.Spicata and androgena. Indian Coconut J. XIII (4):120-132.

Novarianto, H. Miftahorachman and H.T. Luntungan. 1991. Ef-fect of inbreeding on some characters of the Mapanget Tallcoconut. Industrial Crops Res. J.3 (2):15-17.

Nunez, T.C. and V.M. De-Paz. 1994. Growth and development ofF2 pure macapuno palms. Phil. J. Crop Sci. 19(S.1):69.

Perera, L., J.R. Russel, J. Provan and W. Powell. 2000. Use ofmicrosatellite DNA markers to investigate the level of geneticdiversity and population genetic structure of coconut (Cocosnucifera L.). Genome 43(1)15-21.

Rao, P.S.N. and V. Sampath Kumar. 1998. Some botanical curi-osities. Curr. Sci. 75(2):91-92.

Ratnambal, M.J., K. Muralidharan, N.K. Nair, P.M. Kumaran,E.V.V. Bhaskara Rao and R.V. Pillai. 1995. Coconut descrip-tors—part I. CPCRI, Kasaragod.

Satyabalan, K. and J. Mathew. 1984. Identification of prepotentpalms in West Coast Tall coconuts based on the early stagesof growth of the progeny in the nursery. Pp. 15-22 in CoconutResearch & Development (N.M. Nayar, ed.). Wiley EasternLimited.

Stebbins, G.L. 1964. Variation and Evolution in Plants. OxfordBook Co.

Sugimura, Y., D.A. Rocat, C.D. Salud and N. Kamata. 1994a. Aninflorescence mutant appeared in the inbred line of coconut “Markham” Jpn. J. Trop. Agric. 38(2):145-147.

Sugimura, Y., D.A. Rocat, C.D. Salud and N. Kamata. 1994b.Characteristics of coconut palms with plicate leaves. Jpn. J.Trop. Agric. 38(2):119-123.

Sugimura, Y., D.A. Rocat, C.D. Salud and N. Kamata. 1994c.Multi-varietal characteristics of spicata coconut palm. Jpn. J.Trop. Agric. 38(2):264-268.

Teluat, B., C. Aldam, R. Trehin, P. Lebrun, J.H.A. Barker, G.M.Arnold, A. Karp, L. Bauoduin and F. Rognon. 2000. An analy-sis of genetic diversity in coconut populations from acrossgeographical range using sequence tagged microsatellites(SSRs) and AFLP. Theor. Appl. Genet. 100:764-771.

Thankamma Pillai, P.K. and G. Vijayakumar. 1972. Cytology of anabnormal palm. Indian J. Genet. & Plant Breed. 32(2):303-305.

Thomas, C.A. and C. Mathew. 1960. A tri-spatheate coconutpalm. Bull. Indian Central Coconut Committee 7:256-259.

Zeven, A.C. 1964. The Idolatrica. Baileya 12:11.


Genetic variation in populations of cereal cropsin the southern coastal zone of YemenA.A. BawazirCollege of Education, University of Hedhramout for Science & Technology, Mukalla, Republic of Yemen

SummaryGenetic variation in populationsof cereal crops in the southerncoastal zone of YemenA report of germplasm collecting andcharacterization in the agroecologicalzones of coastal Yemen. In particular,landraces and farmer varieties of millet,sorghum and maize were identified.

Key words: Cereals, germplasmcollecting Yemen

ResumenLa variación genética encereales en la zona costeradel sur de YemenUn informe sobre la recolección degermoplasma y la caracterización en zo-nas agroecológicas de la zona costera deYemen. En particular, las especies localesy variaciones de granjeros de mijo, sorgoy maíz fueron identificadas.

RésuméVariation génétique dans lespopulations céréalières de lazone côtière sud du YemenUn rapport de collecte et de caratérisationde materiel génétique dans les zonesagroécologiques de la côte au Yemen.Des variétées traditionnelles de millet, desorgho et de maïs ont particulièrementété identifié.

IntroductionYemen, with its unique geographical location, diverseagroecological zones and long history of agricultural develop-ment, harbours a wealth of plant genetic resources. Of the 3418plant species in the flora of the Arabian Peninsula, 2500 arefound in Yemen. The main crop groups cultivated in Yemen arecereals, legumes, vegetables, coffee and ghat. Cereals are grownon almost 60% of the total cultivated area, with sorghum andmillet occupying 80% of this area.

The southern coastal strip of Yemen varies in width from 1 to80 km. The weather is hot in summer, with an average annualrainfall of <100 mm. Most cultivation is based on spate irriga-tion and is largely confined to the deltas of large wadis in Lahajand Abyan districts (e.g. Tuban, Abyan and Ahwar deltas),where soils are calcareous, silty to sandy loams. However, irriga-tion from wells is more predominant at Gail Bawazir andHadhramout districts (Mukred et al. 1991).

Sorghum is the main cereal crop grown in this zone and isgrown predominantly under spate irrigation as a dual-purposecrop for grain and fodder in the coastal region. However, under-ground water is used to irrigate smaller areas of sorghum in themedium altitude regions of the southern coastal strip, whereasstill smaller high-altitude areas are rainfed (Mu’allem andBawazir 1980).

Millets and maize are minor cereal crops in the coastal strip,and are used for human food (maize is also used as poultryfeed). Millets, similarly to sorghum, are grown under a widevariety of climatic conditions, predominantly cultivated underspate irrigation in coastal regions. Maize is almost exclusively

SHORT COMMUNICATION Plant Genetic Resources Newsletter, 2001, No. 127: 44 - 45

grown under groundwater irrigation in coastal regions. It is arelatively new crop to the area and has good potential whereirrigation water is available (Mu’allem 1988).

Collecting germplasmGermplasm collecting in the southern parts of the country datesback to 1969. Between 1969 and 1987, several collecting expedi-tions were launched by the Agronomy and Plant Breeding Sectionat El-Khod Agricultural Research Center and resulted in a total of351 accessions of landraces and farmer varieties, most of whichwere local sorghum and millet landraces (Mu’allem 1981, 1988,Bawazir 1988, Mukred et al. 1991). However, geographical cover-age was not adequate and some regions were not explored.Moreover, most of the earlier collections were not conserved underappropriate conditions and they lost their viability or were lostaltogether over the years. Nevertheless, some of these earliercollections were evaluated for some agronomic characteristics(Table 1).

Local cultivars of millets belong to Pennisetum glaucum, P.setaceum (Pearl millet), P. rigidum (little millet) and Eleusine coracana(finger millet). They are tall with small heads and reduced tillers.Finger millet is medium in height with small heads split into fivebranches. Three local cultivars were identified, namely: Maseibli(Pearl-type), Kenab (Finger millet-type) and Hibah (Little millet-type). Maseibli is known as Dokhon in the coastal area ofHadhramout. Maseibli and Dokhon are the local names for pearland finger millet, respectively; all over the region Hibah is the localname for the little millet in Socotra Island (Mu’allem 1988).

Table 1. Some morphological characteristics of sorghum germaplasm grown at different climatic regions inYemen (Mu’allem 1981)

Cultivar Climatic region Plant height (cm) Crop duration (days) Panicle type Grain colour(local name) (0–200m)

Beini Coastal areas of Abyan 250–300 100–110 Open WhiteBuker Lahej 250–300 100–110 Open Light redSaif Ahwar (0–200m) 200–260 100–110 Open White


Local cultivars of Zea mays L. are medium in height withsmall yellow-grained cobs and a low yield capacity. One localcultivar in the coastal region was identified, namely Hind(Mu’allem 1988).

Genetic erosionLosing available germplasm is a serious risk for most of thecereals grown in the country because of the introduction of high-yielding exotic cultivars and hybrids. Sorghum is relativelyunthreatened by genetic erosion at present; modern dual-purposevarieties do not impose a threat, but local landraces will bereplaced if and when enough seed of higher-yielding dual-pur-pose exotic varieties is available to farmers (Bawazir 1988,Mukrid et al. 1991).

As there is limited variation in the population of the localvarieties of maize and millet, intrapopulation selection is notexpected to lead to a substantial improvement in yield. Farmersreadily adopted the newly introduced high-yielding varieties ofmaize and millet. High yielding introduced varieties now coveralmost the entire planted area, with limited area sown to theHindi landrace in the coastal region (Mu’allem 1988).

Several varieties of millet have been introduced and includedin the varietal trials to identify high-yielding and early maturingvarieties. Ttwo strains, Dwarf Senegal and Tall Senegal, weredeveloped (Bakshi et al. 1972, Mu’allem 1988).

Loss of local germplasm of maize is imminent; simply be-cause farmers are adopting exotic high yielding varieties at arapid pace and, as a consequence, valuable landraces are beinglost forever (Mu’allem 1988, Mukred 1991). In contrast, locallandraces of millets are persistent and under no danger of beingreplaced by exotic germplasm. Farmers are satisfied with theperformance and adaptation of these landraces, and they prac-tice a yearly cycle of seed selection for the next crop in their ownfields (Mu’allem 1988).

Future actionThere is a need to explore all parts of this unique region inYemen and collect valuable landrace cereals and other cropsadapted to the agroclimatic conditions of the region. Threat ofgenetic erosion is increasing, as new and high-yielding varietiesare being introduced.

There is a need for continuous surveying, collection andclassification, and technical and financial support is needed toestablish a long-term (ex situ) conservation unit. Technical coop-eration and training are needed to classify accessions properlyand to document genetic resources data.

ReferencesBakshi, J.S. et al. 1972. Annual progress report, season 1971/

1972. Agronomy and Plant Breeding Sections, El-kod Agric.Research.

Bawazir, A.A. 1988 Genetic resources of cereal crops in PDRYemen. 1. Sorghum. Plant Genet. Resour Newsletter 72:31.

Mu’allem, A.B. and Bawazir, A.A. 1980 Production of breedersfoundation and certified seeds of sorghum. Report presentedat the 2nd Seed Tech. Course of the National Seed Multiplica-tion Project on Improved Seed production in Coastal Regionof PDRYemen.

Mu’allem, A.B. 1981. Sorghum germplasm in PDRYemen P1antGenet. Resour. Newsletter 47:9-13.

Mu’allem, A.B. 1988. Genetic resources of cereal crops inPDRYemen. 2. Barley, millet and maize. Plant Genet. ResourNewsletter 72:32-33.

Mukred, A.W., L. Guarino, A.B. Mu’allem and A.S. Al-Ghaz.1991. Crop collection in PDRYemen 1988–1989. Plant Genet.Resour Newsletter 83/84: 29-30.


Variability in seed blankness inPistacia atlantica Desf. in a natural habitatH. Mirzaie-Nodoushan? and H.M. ArefiMinistry of Jahade Sazandegi, Forests and Rangelands Research Institute, PO Box 13185-116, Tehran, Iran.Email: [email protected]

SummaryVariability in seed blankness inPistacia atlantica Desf. in anatural habitatPistacia atlantica Desf. is scattered in a vastarea of natural semiarid woodlands ofIran. Although soil conservation andgum production are regarded as themain functions of this species, its seed oilalso merits attention. Seed blankness(hollowness, seeds with empty space in-side) in mature seeds varies among fe-male plants. It is one of the main prob-lems in P. atlantica that bring about de-crease in its value for oil production. Thisstudy was undertaken to investigate theproblem of seed blankness and to deter-mine whether its cause is genetically con-trolled and how much genetic variationexists within a population of the species.Seed was collected from 18 female plantsof a P. atlantica population in a naturalhabitat. The percentage of seed blank-ness and seed weight (g/100 seeds) weremeasured on the collected seeds. Num-ber of male plants located close to each ofthe 18 female plants in an area of 50-mdiameter around the female trees andtheir total and average distances to thefemale plants were measured. Severalmorphological characters of the femaleplants such as plant height, trunk diame-ter at breast height, and crown diameterwere also recorded. Significant differenc-es in seed blankness between the studiedfemale plants were determined throughanalysis of variance. The correlation be-tween seed blankness, average spacingbetween male and female plants, andfemale morphological attributes werecalculated. Female plants differed signifi-cantly in seed blankness, varying from7.8 to 97.5%. Seed blankness did not havea significant correlation with the studiedmorphological attributes, but did have aweak correlation with surrounding maleplants (? =10%).

Key words: Genetic variability, Iran,phenotypic correlation, Pistaciaatlantica L., seed blankness

ResumenVariabilidad en la vaciedad delas semillas de Pistaciaatlantica Desf. en un hábitatnaturalLa Pistacia atlantica Desf está extendidapor una dilatada zona boscosa semiáridade Irán. Aunque la conservación del sue-lo y la producción de goma se considerancomo las funciones principales de estaespecie, el aceite de su semilla tambiénmerece atención. La vaciedad de las semi-llas maduras (los espacios vacíos en suinterior) es variable de unas plantas hem-bras a otras, y es uno de los mayoresproblemas de la Pistacia atlantica, que leresta valor para la producción de aceite.Este estudio se emprendió para investi-gar el problema de la vaciedad de lassemillas y determinar si su causa estácontrolada genéticamente y hasta quépunto hay variación genética dentro deuna población de la especie. Se recogi-eron semillas de dieciocho plantas hem-bras de una población de P. atlantica enun hábitat natural. Se midieron los por-centajes de vaciedad y peso de las semill-as (g./100 semillas) en las semillas recogi-das. También se midieron el número deplantas masculinas situadas cerca de cadauna de las 18 plantas hembras en uncírculo de cincuenta metros de diámetroen torno a éstas y sus distancias totales ymedias hasta ellas. Se registraron asimis-mo varios caracteres morfológicos de lasplantas hembras como altura, diámetrodel tronco a la altura del pecho y diámetrode la copa. Las diferencias significativasen vaciedad de las semillas entre las plan-tas hembras estudiadas se determinaronmediante análisis de varianza. Se calcu-laron la correlación entre vaciedad desemilla, distancia media entre plantasmasculinas y femeninas y atributos mor-fológicos de las plantas hembras. Éstasresultaron ser notablemente diferentesen cuanto a sus semillas vacías, que oscil-aban entre 7,8% y 97,5%. No había unacorrelación apreciable entre vaciedad delas semillas y los atributos morfológicosestudiados, pero había una débil correl-ación con el entorno de plantas masculi-nas (? =10%).

RésuméVariabilité de la proportion degraines vides produites parPistacia atlantica Desf. dansun habitat naturelPistacia atlantica Desf. a une distributionlâche dans une vaste zone semi-aride na-turellement boisée en Iran. Non seule-ment cette espèce joue un rôle importantdans la protection du sol et la productionde résine, mais l’huile extraite des grainesprésente des perspectives intéressantes.La proportion de graines vides (grainescreuses à l’intérieur) à maturation estvariable parmi les arbres femelles et con-stitue l’un des obstacles majeurs limitantl’utilisation de P. atlantica pour la produc-tion d’huile. On ignore si le caractère estgénétiquement déterminé et quel est ledegré de variation génétique existant ausein d’une population donnée. Laprésente étude cherche à répondre à cesquestions. Des graines ont été collectéessur dix-huit arbres femelles d’un peuple-ment de P. atlantica dans un habitat na-turel. Le poids des graines (g /100 graines) et le pourcentage de grainesvides ont été déterminés sur les échantil-lons récoltés. Le nombre d’individusmâles croissant dans un cercle decinquante mètres de diamètre autour dechacun des 18 arbres femelles, ainsi queleurs distances totales et moyennes parrapport aux individus femelles ont étémesurés. Plusieurs caractères mor-phologiques des arbres femelles, tels quela hauteur, le diamètre du tronc (à hau-teur de la poitrine) et le diamètre de lacouronne ont également été enregistrés.Des différences significatives dans lesproportions de graines vides entre lesarbres femelles étudiés ont été déter-minées par une analyse de variance. Lacorrélation entre la proportion de grainesvides, la distance moyenne entre les indi-vidus mâles et femelles et les caractèresmorphologiques femelles a égalementété calculée. Il existe une différence sig-nificative entre la production de grainesvides par les plantes femelles, celle-ci vari-ant de 7,8 % à 97,5 %. En revanche, au-cune corrélation significative n’a été ob-servée entre la proportion de grainesvides et les caractères morphologiquesétudiés. Il existe cependant une faiblecorrélation entre le nombre d’arbresmâles dans le voisinage et la proportionde graines vides (? =10 %).

SHORT COMMUNICATION Plant Genetic Resources Newsletter, 2001, No. 127: 46 - 48


IntroductionPistacia atlantica Desf. is a drought-tolerant tree species, which isadapted to arid and semiarid areas (El-Moslimany 1986). Thisspecies is widely distributed in these area of the world, can evenbe found in harsh areas in which few tree species can be grownand established. This is a remarkable characteristic of the specieswhich makes it suitable for arid and semiarid forestry. Thespecies is a multipurpose tree that can be used in soil conserva-tion, industrial and medicinal applications. Several medicinaluses have been recorded for the extracted gum of the species(Ansari et al. 1993). As well, its wood is used for fuel and homeconstruction, and the fruit feeds wild life. Jordano (1989) hasidentified 26 bird species feeding on the Pistacia species. Thespecies grows in most parts of Iran, and have economic value asa means of livelihood in some parts of Iran.

A limiting factor of this species is seed blankness, whichreduces its efficiency. Many studies have been done on the prob-lem, resulting in several reasons being suggested as the majorcauses of seed blankness. Breakdown of tissue in the growingcotyledon of the pistachio (P. vera) seed has been observed. Lackof embryo sac at anthesis, embryo sac degeneration, and degen-eration of the embryo and endosperm have been considered themajor causes of fruit failure (Grundwag 1975). Deficiency orimbalance of nutrient elements also may be responsible. Thus itmay be possible to control the blanking and semi-blanking prob-lem with soil and water management (Dehghani-Shuroki andSedgley 1996a, 1996b). Even environmental factors such as tem-perature may be a possible cause of seed blankness. Dehghani-Shuroki and Sedgley (1994) hand-pollinated several female treesof P. vera with pollen from three different male plants, includingpollen from P. atlantica , in a staggered manner after anthesis, toinvestigate the pollen tube growth and fruit production of thespecies. They concluded that all of the variables of male parents,time of pollination, and time of harvest were significantly differ-ent. They also concluded that interspecific fertilization (P. atlanticax P. vera) would cause reduction in pollen tube growth due tointerspecific incompatibility and may result in ovule or seedabortion and thus in blanking. Whether these problems are con-stant in individual plants and how much variation exists within aPistacia population are not clarified in the literature. The presentauthors conducted a field study in a natural habitat of thespecies to investigate seed blankness variation in female plantsand its possible correlation with overall and average distance ofadjacent male plants, and some other morphological characters.

Materials and methodsSeed samples were collected from 18 female plants at full matu-rity stage, replicated three times on the same trees, in a naturalhabitat of P. atlantica in an area of 4 ha in Kerman province ofIran. Each sample contained four racemes randomly collectedfrom different parts of the female trees. One-way analysis ofvariance—where treatment is the only criterion for classificationof the data (Steel and Torrie 1987)—was used to test the varia-tion between the trees. Simultaneously, number of male plantslocated adjacent to each of the female plants and their overall andaverage distances to the female plants were recorded in a 50-mdiameter area around the female trees. Several morphological

characters such as plant height, trunk diameter at breast height(TDBH), and crown diameter were also recorded on all of theindividual female plants. Blankness of seed samples was com-pared using one-way analysis of variance of the data. The mainquestion was whether these differences are correlated to themorphological characters of female plants or by different maleplant density. LSD multiple range test was used to classify thefemale trees to indicate significant differences in seed blanknessbetween the female plants. Pearson phenotypic correlations be-tween all combinations of characters were estimated.

Results and discussionAnalysis of variance showed a significant difference betweenfemale trees in seed blankness at 1% probability (Table 1). Blank-ness percentage means and the results of LSD method of meancomparison are presented in Table 2. Blankness varied betweenstudied female plants from 7.8 to 97.5%. Percentage of seedblankness, which was the main characteristic of interest, did notshow significant correlation with any other studied characters,but showed a weak correlation with the number of male plantslocated adjacent to each of the female plants (? =10%). No corre-lation was found between the measured morphological charac-teristics and percentage of seed blankness. Correlation betweenplant height and crown diameter, TDBH and crown diameterwere highly significant (? =1%). Correlation between plant heightand TDBH was significant at 5% probability (Table 3). Thecorrelation values showed that the differences cannot be ex-

Table 1. Analysis of variance on seed blanknesspercentage in Pistacia atlantica Desf.

SOV DF SS MS F

Total 53 33937.62Trees 17 33391.60 1964.21 129**Error 36 546.01 15.17

Table 2. Seed blankness percentage mean and theresults of mean comparison

Tree number Blankness (%)

1 37.73 fg†

2 42.09 ef3 29.43 gh4 35.29 fg5 7.78 j6 80.08 b7 51.44 d8 51.02 de9 66.27 c10 19.08 I11 33.01 fg12 52.66 d13 85.73 b14 33.99 fg15 47.37 de16 97.50 a17 89.08 ab18 22.32 hi

†Means with the same letters are not significantly different.


plained by number of male plants located adjacent to each of thefemale plants or the average intervals of male and female plants.This does not mean that male plant density has no effect on seedformation. It means that the differences in the population understudy may be caused by genetic factors. Therefore, in the breed-ing programmes, particularly in seed orchard establishment, thischaracteristic has to be regarded as a main selection criterion. Itneeds to be mentioned that P. atlantica is a dioecious species andthe genetic potential of male parents in fertilization may bedifferent. Therefore, suitable male parents with maximum flow-ering overlapping to females have to be used in seed orchardestablishment as well as selected female plants. Obviously amain criterion for female plants in seed orchard establishmentwould be less seed blankness as well as other suitable morpho-logical and phenological characteristics.

ReferencesAnsari, S.H., M. Ali and J.S. Qadry. 1993. Essential oils of Pistacia

integerrima galls and their effects on the central nervous sys-tem. Int. J. Pharmacognosy 31:89-95.

Dehghani-Shuroki, Y. and M. Sedgley. 1994. Effect of pistil ageand pollen parent on pollen tube growth and fruit productionof pistachio. J. Hort. Sci. 69:1019-1027.

Dehghani-Shuroki, Y. and M. Sedgley. 1996a. Fruit developmentof Pistacia vera (Anacardiaceae) in relation to embryo abortionand abnormalities at maturity. Aust. J. Bot. 44:35-45.

Dehghani-Shuroki, Y. and M. Sedgley. 1996b. Shell structure andembryo development of Pistacia vera L. and Pistacia atlanticaDesf. (Anacardiaceae) following intra- and inter-specific pol-lination. Int. J. Plant Sci. 57:586-594.

El-Moslimany, A.P. 1986. Ecology and late-Quaternary historyof the Kurdo-Zagrosian oak forest near Lake Zeribar, westernIran. Vegetatio 68:55-63

Grundwag, M. 1975. Seed set in some Pistacia L. (Anacardiaceae)species after inter- and intraspecific pollination. Israel J. Bot.24:205-211.

Jordano, P. 1989. Pre-dispersal biology of Pistacia lentiscus(Anacardiaceae): cumulative effects on seed removal by birds.Oikos 55:375-386.

Steel, R.G.D. and J.H. Torrie. 1987. Principles and Procedures ofStatistics. A Biometrical Approach. McGraw Hill, London.

Table 3. Phenotypic correlation coefficient between studied characteristics in Pistacia atlantica Desf.

Traits Plant height (m) Trunk diameter (m) Crown diameter (m) Blankness (%) Male interval

Trunk diameter 0.56**Crown diameter 0.89*** 0.65***Blankness % –0.24ns –0.05ns 0.36nsMale interval 0.26ns –0.10ns 0.22ns –0.16nsMale number 0.26ns –0.02ns 0.10ns 0.46* –0.14ns

***, **, * significant at P=0.01, 0.005 and 0.10, respectively; ns = not significant. Male interval and male number are the aver-age intervals of male and female plants, and the number of male plants located in a 50–m diameter area around the femaletrees, respectively.

Plant Genetic ResourcesNewsletter

Aims and scopeThe Plant Genetic Resources Newsletter publish-es papers in English, French or Spanish, dealingwith the genetic resources of useful plants, result-ing from new work, historical study, review andcriticism in genetic diversity, ethnobotanical andecogeographical surveying, herbarium studies, col-lecting, characterization and evaluation, documen-tation, conservation, and genebank practice.

ManagementThe Plant Genetic Resources Newsletter is pub-lished under the joint auspices of the Internation-al Plant Genetic Resources Institute (IPGRI) andthe Plant Production and Protection Division ofthe Food and Agriculture Organization of theUnited Nations (FAO).

AvailabilityThe Plant Genetic Resources Newsletter ap-pears as one volume per year, made up of fourissues, published in March, June, September andDecember. Plant Genetic Resources Newsletteris available free of charge to interested librariesof genebanks, university and government depart-ments, research institutions, etc. The periodicalmay also be made available to individuals whocan show that they have a need for a personalcopy of the publication.

Types of paperArticlesAn article will publish the results of new andoriginal work that makes a significant contribu-tion to the knowledge of the subject area that thearticle deals with. Articles, which should be of areasonable length, will be considered by the Edi-torial Committee for scope and suitability, thenassessed by an expert referee for scientific con-tent and validity.

Short communicationsA short communication will report results, in anabbreviated form, of work of interest to the plantgenetic resources community. Short communi-cations in particular will contain accounts of ger-mplasm acquisition missions. The papers will beassessed by an expert referee for scientific con-tent and validity.

Other papersThe Plant Genetic Resources Newsletter willpublish other forms of reports such as discussionpapers, critical reviews, and papers discussingcurrent issues within plant genetic resources.Book reviews will be printed, as well as a Newsand Notes section. Suggestions for books toreview are invited, as are contributions to Newsand Notes.

SubmissionIn the first instance papers may be submitted intypescript form or as an Email message. Thefinal version may be submitted as an Email file oras a Windows-readable file on diskette. Manu-scripts submitted for publication and other com-munications on editorial matters should be ad-dressed to IPGRI's Editorial and PublicationsUnit.

Bulletin des ressourcesphytogénétiques

Domaine d’intérêtLe Bulletin des ressources phytogénétiques pub-lie des articles en anglais, en espagnol et enfrançais, sur les ressources génétiques de plan-tes utiles, fruit de nouvelles recherches, d’étudeshistoriques, d’examens et de critiques concer-nant la diversité génétique, d’études ethnobota-niques et écogéographiques, d’études d’herbiers,d’activités de collecte, de caractérisation etd’évaluation, de documentation, de conservationet les pratiques des banques de gènes.

ParrainageLe Bulletin des ressources phytogénétiques estpublié sous les auspices de l’Institut internationaldes ressources phytogénétiques (IPGRI) et de laDivision de la production végétale et de la protec-tion des plantes de l’Organisation des NationsUnies pour l’alimentation et l’agriculture (FAO)

DistributionLe Bulletin des ressources phytogénétiques paraîtune fois par an en un volume regroupant quatrenuméros publiés en mars, juin, septembre etdécembre. Il est distribué gratuitement aux bib-liothèques des banques de gènes, universités,services gouvernementaux, instituts de recher-che, etc. s’intéressant aux ressources phytogéné-tiques. Il est aussi envoyé sur demande à tousceux pouvant démontrer qu’ils ont besoin d’unexemplaire personnel de cette publication.

Types de documents publiésArticlesUn article contient les résultats de travaux nou-veaux et originaux qui apportent une contributionimportante à la connaissance du sujet dont traitel’article. Les articles, qui doivent être d’unelongueur raisonnable, sont d’abord examinés parle Comité de rédaction qui en évalue la portée etla validité, puis par un expert qui en examine lecontenu et l’intérêt scientifiques.

Brèves communicationsOn entend par brève communication un textecontenant, sous une forme abrégée, les résultatsde travaux présentant un intêrêt pour tous ceuxqui s’occupent de ressources phytogénétiques.Elle contient en particulier des comptes rendusdes missions d’acquisition de matériel génétique.

Autres documentsLe Bulletin des ressources phytogénétiques pub-lie d’autres types de rapport tels que des docu-ments de synthèse, des études critiques et desarticles commentant des problèmes actuels con-cernant les ressources phytogénétiques. Le Bul-letin publie une revue de livres ainsi qu’une sec-tion intitulée Nouvelles et Notes. Les auteurssont invités à envoyer leurs suggestions pour leslivres à passer en revue ainsi que des contribu-tions aux Nouvelles et Notes.

PrésentationEn premier lieu, les documents doivent être sou-mis dactylographiés ou par courrier électronique.La version définitive doit être présentée en fichierde courrier électronique ou sur disquettes com-patibles Windows. Prière d’adresser les manuscritsprésentés pour être publiés et d’autres communi-cations sur des questions de rédaction au Bureaude rédaction de l'IPGRI.

Boletín de RecursosFitogenéticos

Objetivos y temasEl Noticiario de Recursos Fitogenéticos publicadocumentos en inglés, francés y español quetratan de los recursos genéticos de plantas útiles,fruto de nuevos trabajos, estudios históricos,revisiones y análisis críticos relacionados con ladiversidad genética, investigaciones etnobotáni-cas y ecogeográficas, estudios de herbarios,actividades de colección, caracterización y eval-uación, documentación, conservación, y prácti-cas en bancos de germoplasma.

DirecciónEl Noticiario de Recursos Fitogenéticos se publi-ca bajo los auspicios conjuntos del Instituto In-ternacional de Recursos Fitogenéticos y la Di-rección de Producción y Protección Vegetal de laOrganización de las Naciones Unidas para laAgricultura y la Alimentación.

DistribuciónEl Noticiario de Recursos Fitogenéticos aparececomo un volumen anual compuesto por cuatronúmeros, que se publican en marzo, junio, septi-embre y diciembre. Se distribuye gratuitamente alas bibliotecas de bancos de germoplasma, facul-tades universitarias y servicios gubernamentales,centros de investigación, etc. que se interesanen los recursos fitogenéticos. También puedenobtener este noticiario las personas que demues-tren necesitar una copia personal.

Tipos de documentosArtículosLos artículos divulgarán los resultados de traba-jos nuevos y originales que contribuyan de modoimportante al conocimiento del tema tratado.Dichos artículos, que deberán tener una longitudrazonable, serán examinados por el Comité deRedacción en cuanto a su pertinencia e idoneidady posteriormente un experto juzgará su contenidoy validez científicos.

Comunicaciones brevesLas comunicaciones breves informarán de modoconciso sobre los resultados de trabajos de in-terés para las personas que se ocupan de losrecursos fitogenéticos. Las comunicacionesbreves incluirán, en particular, resúmenes sobrelas misiones de adquisición de germoplasma.

Otros documentosEl Noticiario de Recursos Fitogenéticos publi-cará otros tipos de informes, como documentosde trabajo, análisis críticos, y documentos queexaminen cuestiones de actualidad relacionadascon los recursos fitogenéticos. El Noticiario pub-licará una reseña de libros así como una secciónde Noticias y Notas. Las propuestas de librospara reseñar y las contribuciones a la sección deNoticias y Notas serán bien acogidas.

PresentaciónLos documentos deben entregarse, incialmente,en forma de texto mecanografiado o a través delcorreo electrónico. La versión final debe presen-tarse como un archivo de correo electrónico o endisquete compatible con el sistema operativoWindows. Los manuscritos para publicar y otrascomunicaciones sobre asuntos relativos a la re-dacción deberán dirigirse a la Oficina de Redac-ción del IPGRI.

No. 127, September 2001

Contents


Articles

Benefits from giving and receiving genetic resources: the case of wheatK. Cassaday (Mexico), M. Smale (USA), C. Fowler (Norway) and P. W. Heisey(USA)...........................................1

Variabilidad genética en el género Xanthosoma en CubaM. Milián, I. Sánchez, M. García, D. Guerra y A. Corrales (Cuba).................................................................11

Computer tools for spatial analysis of plant genetic resources data: 1. DIVA-GISR.J. Hijmans (Peru), L. Guarino (Colombia), M. Cruz (Peru) and E. Rojas (Colombia)......................................15

Variability and its characterization in Indian collections of blackgram [Vigna mungo (L.) Hepper]S. Gupta, S.R. Gupta, H.K. Dikshit and R.A. Singh (India)............................................................................20

The appearance of chlorophyll defects in cereals during regeneration of genebank accessionsW. Schliephake, M. Grau and A. Börner (Germany).....................................................................................25

Système de reproduction et variabilité morpho-phénologique chez Allium roseum L.R. Jendoubi, M. Neffati, B. Henchi et A. Yobi (Tunisia).................................................................................29

Assemblage of sesame germplasm for conservation and genetic improvement in NigeriaO.A. Falusi and E.A. Salako (West Africa)..................................................................................................35

Unexploited diversity in coconut palm (Cocos nucifera L.)V. Arunachalam, B. Augustine Jerard, M. Elangovan, M.J. Ratnambal, R. Dhanapal, Saran Kumar Rizaland V. Damodaran (India)..........................................................................................................................39

Short communications

Genetic variation in populations of cereal crops in southern coastal zone of YemenA.A. Bawazir (Yemen)...............................................................................................................................44

Variability in seed blankness in Pistacia atlantica Desf. in a natural habitatH. Mirzaie-Nodoushan and H.M. Arefi (Iran).................................................................................................46

Bulletin de Ressources Phytogénétiques...and anti-evolutionist (whose work Darwin nevertheless...

Documents

Transcript of Bulletin de Ressources Phytogénétiques...and anti-evolutionist (whose work Darwin nevertheless...