Il Characterisation and evaluation of okra .-------------------- S. HAMON AND D. H. VAN SLOTEN Introduction Okra (Abelmoschus esculentus) is an important vegetable crop throughout the tropics and subtropics. Its genetic resources, including related cultivated and wild Abelmoschus species, were the subject of joint IBPGR/ORSTOM projects,which resulted in a status report largely based on available literature (Charrier, 1983, 1984). For details on taxonorny, geography, cytology, inter- and intra-specific crossability, etc, reference should also he made to this report. The species nomenclature used in this paper is summarised in Table Il.l. The world collection of oua The composition of the joint ORSTOM/IBPGR okra collection • based in Côte d'Ivoire, is shown in Table 11.2. There are 2,283 accessions. Clearly, the African continent (with 2,029 accessions) and West Africa in particular (with 1,769) is far more heavily represented than other continents and countries. Wild and cultivated species, other than A. esculentus, from Asia are absent in the collection and require collecting in the next few years. Wor1dwide, cultivated okra is largely species A. esculentus, but A. manihot and A. moschatus may he grown as weil (see Table lU). A major discovery was an undescrihed cultivated species, collected mainly in Côte d'Ivoire (Siemonsma, 1982a, 1982b). Hamon & Yapo (1986) detail further the distribution of this latter species which we refer to here as •West African taxon' (WAT), after its present known area of distribution. Figs. 11.1 and 2 map the sampling sites for A. esculentus and WAT. These figures, covering West Africa and part of Central Africa show each species separa tel y to emphasise the differences in cultivation areas throughout the four major climatic zones. From north to south these are:
Transcript
IlCharacterisation and evaluation ofokra
.--------------------S. HAMON AND D. H. VAN SLOTEN
IntroductionOkra (Abelmoschus esculentus) is an important vegetable crop
throughout the tropics and subtropics. Its genetic resources, includingrelated cultivated and wild Abelmoschus species, were the subject of jointIBPGR/ORSTOM projects,which resulted in a status report largely basedon available literature (Charrier, 1983, 1984). For details on taxonorny,geography, cytology, inter- and intra-specific crossability, etc, referenceshould also he made to this report. The species nomenclature used in thispaper is summarised in Table Il.l.
The world collection of ouaThe composition of the joint ORSTOM/IBPGR okra collection
•based in Côte d'Ivoire, is shown in Table 11.2. There are 2,283 accessions.Clearly, the African continent (with 2,029 accessions) and West Africa inparticular (with 1,769) is far more heavily represented than othercontinents and countries. Wild and cultivated species, other thanA. esculentus, from Asia are absent in the collection and require collectingin the next few years.
Wor1dwide, cultivated okra is largely species A. esculentus, butA. manihot and A. moschatus may he grown as weil (see Table lU). Amajor discovery was an undescrihed cultivated species, collected mainly inCôte d'Ivoire (Siemonsma, 1982a, 1982b). Hamon & Yapo (1986) detailfurther the distribution of this latter species which we refer to here as•West African taxon' (WAT), after its present known area ofdistribution.
Figs. 11.1 and 2 map the sampling sites for A. esculentus and WAT.These figures, covering West Africa and part of Central Africa show eachspecies separately to emphasise the differences in cultivation areasthroughout the four major climatic zones. From north to south these are:
174 S. Hamon & D. H. van S/oten 175
]E::ll::o
~oê
oCu
--
Information 011 tlte collectiollTable Il.2 summarises the coverage of passport data in the
ORSTOM/IBPGR okra collection. It is comparatively well documented(cf. Peeters & Williams, 1984). Elevation data are missing because they arenot very important in West Africa. Local names are listed frequently butnot systematically translated and therefore often unusable. Missingpassport data largely relates to material obtained from other genebanksbefore 1981.
Samples acquired from multi-erop coUecting missions fail to list thenumber of fruits, fruit characteristics, and comments on local traditionsindicating crop associations. This produces a major data gap. Furtherinformation can be obtained from local names when they are systematicaUy coUected and translated. Fig. 11.3 has been prepared on the basis ofa translation of local names from the Togo/Benin collecting mission(Hamon & Charrier, 1983). The relative frequencies are shown indecreasing order of importance from top to bottom. Asterisks representthe most commonly used characteristic for a given category. Interestfocuses primarily on the harvest period. Contrast between carly varieties(A. escu/entus) and lale varieties (WAT) is the most usual distinction. Thedate of planling and the length of plant cycle are secondary.
It is not unusual to identify a variety by colour or shape of the fruit,comparing the fruit with a part of sorne familiar animal (e.g. antelopehorn, agouti check, rat tail) or of a human being.
Names describing plant characteristics (height, lcaf type, etc) are lessfrequent, and tend ta he used by ethnie groups who already know a great
desert (village or oasis cullivation), Sahelian (north of latitude 12oN),savannah (betwecn latitudes 8 oN and 12 oN) and rain forest climaticzones.
A. escu/entus (Fig. 11.1) is primarily distributed throughout theintermediate savannah zone betwecn the rain forest and the arid Sahel,The species is less frequently found in the rain forest zone but is, on theother hand fairly weil represented in the Sahel zone. With one exception,the WAT (Fig. Il.2) does not occur in the Sahelian zone since it has a longlife-cycle and usually requires abundant, continuous rainfall. The easternboundary of its distribution is difficult to determine due to lack of samplesfrom Central Africa. At present, the most distant sampling sites are inCameroon. Since a natural interspecific hybrid of the two cultivatedspecies occurs in the central part of Sudan, WAT is possibly more widelydistributed than currently known.
174 S. Hamon & D. H. van S/olen 175
]E::ll::o
~o~.cU
--
Information Oil tlte cOllect;OllTable 11.2 summarises the coverage of passport data in the
ORSTOM/IBPGR okra collection. It is comparatively well documented(cf. Peeters & WiIliams, 1984). Elevation data are missing because they arenot very important in West Africa. Local names are listed frequently butnot systematically translated and therefore often unusable. Missingpassport data largely relates to material obtained from other genebanksbefore 1981.
Samples acquired from multi-erop collecting missions fail to list thenumber of fruits, fruit characteristics, and comments on local traditionsindicating crop associations. This produces a major data gap. Furtherinformation can be obtained from local names when they are systematically collected and translated. Fig. I 1.3 has been prepared on the basis ofa translation of local names from the Togo/8enin collecting mission(Hamon & Charrier, 1983). The relative frequencies are shown indecreasing order of importance from top to bottom. Asterisks representthe most commonly used characteristic for a given category. Interestfocuses primarily on the harvest period. Contrast between early varieties(A. escu/emus) and late varieties (WAT) is the most usual distinction. Thedate of planting and the length of plant cycle are secondary.
It is not unusual to identify a variety by colour or shape of the fruit,comparing the fruit with a part of some familiar animal (e.g. antelopehorn, agouti cheek, rat tail) or of a human being.
Names describing plant characteristics (height, leaf type, etc) are lessfrequent, and tend to be used by ethnic groups who already know a great
desert (village or oasis cultivation), Sahelian (north of latitude 12oN),savannah (between latitudes 8 oN and 12 oN) and rain forest climaticzones.
A. escu/enlus (Fig. 11.1) is primarily distributed throughout theintermediate savannah zone between the rain forest and the arid SaheLThe species is less frequently found in the rain forest zone but is, on theother hand fairly well represented in the Sahel zone. With one exception,the WAT (Fig. 11.2) does not occur in the Sahelian zone since it has a longlife-cycle and usually requires abundant, continuous rainfall. The easternboundary of its distribution is difficult to determine due to lack of samplesfrom Central Africa. At present, the most distant sampling sites are inCameroon. Since a natural interspecific hybrid of the two cultivatedspecies occurs in the central part of Sudan, WAT is possibly more widelydistributed than currently known.
Table 11.2. Currmt starus ofokra 'world collection' at ORSTOM, Centre d'Adiopodoume, Côte d'Ivoire!
Asia- China (Taiwan) 4 4- India 61 61- Philippines 6 6Totals 1375 673 18 8 19 2093
10
190
1 The majority of the accessions in this collection have been obtained from IBPGR and/or ORSTOM germplasm collecting missions(Iargely multi-crop) carried out during the period 1980-5.t No information on country of origin availabJe for A. mtlnihol (3 accessions) and A. moschatus (1 accession); the collection aIso iDCludes :::i10 standard international cultivars -..1
Table 11.2. Current status ofokra •world collection' at ORSTOM, Centre d'Adiopodoume, Cote d'[voire1
Middle-East- Afghanistan 8 8-Iran 16 16- Pakistan 7 7- Saudi Arabia I 1- Syria 4 4
Asia- China (Taiwan) 4 4- India 61 61- Philippines 6 6Totals 1375 673 18 8 19 2093
10
190
1 The majority of the accessions in this collection have been obtained from IBPGR and/or ORSTOM germplasm collecting missions(largely multi-crop) carried out during the period 1980-5.t No information on COUDtry of origin available for A. mtlnihol (3 accessions) and A. moschatus (1 accession); the collection also includes :::i10 standard international cultivars -..I
• --'-----'--.. _.. 0:.... -. _.
~ ._.~ ...._ ~_ _. .... _. liIII
Fig. 11.1 Geograplùcal distribution of Â. eSCIIUl'ltlU in West Arrie.
-.1QO
.-
- '.J
\..", '
,"
. "
.......-.. _-~ .. --" ,,,,,,
-----J, ... , .••
•
.. -....... --•
•
• •..... .O'.
",-- -.. '
:-'·-T111111,
.~
•1
•
'i -~J-• • "... : ..',,,,._.) ..
/>. ••, . ..-
_...---_ ... _----:1,,
li' h.'1111 ell...1 ~.!!!!!!!.!!! nllu'" Il ...1 ."Iu
..-': ~:_"i;"_H'
~-"-'T"
1 1",
F"1g. B.l. Geographie•• distribution of the West Arric.o Taxon (WAn
r 1. 1
1
.'
,,
'-----..--- \1,
l ,.-,,'-r-_.--. H
--. 1 .~.-- ..~ t-= .-dd
-.l\0
,V'
"
,,,
\".1 •,",
',.,
......... - -,FDUST
- ••• .. -J
S'UAltti~H:
O'.
.._.( - -'J
,'P',)
. ""t-•••• '"". ,
... . '".:,p
\~~ .. ;; ,
~ • ,:- -.. ---t1f""","-"'Y"'/•• , •••, *~,. . .. i ... IV -
.'
...-.... __ - 'f..........
'"''• t'_,,,,
"
l''\. ,. .. f -4=,. '-/=T -.-........l '" _.
<"'~,.. -" ". .. .. --- ..
• Sampling sites of WAT tollect~d in West A1rica
deal about the plant. Where a collecting mission finds such names itshould ask. very detailed questions.
Very occasionally, a local name may refer to sorne aspect of the plant'sprovenance, culinary properties or sorne associated crop. OnlyA. moschatus names refer to food taboos.
Charaderisatioo and evaJuationMaU,üds and tMthothThe full oha collection was multiplied initially at the ORSTOM
Station in Adiopodoumé (Iower Côte d'Ivoire). Twenty plants were planted
Fig. 11.3. Recopidon of the West African oUa 11lDdrac:es Inlsed on tlM!traml.tioo of the vemacullt Dames·
in rows, with the control varieties Clemson Spineless (internationalA. escuJentus cv.) and WAT (primitive cv. ORS 520) were planted everytwenty rows. Fungal pathogens, insect pests and nematodes werecontrolled chemical1y. However, leaf curl virus, transmitted by a whitefly,Bemissia tabacci, is very prevalent during the first six months, pcakingbetween February and June. It is impossible to obtain an A. escuJentusplant at that time which will top 50 cm by the end of its cycle. Ali partsof the plant are deformed, inflorescences poor and seeds malform .
Descriptors used in the characterisation and evaluation fall into threecategories: (i) quantitative, (ii) qualitative and (iii) enzymatic. In generalthe published IBPGRjORSTOM descriptor list (Charrier, 1983, 1984)was used, but for the qualitative descriptors more descriptor states wereused.
181Evaluation of okro colleetionsS. Hamon & D. H. van SJoten180
Collecting organisationCollectorCollector's numœrCountry of onginTown/ProvinceLatitudeLongitudeAltitudeVernacular names
Table 11.3. Proportion of missing passport data inoRSTOM / l BPGR okra collection
Quantitative descriptors; Plant morphology and development werecharacterized by plant height, number of internodes, stem diameter andbranching. Measurements were taken systematically between 80 and100 days after planting, coinciding with the end of the growing cycle of theClemson Spineless control cultivar. For very long-cycle plants, mainlyWAT accessions, a second series of measurements was made two or threemonths later.
The day of first fiower opening, the height at first flowering and firstfruiting, and the number of intemodes were also noted for each plant.Fruit-setting parameters (average total number of fruits per plant anddistribution along the main stem and branches) and fruit characteristics(length, width, number ofridges) and seeds (weight of one thousand seeds)were also noted.
••PLANT HEIGHT.lEU SHAPE.POD POSITION 01 THE sni...PRODUCTIOIi LEVEL.SOIlIlB PLACE
... KEW iNTRDDucnDiIDaitr Ir willaii Ir It~.lc ....)
Fig. 11.4. EIectropboretk dJscrimlDating patterns of the two main okrac:uld".ted specles'
Qualitative descr;ptors: There are three main types of qualitativedescriptors: colour. shape and other features. The specifie descriptorswere the colour of the main stem, the petai base, the leaf petiole, the veins,the lamina and the fruit (unripe); the shape of the leaves; and the positionof the fruit on the main stem. These descriptors tend to he highlysubjective.
Enzymat;c descriptors: Isoenzymatic electrophoresis can provide adescription virtually unaffected by the environment and fairly easy todetermine.
Electrophoresis has becn carried out with starch gel according to Second& Trouslot (1980). For okra, eleven systems could be used, in decreasingorder of resolution:
Fair to poor: Esterase, acid phosphatase, peroxiàase, catalase.Within the cuitivated species (A. esculentus and WAT)
enzymatic variability hardly exists. The electrophoreticpatterns of both species are distinct (Fig. lIA), and hencethe method can be used for species identification (Hamon& Yapo, 1986).
Problems encounuyed dilTiIIg c1uuacferisation and eva/uation(1) Species identificationSpecies identification prior to planting is essential to proper
organisation of the work, and passport data on this descriptor are oftenincorrect. Species identification c.'ln he made on fruits at the time ofcollection (whole fruits are the most common method of okra preservation).
Fig. Il.5 shows the main types of fruit found: fruit types 1 to 5 WAl';fruit types 6 to 10 A. escu/entus; fruit type Il the control cultivar ClemsonSpineless.
(if) Sud germinationThe problem of poor germination was the most disturbing factor
in establishing trials. Limited numbers of seeds, low rate of germination,pest attacks and other damage produced trial imbalance for a fairly largenumber of accessions.
i85Evaluation of okra colieClions
Direct identification by seed sampling is impossible for bath species, butelectrophoresis of single seeds can quickly distinguish between the twospecies. The zymograms in Fig. 11.4 are invariant within each of the twospecies and different between them. Sorne samples were found to bemixtures of the two species (Table 11.4). The species can also be identified
from the seedling, but this requires expertise and is not conclusive. Specîesidentification by morphological criteria is definitive only on the basis offlower characteristics (number and shape of epicalyx segments). This iseffective, but it does have the drawback that it oceurs late in the sequenceof the trial.
(iil) Heterogeneityükra, owing to its floral structure and the absence of self
incompatibility. produces much of iLs progeny through selfing. However.cross-pol1ination is frequently mentioned in the literature. The extent ofoutcrossing varies according to the variety, the cropping seasoo, and thelocation (Chandra and Bhatnagar, 1975; Martin, 1983; Tanda, 1985),ranging from 0 to 60 per cent. There is a close correlation between crosspollination and the presence or absence of insects. With strict pest control,contamination is severely restricted. But this is not the case in thetraditional agricultural setting where pesticide treatments are non-existent.The continuous fiowering and the special constraints of the eva!uationprocedure make both systematic bagging and isolation impossible. Thereis, therefore, a certain risk factor.
Table liA lists heterogeneity rates for species and provenances studiedin 1984--5. There are three major types of heterogeneity:
1. Interspecific heterogeneity. This is due to mixing of seeds oftwo distinct species during co!lecting;
2. Intraspecific heterogeneity refers to accessions which aremixed or segregating. This cao result from crossing or frommixing of fruits by the donor or collector. One possibilitymight be to identify a whole fruit as an accession, but, thiswould cause an enormous increase in the number ofaccessions. Table Il.4 shows toat 18 per cent of WAl'and 22
1 1 â 1 ~N
r-OOM~O
~ r-:oOci~~ci
è-'-~t:~
~~-~'è'-~
1;~
"Ë:!;1' .~....~(,Jr ..,l,';S
il: ~~
C)
-.i.--Il1 u
:0C'Of-
•
1~,
186 S. Hamon & D. H. van Sloten Evaluation of okra col/ections 187
ij ,
~I·! .;.1
,l'1111
•
per cent of A. esculentus accessions fall ioto this category;(3) Partial intraspccific heterogeneity refers to heterogeneity foronly one or two traits.
An accession which l is homogeneous is ,a rare o<x:urrence in a piantwhich is reproduced by sceds and is not strictly autogamous.
Resll1ts(1) Uni-variate anaiysisQuantitative descriptors. Table i 1.5 lists the statistics of quan
titative descriptors for the two species. There are marked inter-speciessimilarities and dissimilarities for particular descriptors.. Discriminantanalysis (See page 189). clearly brings out the differences between the twospecies.
Table Il.5. Statistical parameters observed on data recorded asquantitative descriptors
5.0 17.0 .10.2 2.3 22.3 WATFruit diametcr at 0.7 4.8 2.1 0.7 35.1 ESCmaturity 1.2 2.5 0.5 21.7 WAT
Min. = Minimum value Max. = Maximum value Mean = Mean value5D = Standard deviation CV ... Coefficient of variationESC = A. esculenrus WAT = West AfIican Tax.on
Qualitative descriptors: The main markeTS are:Species specifie traits:- Specifie to A. escuientusi
Bronze;(7) stem colour. fruit calour (14), sea green beingcharacteristic of Sudan accessions; darker calour of fruit ridges..
- Specifie to WAT !
Colour of fioral spot always internai; blackish green fruitcolour, fruits slightly or very pendulous; fruits may heprickly; seeds may have a reddish fuzz;
Difference of frequencies:Differences of frequencies basical1y concern three plantaspects. The following are mast common in A. esculentus:green stems, petioles, fruitsentire leaves, no clearly marked lobesshorter branches.
(ù) Bi-varime analysisCorrelations among qUlJ1ltitative variables. Correlations between
quantitative variables were calculated. The strongest and most persistentcorrelations are between early flowering and lirst flowering and fruilingnodes; next, hetween the first fruiting node and plant structure (stemdiameter, number of internodes and number of branches). An unexpectedrelationship involves the weight of 1,000 seeds which is lower for f 'lS setat a oode high on the stem.
In A. esculentus. plant height is c10scly correlated with flowering andfruiting parameters and structure. This is much Jess or not at ail true of theWAT; height is postively correlated with total seed and fruit productionat 80 days. This last descriptor is itself correlated with total fruitproduction. Il can also he seen that the close correlation between fruitingon the stems and on the branches of A. esc14lentus disappears in theWAT.
h is evident that the distribution of quantitative and qualitativevariables and their correlations vary between species. Sorne correlationsare oot constant between accessions, even within the same species.
Relations between early flowering and latitude of sampling site. Fig. 11.6depicts early flowering as a series of cumulative frequencies whilst alsoshowing the environmental zone of the sampling site for both species. Wehave grouped the data in accordance with c1imatic zones as schematisedin Figs. Il.1 and 2. Fig. 11.6 shows that precocity increases as one moves
from humid to more and zones. It aîso shows that A. esculentus ftowerseariier !hat WAT al the same latitude.
Fil. 11.6. FIoweriDa behavlour of West Afrkan okru
Under the evaluation conditions in lower Côte d'Ivoire, the WATaccessions from Cameroon exhibiled major ftowering problems; sorneaccessions had grown two metres in one year after planting withoutputting forth a single flower. This was also true to a lesser extent for sorneaccessions at less than 7° from the c:quatOL Generally speaking, earlyfiowering in both species means less development -- smaller size, shortenedcycle and, for the WAT, fewer branches.
189Evaluation of okru collections
(iil) Muiiivariate analysisComparison between the USDA and Côte d'Ivoire collections.
Factor analysis was carried out on the collection available in 1982. In Fig.Il.7 the seattcr diagram represents 45 per cent of the total variability. AJIvariables were involved in the analysis, but the only ones represented inthe figure are those which aetually contributed to the axes.
A clear contrast between the two cultivated spccies is apparent alonghorizontal axis l. Vertical axis 2 shows intraspecific variability, particularly in colouration, seed production, fruit width and branching. It showsclearly the contrast between the USDA A. esculenrus accessions and thoseFrom Côte d'Ivoire. The Côte d'Ivoire accessions are much morepolymorphie.
Fig. 11.3 shows the varietal recogmtlOn methods as deduced fromtranslating the local names of the samples colleeted. There is a strongsimilarity between the local names and the factorial variables. This showsthat carcful attention should he given to peasant systems of varietyclassification.
Geographie distribution of A. eseulentus variability. The Côte d'IvoireA. eseulentus collection is much more variable than the USDA collection.To undcrstand such differences, a prior comparative analysis of severalcountries is needed. Principal component analysis of the quantitativevariables was undertaken, and Fig. 11.8 shows the seatter diagramcontaining 50 per cent of the total variability. The countries included areBenin, Burkina Faso, Guinea, Mali, Sudan, Togo, Zambia and Zimbabwe.Only thelimits of variability encountered for each country are shown.
A big difference in balloon size is immediately apparent between Togoand Benin with peak variability, and Mali with the smallest balloon. Axisl contrasts two types of plants; one early, small and unbranched withheavy seeds, and the other much hardier. It is interesting to note that Axis2, fruit production, is independent of the oha type and that the mostproductive plants come from Sudan or Burkina Faso.
Comparison of A. eseulentus and WAT, The discriminant analysis usingquantitative variables, fulfils two major objectives. The first is a testclassification of individuals into pre-defined groups according to a selectedcriterion. The second is to establish a classification of variables indescending order of discrimination, in order to select a minimum numberwhich will suffice.
A comparison was made between the two cultivated species. First, ailquantitative variables were compared and, next, on!y the three mostdiscriminant variables (number ofintemodes, plant height and total plant
~ lA. ,seu/entus"utltl.
'NAT
hhl •+12" N. $. 11
IIraqu~ {! 08--12° N. S.Ul S.211
hl. ,.,..t il •HON. S.I81 s.1I1s,.,."1111 (1 ) Il' tU
fig. Il.7. Factor anaIysis 00 the Ivoriu and USDA okn coUectioos
if .. ..
~- 44 ( , 5 Q 2 e JO ., ,.( Q 2 PX _St lf
F'I. JI.&. Principal compoDeDt ana.Iysis 011 A. ellCllklrt1l.J introdudioos
:._....._._---_._--_._-...._-_._-\j • fi : ...ln SOI : h~.. 11 11f: hrU.. fin TU: T". i1 lill: hl... ZI': lI••I, 1~ .AL: '111 llll: ZI.h••, !1_._ _ . :
"'ntfi relFfFlt
Stail ~hzrhr
IH'
---.
Su,,n~.ttl..
fralf ,n~.cflllU lb If,.
!JeliU ,f1000 suil
v:>
Fla. n.9. Eoclldiu distances (variance criteria) between ok...descrlptora
Table 11.6. Discriminant analysis between the two cultivated okra speciesin West Africo
fruit production) were selectod. The pcrccntage of properly dassifiedvariables was determined for each. The results are round in Table
11.6.As cao he seen, the margin of error is usually small and the estimation
with the three variables produces quite comparable results. There is adilference of 0.4 per cent for A. escu/entus and 5.4 per cent for the WAT.
193EvalUa/ion of ok, ra collections
The slight differences can he explained by specifie adaptations or perhapsby introgression l'rom the other species. This is an extremely usefulmethod which can be applied to different kinds of groups such as: planttype, geographic origin, etc.
Regrouping individuals and/or variables. Hierarchical dustering methodsallow data to he c1assified by descriptors or by individuals. In Fig. 11.9 aclassification scheme is presented for 12 descriptors, on the basis of factoranalysis. Descriptors which are very strictly correlated, fruiting or colourof the petiole and stem, have not been taken into consideration.
This classification shows that the choice of descriptor depends on therequired level of precision in describing variability. At the Euclidiandistance level of 0.80 three distinct groups can be distinguished (Fig. Il.9).Already at this level one can see a grouping of descriptors translating thevariabi!ity encountered. A choice at random of one variable in each groupwould already provide a strong indication of available variability. Achoice at level 0.30, which appears acceptable in our situation, wouldallow the elimination of four descriptors.
Establishment of core collectionThe principal ideas behind the concept of the' core' collection are
described in Frankel & Brown (1984); sec: also chapter by Brown in thisvolume. The need to reduce collections does oot appear directly at thelevel of the base collection which is multiplied at least once and conservedin its entirety. However, considering the lack of data on many geneticresources collections, and the extremely costly and time-consumingcharacteri9ation and evaluation of large collections, there is an obviousneed for reduction (Peeters & Williams, 1984; van Sioten, 1987).
With regard to the okra collection. it was decided in 1983 to establishsuch a core collection (irom 200 to 300 accessions). This number was notselectcd in terms of pe:centage of the entire collection, but rather inaœordance with the foilowing objectives:
1. to have a manageable collection scaled down to the needs ofthe breeder and/or other user; and
2. to include the widest possible range of variability.Towards the end of 1985, an okra core collection of 189
accessions was established on the basis of representative variability asdescribed by passport, characterisation and evaluation data, but alsoincluding rare types. This core collection has already been distributed toseveral countries for further evaluation.
Jn this conncclion it should he noted that Ihere are relative!y few okra
breeding programmes in the world and therefore limited opportunities forgene introductions, Le. the transfer of specific genes in breedingprogrammes. There is, however, enormous scope for using the corecollection in adaptation trials in a large numher of different environments(direct plant introduction).
ConclusionsOn the basis of the experience gained in the evaluation of the
ORSTOMjIBPGR okra collection, we are attempting an overview ofthe problems, of the methods of characterisation and evaluation, andpossible solutions which may have application in other crop plants.
The quality of the information obtained at the collecting site is anextremely important factor. It goes without saying that information ongeographic co-ordinates is an absolute necessity. It is possible to improvethe level of information during collecting by observing the following:
1. restrict the mission to one or a very limite<! numher ofspecîes;
2. take sufficient lime to become familiar with the localconditions and customs;
3. request a systematic translation of local names;4. ensure the involvemcnt of women farmers.
The cultivation system use<! is an important source of information,which becomes more important when the crop has a long tradition in theparticular country. One can therefore not expect to obtain the samequality of information in ail areas, but one should know how to profitmost from information available. Wc have secn that the graphicalrepresentation of the variability by means offactoriiU analysis correspondsclosely to the farming systems used and the latter therefore provide acertain orientation in the choice of morphological descriptors.
The choice ofdescriptors is the second critical stage. Irozyme descriptorsare now being use<! more and more. Theil' use eliminates the enviwnmentalinfluences, the necessity for large areas for cultivation, and the method isfairly simple, The authors agree with Crawford (1985) who underlines thelimitations of the use of such markers and cOllsiders the electrophoreticinformation as complementary to standard characterisation. Morphological descriptors are important, since they are of most interest to theagronomist and breeder. Morphologkal polymorphism is not necessarilyassociated Ilor correlated with enzymatic polymorphism. Davis &Gilmartin (1985) emphasize that 5ubstantial morphological variationcould he associated with only minor enzymatic changes, At the same time,adaptation plays an important raIe in the differentiation of e.cotypes. We
195Evaluation of okra collections
ReferencesRorssum-Waalkes. J.van (1966). Malesian Malvaoc:ae revised. BI~a,14,1-251.Chandra, S. & Bhatnagar, S. P. (1975). Reproductive biology of A. ~sCllknlus.
Charrier, A. (1983). Les R.essOlUces Gél'liliqws du Genre Abelmoschus M~d.
(Gombo). International Board for Plant Genetic Resources, Rome.Charrier, A. (1984). GeMlic ksources of lM G~nus Abclmoschus M~d.
(Okra). International Board for Plant Genetic Resources, Rome (Englishtranslation of Charrier, 1983).
Chevalier, A. (1940). L'origine, la culture el les usages de cinq Hibiscus de lasection Abeimoschus. R.evw de &lanÙ/UE Appliquie el d'Agricullure Tropù:al,20,319-28.
Crawford, J. L. (1985). Electrophoretic data and plant speciation. Syslemot/cBOlany. 10,40>'15.
Davis. J. 1. &: Gilmartin. A. J. (1985). Morphological variation and speciation.Systtmalic Bolany, 10(4),416-25.
Frankel, O. H. &: Brown, A. H. D. (1984). CUITenl planl genetic ~uroes-A crilical llppraisal. In GeMlÙ:s: New Frontiers. Proceedings of the XVInternational Congress of Genel.ics. Volume IV, pp. 3-D. Chopra, V. L.,Joshi, B. C., Sharma, R. P. and Bansal, H. C. (ros.), Ollford & IBHP"<Ibl.ishing Co., India. Reprillted ln Crop Genelie ResCPoIpces: Cons~rvalùlll
and Evalualion, pp. 249-57, Holden, J. H. W. & Williams. J. T. (eds.), Allenand Unwin, L.ondon.
HlImon, S. & Cha,rrier, A. (1983). Large 1I8riatioll of okra collecte<! in Beninand Togo. PianI Genelie Re:lOuru.1 News/eller, !56. 52-8.
Hamon, S. &: Yapo. A. (19S6). Perturbation induced within the genusAbelmoschus by the discovery of a second lXlible okra s~jes in WeslAfrica. Acla Horlietdlurae, 182, 133--44.
Martin, F. W. (1983). Natural outeroSling of okl'1l in Puerto Rico. JourMlofAK,iC1JIIlU'~ of lM Universily of~"o Rico, 67, 5i}-2.
Pal, 8. P.. Sing.lJ , H. B. &. Swarup, V. (1952). Tnonomic relationships andbree<iing possibililies of species of AbelmosehUJ related 10 oha(A. l'scu!entus). Bol!Ulïc Gazell". lB, 455--M.
have observed the disappearance of the WAT from arid zones and adecrease in variability in areas where cultural practices have becomerestrictive,
The reduction in the numher of descriptors can he obtained throughmultivariate analysis. A step-wise approach consists ofrunning a principalcomponent factor analysis which projects the variability in a limiteddimension. Then a clustering analysis is done which will he later tested bya discriminant analysis. In this approach the choice made by the researcherleads to a deliherate, but controlled los8 of information,
An effective reduction of a collection, on the understanding that theoriginal collection still needs to he conserved, is a crucial problem. Inokra, a collection reduced to 20Q--400 well-described accessions is likelyto he of a size which can he properly used.
S. Hamon & D.R.van Sioten194
111
•
•
196 S. Hamon & D. H. van Slolen
Peeters, J. P. &. WilIill.lm, J. T. (1984). TowlIrds bettcr use of gencbanks withspecial refercnce to infonnation. P/anl Genette Resources News/eller, 64,22··32.
Second. G. & Trou!iot, P. (1980). EI-.:ctrophorè~ d'cnzymtll du riz (Oryza sp.).Traveaux et documenta de \'ORSTOM, ORSTOM, Paris.
Siemonsma, J. S. (19820). West African okra - morphologka\ andcytogcnetical indications for the existence of Il naturai amphidiploid ofAbelMoschlLf esculelllus (L.) Moench and A. manihot (L.) Medikus.Euphylica, 81, 241-52.
Seimonsrna, J. S. (\982b). La cul/l~re du Gombo (Abelmoschus spp.), Ugumefruit Iropical- avec rifirence spici.2! à la Côte d'Ivoire. Thesil AgnclllturalUniversity, Wageningen, The Netherlands.
Sloten, D. H. van. (\981). The role of curators, brecders and other usens ofaermplasm in characterization and evaluatioD of crop genetic rcsources.Vtb SABRAO Congren.
Tanda, A. S. (1985). Floral biology, pollen dispersal, and foraging behaviourof honeybees in okra. Journal of Agricultwal Research, 1.4, 225-7.
!1..
1
•
The use ofplant geneticresources
Ediled by
A.H. D. BROWNCS/RO Division of Plant Indwtry, Canberra, Al4.'!Jtralia
O.H. FRANKELCS/RO Division of Plant lndustry, Canberra, Australia
D.R. MARSHALLWaite Agricultural Research lnstitute, University of Adelaide, Australia
J. T. WILLIAMSInternational Board for Plant Genetie Resources, FAD, Rome, ltaly
CAMBRIDGE UNIVERSITY PRESS
Cambridge
New York New Rochelle
Melbourne Sydney
12
Part 11
3
88
17
32
68
49
173
105
123
136
197
157
Contributors vii Preface ix Acknowledgements xi
Role of genetfe resource collect:ions ln researdJ and breed.lngGennplasm collections and the public plant breederK. S. Gill
2 Gennplasm collections and the privatt plant breederJ. S. C. Smith & D. N. Duvick
3 Gennplasm collections and the experirncntal biologistR.G. Palmer
•
Contents
Use of collections
International use of a sorghum gennplasm collectionK. E. Prasada Rao, M. H. Mengesha &: V. G. Reddy
5 Current use of potato collectionsJ.G. Th. Hermsen
6 Use of collections in cereal improvement in semi-aridareasJ. P. Srivasrava & A. B. Daman;a
7 Limitations to the use of gennplasm collectionsD. R. Marshall
Siu and structure of collectioœThe case for large collectionsT. T. Chang
9 The case fOi çore collectionsA.H.D. Brown
10 The role of networks of dispersed collectÎonsl'.M. Perret
E~lI1luation
Charact.erisalion and evaluation of ohaS. Hamon & D. H. van S/otenEvaluation of cereals in Europer, FiH'l;hf'rk
Part Il4
Part IV11
Part BI8
Published by the Press Syndicale of the University of CambridgeThe Pill Building, Trumpinglon Street, Cambridge CB2 1RP32 East 57th Street, New York, NY 10024 USA10 Stamford Road, Oakleigb, Melbourne 3166, Australia
ISBN 0 521 34584 7ISBN 0 52l 36886 3 Pbk
UP
Printcd in Great Britain by Cambridge University Prest
~ International Board for Plant Genetic R.csources 1989
First publishcd 1989
British Ubrary cataJogu/Ir, in publkation dala
The use of pilOt aenetic nl8OllI'CeS
1. PIanu. Genes. Variation. CoNCrvationcl exploitation1. Brown, A. H. D.639.9'9
Uhrary of CongreJS callJÛ)glùllg in publication data
The use of plant geoetic rtSOlll'œS/editcd by A. H. D. Brown .. [et a!.].p. cm.
1. Germplum resouroca, Plant. 2. Gennplasm resources, PlantUtilization. i. Brown, A. H. D.SB123.3.U84 1988631S23---;icJ9 88·12292
1\
11
1vi Contents
Il 13 Evaluating the germplasm of groundnut (Arachis hypogaea)and wild Arachis spccies at ICRlSAT
liJ. P. Moss, V. RDmanat/w RJw &: R. W. Gibbons 212
14 Practical çonsiderationl relevant to effective eVllluationJ. T. Wii/lam.s 235
15 Principlcs and strategies of evaluation
I! O. H. Frankel 245Part V Wlld rel.tins or ero,.
16 Collection strategie.ç for the wild relatives of field crops
Il: C.G.D. Chapman 26317 Wild relatives as sources of discasc resistance
J.J. Burdon & A.. M. Jarosz 21101
18 Ecological and genetic considerations in çollecting and usingIl wild relativesG. LAdizinsky 297
Part VI TedtllOlo&kaI or 1deatl8c: iaDo••donI dJat affect the use or
l, leoetic rt!SOUI"CeS
19 ln vitro conservation and germplasm utilisationL.A. Withers 309
20 Screening for rcsistancc to diseasesP. H. Willlam.s 335
21 Restriction fragments as molecular markers for gennplasmevaluation and utilisationR. Bernatsky cl S. D. Tanksley 353
22 Molecular biology and genetic resourcesW. J. Peacoclc 363Index 377
iJ
Contributors
----------------------
R. Bernatsky, Department of Plant Breeding and Biometry, Z52 EmersonHall, Corne\! University, Ithaca, New York, 14853, USA.
A. H. D. Brown, CSIRû Division of Plant Industry, GPO Box 1600,Canberra, ACT, 2601 Australia.
J. 1. Burdon, CSIRO Division of Plant Industry, GPO Box 1600,Canberra, ACT, 2601 Australia.
T. T. Chang, International Riœ Research Institute, PO Box 933, Manila,Philippines.
C. G. D. Chapman, International Board of Plant Genetic Resources,FAOjLNOR, 101 22nd Street, Washington OC 20437, USA.
A. B. Damania, ICARDA, PO Box 5466, Aleppo, Syria.D. N. Duvick, Pioneer Hi-Bred International Ine., 700 Capital Square,
400 Locust Street, Des Moines, Iowa, 50309, USA.G. Fischbeck, Technische Universitât München. I..ehrstühl für PfIan
zenbau, 8050 Freising, Weihenstephan, West Gennany.O. H. Frankel, CSIRO Division of Plant Industry, GPO Box 1600,
India.V. G. Reddy, ICRISAT, Patancheru PÛ, Andhra Pradesh, 502 324,
India.J. S. C. Smith, Pioneèr Hi-Bre<! International Inc., 7301 NW 62nd
Avenue, Johnston, Iowa, 50131, USA.J. P. Srivastava, ICARDA, PO Box 5466, Aleppo, Syria.
S. D. Tanksley~ Department of Plant Breeding and Biometry, 252Emerson Hall, Cornell. University, Ithaca, New York, 14853,USA.
D. H. van Sioten, International Board for Plant Genetic Resources,IBPGR Headquarters, FAO, 00100 Rome, Italy.
J. T. Williams,.International Board for Plant Genetic Resources. IBPGRHëadquarters, FAO, 00100 Rome, Italy.
P. H. Williams, Department of Plant Pathology, University of Wisconsin,Madison, Wisconsin, 53706, USA.
L. A. Withers, Depàrtment of Agriculture and Horticulture, University ofNottingham, School of Agriculture, Sutton Bonington, Loughborough, LEl2 5RD, lJK. Currently: International Board forPlant Genetic Resources, IBPGR Headquarters, FAO, 00100Rome, ltaly.
Preface
---------------------
There can he little doubt that plant breeders have now a greater rangeof genetic diversity available to them than ever before. Moreover, it isavailable to breeders anywhere in the world, subject to sorne technicalconstraints. This is due to the co-operation of national and internationalinstitutions in an international network promoted and co-ordinated bythe International Board for Plant Genetic Resources (IBPGR). Each ofthe important crops or groups of crops, including all the major and manyminor ones, is represented in one or more facilities which act as 'basecollections' charged with responsibility for long-tenn conservation of thegenetic resources of one or more crops. Associated 'active collections'provide the link with the users of collections.
Needless to say, the usefulness ofcollections to plant breeders and otherusers, including evolutionists, plant pathologists, taxonomists and otherexperimental biologists, depends in the first instance on the extent towhich they arc geographicaIly and ecologically representative and on thepresence of genes of particular interest to plant breeders. Collections havebeen enriched by greatly increased collecting activities in recent years,man)' of which were stimulated or organised by IBPGR. Indeed,collections of many more crops are a great deal more comprehensive thancver before.
Then why are they not used by breeders to a greater extent than theyappear to he? Various rcasons have been suggested. Breeders tend to usebreeding materials with which they are familial' and whîch are reasonab!yadapted to their environment, as against alien materials requiring alengthy programme of pre-adaptation. Further, users require infonnationon collections to be presented in a manner that will allow them to identifyaccessions of potential use in their projects. This process involves thedescription, or 'characterisatioi1 " of the material, and its 'evaluation' for
x Preface
characteristics of particular concem to plant breeders. Both characterisation and evaluation have been defined by IBPGR and descriptorlists for many crops have been published. The work involved in theseoperations is considerable, in many instances exceeding the capacity ofnational collections which are close - hence most relevant to users.
This book explores the factors that are likely to limit or to facilitate theutilisation of plant germplasm. It grew out of a workshop convened bythe IBPGR Programme Committee al St Mathieu de Treviers, nearMontpellier, France, 9--12 September 1986.
The book has six parts: in the firsl part threc users representing publicand private plant breeders and experimental biologists deline the raie ofcollections and suggest ways to enhance their usefulness. The second partpresents duce case histories of collections and discusses limitations toeffective use and how they could he remedied. Large collections arecontraste<! in the third part with the recently proposed representative corecollections. A chapter on smaller collections in Europe shows how theassociation of national collections stimulates collaboration betweenbreeders and genebank managers. The fourth part describes the evaluationsystem in three widely differing collections. There follows an assessment ofthe state of management in germplasm collections, and a discllS.'iion of theprinciples of characterisation and evaluation and of the roles of curators.specialists in relevant fields, llnd plant breeders. With the higher prioritynow accorded to wiJd crop relatives, the fifth part examines howcollections are to be broadened and made more representative by theinclusion of crop-relatC".o species. TIle final section outlines recentlydeveloped techniques whlch are begiuning to open up new approaches inail areas of genetic resources work.
