Plant Breeding 106, 319-;328 (1991)@ 1991 Paul Parey Scientific Publishers, Berlin and HamburgISSN 0179-9541

School of Biological Sciences, University of Birmingham (U.K.

The Gene Pools of the Grasspea (Lathyrus sativus L.)

A. G. YUNUS and M. T. JACKSON

With 2 figures and 4 tables

Received August 16, 1990/ Accepted October 29, 1990

AbstractInterspecific hybridization between the grasspea,Lathyrus sativus, and 15 wild species in SectionLathyrus is reported. Only two species, L. am-phicarpos and L. cicera produced viable FI hybridswith low fertility when crossed with L. sativus asmale parent. Crosses with six other species producedpods following pollination when L. sativus was themale, but seedlings were inviable, seeds did notgerminate or pods were empty or had totally shriv-elled seeds. When L. sativus was the female parent,only one cross with L. gorgoni produced an FI, butthe seedling was inviable. The germplasm resourcesof the grasspea are identified, with L. amphicarposand L. cicera placed in the secondary gene pool andthe other species in the tertiary gene pool. Thedefinition of these germplasm resources is discussedin terms of grasspea improvement through plantbreeding.

Key words: Lathyrus sativus -wild species Sect.

Lathyrus -interspecific hybridization -gene pools-germplasm resources -grasspea improvement

reliable grain crops and may be the only foodavailable in some areas when famines occur.This can result in excessive consumption andmay provoke the neurological form of lathyr-ism (GANAPATHY and DWIVEDI1961). There haseven been a ban on its cultivation by the Indiangovernment (RurrER and PERCY 1984) in anattempt to reduce the incidence of lathyrism.The neurotoxic principle has been identified as.8-N-oxalyl-L-a, .8-diaminopropionic acid orODAP (BELL 1964, MURTI et al. 1964). It hasbeen suggested, however, that there is goodscope for selection and development of varie-ties of low toxicity (KAUL et al. 1986).L. sativus is undoubtedly a grain legume withconsiderable potential for improvement. Oneof the steps in this process is the identificationand characterization of the germplasm re-sources of the crop, and a study of its evolu-tion.

The genus Lathyrus is large with 160 speciesand 45 subspecies (ALLKIN et al.1986), dividedin 13 sections (KUPICHA 1983). L. sativus isplaced in Section Lathyrus along with 33 otherspecies, including the recently describedL. belinensis (MAnED and GOYDER 1988). Inearlier classifications, L. sativus had beenplaced in Section Cicercula, but KUPICHA(1983) combined this section with SectionLathyrus, since the species in these two taxaare very similar in general morphology. Thespecies in Section Lathyrus include both annu-al and perennial forms. The species are all OldWorld in distribution, principally throughoutEurope, North Africa and the Near East, butalso reaching as far east as India.

Lathyrus sativus L., known as the grasspea,chickling pea or kheshari dhal, is an ancientOld World cultigen, and perhaps one of thefirst crops to be domesticated in Europe (KIS-LEV 1989), but its origin is not known (BALL1968). SMARTT (1984 a) noted that even thoughL. sativus has been used as a pulse for a longtime, its utilization as a forage crop has re-sulted in little evolutionary progress as a graincrop. Its wide dispersion is due to its utiliza-tion as a forage legume. Its use as a pulse ismainly confined to India where it occupies anarea of more than 1 million hectares (LAL et al.1986). In that country it is one of the most

u.s. Copyright Clearance Center Code Statement: 0179-9541/91/0604-0319$02.50/0

been reported in the genus Lathyrus (MARS-

DEN-JONES 1919, SENN 1938, SAW LWIN 1956,DAViES 1958, CRUICKSHANK 1984, YAMAMOTO etat. 1986, KHA WAjA 1988) during the past seven-ty years. Although a number of hybrids havebeen obtained between species in SectionLathyrus and other sections, only one has beenreported between L. cicera and L. sativus asthe female plant (SAW LWIN 1956).

In this paper we report the first comprehen-sive study of interspecific hybridization be-tween L. sativus and wild species in SectionLathyrus, as well as intraspecific crosses withinL. sativus, in an attempt to define the genepools of this underexploited pulse.

Although there are many Lathyrus species,it does not necessarily mean that availablegermplasm resources outside the biologicalspecies will be extensive (SMAR1T 1986). Thegene pool concept was originally proposed byHARLAN and DE WET (1971) to provide a betterclassification of crop plants and their wild rela-tives. Based on the determination of thebiological species, this concept has found wideapplication in the study of the germplasm re-sources of many crops.

In L. sativus it has been suggested that thereappears to be a considerable primary genepool, rather poorly differentiated in terms ofmorphological characters between what havebeen reported as a wild form and cultigen inIraq (TOWNSEND and GUEST 1974). To datethere have been no experimental taxonomicstudies, and there is little or no information toestablish the existence or extent of secondaryor tertiary gene pools associated withL. sativus. Only a few crossability studies have

Materials and MethodsMaterials: Fifteen wild species from Section Lathy-rus, each represented by a single accession, and 11accessions of L. sativus from a wide geographicaldistribution, were used in the hybridization pro-

Table 1. Accessions of Lathyrus sativus and 15 wild Lathyrus species used in the hybridization programme

Species Birmingham No. Donor Donor No.

57818567067158967230821623610714567385

674463

404429430434435468507558580588

GaterslebenSiena Bot. Gard.Southampton Univ.Southampton Univ.

LAT 139/82261867137867236

868065

LAT 101/78LAT 142/7833618867025182868013VIR-1519

BG-144135048VIR-1321VIR-1336VIR-1247VIR-1516VIR-1519SEL L-4LAT 458/79LAT 454/79

L. amphicaryosL. annuusL. basalticusL. cassiusL. chloranthusL. chrysanthusL. aceTaL. gorgoniL. hierosolymitanusL. hirsutusL. latiJoliusL. marmoratusL. odoratusL. pseudo-ciceraL. tingitanus

L. sativus

Southampton Univ.Bmo Bot. Gard.GaterslebenGaterslebenBordeaux Bot. Gard.Zurich Bot. Gard.Southampton Univ.Blanes-Gerona Bot. Gard.Southampton Univ.Vavilov Inst., Leningrad

Madrid -INIAIzmir Gene BankVavilov Inst., LeningradVavilov Inst., LeningradVavilov Inst., LeningradVavilov Inst., LeningradVavilov Inst., LeningradNBPGR, IndiaGaterslebenGatersleben

~ne Pools of the Gr~sspea (Lathyr~s sativus L. 321

order to have comparative estimates between acces-sions. Total pollen and percentage stainability weredetermined using a haemocytometer.

Pollen tube growth: Studies of in vivo pollen tubegrowth under ultraviolet fluorescence microscopywere carried out using the method described byMARTIN (1959), but with the following modifica-tions. Excised pistils were fixed in Camoy's solutionfor 24 hours, cleared in 2M sodium hydroxide for 2hours at 60 °C, and stained in 0.4 0;0 aniline blue in0.1 M potassium phosphate buffer. The stained pis-tils were mounted in glycerine and observed using aLeitz Orthoplan microscope under ultraviolet lightwith a HBO 200 high pressure mercury lamp at awavelength of 400 nm.

Meiosis: Chromsome pairing at metaphase I wasobserved in putative interspecific hybrids and theintraspecific hybrids in L. sativus. Buds about 2 mmlong were collected, and 3-5 anthers macerated inacetic orcein (1 % orcein in 45 % acetic acid) with-out fixation. Each preparation was heated for fiveseconds, and left to stain for a further three minutes,and then squashed.

ResultsPollen viability

Pollen stainability ranged from as low as 53 %in L. sativus (accession 507) to more than 98 %in L. cassius (Table 2). Most were in excess of80 %. However, when the total pollen countwas taken into consideration, the actual viablepollen could be estimated. L. latifolius had thelowest pollen stainability, but because the totalpollen count was high, viable pollen was esti-mated at 42.2 X 104. The highest viable pollencount, at 60.3 X 104 was recorded inL. odoratus, even though pollen stainabilitywas as low as 74 %. The lowest viable pollencount was recorded in L. marm.oratus, at 3.6 X104.

Interspecific hybridizationSuccessful crosses were indicated by the for-mation of pods and seeds. However, in onlytwo combinations, namely L. amphicarpos xL. sativus and L. aceTa x L. sativus, were ver-ified hybrids obtained (Table 3). Generally thecrosses were more successful when the wildspecies was the female. Ten of the interspecificcombinations were successful in these cases, interms of pod formation, but only six were

gramme (Table 1). Six of the L. sativus accessionswere used for interspecific hybridization, while 10were studied for intraspecific crossability. Fourteenof the wild species were annuals, and L. latifoliuswas the only perennial species. In both types ofcrosses, L. sativus was represented by all threeflower types described by JACKSON and YUNUS(1984), namely blue, white and a mix of blue andwhite. The species identify of all plants was verifiedby reference to published descriptions. The seeds ofthe wild species were scarified prior to sowing, butthis was not necessary for L. sativus. At least fiveplants per accession were used.

Pollinations: On the basis of time to flowering,L. sativus and the wild species could be classifiedinto three groups, and this information was impor-tant in planning the hybridization programme. Ninespecies, L. amphicaryos, L. basalticus, L. chrysan-thus, L. cicera, L. gorgoni, L. hierosolymitanus,L. marmoratus, L. pseudo-cicera and L. sativus,flowered in less than 8 weeks. Flowering occurredbetween 8 and 11 weeks in L. annuus, L. cassius,L. odoratus and L. tingitanus, and in the three re-maining species, L. chloranthus, L. hirsutus andL. latifolius, flowering time was greater than 11weeks and as long as 16 weeks.

Pollinations were carried out in a glasshouse at theUniversity of Birmingham during the summermonths of 1987 and 1988. Buds approaching anthesiswere emasculated and then immediately pollinated,as described by CRUICKSHANK (1984). The sepalcovering the keel was folded back, the keel excisedand the anthers removed. A stigma covered withpollen was removed from the male parent andrubbed on to the stigma of the female parent. Afterpollination the remaining flower parts were keptintact and covered with parafilm to avoid dehydra-tion, as well as possible contamination by foreignpollen. Developing pods could be seen in successfulpollinations after one week, when the parafilm wasremoved. For each plant an unopened flower wascovered with a paper bag to obtain selfed seeds. Eachpollination took between 10 and 15 minutes.

A total of 1,055 pollinations was made, compris-ing 419 interspecific pollinations, 74 intraspecificpollinations of the wild species, 430 intraspecificpollinations between L. sativus accessions, and 132pollinations of all types were made for observationof pollen tube growth. Reciprocal pollinations weremade as possible according to availability of flowersonly between L. sativus and the wild species. Con-trol intraspecific crosses were made with all wildspecies except L. latifolius.

Pollen counts and stainability: Fresh pollen wastaken from at least two plants per accession immedi-ately prior to anthesis to determine the pollen countand stainability. Pollen was mixed with 0.5 ml ofacetocarmine (0.5 % carmine in 45 % acetic acid) in

Table 2. Mean pollen viability in Lathyrus sativusand 15 wild Lathyrus species, and two interspecific

hybrids

% pollen Total viablestainability pollen (x 10~)

Species

82.1

95.5

92.0

98.3

70A

93.9

75.4

87.2

91.7

84.8

57.2

83.7

74.5

89.2

89.5

89.3

69.3

82.4

64.7

53.2

64.2

4.617.18.1

22.617.829.44.6

12.317.67.8

42.43.6

60.316.532.3

24.915.820.614.315.115.6

L. amphicarposL. annuusL. basalticusL. cassiusL. chloranthusL. chrysanthusL. ciceraL. gorgoniL. hierosolymitanusL. hirsutusL. latifoliusL. marmoratusL. odoratusL. pseudo-ciceraL. tingitanus

L. sativus

phicarpos and L. cicera produced viable hy-brids, but with L. hierosolymitanus and L. hir-sutus, the pollinations failed completely whenL. sativus was the male parent. The oppositeresult was obtained in reciprocal crosses ofL. annuus, L. marmoratus, L. pseudo-ciceraand L. tingitanus with L. sativus. These failedcompletely when L. sativus was the female,but pods with shrivelled seeds or empty podswere formed with L. sativus as the male par-ent.

In crosses between L. gorgoni and L basal-ticus with L. sativus as the female the hybridorigin of the progenies must be regarded assuspect. In the reciprocal crosses, even thoughseed set was extremely low in both instances, amorphological comparison of the progeny in-dicated that they may have resulted from self-ing, and in these cases.

Plants from the crosses L amphicarpos XL sativus and L cicera X L sativus were ver-ified as hybrids with low fertility (Table 2) onthe basis of morphological characters. The dif-ferences between the hybrids and female par-ents were more distinct in the case of L am-phicarpos than L cicera, because the formerwas morphologically very different fromL sativus. Hybrid flowers were intermediatein size in both combinations, and the flowercolour was also different from the female par-ent. In both combinations the hybrid flowersranged from light red to slightly bluish-red.Both L. amphicarpos and L. cicera have brickred flowers. Intermediate characters were ob-served with respect to the size and shape ofpods in the hybrids, as well as the length of theleaflets, petioles and tendrils. In both combi-nations the Fl plants grew vigorously andflowered profusely. Most flowers aborted at anearly stage of development, and those thatremained only produced empty pods.

Interspecific hybrids

L. amphicaryos

X L. sativus 433X L. sativus 434X L. sativus 435X L. sativus 468

1.95.21.77.5

0.100.400.150.80

L. ciceTa

x L. sativus 435X L. sativus 507x L. sativus 588

2.31.00.3

0.300.070.03

Pollen tube growth

Observations of pollen tube growth were madeon 82 pollinations representing 26 interspecificcombinations involving 14 species. In thosepollinations which failed completely (Table 3),alien pollen failed to germinate on the stigma.In a compatible combination, such as L. ciceraX L. sativus, abundant pollen tubes could beseen in the style and ovary, and within 24hours after pollination, the pollen tubes had

successful when L. sativus was the female.Seeds were obtained from the crossesL. sativus X L. gorgoni and L. latifolius XL. sativus, but following germination, the Flhybrids were inviable. In the cross L. chloran-thus X L. sativus, the F1 seed failed to germi-nate.

In crosses with five species, L. sativus asfemale produced only empty pods withshrivelled seeds. The reciprocals of thesecrosses gave very different results. Both L. am-

This can be explained by the fact that pollentubes were abundant and comparable to thecompatible cross, L. cicera X L. sativus.

Intraspecific hybridizationWild species: Although the number of pollina-tions of each type was rather small (from 1 to12), all but four species showed normal fertili-ty in sib-matings. The proportion of pods withseeds was lower in L. gorgoni (9/10),L. hierosolymitanus (3/5), L. odoratus (2/5) andL. chloranthus (115). Only one pollination wasmade for L. annuus, but this was successful.L. sativus: The level of success of crosses be-tween different L. sativus accessions was muchhigher than achieved in the interspecific cros-ses, but some genetic barriers do apparentlyexist between them (Fig. 2). Some pollinationsfailed completely, and the level of compatibili-

Meiosis

Chromosome pairing was observed in bothinterspecific and intraspecific hybrids. Regularpairing of 7 bivalents was seen in all intra-specific hybrids of L. sativus (Figs. 1 a and1 b). In the hybrid L. cicera X L. sativus, 6bivalents and 2 univalents or 7 bivalents wereobserved (Figs. 1 c and 1 d). In the hybridL. amphicaryos X L. sativus, meiotic irre-gularities such as multivalents and univalentswere seen (Figs. 1 e and 1 f).

Fig. 1. Metaphase chromo-some pairing in F1 intra-specific and interspecific hy-brids in Lathyrus SectionLathyrus: a and b, L. sativus(434) x L. sativus (429); c andd, L. aceTa (308) x L. sativus(507); e, L. amphicarpos (578)X L. sativus (433); and f,L. amphicarpos (578) XL. sativus (468)

~e_ne Pools of the q~asspea (Lathyrus sativus L. 325

,-~-,-

---

"-t--l,'. ",

."I ...'% I " f~I --"I""':";;r~ -~"#: " ..I

I ""~'-',l.. ..-~ -f,:- ~- ~ 35

'-

Percentage of seeded pods

0% (Pods without seed)

1-35%

-36-70%

-> 70 %

Fig. 2. Crossing polygon of ten accessions of Lathy-rus sativus, based on the percentage of pods formedwith seeds following intraspecific hybridization

ty varied considerably between accessions. Theresults of intraspecific hybridization gave noevidence of geographical origin as a factor inthe success or failure of hybridization withinL. sativus. Only a limited sample of intra-specific hybrids was grown. All plants grewvigorously, flowered profusely and producedmature pods with seeds, as expected on thebasis of the normal meiosis observed. Fiftysamples from the intraspecific pollinations rep-resenting most of the combinations betweenthe 10 accessions were observed for pollen tubegrowth. In almost all cases, pollen tubes couldbe seen in the ovary within 24 hours of pollina-tion. There was no evidence of inhibition ofpollen tube growth in the pistil.

relatives. It is clear, as SMAR1T (1990) haspointed out, that the Harlan and de Wet sys-tem is of equal value in consideration of geneticresources, for purposes of their classification,evaluation and documentation. In contrast,MARSHALL (1990) has argued that the value ofthis system, based as it is on crossability andease of gene transfer, was its direct relevance toplant breeding, but that developing moleculargenetic technologies may make this systemobsolete. The prospect in the future is that thepotential gene pool of any species will includeall life forms. Whilst this prediction may wellapply to the major food crops, such as wheat,maize, rice and potatoes to name but a few, itis highly unlikely that an underexploited cropsuch as the grasspea will attract such highbiotechnological priority. In the meantime,therefore, the identification of the gene poolsof this cultigen remains important, since anadequate knowledge of interspecific hybridiza-tion capability of L. sativus is an essential pre-requisite to effective evaluation of the broaderrange of genetic resources which might beexploited in the improvement of this crop

(SMAR1T 1990).On the basis of morphological variation, 10

species were considered as being more closelyrelated to L. sativus (YUNUS 1990). Seven ofthem were included in the crossability study,but seeds were not available for L. blepharicar-pus, L. hirticarpus and L. stenophyllus. Onlytwo of the 15 wild species in Section Lathyrus,namely L. amphicarpos and L. cicera, have areasonably close biological relationship withL. sativus. Morphological data from DAVIS(1970), JACKSON and YUNUS (1984) and YUNUS(1990) have indicated the close relationship ofthese two species to L. sativus. However, thedegree of interspecific crossability and the fer-tility of the FI hybrids is not sufficiently high,according to the HARLAN and DE WET (1971)gene pool concept, to consider that they andL. s~tivus belong to the same biological

specIes.Cytological studies carried out by YUNUS

(1990) also showed that there is a correlationbetween karyotypes and the relationship be-tween species in Section Lathyrus. In particu-lar, the karyotype of L. sativus was similar toL. amphicarpos and L. cicera, with which FIhybrids were formed, as well as L. basalticus,L. gorgoni and L. marmoratus but which

DiscussionHARLAN and DE WET (1971) considered that thegene pool concept could usefully be applied toease the difficulties in producing satisfactorytaxonomies for cultivated plants and their wild

Plant Breeding, Vol. 106 (4)

The Gene Pools of the G~~spea (Lathyrus sativus ~ 327

Kreuzung mit den beiden Arten L. amphicar-pas und L. cicera wurden lebensfahige Fj-Ba-starde mit geringer Fertilitat erhalten, wenn alsmannlicher Partner L. sativus verwendet wur-de. N ach Kreuzungen yon 6 weiteren Artenmit L. sativus als Bestauber bildeten sich zwarHiilsen aus, aber entweder waren die Hiilsenleer oder batten nur vollig eingeschrumpfteKorner. Wenn Samen ausgebildet wurden, wa-ren diese nicht keimfahig oder die Keimpflan-zen nicht lebensfahig. Wurde L. sativus alsmiitterlicher Partner benutzt, konnten nurnach der Kreuzung mit L. gorgoni Fj-Samenerhalten werden; aber auch bier gingen dieKeimlinge spater zugrunde. Die Einordnungder genetischen Abstammung der Platterbse imRahmen fines Genpool-Systems erfolgt in derWeise, daB L. amphicarpos und L. cicera clem2. Genpool und aIle anderen Arten clem 3.Genpool zugewiesen werden. Diese so aufge-stellten Abstammungs- und Verwandtschafts-verhaltnisse werden im Hinblick auf die Mog-lichkeit finer ziichterischen Verbesserung derPlatterbse diskutiert.

of L. sativus reported by VAVILOV (1951) andKAUL et al. (1986) was found only once in ourseed materials and was not used in the cross-ability programme.

Is L. cicera or L. amphicaryos the wild prog-enitor of L. sativus? On the basis of the datapresented here, this hypothesis is unlikely.ALLKIN et al. (1985) reported the wild origin ofL. sativus as unknown. The only report on thewild form of L. sativus was by TOWNSEND andGUEST (1974) but there has been no clarifica-tion as to whether these plants were truly wildor escapes from cultivation. The studies byJACKSON and YUNUS (1984) gave no indicationof conspecific wild forms, and until the reportsof wild L. sativus from Iraq are confirmed, theexistence of a wild form must be regarded asdoubtful. Simple tests on accessions of hard-seededness and observations on seed coat col-our and seed size could narrow the field, plusstudies of relative reductions of pod dehiscenceat or near maturity.

The germplasm resources profile ofL. sativus shows a tertiary gene pool that ismost extensive, but with a restricted secondarygene pool, with which transfer of genes will bedifficult. Without the identification of genepool1B (wild conspecific races), it is clear thatexploitation of germplasm resources for theimprovement of the grasspea must be concen-trated in gene pool 1A including the landracematerials. There is good evidence from iso-zyme polymorphisms (YUNUS et al. 1990) thatthe landraces are genetically heterogeneous andeven heterozygous at some loci. If this level ofdetected heterozygosity is a reflection of widerheterozygosity in the species, this is an inter-esting finding, since the indications in ourwork were that L. sativus is predominantlyautogamous. Consequently, exploitation ofthis variable gene pool1A through hybridiza-tion between different lines should result inconsiderable heterosis and prospects for selec-tion of favourable characters which might leadto a wider utilization of this grain legume.

The senior author acknowledges receipt of sponsor-ship from the Public Services Department, Malaysia,and the University of Agriculture, Malaysia forstudy leave.

ZusammenfassungDie Genpools cler Platterbse (Lathyrus sati-vus L.)

Berichtet wircl fiber Artkreuzungen zwischencler Platterbse, Lathyrus sativus L., und 15Wildarten der Sektion Lathyrus. Nur nach der

ReferencesALLKIN, R., D. J. GOYDER, F. A. BISBY, and R. J.

WHITE, 1986: Names and synonyms of species andsubspecies in the Vicieae. Issue 3. Vicieae DatabaseProject, Univ. of Southampton, U.K.

--, T. D. MAcFARLANE, R. J. WHITE, F. A.BISBY, and M. E. ADEY, 1985: The geographicaldistribution of Lathyrus. Issue 1. Vicieae DatabaseProject, Univ. of Southampton, U.K.

BALL, P. W., 1968: Lathyrus L. In: T. G. TUTiN, V.H. HEYWOOD, N. A. BURGERS, D. M. MOORE, D.H. VALENTINE, S. M. WALTERS, and D. A. WEBB(eds.), Flora Europaea 2, 136-143.

BELL, E. A., 1964: Relevance of biochemical tax-onomy to the problem of lathyrism. Nature 203,378-380.

--, and L. FOWDEN, 1964: Studies on amino aciddistribution and their possible value in plant clas-sification. In: C. A. LEONE (ed.), TaxonomicBiochemistry and Serology. The Ronald PressCompany, New York.

CRUICKSHANK, M. DE LA P., 1984: Crossability rela-tionships among some species of Lathyrus L. M.Sc. Dissertation, Univ. of Birmingham, U.K.

23*

328

MURn, V. V. S., T. R. SESHADRI, and T. A. VEN-KlTASUBRAMANIAN, 1964: Neurotoxic compoundsof the seeds of Lathyrus sativus. Phytochemistry 3,877-891.

RurrER, J., and S. PERCY, 1984: The pulse thatmaims. New Scientist 23, 22-23.

SAW LWIN, 1956: Studies in the genus Lathyrus.M.Sc. Thesis, Univ. of Manchester, U.K.

SENN, H. A., 1938: Experimental data for the revi-sion of genus Lathyrus. Amer. J. Bot. 25, 67-78.

SIMOLA, L. K., 1966: Chemical and morphologicalcharacteristics of some endemic species of Lathy-rus. Suomen Kemistilehti B 39, 148-150.

--, 1967: The effect of some non-protein aminoacids on pollen germination and pollen tube growthin five species of the Vicieae. Planta 77, 287-297.

SMARTT, J., 1979: Interspecific hybridization in thegrain legumes -a review. Econ. Bot. 33,329-337.

--, 1980: Evolution and evolutionary problemsin food legumes. Econ. Bot. 34,219-235.

--, 1984a: Evolution of grain legumes. I.Mediterranean pulses. Expl. Agric. 20, 275-296.

--, 1984b: Gene pools in grain legumes. Econ.Bot. 38, 24-35.

--, 1986: Evolution of grain legumes. VI. Thefuture -the exploitation of evolutionary knowl-edge. Expl. Agric. 22, 39-58.

--, 1990: Grain Legumes. Cambridge UniversityPress, Cambridge.

TOWNSEND, C. C., and E. GUEST, 1974: Flora ofIraq, Vol. 3, Leguminales. Ministry of Agricultureand Agrarian Reform, Baghdad.

VAVILOV, N. I., 1951: Phytogeographic basis ofplant breeding. In: F. VERDOORN (ed.), The Ori-gin, Variation, Immunity and Breeding of Culti-vated Plants. Chronica Bot. 13, 13-54. TheRonald Press Company, New York.

YAMAMOTO, K., T. FUJIWARA, and I. D. BLUMEN-REICH, 1986: Karyotypes and morphological char-acteristics of some species in the genus Lathyrus L.Jap. J. Breed. 34,273-284.

YUNUS, A. G., 1990: Biosystematics of LathyrusSection Lathyrus with special reference to the grasspea, L. sativus L. Ph.D. Thesis, Univ. of Birming-ham, U.K.

--, M. T. JACKSON, and J. P. CATTY, 1990:Phenotypic polymorphism of six enzymes in thegrasspea (Lathyrus sativus L.). Submitted for publi-cation.

Authors' addresses: A. G. YUNUS, Farm Office,University of Agriculture, 43400 Serdang, Selangor(Malaysia); M. T. JACKSON, International Rice Re-search Institute, P.O. Box 933,1099 Manila (Philip-pines).

DAVIES, A. J. S., 1957: Successful crossing in thegenus Lathyrus through stylar amputation. Nature180,612.

--, 1958: A cytogenetic study in the genusLathyrus. Ph.D. thesis, Univ. of Manchester, U.K.

DAVIS, P. H., 1970: Lathyrus L. In: P. H. DAVIS(ed.), Flora of Turkey and the East Aegean Islands3, 328-369.

GANAPATHY, K. Y., and M. P. DWIVEDI, 1961:Studies on clinical epidemiology of lathyrism.ICMR. Rewa.

HARLAN, J. R., andJ. M. J. DE WET, 1971: Towarda rational classification of cultivated plants. Taxon20,509-517.

HITCHCOCK, C. L., 1952: A revision of the NorthAmerican species of Lathyrus. University ofWashington Publications in Biology 5.

JACKSON, M. T., and A. G. YUNUS, 1984: Variationin the grass pea (Lathyrus sativus L.) and wildspecies. Euphytica 33, 549-559.

KAUL, A. K., M. Q. ISLAM, and A. HAMID, 1986:Screening of Lathyrus germplasm of Bangladesh forBOAA content and some agronomic characters. In:A. K. KAUL and D. CQOMBES (eds.), Lathyrus andLathyrism. Third World Medical Research Found-ation, New York.

KHAWAjA, H. I. T., 1988: A new interspecificLathyrus hybrid to introduce the yellow flowercolour into sweet pea. Euphytica 37, 69-75.

KISLEV, M. E., 1989: Origins of the cultivation ofLathyrus sativus and L. aceTa (Fabaceae). Econ.Bot. 43, 262-270.

KUPICHA, F. K., 1983: The intrageneric structure ofLathyrus. Notes Roy. Bot. Gard. Edinb. 41,287-326.LAL, M. S., I. AGRAWAL, and M. W. CHITALE,1986: Genetic improvement of chickling vetch(Lathyrus sativus L.) in Madhya Pradesh, India. In:A. K. KAUL and D. COOMBES (eds.), Lathyrus andLathyrism. Third World Medical Research Found-ation, New York.

MARSDEN-JONES, E. M., 1919: Report to the Scien-tific Committee. J. Roy. Hort. Soc. 45, xcii.

MARSHALL, D. R., 1990: Crop genetic resources:Current and emerging issues. In: A. H. D. BROWN,M. T. CLEGG, A. L. KAHLER, and B. S. WEIR(eds.), Plant Population Genetics, Breeding, andGenetic Resources. Sinauer Associates Inc., Pub-lishers, Sunderland, Massachusetts.

MARTIN, F. W., 1959: Staining and observing pollentubes in the style by means of fluorescence. StainTechnol. 34, 125-128.

MAXTED, N., and D. J. GOYDER, 1988: A newspecies of Lathyrus (Leguminosae -Papilionoideae) from Turkey. Kew Bull. 43,711-714.