Euphytica 55: 33-42, 1991.@ 1991 Kluwer Academic Publishers. Printed in the Netherlands.

Phenotypic polymorphism of six enzymes in the grasspea(Lathyrus sativus L.)

A.G.

Yunusl, M. T. Jackson2 & Janet P. CattySchool of Biological Sciences, The University of Birmingham, Edgbaston, Birmingham B152TT, UK,.Present addresses: 1 Farm Office, University of Agriculture, 43400 Serdang, Selangor, Malaysia,. 2 International

Rice Research Institute, P.O. Box 933,1099 Manila, Philippines

Received 19 March 1991; accepted in revised form 6 May 1991

Key words: Lathyrus sativus, grasspea, isozyme phenotype, phenotypic polymorphism, genetic diversity,genetic resources

Summary

The phenotype variation in six enzymes, 6-phosphogluconate dehydrogenase (6-PGD), malate dehydroge-nase (MDH), peroxidase (PRX), isocitrate dehydrogenase (IDH), galactose dehydrogenase (GD) andglutamate oxaloacetate transaminase (GOT), was investigated using horizontal starch gel electrophoresis in52 accessions of the grasspea, Lathyrus sativus. Phenotypic polymorphism was observed for all six enzymes.High phenotypic polymorphism (Pj) was observed for PRX and 6-PGD, while there was little polymorphismfor GOT, with only two accessions showing variation. There was no correlation between phenotypicpolymorphism and region of origin, or groupings of accessions made on the basis of flower colour. Tentativegenetic interpretations of banding patterns are given for five of the enzyme systems. The level of apparentheterozygosity was higher than expected in this predominantly autogamous species. The level of variation inthe grasspea is discussed in terms of its potential for exploitation through plant breeding.

Introduction

Allozyme markers are often used to investigatesystematic problems or to measure levels of varia-tion within and among populations (Hamrick &Godt, 1990). Furthermore an understanding of theway genetic variation is partitioned among pop-ulations is of primary importance for the conserva-tion of genetic diversity of plant species. Few stud-ies on the grasspea, Lathyrus sativus L., a minorpulse crop grown principally in the Indian sub-continent and also in North Africa and other partsof the Mediterranean basin, have been carried outto evaluate its variation patterns. On the basis ofmorphological characters, Jackson & Yunus (1984)showed that the grasspea is differentiated into sev-

eral distinct forms, primarily on the basis of flowercolour, seed size and size of leaves.

Genetic analysis of isozymes has been carriedout in genera of the tribe Vicieae. The garden pea(Pisum sativum) has become an important modelfor isozyme studies (Przybylska et al., 1982; Weed-en, 1985; Parzysz et al., 1986; Weeden & Marx,1987; Weeden, 1988). Other studies have been car-ried out with Lens (Zamir & Ladizinsky, 1984;Pinkas et al., 1985; Hoffman et al., 1986; Tadmor etal., 1987) and Vicia (Yamamoto & Plitmann, 1980;Mora et al., 1983; Suso & Moreno, 1986; Manciniet al., 1989). Only one study of isozyme diversity inLathyrus has been published by Yamamoto et al.

(1986).The gene pools of L. sativus (sensu Harlan & De

34

zyme phenotypic polymorphism is a useful initialstep for identifying genetic variation in plant spe-cies, even when the genetic control of bandingpatterns has not been determined. Since the genet-ic basis of many isozymes has been reported inmany plant species, including those in severalclosely-related genera of the tribe Vicieae, a tenta-tive genetic interpretation of isozyme patterns inthe grasspea can be made by extrapolation.

Materials and methods

Isozyme analysis was carried out on 52 accessionsof L. sativus (Table 1), the majority of which werestudied for morphological variation by Jackson &Yunus (1984). The accessions were obtained frombotanic gardens and research institutes, and repre-sent the geographic range of this cultigen, fromPortugal to Afghanistan and India, as well as in-

Wet, 1971) were described by Yunus & Jackson(1991). It is clear that the primary gene pool of L.sativus is extensive, including both primitive land-race forms as well as more advanced varieties, eventhough formal plant breeding appears to have hadlittle impact on the development of this cultigen.The wild progenitor of L. sativus has not yet beenidentified with any certainty. The secondary genepool is rather limited, comprising two species, L.amphicarpos and L. cicera. All other Lathyrus spe-cies can be placed in the tertiary gene pool. Futureimprovement of the grasspea will consequently bedirected at exploitation of the primary gene pool.Given the variation in morphological characters inthis cultigen, a study of isozyme variation may beconsidered as a useful adjunct to more convention-al studies of plant morphology.

In this paper we report the study of isozymediversity in 52 accessions of L. sativus representingits wide geographical range. The analysis of iso-

Table 1. The geographic origin and flower colour of the L. sativus accessions studied for isozyme diversity

Acc. no. Flower colour Provenance Acc. No. Flower colour Provenance

404406428429432433434435438439441442444448452454455459466467468472473474475476

MixedMixedMixedBlueWhiteMixedMixedWhiteBlueBlueBlueWhiteWhiteBlueBlueBlueBlueWhiteMixedWhiteBlueMixedMixedWhiteWhiteBlue

479480481484485488489491492494495498499502504506507556558565567580585586587588

MixedBlueWhiteWhiteWhiteMixedWhiteBlueMixedBlueWhiteWhiteBlueMixedWhiteBlueWhiteWhiteBlueBlueWhiteBlueBlueBlueMixedWhite

USSR (Georgian SSR)USSR (Tajik SSR)USSR (Ukrainian SSR)USSR (Krasnodar Territory)USSR (Kuibyshev region)

BulgariaCzechoslovakiaIndiaCzechoslovakiaUSSR (Azerbaijan SSR)USSR (Orel region)USSR (Kuibyshev region)AustraliaUSSR (Krasnodar territory)Australia

AfghanistanChileIndiaIndiaTunisiaCreteIranIranIsraelCreteTunisia

SpainSpainTurkeyTurkeyPortugalHungaryHungaryUSSR (Tambor region)IndiaIndiaIndiaUSSR (Voronezh region)USSR (Georgian SSR)USSR (Tajik SSR)USSR (Azerbaijan SSR)TurkeyAfghanistanUSSR (Mordovian SSR)USSR (Azerbaijan SSR)USSR (Orenburg region)FranceUSSR (Moldavian SSR)BulgariaUSSR (Tatar SSR)USSR (Tambor region)Turkey

35

in this slice. For Buffer System A, the second slicewas used for 6-PGD, and the third for MDH. Thesecond slice of the cathodal strip was used for PRX.For Buffer System B, IDH was stained in the sec-ond slice, GD in the third and GOT in the fourth.Stain recipes were obtained from Shaw & Prasad(1970) for PRX and modified for 6-PGD (0.5 MTris-HCI pH 7.15 ml, distilled water 45 ml, 6-phos-phogluconic acid Na3 10mg, NADP 5 mg, NBT10mg, PMS 2mg), from Tanksley & Orton (1983)for GD and GOT, and from Wendel & Weeden(1990) for MDH. For IDH, the following recipefrom Nickerson-Zwaanesse, B.V. (unpublished)was used: 0.IM-Tris-HClpH7.550ml, 1M MgClz0.5 ml, DL-isocitric acid 50 mg, NADP 8 mg, NBT10mg, PMS 2mg. Variation in banding patternswas determined by the migration from the origintowards the anode, or cathode for PRX, and thewidth of the bands.

An assessment of isozyme phenotypic polymor-phism was made using the overall banding pat-terns. Phenotypic polymorphism (Pj) and weightedphenotypic polymorphism (P) were calculated ac-cording to Kahler et al. (1980). A tentative geneticinterpretation of the banding patterns was made,based on the reported structure of each enzyme indifferent plant species (Weeden & Wendel, 1990),and particularly in related genera such as Pisum,Lens and Vicia, where the information was avail-able. L. sativus is a diploid species (2n = 14).

Results

All observed phenotypes for the six enzymes stud-ied are shown diagrammatically in Fig. 1, and thenumber of accessions and frequencies in the totalnumber of plants in which they were observed aregiven in Table 2.

6-phosphogluconate dehydrogenase (6-PGD, EC1.1.1.44). Sixteen phenotypes were observed forthis enzyme. Phenotype 1 with two bands was themost common, and observed in all accessions ex-cept 565 from Tunisia. In three accessions, 556 and558 from India and 580 from Iran, all 10 plants hadphenotype 1. Phenotypes 2, 3, 5 and 6 were observ-

cluding three accessions from Australia and Chile.They also encompassed the range of flower col-ours, namely white, blue, and blue and white, de-scribed by Jackson & Yunus (1984).

Ten seedlings per accession were studied. Leafsamples were taken from 3-4 week old seedlingsgrown in a glasshouse between March and August1989. Only the youngest leaves were used. Crudeextracts were prepared by macerating leaves in50 ILl of extraction buffer (O.lM Tris-HC1 pH7.2with 13mM (0.1% v/v) 2-mercaptoethanol). Theperspex extraction trays were kept on crushed iceduring maceration to prevent denaturation of theenzymes. Extracts were absorbed on to wicks madefrom Whatman No. 3MM chromatography paper.Horizontal electrophoresis was carried out in 12%starch gels. Two buffer systems were used: BufferSystem A 0.005 M L-histidine monohydrochloridepH 7.0 (gel buffer), and 0.2 M trisodium citrate pH7.0 (electrode buffer) (Nickerson-Zwaanesse,B.V., unpublished); Buffer System B 0.009M Trisand 0.003 M citric acid pH 7.0 (gel buffer), and0.135M Tris and 0.043M citric acid pH 7.0 (elec-trode buffer) (Shaw & Prasad, 1970). Thirty sam-ples were run on each gel, plus two wicks dyed withbromophenol blue to act as a marker control. Ref-erence extracts from the L. sativus accession 476were run on each gel for comparison.

Electrophoresis was carried out at 40 C with aconstant current of 30 mA until the tracker dyemoved about 1 cm from the origin, after which thewicks were removed and the current reduced to25 mA. Each electrophoretic separation took ap-proximately 6 hours for both the buffer systemsused.

Six enzymes were selected for detailed analysisafter a preliminary survey of 13 enzymes, since theygave consistent results with this species. The sixenzymes assayed were 6-phosphogluconate dehy-drogenase (6-PGD), malate dehydrogenase(MDH-NAD) and peroxidase (PRX) with BufferSystem A, and isocitrate dehydrogenase (IDH-NADP), glutamate oxaloacetate transaminase(GOT) and galactose dehydrogenase (GD) withBuffer System B.

Each gel was cut into four slices. The top slicewas discarded since most enzymes did not stain well

36

PRX

Phenotype

Fig. 1. Phenotypes of six isozymes from 52 accessions of the grasspea, Lathyrus sativus.

ed in more than 30% of the accessions, and pheno-type 15 in 13% of the accessions. The remaining 10phenotypes were less common in less than 6% ofthe accessions. Of these phenotypes, 7, 11, 13 and

14 were found in accessions only from the SovietUnion. Phenotypes 4 and 16 were seen in three andtwo accessions respectively from widely separated

geographic regions.

37

The isozymes of 6-PGD have been reported asdimers in castor bean (Simcox & Dennis, 1978) andin spinach (Schnarrenberger et al., 1973). The twodistinct zones of activity in L. sativus can be in-terpreted as two loci, A and B. At locus B a singleband representing one of two alleles was present,and one of these was seen in nearly 98% of allplants. At locus A the patterns can be interpretedas several alleles and both single and multiplebands were present. Almost 58% of plants werehomozygous at locus A. However, almost 40% ofplants showed three bands in this zone, and can beinterpreted as heterozygous for this locus. In fourphenotypes (13-16) only two bands were seen in

this zone. There are several explanations for thisphenomenon. Such a pattern may represent homo-zygosity with gene duplication (Gottlieb, 1982), orheterozygosity with a null allele (Hvid & Nielson,1977). Another possibility is co-migration of bands,but this and other explanations can be verified onlyby formal genetic analysis.

Malate dehydrogenase (NAD) (MDH, EC 1.1.1.37).There were only three phenotypes, and phenotype1 with five bands was represented in all accessions,and almost 98% of plants studied. Phenotype 2 wasfound only in eight accessions, and in no more thantwo plants per accession. Phenotype 3 was rare and

Table 2. The frequency of isozyme phenotypes for 6 enzymes in 52 accessions and individuals of L. sativus (n = 520)

Accs. % popn. Accs. % popn Accs. % popn. % popnAccs. Accs % popn Accs % popn

5221

99.4<1.0<1.0

52101

97.91.9

<1.0

4518

3

9

74.017.51.27.3

5211253121

80.07.111.3

<1.0<1.0<1.0<1.0

5129173

20231111213

172

57.711.06.21.18.58.0

<1.0<1.0<1.0<1.0<1.0<1.0<1.0<1.0

2.7<1.0

45331

2362

2230

4

2124

98

1284

127

1117118171212

30.2<1.0< 1.0

1.27.31.2

<1.08.113.8

<1.0<1.0<1.0<1.0<1.0

2.51.94.41.7

<1.03.51.53.76.22.12.1

<1.01.5

<1.0<1.0<1.0<1.0

23

456789

10111213141516171819202122232425262728293031

38

only observed in one plant from accession 504 fromAustralia. There were three zones of enzymic ac-tivity. The zone nearest to the origin generally haddiffuse bands, but in one plant only two clear bandswere observed (phenotype 3). The other two zoneswere close together but were usually separated asshown in phenotype 1, where the zone nearest tothe anode consisted of two bands, and the otherhad three bands. Occasionally, a single band wasdetected between these two zones (phenotype 2).

In MDH (NAD) both three and four dimeric locihave been reported in maize (Goodman et al.,1980). In L. sativus the banding patterns can beinterpreted as three polymorphic loci. Phenotype 1was the most common with two bands at locus Cand three bands at locus B. The latter locus certain-ly shows heterozygosity. Another explanation ofthe multiple bands could be gene duplication, sincethis phenomenon has been reported for MDH

(Gottlieb, 1982).

A maximum of four bands was observed and sevenphenotypes scored. Phenotype 1 was the most com-mon and seen in 80% of all plants; 19 accessions(37%) had only phenotype 1. Phenotypes 2 and 3were less common, represented in 11 and 25 acces-sions respectively, and in 7% and 11 % of all plants.The rare phenotype 7 was seen in accession 473from Bulgaria.

A single isozyme locus and dimeric structure forIDH were reported in Pisum sativum (Weeden,1985; Ni et al., 1987). Using these reports as mod-els, the patterns seen in L. sativus appear to repre-sent a single locus, but it is difficult to suggest whatthe diffusely stained region might represent, sincein Beta, for instance, it sometimes shows threebands or diffuse staining (Adel Sabir, personalcommunication). The multiple bands seen in six ofthe phenotypes indicate heterozygosity, yet thepresence of two and four bands in phenotypes 3 and6, and 2, respectively, are difficult to explain. Onlyone plant from a total of 520 studied (0.2%) hadphenotype 7 with a single band, which can be in-terpreted as a homozygous locus.

Galactose dehydrogenase (GD, EC1.l.l.48). Thestaining method for this enzyme exhibits achromat-ic bands. Only one zone of activity was observed,and the bands can be interpreted as four alleles atthe same homozygous locus. Phenotype 1 was themost common, in 74% of plants studied, and ob-served in all but five accessions (455, 481, 485, 499and 567). Almost 50% of the accessions had onlyphenotype 1. There were 18 accessions with pheno-type 2, but only three accessions (455, 506 and 585)with phenotype 3. Phenotype 4 occurred in nineaccessions, and in all 10 plants from the Australianaccession 499. This is one of the least frequentlystudied enzymes in plant species. Given the singleband patterns observed, this enzyme could beeither a monomer or a dimer in L. sativus.

Cathodal peroxidase (PRX, EC 1.11.1.7). Perox-idase exhibits both anodal and cathodal bands, butonly the cathodal region was analysed because ano-dal bands were diffuse and difficult to interpret. InL. sativus, PRX banding patterns were complexwith a total of seven bands, but the number ofbands in a phenotype ranged from one to five. Thecomplex banding patterns were represented by 31phenotypes. The maximum number of phenotypesin a single accession was eight, and the minimumwas two. Phenotypes 1, 5, 8, 9 and 23 were com-mon, observed in more than 30% of the accessions.Fifteen phenotypes were present in fewer than10% of the plants studied, and 14 were found infewer than 1 % of the plants.

Weeden (1985) and Weeden & Marx (1987) re-ported a similar complexity in the dimeric perox-idase system in P. sativum which has four isozymeloci. Due to the complexity of the banding patternsin L. sativus it was not possible with any confidenceto make a genetic interpretation. Glutamata oxaloacetate transaminase (GOT, EC

2.6.1.1). Most of the plants analysed showed onedarkly stained zone with another two zones near tothe origin. Since these showed little enzymic activ-ity they could not be analysed consistently. Threephenotypes were observed and the single band of

/socitrate dehydrogenase (NADP) (/DH, EC/././.42). The banding pattern of IDH (NADP)was clearly seen in one well-defined zone but with afaint and diffuse stained region closer to the origin.

39

Table 3. Phenotypic polymorphism (Pj) and weighted polymorphism (P) for 6 enzymes

Accession 6-PGD MDH PRX IDH GD GOT Average Pj Average P

404406428429432433434435438439441442444448452454455459466467468472473474475476479480481484485488489491492494495498499502504506507556558565567580585586587588

Mean

0.320.580.540.460.180.460.560.500.760.540.620.580.620.180.740.660.320.420.320.460.460.640.480.480.720.340.620.580.660.460.700.460.560.460.740.480.660.660.320.340.580.540.780.000.000.760.700.000.660.740.580.58

0.51

0.000.000.180.180.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.180.000.000.000.180.180.000.000.000.000.180.320.000.000.000.000.000.000.000.180.000.000.000.000.000.320.000.000.000.000.00

0.04

0.820.760.320.800.460.600.760.180.760.680.760.860.820.600.760.860.840.800.700.800.700.820.820.720.680.540.780.460.820.580.800.760.540.480.780.820.680.580.680.860.720.580.460.540.480.660.720.560.840.450.780.86

0.69

0.420.000.000.320.340.340.000.320.000.480.480.480.000.420.000.180.540.180.000.180.000.000.580.580.000.180.560.180.420.420.320.180.180.320.000.540.000.560.180.000.000.180.180.180.000.000.420.000.460.480.000.00

0.23

0.000.000.000.500.000.180.000.480.000.480.420.000.000.000.320.000.320.340.000.000.000.000.480.000.000.180.180.000.000.480.000.500.420.420.540.000.000.420.000.000.000.560.000.000.000.420.000.000.620.000.480.00

0.17

0.000.000.000.340.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.000.180.000.000.000.000.000.000.000.000.000.00

0.01

0.260.220.170.430.160.260.220.250.250.360.330.320.240.200.300.280.340.290.170.240.190.270.390.300.230.240.390.200.320.320.300.650.340.330.340.310.220.370.200.200.250.340.240.120.080.310.360.090.430.280.310.24

0.090.050.090.350.060.120.050.180.060.210.190.130.060.080.130.080.180.140.030.070.040.110.220.120.060.150.220.070.110.200.100.230.250.180.180.120.050.210.060.040.100.240.080.040.010.150.200.020.250.110.160.06

40

Table 4. Phenotypic polymorphism (Pj) for 6 enzymes in 52 accessions of L. sativus grouped according to geographical origin

Region No. accs 6-PGD MDH PRX IDH GD GOT Average Pj Average P

W. & S.W. EuropeC. & S. EuropeUSSRW.AsiaTunisiaIndiaChileAustralia

48

20926

0.390.570.540.470.670.350.780.45

0.000.100.120.060.000.000.000.09

0.690.720.710.640.760.670.460.70

0.190.210.260.260.000.240.180.09

0.000.330.110.400.210.220.000.00

0.000.000.000.060.000.000.000.00

0.210.320.270.300.270.250.240.22

0.060.180.090.180.100.120.180.082

For MDH only 17% of accessions were polymor-phic, with an average value of 0.04, and for GOTthe value was even lower at 0.01, represented byonly 4% of the accessions. Among the accessions,the weighted average polymorphism (P) rangedfrom 0.01 (accession 558 from India which waspolymorphic only for PRX) to 0.35 (accession 429from Turkey which was the only one polymorphicfor all six enzyme systems). Only one other acces-sion, 580 from Iran was polymorphic for a singleenzyme, PRX, with an average Pj of 0.09, and P of0.02. Grouping accessions according to geograph-ical areas of origin (Table 4) or according to flowercolour forms (Table 5) as indicated by Jackson &Yunus (1984), showed no significant differencesbetween average polymorphism or the weighted

average.

phenotype 1 was seen in over 99% of plants. Thiswas the least variable enzyme studied, with onlytwo accessions (429 and 506) showing variation.Accession 429 from Turkey had all three pheno-types, whereas 506 had the single-banded pheno-types 1 and 2.

In Pisum sativum, four loci are present (Weed-en, 1985), but in Vicia faba (Mancini et al., 1989)and Lens culinaris (Zamir & Ladizinsky, 1984)only three GOT loci have been observed. In spin-ach and pea, the enzyme is a dimer (Huang et al.,1986; Weeden & Marx, 1987). In L. sativus, threeputative loci were detected, and at locus C whereanalysis was possible, there were apparently threealleles. Only one plant was heterozygous, and allother plants were homozygous.

Phenotypic polymorphism. Peroxidase showed thehighest average polymorphism (Pj) with a value of0.69, and all accessions were polymorphic for thisenzyme (Table 3). The second highest value wasshown by 6-PGD, with a value of 0.51, and 94% ofaccessions were polymorphic. IDR and galactosedehydrogenase were less polymorphic with valuesof 0.23 and 0.17 respectively, and 63% and 40% ofaccessions were polymorphic for these enzymes.

Discussion

The grasspea, L. sativus, is a morphologically vari-able species, based on an analysis of floral andvegetative characters (Jackson & Yunus, 1984).The isozyme data presented here also show consid-erable polymorphism although there is no apparent

Table 5. Phenotypic polymorphism (Pj) for 6 enzymes in 52 accessions of L. sativus grouped according to flower colour

Flower colour No. accessions 6-PGD MDH PRX IDH GD GOT Average Pj Average P

WhiteBlueMixed blueand white

0.540.470.51

0.660.680.74

0.250.250.16

0.110.210.18

0.270.280.27

192013

0.040.020.06

0.00

0.03

0.00

0.120.130.13

41

then be possible to make selections for forms withlower concentrations of the neurotoxin, ~-N-ox-alyl-L-a,~-diaminopropionic acid or ODAP,which causes neuro-lathyrism, or indeed to selectnon-toxic forms which might allow wider utiliza-tion of this minor pulse. Given its agricultural im-portance in India and in several other countries,and given that with increasing stresses being placedon the agricultural environment world wide nowand in the future (Jackson et al., 1990), the grass-pea clearly has the potential to increase in impor-tance as a grain legume. What is needed is a con-certed breeding effort to exploit the variationwhich we have demonstrated exists in this under-exploited crop genetic resource.

Acknowledgement

The senior author acknowledges receipt of spon-sorship from the Public Services Department, Ma-laysia, and the University of Agriculture, Malaysiafor study leave.

References

correlation with morphology. Furthermore the iso-zyme variation cannot be interpreted on a geo-graphical basis.

Although we did not carry out formal geneticanalyses of these isozyme banding patterns, anal-ogy with similar systems in closely related generasuch as Pisum, Lens or Vicia and other plant spe-cies had enabled us to speculate on the extent ofgenetic variation at particular isozyme loci. Eventhough our interpretations must be regarded astentative until further analyses are undertaken bystudy of progenies, it is clear that the banding pat-terns, which we have observed, do represent allelicvariation at several loci.

The important feature to emerge from this studyis that L. sativus, an autogamous species as we havenoted during our own experimental work, is varia-ble with regard to four of the enzymes studied, andseveral of the hypothesized loci appear to be heter-ozygous. The frequency of putative heterozygousloci is greater than we would have expected givenour knowledge of the breeding system of L. sativus,and by analogy with the breeding systems of othergrain legumes in which there is a high frequency ofautogamy. Differences in terms of alleles or allelicfrequencies might be expected on a geographicbasis, on a large or small scale, since in other spe-cies correlations between isozyme phenotypes andenvironment have been reported in A vena barbata(Kahler et al., 1980) and Triticum dicoccoides inIsrael (Nevo et al., 1982). Furthermore, allozymevariation in 7: dicoccoides is correlated with spike-let variation, even on a mosaic pattern betweencollecting sites.

A high level of phenotypic polymorphism typ-ifies some L. sativus accessions. Given the fact thatthe secondary gene pool of this species is ratherlimited, and utilization of the tertiary gene pool isdifficult because hybridization is difficult, it is ap-parent to us that exploitation of the primary genepool of L. sativus is the most appropriate strategyto employ for the improvement of this cultigen. Ifthe level of apparent heterozygosity in L. sativuswhich we have detected is a reflection of widerheterozygosity in this species, then crosses betweendifferent forms are likely to lead to the release of aconsiderable amount of hidden variation. It might

Goodman, M.M., C.W. Stuber, C.N. Lee & F.M. Johnson,1980. Genetic control of malate dehydrogenase isozymes inmaize. Genetics 94: 153-168.

Gottlieb, L.D., 1982. Conservation and duplication of isozymesin plants. Science 216: 373-380.

Hamrick, J .L. & M.J. Godt, 1990. Allozyme diversity in plantspecies. In: A.H.D. Brown, M. T. Clegg, A.L. Kahler & B.S.Weir (Eds.), Plant Population Genetics, Breeding, and Ge-netic Resources. pp. 43-63. Sinauer Associates Inc., Sunder-land, Massachusetts.

Harlan, J .R. & J .M.J. De Wet, 1971. Toward a rational classifi-cation of cultivated plants. Taxon 20: 509-517.

Hoffman, D.L., D.E. Soltis,F.J. Muehlbauer&G. Ladizinsky,1986. Isozyme polymorphism in Lens (Legurninosae). Sys-tem. Bot. 11: 392-402.

Huang, A.H.C., K.D.F.. Liu & R.J. Youle, 1976. Organelle-specific isozymes of aspartate-a-ketoglutarate transaminasein spinach leaves. Pl. Physiol. 58: 110-113.

Hvid, S. & G. Nielson, 1977. Esterase isoenzyme variants inbarley. Hereditas 87: 155-162.

Jackson, M. T. & A.G. Yunus, 1984. Variation in the grass pea(Lathyrus sativus L.) and wild species. Euphytica 33: 549-559.

Jackson, M., B.V. Ford-Lloyd & M.L. Parry (Eds.), 1990. Cli-

42

and pentose phosphate pathways in proplastids from the de-veloping endosperm of Ricinus communis. Pl. Physiol. 61:871-877.

Suso, M.J. & M. T. Moreno, 1986. Isoenzymatic polymorphismof superoxide dismutase (SOD) in Vicia faba and its system-atic implication. FABIS Newsletter 16: 3-5.

Tadmor, Y., D. Zamir & G. Ladizinsky, 1987. Genetic mappingof an ancient translocation in the genus Lens. Theor. Appl.Genet. 73: 883-892.

Tanksley, S.D. & T.J. Orton (Eds.), 1983. Isozymes in PlantGenetics and Breeding. Part A. Elsevier, Amsterdam.

Weeden, N., 1985. An isozyme linkage map for Pisumsativum.In: P. Hebblethwaite, M.C. Heath & T. Dawkins (Eds.), ThePea Crop, pp. 55-66. Butterworths, Guildford.

Weeden, N., 1988. Polymorphic isozyme loci identified in Pi-sum. Pisum Newsletter 20: 46-48.

Weeden, N. & G.A. Marx, 1987. Further genetic analysis andlinkage relationships of isozyme loci in pea. J. Hered. 78:153-159.

Weeden, N. & J.F. Wendel, 1990. Genetics of plant isozymes.In: D.E. Soltis & P.S. Soltis (Eds.), Isozymes in Plant Biol-ogy. pp. 46-72. Chapman & Hall, London.

Wendel,J.F. & N.F. Weeden, 1990. Visualization and interpre-tation of plant isozymes. In: D.E. Soltis & P.S. Soltis (Eds.),Isozymes in Plant Biology. pp. 5--45. Chapman & Hall, Lon-don.

Yamamoto, K. & U. Plitmann, 1980. Isozyme polymorphism inspecies of the genus Vicia (Leguminosae). Japan. J. Genet.55: 151-164.

Yamamoto, K., T. Fujiwara&I.D. Blumenreich, 1986. Isozym-ic variation and interspecific crossability in annual species ofthe genus Lathyrus L.. In: A.K. Kaul & D. Coombes (Eds.),Lathyrus and Lathyrism. Third World Medical ResearchFoundation, New York.

Yunus, A.G. & M. T. Jackson, 1991. The gene pools of thegrasspea (Lathyrus sativus L.). Plant Breeding (in press).

Zamir, D. & G. Ladizinsky, 1984. Geneticsofallozyme variantsand linkage groups in lentil. Euphytica 33: 329-336.

matic Change and Genetic Resources. Belhaven Press, Lon-don.

Kahler, A.L., R.W. Allard, M. Krzakowa, C.F. Wehrhahn &E. Nevo, 1980. Associations between isozyme phenotypesand environment in the slender wild oat (Avena barbata) inIsrael. Theor. Appl. Genet. 56: 31-47.

Mancini, R., C. De Pace, G. T. Scaracia-Mugnozza, V. Delre &D. Vittori, 1989. Isozyme gene markers in Vicia faba L.Theor. Appl. Genet. 77: 657-667.

Mora, T.D., J .A. Gonzalez & J .M. Serradilla, 1983. Prelimina-ry studies on genetic determination of electrophoresis alloen-zyme bands in Vicia faba L. FABIS Newsletter 7: 11-12.

Nevo, E., E.M. Golenberg, A. Beiles, A.H.D. Brown & D.Zohary, 1982. Genetic diversity and environmental associ-ations of wild wheat Triticum dicoccoides, in Israel. Theor.Appl. Genet. 62: 241-254.

Ni, W., E.F. Robertson & H.C. Reeves, 1987. Purification andcharacterization of cytosolic NADP specific isocitrate de-hydrogenase from Pisum sativum. Pl. Physiol. 83: 785-788.

Parzysz, H., L. Gruchala & J. Przybylska, 1986. Isoenzymevariation in the genus Pisum. III. Electrophoretic patterns ofesterases from cotyledons of ungerminated seeds. Genet.Polon. 26: 297-301.

Pinkas, R., D. Zamir & G. Ladizinsky, 1985. Allozyme di-vergence and evolution in the genus Lens. Pl. Syst. Evol. 151:131-140.

Przybylska, J., S. Blixt, C.H. Parzysz & Z. Zimniak-Przybyl-ska. 1982. Isozyme variation in the genus Pisum. I. Electro-phoretic patterns of several enzyme systems. Genet. Polon.23: 103-121.

Schnarrenberger, C., A. Oeser & N.E. Tolbert, 1973. Twoisoenzymes each of glucose-6-phosphate dehydrogenase and6-phosphogluconate dehydrogenase in spinach leaves. Arch.Biochem. Biophys. 154: 438-448.

Shaw, C.R. & R. Prasad, 1970. Starch gel electrophoresis ofenzymes -a compilation of recipes. Biochem. Genet. 4:297-320.

Simcox, P .D. & D. T. Dennis, 1978. Isozymes of the glycolytic