O. Dolstra

Foundation for Agricultural Plant Breeding (SVP), Wageningen

Synthesis andfertilit y of xBrassicoraphanus andway so f transferring Raphanuscharacter s to Brassica

Centre for Agricultural Publishing and Documentation

Wageningen - 1982

'zcé^l^ CIP-GEGEVENS

Dolstra, 0.

Synthesis and fertility of xßrassicoraphanus and ways of transferring Raphanus characters to Brassica / 0. Dolstra. - Wageningen: Pudoc. - 111. - (Agricultural research reports; 917) Ook verschenen als proefschrift Wageningen, 1982. - Met lit. opg. ISBN 90-220-0805-3 SISO 632 UDC 633 Trefw.: plantenteelt.

ISBN 90 220 0805 3

The author graduated on 29 October 1982 as Doctor in de Landbouwweten­ schappen at the Agricultural University, Wageningen, the Netherlands, on a thesis with the same title and contents.

© Centre for Agricultural Publishing and Documentation, Wageningen, 1982.

No part of this book may be reproduced or published in any form, by print, photoprint, microfilm or any other means without written permission from the publishers. Abstract

Dolstra,0. ,1982 .Synthesi san dfertilit yo fxBrassicoraphanu san dway so ftransfer ­ ringRaphanu scharacter st oBrassica .Agric .Res .Rep . (Versl.landbouwk .Onderz. ) 917,ISB N90-220-0805-3 ,(viii )+ 9 0p. ,3 8tables ,1 4figs ,11 4refs ,Eng .an dDutc h summaries. Also:Doctora lthesis ,Wageningen .

About25 0intergeneri chybrid swit hth egenom econstitutio nAR ,AAR Ro rAR Rwer eob ­ tained fromove r1 500 0crosse sbetwee nBrassic a campestris (AAo rAAAA ,femal eparent ) andRaphanu ssativu s(R Ro rRRRR) .Th epoo rcrossabilit ywa sshow nt ob ea consequenc e ofvariou sbreedin gbarriers .Th ediploi dan dtetraploi dhybrid swer estudie d inform , fertility,chromosom eassociatio na tmeiosi san dcrossabilit ybot hamon gth ehybrid s andi ncrosse swit hth eparenta lspecie san dBrassic ananus .xBrassicoraphanu s (AARR) wasbackcrosse dtwic ewit hB .campestri s (AA),an dwa sals ocrosse dan dbackcrosse d withB .napu s (AACC).For man dchromosom eassociatio na tmeiosi si nAA Ran dAAC Rhy ­ bridswer estudied . Twopopulation so fxBrassicoraphanu s (AARR)wer epropagate d andstudie d inmor ede ­ tail,partl yfo rpossibl esuitabilit ya sa ne wcrop .Mos tplant swer ehighl ysteril e becauseo fbreedin gbarriers ,whic hwer esimila rt othos ei nth eorigina lintergeneri c crosses.Th eincreas ei nfertilit yan dchromosoma lstabilit ywer estudie d inth efirs t threegeneration so fth epopulations .

Freedescriptors :Brassic a campestris,Raphanu s sativus,incongruity ,breedin gbar ­ riers,embry oculture ,intergeneri chybrids ,xBrassicoraphanus ,fertility ,matin g system,introgression ,Brassic anapus ,allosyndesis ,forag ecrop . Contents

1 General introduction 1 1.1 Rationale 1 1.2 Species and allopolyploid hybrids 1 1.3 Scopeo fth eresearc hprogramm e 2

2 Crosses between Brassica campestris and Raphanus sativus 5 2.1 Introduction 5 2.2 Material andmethod s 6 2.3 Results 9 2.3.1 Settingo fsiliqua e and seeds 9 2.3.2 Seed size 9 2.3.3 Productiono fhybrid s andmatromorph s 10 2.3.4 Useo fgeneti cmale-sterilit y 13 2.3.5 Cultureo fembryo s 14 2.3.6 Genome constitution andnumbe ro fchromosome si n theprimar y hybrids 15 2.4 Discussion 17

3 Nature of breeding barriers 19 3.1 Introduction 19 3.2 Material andmethod s 19 3.3 Results 20 3.4 Discussion 22

4 Primary allodiploid ARhybrids and their progeny 24 4.1 Introduction 24 4.2 Material andmethod s 25 4.3 Results 26 4.3.1 Formo fA Rhybrid s 26 4.3.2 Chromosome association atmeiosi s 28 4.3.3 Male and seed fertility 30 4.3.4 Crossabilityo fA Rhybrid swit hvariou s species 31 4.3.5 Somatic andmeioti c doublingo fchromosom e number 33 4.3.6 Progenyo fcrosse sbetwee nA Rhybrid s 33 4.4 Discussion 34 5 xBrassicoraphanus: form, fertility and crossability with various species 37 5.1 Introduction 37 5.2 Material andmethod s 38 5.3 Results 38 5.3.1 Form and fertilityo fprimar y hybrids 38 5.3.2 Crossabilitywit h Brassica campestris, Raphanus sativus and B. napus 40 5.4 Discussion 42

6 xBrassicoraphanus: sterility factors and mating system 44 6.1 Introduction 44 6.2 Material andmethod s 45 6.3 Results 45 6.3.1 Meiosis 45 6.3.2 Barriersi nth eprogami c phase 49 6.3.3 Barriers after fertilization 52 6.3.4 Mating system 54 6.4 Discussion 55

7 Improvementin fertility of xBrassicoraphanus 57 7.1 Introduction 57 7.2 Material andmethod s 57 7.3 Results 59 7.3.1 Chromosome numbers 59 7.3.2 Changesi nfou r fertility characteristics 60 7.3.3 Effectso fchromosom e numbero nfertilit y 64 7.4 Discussion 64

8 AAR hybrids and introduction of alien chromosomes into Brassica campestris 66 8.1 Introduction 66 8.2 Material andmethod s 66 8.3 Results 67 8.3.1 Form and chromosome associationi nAA R hybrids 67 8.3.2 Attemptst ointroduc e alienchromosome s 69 8.4 Discussion 71

9 Ways of transferring Raphanus characteristics to Brassica napus 72 9.1 Introduction 72 9.2 Material andmethod s 72 9.3 Results 73 9.3.1 Hybrids from crossesbetwee n x-Brassicoraphanus and Brassica napus 73 9.3.2 Chromosome associationi nAAC Rhybrid s 75 9.3.3 Firstbackcros s generation 76 9.4 Discussion 76

10 Prospects 78 10.1 Resistancet obee teelwor m 78 10.2 xBrassicoraphanus 78 10.3 Introgression 79

Suvomary 80

Samenvatting 83

References 86 1 General introduction

1.1 RATIONALE

Likemos tothe rcruciferou s crops, Brassica spp.ar ehost s toth ebee t eelworm (Heterodera schachtii Schm.)(de nOuden ,1954 ;Steele , 1965), which isdistribute d allove rth eworl d (Stelter,1973 )an d isth emos t seriousnematod e of sugar-beet (Beta vulgaris L.) inEurop e (Heijbroek, 1979). Inth eNetherlands , forexample ,a quarte r ofth e sugar-beetare a hasbee n infectedb y thisnematode ,especiall y inth e south-west and south-east (Heijbroek, 1979). Problems arisemainl y fromto onarro wcro p rotations.Th e growingo fsugar-bee t iseconomicall y attractive toth e farmers, and theywil l try tocontinu ewit h suchrotations ,despit e dis­ advantages from aphytosanitar y pointo fview . Underpresen tcircumstances ,n oplac e islef t for Brassica crops in rotationswit hbeets ,unles s resistantvarietie s arebre dwit hpreferabl y the ability toreduc e thepopulatio n density ofth eeelworm . Inth e genus Brassica, however,n o resistance hasbee n found sofar (Baukloh,1976 ; Lubberts& Toxopeus , 1982). Inth e 1960s,report s appeared inGerma n farming journals onresistanc e in Raphanus sativus L.var . oleiformis, theoil-see d radish.Th e early contradictory reportswer e reviewedb y Hirling (1976).Wit h oil-seed radish theeelwor m usuallymultiplie d slow­ ly,bu t ina fe wreport s rather quickly. Inrecen tyears ,th epresenc e ofresistanc e in Raphanus hasbee nconfirme db yBauklo h (1976)an dToxo ­ peus& Lubbert s (1979). Baukloh (1976)studie d the inheritance and assum­ ed asingl e dominantgen e forresistance . Thegenu s Brassica includesmajo r crops for fodder,oil-see dproduc ­ tion, and greenmanure ,som eo fwhic hmigh t fitquit ewel l into cropro ­ tationswit h sugar-beet.S o aprogramm ewa s started onintergeneri c hy­ bridization between Brassica spp.an dnematode-resistan t formso f R. sa­ tivus with theultimat e goal oftransferrin g resistance to Brassica. A second argument forthi sprogramm e was thepresenc e ofresistanc e to club-root (Plasmodiophora brassicae Woron.)i noil-see d radish.

1.2 SPECIESAN DALLOPOLYPLOI D HYBRIDS

Theprogramm e on intergeneric hybridization includes the species R. sativus (2n= 18 ,RR) , B. campestris (2n= 20,-AA )an d B. napus (2n =38 , AACC). Their genome constitutions and interrelationships areoutline d in Figure 1 (Mizushima, 1980). B. napus is anatura l allotetraploid andi s composed oftw o genomes of B. campestris andtw o of B. oleracea (2n =18 , CC). Table 1enumerate s the subspecies orbotanica lvarietie s distin­ guished inthes e species.Accordin g toHel m (1957), R. sativus comprises fivebotanica lvarieties ,whic h includevariou svegetable s andagricul ­ tural crops.Th ebes tknow n inEurop e are radish,gian tradish ,blac k radish, and oil-seed or fodder radish.Th e latter ishoweve r mainly used asgree nmanure .Mougri-radis h isgrow n in south-eastAsia , especially for its leaves andyoun gpod s (Banga, 1976). For B. campestris, the clas­ sification ofOlsso n (1954)ha sbee n followed.H e distinguished various subspecies,whic hwer e classified by other taxonomists as separate spe­ cies. B. campestris containsman y oriental leafyvegetables , and several root, oilseed and fodder crops. B. napus also includesmajo r root,oil ­ seed and fodder crops,althoug h itsdomesticatio n started only afe whun ­ dred years ago (Toxopeus,1974a ;McNaughton , 1976). Since the discovery ofth egenom e constitution of B. napus (U, 1935), this speciesha s often beenresynthesize d from itsdiploi d ancestors.Thi sha s increased genetic variation inth eexistin g crops andeve nresulte d ina headin g type cal­ led 'Hakuran' (Nishi, 1980). More detailed information onth e four spe­ cies inth epresen t study is found inBoswel l (1949), Herklots (1972), Simmonds (1976)an dNish i (1980).

Intergeneric hybridsbetwee n R. sativus and Brassica species havea long andwel l knownhistory . Atpresen tvariou smeioticall y stableal ­ lopolyploids areknow n (Fig.1) .Plant s with thegenom e constitution RRCCwer e already obtainedb y Karpechenko (1927; 1928)an dname d Raphanobrassica. Someyear s later,Terasaw a (1932)wa s the firstt opro ­ duce allotetraploids withth e genome constitutionAARR ,whic hh e called Brassicoraphanus. Inthi s study, thename s xRaphanobrassica and xBrassi- coraphanus areuse d forplant swit hth egenom econstitutio n RRCC and AARR, respectively, irrespective ofth e femaleparen t inth e original crosses.Accordin g toth e international rules fornomenclature , those names should beprecede d bymultiplicatio n sign (International Bureau forPlan tTaxonom y andNomenclature , 1980). A less known type isth e allohexaploid with the constitutionAACCRR ,whic hwa s recently produced by Clauss (1978).

1.3 SCOPE OFTH ERESEARC H PROGRAMME

Thepresenc e of resistance toth ebee teelwor m in R. sativus wasth e mainreaso n forstartin g aprogramm e onintergeneri c hybridization in order to transfer this trait toth egenu s Brassica, especially to B. campestris and B. napus. Those species includemos to fth emajo r arable Brassica crops grown inth eNetherlands . Figure 1indicate s thatvariou s ways for transferring genes from Raphanus to Brassica mayb e feasible. Table 1.Lati nan dcommo nEnglis hname so fsubspecie san dbotanica lvarietie si n Raphanussativu sL. ,Brassic a campestrisL .an dBrassic anapu sL .

Species Subspecies/ Commonname(s ) botanicalvariet y

R. sativusL . var.gayanu sWebb , wildradis h var.oleiformi sPers . oil-seed orfodde rradis h var.mougr iHel m Mougri-radisho rrat-taile d radish var.nige r (Miller) blackradish ,gian tradish , Pers.sens ulat . Japaneseradis h var.sativu sHel m radish

B.campestri sL . ssp.eu-campestri sOlsso n wildtyp e ssp.oleifer a (Metzg.)Sinsk . turnip-rape ssp.rapifer a (Metzg.)Sinsk . turnip ssp.chinensi s (L.)Makin o Chinesemustar d ssp.pekinensi s (Lour.)Olsso n Pe-tsai,Chines ecabbag e ssp.narinos a (Bailey)Olsso n ssp.nipposinic a (Bailey)Olsso n ssp.dichotom a (Roxb.)Olsso n toria ssp.perviridi sBaile y tendergreen ssp.triloculari s (Roxb.)Olsso n yellow-seeded sarson

B.napu sL . var.napobrassic a (L.)Peterm . swede ssp.oleifer a (Metzg.)Sinsk . oil-seed and fodder rape

R. sativus

x Brassicoraphanus x Raphanobrassica

B.campestris B. napus B.oleracea Figure 1.Genom edesignations ,chromosom enumber san d interrelationships ofRaphanu s sativus,som eBrassic a speciesan dallopolyploi dhybrids . Direct introgression into B. napus by crossing this specieswit h R. sa- tivus was apparently difficult, although somehybrid shav ebee n obtained (see for instance Yarnell, 1956). Amor epromisin g firstste pwa sth e cross B. campestris x R. sativus, because (1)th e crossability isbette r (Yarnell, 1956)an d (2)th e resulting allotetraploid hybrids mayb ea bridgebetwee n R. sativus and B. napus. So alleffort swer econcentra ­ tedo nthi sparticula r cross (Chapter2) . Most initial hybrids grewvigorousl y and suggested thatth eallote ­ traploid hybrids (= xBrassicoraphanus) mightb e suitable asa fodder crop,combinin g rapid growth andbee teelwor m and club-root resistance of oil-seed radishwit h thebette r palatability ofth e Brassica parent. Aprerequisit e for successful selectionwithi n such a 'newspecies 'i s the availability of abroa d geneticvariation , i.e. toproduc e hybrids involving many genotypes ofth eparenta l species.S o thenatur e ofbreed ­ ingbarrier s andway s of improving crossability between B. campestris and R. sativus were studied (Chapter 2an d3) . Severalpropertie s ofth eprimar y hybrids,e.g . form, fertility, crossability andmeiosi s aredescribe d inChapte r 4 and 5.Possibl e causes of sterility and improvement of fertility in xBrassicoraphanus aregive n inChapte r 6an d 7. We tried totransfe r Raphanus genes to Brassica, intw oways .Th e firstapproac hwa s toproduc e in B. campestris monosomic addition lines carrying anextr a Raphanus chromosome.Th e intentionwa s to test such lines forresistanc e tovariou s diseases and afterwards totransfe r the resistance genes from separate Raphanus chromosomes ina mor e directed attempt into the B. campestris genome (Chapter 8). The second approach was abackcros sprogramm e of y-Brassicoraphanus with B. napus asth ere ­ currentparen t (Chapter9) . 2 Crosses between Brassica campestris and Raphanus sativus

2.1 INTRODUCTION

Sinceth e1920 s several researchworkers ,especiall y fromJapa nhav e reported thehybridizatio n of B. campestris and R. sativus (Table2) . Terasawa & Shimotomai (1928)obtaine d the firstdiploi d hybrids.Mor e reports onsuccessfu l hybridization followed. Initially thecrosse swer e carried outa tth ediploi d levelonl ybu tlate r autotetraploid formso f eitheron eo rbot h specieshav ebee nuse d aswell ,especiall y during the lasttw odecade s (Table 2). Crossabilitywa spoo r inal lcombinations . Thebes tresult swer enevertheles s obtained fromcrosse sbetwee ndiploi d forms,usin g B. campestris asth e femaleparent . Inthos ecrosses ,ther e wasmarke dvariatio n inth enumbe r ofhybrid sproduce d per pollination (Table2) . As fara sknown ,th eprimar yhybrid s reportedb yvariou s authors (Table2 )ha dnearl y alwaysth eexpecte d genome constitution,bu tsom e deviantplant swer e found in2 xx 2x crosseswit h R. sativus asth efe ­ maleparent .U et al. (1937),Morri s &Richhari a (1937)an dMizushim a (1950a)obtaine d amongthei rhybrid s five allotriploid plants comprising twogenome s of Raphanus andon eo f Brassica. Theseplant sprobabl y arose from afertilize d 2n egg cell. Apart form an interest inth e intergeneric hybrids assuch ,th e crossesmentione d inTabl e 2hav ebee nmad emainl y fortw oreasons :t o study thecause s ofmatromorph y (Nishie t al., 1964;Tokumasu ,1965 ; 1970a;Opeft a& Lo ,1978 )an dt oge t informationo n interspecific incom­ patibility (Kakizaki,1925 ;Becker ,1951 ;Hinat a et al., 1974).Usuall y progeny ofcrosse sbetwee n B. campestris and R. sativus consist partly ofmatromorphi cplants .Tokumas u (1965;1970a )showe d thatsuc hplant s wereno tcompletel y homozygous asNish i et al. (1964)assumed .Matromor ­ phy inCrucifera e isa kin d opdiploi dparthenogenesi s andha sbee nob ­ served inman y interspecific and intergeneric crosseswithi nthi s family (Eenink, 1974). Thischapte rpresent s resultso fintergeneri c crossesbetwee ndiploi d and tetraploid formso f B. campestris and R. sativus. The latter species was alwaysuse d asth epolle n donor,becaus e inthi sdirectio nth ecross - abilitywa smuc hbette r than inth ereciproca l (Table 2). Many accessions of B. campestris wereuse d inorde r tominimiz e thechanc e ofusin gpoor ­ lycrossabl e genotypes only.Th emal eparent svarie d less,bein gconcen - Table2 .Lis to freporte d crossesbetwee ndiploi dan dtetraploi d formso f Brassica campestris(A Aan dAAAA )an dRaphanu s sativus(R Ran dRRRR) .

Cross Numbero f Numbero f Numberfract ­ Source pollinations hybrids ionhybrid s topollinat ­ ions(% )

AAx R R 33 0 0 Kakizaki (1925) 90 6 6.7 Terasawa& Shimotoma i (1928) 428 12 2.8 Morris& Richhari a (1937) 217 0 0 Becker (1951) 903 167 18.5 Mizushima (1952) 800 3 263 3.3 Nishie tal . (1964) 286 31 10.8 Tokumasu (1970b) 82 0 0 Hinatae tal . (1974) 1328 7 44 0.3 Opena& L o(1978 ) 211 1 13 0.6 Namai (1980) 24 0 0 Takeshitae tal . (1980)

RRx A A 39 0 0 Kakizaki (1925) 30 0 0 Terasawa& Shimotoma i (1928) 1 >2 0.2 Ue tal . (1937) 413 2 0.5 Morris &Richhari a (1937) 328 0 0 Becker (1951) 477 5 2 0.04 Nishi et al. (1964) 417 0 0 Tokumasu (1965) 111 1 0 0 Namai (1980) 19 0 0 Takeshita et al. (1980)

AAx RRR R •} 2 Ellerström (pers. commun.)

RRRRx A A ? 0 Ellerström (pers. commun.)

AAAAx R R 481 0 Nishi et al. (1964)

RRx AAA A 471 0 Nishi et al. (1964)

AAAAx RRR R 58 0 Tokumasu (1970a)

RRRRx AAA A 132 0 Tokumasu (1970a)

Remark: Intergeneric hybrids were also reported by Morris (1936), Nishiyama (1946), Hosoda (1946; 1947) Mizushima (1950a), Clauss (1978), McNaughton &Ros s (1978), Ellerström (1978), Olsson &Ellerströ m (1980), Namai et al. (1980).

trated to fodder radish because of the requirement of combined resistance to beet eelworm and club-root. Some attempts were made to improve the re­ sults of these intergeneric crosses by culture of immature embryos in vitro.

2.2 MATERIAL AND METHODS

The diploid and tetraploid genotypes of B. campestris and R. sativus used for intergeneric hybridization are listed in Table 3. The fodder radishes 'Siletta', 'Levana', 'Palet' and RS14 were chosen because they were partially resistant to beet eelworm. In most tests for resistance Table3 .Description ,ploid yan dorigi no fBrassic aan dRaphanu saccession suse d forintergeneri ccrosses .

Designation Species/subspecies Description Ploidy Origin orbotanica l variety.

B.campestri s Al ssp. rapifera inbredlin e BC183 2x A2 ssp. rapifera inbredlin e BC18 2 2x A3 ssp. rapifera inbredlin e BC171 2x A4 ssp. rapifera inbredlin e BC172 2x A5 ssp. rapifera inbredlin e BC181 2x A6 ssp. rapifera inbredlin e BC176 2x A7 ssp. rapifera inbredlin e BC184 2x A8 ssp. rapifera cv.Trofe e 2x A9 ssp. oleifera boterzaad 2x A10 ssp. pekinensis 2x All ssp. chinensis Shan-Jue-Man-Tsien-Tsaj 2x A12 ssp. chinensis 2x A13 ssp. chinensis S line fromA1 2 2x AA1 ssp. rapifera cv. Novitas 4x 4x AA2 ssp. rapifera cv.Tarond a 4x AA3 ssp. rapifera cv.Tigr a AA4 ssp. pekinensis 2243 4x 4x AA5 ssp. nipposinica 2245 4x AA6 ssp. perviridis 2240 ssp. narinosa 2260 4x AA7 4x AA8 ssp. chinensis 2242

R.sativu s Rl var. oleiformis cv.Silett a 2x R2 var. oleiformis cv.Levan a 2x R3 var. mougri Rs4.0 1 2x R4 var. niger Ra74/6 5Chin a 1956:Kanton3 52 x R5 var. oleiformis Rs6.0 4 2x R6 var. niger Ra 113/64Japanes ewhit e NewDelh i 2x RR1 var. oleiformis cv.Pale t 4x RR2 var. oleiformis RS14 4x

*] 1.Stichtin gvoo rPlantenveredelin g (SVP),Wageningen ,Th eNetherlands . 2.Zwaa n& D eWiljes ,Zaadteel te nZaadhande lBV ,Scheemda ,Th eNetherlands . 3.Zentralinstitu t fürGeneti kun dKulturpflanzenforschung ,Gatersleben ,DDR . 4.Cebeco-Handelsraad ,Rotterdam ,Th eNetherlands . 5.Zelde rBV ,Ottersum ,Th eNetherlands . 6.Kon .Zaadteel te nZaadhande lSlui se nGroo tBV ,Enkhuizen ,Th eNetherland s 7.Swedis hSee dAssociation ,Svalöv .Sweden . 8.Kon .Kweekbedrij fe nZaadhande lD.J .va nde rHav eBV ,Kappelle , TheNetherlands . 9.Scottis hPlan tBreedin gStation ,Pentlandfield ,Unite dKingdom .

usually some multiplication of this nematode occurred, but to a relati­ vely low degree (Toxopeus, pers. commun.). Fodder-radish varieties often have a high degree of resistance to club-root (Crute et al., 1980). The German variety Siletta was resistant to many isolates of club-root (Toxopeus, 1974b). The fodder radishes 'Levana1 and 'Palet', are direct derivatives from 'Siletta' and have probably the same genes for resis­ tance. Mosto fth eplan tmateria l wasgrow n inrathe rwel l conditionedgreen ­ houses attemperature s varying from 15t o2 0 °Can d adaylengt hbetwee n 14an d 17h .Th eplant swer e grown inpot s ofcapacit y about 2\ litres. The intergeneric crosseswer emad eb y hand inth eperio d from January to May of 1975 and 1976.Ope n flowers andyoun g flowerbud swer e removed fromth e inflorescences and subseguently the resto fth ebud swa semas ­ culated.Th e inflorescences,wit h anaverag e ofabou t 15bud swer e bagged.The ywer epollinate d between 13:00 and 15:00 immediately after emasculation in 1975 andtw o tothre e days after emasculation in 1976b y putting asmal l amounto f apolle nmixtur e ontoth e stigmaswit h asmal l softbrush .Suc hmixture swer e composed ofalmos tequa l amounts of fresh pollen collected from 25t o4 0plant spe r accession.A tth emomen to f pollination in1976 ,th e oldestbu dwa smarke d andbud s andope n flowers ofeac h inflorescence were counted. Threet o fourweek s later thedevel ­ oped siliquae were counted. Atharvest ,th etota lnumbe r ofsiliqua e andth enumbe r ofseed swer e counted foreac h inflorescence.Th e largestdiamete r of each seedwa sde ­ termined as ameasur e of seed size.Larg e seedswer e germinated inPetr i dishes onwe t filterpape r at2 3 °Ci nth e dark. Small and shrivelled seedswer e firstdisinfecte d for2 0min .i n aqueous sodium hypochlorite ofmas sconcentratio n 3g/1 , followedb y thorough rinsing in sterilized water.Th e disinfected seedswer ethe nals o germinated at2 3 °C inth e darkbu t insmal lPetr i dishes ona soli d nutrientmediu m ofth ecompo ­ sitiongive nb yGuh a &Maheswar i (1964). Harberd'smetho d ofembry o culture (1969,-1971)wa sused .Th edevelop ­ mental stages ofth eculture d embryoswer e established according toth e drawings ofWilma r& Hellendoor n (1968)fo r B. oleracea. Initially all developed ovules were cultured,whethe r anembry owa spresen t ornot , but only ovules with avisibl e embryo havebee nculture d since1976 . The distinctionbetwee n hybrids andmatromorphi cplant swa s feasible onth ebasi s ofplan tmorphology , aswa s confirmed by countingchromo ­ somes inroot-ti p ormeioti c cells.Th e root-tipswer e treated and squashed forcountin g thechromosome s according toth emetho d of Jochemsen &M^ynie c (1974). Forcountin g inmeioti c cells youngbud s were fixed ina solutio no fthre epart sb yvolum e of aqueous ethanol (volume fraction ofethano l 0.96) ando fon epar tglacia l acetic acid andplace d at4 °C for afe w days.Th ebud swer ethe ntransferre d to aqueous ethanol (volume fraction ofethano l 0.7) and kept in storage at 4 °C for afe wweeks .Th ebud swer e subsequently stainedb y Snow'smeth ­ od (1963)fo r about1 4h at6 0°C . Ina smal l experiment in197 7th epossibilit y was tested ofusin gge ­ netically male-sterile B. campestris plants for intergeneric hybridiza­ tion. Fourmale-steril e F? plants from acros sbetwee n amale-steril e plant from accessionA 5 and aplan to fAl lwer e transplanted intoiso - lationcage swit h floweringplant so fR. sativus. Honey-beeswer euse d forpollination .

2.3 RESULTS

2.3.1 Setting of siliquae andseeds

Settingo fsiliqua ewa s assessed foral l thecrosse si n197 6 threet o fourweek s afterpollinatio nan da tharvesting .Th eresult s for2 xx2x , 2xx 4x , 4xx 2 xan d4 xx 4 x aresummarize di nTabl e4 fo reac h Brassica accession, irrespectiveo fpollinator .Afte r21-2 8day s the Brassica ac­ cessions differedver ymuc hi nsettin go fsiliquae .Th eploid yo fth e pollinatorsha dapparentl yn olarg eeffec t (Table4) . A lowproportio no fsiliqua e shrivelled away after initial development andwer eno tharveste d (Table 4). Theremainde r onlypartl y contained seeds.Thos e seedsgav e sometimes riset omatromorphs ,especiall yi n crossesi nwhic hth enumbe r ratioo fdevelope d siliquaet opollinate d flowerswa s comparatively low,a so n A8,A9 ,AA2 ,AA 3an dAA4 .

2.3.2 Seed size

Nearly all seedsobtaine d from the crossesi n197 5an d197 6wer e

Table4 .Numbe r fractiono fsiliqua ean do fsiliqua ewit hseed st opollinations ,an d number ratioo fseed spe rpollinatio ni nth ecrosse smad ei n197 6betwee ndiploi dan d tetraploid formso fBrassic a campestrisan dRaphanu ssativu s (Tables5 ,6 an d7) .Th e settingo fsiliqua ewa sassesse d21-2 8day safte rpollinatio nan da tharvest .Th ecolum n subheadings indicateth egenom econstitutio no fth epollinator .

Female Number fraction of sili quae Numb errati o Df Number ratioo f parent topollination s (%) sili quaewit h seeds seeds to top ollinations pollinat] (%) ions after21-2 8day s at harvest

RR RRRR RR RRRR RR RRRR RR RRRR

B.campestri s(2x ) A8 29 20 18 19 10 16 0.18* 0.35* A9 22 26 22 23 10 23 0.16* 0.45* A13 80 69 72 61 27 13 0.37 0.22

B.campestri s (4x) AA2 9 13 9 13 4 8 0.08* 0.12* AA3 2 3 2 3 1 2 0.02* 0.02 AA4 5 6 5 1 3 0 0.03* 0 AA5 49 45 15 25 0 5 0 0.06 AA6 48 53 31 47 4 17 0.04 0.20 AA7 51 48 16 14 3 1 0.03* 0.01* AA8 43 40 33 33 1 3 0.01 0.05

•]Seed sgav emostl y riset o matromorphs. Cumulative frequency Wo)

2.0 2-5 3.0 Seed size (mm)

Figure 2. Cumulative frequency distribution of seed size of hybrid and matromorphic seeds obtained from 2x x 2x and 4x x 4x crosses between Brassica campestris and Raphanus sativus,

measured. The seedswer emostl y well-filled andwer e ovoid-spherical, although somewer e shrivelled. The seed size ofhybri d andmatromorphi c seeds obtained from 2xx 2 xan d 4xx 4 x ispresente d inFigur e 2.Th e hybrid seedswer e onaverag e smaller than seedsyieldin g matromorphic plants,bu t therewa s to some extent anoverla p in seed size ofhybrid s andmatromorphs , especially for 2xx 2 xcrosses .Matromorphi c seeds from 4x x2 xan d 2xx 4 x crosses were similar in size to thato f4 xx 4 xan d 2xx 2 xcrosses ,respectively . The fewhybri d seeds from those crosses were also smaller andwer e atmos t 1-2 mm indiameter . The germinability ofth e seedswa s related to seed size.Th e seeds not germinating orwhos e seedlings diedbefor e thecotyledo n stagewer e on average smaller thanth e germinated hybrid seeds andwer e probably mainly hybrid seeds.Th e respective ratios of seeds notgerminatin g to hybrid seeds in2 xx 2x ,4 xx 2x ,2 xx 4 x and4 xx 4 xwer e approximately 1, 7,1 8 and 1,indicatin g thatespeciall y in4 xx 2 x and2 xx 4 xth e germinability ofhybri d seedswa spoor .

2.3.3 Production of hybrids and matromorphs

The results of all intergeneric crossesmad e in197 5 and 1976ar e given inTabl e 5,6 an d 7.Tabl e 5list s thecrosse s betweendiploids . In197 5nearl y all available plants frommos t Brassica accessions were crossed as femaleswit h the Raphanus accessions,R l and R3. Usually each

10 Table5 .Result so fcrosse sbetwee ndiploi d formso fBrassic a campestris (9)an d Raphanussativu si n197 5an d1976 .

Cross Numbero f Numbero f Numbero f Numbero f Numberrati oo fhybrid s pollinations femaleplant s hybrids matromorphs topollination s(% )

tested yielding average maximum hybrids

1975 Al X Rl 488 18 12 21 2 5 17.1 R3 267 13 4 8 17 3 0 11.9 A2 X Rl 207 8 0 0 27 0 0 R3 205 8 0 0 6 0 0 A3 X Rl 429 17 3 6 25 1 4 7.4 R3 462 17 2 6 48 1 3 8.0 A4 X Rl 249 10 0 0 8 0 0 R3 283 10 2 2 11 0 7 3.3 A5 X Rl 443 18 0 0 12 0 0 R3 400 17 2 3 11 0 8 7.4 A6 X Rl 364 15 2 2 3 0 5 6.3 R3 344 14 3 4 3 1 2 15.0 A7 X Rl 423 18 2 4 7 0 9 10.0 R3 432 17 3 11 21 2 5 36.0 A10 X Rl 44 2 0 0 0 0 0 All X Rl 346 14 3 3 20 0 9 5.6 R3 295 13 0 0 30 0 0 A12 X Rl 102 1 1 4 0 3 9 3.9

1976 A8 x Rl 239 10 0 0 41 0 0 A9 X Rl 105 7 1 1 16 1 0 5.0 A13 X Rl 182 11 10 44 8 24 2 43.8 R2 129 4 3 18 1 14 0 122.2

Total 6438 128 336 2.0 122.2

plant was pollinated with both pollinators on the same day and an almost equal number of pollinations was made for each pollinator. The number ratio of hybrids per pollination was low in both years and ranged from 0 to 24.2 % between crosses. A large variation in hybrid production was also found within the Brassica accessions (Table 5).Usuall y a low pro­ portion of the female plants actually gave rise to hybrids (Table 5).I n 1975, no striking differences were observed between the two pollinators in production of hybrids from crosses with various Brassica accessions (Table 5).Crosse s of A13 gave rise to a substantial number of hybrids; nearly all plants from that accession produced some hybrids in crosses with the accessions Rl and R2. A13 was an S, line obtained by selfing a particular A12 plant, which in 1975 showed a good fruit-set and good pro­ duction of hybrids, i.e. number ratio of hybrids to pollinations (cross- ability) 3.9 %i n crosses with Rl. The results with A13 suggest that se­ lection within B. campestris may give rise to a considerable increase in crossability. If we exclude the results of crosses with A13, the mean crossablility over the two years was about 1 %. The crossability of the

11 Table6 .Result so fcrosse sbetwee nautotetraploi d formso fBrassic a campestris (o)an d Raphanus sativusi n197 5an d1976 .

Cross Numbero f Numbero f Numbero f Numbero f Number ratioo fhybrid s pollinations femaleplant s hybrids matromorphs topollination s(% )

tested yielding average maximum hybrids

1975 Ml x RR2 629 29 0 0 1

1976 AA2x RR1 138 8 0 0 12 0 0 RR2 133 8 0 0 18 0 0 AA3x RR1 73 6 0 0 0 0 0 RR2 128 7 1 1 1 0.8 5.0 AAAx RR1 121 7 0 0 0 0 0 RR2 26 2 0 0 0 0 0 AA5x RR1 39 6 3 3 1 7.7 25.0 RR2 107 10 1 1 0 0.9 9.1 AA6x RR1 455 18 13 39 5 8.6 53.3 RR2 74 6 3 6 1 8.1 30.0 AA7x RR1 158 8 0 0 3 0 0 RR2 52 6 0 0 0 0 0 AA8x RR1 398 18 2 12 2 3.0 57.9 RR2 68 6 0 0 0 0 0

Total 2599 62 44 2.4 57.9

bestA1 3plan twa s evenmor e than 100% (Table5) . Table 6list s crosses between tetraploids.Th emea n crossability was slightly better thanbetwee n diploids;o n average 2.4 %wit h arang e 0-8.6 %. Four tetraploid accessions of B. campestris gaven ohybrid s at all; ofthos e accessions thatproduce d hybrids,usuall y only afe w plants did so,AA 6 being awelcom e exception. Femaleplant s with good crossability were found inAA5 ,AA 8 and especially AA6. Crosses between diploid forms of B. campestris and tetraploid R. sativus gave rise to 5hybrid s from 3132pollination s on about13 0 Brassica plants (Table 7).Th emea ncrossabilit y was only 0.16 %. Incrosse swit h tetraploid formso f B. campestris as femaleparent s anddiploi d forms of R. sativus aspollinator , crossability waspoo rto o (Table 7).I ntotal ,141 1pollination s onsevera l Brassica plants gave rise toon ehybri d only,whic h amounts to acrossabilit y of 0.07 %. Matromorphs were frequently found in2 xx 2 x and 2xx 4 x crosses and less frequently in4 xx 2 x and4 xx 4 xcrosse s (Table 5,6 an d 7). The respective number ratios ofmatromorph s topollination s were 5.2,8.1 , 1.6 and 1.7 %. So diploid forms of B. campestris gave rise tomatro ­ morphsmor e easily than tetraploid forms.

12 Table7 .Result so f2 xx 4 xan d4 xx 2 xcrosse sbetwee nBrassic acampestri s(o )an d 75an d 1976.

Cross Year Number of Number of Numbero f Numbero f pollimition s female plants hybrids matromorphs

tested yielding hybrids

2x X ix Al X RR2 1975 441 17 0 0 52 A2 X RR2 1975 213 8 0 0 12 A3 X RR2 1975 413 17 0 0 44 A4 X RR2 1975 189 8 0 0 3 A5 X RR2 1975 464 18 0 0 12 A6 X RR2 1975 353 15 2 2 3 A7 X RR2 1975 437 18 0 0 21 A8 X RR1 1976 69 4 0 0 15 RR2 1976 99 7 0 0 40 A9 X RR1 1976 31 2 0 0 14 All X RR2 1975 319 14 0 0 30 A13 X RR1 1976 38 4 0 0 1 RR2 1976 66 7 2 3 7

Total 3132 254

4xx 2 x AA1x R 3 1975 727 29 0 AA2x R l 1976 97 7 0 AA3x R 2 1976 161 7 0 3 AA4x R l 1976 58 5 0 2 AA5x R l 1976 9 1 0 0 R2 1976 32 6 0 0 AA6x R l 1976 107 12 ] 0 AA7x R l 1976 10 1 0 0 R2 1976 27 4 0 1 AA8x R l 1976 114 12 0 1 R2 1976 69 6 0 0

Total 1411 22

2.3.4 Use of genetic male-sterility

To avoid time-consuming hand-pollination,us eo fgeneticall ymale - sterileplant s of B. campestris was tested in1977 ,thoug ho na limite d scale.Fou rmale-steril eplant s growni n four isolationcage s together with aflowerin g diploid ortetraploi d accessiono f Raphanus gaveris e toa considerabl e number ofsmal l seeds (Table 8). Unfortunately these cageswer eno tcompletel ypolle nproo f and someundesire d cross-pollina­ tionoccurre d from floweringplant s of B. carinata (BBCC)growin g quite neart oth ecages .Therefor e deviantplant swit h thegenom e constitution ABCwer e recovered aswel l asth eexpecte d intergenerichybrid s andmatro ­ morphs.

13 Table 8. Results of crosses between male-sterile plants of Brassica campestris (2x) and diploid or tetraploid forms of Raphanus sativus (1977). Plants were grown in isolation cages and were polling ted by b ees.

Parent No Poll: nator Number of seeds Number of

obtained sown plants hybri ds matromorphs

true false"'

1 R4 156 15 12 0 10 2 2 R5 2 2 0 0 0 0 3 R6 70 15 9 8 0 1 4 RRl 15 15 5 1 3 1

"]Numbe r of hybrids with the genome constitution ABC obtained by illegitimate pollination.

Table 9. Setting of siliquae, ovules and hybrid and matromorphic embryos in crosses in 1976 between diploid and tetraploid forms of Brassica campestris (AA and AAAA) and Raphanus sativus (RR and RRRR), as well as the developmental stage of the hybrid embryos 3-4 weeks after pollination.

Female Number of Embryo stage of hybrid embryos"* Number ratio parent of hybrid polli- siliquae developed embryos" ? G H LH ET T LT WS M embryos to nations ovules pollinations (%) AA x RR 61 18 24 14 7 11 A9 109 24 37 21 8 7 A13 62 54 7 46 29 74

AA X RRRR 54 10 21 16 2 A9 84 32 64 64 0 A13 25 18 0 2

AAAA x RR AA5 126 50 10 1 5 1 3 4 AA6 164 78 59 2 29 2 10 7 2 3 18 AA7 14 10 1 0 0 0 AA8 106 57 41 0 18 16 9 1 17

AAAA x RRRR AA5 310 137 57 3 29 11115 3 2 15 9 AA6 251 131 187 14 119 2 12 23 19 18 12 5 26 47 AA7 97 94 42 1 20 3 3 2 3 5 2 2 21 AA8 175 81 128 6 61 1 15 25 7 3 4 1 4 35

"]M matromorphic embryo;H hybrid embryo. "] ?unidentified ; G globular; H heart + early heart; LH late heart; ET early torpedo; T torpedo; LT late torpedo; WS walking stick; M mature I+ II (Wilmar & Hellendoorn, 1968).

2.3.5 Culture of embryos

Embryo culture depends firsto nth e availability of sufficient hybrid embryos.Tabl e 9summarize s data from intergeneric crosses made in197 6 forembry o culture.Ther ewer e largedifference sbetwee n femaleparent s innumbe r ratio ofdevelope d ovules topollinations ,an dnumbe r ratioso f matromorphican dhybri d embryos topollinations .Siliqua enearl y always contained one ormor e developed ovules,whic h ingenera lwer erathe r small and oftenmilk y oreve nbrow n and shrivelled. Embryos obtained from such ovuleswer e classified ashybri d embryos.However , largeovule swer e found too, especially incrosse swit hA 8 andA9 .Thes e ovules contained large embryos,whic hwer eclasse d asmatromorphic .A s fara splant swer e raised from theseovules ,th e largeone s gaveexclusivel ymatromorph san dth e small oneshybrids .No t alldevelope d ovulesha d adevelope dembryo . Successo fculturin g embryos invitr o further depends onth e sizean d conditiono fth eembryos .Th ehybri d embryos cultured in197 6 differed verymuc h indevelopmenta l stage (Table 9). Classification ofth e embryos by the scheme ofWilma r& Hellendoor n (1968)wa s usuallypossibl e except fora fewembryo s ofirregula r shape.Th edevelopmenta l stage ranged from globular tomatur e I+ II.Th egrea tmajorit y ofth e embryos had already reachedthei r final stage,a sjudge d fromth econditio no fth e ovules.Thi smean s thatth e developmentusuall y stoppedbefor e embryo maturity. Theresult s ofcultur e ofembryo s from crossesbetwee n B. campestris and R. sativus invitr o are summarized inTabl e 10.I n197 5 thetechni ­ quewa s tested onembryo s from crosses ofA1 0 andA1 2wit h 'Siletta'. The developmental stages ofthos eembryo s ranged from globular toma ­ ture.Th eproportio n ofplant s raised from embryoswa s about3 1% ,in ­ cludinghybrid s andmatromorphs . In1976 ,al lembryo smentione d inTa ­ ble 9wer eculture d invitro .Hybrid s andmatromorph swer eraise d from 2xx 2 xan d4 xx 4 xcrosse s andonl ymatromorph s from 2xx 4 xcrosse s (Table 10).Th ematromorph s were obtained fromcomparativel y largeem ­ bryos.O n average, 11% of allembryo s cultured in197 6gav eris e toa plant.Th enumbe r ratio ofhybrid s topollination swa s unaccountably low incompariso n toth ehybri d productionwithou tembry o culture (Table5 , 6 and 7). Initiallymos tembryo s grew fairlywel l inth e liquidmedium , but afterwardsman y turnedyello w anddied .

2.3.6 Genome constitution and number of chromosomes in the primary hy­ brids

The chromosome number of 140primar yhybrid swa sdetermine d (Table11) . Intergeneric crossesbetwee n diploids gaveonl yplant swit h 19chromoso ­ mes, although several hybridsha droot-ti pcell swit hdouble dnumbe ro f chromosomes.Apparentl y allembryo s from these crossesha d theexpecte d genome constitutionAR . Crosses between tetraploids gaveris e tohybrids ,whos e chromosome number ranged from 35t o40 .Abou t4 8% ha d 38chromosomes ,th eexpecte d

15 Table 10.Summar yo fth eresult so fcultur eo fembryo sobtaine d fromcrosse sbetwee n diploid andtetraploi d formso fBrassic a campestris (AAan dAAAA )an dRaphanu ssativu s [RRan dRRRR) .Th eembryo swer e isolated 21-28day safte rpollination .

Cross Year Numbero f Numbero f Numbero fNumbe ro fNumbe r ratioo f pollina- hybrids matro- hybridst o embryos ovules morphs pollinations(% ) cultured cultured"

AAx R R AIOx R i 1975 69 1 39 0 0 A12x R I 1975 191 132 76 38 19. A8 xR I 1976 61 21 0 0 A9 xR i 1976 109 29 0 0 A13x R i 1976 36 36 7 19. R2 1976 26 17 3 11. Total 492 236 48 9. AAx RRR R A8 xRR l 1976 37 0 RR2 1976 17 10 1 A9 xRR 2 1976 84 63 19 A13x RR l 1976 8 1 0 RR2 1976 17 1 0 Total 163 83 20 AAAAx R R AA5x R i 1976 6 1 R2 1976 120 5 AA6x R i 1976 164 31 AA7x R 2 1976 14 0 AA8x R I 1976 106 18 Total 410 55

AAAAx RRR R AA5x RR l 1976 143 20 0 RR2 1976 167 12 2 AA6x RR l 1976 207 108 4 RR2 1976 44 15 2 AA7x RR l 1976 34 10 0 RR2 1976 63 11 0 AA8x RR l 1976 175 67 2 Total 833 243 10 >]Ovule swit hn ovisibl eembryo .

Table 11.Chromosom enumber san dprobabl egenom econstitution s ofprimar y hybridsfro mcrosse sbetwee ndiploi dan dtetraploi d formso fBrassic a campestris (AAan dAAAA )an dRaphanu s sativus (RRan dRRRR) .

Cross Numbero f Genomeconstitutio nan dchromosom enumbe r hybrids classified AR ARR AARR

19 28 29 35 36 37 38 39 40

AA xR R 85 85 AA XRRR R 6 AAAAx R R 1 1 AAAAx RRR R 48 1 1 12 23 10 1 Total 140 85 1 1 13 25 10 euploidchromosom e number ofplant swit h the genome constitutionAARR . Chromosomenumber sdeviatin g from3 8migh tresul tfro m fusiono faneu - ploid gametes derived from eithero fth etw oparents ,especiall y thefe ­ maleparent . The2 xx 4 xcrosse s resulted intw otetraploi d and fourtriploi dhy ­ bridswit hprobabl e genome constitutions AARR andARR , respectively. One ofth e triploidswa s hyperploid (2n= 29) .Th etw o allotetraploids werepresumabl y derived from a fertilized 2n egg cell. The4 xx 2 xcrosse s gave rise toon ehybri dwit h 37chromosomes .A plausible explanation forthi sunexpecte d chromosome number is fusion of ananeuploi d egg-cell and a 2n malegamete .

2.4 DISCUSSION

Inthi sstudy ,th e crossability betweendiploi d formso f B. campes­ tris and R. sativus was intermediate totha t foundb y otherworker s (Table 2). Themea ncrossabilit y atth etetraploi d level,however ,wa s better thani npreviou s reports.Th eresult s of2 xx 4 xan d4 xx 2 x crosseswer epoo r incompariso nt o thoseo fcrosse s atth ediploi d and tetraploid level.S other e isapparentl y aneffec to fploid y oncross - ability inthi s intergeneric cross.Suc heffect shav ebee ndescribe d in various interspecific and intraspecific crosses amongCrucifera e (e.g. Howard, 1942b). Inth ecrosse s studied,th e optimum ratio ofth egenom e number inembry o andendosper m isapparentl y 2t o3 .Th eoccurrenc e of anunexpecte d highproportio n ofallotetraploid s among thehybrid s from 2xx 4 xan d4 xx 2 x supports thisview . Theviabl e seed from intergeneric crosses consisted ofhybri d andma - tromorphic ones.Usuall y small seeds gave rise tohybrids ,a swa sear ­ lierreporte db yNish i et al. (1964)an dTokumas u (1970b).Plum p seeds resulted inmatromorphs . Manymor ematromorph s wereobtaine dpe rpollinatio n incrosse s ondi ­ ploid formso f B. campestris thano ntetraploids .Althoug h illegitimate self-pollination and cross-pollination cannotb eprecluded , thediploi d forms seemed tohav e abette rparthenogeneti c ability of 2n eggs thanth e tetraploids afterpollinatio nwit h Raphanus. Clear-cut differencesbetwee n andwithi n B. campestris accessionsex ­ isted incrossabilit y with Raphanus. Only asmal lproportio n ofth efe ­ maleplant suse d inthi s studyproduce d hybrids.Variatio n incrossabil ­ itycoul db ecause db yvariatio n inenvironmenta l conditions.Howeve ra s the femaleplant swer e grown and treated ina simila rway ,ther emus tb e genotypic variation incrossability . The finding thatnearl y allplant s from anS . line (A13), whichha dbee nobtaine db y selfing aplan twit h good crossability, alsoha d agoo d crossability, supports thisvie w and proves thatselectio n among the femaleparent s for improved crossability

17 couldb e effective.Th e effecto fpollinato r genotype oncrossability , was lesspronounced . Possible genotypic variationwithi n themal eacces ­ sions,however ,ma yhav ebee nmaske d by theus e ofbulke dpollen . Theproductio n of largenumber s ofhybrid s istime-consuming . Effi­ ciency couldb e improvedb y theus e ofgeneticall y male-sterile plants of B. campestris. To facilitate large-scale production ofhybrid s it mightb ewort h trying tocombin emale-sterilit y with agoo d genetically determined crossability. Another interesting approach isembry o culture.Developmen t ofhybri d embryosmostl y stopsbefor e embryomaturity . The results ofembry o cul­ ture inthi s study,however ,wer e disappointing, although in 1975 itcon ­ siderably increased thenumbe r ofhybrid s raisedpe rpollination . 3 Nature of breeding barriers

3.1 INTRODUCTION

Only afe wbrie f reports existo nbreedin gbarrier s inth eprogami c phase ofcrosse sbetwee n B. campestris and R. sativus. Oelke (1957)ob ­ served thatpolle n of several unnamed cruciferous specieswa s inhibited onradis h stigmas.On e ofthos e specieswa s B. campestris spp. rapifera (Sampson, 1962). Itspolle n germinated butdi dno tpenetrat e the radish stigma.Reciproca l crossesbetwee n the two specieswer e studied atth e ScottishPlan tBreedin g Stationb y fluorescence microscopy. 'Pollenger ­ mination and growthwer e fairly goodbu tver y fewovule s starteddevel ­ opment' (ScottishPlan tBreedin g Station, 1971,p .28) . Self-incomptability in B. campestris and R. sativus isgoverne db yth e sporophytic actiono fon eS-alleli cserie s (reviewedb yHinat a &Nishio , 1980). Incontras t toLewi s &Crow e (1958), Sampson (1962)observe d that mostcrosse sbetwee n self-incompatible species inCrucifera ewer e incom­ patible, except inth ebu d stage of flowers.Oelk e (1957)ha d also found thatbu dpollinatio n had abeneficia l effect incrosse sbetwee n kohlrabi (B. oleracea var. gonguloides L.) and radish,usin g the former aspis ­ tillateparent .Thu sbu dpollinatio n seemed anattractiv emetho d for overcoming interspecific breedingbarriers . Thepresen t studywa s intended toclarif y thenatur e ofbreedin gbar ­ riers during theprogami cphas e of fertilization incrosse sbetwee ndi ­ ploid and tetraploid formso f B. campestris (9)an d R. sativus andt o testth e significance ofbud-pollination .

3.2 MATERIAL ANDMETHOD S

Parto fth e 1976 crossingprogramm e was anexperimen to nbreedin gbar ­ riers. Plantmaterial ,growin g conditions andcrossin gmethod swer ea s described inChapte r 2, flowersbein gpollinate d 2-3 days afteremascu ­ lationwhe n some flowerbud s had reached theopen-flowe r stage andth e restwer e still inth ebu d stage.Th e oldestbu d ofeac h inflorescence wasmarke d and three days later the thirdpisti lbelo w and aboveth e marked pistilwer e fixed inCarno y solution (6volume s of aqueousetha ­ nol (volume fraction 0.96), 3volume s ofchlorofor m and 1volum e ofgla ­ cial acetic acid)an d stored ina refrigerator . The resto fth epistil s ofeac h inflorescence were lefto nth eplant s anduse d forembry o cul-

19 ture afterwards.Th e results ofthes e cultures are reported inTabl e10 . The fixedpistil swer e subsequently softened inNaO H solution ofsub ­ stance concentration 1mol/ 1 for 1h at6 0 °Can d stained ina naqueou s solutionwit hmas s concentration of anilineblu e 2g/l andK-PO .2 0 g/1. Thepistil swer e gently squashed ina dro p ofglycero l and studied under a fluorescencemicroscope . Inal lpistils ,th eproportion s ofgerminate d pollen and thepropor ­ tions ofpolle n grainswit h apolle n tube thatha dpenetrate d the stigma were counted among arando m group of about 50polle n grains per stigma. Although usually over 1000polle n grains were applied oneac hstigma , mostly afe whundre d grainswer e stillpresen t and sometimes evenles s than 50.Th e reduction innumbe r ofpolle n grainswa sprobably , atleas t inpart ,cause d by the loss ofungerminate d pollen during fixation,ma ­ ceration and staining. Inal lpistils ,th epolle n tubeswer e counted in the transitional areabetwee n style and ovary and the ovules fertilized, i.e. theovule swit h apolle ntub e inthei rmicropyle ,wer ecounte d for eachovary .

3.3 RESULTS

Intergeneric incompatibility was studied in 2x x2x ,2 xx 4x ,4 xx 2 x and4 xx 4 xcrosse s between B. campestris and R. sativus. Inal lcrosses , typical incompatibility reactions occurred onth e stigmas similar toin ­ compatibility reactions after self-pollination. Usually more than 80% o f thepolle n stillpresen t ona stigm a hadgerminated . Germinationwa s somewhatbette r on stigmaspollinate d inth eopen-flowe r stage thani n thebu d stage.Th epolle n tubeswer emostl y shortan d strongly fluores­ cent,becaus e ofcallos e formation (Fig.3A) .Callos e deposits wereusu ­ ally alsoobserve d inman y stigmapapillae .Suc hreaction s apparently prevented thepenetratio n ofpolle n tubes into the style.Howeve r acer ­ taindegre e ofpenetratio n was usually observed afterbu d and open-flower pollination. Some abnormalities occasionally occurred within stigma papillae, such asbranchin g ofpolle n tubes and inhibition ofpolle n tubegrowt h (Fig.3B) . Nobarrie r seemed to exist inth e style,an don e ormor epolle n tubes entered theovar y innearl y allpistils ,irrespectiv e ofth emod e ofpol ­ lination (Table 12).A hig hproportio n ofpistil s inal l crosses hadmor e thante npolle ntube spe r style.Bu dpollinatio n didno t increase thenum ­ ber ofpolle n tubespe r style.Th e results for thevariou s crosses indi­ cate thatther ewa s nomarke d effecto fploid y onpenetratio n ofpolle n tubes into thestyle . Inth e ovaries,mos tpolle n tubesusuall y grew fairly straight toth e bottom,wher e growthbecam e erratic.Usuall y only afe wpolle n tubes grew inth e direction of anovul e andpenetrate d through themicropyl e (Table13 ]

20 Figure3 .Incompatibilit y reactionsi nth eprogami cphas eo fcrosse sbetwee nBrassic a campestrisan dRaphanu ssativu s (ultravioletfluorescence) . A.Inhibitio no fgrowt ho fpolle ntub eo na stigm apapill a (x490). B.Inhibitio no fgrowt ho fpolle ntub ewithi na stigm apapill a (X490).

Table 12.Numbe ro fpolle ntube senterin gth eovar yafte rbu dan dopen-flowe r pollinationi ncrosse sbetwee ndiploi d andtetraploi d formso fBrassic a campestris (AAan dAAAA )an dRaphanu ssativu s (RRan dRRRR) .Th edat aar ebase do nobservation s inth esam epistil sa si nTabl e13 .

Cross Modeo f Numbero f Numbero fpistil s pollination pistils withn o with 1-9 with£1 0 tubes tubes tubes

AA XR R Bud 13 1 5 7 Flower 13 1 6 6

AA XRRR R Bud 12' 0 3 7 Flower 13 1 4 7

AAAAx R R Bud 18 1 5 13 Flower 17 1 5 11

AAAAX RRR R Bud 44 4 12 28 Flower 44 2 9 33

*]Dat afro mtw opistil smissing .

21 Table 13.Numbe ro ffertilize d ovulesan dth enumbe ro fpistil s inwhic hfertilizatio n wasobserve d incrosse sbetwee ndiploi dan dtetraploi d formso fBrassic a campestris (AAan dAAAA )an dRaphanu s sativus (RRan d RRRR).

Cross Numbero f Numbero f Fertilized Pistilswit hovule s pistils ovules ovules fertilized* * observed observed" number proportion(%) number proportion(%)

AA xR R A8 X Rl 9 203 11 5.4 3 33 A9 X Rl 17 362 17 4.7 8 47

AA X RRRR A8 X RR1/RR2 12 285 9 3.2 5 42 A9 X RR1 13 278 12 4.3 8 62

AAAA X RR AA6 X Rl 19 424 14 3.3 10 53 AA8 X Rl 16 276 17 6.2 7 44

AAAA X RRRR AA4 X RR1 9 141 2 1.4 2 22 AA5 X RR1 5 108 1 0.9 1 20 AA6 X RR1 27 637 30 4.7 15 56 AA6 X RR2 6 130 5 3.8 3 50 AA7 X RR1/RR2 12 248 23 9.3 9 75 AA8 X RR1 29 512 17 3.3 11 38

"]Onl yeasil yvisibl eovule si neac hovar ywer eassessed . '"'"]Fertilizatio n isunderestimate d becausesom eovule swer eno tvisible .

Occasionally erratic growth ofpolle n tubeswa s also observed near theen ­ trance toth emicropyle .Th eproportio n ofovule s fertilized was low, 0.9 to 9.3 %. Inman ypistils ,n o fertilization was observed, and thenumbe r of fertilized ovules perpisti l averaged less thantw o inth ebes tcros s (Table13) .

3.4 DISCUSSION

Inth eprogami c phase ofth e intergeneric crosses studied, therewer e obviousbarrier s tobreeding :poo rpenetratio n ofpolle n tubes into the stigmas anderrati c growth ofth epolle n tubewithi n theovaries .Th e firstbarrie r was not fully effective and inmos t crosses aconsiderabl e number ofpolle n tubes reached theovaries .Thes e observations agree quitewel l with those atth e Scottish PlantBreedin g Station (1971). The poor development of ovules observed inScotlan d and inthi s study (Ta­ ble 9)seem s tob e causedprimaril y by scanty fertilization. Apparently the ovules failed to attractpolle n tubes.Suc h abreedin g barrierwa s also observed ininterspecifi c crosses in Arabidopsis (Berger, 1968).A ratherhig hproportio n of fertilized ovules apparently gave rise tode ­ veloped ovules inwhic h anembry o frequently was present (Table 9).Se ­ lectionwithi n B. campestris for improved crossability mightb e achieved

22 by selection forgenotype swit h arelativel y good fertilization asde ­ tectedb y ultravioletmicroscop y or forgenotype swit h ahig hproportio n ofdevelope d ovules orembryo spe rpollination . Inth eprogami c phase,ther ewer e no striking differencesbetwee n2 x x 2x,2 xx 4x ,4 xx 2 x and4 xx 4 xcrosse s forth enumbe r ofpolle ntu ­ bes observed inth e style and success of fertilization. The effecto f ploidy oncrossabilit y (Section2.4 ) apparently came after fertilization andno tbefor eit . Budpollinatio nwa s certainly not aneffectiv emetho d ofcircumven ­ ting thebreedin gbarrier s onth e stigma,a swoul db e expected fromth e results of Sampson (1962),becaus e therewa sn omarke d differencebe ­ tweenbu d and open-flower pollination indegre e ofth e incompatibility reactions onth e stigma andpenetratio n ofpolle n tubes into thestigma . A role ofth eS-locu si ninterspecifi c incompatibility orincongru ­ ity asdefine db yHogenboo m (1973;1975 )wa s assumed, for instanceb y Sampson (1962), butshoul db equeried ,becaus e theactivitie s oftha t locus are found instigma s ofope n flowers andolde rbuds . Perhaps the buds studiedwer e too old and theS-locu swa s already operative.Howeve r the results ofNama i (1980),wh o recently studied theeffec to f ageo f buds incrosse sbetwee n B. campestris and R. sativus, indicate that crossability ofyounge r buds iseve n less.

23 4 Primary allodiploid AR hybrids and their progeny

4.1 INTRODUCTION

Several authors (Table 2)hav e obtained AR hybrids from crossesbe ­ tweendiploi d forms of B. campestris and R. sativus. Terasawa &Shimo - tomai (1928), Tokumasu (1970b)an dOpen a& L o (1978)hav edescribe d such hybrids from crossesbetwee n orientalvegetabl e forms of Brassica and common radisho rJapanes e radish.Th eplant swer e intermediate between theparenta l species inman y respects with alea f rosette inth evege ­ tative stage andhardl y show any thickening ofth eroots ,i f atall .Th e white andpurpl e flower of R. sativus proved tob e dominant overth e yellow of Brassica. Clauss (1978), however, obtained onlyyellow-flower ­ edhybrid s from similar crosses.Th e siliquae ofth ehybrid s consisted oftw opart s asi n Brassica, i.e. avalvat e part and arathe r large in- dehiscentpart ,th ebeak .Bot hpart s contained ovules (Terasawa& Shi - motomai, 1928;Tokumasu , 1970b). Tokumasu (1970b)observe d also thatth e hybrids grewvigorousl y inth e later stages ofdevelopment . Various authors studied meiotic chromosomebehaviou r in suchhybrid s (Terasawa& Shimotomai, 1928;Richharia , 1937;U et al., 1937;Mizushi - ma, 1944b; 1950a;Tokumasu , 1970b). The datapresente d varyconsidera ­ bly.Accordin g toU et al. (1937), Terasawa found amaximu m of fivebi ­ valents per cell. Inthei r ownhybrids ,U e t al. (1937)counte d nomor e than threebivalent s andRichhari a (1937)eve n found amaximu m ofeigh t bivalents per cell. Tokumasu (1970b)observe d sixbivalent s per cell at most and also found alo w frequency oftrivalent s andquadrivalents .Th e causes ofth edifference s indegre e ofchromosom e association areun ­ known,bu tmigh tb e inpar tgenetic . Another intriguing aspect ofallodiploi d hybrids isthei r fertility. Morris (1936)an dTokumas u (1970b)studie d male fertility. Morris found that allhybrid s studied produced aconsiderabl e amount ofdiploi dger - minablepollen .Polle n stainability varied considerably from day toda y andprobabl y fromplan t toplant .Tokumas u (1970b)als o noticed arathe r highproportio n of stainable pollen,varyin g from 4t o3 0% wit h amea n of 11% .Howeve r seed fertility ofth ehybrids ,a sdetermine d byinter - mating,wa sver ypoor .Terasaw a & Shimotomai (1928)obtaine d someF _ plants,whic h differed in form and flower colour.Als oMorri s (1936)an d Mizushima (1952)obtaine d F_plants .Howeve r Tokumasu (1970b)di dno t succeed inobtainin g aprogen y fromF .hybrids .

24 Thecrossabilit y ofth ediploi d hybridswit h thetw oparenta l species, B. campestris and R. sativus was equallypoo r (Terasawa& Shimotomai,1928 ; Tokumasu, 1970b). Only Terasawa & Shimotomai (1928)obtaine d afe wcom ­ pletely sterileplant s frombackcrossin g F,hybrid s as femaleswit hth e Brassica parent.Backcrosse s with R. sativus failed. Thischapte rpresent s observations onvariou s aspects ofth ediploi d Fi hybrids obtained inthi s study, such as form,meioti c association of chromosomes,polle n and seed fertility.Als o reported arechromosom e num­

ber and fertility ofF 2plants ,attempt s todoubl e the chromosome number ofF .hybrid s and crosses ofF .hybrid swit h B. campestris, R. sativus and B. napus respectively.

4.2 MATERIALAN DMETHOD S

Allhybrid s studiedwer e obtained from the crossesbetwee n thedip ­ loid accessions of B. campestris and R. sativus. Allhybri dplants ,a s far aswa s known,ha d 19chromosome s andth egenom e constitutionwa s probablyAR .Th eplant swil lb e referred to asA R hybrids or rapifera hybrids, for instance,referrin g toth e Brassica parent inth e original intergeneric crosses.Th ehybrid swer e grown inth e samewa y asthei r parents, and under similar conditions,bu tmos twer etransplante d into soilo f aheate d greenhouse justbefor e flowering.Non-boltin g plants were firstplace d ina col d greenhouse for about4 0day s inorde rt o induce flowering. Pollen stainability was assessed on freshpolle n collected fromva ­ rious flowersb y ametho d ofSas s (1964)wit h arando m sampleo f about 200polle n grainspe rplant .Som e slideswer e alsouse d formeasurin g pollen grainsize .Th emethod s used forstudyin gmitosi s andmeiosi s havebee ndescribe d in Section2.2 .Al l data onth econstitutio n of sporad cellswer eobtaine d from anthers squashed ina dro p ofpropiono - iron-alum haematoxylin (Henderson& Lu , 1968). SeveralA Rhybrid swer evegetativel y propagated from cuttings,cul ­ ture invitr o of stem andpetiol e explants (Kartha et al., 1974)o ro f axillary buds. Propagationb y cuttings,whic hwer etreate d with atalcu m powder containing indolebutyric acid (IBA)a ta substanc e contento f 5g/kg, gaveusuall y rise towea k and smallplants .Th e twomethod s of culture invitr o resulted inwel l developedplants .Th e overall results of propagation through stem andpetiol e expiants,however ,wer epoor ; some shoot formation occurred on acultur emediu m ofMurashig e & Skoog (1962)containin g 6-benzylaminopurin e (BA)an d 1-naphthaleneacetic acid (NAA)a tsubstanc e concentrations of 10an d 0,5 mmol/m3, respectively. Culture conditions andmediu m for invitr o culture of axillarybud swer e the same asi nKarth a et al. (1974). Themetho d gave relatively goodre ­ sults ona mediu m containing BA at5 mmol/m 3 andNA A at1 mmol/m 3.

25 Cuttings ofvariou s ARhybrid s were transplanted inisolatio n cages (2.5m x2. 5 m)place d outside and intercrossed with thehel p ofhoney ­ bees toobtai n anF generation. Subsequently, theF 2plant s weremulti ­ plied ina simila rwa y incage s ina greenhouse . All axillary buds ofroote d cuttings of several AR-hybrids were treated with colchicine in four substance concentrations ranging from2 to 6g/1 . The colchicine treatmentswer e carried out according toth e traganth-slimemetho d of Schwanitz (1949)a tabou t 18 °Can dwit hhig h airhumidity .

4.3 RESULTS

4.3.1 Form of AR hybrids

Intergeneric crossesbetwee n B. campestris and R. sativus gaveris e tohybrid s andmatromorphs .Th e distinctionbetwee n the two types proved tob e easy andwa susuall y possible ina nearl y developmental stage.Fo r example,hybrid s usually had awhitis h hypocotyl andthei r firstlea f waspubescent . Incontrast ,matromorph s fromcrosse sbetwee n B. campes­ tris ssp. chinensis and fodder radish, for instance,ha d plain greenan d more robusthypocotyl s andglabrou s leaves.Th e hybrids subsequentlyde ­ veloped aleaf-rosett e with several slightly pubescent leaves (Fig.4A) . The shape ofth e leaveswa s intermediate,bu tver yvariable ,rangin g from simple topinnate .Hybrid s usually had slightly pubescentpetioles ,whic h were sometimes slightly red from anthocyanins. Some hybrids from crosses with turnip showed aswolle n turnip-like root.O n thewhole , thehybrid s proved tob evigorous ,althoug h afe wwer e stunted or even deformed. The transition fromvegetativ e togenerativ e stage did notrequir e cold treatment,excep t for somehybrid s originating from crosses with stubble turnip, afor m of B. campestris ssp. rapifera thatneed s cold for flower induction.Th e flowering time ofA Rhybrid s was intermediate between that ofth eparenta l species. Inth egenerativ e stage,hybrid s usually had anerec tmai n stem witha terminal inflorescence (Fig.4B) .Afterward s many lateralbranche s with secondary inflorescences developed andth eplant s reached aheigh to f 1.3-1.7m .Th e indeterminate flowering observed inA R hybrids,probabl y causedb y ahig hdegre e ofsterility , resulted finally in awil d dense bunch ofbranches .O nth e stem and lateral branches, asligh tpubescenc e and some anthocyanin colorationwa s occasionallypresent . The inflorescences ofA R hybridswer emor e similar to B. campestris thant o R. sativus innumbe r of flowers, flower size and thepositio n of flowers and flowerbud s (Fig. 4C). Thebud s were usually slightly pubescent.Mos tA R hybrids hadwhit e orpurpl e flowers with clear-cut veins, like Raphanus, but someplant s hadplai nwhit e flowers oreve n

26 Figure 4. Hybrids with the genome constitution AR from the cross Brassica campestris ssp. chinensis x Raphanus sativus cv. Siletta. A. Vegetative stage. B. Generative stage. C. Inflorescence.

27 yellow flowers,lik e Brassica. Intotal ,abou t 11% of all floweringA R hybridswer e yellow-flowered. The flowers mostly had strikingveins ,bu t a fewha dplai nyello w flowers. Ina fe whybrid s withwhit e orpurpl e flowers, sectorial chimaeras for flower colourwer e observed;usuall ya small sector ofa peta lbein gyellow . SomeA R hybrids showed some setting of smallbivalve d siliquaewit ha largebeak .

4.3.2 Chromosome association at meiosis

Earlymeioti c stages inA R hybridswer e quite regularbu t indiaki - nesis andmetaphas e I,frequen tunivalen t formation disturbed meiosis (Table 14). Datapresente d inth eTabl e 14shoul db e interpreted with caution.Analysi s ofchromosom e association insuc hhybrid s wasdifficult , because thehig h frequency ofunivalent s sometimes hindered recognition ofmeioti c stages (metaphase Ian d anaphase I, for example). Itals opro ­ ved difficult todistinguis hbetwee n truebivalent s and secondary asso­ ciations ofunivalents . No chromosome associationwa s observed in 67% o f allpolle nmothe r cells studied in sixA Rhybrids .Cell swit h oneo r twobivalent s occur­ red in2 1% and8 % ofth epolle nmothe r cells,respectively , andtri - valents in2. 6 %o f thecells .Th emaximu m number ofboun d chromosomes waseigh t ina cel lwit h fourbivalents .Mos tbivalent s were rod-shaped (Fig. 5A,B) ,an d all trivalents were chains.Th e differences inchromo -

Table 14.Distribution s of meiotic configurations in diakinesis or metaphase I in six AR hybrids from crosses between Brassica campestris ssp. chinensis and Raphanus sativus cv. Siletta.

Configurations" Number fract Lons(% ) of named hyb rids Number of cells

EC119 EC186 EC209 EC274 EC314 75 1040-9 Observed Expected**

19 I 78.4 63.9 62.0 75.5 65.7 62 8 361 336 17 I+ 1I I 15.7 26.4 24.0 13.8 21.2 24 4 115 158 16 I+ 1 III 1.4 1.1 2.0 1 7 7> 37 15 I+ 2 II 5.9 5.6 12.0 6.4 10.1 8 7 44! 14 I+ 1 II + 1 III 2.0 1.1 1 7 6 13 I+ 3 II 1.4 1.1 1.0 l> 12 I+ 2 II + 1 III 1.4 0 6 1 11 I+ 4 II 1.1 ?>

Total number of cell

51 72 50 94 99 172 538

'']I , II and III are symbols for univalents, bivalents and trivalents, respectively. c]Expectatio n based on a Poisson distribution for a mean number of 0.47 bivalents per cell.

28 B

%* • • • *fm 1 • • • • g# f *

•I « ê $

Figure5 .Chromosom eassociation ,dya dformatio nan dstainabl epolle ni nA Rhybrid s (2n= 19) . A.Polle nmothe rcel la tmetaphas eI wit h1 5I + 2 I I (X1540). B.Polle nmothe r cella tmetaphas e Iwit h1 7I + 1 I I (X1540). C.Spora dstag ewit hdyad san don etetra d (x550). D.Stainabl epolle n(X559 ) Table 15.Frequenc y distributiono fpolle nstainabilit y inA Rhybrid sfro m crossesbetwee nBrassic a campestrisan dRaphanu ssativus .Dat a from1975 .

Pollenstainabilit y(% ) Numbero fplant s Number fractiono fplant s(% )

0 53 62 1-10 16 19 11-20 8 9 21-30 6 7 31-40 2 2 Total 85

some associationbetwee n theA R hybridswer e onlysmall . Themea nnumbe r ofbivalent s percel l inth e sixA R hybridswa s 0.47 (one trivalent calculated astw obivalents) . The frequency distribution ofth enumbe r ofbivalent s observedpe r cell deviates significantly from a Poisson distribution (P< 0.01 ) for JJ= 0.47,becaus e of aneviden t excess ofcell swit h exclusively univalents andcell swit h one trivalent ortw obivalents .Chromosom e association isapparentl y not simply amat ­ ter ofchance ;som e chromosomes probably pairmor e easily thanothers .

4.3.3 Male and seed fertility

Themajorit y ofA R hybridswer e completely male-sterile,bu t 38% o f theplant s studied produced some stainable pollen grains (Table 15,Fig . 5D). AllA Rhybrid s from crosses withmale-steril e turnips, forin ­ stance,wer epartl y male-fertile. The density function ofproportio n of pollen grains stained showed that twoplant s produced over 30% staina ­ blepollen .Stainabl e pollenwa s usually ovoid-spherical and had adia ­ meter of35. 2 \im(mea n for 5A R hybrids). Cytokinesis inA Rhybrid swa s irregular, resulting invariou s types ofspora d :beside s normal tetrads,monads ,dyads ,triads ,pentad s and hexads (Table 16).Th e size ofth ecell s ina spora d usually variedcon ­ siderably, but dyads usually contained two cells ofequa l size (Fig. 5C). Relatively high frequencies ofdyad s occurred insevera l chinensis hy­ brids and innearl y all rapifera hybrids.Mos to fthes eplant s alsoha d a relatively highproportio n of stainablepollen . Sodya d formation is prerequisite butno t aguarante e forpolle n stainability. One rapifera hybrid, 75.1283-1,als oproduce d andexceptionall y high number of monads (47ou to f98) . Thehig hproportio n ofdyad s inman yA R hybrids canb e explained by the occurrence ofrestitutio n nuclei atinterkinesis , assuc hnucle i were observed mainly inplant swit h ahig h dyadproductio n or ahig h pollen stainability. Thewel l filled and stainablepolle n ofA R hybridswer egerminable .

30 Table 16.Distributio no fsporad swit hdifferen tnumbe ro fcell spe rspora d insom e ARhybrid soriginatin g fromcrosse sbetwee nBrassic a campestrisssp .chinensi so rssp . rapiferaan dRaphanu s sativus,cv .Silett ao rvar .mougri .

Plants Total Numbero fsporad s Numberfract - Pollenstainab - number iono fdyad s ility(% ) ofsporad s (%)

chinensis hybrids EC119 87 0 15 10 52 10 17 0 EC138 58 0 1 6 48 1 2 2 1 EC186 129 2 32 8 84 2 1 25 6 EC189 74 2 58 7 7 0 0 78 24 EC200 65 2 12 0 48 3 0 18 0 EC263 137 0 43 21 71 2 0 31 7 EC274 63 0 0 8 44 11 0 0 0 EC277 69 0 4 5 52 7 1 6 0 EC282 68 1 25 2 40 0 0 37 32 EC305 57 0 20 7 30 0 0 35 11 EC314 55 3 23 2 27 0 0 42 17 75.1040-9 72 0 5 6 60 1 0 7 0 75.1088-1 94 2 6 2 81 3 0 6 0

rapiferahybrid s 75.1283-1 98 47 49 1 1 0 0 50 75.1308-2 66 0 34 9 23 0 0 52 75.1310-2 72 1 37 11 22 1 0 51 6 75.1427-3" 79 0 51 8 20 0 0 65 20 75.1671-1 56 0 1 1 53 1 0 2 0 75.1682-5-' 62 0 41 1 20 0 0 66 25

Total 1461 6045 7 115 783 42 31

"] Themarke dA Rhybrid soriginate d fromR . sativusvar .mougri .

In crosses of AR hybrids as male parents with B. campestris ssp. chinen­ sis and B. napus, for example, germinated pollen could be demonstrated by ultraviolet microscopy. Two days after pollination, pollen tubes were often observed in the style and some of them had even reached the micro- pyle of an ovule. Seed fertility of AR hybrids was poor. Many attempts were made to pro­ pagate such hybrids by intercrossing, even with culture of immature em­ bryos in vitro (Table 17).Th e considerable efforts with intercrossing of AR hybrids resulted in only three plants. However, propagation in isola­ tion cages of some AR hybrids or cuttings of them still gave rise to sev­ eral more seeds (Table 18).I n total, 36 % of the hybrids produced seeds in this way.

4.3.4 Crossability of AR hybrids with various species

The results of crosses between AR hybrids and B. campestris, B. napus and R. sativus, respectively were poor (Table 17).Fro m crosses of AR hybrids (2n = 19) with B. campestris (2n =20) as pollinator, only one hybrid was obtained with 47 chromosomes. The reciprocal cross gave only

31 Table 17.Number s of seeds and embryos obtained by intercrossing AR hybrids and by crossing AR hybrids with Brassica campestris (AA),Raphanu s sativus (RR),Brassic a napus (AACC) and AARR hybrids. All embryos were cultured invitro .

Cross Number of Number of Number of Number of plants pollinat­ ions seeds plants embryos plants hybrids matromorphs obtained cultured obtained

AR X AR ca 2500 5 2 11 1

AR x AA ca 350 1 1 2 0 1 0 AR X AACC 120 4 2 3 1 3 0

AA X AR ca 900 13 4 26 4 0 8 RR x AR 66 0 0 1 0 0 0 AACC X AR 811 105 84 15 4 1 87

AARR X AR ca 1000 12 6 2 0 6 0 AR X AARR ca 250 3 0 0 0 0 0

matromorphs, some ofwhic hwer e raisedb y embryoculture . Pollination ofA Rhybrid sb y B. napus gavethre ever y similarhybrids , all derived from oneparticula r rapifera hybrid.Th eplant s had aste m like thato f forage rape and leaves andpetiole swer epubescent . Inth e generative stage,thes eplant s produced only afe w flowers,whic hwer e completelywhite .Unfortunatel y thechromosome s ofthes ehybrid swer e notcounted . The reciprocal crosses gave animpressiv e number ofplants , partly raisedb y embryo culture.Onl y oneplan twa s atru ehybri d with 2n =55 ;th eother swer ematromorphs .Th ehybri dwa s obtained froma small seed;i tgre w irregularly, hadwhit e flowers and evenproduce d stainablepollen . Crosses of fodder radish (9)an dA R hybridswer emad e ona smal lscale , resulting inonl y one embryo (Table 17).Cultur e of this embryo invitr o failed.

Table 18.See d production in the F1 and F generation of two populations of AR hybrids. The F and F„ plants 1fro m each population were propagated in t P cages and pollinated by Dees.

Origin Gene­ Numb er of plants Number of seeds population ration test ed produci ng seeds

rapifera hybrids 21 9 45 24 10 53 chinensis hybrids 15* 4 23 Fi 15 10 441 F2 *] 23 cuttings

32 4.3.5 Somatic and meiotic doubling of chromosome number

Variousmethod swer euse dt odoubl eth echromosom e numbero fA Rhy ­ brids.Th efirs tmethod , colchicine treatmento faxillar ybud so f9 0 rootedcutting s from2 8A Rhybrids ,resulte di ndoublin gi n3 2cuttings . Recognitiono fsuc hdouble d sectorsprove d rathereasy ,becaus e flowers were largeran dpolle nwa s usuallymor e stainable.A treatmen twit h col­ chicine solutiono fmas s concentration4 g/ 1(i n traganth slime)gav e thebes tresults ,bu tdifference s fromothe r colchicine concentrations weresmall . Another approachfo rdoublin g the chromosome numberso fA Rhybrid s wast ointercros sA Ran dAAR R hybrids.Doublin gwa sexpecte di nthi s wayb ymeioti c nuclear restitutioni nth eA Rhybrids .Tabl e1 7show s thatn oplant swer eobtaine di fA Rhybrid swer euse da sfemal eparent . Thereciproca l cross gaveris et osi x vigoroushybrids .Tw oo fthes e plantsha d3 8chromosome san don eplan tha d36 ;th echromosome so fth e otherswer eno t counted. Chromosome doubling occurredb ychanc ei na ste mexpian to fon eA R hybrid culturedi nvitro .

4.3.6 Progeny of crosses between AR hybrids

Seeds from crossesbetwee nA Rhybrid swer eproduce db yhand-pollina ­ tion (Table 17)o rinsec tpollinatio n (Table 18).O fth e seeds,abou t 64% germinate dan dmos to fthos egav eris et oviabl eplants .Th e off­ springo fsuc hcrosse svarie d widelyi nchromosom e number (Table19) . However,a remarkabl y highproportio no fth e plantsha dabou t5 7chromo ­ somes, especiallyi fobtaine d from crossesbetwee n chinensis hybrids. Suchhexaploid sprobabl yha dth egenom e constitutionAAARRR .Th e remai­ ningplant smostl yha daneuploi d chromosome numbersan dwer e genetically lessbalance d thanth ehexaploi dplants . Theprogen yo fcrosse sbetwee nA Rhybrid s alsovarie dwidel yi nform , not surprisinglyi nvie wo fth e variationi nchromosom e number.Som e

Table 19.Somati c chromosomenumber si nprogen yo fA Rhybrid sproduce db yintercrossin g chinensishybrid san drapifer ahybrid s respectively.

Population Chromosomenumbe r Numbero f plants 29 31 35 36 37 38 39 40 46 52 54 55 56 57 58 classi­ fied rapifera hybrids 21121231 1 51 20 chinensis hybrids 2 1 1 1 2 6 13 Total 41 121231 1 1 1 12 11133

33 Figure6 .Tw ohexaploi dplant swit hprobabl egenom e constitutionAAARR R obtainedb y intercrossingA Rhybrid sderive d from crossesbetwee n stubble turnip (2x)an dRaphanu s sativus(2x) .

plants showed serious growth irregularities, but others grew vigorously, especially the hexaploids (Fig. 6).Th e hexaploid plants usually had plump flower buds, large white or purple flowers, and produced some stainable pollen. For propagation, most offspring of rapifera and chinensis hybrids was transplanted into three isolation cages with bees for pollination (Table 18). Over 50 % of the offspring produced some seed. The seed production per plant was 2.2 and 29.4 in the offspring of rapifera and chinensis hybrids, respectively. Most of the seed-producing plants were hexaploids with about 57 chromosomes.

4.4 DISCUSSION

The form of AR hybrids obtained by other authors (Table 2) and of the ones in the present study were similar. Some of the present hybrids, however, had yellow flowers, as reported by Clauss (1978). In hybrids with white and purple flowers, the synthesis of flavonoids was apparently suppressed, as in R. sativus (Bonnet, 1978). This suppressor system is perhaps inactivated in the yellow-flowered hybrids by one or more genes

34 from theA genome. Interactionbetwee n geneso fth e twogener a alsooccur ­ red inpartiall ymale-fertil e hybrids,whic h originated fromcrosse sbe ­ tweenmale-steril e turnips and R. sativus. Pollen sterility inturni p proved tob e inherited ina simpl emonofactoria l and recessivewa y (Dolstra,unpublished) . Sooccurrenc e ofpolle n fertility inA Rhybrid s ofmale-steril e turnips indicated thatth eeffec to fth e gene forpolle n sterility inthes ehybrid s iscompensate d by gene(s)o f Raphanus. Only alo wdegre e ofchromosom e associationwa s found atmeiosi si n AR hybrids. Incompariso nwit hth e results ofRichhari a (1937)an d Tokumasu (1970b), the frequency ofcell swit h onlyunivalent swa s high andth emaximu m number of associated chromosomes per cell low.Genotypi c and environmental variationma y account forsuc hdifferences . Inexten ­ sive studies onchromosom e association ininterspecifi c hybrids and haploids,Mizushim a (1980)foun d amaximu m oftw o autosyndetic chromo­ somepair s inth eA andR genome together. Iftha testimat e iscorrect , themaximu m number ofallosyndeti c chromosomepair snotice d inth epre ­ sentA R hybrids istwo .Highe rvalue shav ebee n found (Mizushima, 1980). A remarkablephenomeno n istha t ahig hproportio n ofth eA Rhybrid s produced aconsiderabl e proportion of stainablepollen .Th emai ncaus e isth eoccurrenc e ofdyad s inth e sporad stage,probabl yb y theforma ­ tiono frestitutio n nuclei ininterkinesis ,a ssuggeste db yMorris , (1936).However ,Tokumas u (1970b)observe d low frequencies ofdyad s in threehybrid swit h arathe r goodpolle n stainability. Themechanis m of formation ofrestitutio n nuclei is still obscure.A possibl e causemigh t be apoorl y developed spindle asa consequenc e ofscant ypairin g ofchro ­ mosomes.However ,variatio n inth edegre e ofdya d formationwa sno trela ­ ted toth edegre e ofchromosom e association. Intercrossing ofA R hybrids gave aver ypoo r seed set.Th e onlyfea ­ siblemetho d ofobtainin g seedwa s togro w thehybrid s inisolatio n and tous ebee s forpollination .A remarkable number ofF .plant sproduce d inthi swa y had ahig hnumbe r of chromosomes comparedwit h the results ofTerasaw a & Shimotomai (1928). Suchplant swit h about 57 chromosomes were frequently found andpresumabl y had thegenom e constitutionAAARRR . Fusiono f 4n eggcell s andunreduce d male gametesma y give rise tosuc h plants.A hybri d with 47chromosome s from acros sbetwee n anA R hybrid (2n =19 )an d B. campestris (2n =20 )possibl y also originated by fusion of ahypoploi d in gamete and anormall y reduced Brassica gamete. InCru - ciferae, theoccurrenc e of 4n gameteswa s reported earlier forhaploid s of B. napus (Heyn, 1974). Chromosome doublingb y colchicine treatmentwit h thetraganth-slim e method of Schwanitz (1949)wa spossible ,bu tha d thedisadvantag e that thedouble d sectorsmostl y were rather small. Asprimar yAAR R hybrids have apoo r fertility (Section 5.3.1), seedproductio n on suchdouble d sectors often failedbecaus e of ashortag e of flowers.

35 Theproductio n ofviabl e unreduced pollen byman yA R hybridsmake s ittemptin g todoubl e thechromosom e numbers ofthes ehybrid smeiotical - lyb yusin g them aspollinato r incrosse s withAAR R hybrids.However , suchcrosse s gave ratherpoo r results,probabl ybecaus e ofth epoo rfer ­ tility ofth eAAR R hybrids used.Th eus e ofmor e fertile xBrassicorapha- nus may improve the seed-set considerably.

36 xBrassicoraphanus: form,fertilit y andcrossabilit y with various species

5.1 INTRODUCTION

Nopublishe d datawer e available onth e agriculturalvalu e of xBrassi­ coraphanus, althoughsuc h allotetraploid hybridshav ebee nproduce d in variousways .Mizushim a (1968)expecte d thatsynthesi s of xBrassicora­ phanus by crossing turnips and radishcoul db e apromisin gwa y ofobtain ­ ing ane w rootcrop . Allotetraploids with thegenom e constitutionAAR Rwer e first found in

anF 4generatio no fsom eA Rhybrid s (Terasawa, 1932). The commonmethod s ofsynthesi swer edoublin g ofth echromosom e number ofA Rhybrid sb y col­ chicine treatment (Mizushima, 1950b;Tokumasu , 1976;Sectio n4.3.5 )an d crossing oftetraploi d formso fth e twoparenta l species (Nishiyama,1946 ; Tokumasu, 1970b;Olsso n& Ellerström , 1980;McNaughto n &Ross , 1978;Sec ­ tion 2.3). Less attractivemethod s are2 xx 4 xcrosse sbetwee n B. campes- tris and R. sativus, although itwa spossibl et oobtai nhybrid s inthi s way (Section 2.3.6). McNaughton (ScottishPlan tBreedin g Station, 1976, p. 16)trie d toproduc e xBrassicoraphanus indirectlyb y crossing B. napo- campestris, an artificial hexaploid with the genome constitutionAAAACC , with autotetraploid formso f R. sativus (RRRR). Inth e subsequentgenera ­ tionso fth eAARR Chybrids ,a gradua l losso fth eC genom ewa sexpected . Manyhundred so fpollination s gaveonl y oneseed . The formo f theAAR R hybridswa spoorl y described. Inman y respects suchhybrid swer e knownt ob e intermediate between theparenta l species and similar toth ecorrespondin g AR hybrids (Terasawa, 1932). Anundesi - redtrai to fsuc hplant swa sth epoo r fertility (Mizushima,1950b ; Tokumasu, 1976;Olsso n& Ellerström , 1980). The crossability of xBrassicoraphanus as afemal eparen twit hth e diploidparenta l specieswa s ratherpoor ,bu t seedswer eobtaine d (Tera­ sawa, 1933;Nishiyam a & Inomata, 1966;Kat o& Tokumas u 1979; 1980). The reciprocal crossesgav en oseed s atal l (Nishiyama& Inomata, 1966). The failureo fsuc hcrosse swa s dueeithe rt oth edeat ho fhybri d embryoso r to abnormal development ofth eendosper m and inhibition ofembry odevel ­ opment (Nishiyama & Inomata, 1966). Kato& Tokumas u (1979)reporte d that xBrassicoraphanus withwhit e flowers showed muchmor ecross-affinit y with white-flowered R. sativus thanwit hdiploi dyellow-flowere d B. japonica Sieb. (= B. campestris ssp. nipposinica). Therevers ewa s found foryel ­ low-flowered xBrassicoraphanus. Pollination ofwhit e flowered xBrassico-

37 raphanus with tetraploid R. sativus gave rise to 1.1 seedspe rpollina ­ tion,wherea spollinatio nwit h tetraploid B. japonica gaven o seeda t all (Kato& Tokumasu , 1978). Terasawa (1933)teste d thecrossabilit y of xBrassicoraphanus with B. napus (AACC), B. cernua (2n =36 ,AABB )an d B. oleracea (CC),usin g the Brassica spp.a spollinators .A n exceptionally high seed-setwa sobtai ­ ned from crosses with B. napus, arathe r good seed-setwit h B. cernua andn o seedwit h B. oleracea. The form and fertility ofprimar y allotetraploid hybrids was studied andth e crossability withvariou s species was tested.

5.2 MATERIAL ANDMETHOD S

Theprimar y allotetraploid hybrids described below arose directly from theintergeneri c crossesmentione d inTable s 6an d 10.Th e data presented on seed-seto fprimar y AARR hybrrds,however , also include crosseswit h 'doubled'A R hybrids (Section 4.3.5). Theparenta l material of xBrassi- coraphanus used incrosse s withvariou s species consisted in 1976 and 1977o fprimar y hybrids and in 1978o fplant s from the second generation. Sometimes thevariou s hybrids arebelo w called, for instance, chinensis or perviridis hybrid, referring toth e Brassica parent inth e original intergenericcross . Most accessions ofth evariou s species crossed with xBrassicoraphanus werementione d inTabl e 3,excep t fortw ovarietie s of spring oil-seed rape. Seed of 'Zephyr', aCanadia nvariety ,wa s kindly supplied byDr . G.R. Stringam (Research Station CanadaDepartmen t ofAgriculture ,Saska ­ toon, Canada)an d seed of 'Tantal',a Frenc hvariety ,b y ZelderB V (Ottersum,Netherlands) . Plantmateria l was treated normally, excepttha t theAAR Rplant swer e transplanted before flowering into the soil ofa heate d greenhouse.Cross ­ ing technique and othermethod s were the same as inpreviou sexperiments . In1976 ,flowe rbud swer emostl y pollinated immediately after emasculation but, since 1977,tw o or three days after emasculation,whe n somebud s had opened. The reason forthi s change oftechniqu e was the questionable significance ofbu dpollinatio n in intergeneric hybridization.

5.3 RESULTS

5.3.1 Form and fertility of primary hybrids

Growthhabi to fF AARR hybridswa s similar totha to fA R hybrids.I n thevegetativ e stage,al lha d aleaf-rosette , irrespective ofth e Bras­ sica parentuse d (Fig.7) .Lea fshap ewa s intermediate butvarie d widely. Leaves andpetiole s were lesspubescen t than of Raphanus. Hairinesspro -

38 Figure7 .Allotetraploi d hybridsfro mcrosse so fa tetraploi d fodderradish , RS14. A.Wit ha diploi d inbred lineo fturni p(o) . B.Wit htetraploi d turnipcv .Tigr a(o) . C.Wit ha tetraploi d formo fBrassic a campestrisssp .perviridi s(o) . ved to be an excellent marker in early stages of development to distin­ guish hybrids from matromorphs in crosses between oriental vegetable forms of B. campestris and R. sativus, because the former were glabrous. Similarity in growth habit to AR hybrids was also remarkable in the generative stage. All hybrids had a main flower stalk, which later pro­ duced many lateral branches. The flowers were somewhat larger than in tetraploid B. campestris and were white or purple. However, two hybrids, both with 39 chromosomes, had yellow flowers. Most flowers, whatever the colour, had petals with strikingly dark veins, although plants with plain white or plain yellow flowers occurred too. Variation in the ex­ tent of flowering was frequently observed. Some plants produced only a

39 B

,fmm' W5ÊÈF

Figure 8. Siliquae and seeds of F hybrids with genome constitution AARR. A. Siliquae of a hybrid from the cross Brassica cainpestris ssp. chinensis (4x) and fodder radish cv. Palet. B.Seed so fB .campestri s ssp.perv iridi s (leftrow) ,a nAAR Rhybri d (middle row)an d Raphanus sativuscv .Pale t (rightrow) .

few flowers, if at all; whereas the flower buds generally dried and dropped prematurily, others flowered abundantly. Siliquae of y-Brassico- raphanus were intermediate between those of the parents (Fig. 8A).Eac h consisted of a dehiscent bivalved part with a septum and a rather large indéhiscent part, the beak or rostrum. The mean number of ovules per pistil was about 19, with some in the beak. The fertility of F, AARR hybrids will be described in more detail in Section 7.3.2. In summary, the hybrids showed a large variation in pollen stainability and varied widely in setting of siliquae. The seed- set, on the other hand, was very poor in crosses between AARR hybrids; on average 0.008 seed per pollination. Usually, these seeds were rather well-filled and were similar in diameter and 1000-grain weight to B. campestris (Fig.8B) .

5.3.2 Crossability with Brassica campestris, Raphanus sativus and Brassi­ ca napus

xBrassicoraphanus was used extensively in crosses with the parental species and B. napus (Table 20).Th e crosses in which diploid forms of

40 Table20 .Result so fcrosse so fxBrassicoraphanu s (AARR)wit hBrassic a campestris (AAan dAAAA) ,Raphanu ssativu s (RRRR)an dBrassic anapu s (AACC).

Cross YearNumbe ro f

femalesee d polli­ sili- seeds hy- matro- plantsprodu -na tnat ­ quae brids morphs cing ions obta­ sown plants ined

AARRx A A chinensishybrid sx B.c.ssp .chinensi s 1976 4 1 ca 85 34 19 19 9 1977 4 0 90 21 0 0 0 chinensishybrid sx B.c.ssp .oleifer a 1976 1 1 ca 6C 7 5 5 3 perviridishybrid sx B.c.ssp .chinensi s 1977 6 1 112 6 4 4 3 1976 1 0 7 0 0 0 0 rapiferahybrid sx B.c.ssp.chinensi s 1976 3 0 16 2 0 0 0 rapiferahybrid sx B.c.ssp .oleifer a 1976 1 0 ca11 0 0 0 0 0

AARRx AAA A perviridishybrid sx B.c.ssp .chinensi s 1977 1 0 19 6 0 0 0 chinensishybrid sx B.c.ssp .chinensi s 1977 2 0 19 0 0 0 0 perviridishybrid sx B.c.ssp .perviridi s 1978 6 0 113 0 0 0 0

AARRx RRR R perviridishybrid sx 'Palet ' 1976 1 1 12 6 1 1 1 1978 8 0 98 25 0 0 0

AARRx AAC C chinensishybrid sx 'Zephyr' 1977 5 3 51 13 20 10 9 'Tantal' 1977 3 1 27 1 1 0 0 perviridishybrid sx 'Zephyr ' 1977 12 6 185 52 115 48 39 'Tantal' 1977 6 1 48 7 7 0 0 rapiferahybrid sx 'Zephyr ' 1977 2 1 19 1 1 0 0 'Tantal' 1977 4 0 83 0 0 0 0

AAAAx AAR R B.c.ssp.perviridi sx perviridi shybrid s 1978 4 0 94 50

RRRRx AAR R 'Palet'x perviridi shybrid s 1978 6 0 70

AACCX AAR R 'Zephyr'x perviridi shybrid s 1977 8 8 134 33 208 99 0 98 'Zephyr'x chinensi shybrid s 1977 8 8 215 63 256 132 4 118

B. campestris were used as pollinators gave an average of 0.06 seed and 0.03 hybrid per pollination. In 1976, only one chinensis hybrid produced seed incrosse swit h B. campestris ssp. chinensis and ssp. oleifera. This plant originated from an AR hybrid, whose chromosome number had doubled during regeneration of plantlets in vitro from stem expiants. In total, this plant gave rise to 12 backcross hybrids. In addition, four hybrids from crosses between this plant and B. campestris ssp. chinensis (2x) were obtained by culture of 43 embryos in vitro (from 13 pollina­ tions only). In 1977, only one perviridis hybrid produced seed. Crosses between xBrassicoraphanus (female) and tetraploid forms of B. campestris (male) gave very few siliquae and no seed (Table 20).Th e re-

41 ciprocal crosses,however , resulted ina muc hbette r setting ofsiliquae , though theywer e seedless.Crosse sbetwee n xBrassicoraphanus (female) and tetraploid forms of R. sativus (male)gav e good setting of siliquae and one seed,bu t the reciprocal crosses yielded only one siliqua andn o seed. Inbot h sets ofreciproca l backcrosses, the reciprocals differ dis­ tinctly insettin go f siliquae,bu t inopposit edirections . Incrosse sbetwee n xBrassicoraphanus (female)an d B. napus, especial­ ly 'Zephyr', amuc hbette r seed-setwa s obtained; on average about 0.35 seedpe rpollination . Several females gave seed, although therewa swid e variationbetwee n females inth e setting of siliquae and seeds.Th emea n seedproductio n ofth ebes t female incrosse swit h rapewa s 3.4 seedpe r pollination.Th e seeds from these crosseswer e larger than frominter ­ crossing xBrassicoraphanus andwer eo n average 2.2 mm indiameter .Th e reciprocal cross,AAC Cx AARR ,gav e riset o evenmor e seedpe rpollina ­ tion,o n average 1.3.Howeve r nearly all22 0plant s raised from these seedsturne d outt ob ematromorphs .Onl y fourplant swer ehybrid s and arose from small seedswit h adiamete r less than2. 0 mm.

5.4 DISCUSSION

Theprimar yAAR R hybridsmostl y had awel lbalanced , rather vigorous and regular growth.Th e growing conditions,however , dono tpermi tfar - reaching conclusions aboutth e agronomic value of xBrassicoraphanus, al­ though thevigou r ofth ehybrid s indicated potential as aforag ecrop . Thelesse rhairines s than fodderradis hwoul dmak e themmor epalatabl e forcattle .Morphologica l characteristics, like abundant flowering anda rather highpotentia l seedproductio npe r siliqua indicate that xBras- sicoraphanus could yield amuc h largernumbe r ofsee d than fodder radish but the seed size ismuc h smaller thano f fodder radish.Threshin gdiffi ­ culties, as in R. sativus, canb e expected to some extent in xBrassico- raphanus, because some ovules are inth e indéhiscentbea k ofth esiliquae . Giventh e appearance ofth ehybrid s and the large diversity in B. campes- tris and R. sativus, thecreatio n ofattractiv e andvariabl e sourcemate ­ rial forth eproductio n ofa ne w forage crop, xBrassicoraphanus, ispossi ­ ble. Themai nhindranc e inachievin g this goal isth ever ypoo r fertility ofnewl y synthesizedhybrids . Mostcrosse sbetwee n xBrassicoraphanus (female)an ddiploi d formso f B. campestris failed to setseed .Onl y twowhite-flowere dhybrid sprodu ­ ced seed;on eplan t evenha d anexcellen t setting of siliquae andseeds . The differences observed incrossabilit y were certainly notrelate dt o the flower colour of xBrassicoraphanus as such, as suggested byKat o & Tokumasu (1979). The twoyellow-flowere dplant s set fruitpoorl y and gave no seed atall . Reciprocal crosses of xBrassicoraphanus with tetraploid forms of B.

42 campestris ando f R. sativus werenearl y allunsuccessful .Th ediffer ­ ences insettin go fsiliqua ewer e remarkable;crosse swit h J?, sativus (4x)gav eth ebes tresult swit h xBrassicoraphanus as femaleparent .Th e reverse held true forcrosse swit h B. campestris (4x). These contrasts in fruitsettin gmigh treflec tth e effecto fselectio n forbette rcross - able genotypes in B. campestris and R. sativus, whichprobabl y occurred during the synthesis of x-Brassicoraphanus. Setting ofsiliqua e ando fhybri d seedwer equit egoo d incrosse sbe ­ tween xBrassicoraphanus (female)an d B. napus. The reciprocal crosses also resulted ina goo d seed-set,bu tmos tseed sgav e riset omatromorphs . Although somematromorph s mighthav eoriginate d from illegitimate sel- fings,matromorph ymus toccu r frequently inth e rapevariet y 'Zephyr'. Theoccurrenc e ofhexaploi d hybrids inoffsprin g ofthes e crosses further supports thisvie w (Table 36).Strikingl y largenumber s ofmatromorph s were also obtained incrosse sbetwee n B. napus andA R hybrids (Table17) . SoA R andAAR Rhybrid s apparentlyhav ea nexcellen t ability toinduc e parthenogenesis in B. napus. R. sativus probably has asimila r ability inrap e (McNaughton& Ross , 1978). Thereciproca l difference inproductio n ofhybri d seed incrosse sbe ­ tween xBrassicoraphanus and B. napus was strikingly similar to aclassi c example ofunilatera l incompatibility, namely thecros sbetwee n B. olera- cea and R. sativus. Inbot hcombinations ,crosse s aremor e difficultwit h species having aC genom e for femaleparent ; B. napus (AACC)an d B. olera- cea (CC),respectively . The good seed-seto f x-Brassicoraphanus plants afterpollinatio nb y B. napus showed thatth epoo r fertility of xBrassicoraphanus wasno tcause d by adisturbe dmegasporogenesis .

43 6 xBrassicoraphanus:sterilit y factors and mating system

6.1 INTRODUCTION

Poor fertility is anorma l phenomenon inearl y generations of xBrassi­ coraphanus (Section 7.3.2). Small-scale cytogenetic studies have shown thatmeiosi s inthi shybri d ismostl y regular:usuall y onlybivalent s were found atmetaphas e I (Terasawa, 1932;Mizushima , 1944b, 1950b;Nishiyam a & Inomata, 1966;Tokumasu , 1976). Tokumasu (1976), forexample , foundonl y bivalents in6 8% ofth epolle nmothe r cells examined; theothe r cellscon ­ tained someunivalent s ormultivalents .H e concluded thatth e frequency of meiotic irregularities was too low tocaus e suchpoo r fertility. Nishiyama &Inomat a (1966)showe d thatmeioti c irregularities mightreduc e fertility in later generations.The y examined meiosis intw o strains originating from the amphidiploidsproduce d byTerasaw a (1932)an d found thaton e strain withhighl y fertilepolle n andhig h seed-setha d regularmeiosi s andth e other strainwit hpolle n of low fertility had slightly irregular meiosis (a fewunivalents ,sometime s fragments). Ingeneral ,meiosi s in xBrassi­ coraphanus gave rise tobalance d gametes,whic hwer eofte nviable .Fo r example, theprimar y AARR hybrids usually produced functional andnormal ­ lyreduce d eggcells ,a s indicated by thegoo d crossability with B. napus (Section5.3.2 ). So sterility in xBrassicoraphanus mustb e caused bybarrier s inth e progamic phase and after fertilization. Kato (1971)showe d that 50t o 70% of allembry o sacs (ina fe wplant s ofprobabl y an advanced genera­ tion)ha dbee n fertilized fiveday s afterpollination . Fifteen days after pollination, the embryos had reached theglobula r orearl y heartstage . Subsequently, the development ofembryo swa s retarded incompariso n to that inth eparenta l species andman y embryos stopped growing anddege ­ nerated. Tokumasu (1976)an d Kato& Tokumas u (1976)observe d asudde n increase in fertility inassociatio nwit h achang e in flower colour fromwhit e to yellow.Th e authors concluded that asterilit y gene controlling thedeve ­ lopment ofembryo s (orendosperm )wa s closely linked toth e white-flower gene and that the increase in fertilitywa s causedb y aninterchang ebe ­ tween a Brassica and a Raphanus chromosome.Thi s interchange finallyre ­ sulted ina duplication-deficienc y type ofplant , lacking the sterility gene aswel l asth ewhite-flowe r gene.However , the sterility ofth e white-flowered plants may alsob e caused by adiscordanc e between the

44 cytoplasm of Brassica andth ewhite-flowe r gene (orrathe r genes linked to it)fro m Raphanus. Recently the same authors suggested thatther e mustb e stillothe rgene s affecting fertility (Tokumasu& Kato , 1980). Sterility isa seriou s andpoorl yunderstoo d problem. Sovariou sas ­ pects of itwer e studied:th echromosoma l behaviour inmeiosis ;th e occurrence ofbreedin gbarriers ;th epresenc e and significance ofself - incompatibility.

6.2 MATERIALAN DMETHOD S

Meiosiswa s examined inG .an dG _plant s from theP H and CHpopulatio n of x-Brassicoraphanus (Section7.2) .Th e cytological techniques werede ­ scribed inSectio n2.2 . Breedingbarrier s andexten to fself-incompatibilit y in x-Brassico­ raphanus were tested ina nexperimen twit hseve nP Hplant s and fiveC H plants ofG_ .Abou t fivebud s of four inflorescences oneac ho fthes e plantswer eemasculate d andbagged .Thre eday s later,tw o inflorescences ofeac hP H and CHplan twer epollinate d withbulke dpolle n from all available PHo rC Hplants ,respectively .A tth e sametime ,th etw oremai ­ ning inflorescenceswer e selfed.Th epistil s ofthos e inflorescences were fixedtw oday s afterpollinatio n and later studiedb yultraviole tmicros ­ copy forpolle ngermination ,penetratio n ofpolle ntube s intoth estyle , thenumbe r ofpolle n tubespe r style and thenumbe r ofovule swit h apol ­ lentub ei nth emicropyl e (Section 3.2). Thewhol e testwa s repeated threeweek slater .

SixG 2 plants ofth eP Hpopulatio n and threeG .plant s ofth eC Hpopu ­ lationwer e tested separately for setting of siliquae and plant-to-plant variation innumbe r ofovule s developed per siliguaan d development of embryos.A total of 60t o8 0emasculate d flowerspe rplan twer epollina ­ tedwit hbulke dpolle n from several PHo rC Hplants ,respectively . Thirty days later,th edevelope d siliquae andovule swer ecounte d andth edeve ­ lopmental stageo f allembryo swa s determined according toth e scheme of Wilmar& Hellendoor n (1968).

6.3 RESULTS

6.3.1 Meiosis

Meiosis in x-Brassicoraphanus was similar totha t inth ediploi dpar ­ ental species and showed almostexclusivel y bivalents (Fig.9A ,B) .A random sample ofeuploi d G plants from theP H andC Hpopulatio n showed bivalents only in7 9an d 87% ofth epolle nmothe r cells,respectively , and forG 2plant s these figureswer e 91an d 94% (Table21) . Themai n irregularity inmeiosi s ofeuploi dplant s (2n= 38 )wa sth e

45 B

ïy

* « .v

•1 *

•

i • # »

•

Figure9 .Meiosi s inxBrassicoraphanu s (AARRhybrids) . A.,B .Polle nmothe r cellsa tmetaphas e Iwit h 19I I (X1540). C.Polle nmothe r cella tanaphas e I(X1090) .

occurrence ofunivalents .Tw ounivalent s were found in about 13% ofal l cells examined and four inonl y one cell.Th e only other irregularity, multivalents,occurre d sporadically; only 1.4 % of all cells examined had aquadrivalent , usually ofchai n configuration. Aneuploid plants showedmor emeioti c irregularities thaneuploi d ones. Thehypoploi d plants (2n= 37 )ha d oneunivalen t and 18bivalent s inabou t 82% ofth ecell s (Table 22).Mos t ofth eremainin g cellsha d three or fiveunivalents .Multivalent s were observed sporadically, as inth eeuploids . Hyperploid plants (2n =39 )showe d trivalents in about2 5% of the cells studied,bu tmos t cells contained one,thre e or even fiveunivalent s as well asbivalent s (Table 23).Quadrivalent swer eno tnoticed .

46 Table21 .Chromosom eassociatio na tdiakinesi san dmetaphas eI i npolle nmothe r cellso feuploi dG andG plants (2n= 38 )fro mth eP Han dC Hpopulatio no f xBrassicoraphanus.

Plant Numbero f Numbero fcell swit h cellsex ­ amined 1911 1811+ 2 1171 1+ 4 1171 1+ 1IV 1IV+ 161 1+ 2 1

PHpopulation/ G PH3 50 42 6 PH5 25 22 2 PH6 25 22 3 PH9 25 22 3 PH18 25 22 3 PH10 16 5 11 PH12 9 1 PH15 25 18 76.2079-2 7 6 76.2257-4 25 21 3 75.2326-1 25 19 5 76.2256-1A 25 22 2 76.2256-1B 25 22 3 Total 307 244 56

PHpopulation/G, , PH2-1 25 22 2 PH3-1 22 18 4 PH5-1 25 23 2 PH7-1 25 24 1 PH7-3 25 22 3 PH9-6 25 25 Total 147 134 12

CHpopulation/ G CHI 25 24 1 CH2 25 21 4 CH3 25 20 3 76.2104-2 25 22 3 Total 100 87 11

CHpopulation/G, , CH3-4 25 22 CH4-3 25 25 CH4-6 50 47 Total 100 94

-"]I ,II ,II Ian d IVar esymbol sfo runivalents , bivalents,trivalent san d quadrivalents,respectively .

At anaphase I, the homologous chromosomes separated synchronously and were usually regularly distributed to the poles (Fig. 9C).Laggard s and misdivisions of chromosomes were observed, though rarely (Table 24).Abou t 88 % of all pollen mother cells in the euploid plants showed an equal chromosome distribution (19/19). Like the first meiotic division, the second meiotic division was quite regular with occasional laggards and misdivisions. The high frequency of pollen mother cells with univalents in the aneu- ploid, probably monosomic and trisomie, plants resulted in more half chro-

47 Table 22.Chromosom e association at diakinesis and metaphase I in pollen mother cells of hypoploid G and G plants (2n = 37) from the PH and CH population of XBrassicoraphanus.

Plant Numbero f Number of cells with* cells examined 1811+ 11 1711 +3 1 1611 +51 other* *

PHpopulation/G , PHI 25 24 1 PH4 25 24 1 PH14 25 23 1 1 76.2147-2 13 7 4 1 1 76.2229-1 25 25 76.2229-12 4 1 3 76.2237-7 25 2 14 9 76.2251-2 25 23 2 Total 167 129 26 10 2

CHpopulation/ G CH-5 25 20 5 76.2104-7 25 23 1 1 76.2202-2 25 25 Total 75 68 6 1

CHpopulation/G, , CH2-1 14 12 1 1

*] I, II,II I and IV are symbols for univalents,bivalents , trivalents and quadrivalents, respectively. »*]Othe r chromosome configurations, i.e. 1 III + 15 II + 4 I, 1I V + 16 II + 1 I, or 1 III + 16 II + 2I

Table 23.Chromosom e association at diakinesis and metaphase I in pollen mother cells of hyperploid G and G plants (2n = 39) from the PH and CH population of xBrassicoraphanus.

Plant Number of Number ofc e 11s with " cells examined 1811 + IUI 1911+ 11 1811 +3 1 1711+ 5 1

PHpopulation/ G PH8 26 2 16 6 2 PH11 25 6 18 1 76.2154-5 25 8 14 3 76.2237-11 25 2 9 12 2 Total 101 18 57 22 4

CHpopulation/ G CH7 25 5 20 76.2104-3 2 1 1 Total 27 6 21

CHpopulation/G, , CH1-1 25 10 14 1 CH4-5 18 7 8 3 Total 33 17 22 4

*] I, II, III are symbols for univalents, bivalents, trivalents respectively.

48 Table24 .Chromosom edistributio na tlat eanaphas eI an dmetaphas eI Ii npolle n mothercell so fG .plant sfro mth eP Han dC Hpopulatio no fxBrassicoraphanus •

Plantswit h2 n= 3 8 Plantswit h2 n= 3 7 Plantswit h2 n 39 distribution numbero f distribution numbero f distribution numbero f cells cells cells

19 :1 9 169 19 :1 8 64 19: 2 0 29 18: 2 0 16 20 :1 7 1 19%:19 % 8 19 :18 * 4 18 :18 * 9 19: 19 * 3 17 :2 1 1 18%:18 % 17 Other 1 Other 2 Other 13

Totalnumbe r ofcell s 192 104 41 examined

Numbero f laggardspe r 0.021 0.13 0.07 cell

Numbero f misdivisionspe r cell 0.016 0.26 0.20 *]plu son elaggar d

mosomes and laggardspe r cell inth e subsequentmeioti c stages thani n euploidplants .Th eprobabl e cause ofhal fchromosome s was centromere misdivisioni nth eunivalents . Inth eeuploids ,rathe rhig h frequencies ofunbalance d gameteswer eproduced ,becaus e ofth eoccurrenc e ofuni ­ valents.Mos tunivalent s were included inon eo fth eanaphas e groups of chromosomes. Univalent formationwa s apparently undergeneti c control,becaus emos t G-plant swit hman yunivalent s were derived from the same Brassica plant, i.e. PH10 andPH1 2 inTabl e21 ,plan t 76.2237-7 inTabl e 22an dplan t 76.2237-11 inTabl e23 .No t surprisingly, thehalf-si b families derived fromthes eplant s contained ahig h frequencyo faneuploi d plants.

6.3.2 Barriers in the progamic phase

Stigmas ofth ecross-pollinate d flowers ofG _plant s from theC Han d PHpopulatio n always showed serious incompatibility reactions twoday s afterpollination . Strongultraviole t fluorescence due tocallos edepo ­ sits occurred inman y stigmapapilla e (Fig.10A) ,an dpolle ngrain smost ­ ly failed togerminat e orproduce d ashor tultraviolet-fluorescen tpolle n tube. Only fewtube s actually penetrated thestigma .Th e final resulto f these incompatibility reactionswa s apoo rpenetratio n ofpolle n tubes into style and ovary.Th enumbe r oftube s countedpe r stylevarie d from 0 to2 2an dwa s ona naverag e about10 .I twa s nearly always less than

49 Figure 10.Ultraviole t fluorescence photographs of the progamic phase in xBrassicoraphanus 2 days after pollination. A. Callose in stigma papillae after cross-pollination (xlOO). B. Callose in stigma palillae after self-pollination (x24). C. Fertilized ovule (xlOO). D. Bundle of pollen tubes in the style after self-pollination (xlOO).

50 the number of ovules in an ovary (about 19). Significant differences were observed between individual plants (Table 25). In a few plants, how­ ever, there was rather large variation between the test dates, for in­ stance plant PH9-4. The pollen tubes that penetrated apparently grew unhindered through the style and nearly all reached the ovary. Many tubes subseguently en­ tered an ovule through the micropyle. Such ovules with a tube in their micropyle were classed as fertilized, although it was unknown whether fertilization had occurred (Fig. IOC). There were large and significant differences in the number of fertilizated ovules per pistil between PH plants but not between the CH plants. Number of fertilized ovules per pistil ranged from 1.2 in PH7-2 to 14.9 in PH3-2. Except for PH9-4, this number was nearly equal on the two evaluation dates. The low number for PH9-4 in the second test was obviously associated with the few pollen tubes in the style at that date; it accounted completely for the diffe­ rence between the overall means of 6.3 and 4.9 fertilized ovules per ova­ ry (against 5.5 and 5.3 respectively if the values for PH9-4 were ex­ cluded) . Table 25 shows a large variation in the ratio of number of fertilized ovules and number of pollen tubes in the style, i.e. the fertilization

Table 25. Results of ultraviolet-fluorescence studies on breeding barriers in the progamic phase of fertilization after cross-pollination in the PH and CH population of xBrassicoraphanus. These tests were repeated after three weeks. Each value represents the mean of 4 or 5 flowers.

Plant Number of pollen Number of fertilized Fertilization tubes per style ovules per ratio (B/A) (A) ovary (B)*

Test1 Test2 Test 1 Test2 Test1 Test2 PHpopulatio n PH3-2 14.60 17.50 14.00 15.75 0.96 0.91 PH5-1 4.75 11.00 3.25 5.00 0.80 0.42 PH7-1 9.80 6.25 6.00 3.00 0.57 0.50 PH7-2 11.80 11.60 1.40 1.00 0.12 0.08 PH7-3 12.20 11.75 11.00 10.00 0.89 0.83 PH9-2 5.00 8.00 2.00 2.40 0.47 0.31 PH9-4 17.40 2.80 17.00 1.40 0.98 0.36

CHpopulatio n CH1-1 11.00 17.75 5.60 5.50 0.52 0.30 CH4-3 6.50 12.40 3.00 3.60 0.47 0.28 CH4-4 5.60 12.25 3.60 3.25 0.64 0.37 CH6-1 6.80 5.60 5.20 3.00 0.78 0.60 CH7-1 9.75 - 3.75 - 0.39 - Mean 9.60 10.63 6.32 4.90 0.63 0.47 Leastsignifican t difference (0.05) 3.36 5.70 2.80 2.96 0.27 0.26

"]Fertilize d ovulewa sdefine da sovul ewit hpolle ntub ei nit smicropyle .

51 ratio.Thi svariatio n didno tmerel y reflectvariatio n inth enumbe ro f pollen tubes inth e stylebecaus e theplan tmean s forthes e traitswer e not significantly correlated (r= 0.18). The fertilization ratio andth e number ofpolle n tubes inth e stylewer e apparently two independent causes ofvariatio n innumbe r of fertilized ovules per ovary.Thi s agrees with theobservatio n that inplant swit hpoo r fertilization theorienta ­ tiono fpolle n tube growthwithi n theovar ywa s oftendisturbe d asi n theorigina l intergenericcrosses .

6.3.3 Barriers after fertilization

Ina sampl e of xBrassicoraphanus plants,th e setting ofsiliqua e and thedevelopmen t ofovule s and embryoswer e assessed thirty days after cross-pollination. By then,nearl y allembryo s had reached the final size and stage.Th e setting ofsiliqua e ranged from 38% inCH4- 4t o9 7% in PH7-3 (Table 26). Theplant s also differed inmea nlengt h andoute rap ­ pearance ofth e siliquae (Table 26).Smal l siliquaewer e oftenyello w green and sometimes slightly shrivelled. Other fruits,mainl y the larger ones,wer enorma l green and often swollen.A positiv e correlation of 0.65 was found forplant sbetwee n setting and average length ofsiliquae . Thevariatio n inbot h characteristics probably reflected variation in thenumbe r of fertilized ovules,becaus ebot h traitswer e significantly correlated with thenumbe r ofdevelope d ovulespe r siliqua (r - 0.93 and r =0.65 , respectively).

Table 26. Development of siliquae, ovules and embryos on G plants from the PH and CH population 30 days after pollination with bulked pollen mixture from various plants of the PH or CH population, respectively.

Plant Number of Numb er of Mean Nuinbe r of Number of Number of polli­ sili quae length ovu ] es ovules embryos nations obtained of sili­ deve loped collapsed per siliqua quae (cm) per sili- after qua initital development

PH population PH3-2 82 77 4.3 6.6 5.2 1.3 PH4-4 62 44 3.7 2.0 1.2 0.6 PH7-2 62 37 1.7 0 0 PH7-3 63 61 2.5 1.7 0.6 0.9 PH9-4 62 51 2.3 0.7 0.6 0.04 PH12-2 61 54 4.3 7.5 1.9 3.8

CH populs tion CH4-4 61 23 1.7 0.4 0.4 0 CH4-6 64 59 5.1 8.0 7.6 0.3 CH6-1 86 46 2.8 1.2 1.2 0.02

Mean 3.2 3.1 2.3 0.77

52 Thetrai tdevelope d ovulespe r siliqua (Table26 )include d all ini­ tially developed ovules.Ther ewa s enormousplant-to-plan tvariatio n in it.O nth ewhole ,th enumbe r ofovule s developed per siliquawa s lower thanth enumbe r ofovule swit h apolle n tube inthei rmicropyl e (Table 25). Ahig hproportio n ofth e initially developed ovules later collapsed and shrivelled (Table26) . Mostunshrivelle d ovules contained anembryo ,bu t about2 6% ofthe m contained no embryo.Th enumbe r ofembryo spe r siliquawa s only0.77 , whichwa s far lesstha nth enumbe r ofinitiall y developed ovules.How ­ ever individual plants differedwidely ;PH12- 2produce d 3.8 embryospe r siliqua,wherea s inothe rplant s not asingl eembry o couldb e detected. Thedevelopmenta l stageo fal lembryo swa s classified accordingt o the system ofWilma r& Hellendoor n (1968).Althoug h this systemwa sde ­ vised forBrussel s sprouts,i tcoul db ereadil y applied to xBrassico- raphanus. Thenumbe r ofembryo s isolatedpe rplan tdiffere d greatly,bu t nearly allplant sha d analmos tcontinuou s frequency distribution ofth e embryos overth evariou s developmental stages,indicatin g thatther ei s not aspecifi c stage atwhic hgrowt h and development stopped.Th ediffe ­ rences infrequenc y distributionbetwee nplant s suggesttha tsom evaria ­ tionexiste d inth edegre e ofinhibitio n ofgrowt h anddevelopmen t (con­ trastPH7- 3 and PH12-2). Theconditio n ofth e ovules indicated thatmos t embryosha d already stopped developing orcomplete d development.Unde r normal conditions,onl y theovule swit hembryo s inth ewalkin g sticko r mature I+ IIstag ewoul dprobabl y give rise toviabl e seed (Table27) .

Table27 .Classificatio nb yth eschem eo fWilma r& Hellendoor n (1968) ofal lembryo sobtaine d fromth ecrosse smentione d inTabl e25 .

Plant Numbero f Developmentalstage * embryos ET LT WS

PHpopulatio n PH3-2 97 4 46 25 16 1 3 2 PH4-4 26 6 7 4 5 3 1 0 PH7-3 55 2 8 13 16 6 6 4 PH9-4 2 1 0 0 1 0 0 0 PH12-2 203 7 90 62 26 6 1 1

CHpopulatio n CH4-6 18 0 0 7 3 5 0 3 CH6-1 1 0 0 0 1 0 0 0

Total 402 30 151 111 68 21 11 10

'']G globula rstage ;H hear t+ earl yhear tstage ;E Tearl ytorped o stage;T torped ostage ;L Tlat etorped ostage ;W Swalkin gstic k stage;M matur eI + II .

53 6.3.4. Mating system

Theparenta l species of xBrassicoraphanus arepredominantl youtbree - ders,becaus e ofthei r self-incompatibility systems accompanied by insect pollination. xBrassicoraphanus plants also attract insects,bu tther e wasn opublishe d data onwhethe r they are self-incompatible.T o study this critical aspecto f fertilization, two series of selfingswer emad e onth e sameplant s asuse d fortest s onbreedin g barriers inth eproga ­ micphas e of fertilization. Themea nnumbe r ofovule s fertilized perova ­ rywit h selfing and cross-pollination werepresente d inTabl e 28.Fou r plantswer e completely self-incompatible (Fig. 10B),an dthree ,PH3-2 , PH7-2 and CH4-3, had amuc h smallernumbe r ofpolle n tubespe r style after selfingtha n after cross-pollination. Absence ofpolle n tubes inth estyl e couldb e caused by apoo rpolle ngerminability .However ,polle nstainabil - itywa s notrelate d to self-incompatibility andgerminate d pollen grains were observed on selfed stigmas of allplant s examined by ultraviolet microscopy. Sosuc hplant swer eprobabl y truly self-incompatible, aswoul d be expected for allotetraploids derived from two specieswit h a sporophytic incompatibility system. Incontrast , fiveplant swer e fully self-compatible. Inthes eplants ,th enumbe r ofpolle n tubespe r style after selfingwa s evenhighe r than that after cross-pollination (Fig.10D) .However , the progamic barrier onth e stigma (Section6.3.2 )wa s observed inal l stig-

Table28 .Compariso no fXBrassicoraphanu splant s fromth eP Han dC Hpopula ­ tionsfo rnumbe ro fpolle ntube spe rstyl ean dnumbe ro fovule sfertilize d perovar yafte rself-pollinatio n andafte rcross-pollination .Eac hvalu e represents themea no f8-1 0flowers .

Plant Pollen Numbero fpolle n Numbero fovule s stainability tubespe rstyl e fertilizedpe rovar y (%) crossed"' selfed crossed"' selfed

PHpopulatio n PH3-2 74 16.05 0.20 14.88 0.00 PH5-1 66 7.88 0.00 4.13 0.00 PH7-1 39 8.03 8.10 4.50 6.20 PH7-2 64 11.70 1.75 1.20 0.00 PH7-3 61 11.98 0.00 10.50 0.00 PH9-2 28 6.50 0.00 2.20 0.00 PH9-4 62 10.10 13.90 9.20 10.10

CHpopulatio n CH1-1 33 14.38 15.45 55 5.70 CH4-3 63 9.45 1.50 30 1.00 CH4-4 37 8.93 9.45 43 3.35 CH6-1 28 6.20 0.00 10 0.00 CH7-1 55 9.75 10.10 75 2.50

Mean 51 10.08 5.04 5.56 2.40

Data fromTabl e25 .

54 > mas,whethe r self-o rcross-pollinated . Soself-pollinatio n ofself-in ­ compatible plants only further reducedpenetratio n ofpolle n tubes into thestyle .

6.4 DISCUSSION

Inth epopulation s studied, sterility had several causes.Th emai n factorswer e incompatibility reactions onth e stigma,disorientatio n of pollen-tube growthwithi n theovary ,poo r settingo rgrowt h ofsiliguae , andpoo r growth and arrested development ofembryos .Al l these factors dono toperat e equally inal lplant s orhav eth e same impacto nfertil ­ ity.A disorientation ofpollen-tub e growth, forexample ,wa s absenti n someplants .Th emajo r causeo f sterilitywa s undoubtedly poor growth and arrested development ofembryos ,probabl yb ymalfunctionin g ofendo ­ sperm, asconclude db yKat o &Tokumas u (1976). The sterility factorswer e quite similar toth ebreedin gbarrier sob ­ served inth eorigina l intergeneric crosses. Interm so fth etheor y of incongruity (Hogenboom 1973; 1975), thismean s thatth ebarrie r andpene ­ tration capacity of xBrassicoraphanus matchedpoorly . Soresidua l fac­ tors ofth e incongruity system of B. campestris and R. sativus were at leastpartiall y responsible forsterility .Th e situation insuc hnewl y established allotetraploids isprobabl y like adoo rwit h adoubl e lock: thebarrie r capacities of B. campestris and R. sativus. Fertilization only occurs ifth ekey ,th e combinedpenetratio n capacities ofthos e specieswithi n xBrassicoraphanus, fitsbot h locks.A consequence ofthi s hypothesis istha t selection foreasie rcrossabl e genotypeswithi nth e parental specieso f y-Brassicoraphanus may alsohav e apositiv e effecto n fertility of allotetraploids derived from suchgenotypes . Chromosome association inmeiosi sagree d quitewel lwit htha tdescri ­ bedb yTerasaw a (1932), Mizushima (1950b)an dNishiyam a& Inomata (1966). Theonl y irregularitieswer e acertai ndegre e ofuni -an dmultivalen t formation.Tokumas u (1976), incontrast , foundman ymor e multivalents perpolle nmothe r cell. Sinceon eo fhi sF .hybrid swa s amixoploi d (2x/4x)wit h adiploi d sectorproducin g stainablepollen ,hi sF _plant s mayhav eoriginate d from a4 xx 2 xcross .Suc ha norigi nwoul d givea more satisfying explanation forth e remarkably high frequency ofmulti ­ valents. The degree ofunivalenc e in xBrassicoraphanus was farles stha n inoctoploi d andhexaploi d triticale (Scoles& Kaltsikes , 1974), sotha t aneuploidy in xBrassicoraphanus seems ales s importantproblem . There was some geneticvariatio n inthi s characteristic of xBrassicoraphanus and so selection againstplant swit ha comparativel y highdegre e ofde - synapsis should reduce the frequency ofaneuploids . Fertility problems inth enewl y established populations areprobabl y under genetic control and atleas t some genetic variation seems toexis t

55 inal l components of sterility.A simplebreedin gprogramm e toaccumul ­ ate fertility factors should improve fertility, asmus thav ehappene d in theextensiv ebreedin gprogramme s of xRaphanobrassica (McNaughton,1973 ; Ellerström &Zagorscheva , 1977;Iwas a& Ellerström , 1981). The question remains whether thepopulation s contain sufficient genetic variation for all sterility factors. Ifnot ,structura l changes inth ekaryotyp e (e.g. loss of achromosom e segment carrying asterilit y gene)coul db ebenefi ­ cial (Tokumasu, 1976). However, aseriou s disadvantage ofsuc h changes isth epossibl e loss ofusefu l genetic information.Fo r instance,intro ­ duction into thepresen tpopulation s of xBrassicoraphanus of thedefi ­ ciency thatwa sdescribe db y Tokumasu (1976)t o improve the fertility would probably soonresul t incompletel y homozygous populations forthi s deficiency,becaus e ofth e selective advantage ofremova l of asterilit y factor. Its genetic load,however , isunknow n and so introduction into thebreedin g programme should be avoided. Mizushima (1950b)reporte d that xBrassicoraphanus was self-sterile. Thepresen t results,however , show that such allotetraploids couldb e self-compatible, despite thepresume d presence ofa one-locu s sporo- phytic self-incompatibility system inth eparenta l species of xBrassi­ coraphanus. The occurrence of self-compatible plantsha s alsobee nob ­ served inothe r artificial cruciferous allotetraploids derived from self-incompatible diploid species,e.g . in xRaphanobrassica (Howard, 1942a). The occurrence ofself-compatibl e plants inou rpopulation s indi­ catestha tth eself-incompatibilit y systems ofth eparenta l specieswer e sometimes entirely orpartiall y eliminated. Thepresenc e of self-incom­ patibleplant s does notnecessaril y imply that self-incompatibility mech­ anisms ofth eparenta l species were actively involved; poorly matching barrier andpenetratio n capacities couldb e responsible aswell . Ifso , improvement inmatchin g betweenbarrie r andpenetratio n capacities should increase theproportio n of self-fertileplants .

56 7 Improvement in fertility of xBrassicoraphanus

7.1 INTRODUCTION

Primary allotetraploidhybrid s derived fromcrosse sbetwee n B. campes- tris and R. sativus werepoorl y fertile orproduce d no seeds atal l (Hosoda, 1946,1947 ;Mizushima , 1950b,1952 ;Tokumasu , 1976;Olsso n & Ellerström, 1980). Inm yhybrids ,fo rexample ,th emea n seed-setpe rpol ­ linationwa s only about0.008 .Mizushim a (1950b)eve n failed toobtai n any seeds atall .I nth e subsequentgeneration s ofnewl y established xBrassicoraphanus, Olsson& Ellerströ m (1980)observe d deterioration and extinctionb y the fifthgeneration .Othe r authors,however ,obtaine drath ­ er fertile strains of xBrassicoraphanus (Terasawa,1932 ;Tokumasu ,1976 ; Tokumasu& Kato , 1980). Terasawa (1932)reporte d theoccurrenc e of allotetraploids inth eF . ofa cros sbetwee n B. campestris ssp. chinensis (2x)an d R. sativus (2x). Mosto fthes eplant swer ecompletel y fertile. In1966 ,tw o strains of Terasawa'smateria l still existed (Nishiyama& Inomata, 1966). Oneo f them (2n =38 )ha dhighl y fertilepolle n and ahig h seed-set.Anothe r fertile straino f xBrassicoraphanus produced byTokumas u (1976)showe da sudden increase in fertility inF _ derived from twoF .hybrid s ofwhic h thechromosom e numberwa s doubledwit hcolchicine .Th e fertility inthi s straingraduall y improved even further inth e subsequent generations (Kato& Tokumasu , 1976). Recently the same authors obtained ane wrathe r fertile strain from anallodiploi d of acros sbetwee n atriploi d B. japon- ica (AAA)an d diploid R. sativus (RR).On eplan tobtaine db yopen-pollina ­ tion (possiblywit h the fertile strainproduce d earlier asth epolle n donor)wa s anallotetraploi d andproduce d aboutthre e seedspe rsiliqua . All existing strains of xBrassicoraphanus with arathe r goodfertil ­ ityhav e anarro wgeneti cbasis . Incontrast ,th ehybrid s obtained in thepresen t studyhav e amuc hbroade r origin. Improvement in fertility isprerequisit e forexploitatio n ofan yagricultura l value of xBrassico­ raphanus . So fertility improvement and chromosomal stabilitywer e stud­ ied inearl y generations oftw opopulation s composed ofprimar yhybrids .

7.2 MATERIALAN DMETHOD S

Thetw o xBrassicoraphanus populations coded PH andCH ,wer e composed ofprimar y allotetraploids.Th eP Hpopulatio n consisted of4 8 perviridis

57 Table 29.Siz e of the PH and CH population in the three generations studied and the number of isolation cages used in each generation.

Generation PopulatioPopulationn NumbeNumber orf of Number of isolation plant;»* cages

G PPHH 3 3 48 CH 1 14 G2 PPHH 3 3 82(-1782( ) CH 1 30(-9) G PPHH 6 6 180 80 CH 1 25

Inbracket s the number of plants eliminated because of highly deviant chromosome numbers.

hybrids from crossesbetwee n B. campestris ssp. perviridis (4x)an dth e tetraploid fodder radish accessions 'Palet'an dRS1 4 (Tables 6an d10) . The CHpopulatio nwa s composed of 14 chinensis hybrids derived from crossesbetwee n B. campestris ssp. chinensis (4x)an d 'Palet' (Tables6 and10) . Bothpopulation s werepropagate d for three generations inisolatio n cages (Table 29). Ineac hcage ,rando mmatin gwa s stimulated byhoney ­ bees for almostth ewhol e floweringperiod . The generationsG _ andG 3 of thetw opopulation s included allhalf-si b families obtained inth epre ­ vious generation.T o avoid inbreeding asmuc h aspossible ,th e samenum ­ ber ofseedling swa s raised foreac hhalf-si b family, ifsufficien t seeds were available,an d forP Hequall y distributed over the cages.S oP Hi n G.,G _ andG ,consiste d of3 , 3an d 6subpopulations , respectively,where ­ asC Hwa spropagate d ason epopulation . Someplant s with exceptionally high orlo wchromosom e numbers wereexclude d fromG ~ andmultiplie d sepa­ rately (Table 29).I fpossible ,thes eplant s were replacedb y otherplant s from the samehalf-si b families. The isolation cagesuse d forpropagatio n were 2.5mx2.5mor3mx

3m andwer emounte d in aheate d greenhouse (18 °C)fo r G-j^an dG 2,an d outdoors forG. .Th eplant s were raised inpot s of about2, 5 litres and were transplanted into thecage sbefor e flowering.Th emai n flowering period was from February toJun e inG ,an dG _ from June toAugus t inG 3. Four fertility traits,i.e .polle n stainability, number ofsiliqua e perplant ,numbe r of seedspe rplan t and seedproductio npe rsiliqua , were studied inth e firstthre e generations ofth epopulations .Polle n stainability wasusuall y assessed twice ata time interval of several weekswit h lactophenol acid fuchsinb y themetho d ofSas s (1964). Pollen preparations from one flowerwer euse d forG. .; abou t 500polle n grains were examinedpe r flower. InG andG ,a rando m sample of20 0 grains was examined inbulke dpolle n from about fiveflowers . The remaining fertility traitswer emonitore d atharvest ,agai n for eachplan t separately. InG. .an dG_ ,al l siliquae developed wereharves ­ ted, counted and shelledb yhand .Al l seedswer e thencounte d andth e number ofseed spe r siliquawa scalculated .Th e sameprocedur ewa suse d forG _ whenth enumbe r of siliquaepe rplan twa s less than 100.I fmor e than 100siliqua ewer epresen t inG 3,al l siliquaewer eweighte d and seed productionpe r siliquawa s estimated ina rando m sampleo f 100siliquae , whichwa s alsoweighte d inorde rt oestimat e the totalnumbe r ofsiliquae . Chromosomes werecounte d atmeiosi so rmitosi s ofG . andG ~plants .

7.3 RESULTS

7.3.1 Chromosome numbers

The chromosomes of4 6G^ ^plant s and 118G -plant s from theP H andC H populations were counted (Table 30).I nG, ,th echromosom e numbers ranged from 35t o3 9an d amongG _plant s from3 4t o 58.Bot hpopulation s inG_ , had aremarkabl ebimoda l frequency distribution forchromosom e number withmaxim a at3 8an d 57.Mor etha n 12% ofG plants classified had over 50chromosomes . InG_ ,th e frequency ofeuploi dplant s (2n= 3 8o r 57) was higher than inG .bu tth eproportio n ofplant swit h 38 chromosomes hadhardl y increased.Th enumbe r ofhypoploi d plants,especiall y inth e PHpopulation ,wa smuc hhighe r thanth enumbe r ofhyperploi d plants,es ­ pecially inG 2. The frequency ofchromosom e numbers inoffsprin g ofal lseed-produc ­ ingG ,plant s fromth e twopopulation s studied isliste d inTabl e31 .G . plantswit h 37chromosome s gavepredominantl y euploidplant s (2n =38), like theeuploi d G-j^plants .Nearl y allhypoploi d plants inth e half-sib progenies ofeuploi dG ]Lplant swer ederive d fromPH1 0 and PH12 (Table 31). Thechromosom e number inoffsprin g ofG .plant swit h 39chromosome s was quitevariable .Mos t aberrantP Hplant swer e derived from thehybri d PH11,whic h originated from acros s onth e same Brassica planta sPH1 0 and PH12, twohybrid s thatshowe d acomparativel y high frequencyo funi ­ valents.

Table 30.Frequenc y of chromosome numbers in G and G of the PH and CH population.

Genera­ Popula­ Number of Chromosome numb er tion tion plants 34 35 36 37 38 39 40 41 43 45 52 54 56 57 58

CH 10 3 5 2 Gl PH 36 1 1 11 16 7 CH 32 1 2 4 4 2 G2 15 1 1 2 FH 86 1 5 9 12 47 1 2 1 1 1 5 1

59 Table 31. Frequency of chromosome numbers in half-sib families of G plants with various numbers of chromosomes in the PH and CH population.

Seed parents (G plants) Chromosome number G. plants" Number of G plants Chromo­ Popula­ Number 34 35 36 37 38 39 40 41 43 45 52 54 56 57 58 some tion of number plants

37 CH 1 1 1 2 PH 3 2 3 11 16 38 CH 4 19 3 1 1 15 PH 11 4 6 3 28 1 13 46 (3) (6) (3) (1) (1) (1)

39 CH 1 3 1 11 3 1 10 PH 2 1114 7 11 1 2 1 20 (1) (1) (1) (1) (2) (1) (2) (1) unknown CH 1 12 11 PH 2 2 1 1

']I nbracket s the number of plants from PH10,PH1 1 and PH12.

Ina fe whalf-si b families,G „plant swit h 52-58 chromosomes were found.O fth e 25 families raised, only 7gav e rise to such aberrant plants. The occurrence ofhexaploi d andhighl y aneuploidplant s inG _ seemed a serious threatt oth echromosoma l stability ofth eP H andC Hpopula ­ tion.Therefore , allG _plant swit hmor e than3 9 and less than3 7chro ­ mosomeswer e excluded from the twopopulations . Some aneuploids,however , were noteliminate d because their chromosome numberwa s not established properly before transplantation into the isolationcages . InG ,o fth eP H andC Hpopulations ,n ohexaploid s weredetected :mor ­ phological observations,measurement s ofpolle n size and checks ofchro ­ mosome number ina fewsuspect s revealed nohexaploids .

7.3.2 Changes in four fertility traits

Changes inpolle n stainability, setting of siliquae,see d production and seed-setpe r siliquawer e studied inth e firstthre e generations of theP Hpopulatio n (Fig.11 )an dC Hpopulatio n (Fig.12) .Th e results wereexpresse d ascumulative-frequenc y distributions.Th e actual number ofplants ,o nwhic h the graphs arebased ,ma y differ slightly from each otherbecaus e ofmissin g data andbecaus e a fewplant s diedprematurel y and soprovide d no data.Th ethre egeneration s ofth etw opopulation s were successively studied in aperio d of threeyears .S odifference sbe ­ tween generations canb e attributed to some extentt oyea r influences. Theplant-to-plan tvariatio n inpolle n stainability wasver y large

60 Cumulative frequency Cumulative frequency (%) 100 100

o—o G1 (N=47) G2(N=64) D—D G3(N=178)

0 10 20 30 40 50 60 70 80 90 100 100 1000 10000 pollen stainabilityC/o) Number of siliquae per plant

Cumulative frequency Cumulative frequency C/o) 100 100

(N=27) G1 ( N=47) (N=49) A 62 ( N =64 ) (N=170) D G G3( N= 180 )

—Klog)0 l-^V-- 10 100 250 0 1 ,°9) Number of seeds per 100 siliquae 10 100 1000 10000 Number of seeds per plant

Figure 11. Cumulative frequency distributions of various fertility traits in the first three generations of the PH population of xBrassicoraphanus. A. Pollen stainability (%). B. Number of siliquae per plant. C. Number of seeds per 100 siliquae. D. Number of seeds per plant. The population means are marked with an arrow. N = number of plants in the generation.

(Fig. IIA and 12A). About 35 %o f G1 plants of the PH population pro­ duced no stainable pollen at all. This male-sterility had various causes, such as the absence of pollen shedding or even of flowers in a few plants. In the subsequent generations, the proportion of male-sterile plants grad-

61 frequency Cumulative frequency (%) 100

(log) O 10 20 30 40 50 60 70 80 90 100 100 1000 10000 Pollen stainability (%) Number of siliquae per plant Cumulative frequency Cumulative frequency (%) (7=) 100 100r

G1 (N =1 4 ) "G2 ( N= 2 1 ) G3( N--25)

(log) 10 100 250 100 1000 10000 of seeds per 100 siliquae Number of seeds per plant

Figure 12. Cumulative frequency distributions of various fertility traits in the first three generations of the CH populations of xBrassicoraphanus• A. Pollen stainability (%). B. Number of siliquae per plant. C. Number of seeds per 100 siliquae. D. Number of seeds per plant. The population means are marked with an arrow. N = number of plants in the generation.

ually decreased and the mean pollen stainability of the PH population im­ proved from 24 % in G. to 49 % in G_. The CH population had a higher pollen stainability initially, mainly because of fewer male-sterile plants; there was no improvement in subsequent generations; the G„ even showed a decline.

62 Onth ewhole ,polle n stainabilitywa s highlyvariable . Ina separat etest , a coefficiento fvariatio n ofabou t3 0% wa s found.S odat amus tb einter ­ preted cautiously. TheP Hpopulatio n alsoshowe d anincreas e inth ethre e successivege ­ nerations inth enumbe r ofsiliqua eharveste dpe rplan t (Fig. IIB). The proportion ofplant swithou t anysiliqua e atal ldecrease d considerably. The samewa s true forth eC Hpopulatio n (Fig.12B) .Th eG^ ^plant s ofth e PHpopulatio nproduce d only few siliquae.Howeve r astrikin g improvement occurred inth enex ttw ogeneration s by areductio n inth eproportio no f plantswithou t siliquae andplant swit h few siliquae.Th emaximu mproduc ­ tion inth e respective generations was 1482,138 6 and 1947 siliquaepe r plant.Th edat a onth eC Hpopulatio n areles s indicativebecaus e fewer plantswer e available ineac hgeneration .However ,G .showe d aseriou s drop inproductio n of siliquaepe rplant ,thoug hrecoverin g to levelo f G. followed inth enex tgeneration . Themos tcrucia l traiti sundoubtedl y seedproduction .Th eP H andC H populations hadver y low fertility inG .an dproduce d only 6.5 and 5.6 seedspe rplant ,respectivel y (Fig.11 D and 12D,respectively) . Inth e following generations ofth eP Hpopulation , aconsiderabl e improvement occurred, theG -plant sproducin g about 72seed spe rplant .Th e frequen­ cydistribution s for seedproductio n inth eP Hpopulatio n showeda striking reduction inth eproportio n ofcompletel y seedlessplant s in successive generations.Th e lastgeneratio n showed amarke d shiftto ­ wards ahighe rproductivity .Th e CHpopulation ,o nth eothe rhand , hada setback insee dproductio n inG _paralle l toth edecreas e inproductio n ofsiliquae ,bu tth eresult s ofG _wer emuc hbetter ,eve ntha nG. . The frequency distributions ofseed-se tpe r siliqua differed only slightlybetwee nG .an dG _ ofth eP Hpopulation ; about 0.05 seedwer e obtained inbot hgeneration s (Fig.11C) .Th eG_ ,however , showeda striking improvement. Someplant seve nha d aproductio n ofove ron esee d per siliqua, andth e frequency ofplant swit h only seedless siliquae further decreased.Th e range of seed-setpe r siliqua inth e threesucces ­ sivegeneration so fth eC Hpopulatio nwa s aboutth esam e asi nG .o fth e PHpopulation .Afte r asetbac k in seed-set inG _b y increase inth enumbe r ofplant swit h seedless siliquae,th e seed-set inG ,greatl y improved. Insummary , seedproductio n inth eP Hpopulatio n improvedwit ha nin ­ crease insiliqu aproductio n and seed-setpe r siliqua.Th e two traits were only slightly correlated (r= 0.1 7 in G,), butwer ebot hpositivel y correlated with seedproductio n (r= 0.5 3 andr =0.7 0 inG, ,respective ­ ly). Similar relationships were found inG _ ofth eC Hpopulation .Th eim ­ provement in fertility ofthi spopulatio nwa sbrough t aboutentirel yb y an improved seed-setpe r siliqua.Th e fourth fertility trait,polle n stainability, was independent of seedproductio n (r= 0.2 3 and r= -0.0 7

63 7.3.3 Effects of chromosome number on fertility

Chromosome numbervarie d inth eG , andG _ ofth e twopopulation s of xBrassicoraphanus studied.Th e impact ofthi s variation onvariou s fer­ tility traits issummarize d inTabl e 32.Euploi d plants had, onaverage , more stainablepolle n andproduce d considerably more siliquae and seeds than aneuploids.A n exceptionwa s the class ofplant s with 36chromosomes . Sixo fthos eplants ,however ,wer e derived from the sameG .plant ,whic h had acomparativel y good fertility. Soth e rather good fertility ofthi s classwa sprobabl y not causedb y adirec t effecto fchromosom e number on fertility. Inconclusion , aneuploidy had asmal lnegativ e effecto n all fertility traits. So selection for fertility favours euploid xBrassicoraphanus plants. Cytological selection foreuploid y seemsno tnecessary .

7.4 DISCUSSION

Thepopulation s responded fairly strongly tonatura l selection forim ­ proved fertility. TheP Hpopulatio n responded with agradua l improvement inal l thetraits ;th e appearance ofth eplant s improved too,resultin g inG ,o fgreate r uniformity, abundant flowering, and only afe wplant s of abnormal form.Th e CHpopulation , onth eothe r hand, showed less obvious trends. InG ,a nobviou s loss ofvigou rwa s accompanied by aseriou s setback in all fertility characteristics.Th epopulation ,however ,reco ­ vered to some extent inG_ . A largevariatio n in all fertility traits studied was stillpresen t inG- .o fbot hpopulations .Polle n stainability and fruit settingvarie d from zero to almost 100% . Incontrast , seed-setpe r siliquawa s always farbelo wpotential . InG, , for example,th ebes t seed-set observed was about1 an d 0.3 seedpe r siliqua inth eP H and CHpopulation ,respect -

Table 32.Effect s of chromosome number on variation in some fertility traits in G and G of the XBrassicoraphanus populations PH and CH.

Characteristic Chromosome n imber

35 36 37 38 39 40 41 45

Number of plants* 1 7 27(2) 80(2) 14 1 1 1 Mean pollen stainability 15 57 28 41 32 13 0 19 Number of siliquae per plant 3 284 75 159 120 0 4 13 Number of seeds per plant 0 9 7 2.5 13.4 8 1 0 0 0 Number fraction of plants with siliquae(% ) 100 86 56 81 57 0 100 100 Number of seeds per 100 siliquae 0 4 1 3.6 5.8 5 7 0 0 0

'*]I nbracket s the number of plants,whic h died before maturity.

64 ively,wherea s thenumbe r ofovule spe r siliquawa s about 19.Th e large variation insettin g of siliquae and ofseed spe r siliqua inG , suggested that succeeding generations would furtherimprove . Prerequisite forcommercia l use ofx Brassicoraphanus isa stabl echro ­ mosome number. InG. .an dG_ ,mos tplant s had about3 8chromosomes ,bu t severalplant sha d anaberran tnumber .Th emos tstrikin gphenomeno nwa s the occurrence ofhexaploi d plantswit h about 57chromosome s insom eG _ families frombot hpopulations .Suc hplant sprobabl y arose from fusion of anunreduce d and areduce d gamete,resultin g inplant swit h thegenom e constitutionAAARRR .Th e occurrence ofunreduce d gametes indicates that ^Brassicoraphanus may form unreduced diplosporic orapospori c embryosacs as xRaphanobrassica (Ellerstrom& Zagorscheva , 1977). Thehexaploi d G_plant s usually showed only rather small differences in appearance compared toth e tetraploid counterparts,suc h aslarge r flowerbud s and flowers, andusuall y had somewhat larger stainablepollen . InG 3,n o hexaploidswer e detectedb yvisua l screening andb ymeasurin gpolle nsize . The occurrence of suchhexaploid s couldbadl y affectchromosoma lstabil ­ ity innewl y created populations.Th e deterioration and subsequentextinc ­ tioni nEllerstróm' smateria l (Olsson& Ellerstrom ,1980 )migh thav ebee n causedb ynon-reduction . Theothe rimportan taspec to fchromosoma l stability,th e aneuploidy, was already present inG, . Intha tgeneration ,th eoccurrenc e of aneuploid plantswa s largely enhancedb y theus eo fautotetraploi d formso fth epar ­ ental species inth eproductio n of xBrassicoraphanus. InG~ ,mos tplant s derived frommonosomi c andtrisomi eG- .plant sturne d outt ob eeuploid , indicating thatbalance d gametes andzygote s were fitter and had aselec ­ tiveadvantage .Howeve r several aneuploids were still found inG_ ,proba ­ blybecaus e of formation ofunivalent s inmeiosis . Insummary , sterility inearl ygeneration s of y-Brassicoraphatms ispre ­ dominantly causedb yvariou sbreedin gbarriers ,whic hwer e geneticallyde ­ termined.Meioti c irregularities likenon-reductio n and desynapsisproba ­ blyha d no effecto n fertility assuch ,bu tma y give rise to aneuploid offspringwit h reduced fertility.Howeve r non-reductionan d desynapsis seem of littleconsequence .

65 8 AAR hybrids and introduction of alien chromosomes into Brassica campestris

8.1 INTRODUCTION

Introgression of Raphanus genes into B. campestris hasno tye tbee n accomplished. The aim ofth epresen tprogramme ,wa s toproduc e monosomic additions in B. campestris, eachplan thavin g thenorma lcomplemen t of chromosomes plus anextr achromosom e from R. sativus. The expectation was thatsuc hplant swoul d serve assteppin g stones fortransfe r ofpar ­ ticular Raphanus genes,e.g . resistance tobee t eelworm ort oclub-root . Themetho d chosen forproductio n ofsuc h additions included twosteps : production ofAA R hybrids andbackcrossin g with B. campestris (AA)a s recurrentparent .Thi s chapter reports attempts toachiev e the aim and discusses theproblems ,tha tma yhampe r successful transfer of Raphanus genes into B. campestris.

8.2 MATERIALAN D METHODS

The attempts toproduc eAA R hybridswer e summarized inTabl e 20.I n total, 19 firstbackcros s hybrids (BC.) includin g 4plant sproduce db y embryocultur ewer e obtained from crosses oftw oprimar yAAR Rhybrids , 75.1040-9 (2n= 38 )an d 76.2229-13 (2n =39 )wit h B. campestris (2x). The allotetraploid, 75.1040-9,originate d from doubling invitr o ofth e chromosome number ofa diploi d hybrid from acros sbetwee n B. campestris ssp. chinensis (2x)an dcv .Silett a (2x).Th e other allotetraploid, 76.2229-13,wa sobtaine d from acros sbetwee n B. campestris ssp. perviri- dis (4x)an d cv.Palet .Th eAA R hybrids were studied for form, chromosome number, fertility incrosse swit h B. campestris ssp. chinensis (2x)an d inparticula r forchromosom e association inmeiosis . TheAA R hybrids and B. campestris ssp. chinensis (2x)wer e crossed in thewinter s 1976/77 and 1977/78 ina heate d greenhouse, inwhic h all parentplant swer e growntogether . Inth e firstwinter ,newl y opened flowers ofth eAA R hybridswer ehand-pollinate d withbulke d pollen from therecurren tparent , B. campestris, two or three times awee k overa period ofabou tthre emonths . Inth e secondwinter , theparent swer e grown together ina nisolatio n cagewit hbee s forpollination . The secondbackcros s generation (BC_)wa s grown ina simila rwa yt o theBC ,generatio n in1977/78 ;i na nisolatio n cagewit h the recurrent parent andwit hbee s forpollination .

66 Figure 13.Backcros sderivative so fa 'doubled'F hybrid fromth ecros sBrassic a campestrisssp .chinensi s (2x)x Raphanu ssativu scv .Silett a (2x)an dB .campestri s ssp.chinensi s (2x)a sth erecurren tparent . A.Firs tbackcros sderivativ e (AAR,2 n= 29 )i nth evegetativ estage . B.Firs tbackcros s derivative (AAR,2 n= 29 )i nth egenerativ estage . C.Secon dbackcros s derivative (2n= 2 5plu s fragment).

8.3 RESULTS

8.3.1 Form and chromosome association in AAR hybrids

Of 12 BC-. plants, ten plants had 29 chromosomes and were apparently true allotriploids with the genome constitution AAR. The two remaining plants had 28 chromosomes plus a fragment and 31 chromosomes, respec­ tively. The latter occurred in the progeny of plant 76.2229-13.

67 TheBC- .plant s derived from crossesbetwee n 75.1040-9 (4x)an d B. campestris ssp. chinensis or ssp. oleifera grew quitevigorousl y and were inman y respects intermediate between theparents .Al lplant s formed aleaf-rosett e inth evegetativ e stage andha d shallowly lobed and slightly pubescent leaves (Fig.13A) .Th e incisions ofth e leaves weremor epronounce d inplant s from crosseswit h B. campestris ssp. oleiferea. Inth e generative stage,th eBC . plants initially had an erectmai n stemwit h atermina l inflorescence (Fig.13B )an dreache da heighto fabou t1. 5 m.Afterward sman y secondarybranche swer e formed, resulting in adens ebush yplant .Th ehybrid s derived fromplan t 75.1040-9 (4x)flowere d abundantlywit hplai nwhit e flowers likethos e ofplan t75.1040-9 . Incontrast ,th ethre e BC.plant s originating fromplan t 76.2229-13 (4x)showe d disturbed growthwit h irregularly shaped leaves.Onl y oneo f theseplant s flowered; the flowerswer eveine d andpurple . Noneo f theAA R hybridsproduce d stainablepollen , the anthers were already shrivelled beforepolle n shed.Th edevelopmen t of floralpart s was apparently quitenormal ,althoug hpoorl y developed ovules wereob ­ served ina fewhybrids .Afte rpollinatio n withpolle n from B. campes­ tris, afe w siliquae developed; theywer ebivalve d with awel l developed beak andha d amea n length of about2. 6 cm. Chromosome association atmetaphas e Io fmeiosi swa s studied inthre e closely related allotriploid plants (Table 33). Thepresenc e of atriva ­ lent ina fe wpolle nmothe r cells andth e largeproportio n ofpolle n mother cells (86% )wit h1 0bivalent s and 9univalent s (Fig.14A )indi ­ cate thathomologou s pairing occurred almostexclusively .Howeve r the occurrence ofpolle nmothe r cellswit h elevenbivalent s showstha tsom e autosyndetic pairingbetwee nR-chromosome s occurred (Fig.14C) .Th e pres­ ence of afe wchain-trivalent s further suggested that allosyndetic pairing betweenA andR chromosome s waspossibl e (Fig.14B ;Tabl e33) .

Table33 .Chromosom eassociatio na tmetaphas eI i nthre e AARhybrid s (2n= 29 )fro mcrosse sbetwee na homozygou s xBrassicoraphanusplan t (75.1040-9 (4x))an dBrassic a campestrisss p chinensis(2x) .

Chromosome Number of cellsobserve d configurations^- Hybrid 1 Hybrid 2 Hybrid 3

11I + 9I I 3 1 0 9 I+ 10I I 57 34 29 8 1+ 9I I+ 1II I 1 2 0 7 I+ 1 1I I 7 3 2 Total 68 40 31

'']I ,I Ian d IIIar esymbol sfo runivalents ,biva ­ lentsan d trivalents,respectively . } A fr. B

WL

^ ,. • *•

• 0 40

• • # * JL »

Jife

- # I

Figure14 .Chromosom eassociatio na traetaphase Ii nAA Rhybrid sfro ma cros sbetwee n a'doubled 'A Rhybrid ,76.1040- 9(4x) ,an dBrassic acampestri sssp .chinensi s (2x). A.1 0I I+ 9 I (X1540) . B.1 II I+ 9 I I+ 8 I (X1540) . C.1 1I I+ 7 I (X1090) .

The differences inchromosom e associationbetwee n thethre eAA R hybrids weresmall .On ereaso nma yb etha tth eplant sha d anidentica l seto f R andA chromosomes .

8.3.2 Attempts to introduce alien chromosomes

In197 7an d 1978,thousand s of flowers ofth eAA R hybridswer epollin ­ atedb yhan d orb ybee swit hpolle n from B. campestris (2x).Th e results ofthes e crosses,i.e . BC~, are summarized inTabl e 34.Mostl yn o siliquae

69 Table 34. Seed-set on several AAR hybrids using Brassica campestris ssp. chinensis (AA)a s pollinator.

Origin of Year Number of Total number AAR hybrids AAR hybrids of seeds

75.1040-9 (4x)x 1977 n 5 B.c. ssp. chinensis 1978 2 0

75.1040-9 (4x)x B.c. ssp. oleifera 1978 3 27

76.2229-13 (4x)x B.c. ssp. chinensis 1978 3 0

oronl y afe wdevelope d oneac hplant ,excep t onon eAA R hybrid derived from acros sbetwee n 75.1040-9 (4x)an d B. campestris ssp. oleifera. De­ spiteconsiderabl e efforts,onl y 32seed swer eproduce d from sixo fth e plants only.

The BC2 generation consisted ofeleve nplants ,o fwhic h three died beforematurity . In1978 ,anothe r three BC_ plantswer e raisedb y cul­ ture of2 6 embryos invitro .Th e chromosomes ofmos t full-grownBC _ plantswer e counted (Table 35).Mos tha d afe wchromosome s less thanth e original AAR hybrids,bu t twoplant s had considerably morechromosomes . Son omonosomi c additionswer e obtained.Th e ESC,,plant swer e usually weak and flowered poorly.Al lplant swer emal€:-steril e andha dwhit e flowers, excepttw oplant swit hpal eyello w flowers.Figur e 13C shows theonl y BC? plantwit h 25chromosome s (plus a fragment).Thi splan tha d rather irregular growth, andmos t Raphanus characteristics were apparent­ lylost . All BC_ plantswit h 29o r fewer chromosomes werebackcrosse d againwit h B. campestris ssp. chinensis (2x).Despit e large-scale pollinations,non e ofthes eplant s producedseeds .

Table 35.Frequenc y of chromosome numbers in full-grown plants obtained from crosses between AAR hybrids and Brassica campestris (2x) in 1977 and 1978.

Year Number of p lants with 2n = Total number 25+f * 26 27 28+f 29 44 48 9 of plants

1977 1 1 1 1 4 1978 1 3 1 2 7 Total 1 1 3 1 1 1 1 2 11

f - fragment 8.4 DISCUSSION

Theintroductio n ofsingl e Raphanus chromosomes into B. campestris pro­ ved asye t impossible.M ydifficultie s inobtainin g aBC 2 generationwer e similar tothos eo fBannero te tal . (1974)an dMcCollu m (1979)i nattempt s tocros sCC Rhybrid swit h B. oleracea (CC).M yproblem smigh tb e overcome byusin gmor eAA R hybrids and awide r rangeo fpollinators . InBC, , thevariatio n inchromosom enumbe rwa s large,a sother s found forCC Rhybrid s (Bannerote t al., 1974;McCollum , 1979). Even repeated backcrossing by those authorswit h B. oleracea asrecurren tparen tdi dno t resulti nmonosomi c additions.Bannero te tal . (1974)obtaine d sucha n additionb y culture invitr o ofimmatur e embryos.Th e same approachma y benecessar y toobtai n additions in B. campestris. Analysis ofchromosom e associationdurin gmeiosi s inAA R hybrids sho­ wed thatonl yhomologou sA chromosomespaire d asa rule .Allosyndeti c and non-homologous autosyndeticpairin gwa srare .Th epresen tdat aconfirme d thehypothesi s ofMizushim a (1980)tha tth eR genom ema ygiv eris et oon e autosyndetic bivalent.Th e frequencywa s about0.08 6bivalent s per cell. Thedegre e of allosyndeticpairin gwa spresumabl y evenlowe rbu t atleas t oneR chromosome seemed sometimes tob eboun d to anA chromosome.Thi s agreeswit hth e findings onchromosom e association ina nAR Rhybri db y Mizushima (1980). Comparisono fchromosom e association inA R andAAR R hybrids indicates thatth edegre e ofallosyndesi s inth eAA R hybridswa s reduced,becaus eo fpreferentia l pairingbetwee nhomologou sA chromosomes inth ehybrid s (Tables 14an d 33). Preferentialpairin gprobabl y affected especially thedegre e ofallosyndesis ,becaus e thedegre e of autosyndesis withinth eA genomewa s already small (Mizushima,1980 )an dther ewa sn o reason for adecreas e inautosyndesi swithi nth eR genome . Inconclusion ,transfe r of Raphanus characteristics to B. campestris by themetho d chosen isver y difficultan dtime-consuming ,becaus e of difficulties ofintroducin g singleR chromosome s into B. campestris and theexpecte d lowdegre e ofrecombinatio nbetwee nR andA chromosomes .

71 9 Ways of transferring Raphanus characteristics to Brassica napus

9.1 INTRODUCTION

Direct transfer of Raphanus characteristics to B. napus by crossing the two species andbackcrossin g theF .hybrid swit h the latter isdif ­ ficultbecaus e ofth epoo r crossability between these species (Chopinet, 1942; 1944;Fukushima , 1945;Becker ,1951 ;Heyn , 1973;Valdivi a &Badilla , 1977;McNaughto n &Ross , 1978;Rousselle , 1979;Takeshit a et al.,1980 ; my results). Some authors,however , obtained afe whybrids .Takeshit a et al. (1980)culture d several ovules from reciprocal crosses invitr o and raised oneplant .Crossabilit y couldb e somewhatbetter , if tetraploid formso f R. sativus were used (Turesson& Nordenskiöld , 1943;Chopinet , 1944). Hybrids from suchcrosses ,however , appear less suitable as startingmateria l fortransfe r of Raphanus genes,becaus e ofpreferen ­ tialpairin gbetwee nR chromosomes andprobabl y apoo r crossability with B. napus. Alternative ways for introgression arepossibl e with thevariou sin - tergeneric hybrids from crosses between Brassica species and R. sativus, i.e. allopolyploidswit hth egenom e constitution CCRR,AARR , orAACCRR . An attractivewa y isa backcros s programme of xRaphanobrassica (CCRR) with B. napus, because xRaphanobrassica hybridizes readilywit h that species (Karpechenko, 1937;Rousselle , 1979). Inthi sway , Stefansson (Winnepeg,Canada ,cite d byHeyn , 1978)ha s succeeded in transferring thewhite-flowe r traito fradis h torape .Simila rwor k isi nprogres s in Germany (Heyn,1978 )an d France (Rousselle, 1979)t o transfer fertility- restorer genes from Raphanus to rapewit h cytoplasmic male sterility. Heyn (1978)als o crossed suchmale-steril e rape;wit h an allohexaploid withth e genome constitution AACCRR, thatha dbee nproduce db yChopinet . As xBrassicoraphanus seemed agoo dbridg e fortransfe r of Raphanus characteristics to B. napus, abackcros s programme with B. napus as recurrentparen twa s setu pwit h special attention to form, fertility and chromosome association atmeiosi s ofhybrid sbetwee n xBrassicoraph- anus and B. napus.

9.2 MATERIAL AND METHODS

Initialmateria l for thebackcros s programmewa s 24male-steril e hybrids from crosses of xBrassicoraphanus (perviridis and chinensis

72 hybrids)wit h 'Zephyr', aCanadia n rapevariet y (Table 20).Befor e flowering, theseplant swer e transplanted into two isolationcage s in a heated greenhouse, inwhic h continuously floweringplant s of 'Zephyr', the recurrentparen t inth ebackcros sprogramme ,wer epresent .Honey ­ beeswer e kepti nth ecage s forpollinatio n and soman y flowers ofth e hybridplant swer epollinate dwit h 'Zephyr'pollen .Som eembryo swer e raised invitr ob y themetho d ofHarber d (1969,1971 )i norde rt o shorten thelife-cycle . The firstbackcros s generation,partl y raisedb y embryo culture,wa s grownunde r isolation ina heate d greenhouse togetherwit h 'Zephyr'.Abou t 100hand-pollination s were doneo neac hplan tt oobtain d asecon d backcross generation.

9.3 RESULTS

9.3.1 Hybrids from crosses between xBrassicoraphanus and Brassica napus

Table 36 summarizes the data on somatic chromosome numbers inhybrid s from reciprocal crossesbetwee n xBrassicoraphanus and rape.Plant swit h 38chromosome s presumably had the expected genome constitution AACR and are called AACR hybrids.A small group ofplant s from crosses with B. napus as female, however, had about 57chromosome s andprobabl y had the genome constitution AAACCR. Also aplan twit h 31 chromosomes was found, whose originwa s notclear . MostAAC R hybrids grewvigorousl y andha d abalance d habit ofgrowth . Theplant s exhibited many characteristics ofrape , such as anobviou s

Table36 . rape 'Zephyr'an dsettin go fsiliqua ean :idseed so fth eF after pollinationwit hrap e 'Zephyr' (BC,). 'V Originó f Chromosome Numbero f BC F hybrids numberi n F plants F, Numbero f Numbero f Numbero f seed-pro­ siliquae seeds ducing harvested plants perviridishybrid sx 31 1 0 0 1 Zephyr ' 38 14 129 44 1 142 52 chinensishybrid sx 38 3 61 24 'Zephyr ' ? 1 63 37

'Zephyr'x chinensi s 38 1 378 57 hybrids 57 1 339 102 58 2 0 0

73 Figure 15.Hybrid s from crosses between Brassica napus and AARR hybrids. A. Offspring of crosses between rape cv. Zephyr and chinensis hybrids in the vegetative stage. Top row from left to right: rape;hybri d with 2n = 38;hybri d with 2n =57 . Bottom row from left to right; two hybrids with 2n = 58;rape . B. AACR hybrid in the generative stage.

stem inth evegetativ e stage and leaves thatwer e dull greenbecaus e of amarkedl y waxy layer (Fig. 15A).Mos tAAC R hybrids grew rathertal l (upto 2m )an d flowered abundantly (Fig.15B) ,bu t a fewdi d not flower atall .Th e flowers ofth ehybrid s were somewhat larger than in xBrassi- coraphanus andwer ebrigh t andwhite .Marke dveins ,a s in xBrassicorapha- nus, were absent.A few flowers had ayello w sector onth epetals .Th e siliguaeo fAAC R hybridswer ebivalved , as in Brassica, andha d alengt h of about3 cm , including abea k of about1. 5 cm.Th e aberrantplan twit h 31chromosome s remained small,gre w irregularly anddi dno t flower at all. The small group ofhexaploid s consisted ofon ewel l balanced andtw o irregularly shaped plants (Fig. 15A).Th e former had 57chromosomes ,wa s similar to arelate d AACR hybrid and flowered abundantly. The two other plants had 58chromosomes , floweredpoorl y andha dplai nwhit e flowers, although oneplan t alsoproduce d some yellow flowers.Onl y the hexaploid with 57chromosome s yielded siliquae,whos e size and shapewer e similar to those ofth eAAC Rhybrids . All F, hybrids were completely male-sterile and, after pollination

74 withpolle n fromrape ,se tsiliqua e and seedpoorly : abouthal fth eAAC R hybrids andon ehexaploi d yielded seed (Table 36). However aconsiderabl e numbero fseed swa s obtained,becaus e ofth e large-scalepollinations . Mostproductiv e were thetw oF .hybrid swit h 'Zephyr'a s femaleparent . Thetota lnumbe r ofsiliqua eharveste d and ofseed spe rhybri dwer e somewhatreduced ,becaus e 53 siliquae ofAAC R hybrids and4 0 siliquae of thehexaploi dwit h 2n =5 7wer euse d forembry o culture.Th e respective numbers ofembryo s obtainedwer e4 9an d 17.Mos tembryo sha dreache d the walking-stick stage (Wilmar& Hellendoorn , 1968)o rwer ematur e already. Intotal ,2 1 and 5plants, respectively ,wer e obtainedb y embryoculture .

9.3.2 Chromosome association in AACR hybrids

Meiosis inAAC R hybrids (Zn =38 )wa s analysed in afe wpolle nmothe r cells (Table 37).I fonl yhomologou s pairing were tooccur , tenbivalent s (twoA genomes)an d 18univalent s (oneC an don eR genome )woul db eex ­ pected.Th e degree of association, however, was higher, themea n numbers per cell ofbivalents , trivalents and quadrivalents being 10.59, 0.75 and 0.02, respectively. Consequently thenumbe r ofunivalent s percell , 14.48, was less than expected. The number ofchromosom e configurations exceeded 10 in8 7% of allpolle nmothe r cellswit h amaximu m of 13biva -

Table37 .Chromosom eassociatio n inmetaphas eI o ftw oAAC R hybrids (2n= 38 )fro mcrosse sbetwee nxBrassicoraphanu san d rape.

Chromosome Numbero fpolle nmothe rcell s configurations* Plant1 Plant2 Total

18I + 7I I+ 2 II I

18I + 10 II 0 1 17I + 9 II + 1 III 1 0 16I + 8 II + 2 III 1 0 16I + 9 II + 1 IV 1 0 15I + 7 II + 3 III 0 1

16I + 11 II 8 1 9 15I + 10 II + 1 III 4 2 6 14I + 9 II + 2 III 0 2 2 13I + 8 II + 3 III 0 1 1

14I + 12 II 5 4 9 13I + 11 II + 1 III 2 3 5 12I + 10 II + 2 III 0 3 3

12I + 13 II 2 0 2 11I + 12 II + 1 III 0 1 1

Total 25 19 44

'•]I ,II ,II Ian d IVar esymbol sfo runivalents ,bivalents , trivalentsan dquadrivalents ,respectively .

75 lents andmultivalents .Thi snumbe r indicates thedegre e of autosyndesis and allosyndesiswithi n andbetwee n the Can dR genome , respectively. Thenumbe r oftrivalent s ina cel l ranged up to3 .A n increasing number ofmultivalent s ina cel lwa s associated with areductio n innumbe r of bivalents intha tcell . Somos ttrivalent s probably resulted frompairin g between one Co rR chromosom e andon epai r ofA chromosomes .

9.3.3 First backcross generation

A fewB C plants derived fromAAC R andAAACC R hybridswer e raised, partlyb y embryo culture (Table 38).Bot h types ofhybri d gave riset o highlyvariabl e offspring. Someplant s showed slight abnormalities,suc h asdeforme d leaves,bu t all flowered. The flowerswer eusuall ywhite , although inth e firstbackcros s generation derived fromAAC R andAAACC R hybrids somewer eyello w (Table 38).Als owhit e flowerswit h ayello w sector occurred occasionally inth e firstbackcros s generation, even more frequently than inth eparen thybrids . Apromisin gphenomeno nwa s the occurrence ofsom epartl y male-fertile plants inth e firstbackcros s generation (Table 38).On eplant ,derive d from theonl y fertileAAACC R hybrid, evenha d apolle n stainability of 59% .Th e seed seto fth e firstbackcros s generationwa s low incrosse s withrape .

9.4 DISCUSSION

Introgression of Raphanus characteristics into rapeb y repeatedback - crossing of xBrassicoraphanus with rape asth e recurrentparen t isprob ­ ably feasible.Majo r drawbacks are low fertility in successive backcross generations, and rather infrequent associationbetwee n Raphanus and Brassica chromosomes.Th e adapted method for large-scale crossing ofAAC R

Table 38.Flowe r colour, pollen fertility and seed fertility in BC plants from about 100 hand-crosses of either AACR or AAACCR hybrids with Brassica napus 'Zephyr' and raised from seed or by embryo culture.

Cross Number of Mode of Flower CO lour Number of flowering culture plants white ye How plants plants with produ­ seeds some cing stainable seed pollen

AACR x AACC 15 embryo 12 3 2 5 20 14 seed 10 4 1 9 41

AAACCR x AACC embry o 75

76 andAAACC Rhybrid swit h rapeb y growinghybrid s andrap e together ina n isolation cageprove d tob eusefu l forproductio n of afirs tbackcros s generation ofa reasonabl e size,despit epoo r fertility ofF .hybrids . The infertility insuccessiv ebackcros s generations,whic hprobabl ywil l be less severe,ma yb e overcome ina simila rway . Successful transfer of Raphanus genes dependsmainl y onth epairin g abilitybetwee n Raphanus and Brassica chromosomes.Th echance s forpair ­ ingan dexchang ebetwee nR andA chromosomes inAAC Rhybrid s areprobabl y assmal l asi nAA Rhybrid sbecaus e ofpreferentia l pairingbetwee nhomo ­ logousA chromosomes.Th erathe rhig h frequency ofmultivalent s wasprob ­ ablymainl y due topairin gbetwee nA and Cchromosomes ,becaus e inAA C hybrids similar numbers ofmultivalent s occurred (Namai,1976 ;1978 )an d inAA Rhybrid s only afew .O nth e otherhand ,AAC R hybrids aremor e favourable forexchang ebetwee n Ran dC chromosomes .Th e frequency of pairingbetwee n suchchromosome s isdifficul t toestimat ebecaus e the degree of autosyndesis inth eR andC genom e isunknown .O n thebasi so f detailed genomic analyses,Mizushim a (1980)estimate d thatth emaximu m number ofautosyndeti c pairswa s two inth e Cgenom e and one inth eR genome.Th ebivalen t frequency inth eAAC R hybrids due to non-homologous autosyndesis wasprobabl y low,becaus e suchpairin g onlyoccur ssporadic ­ ally inAA R andAA Chybrid s (myresults ;Namai ,1976 ; 1978). Consequently mostpairin g involving Co rR chromosomes ,o n average atleas t1.3 4biva ­ lentspe rpolle nmothe r cell,wa sprobabl y causedb y allosyndesis.Th e number ofcell s examinedwa s too small to justify conclusions aboutth e maximum number of Raphanus chromosomes thatcoul dpai rwit h aC chromo ­ some.Howeve r theobservatio n ofu pt o 9bivalent s inC Rhybrid s indi­ cates that several Rchromosome s arecapabl e ofpairin g allosyndetically (Harberd &McArthur , 1980;Namai ,1976 ;1980 ;McNaughton , 1973).

77 10 Prospects

10.1RESISTANC E TOBEE TEELWOR M

The absence ofresistanc e tobee teelwor m in Brassica and itspres ­ ence insom e fodder radishvarietie s was themai nmotiv e forth epresen t researchprogramme .Littl ewa s known aboutth e inheritance andmechanis m ofth e resistance whenth eprogramm e started. Baukloh (1976)presume d simplemonofactoria l dominantinheritance . Toxopeus& Lubberts (1979)showe d that themultiplicatio n rateo fbee t eelwormwa s lowo nmos t forms of R. sativus. Raphanus populations respond­ edwel l to selection for resistance andToxopeu s and his colleagues at theFoundatio n forAgricultura l PlantBreedin g (SVP), Wageningen,hav e produced populations onwhic h thenematod emultiplie d much slower than onth e original stock (Lubberts &Toxopeus , 1982). Recent studieshav e showntha tresistanc e results from ashif ti nth e sex ratio of the eel­ worm in favouro fth emale s (Veerman, 1981)Th eeffec to fgrowin gpar ­ tially resistant fodder radishpopulation s onbee teelwor m populations inth e soil isbein g studied inth eNetherlands . Thepresen tpopulation s of xBrassicoraphanus also showed moreresist ­ ance than Brassica species.Th e cystproductio npe rplant ,however ,wa s highlyvariabl e and themultiplicatio n rate ofth eeelwor mwa s on average considerably higher than for the Raphanus parents.T o improve resistance, selection forresistanc e isproceedin g inthos epopulations .Th eexpe ­ riencewit h Raphanus suggests thatconsiderabl e improvement ispossible .

10.2xBRASSICORAPHANU S

Insynthesizin g ane w forage crop, xBrassicoraphanus, several problems arose sucha sth epresenc e ofbreedin g barriersbetwee n theparenta l spe­ cies,poo r fertility inth e firstthre e generations andoccurrenc e of aneuploid andhexaploi d plants.Extensiv e studies have showntha tsuc h difficulties hamper the rapid establishment of fertile andbroadl y based populations.Howeve rproductio n of suchpopulation s islikel y tob e pos­ sibleb y selection for improved fertility inth e availablebreedin gpopu ­ lations and resynthesis of ^Brassicoraphanus inorde r tobroade n thegene ­ ticbasis .Resynthesi s mightb e facilitated by exploitation ofth e genetic variation incrossibilit y in Brassica campestris and Raphanus sativus, use ofmale-steril e B. campestris plants incrosse s and improvement of embryo-culture techniques forsuc hcrosses . The agricultural value of x-Brassicoraphanus could notye tb e assessed, because thenumbe r ofseed s availabledi dno tpermi t field trials.Howeve r growthhabi t and life-cycle arereason s foroptimis m on suitability asa foragecrop .Howeve r yield and qualitative aspects likepalatability ,di ­ gestibility, crudeprotei nconten tan dpossibl e occurrence oftoxi ccon ­ stituentsmus tb e tested as for xRaphanobrassica (McNaughton, 1979). Insummary , xBrassicoraphanus mightbecom e aproductiv e foragecrop , highly resistantt obee teelworm ,powder ymilde w (Erysiphe cruciferarum) andt oclub-root . Inth e agricultural systems ofNort hWester nEurope , such acro pmigh t fitint oth e samenich e asstubbl e turnips,fodde r rape and fodder radish,catc h cropsgrow n for forageproductio n orgree nma ­ nuring.

10.3 INTROGRESSION

Firststep s fortransfe r of Raphanus traits to B. campestris and B. napus weremad ebu tintrogressio n ofresistanc e tobee teelwor m andclub - roothav eno tye tbee n accomplished because ofth e long-term nature of such aims and theproblem s encountered. Transfer of Raphanus genes to B. campestris proved tob edifficult , becauseproductio n of alienmonosomi c additions failed.Onl ywit hconsi ­ derable efforts suchaddition smigh tb eobtained .Alie n chromosomes in suchplants ,however ,woul d paironl y sporadically with Brassica chromo­ somes, ifa tall .S oth ewa ychose nha sprove dunattractive . Theprospect s forintrogressio n of Raphanus genes into B. napus were morepromising ,x Brassicoraphanus proved ausefu lbridg ebetwee n R. sati- vus and B. napus, because thecrossabilit y with B. napus was rathergoo d and thecondition s forexchang ebetwee n Raphanus and Brassica chromosomes weremor e favourable inth eAAC R hybrids.Howeve r difficulties inachie ­ vingbackcrosse s restricted successt osom eextent . Finally increased knowledge ofth emechanis m ofresistanc e toth ebee t eelworm made transfer far less feasible than itseeme d tob e atth e start ofth eprogramme .Th epartia l andpossibl ypolygeni c character ofresist ­ ancewil l hamper transfer to B. campestris and B. napus.

79 Summary

InDutc h agriculture, cruciferous arable crops areo f littlesignifi ­ cance, although the genus Brassica does comprise attractive crops like rape.Th e small interest isno t surprising in areaswit h anemphasi s on cultivation of sugar-beets,becaus ebee t eelworm (Heterodera schachtii) is able tomultipl y on Brassica crops.Breedin g ofcruciferou s cropswit h resistance tothi snematod ewoul d remove amajo r objection to suchcrops . This consideration was themai nmotiv e forth epresen tprogramm e onin - tergeneric cruciferoushybrids . The investigations weredirecte d totransfe r ofgene s from fodderra ­ dish (Raphanus sativus) to forage rape or oil-seed rape (Brassica napus) and stubble turnip (Brassica campestris), especially genes for resistance tobee teelwor m and club-root, andt oproductio n of afertil e artificial allotetraploid, xBrassicoraphanus. Suchhybrid s from thecros s B. cam­ pestris x R. sativus couldhav e agricultural value as aforag e crop,b y combining the rapid growth, resistance tobee teelwor m and club-root of fodder radish andpalatabilit y of foliage of B. campestris. On alarg e scale,2 xx 2x ,4 xx 4x ,4 xx 2 x and 2xx 4 x crossesbe ­ tween B. campestris (AAo rAAAA )an d R. sativus (RRo rRRRR )wer emade , alwayswit h the former species as femaleparent .Thi s resulted in0.0 20 , 0.024, 0.001 and0.00 2hybri d perpollination , respectively: intotal , 196hybrids .Moreover , another 58hybrid swer e obtained by embryo cul­ ture. Thepoo r crossability between the two species resulted frompoo r fer­ tilization and arrested embryo growth and development.O n the stigma, growth andpenetratio n ofmos tpolle n tubes into stigmapapilla e were disturbed. Such events limited considerably thenumbe r ofpolle ntubes , penetrating the style.Eve n sote no rmor epolle n tubespe r stylewer e counted inmos tpistils .Bu dpollinatio n had nopositiv e effect onthi s number.Poo r fertilization was predominantly duet oerrati c growth of pollen tubeswithi n theovary . Inaggregate , 1t o 2ovule s perpisti l were fertilized.Th enumbe r ofdevelope d ovules in2 xx 2 x and4 xx 4 x crosseswit h rather easily crossable accessions of B. campestris wasal ­ most ashigh .Th e ovuleswer e small and often contained anembry o with arrested growth (Table 9).Howeve r some ovuleswer ebigge r and contained a normally developed embryo.Th e embryos from those ovules gave riseonl y tomatromorphs , and theembryo s from small ovules only tohybrids . For large-scale production ofhybrids ,cultur e ofembryo s invitr o seemed attractive,sinc eotherwis e agrea tman y embryos diedbefor ematu ­ rity.Result s ofembry o culture,however ,wer e onth ewhol e ratherpoor . A larger effectwa s obtainedb y selection ofmor e easily crossable geno types of B. campestris. The laboriouswor k ofcrossin g couldb ecu tdow n byusin g geneticmale-steril e plants of B. campestris. Thegenom e constitution ofth ehybrid swa sAR ,AR R orAARR .Th e cros­ ses 2xx 4 x and4 xx 2 xgav eris e toa fe wAAR Rhybrids .Th eintergene - richybrid swer e ingenera lvigorous , formed alea frosett e atth evege ­ tative stage andha dwhite ,purpl e oryello w flowers,whic hwer emostl y distinctly veined. Despite derangedmeiosi swit h only 0.47 bivalents per cell, 38% o f allA Rhybrid s stillproduce d some stainablepolle nb y formation of 2n gametes,probabl y asa consequenc e of formationo frestitutio nnucle i during interkinesis. Intercrossing ofA Rhybrid s resulted infe woff ­ springwit h abouthal fhexaploid s (AAARRR). Intercrossing ofsuc hoff ­ springgav eris et o afe wseeds . Insom eo fth eA R hybrids,th enumbe r ofchromosome s was doubledb y colchicine treatments ofroote d cuttings orb y crossing ofA R hybrids withAAR R hybrids (meiotic doubling). TheAAR R hybridswer ehighl y sterile;crosse sbetwee n suchhybrid s resulted in0.00 8see dpe rpollination .Th e low fertilitywa sno tcause d bymeioti c irregularities,bu tb ybreedin gbarriers ,whic hwer ever y sim­ ilar toth ebarrier s inth eorigina l intergeneric crosses.S oth eincong ­ ruity systems of B. campestris and R. sativus were apparently stillac ­ tive in xBrassicoraphanus. Someplant sprove d self-compatible, i.e.th e number ofpolle n tubes inth e style after self-pollination wasn oles s than after cross-pollination. Twopopulation s of xBrassicoraphanus were established andexpose d for three generations tonatura l selection.Bot hpopulation s improved forpolle nstainability ,productio n ofsiliqua e and seeds,an d seed-set per siliqua.Howeve r fertilitywa s still lowafte r threegenerations . Thebes tpopulatio n gave 72seed spe rplant .Variatio n in fertility traitswa s still large,s otha ta considerabl e improvementwa s stillex ­ pected inlate rgenerations . Beside good fertility, astabl enumbe r ofchromosome s isprerequisit e forcommercia lus eo f xBrassicoraphanus. The occurrence ofhexaploid s (AAARRR)an d the formation ofunivalent s endanger chromosomal stability. The dangerwil lprobabl y decrease as the fertility of xBrassicoraphanus improves. TheA R andAAR R hybridswer e crossedwit hvariou s species (Tables 17 and 20). xBrassicoraphanus wasbackcrosse d twicewit h B. campestris to produce aserie s ofmonosomi c additions in B. campestris. Unfortunately, theB C plants didno thav e the desired number ofchromosome s andwer e completely sterile. xBrassicoraphanus and B. napus hybridized readily, if

81 the intergeneric hybridswer euse d as femaleparent .Th e reciprocal cros­ ses resulted mainly inmatromorph s and only some hybrids,mostl yhexa - ploid,wit h thegenom e constitutionAAACCR . The F hybrids (AACR and AAACCR)wer e subsequently twicebackcrosse dwit h B. napus. Studies of chromosome association inth emeiosi s ofAA R andAAC R hybrids showed, thatth e former typeswer e ratherpromisin g for the introgression of Raphanus genesint o a Brassica species. Samenvatting

Ind eNederlands e landbouw isd eteel tva nkruisbloemig e landbouwge­ wassenva ngering ebetekenis ,hoewe l hetgeslach t Brassica aantrekkelijke landbouwgewassen omvat,zoal s koolzaad. Instreke nme tee n intensieve teeltva nsuikerbiete n isd egering ebelangstellin gvoo r landbouwcrucife- renechte rnie tz overwonderlijk ,omda the tbietecystenaaltj e (Heterodera schachtii) zicho pSrassica-gewasse n vermeerdert.He tkweke nva n rassen vanlandbouwcrucifere nme tresistenti e tegendi t aaltjeza lnaa rverwach ­ ting eenbelangrijk ebeperkend e factorvoo rd eteel tva ndergelijk ege ­ wassenwegnemen .Dez e overwegingvormd ehe thoofdmotie fvoo rhe tverwe ­ zenlijkenva nee nsoortkruisingsprogramma . Hetonderzoe kwa s gerichto pprobleme ndi e samenhangenme td eover ­ drachtva ngene nui tbladramena s (Raphanus sativus) naarbladkoo l of koolzaad (Brassica napus) enstoppelkno l (Brassica campestris), inhe t bijzonder genenvoo r resistentie tegenhe tbietecystenaaltj ee nknolvoet , enhe tproducere nva n een fertiele kunstmatige allotetraploid, xBrassi- coraphanus. Zulkehybride nui td ekruisin g B. campestris x R. sativus kunnennaa rverwachtin g landbouwkundige waarde hebben alsvoedergewa s doorcombinati eva nd esnell egroe ie nresistenti e tegenhe tbietecysten ­ aaltje enknolvoe tva n R. sativus metd e smakelijkheid van B. campestris. Tussen B. campestris (AAo fAAAA )e n R. sativus (RRo fRRRR )zij no p grote schaal 2xx 2x ,4 xx 4x ,4 xx 2 xe n2 xx 4 x kruisingengemaakt ; steedsme td eeerstgenoemd e soort alsmoeder .Di t leverderespectieve ­ lijk 0,020, 0,024, 0,001 en0,00 2hybride npe rbestuivin g op;i ntotaa l 196hybriden .Toepassin gva nembryocultuu rleverd eno geen s 58hybriden . De slechte kruisbaarheid tussend etwe e soortenwa she tgevol gva n (1)ee n slechtebevruchtin g en (2)ee ngeremd e embryogroeie n-ontwikke ­ ling.O pd e stempelware nd egroe i end epenetrati e ind estempelpapil - lenva nd emeest epollenbuize n insterk emat e gestoord.Dergelijk ereac ­ tiesveroorzaakte n een aanzienlijke beperkingva nhe t aantalpollenbui ­ zen, datd e stijlwis tbinne n tedringen .Toc hwerde n inhe tgrootst e deelva nd e stampers tieno fmee rpollenbuize npe r stijl geteld.Knop - bestuiving had geenpositiev e invloed opdi taantal .D e slechtebevruch ­ tingwer dvoora lveroorzaak t door eengestoord e oriëntatie vand epollen - buisgroei inhe tvruchtbeginsel .Gemiddel d werden 1à 2 zaadknoppe npe r stamperbevrucht .He t aantal uitgegroeide zaadknoppen wasbijn a netz o grooti n2 xx z xe n4 xx 4 xkruisinge nme t goed kruisbare B. campestris- herkomsten.D e zaadknoppenware n kleine nbevatte nvaa k eenembryo ,waar - vand egroe imeesta l voortijdig stagneerde (Tabel 9).Sommig e zaadknoppen waren echter aanzienlijk groter enbevatte n eennormaa l ontwikkeld embryo. Uitdez e grote zaadknoppenwerde n door embryocultuuruitsluiten dmatro - morfenverkrege n enui td eklein e zaadknoppen alleenhybriden . Voor grootschalige produktieva nhybride n leek invitr o opkweekva n hybride-embryo's aantrekkelijk, omdater gvee l embryo's inee nonvol ­ groeid stadium afsterven.D e resultatenva n embryocultuur waren echter overhe tgehee l nogal teleurstellend. Betere resultatenwerde n verkregen door de selectieva n goed kruisbare B. campestris-genotypen. Hettijdro ­ vende kruisingswerk konvoort sworde nvermede n door genetisch mannelijk steriele B. campestris-plantent egebruiken . De genoomsamenstelling van deverkrege n hybridenwa sAR ,AR R ofAARR . Hetwa s opmerkelijk dat2 xx 4 x en4 xx 2 xkruisinge n enkeleAARR-hybri - denopleverden .D e geslachtshybriden waren overhe t algemeen groeikrach- tig,vormde n inhe tvegetatiev e stadium eenbladroze t enhadde nwitte , paarse ofgel ebloemen , diemeesta l duidelijk geaderdwaren . Ondanks eenvolledi g gestoorde méioseme tgemiddel d slechts 0,47bi ­ valentenpe r celproduceerd e tochno g3 8% va n alleAR-hybride n enig kleurbaar stuifmeel.Di thiel d verband methe tontstaa nva n 2n-gameten waarschijnlijk alsgevol gva nd evormin gva n restitutiekernen tijdens de interkinese.Onderling e kruisingva nAR-hybride n leverde een kleinenako ­ melingschap op,di evoo r ietsmee r dan dehelf tui thexaploide n (AAARRR) bestond.Onderling e bestuivingva nd eplante nui tdez e nakomelingschap gafee nklei n aantalzaden . Vanee ndee lva n deAR-hybride nwer dhe t aantal chromosomen verdubbeld door (1)okselknoppe nva nbeworteld e stekkenme t colchicine tebehande ­ lene n (2)doo r kruisingva nAR-hybride nme tAARR-hybride n (meiotische verdubbeling). DeAARR-hybride nware n inhog emat e steriel;onderling e kruisingen leverden gemiddeld 0,008zade npe rbestuivin g op.D e geringe fertiliteit werd nietveroorzaak tdoo r onregelmatigheden ind eméiose ,maa r door kruisingsbarrières dievee l overeenkomstvertoonde nme t debarrière s in deoorspronkelijk egeslachtskruising .Di twijs tero p dat in xBrassicora- phanus de incongruentiesystemen van B campestris en R. sativus nog actief zijn.Ee n deelva n deplante n bleekbi j zelfbestuiving zelf-compatibel, d.w.z.he t aantalpollenbuize n ind e stijlwa sn a zelfbestuiving niet lager dann a kruisbestuiving. Tweepopulatie s van xBrassicoraphanus zijn samengesteld enzij ngedu ­ rende drie generaties mino fmee r overgelaten aannatuurlijk e selectie. Beidepopulatie s vertoondenverbeterin g vanpollenfertiliteit , hauw-e n zaadproduktie enzaadzettin g per hauw.N a drie generaties was hetferti - liteitsniveauechte r nog laag.D ebest epopulati e leverde gemiddeldon ­ geveer 72zade npe rplan t op.D evariati evoo r deverschillend e fertili- teitskenmerkenwa s nog groot,zoda tnaa rverwachtin g de fertiliteit in devolgend e generaties nog aanzienlijk zalverbeteren . 84 Naastee ngoed e fertiliteit isee n stabiel chromosoomaantal eennood ­ zakelijkevoorwaard e voor commercieel gebruikva n xBrassicoraphanus. Het ontstaanva nhexaploide n (AAARRR)e nd evormin gva nUnivalente n zijnee n potentieel gevaarvoo r dechromosomal e stabiliteit.Naa rverwachtin gwor ­ dendez e gevarenminder , indiend e fertiliteitva n xBrassicoraphanus ver­ betert. DeAR -e nAARR-hybride n zijnme tdivers e soortengekruis t (Tabel 17 en 20). xBrassicoraphanus istweemaa l met succes teruggekruistme t B. cam- pestris, teneinde in B. campestris eenreek smonosom e additiest eprodu ­ ceren.D eT_-plante nhadde n evenwel niethe tgewenst e aantal chromosomen enware nvolledi g steriel.Kruisin gva n xBrassicoraphanus met B. napus verliep opvallend goedme td egeslachtshybrid e alsmoeder .D e reciproke kruisingen leverden daarentegen hoofdzakelijk matromorfen ope n slechts enkele,mees thexaploid e hybridenme td egenoomsamenstellin g AAACCR.D e F,-hybriden (AACR enAAACCR )werde nvervolgen s tweemaalme t succesterug - gekruistme t B. napus. Studiesva nd e chromosoomassociatie ind eméios e vanAAR - enAACR-hybride n tonen aan,da td e laatstgenoemde typenvri j kansrijk zijnme tbetrekkin g totd e introgressie vani?apiïa/2us-gene ni n eenBrassica-soort .

85 References

Banga, 0., 1976.Radish . Raphanus (Cruciferae). In: Simmonds. p.60-62. Bannerot,H. , L. Boulidard, Y. Cauderon & J. Tempe, 1974.Cytoplasmi c male sterility transfer from Raphanus to Brassica. In: Wills & North, p. 52-54. Baukloh, H., 1976.Untersuchunge n zur Wirtspflanzeneignung der Kruziferen gegenüber dem Rübennematoden, Heterodera schachtii (Schmidt), unter besonderer Berücksichti­ gung der Resistenzzüchtung. Dissertation, Georg-August Universität, Göttingen. 72p . Becker, T., 1951. Siebenjährige blütenbiologische Studien an den Cruciferen Brassica napus L., Brassica rapa L., Brassica oleracea L., Raphanus L. und Sinapis L. Zeit­ schrift für Pflanzenzüchtung 31:72-103 . Berger, B., 1968.Entwicklungsgeschichtlich e und chromosomale Ursachen der verschiede­ nen Kreuzungsverträglichkeit zwischen Arten des Verwandtschaftskreises Arabidopsis. Beiträge zur Biologie der Pflanzen 45:171-212 . Bonnet, A., 1978.A yellow petal mutant in cultivated radish (Raphanus sativus L.) heridity of the character. Cruciferae Newsletter 3:43 . Boswell, V.R., 1949. Our vegetable travelers.Nationa l Geographie Magazine 46: 188-217. Chopinet,R. , 1942. Sur quelques hybrides expérimentaux interspécifiques et inter­ génériques chez les Crucifères. Comptes rendus de l'Académie des Sciences, Paris 215: 545-547.Cite d by Yarnell, 1956. Chopinet, R., 1944.Hybride s intergénériques Raphano-Brassica. Revue Horticole, Paris 116: 98-100. Clauss, E., 1978.Allohexaploid e Gattungsbastarde vom Typ Brassico-Raphano-Brassica. Archiv für Züchtungsforschung 8: 297-302. Crute, I.R., A.R. Gray, P. Crisp & S.T. Buczacki, 1980.Variatio n in Plasmodiophora brassicae and resistance to clubroot disease in brassicas and allied crops.A cri­ tical review. Plant Breeding Abstracts 50: 91-104. Eenink, A.H., 1974.Matromorph y in Brassica oleracea L. 1. Terminology, parthenogenesis in Cruciferae and the formation and usability of matromorphic plants. Euphytica 23: 429-434. Ellerström, S., 1978. Species crosses and sterility in Brassica and Raphanus. Cruciferae Newsletter 3: 16-17. Ellerström, S. & L. Zagorscheva, 1977. Sterility and apomictic embryo-sac formation inRaphanobrassica . Hereditas 87:107-120 . Fukushima, E., 1945. Cytogenetic studies on Brassica and Raphanus. I Studies on the intergeneric F hybrids between Brassica and Raphanus. Journal of the Department of Agriculture,Kyüsy ü Imperial University 7:281-400 . Guha, S. & S.H. Maheshwari, 1964. Invitr o production of embryos from anthers of Datura. Nature 204:497 . Harberd, D.J., 1969. A simple effective embryo culture technique for Brassica. Euphy­ tica 18:425-429 . Harberd, D.J., 1971.Addendu m to 'A simple effective embryo culture technique for Brassica'. Euphytica 20:138 . Harberd, D.J. & E.D. McArthur, 1980.Meioti c analysis of some species and genus hybrids in the Brassiceae. In: Tsunoda, et al., p. 65-87. Helm, J., 1957.Übe r den Typus der Art Raphanus sativus L., deren Gliederung und Synonymie. Kulturpflanze 5: 41-54. Henderson, S.A. & B.C. Lu, 1968.Th e use of haematoxylin for squash preparations of chromosomes. Stain Technology 43:233-236 . Herklots, G.A.C., 1972.Vegetable s in South-East Asia. George Allen & Unwin Ltd, London, p. 182-224. Heijbroek, W., 1979. Results of the sugar beet nematode enquiry, organised by the 'Pests and diseases' study group of the I.I.R.B. Comptes rendu d'Institut Interna­ tional de Recherches Betteravières, 42 Congrès d'hiver, Bruxelles Fébrier 1979. p. 141-147. Heyn, F.W., 1973.Beiträg e zum Auftreten unreduzierter Gameten und zur Genetik einiger Merkmale bei den Brassiceae. Dissertation, Georg-August Universität, Göttingen. 103p .

86 Heyn,F.W. , 1974.Unreduce d gametesan dsom egeneti cpeculiaritie s inth eBrasslceae . In:Will s& North ,p .69-73 . Heyn,F.W. , 1978.Introgressio no frestore rgene sfro mRaphanu s sativusint ocytoplas ­ micmal esteril eBrassic anapu san dth egenetic so ffertilit y restoration.Procee ­ dings5t hInternationa lRapesee dConference ,Malmo' ,Sweden , Volume 1,Jun e12-16 . p.82-83 . Hinata,K. ,N .Konn o& U .Mizushima ,1974 .Interspecifi c crossability inth etrib e Brassiceaewit hspecia lreferenc et oth eself-incompatibility .Tohok uJourna lo f AgriculturalResearc h25 :58-66 . Hinata,K .& T .Nishio ,1980 .Self-incompatibilit y incrucifers .In :Tsunod ae tal . p.223-234 . Hirling,W ., 1976 .Di eWirtseignun g desÖlrettich s (Raphanusoleiferu sL. )fü rda sRii - benzystenälchen (Heteroderaschachti iSchmidt) .Zeitschrif tfü rPflanzenkrankheite n undPflanzenschut z83 :631-636 . Hogenboom,N.G. ,1973 .A mode lfo rincongruit y inintimat epartne rrelationships .Euphy - tica22 :219-233 . Hogenboom,N.G. ,1975 .Incompatibilit yan dincongruity :tw odifferen tmechanis m forth e non-functioningo fintimat epartne rrelationships .Proceeding so fth eRoya lSociet y ofLondo nB .Biologica lScience s 188:361-375 . Hosoda,T. ,1946 .Fertilit yo fcolchicine-induce d amphidiploidsbetwee nBrassic aan d Raphanus.Agricultur ean dHorticulture ,Tokyo ,21 :51 5 (Abstractfro mHeridit y4 : 400). Hosoda,T. ,1947 .O nth efertilit yo fRaphanu s- Brassic aan dBrassic a -Raphanu sob ­ tainedb ycolchicin etreatment .Japanes eJourna lo fGenetic s22 :52-53 . Howard,H.W. , 1942a.Self-incompatibilit y inpolyploi d formo fBrassic aan dRaphanus . Nature 149:302-303 . Howard,H.W. , 1942b.Th eeffec to fpolyploid yan dhybridit yo nsee dsiz ei ncrosse s betweenBrassic a chinensis,B .carinata ,amphidiploi dB .chinensis-carinat a andauto - tetraploidB .chinensis .Journa lo fGenetic s43 :105-119 . InternationalBurea uo fPlan tTaxonomy ,1980 .Internationa lcod eo fnomenclatur efo r cultivatedplant s- 1980 .Regnu mVegetabile ,Volum e 104:3 2p . 5thInternationa lRapesee d Conference,Malmö ,Sweden ,1978 .(Proceedings )Jun e12-16 , 1978,2 Volumes . Iwasa,S .& S .Ellerström ,1981 .Meiosi sdisturbances ,aneuploid yan dsee d fertilityi n Raphanobrassica.Heridita s95 :1-9 . Jochemsen,G .& W .Mlyniec ,1974 .A neffectiv e squashtechniqu e forroo ttip so fculti ­ vated crucifers.Genetic aPolonic a 15:443-447 . Kakizaki,Y. ,1925 .A preliminar y reporto fcrossin gexperiment swit hcruciferou s plants,wit hspecia lreferenc et osexua lcompatibilit yan dmatroclinou shybrids . JapaneseJourna lo fGenetic s3 :49-82 . Karpechenko,G.D. ,1927 .Th eproductio no fpolyploi d gametesi nhybrids .Heredita s9 : 349-368. Karpechenko,G.D. ,1928 .Polyploi dhybrid so fRaphanu s sativusL .x Brassic aolerace a L.Zeitschrif tInduktiv eAbstammungs -un dVererbungslehr e48 :1-85 . Karpechenko,G.D. ,1937 .Increasin gth ecrossabilit y ofa specie sb ydoublin g itschro ­ mosomenumber .Bulleti no fApplie dBotany ,o fGenetic san dPlan tBreeding ,Ser .I I 7:37-51 . Kartha,K.K. ,O.L .Gambor g& F .Constabel ,1974 .I nvitr oplan tformatio nfro mste m expiantso frap e (Brassicanapu scv .Zephyr) .Physiologi aPlantaru m 31:217-220 . Kato,M ., 1971 .Cytologica lan dhistologica l studieso fsee ddevelopmen tan dfailur e inth eamphidiploi d (Brassicajaponic ax Raphanu s sativus).Proc .Shikok uMeeting , Jap.Soc .Breed .5 :17-19 . Kato,M .& S .Tokumasu ,1976 .Th emechanis mo fincrease d seedfertilit yaccompanie d withth echang eo fflowe rcolou ri nBrassicoraphanus .Euphytic a 25:761-767 . Kato,M .& S .Tokumasu ,1978 .Nucleu s substitutiono fBrassic awit hRaphanus .Crucife - raeNewslette r 3:42 . Kato,N .& S .Tokumasu ,1979 .Chang eo fcross-affinit ywit hparenta lspecie saccompa ­ niedth echang eo fflowe rcolou ri nBrassicoraphanus .Crucifera eNewslette r4 :34 . Kato,M .& S .Tokumasu ,1980 .Nucleu s substitutiono fBrassic ajaponic a Sieb,wit h Raphanus sativusL .an d itsresultan tchlorophyl ldeficiency .Euphytic a29 :97-106 . Lewis,D .& L.K .Crowe ,1958 .Unilatera l interspecificincompatibilit y inflowerin g plants.Heredit y 12:233-256 . Lubberts,J.H .& H .Toxopeus ,1982 .He tontwikkele nva nbietecystenaaltje s resistente bladrammenase ngel emosterd .Zaadbelange n (inpress.) . Marrewijk,N.P.A .va n& H .Toxopeu s (Eds), 1979.Proceeding so f Eucarpia-Cruciferae 1979- Conference ,Wageningen ,1 ,2 an d3 October .Pos tConferenc eEdition ,SV Pan d RIVRO,Wageningen . 184p . McCollum, G.D., 1979. Sterility in successive backcrosses of Raphanobrassica (2n = 4x = 36)wit h recurrent parent Brassica oleracea (2n = 2x = 18).Canadia n Journal of Genetics & Cytology 21:479-485 . McNaughton, I.H.,197 3 Synthesis and sterility of Raphanobrassica. Euphytica 22:70 - McNaughton, I.H.,197 6 Swedes and rapes. Brassica napus (Cruciferae). Tn: Simmonds. p.53-5 6 McNaughton, I.H.,197 9 . The current position and problems in the breeding of Raphano- brassica (Radicole) as a forage crop. In: Marrewijk & Toxopeus. p. 22-28. McNaughton, I.H.& C. L . Ross, 1978. Inter-specific and inter-generic hybridization in the Brassicae with special emphasis on the improvement of forage crops. In: Scottish Plant Breeding Station, p. 75-110. Mizushima, U., 1944a. Studies on some auto- and allopolyploids made in Brassica, Sina- pis, Eruca and Raphanus. Agriculture and Horticulture 19: 743-744 (Abstract from Heridity 4: 399-400). Mizushima, U., 1944b. Some amphidiploids in Cruciferae. I-IV. Breeding and Horticulture 2: 401-404,441-444 , 481-483,515-51 6 (Abstracts from Heridity 4:400) . Mizushima, U., 1950a. Karyogenetic studies of species and genus hybrids in the tribe Brassiceae of Cruciferae. Tohoku Journal of Agricultural Research 1: 1-14. Mizushima, U., 1950b. On several artificial allopolyploids obtained in the tribe Bras­ siceae of Cruciferae. Tohoku Journal of Agricultural Research 1: 15-27. Mizushima, U,, 1952,Karyo-genetica l studies on Brassiceae. Gihôdô, Tokyo. 112 p. Cited by Tokumasu, 1970b. Mizushima, U., 1968. Phylogenetic studies on some wild Brassica species. Tohoku Journal of Agricultural Research 19: 83-99. Mizushima, U,, 1980.Genom e analysis in Brassica and allied genera. In: Tsunoda et al. p. 89-106. Morris, L.E., 1936. Pollen germination in Brassica chinensis x Raphanus sativus F - hybrids. Journal of Genetics 33:433-441 . Morris, L.E. & R.H. Richharia, 1937.A triploid radish x turnip hybrid and some of its progeny. Journal of Genetics 34:275-286 . Murashige, T. & F. Skoog, 1962- Revised media for rapid growth and bioassays with to­ bacco tissue culture. Physiologia Plantarum 14:423-426 . Namai, H., 1976. Cytogenetic and breeding studies on transfer of economic characters by means of interspecific and intergeneric crossing in the tribe Brassiceae of Cru­ ciferae. Memoirs of the Faculty of Agriculture, Tokyo University of Education 22: 101-171. Namai, H., 1978.Aspec t on transfer of economic characters by means of interspecific and intergeneric crossing in cruciferous crops. In: 5th International Rapeseed Conference,Volum e 1.p . 127-130. Namai, H., 1980.Effec t of flower bud age at bud pollination on cross fertility in re­ ciprocal intergeneric crosses, Raphanus sativus x Brassica oleracea and B. campes- tris x R. sativus. Cruciferae Newsletter 5: 26-27. Namai, H., M. Sarashima, & T. Hosoda, 1980. Interspecific and intergeneric hybridiza­ tion in Japan. In:Tsunod a et al. p. 191-203. Nishi, S., 1980.Differentiatio n of Brassica crops inAsi a and the breeding of 'Haku- ran'. A newly synthesized leafy vegetable. In: Tsunoda et al. p. 133-150. Nishi, S., T. Kuriyama & T. Hiraoka, 1964. Studies on the breeding of Crucifer vegeta­ bles by interspecific and intergeneric hybridization. 1. Special reference to the utilization of matroclinous hybrids accompanied with pseudogamous phenomena in those hybridizations. Bulletin of the Horticultural Research Station, Japan, Ser. A 3: 161-250. Nishiyama, I., 1946. On intergeneric hybridization between tetraploid strains of radish and Brassica. Japanese Journal of Genetics 21: 115.Cite d in Plant Breeding Ab­ stracts 22:no .496 . Nishiyama, I. & N. Inomata, 1966. Embryological studies on cross-incompatibility be­ tween 2x and 4x in Brassica. Japanese Journal of Genetics 41: 27-42. Oelke, J., 1957. Zur Physiologie der Selbst- und Kreuzungssterilität beim Radieschen (Raphanus sativus L.). Der Züchter 27:358-369 . Olsson, G., 1954. Crosses within the campestris group of the genus Brassica. Hereditas 40: 398-418. Olsson, G. & S. Ellerströ'm, 1980. Polyploidy breeding inEurope . In: Tsunoda et al. p. 167-190. Opena, R.T. & S.H. Lo, 1978.Derivatio n of matroclinal diploids in Chinese cabbage and evaluation of their significance in breeding. Journal of the American Society for Horticultural Science 103:820-823 .

88 Ouden,H .den ,1954 .He tbietecystenaaltj e enzij nbestrijding .2 .Waardplante ne nhu n betekenisvoo rd ebietenteelt .Mededelinge nInstituu tvoo rRationel eSuikerproducti e 24:123-136 . Richharia,R.H. ,1937 .Cytologica l investigations ofRaphanu s sativus,Brassic aolera - ceaan dthei rF andF hybrids.Journa lo fGenetic s34 :19-44 . Rousselle,P. , 1979.Stud yo fa cytoplasmi cmal esterilit yafte rinterspecifi ccrosse s betweenBrassic anapu san dRaphanu s sativus.In :va nMarrewij k& Toxopeus .p .100-101 . Sampson,D.R. , 1962.Intergeneri cpollen-stigm a incompatibility inth eCruciferae . CanadianJourna lo fGenetic s& Cytolog y4 :38-49 . Sass,J. , 1964.Botanica lmicrotechnique .Iow aStat eUniversity ,Ames ,Iowa .22 8p . Schwanitz,F. , 1949.Ein eneu ewirkungsvoll e undsparsam eMethod ede rColchicinbehand - lung (Colchicin-Traganth-Schleim).De rZüchte r 19:301-302 . Scoles,G.J .& P.J .Kaltsikes ,1974 .Th ecytolog yan d cytogeneticso fTriticale .Zeit ­ schriftfü rPflanzenzüchtun g 73:13-43 . ScottishPlan tBreedin gStation ,1971 .Fiftiet hAnnua lRepor t1970-1971 . ScottishPlan tBreedin gStation ,1976 .Fifty-fift hAnnua lRepor t1975-1976 . ScottishPlan tBreedin gStation ,1978 .Fifty-sevent hAnnua lRepor t1977-1978 . Simmonds,N.W. , (Ed.),1976 .Evolutio no fcro pplants .Longman ,Londo nan dNe wYork . 408p . Snow,R ., 1963 .Alcoholi chydrochlori c acid-carminea sa stai nfo rchromosome si n squashpreparations .Stai nTechnolog y 38:9-13 . Steele,A.E. ,1965 .Th ehos trang eo fsuga rbee tnematod eHeteroder a schachtiiSchmidt . Journalo fth eAmerica nSociet yfo rSuga rBee tTechnolog y 13:573-663 . Stelter,H. ,1973 .Di eArte nde rGattun gHeteroder a (Nematoda:Tylenchidae )un dihr e Verbreitung.Pedobiologi a 13:40-61 . Takeshita,M. ,M .Kat o& S .Tokumasu ,1980 .Applicatio no fovul ecultur et oth eproduc ­ tiono fintergeneri c hybrids inBrassic aan dRaphanus .Japanes eJourna lo fGenetic s 55:373-387 . Terasawa,Y. ,1932 .Konstant eamphidiploid eBrassico-raphanu sBastarde .Proceeding s ImperialAcadem y8 :312-314 . Terasawa,Y ,1933 .Crossin gbetwee nBrassico-raphanu s andBrassi a chinensisan dRapha ­ nussativus .Japanes eJourna lo fGenetic s8 :229-230 . Terasawa,Y .& N .Shimotomai ,1928 .Bastadierungsversuch ebe iBrassic aun dRaphanus . ScientificReport so fth eTohok uImperia lUniversity ,Ser .4 (Biology)13 :827-841 . Tokumasu,S. ,1965 .O nth eorigi no fth ematromorphi cplant so fBrassic ajaponic aob ­ tained fromth ecros sbetwee nBrassic aan dRaphanus .Journa lo fth eJapanes eSociet y forHorticultura lScienc e34 :223-231 . Tokumasu,S. ,1970a .Studie so npseudogam yi nth ecas eo ftetraploi dplant si nRaphanus . JapaneseJourna lo fBreedin g20 :15-21 . Tokumasu,S. ,1970b .Intergeneri chybrid sbetwee nBrassic ajaponic aan dRaphanu ssati ­ vus.Memoir so fth eColleg eo fAgriculture ,Ehim eUniv .14 :285-302 . Tokumasu,S. ,1976 .Th eincreas eo fsee dfertilit yo fBrassicoraphanu sthroug hcytolo ­ gicalirregularity .Euphytic a24 :463-470 . Tokumasu,S .& M .Kato ,1980 .Cytology ,fertilit yan d flowercolou ri nhybrid sbetwee n Brassicajaponic aan dRaphanu ssativus ,an dth esynthesi so fne wBrassicoraphanus . JapaneseJourna lo fBreedin g 30:11-19 . Toxopeus,H. ,1974a .Outlin eo fth eevolutio no fturnip san d colesi nEurop ean dth e origino fwinte r rape,swede-turnip s and rapekales .In :Will s& North ,p .1-13 . Toxopeus,H. ,1974b .Th ecodin go frace so fPlasmodiophor abrassica e (Woron.).In : Wills& Nort hp .90-93 . Toxopeus,H .& H .Lubberts ,1979 .Breedin g forresistanc et oth esuga rbee tnematod e (Heterodera schachtiiSchm. )i ncruciferou scrop s (summary).In :va nMarrewij k& Toxopeus.p .111 . Tsunoda,S.K. ,K .Hinat a& C .Gómez-Campo , (Eds),1980 .Brassic a cropsan dwil dspecies . JapanScientifi cSocietie sPress ,Tokyo .35 4p . Turesson,G .& H .Nordenskió'ld ,1943 .Chromosom edoublin gan d crosscombination si n somecruciferou splants .Annal so fth eAgricultura l Collegeo fSweden ,11 :201-206 . U,N. ,1935 .Genome-analyse s inBrassic awit hspecia lreferenc et oth eexperimenta l formationo fB .napu san dpeculia rmod eo ffertilization .Japanes eJourna lo fBotan y 7:389-452 . U,N. ,U .Midusim a& K .Saito ,1937 .O na diploi d andtriploi dBrassic a -Raphanu s hybrids.Cytologi a8 :319-326 . Valdivia,B.V.A .& M.M.E .Badilla ,1977 .Artificia lan dnatura lcrosse sbetwee nrap eo r colza (Brassicanapu sL. )an dwil dan dcultivate d crucifers.Agricultur aTécuic a37 : 25-30.

89 Veerman, I., 1981.Onderzoe k naar het mechanisme van de resistentie tegen Heterodera schachtii (Schmidt) bij verschillende waardplanten. Verslag doctoraalvak, Vakgroep Nematologie van de Landbouwhogeschool, Wageningen. 21 p. Wills, A.B. & C. North (Eds), 1974.Proceeding s of Eucarpia 'Cruciferae 1974', Dundee 25-27 September 1974. Scottish Horticultural Research Institute, Dundee, 134p . Wilmar, J.C. & M. Hellendoorn, 1968.Embry o culture of Brussels sprouts for breeding. Euphytica 17: 28-37. Yarnell, S.H., 1956. Cytogenetics of the vegetable crops. 2. Crucifers. Botanical Re­ view 22:81-166 .

90