Pseudogamic production of dihaploids and monoploids in Solanum tuberosum and some related species

E.W.M. van Breukelen

NN08201.838 E.W.M, van Breukelen

Pseudogamic production of dihaploids andmonoploid s in Solanum tuberosum and some related species

Proefschrift terverkrijgin gva nd egraa dva n doctori nd elandbouwwetenschappen , opgeza gva nd erecto rmagnificus , dr.H.C .va nde rPlas , hoogleraari nd eorganisch e scheikunde, inhe topenbaa rt everdedige n opwoensda g2 2apri l198 1 desnamiddag st evie ruu ri nd eaul a vand eLandbouwhogeschoo l teWageninge n

Centre for Agricultural Publishing and Documentation

Wageningen - 1981 Abstract

Breukelen, E.W.M, van (1981) Pseudogamic production of dihaploids and monoploids in Solanum tuberosum and some related species. Agric. Res. Rep. (Versl. landbouwk. Onderz.) 908. ISBN 90 220 0762 6, (viii) + 121 p., 34 tables, 8 figs, 237 refs, Eng. and Dutch summaries. Also: Doctoral thesis, Wageningen. Attempts were made to maximize frequencies of dihaploids from Solanum tuberosum, obtained through pseudogamy after pollination with S. phureja. Factors influencing diha- ploid frequencies were studied: genetics of the pollinator effect, genetics of the seed parent effect and interaction between the two effects on dihaploid frequencies. Temperature influences were determined in a growth chamber experiment. The mechanism of dihaploid formation was studied with the aid of cytological techniques. The pollinator effect was confirmed. Five or more loci were involved and the within-locus interaction was intermedi­ ate. High numbers of hybrids had a negative but small effect on numbers of dihaploids. The seed parent effect was also confirmed. The frequency of dihaploids was determined by the sporophyte rather than by the gametophyte of the seed parent. No interaction was found between the pollinator and seed parent effect on the dihaploid frequency. Low temperature had a positive effect on the dihaploid frequency via the pollinator, but no effect was found via the seed parent. Not the 2n-pollen but the n-pollen proved to be instrumental in dihaploid induction. Monoploids were produced from diploid S. tuberosum and S. verrueosum using several S. phureja genotypes as pollinator. The n-pollen induced the haploids in this case as well. Doubled monoploids were obtained with good female fertility.

Free descriptors: Solanum tuberosum, S. phureja, S. Verrucosum, haploid, diploid, mono­ ploid, pseudogamy, parthenogenesis, genetics, cytogenetics, pollen tube mitosis, certation, 2n-gametes, growth chamber, breeding, chromosome-doubling.

This thesis will also be published as Agricultural Research Reports 908.

(£) Centre for Agricultural Publishing and Documentation, Wageningen, 1981.

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. Stellingen

1. In tegenstelling tot de opvatting van Irikurae n Sakaguchi ishe tmoge ­ lijko mhe taanta l chromosomenva nd edihaploid e aardappel tehalvere n door pseudogamie. Y. Irikura& S.Sakaguchi ,1972 .Potat oResearc h 15:170-173 . Ditproefschrift .

2.He tverdien t aanbeveling om te onderzoeken ofU.V.-bestralin g tijdens de pollenbuisgroei een positieve invloed heeft op het aantal monoploiden dat gevormdword tui tkruisinge n tussen twee diploiden.

3. Diploide regeneranten uit microsporen die homozygoot zijnvoo r eenken ­ merkwaarvoo r deoude rheterozygoo t is,zij nnie tnoodzakelij k uit haploide microsporen ontstaan. E. Jacobsen& S.K. Sopory, 1978.Theoretica l andApplie d Genetics52 : 119-123.G .Wenzel ,1980 .In :D.R . Davies &D.A . Hopwood (eds.), ThePlan tGenome ,p .185-196 .

4. De door Hermsen geformuleerde maat voor efficiëntieva ndihaploiden-in - ductie kan worden verbeterd door het aantal dihaploidenpe r 100zade nmin ­ derzwaa rme e telate ntelle n danhe taanta l dihaploidenpe r bes. J.G.Th. Hermsen& J .Verdenius ,1973 .Euphytic a 22:244-259 .

5. Boeren in ontwikkelingslanden zijn eerder geholpen met voorzieningen voorgezon d zaaizaad danme tverbeterd e rassen.

6.Bezuiniginge n ophe tNederlands everedelingsonderzoe k zulleno p den duur hunweersla ghebbe n opd euitvoe rva n zaaizaad enpootgoed .

7.Al spleiotropi e het faillissement isva nd egenetic ava nhe teiwitgehal ­ teva ntarwe ,i sd eoogst-inde x het faillissementva n degewasfysiologie . Th.Kramer ,1980 .Landbouwkundi g Tijdschrift 92:279-284 .

8. De neoscholastieke leer van het hylemorfisme kan een dynamische werke­ lijkheid nietbeschrijven . 9. De produktieva n kunstvoeding voor zuigelingen zou inontwikkelingslan ­ dennie tal sbijdrag e tothe tnationaa lproduk tmoete nworde n gerekend.

10. In een land waar geen grote zoogdieren meer in het wild voorkomen, is het gemakkelijk om voorstander te zijn van de bescherming van deze dieren elders.

11. Bij de voortgezette verkleining van elektronische apparatuurword the t steedsmoeilijke rzic hachte ree ncompute r teverschuilen .

Proefschriftva nE.W.M ,va nBreukele n Pseudogamicproductio n ofdihaploid s andmonoploid s in Solanum tuberosum and somerelate d species Wageningen,2 2 april198 1 Woord vooraf

De werkzaamheden aan dit proefschrift vielen uiteen in twee periodes: het verzamelen en verwerken van gegevens in Wageningen en het meeste schrijfwerk inNairobi .O m degegeven s teverkrijge n zijn 50 000 kruisingen gemaakt en meer dan 600 000 zaden éénvoo r éénbekeken .Ee nklei ndee l van de zaden is uitgezaaid om de zaailingen te beoordelen, maar dat waren er toch nog vele duizenden. Inbeid e periodes heb ikva nverschillend e mensen veel hulp gehad eni kwi l ze graagbedanken . Professor dr.ir .J.G.Th . Hermsen confronteerde mijtijden smij ningeni ­ eursstudie met de raadsels van de aardappel-dihaploïden. Dit leidde - via een onderzoeksvoorstel ingediend bij de Faculteitsraad - tot het hier ge­ presenteerdewerkstuk . Ikdan khe m alspromoto rvoo rzij ngrot eenthousias ­ me end ekritisch e zin,waarme e hij allerlei theorieënme tm e doorsprak. Professor dr. ir. J. Sneep ben ikerkentelij k voor demi j gebodengele ­ genheid ditonderzoe k teverrichte n ophe t IvPe nvoo r zijnbelangstelling . Dr. M.S. Ramanna dank ik voor de geduldige manier waarop hij me in de praktische cytologie inleidde en de vele gesprekken die we hadden over de interpretatieva ncytologisch e aspectenva ndihaploïde n enmonoploiden . De kennis van ing. J. Verdenius van het Solanum-materiaal is van groot nut geweest. De uitstekende opkweek en verzorging van de planten door E.va n de Scheur,J .Rijkse ne nChr .Lo owaardee r ikzeer .Ee n deelva n het grote aantal kruisingen is gemaakt door J. Dregmans,J . Wilmer enP . Korte als stagiaire ofvakantiehulp .Bi jhe tsaaie ,bijn a eindeloze uitwassen van zaden werd ik geholpen door E.va n Rijckevorsel,J . Eerbeek enD .Kelholt . E.va n Rijckevorsel heeft ook veel chromosoomtellingen gedaan,waarbi jon ­ dermee r deeerst emonoploide n als zodanig herkendwerden . Ing.Z . Sawor en J. de Hamer hebben ook een nuttige bijdrage geleverd aan het cytologische werk. Ir. L.S.Anem a ben ik dankbaar, dat ikd egegeven sva n zijningenieurs - onderzoekhe b kunnengebruiken . Voor het werk in klimaatcellen werd gebruik gemaaktva nhe t fytotron in de afdeling Landbouwplantenteelt en Graslandcultuur. De hulp van ing. K. Schölte en van J.H. Möhring bij het terbeschikkin g stellenva n de kli­ maatcellen end e afstelling ervan isdoo rmi jzee r gewaardeerd. Demeest e foto's zijnverzorg d doorH .Sengers . Bijdr .A.C .va nEijnsberge n enir . I.Bo sko n ikaltij d aankloppenvoo r statistischadvies . J.S. de Block ben ik dankbaar voor zijn waardevolle suggesties voor de verbeteringva nd eEngels etekst .R.J.P .Aalpo l dank ikvoo rd e redactie en uitgaveva nhe tproefschrift . Marryheef tm evee lmorel e steungegeve n ind e lange schrijffase. Ze gaf ookvee lpraktisch e steundoo rhe tbecommentariëre n en typenva nee neinde ­ lozereek sversie sva nhoofdstukken . En tenslotte dank ik de cheetah's inhe tNairob i National Park, dieva n tijd tottij dvoo rd enodig e ontspanning zorgden,oo k alsz eonvindbaa rwa ­ ren. Curriculum vitae

De auteurwer d op 15oktobe r 1943 geboren inDe nHaa g en slaagde in 1961 voor het eindexamen Gymnasium-ß aan het Aloysius College aldaar. In 1966 behaalde hij het licentiaat in de filosofie aan de Filosofische Faculteit Berchmanianum teNijmegen ,waarn a hijzij n studie aand e Landbouwhogeschool te Wageningen begon. De praktijktijd werd doorgebracht ophe tPlan tBreed ­ ing Station te Njoro, Kenya. In april 1973 studeerde hij met lof afi nd e studierichting plantenveredeling met als keuzevakken erfelijkheidsleer en planteziektenkunde. Van juni 1973 tot januari 1977 was hij als promotie­ assistent verbonden aan het Instituut voor Plantenveredeling (IvP)va n de LH. Sinds april 1977 is hij alslecture rverbonde n aand eMSc-opleidin g in plantenveredeling aand e landbouwfaculteit vand eUniversitei tva nNairobi , Kenya. Tevens doet hij daar onderzoek aanresistenti eva n Phaseolus vulga­ ris tegenroest . Contents

Introduction andterminology 1.1 Introduction 1.2 Terminology

2 Literature review 4 2.1 Introduction 4 2.1.1 Sourceso fhaploid s 4 2.1.2 Solanum haploids 5 2.1.3 Screening forhaploid s 6 2.2 Useo fhaploid s 7 2.2.1 Haploidsi nresearc h 7 2.2.2 Haploids inbreedin g 8 2.2.3 Solanum haploidsi nresearc h 9 2.2.4 Solanum haploids inbreedin g 9 2.3 Influence ofth eparent s on frequencies of haploids frompseudogam y H 2.3.1 Influenceo fth epollinato r 12 2.3.2 Influenceo fth e seedparen t 14 2.3.3 Interactionbetwee npollinato r and seedparen t influence 16 2.4 External influenceo nhaploi d frequencies 16 2.5 Mechanisms ofpseudogami chaploi dproductio n 18 2.5.1 2xx2 xcrosse s 19 2.5.2 4xx2 xcrosse s 19 2.5.2.1 Triploidhybrid s 19 2.5.2.2 Tetraploidshybrid s 21 2.5.2.3 Dihaploids 21

3 Pollinator influence on dihaploid andhybrid frequencies 25 3.1 Introduction 25 3.2 Material andmethod s 26 3.3 Results 28 3.3. .1 Influence ofth epollinato r on dihaploid andhybri d frequencies 30 3.3. .2 Relationshipbetwee ndihaploi d andhybri d frequencies 31 3.3. .3 Inheritance ofdihaploi d inducing ability 33 3.3 .4 Inheritance ofhybri dproducin g ability 34 3.4 Discussion 35 3.4.1 Influence ofth epollinato r on dihaploid andhybri d frequencies 35 3.4.2 Relationship betweendihaploi d andhybri d frequencies 36 3.4.3 Inheritanceo fdihaploi d inducing ability 37 3.4.3.1 Modeo finheritanc e 38 3.4.3.2 Numbero floc i 39 3.4.3.3 Test forintermediat e inheritance 39 3.4.3.4 Transgression 40 3.4.4 Inheritanceo fhybri dproducin g ability 40

4 Seed parent influence on dihaploid andhybrid production 41 4.1 Introduction 41 4.2 Material andmethod s 41 4.3 Results 43 4.3.1 Influence ofth esee d plant ondihaploi d andhybri d frequencies 43 4.3.2 Cytoplasmic inheritance 45 4.3.3 Sporophyticvs .gametophyti cdeterminatio n 46 4.4 Discussion 47 4.4.1 Influence ofth esee d parent ondihaploi d andhybri d frequencies 47 4.4.2 Cytoplasmic inheritance 49 4.4.3 Sporophytic vs.gametophyti c determination 49 4.4.4 Modeo finheritanc eo fdihaploi d producing ability 50

5 Interaction between pollinator and seed parent influence 51 5.1 Introduction 51 5.2 Material andmethod s 52 5.3 Results 53 5.3.1 Dihaploid production 53 5.3.2 Hybridproductio n 59 5.4 Discussion 60 5.4.1 Dihaploids production 60 5.4.2 Hybridproductio n 61

6 Influence of temperature on frequencies of dihaploids andhybrids 62 6.1 Introduction 62 6.2 Material andmethod s 62 6.3 Results 64 6.3.1 Treatments with fixed temperatures 65 6.3.2 Treatments inwhic h the seedparen twa s transferred 67 6.3.3 Treatments inwhic h thepollinato r was transferred 67 6.4 Discussion 69 7 The mechanism of dihaploid induction 72 7.1 Introduction 72 7.2 Material andmethod s 72 7.3 Results 74 7.3.1 Determinationo f2 npolle n frequencies 74 7.3.2 Ovule lethality 75 7.3.3 Correlationbetwee ndihaploid s andhybrid spe rberr y 76 7.3.4 Pollentub emitosi s 76 7.3.5 Certation 79 7.3.6 Delayedpollinatio n 79 7.4 Discussion 79 7.4.1 Relationship between2 npolle n andhybrid s 79 7.4.2 Ovulelethalit y 80 7.4.3 Therol eo f2 npolle ni ndihaploi d induction 81 7.4.4 Pollentub emitosi s 82 7.4.5 Certation anddihaploi d frequencies 83 7.4.6 Themechanis mo fdihaploi d induction 84

8 Nonoploids 85 8.1 Introduction 85 8.2 Material andmethod s 85 8.3 Results 88 8.3.1 Pollinator influence 90 8.3.2 Seedparen t influence 91 8.3.3 Doubledmonoploid s 92 8.4 Discussion 93 8.4.1 Pollinator influence 94 8.4.2 Seedparen tinfluenc e 94 8.4.3 Doubledmonoploid s 96

8.4.4 Mechanismo fmonoploi d formation 96

9 Conclusion 99

Summary 101

Samenvatting 104

References 108 1 Introduction and terminology

1.1 INTRODUCTION

Haploidy plays an important role in the life ofplants .Th e alternation of haplophase and diplophase makes recombination of genetic material pos­ sible. In lower plants the haplophase is the longest phase in the life cycle. It was reduced during evolution in size and duration, so that the haplophase inhighe rplant sconsist s ofa fewcell s onlywit h ashor tlife ­ span. The success of the diplophase inhighe rplant s indicates,tha t there are advantages forplant s inhavin g chromosomes inpairs . With the systematic production of haploids in higher plants by biolo­ gists and breeders, the haplophase is extended again and can be as impor­ tant as the diplophase. Artificial life cycles canb emad ewit h haplophase anddiplophas e ofequa l length andimportance . Haploids canb euse d inbasi c research. Inplan tbreedin g the advantages of both the n and 2n condition can be combined. The sterile haploids from heterozygous diploids and allopolyploids can be used toproduc e homozygous plants. More applications can be expected from autopolyploids asth epoly ­ haploid is usually a well growing plantwhic h canb e fertile.Fo rsuccess ­ ful breeding it is necessary thatplan t material is available with suffi­ cient diversity. Systematic work with haploids can only be carried out if manyhaploid s canb eproduce d fromman ydifferen t2 ngenotypes . Much research hasbee n done alreadyo nth epseudogami c production ofdi - haploids in Solanum. Dihaploids of autotetraploid Solanum tuberosum (2n=48) can be produced in large numbers by pollination with diploid S. phureja (2n=24)(Houga se t al., 1958). The genotype ofbot h seedparen tan dpollina ­ tor proved to be important for thedihaploi d frequencies obtained (Gabert, 1963). Cytological studieshav e shownth e importance ofvita l endosperm for the development ofdihaploi d embryos (VonWangenhei m et al., 1960). Over the years the frequencies of dihaploids were increased from 80 dihaploids per 100 berries (Gabert, 1963) to 400 dihaploids per 100 berries (Hermsen & Verdenius, 1973). This led to the questionwhethe r themaximu m rateo fdi ­ haploid formationha dbee nreached , and ifnot ,ho w itcoul d furtherb e im­ proved. The availability ofa largenumbe r of S. phureja genotypes with adiver ­ sity in dihaploid inducing ability at the Institute of Plant Breeding (IvP), Wageningen, together with good potato crossing facilities provided anexcellen topportunit y forresearc ho nhaploi d production inS . tuberosum andothe r Solanum species. The objective of this study was toanswe r thequestio nwhethe rwit h the available S. phureja genotypes thelimi to finducin g dihaploids in S. tube­ rosum was reached.Thi swa s tob edon eb y investigating the genetic and cy­ togenetic background of the differences in dihaploid inducing ability in S. phureja genotypes and the differences in dihaploid producing ability of cultivars of S. tuberosum. In addition the influence ofth e environment on dihaploid frequenciesha d tob etake nint oaccount . Accordingly experiments were set up to determine the influence of dif­ ferent seed parent and pollinator genotypes on dihaploid formation during fourcrossin gseasons .Cytologica lexperiment s were included to obtainmor e insighti nth emechanis m ofdihaploi d formation. Inth ecours e of the study itwa s discovered thatpseudogam y could produce S. tuberosum monoploids as well. After initial success thispar t ofth estud ywa s expanded because of itsow nimportanc e and sincei twa shope d thatmonoploid s would throw light onth eorigi no fdihaploid s aswell . Literature onhaploid s ispresente d inChapte r 2.A general part is fol­ lowed by literature relevant for the experimental chapters.Experiment s to gain information on the genetics of the influence of pollinator and seed parent are reported in Chapter 3 and 4 respectively. The interaction be­ tween the two parental influences is dealtwit h inChapte r 5. InChapte r6 experiments are described which were carried out in growth chambers under controlled conditions.Thes eexperiment s aimed atdeterminin g the effect of thetemperatur e ondihaploi d frequencies.Cytolog y and themechanis m ofdi ­ haploid formationar ecovere d inChapte r 7.Attempt s to induce S. tuberosum monoploids by pseudogamy are reported in Chapter 8 alongwit h a discussion about the origin of monoploids. In the last chapter questions about the maximum possible level of dihaploid production and theefficienc y oftech ­ niques fordihaploi d inductionar e discussed.

1.2 TERMINOLOGY

The terminology used in this study follows the recommendations of De Fossard (1974) as much as possible. The symbol x is used for the basic chromosome number and for multiples of it, e.g. x = monoploid, 2x =di ­ ploid, 3x = triploid. The symbol n is used for the gametic (gametophytic) numbero fchromosome s and2 n forth e somatic (sporophytic)number . A major source of confusion in ploidy terminology is thedoubl e meaning of the word 'diploid'. It can mean both 2x (in the series monoploid, di­ ploid, triploid) and 2n (as opposed to haploid (n)).Her e especially the recommendationo fD eFossar d (1974)i s followed 'torestric tth eus eo fth e word 'diploid' to the 2x condition, to use the word 'haploid' for the n conditionan dt oleav eth e2 n conditionundescribe d by aword' . Haploidplant s aresporophyte s carrying the gametophytic chromosome num- ber (Riley, 1974), or the gametic chromosome number of a species (DeFos - sard, 1974). As thehaploi d isa sporophyt e its chromosome number should be indicated as 2n. On the other hand haploids have the gametic chromosome number of the parent and therefore theycoul d be indicated by n.Th echro ­ mosome number of a haploid plant is denoted by 2n in this study toempha ­ size that it concerns a sporophyte and to distinguish a haploid from the parental gametes. '> 'Haploid' is a relative concept. A haploid hashal f thenumbe r ofchro ­ mosomes of the parent, irrespective of theploid y level ofth eparent .Ha ­ ploids of different ploidy levels are distinguished byprefixe s indicating their ploidy level, e.g. dihaploids and trihaploids are haploids carrying two and three genomes respectively. The word 'haploid' isuse d forhaploi d sporophytes ingeneral ,irrespectiv e ofthei rploid ylevel . The words 'diploid' and 'dihaploid' can beuse d for the sameplant ,th e first indicating the2 xcondition , the second also indicating thatth epar ­ ent was a tetraploid. In this study the word 'dihaploid' is only used for 2xplant s derived directly from4 xplants ;i n allothe r cases 2xplant s are called 'diploids'. The same distinction as between 'diploid' and 'dihaploid' could alsob e made for 'monoploid' and 'monohaploid1. A s there is no reason for confu­ sion, only thewor d 'monoploid' isused . Haploids in Solanum are often formed throughpseudogamy , afterpollina ­ tion. Plants used as pollen source with the aim to induce dihaploids are called 'pollinators'.A goodpollinato r is aplant ,whic h induces dihaploids in a relatively high frequency. The potential ofpollinator s to inducedi ­ haploids iscalle d dihaploid inducing ability (d.i.a.)an d thepotentia l of seed parents to produce dihaploids is called dihaploid producing ability (d.p.a.). 2 Literature review

Thefirs tpar to fth eliteratur erevie w(Section s2. 1an d2.2 )give sge ­ neralinformatio nabou tsource so fhaploid san dthei ruse swit hemphasi so n Solanum. The otherpart sdea lwit hliteratur e relatedt oth eexperimenta l chaptersabou tparenta l influence andexterna l influenceo ndihaploi dfre ­ quenciesan dabou tth emechanis mo fpseudogamy .

2.1 INTRODUCTION

2.1.1 Sources of haploids

Thefirs trepor to na haploi dplan twa sb yBlakesle ee tal . (1922).The y describeda Datura stramonium plant.I tha d1 2chromosome san dwa sobtaine d after cold treatment.Th eoldes tknow nhaploi dprobabl y isth e 'gracilis' type of Thuja plicata, an ornamental tree,whic h was first described in 1896b yBeissne ran dlate rrecognize da sa haploi db yPohlhei m (1968). Aftera perio d inwhic hsevera l singlehaploid swer efoun dan disolate d work on them was done,e.g . byJ0rgenso n (1928)wit h Solanum nigrum, the search forhaploid s aswel l as their production started to be donemor e systematically.Chas e(1949 ,1969 )worke dextensivel ywit hmaize .H eclear ­ lysa wth eapplication s forplan tbreedin g (Chase,1963b ,1964a) .Houga s& Peloguin (1957)wer e pioneers in research on dihaploids in S. tuberosum. The expansion of thehaploi dwor kca nb esee ni nth ereview so fKimbe r& Riley (1963), Magoon& Khanna (1963), Chase (1969)an d in the symposium 'Haploids inhighe rplants 'hel d atGuelph ,Canad a (Kasha,1974) .Haploid s are now available inman y cultivated species,e.g. : Gossypium barbadense (Harland, 1936), Lucopersicon esculentum (Cooper& Brink , 1945), Capsicum frutescens (Morgan& Rappleye ,1954) ,Prunu s persica (Hesse,1971) , Populus alba (Kopecky, 1960), Medicago sativa (Bingham, 1969), Beta vulgaris (Kruse,1961) , Theobroma cacao (Dublin,1974 )an d Secale cereale (Miintzing, 1937). After the more conventionalmethod so fhaploi dproductio n likedistan t crosses, interploidy crosses and experimentation with temperatures,les s conventional schemes were developed. SinceGuh a& Maheshwar i (1964,1966 ) produced embryos andhaploi dplant sb yculturin g antherso f Datura innox- ia, anthercultur e andeve npolle ncultur eopene du p aric h field forth e large-scale production of haploids, especially in tobacco (Bourgin & Nitsch,1967 ;Nitsc h& Nitsch , 1969). Reviewarticle swer ewritte nb ySun - derland(1974 )an dNitsc h (1974). Thediscover yo fa gen e inmaiz efo rindeterminat egrowt hi nth efemal e gametophyte,whic hlead st ohig hfrequencie so fandrogenesis ,increase dth e possibilitieso fpseudogameti candrogenesi s (Kermicle,1969 ,1974) . Kasha& Ka o(1970 )place dth epoorl ygrowin gembryo sfro mth ecros s Hor- deum vulgare x H. bulbosum ona medium . Thechromosome s of H. bulbosum were selectively eliminated duringth eearl ydevelopmen to fth eembry o(Subrah - manyam& Kasha ,1973 )an dhaploi d B. vulgare plantsresulted .Thi sproces s isunde rgeneti ccontro l (Ho& Kasha ,1975) . Semigamyappear st ob ea ver yefficien tsyste mo fhaploi dproduction .I n cottoni tca nlea dt otw otype so fhaploid sfro mth esam ecross ,a sth enu ­ cleid ono treall y fusea tfertilization .Th eresul ti sa chimeri cembryo , growingint oa plan twit hmaterna l andpaterna lhaploi dsectors ,whic hof ­ tendoubl espontaneousl y (Turcotte& Feaster ,1967 ,1969 ,1974) . Tsunewakie tal .(1974 )use dalie ncytoplas mt ostimulat ehaploi dforma ­ tion. Lineso fa whea tvariety ,eac hwit ha differen t Aegilops cytoplasm, werepollinate dwit hnorma lwheat .Thi swa yman ytrihaploid swer eobtained . Kihara & Tsunewaki (1962)ha d done similar workwit h aliencytoplas mbe ­ fore. The influence ofchemical so nhaploi d frequenciesi sdeal twit hi nSec ­ tion2.4 .

2.1.2 Solanum haploids

Thefirs treport so nhaploid si ntube rbearin g Solanum concernedindivi ­ dual plants frominterspecifi c crosses.Lam m (1938)reporte d a2x-4 xtwi n from S. x chauca x s. tuberosum, Ivanovskaja (1939) ahaploi d from the cross S. tuberosum x s. phureja with maternal characters. Interspecific crossesyielde dhaploid si n S. demissum (Bains& Howard ,1950 ;Dodds ,1950 ) andS . polutrichon (Marks,1955 )a swell . Arepor to na singl e S. tuberosum dihaploidfro ma cros swit h S. phureja in 1957 (Hougas & Peloguin, 1957), followedb y2 8dihaploid s inth enex t year(Houga se tal. ,1958) ,wa sth estar tfo rusin g S. phureja asa pollina ­ tor systematically. This species was also a successful pollinator with other Solanum species: S. chacoense (Hermsen, 1969), S. acaule (Hermsen, 1971)an d S. andigena (Del aPuente ,1973 ;Hermsen ,unpubl.) . Otherdiploi d species did also induce dihaploids. Jakubiec (1964) used S. vernei and S. megistacrolobum with some success. S. stenotomum was as successful as S. phureja inexperiment so fBuketov a (1970)an dBuketov a& Yashin a (1971). Budin& Broks h(1972 )obtaine dhaploid swit h S. goniocalyx and S. canasense as pollinators. Dihaploids of S. tuberosum wereals oabl et oinduc eothe r dihaploids (Hermsene tal. , 1974a). Dihaploids froma cros sbetwee nculti - varso ftetraploi d S. tuberosum werereporte d (Cooper& Rieman ,1958 ;Rie - mane tal. ,1959) ,bu tneve rfollowe dup . Within S. phureja certain genotypes proved to have a higher dihaploid inducing ability (d.i.a.)tha n others.Gaber t (1963;Houga s et al., 1964) tested a S. phureja population ford.i.a .an d found three superiorpollina ­ tors. His material has been used by several workers (Jakubiec, 1964;Va n Suchtelen, 1966; Frandsen, 1967) and was improved by Hermsen & Verdenius (1973), who incorporated an embryo marker and selected further for high d.i.a. The first attempts to apply anther culture to Solanum were only partly successful. Irikura & Sakaguchi (1972)announce d a monoploid from diploid S. verrucosum, butfro mtetraploi d S. tuberosum Dunwell& Sunderland (1973) could report only dihaploid embryo formation. Irikura (1975a,b)late r pro­ ducedmono-, di -an dtrihaploi dplant s from anthers of 17tube rbearin g and 2 non tuber bearing Solanum species.H e obtained, however, only one diha­ ploid fromtetraploi d S. tuberosum. Hewa sno t successful with anthers from pure S. tuberosum dihaploids, but the closely related S. phureja and S. stenotomum yieldedmonoploids . The first pseudogamic monoploids from pure S. tuberosum have been re­ portedb yVa nBreukele ne t al. (1975), apartmayb e from asingl e Solanum mo­ noploid mentioned by Frandsen (1968), where androgenesis could not be ex­ cluded.Va nBreukele ne t al. (1975,1977 )reporte d 82monoploid s from S. tu­ berosum and 3 from S. verrucosum, all from crosseswit h S. phureja. Jacob- sen (1978) and Frandsen & Wenzel (mentioned in Wenzel, 1979) also used pseudogamyt oproduc emonoploids .The yproduce d two and fivemonoploid sre ­ spectively. Two S. tuberosum monoploids from anther culture were reported by Fo- roughi-Wehr et al. (1977), twoother sb yJacobse n& Sopory (1978).Wenze l et al. (1979) did not have success with pure S. tuberosum diploids,bu tob ­ tained monoploids from anthers of S. phureja and S. phureja-dihaploid S. tuberosum hybrids.Recentl y (Wenzel, 1980) they reported alarg enumbe r ofmonoploid s from sixclones .Als o diploids wereproduce d whichwer espon ­ taneously doubled according to them. The highest frequency of regenerated plantswa s onepe r fourplate d anthers.

2.1.3 Screening for haploids

Haploid detection evolved with the techniques ofhaploi d induction.Th e first haploids were detected because of aberrant seed size (Gaines& Aase , 1926), small plant size, or the fact that the plant differed from thehy ­ brid progeny and resembled one of theparent s only.Th e finalproo f ofha - ploidywa s achromosom ecount . For large-scale hybrid production markers were developed to distinguish haploids from hybrids more easily. A dominant character in the pollinator was used to find maternal haploids andi nth e seedparen t todetec tandro - genetic haploids. Seedling markers as hypocotyl colourationhav ebee nuse d (Peloquin & Hougas, 1959; Bingham, 1969), but embryo or seed markers were introduced where possible for even earlier detection. Different seed markershav ebee nuse d inmaiz e (Chase,1949 ,1969 ;Co e& Sarkar, 1964), an embryomarke r inmaiz e (Nanda& Chase,1966 )an d anembry omarke r inpotat o (Hermsen & Verdenius, 1973). Screening for polyembryony increased the chances of findinghaploid s inman y species (Review:Lacadena , 1974). Haploids can be.distinguishe d from other plants by their growth habit. Usually they are smaller and their leaves are thinner and narrower. They have smaller cells,whic h results in a higher stomata density in the epi­ dermis (Capote t al., 1968)an d smallerpolle n grains (Kostoff, 1942). Num­ bers ofchloroplast s inth eguar d cells ofth e stomata are lowerwit h lower ploidy levels. This has been used successfully forprescreenin g byButter - fass (1958), Rothacker et al. (1966), Frandsen (1967, 1968), Najcevska & Speckmann (1968), Broksh (1969), Chaudhari & Barrow (1975)an dVa nBreuke - lene t al. (1975, 1977). The finalproo f forhaploid y is achromosom e count.Ofte nth e counts are done on root tips of seedlings in an early stage of development of the plant. When a plant is suspected to be chimerical because of spontaneous chromosome doubling,however ,diameter s ofpolle ngrain s are a fairindica ­ tion,bu tonl y counts frommeioti c cells are decisive.Roo tcell s originate from another tissue layer in the plant than the layer which eventually formsth egamete s (Klopfer, 1965).

2.2 USEO FHAPLOID S

2. 2. 1 Haploids in research

Haploids offer special opportunities forgeneti c studies.Th edirec tus e ofmonoploid s is limited,especiall ybecaus e ofthei r sterility.Pairin g of chromosomes at meiosis can give information on homology within the basic chromosome set of a species. Monoploids are goodmateria l formutatio nre ­ search. Spontaneous and induced mutations can be detected directly, asre ­ cessive alleles are not obscured by dominant alleles (Melchers, 1960;De - vreux & De Nettancourt, 1974). These mutants can contribute to our know­ ledgeo fplan tphysiolog y (Zenk, 1974), notonl y asplants ,bu t also inha ­ ploid cell cultures (Binding, 1974). Haploids can be used to construct a more complete polyploid series for study of gene dosage effects. Seven ploidy levels are now available in Medicago sativa (Bingham & Saunders, 1974). Study of the non-homologous pairing at meiosis in monoploid pollen mother cells cangiv e information onth ebasi c chromosome number and onre ­ lationshipsbetwee ngenome s inallopolyploid s (Sadasivaiah, 1974). Inheritance studies aremuc heasie r toperfor m with dihaploids thanwit h tetraploids, because of the smaller populations needed to detect recessive alleles. 2. 2. 2 Haploids in breeding

Somehaploid sca nb euse d asthe y are,mainl y inornamentals .A Pelargo­ nium cultivar,whic hwa s attractivebecaus e of its small size,prove d tob e a haploid (Daker, 1966), as was the casewit h the 'gracilis' type in Thuja plicata (Pohlheim, 1968). Most haploids, however, are used as a tool in breeding.Th egrea tpotentia l ofhaploid s isth e fastproductio n ofhomozy ­ gous lines.Thes e canb eproduce d fastertha nb y inbreeding and saveon e to three years in maize (Chase, 1952b) or many more years in forest trees (Winton& Stettler, 1974). More importanttha ntim e saving isth e fact that haploids are the way to homozygosity for self-incompatible species, when inbreeding is impossible. Even completely homozygous tetraploids can thus bemade .A n increase inhomozygosit y inautotetraploid s canb e accomplished via a tetraploid-diploid-tetraploid cycle,whic h is comparable to 3.8 gen­ erations of selfing (Obajimi & Bingham, 1973). Androgenesis saves many backcrosses ina progra m totransfe r cytoplasm (Chase, 1963a). The production ofhaploid s on a larger scale gaveris et ohig hexpecta ­ tions for the use of haploids inpractica l breeding. Chasepropose d breed­ ing schemes for allopolyploid cotton andwhea t (Chase,1964a )an d theauto - tetraploid potato (Chase, 1973). Later he reviewed the utilization of ha­ ploids in breeding diploid species (Chase, 1974). Collins & Legg (1974) discussed the potential of haploids for breeding allopolyploids.The y pro­ duced doubled haploids of tobacco which compared well with the parent va­ rieties. By this method they eliminated many undesirable alleles.Nakamur a et al. (1974a,b)produce d three promising tobacco lines in their breeding program via haploids. In asparagus entirely male Fl progenies have been madewit h agrea thomogeneit y (Thévenin, 1974). Inoi l seed rape theculti ­ varMari sHaplo nwa sderive d from ahaploi d (Thompson, 1972). Reinbergs etal .(1975 )sa wthei rdouble d haploids as fixed gametes.The y used them toestimat e theyiel dpotentia l ofbarle y crosses inearl ygener ­ ations (Reinbergs et al., 1976). Their lines had the sameyiel dpotentia l as randomlychose nF 4line s fromth esam ecros s (Songe t al., 1978). So,wit h a threeyea rgai n intime ,the ydi dno tloos eyiel d potential.Thi s confirmed a computer study of Walsh (1974), who also pointed out that the pedigree method isbette r iflinkag e isinvolved . In autogamous crops the good homozygous linesma ybecom e cultivars,bu t in other crops heterozygosity is necessary. Dunbier & Bingham (1975)mad e two populations from haploid derived autotetraploids with the same gene frequencies, but different levels of heterozygosity. Themor e heterozygous doublecros sha d thehighes tyield ,presumabl ybecaus e ofth ehighe r number of tri- and tetra-allelic loci. Doubled haploids from these crops canonl y beuse d for ahybri d program. 2. 2. 3 Solanum haploids in research

Bothmonoploid s and dihaploids in Solanum have theirow n applications in research just as other haploids. Non-homologous chromosome association at meiosis has been observed in monoploids (Van Breukelen et al., 1975). The most important application of monoploids is the fastproductio n ofhomozy ­ gous diploidplants . S. tuberosum dihaploids are smaller than tetraploids, but they may be vigorous and give areasonabl e tuberyield . Leaves are smaller andnarrowe r (Hougas & Peloquin, 1957). The flowering behaviour may be comparable with that of tetraploids,bu t the net female fertility is lower (Carroll &Low , 1975). Male fertility is greatly reduced (Van Suchtelen, 1966; Carroll & Low, 1976;Hermsen ,unpubl.) . Thepercentag e ofplant sproducin g functional pollen canb e aslo w as3 % (Peloquin et al., 1966). Simple segregation ratios in dihaploids (Hougas & Peloquin, 1958) have led to several articles ongeneti cmarker s (Kessel& Rowe ,1974 ;Hermse n et al., 1978a), localization ona particula r chromosome (Hermsen et al., 1973), genetics ofself-compatibilit y (Hermsen, 1978a,b;Hermse n etal. ,1978b )an d gene-centromere mapping (Mendiburu & Peloquin, 1969; Ross &Langton ,1974 ; Mok et al., 1976;Mendibur u &Peloquin , 1979). Dihaploid plants aremor e suitable formutatio n studies than tetraploids (Van Harten, 1978). They also can produce aneuploids after crosses with triploids as seed parents (Vogt & Rowe, 1968;Wagenvoor t & Lange, 1975). Aneuploid diploids can also be found amongst the normal dihaploids from a 4xx 2xcros s (Hermsen et al., 1970).

2. 2. 4 Solanum haploids in breeding

The two main advantages of the use ofdihaploid s inpotat o breeding are (a)thei r disomic inheritance, which facilitates theproductio n ofhomozy ­ gosity for selected characters and (b)th e increased possibilities tomak e crosses with wild or primitive diploid Solanum species (Hougas &Peloquin , 1958, 1960). With these crosses the diversity inth epotat o germ plasm can be greatly increased, for instance with respect to resistance to diseases andpests . Themos tseriou sproble m forth eus e ofdihaploid s isthei rmal esteril ­ ity.Thi s isadde d toth enorma lproblem s ofdiploids : self-incompatibility andmal e sterility insom e interspecific hybrids (Ross et al., 1964;Abdall a & Hermsen,1972 ;Perez-Ugald ee t al., 1964)o runilatera l incongruity (Herm­ sen etal. ,1974b) . According toCarrol l (1975)thi smal e sterility canoc ­ cur invariou s degrees indifferen tgenotypes . Dihaploids from one tetraploid plant form a 'gametic sample'.The y have the genotypes of the gametes and show which characters are hidden in the tetraploid genotype.The y canals oprovid e anestimat e ofth edegre e ofhe - terozygosity for certain characters. This gives dihaploids an application in conventional breeding. The gametic sample is an aid in choosing the rightparent s in4 xx 4 xcrosses . The dihaploids can also be used themselves in breeding. Rothacker & Schäfer (1961)envisage d to make tetraploids with ahig h degree ofhomozy ­ gosity as parents for abreedin g program, exploitingheterosis .Dihaploid s would be crossed with wild Solanum species, made homozygous for selected characters and then doubled. The homozygous tetraploids would constitute good breeding material as the frequencies of genes for desired characters inth eprogen ywil lb ehigh . Chase (1963b) introduced the name 'analytic breeding.'fo rth e procedure of reducing a polyploid to its diploid components, crossing and selecting atth ediploi d level,resynthesizin g thepolyploi d form and testingit . The exploitation of 'unreduced gametes' or rather '2ngametes 'ha s been the subjecto fman y studies ofPelogui nc.s. .The y investigated thecytolo ­ gy of 2n gamete formation and distinguished between two meiotic restitu­ tions: first division restitution (FDR) and second division restitution (SDR). FDR gametes transmit the genotype ofth eparen t largely intact,in ­ cluding thenon-additiv e gene actions.Thei rus eprovide s apowerfu l breed­ ingtechniqu e (Wange t al., 1971). SDR increases homozygosity. The cytology and geneticso f2 npolle n formationha sbee n studied furtherb yMo k &Pelo ­ guin (1975). According to them three mechanisms of2 npolle n formation can be distinguished and they are simply inherited. FDRmaintain s existinghe ­ terozygosity, which is important in autotetraploids. Mendoza & Haynes (1974) showed that heterozygosity was more important than dominance for yield inpotato .Dunbie r& Bingha m (1975)demonstrate d inalfalf a thattri - and tetra-allelic loci increased yields.Doublin g ofth echromosom e number either spontaneously orwit hcolchicin e increases the degree of homozygosity andmigh t influenceyiel d innegativ e direction. Tous e thepotentia l of2 ngamete swit hFD RMendibur u et al. (1974)pro ­ posed two ways of sexual polyploidization in potato breeding: unilateral (4x x 2x) and bilateral (2x x 2x).I n a 4x x 2x cross an adapted variety would be crossed with a selected diploid,preferabl y with FDR, toobtai na highheteroti c response (Mendiburu& Peloquin , 1971). Inreciproca l crosses between 4x and 2x plants Kidane-Mariam & Peloquin (1974) associated the higher yields in 4x x 2x crosses with FDR in the diploid. It seems,how ­ ever, that negative interaction between S. tuberosum genes and S. phureja cytoplasm couldno tb erule d outcompletely . Thepropose d 2xx 2 xcros s (Mendiburue t al., 1974)i s amor e interesting type. It can be applied in manyways .Fo r instance,a combinatio n ofdiha - ploid S. tuberosum, s. phureja, dihaploid S. andigena and S. chacoense was suggested: S. chacoense was incorporated toavoi dcytoplasmi cmal esterili ­ ty. Breeding was to be done at the diploid level, followed by two crosses between twoparent s toproduc e two intermediatediploi d hybrids,whic hwer e

10 again to be crossed to give a tetraploid that could be multiplied vegeta- tively. FDR in 2n pollen and 2n egg formation in the last crosswoul dac ­ complish that the diploid genomes were incorporated largely intact in the tetraploid. Cytological work or test crosses would be needed dur^n^ the process to screen forplant s with FDR (Mendiburu et al.. I.e.). This isa promising breeding scheme, as it can combine genes from four sources. The cytogenetic explanation of FDR and SDR has been questioned by Ramanna (1979),wh o suggested thatsevera l cytological observations,e .g . parallel spindle, could not be interpreted as genetical processes.Hi s owninvesti ­ gations with diverse material also showed that various processes can lead to restitution in meiosis, even in one anther.Unravellin g the genetics of FDRwoul db emor ecomplicate d thanwa s suggested.Eve ni fth egeneti cback ­ ground ofFD Rwer eno ta ssimpl e asMo k &Peloqui n (1975)stated , thehete ­ rosis reported inth eexperiment s iswort h exploiting. Hermsen (1974, 1977)propose d a modification ofthi s scheme:instea d of maintaining the tetraploid vegetatively, it can be reconstituted in the form of hybrid seeds from the vegetatively propagated diploid hybrids.Po ­ tatoes from seeds might carry a promise forth e future intropica l regions whereviru s is ayea r roundproble m andwher e enoughlabou r is available to transplant seedlings from anurser y to the field. Screening for resistances in vitro is a recently developed method in which protoplast techniques are used. Dihaploid potato cellshav ebee nre ­ generated intoplant s (Behnke, 1975). Behnke (1979)selecte d irradiated and non-irradiated dihaploid calli for resistance to Phytophtora infestans in vitro. Plants from surviving calli proved to be more resistant than con­ trols.Wit h refinement ofi nvitr o techniques selectionmigh t inth e future takeplac e inlaboratorie s rather than infield s orgreenhouses . Wenzel et al. (1979) incorporated such laboratory techniques in their proposed potatobreedin g scheme.Dihaploid s were tob eproduce d fromtetra - ploids through pseudogamy. Monoploids would be madevi a anthercultur e and doubled to obtain homozygous diploids. Selected diploids should then be crossed in pairs. Protoplasts from their progenieswer e tob e isolated and fused inpairs .Afte r regeneration highly heterozygous tetraploids would be obtained, provided that the diversity of the original tetraploids was big enough.

2.3 INFLUENCEO FTH EPARENT SO NFREQUENCIE S OFHAPLOID S FROM PSEUDOGAMY

Pseudogamy is the development of seeds after pollination, but without fertilization ofth eeg g cell.Pseudogam y is sometimes calledparthenogene ­ sis, but this indication is less precise, as it also includes cases where no pollination is involved. Pseudogamic haploidy oftenoccur s after inter­ specific or intraspecific crosses where different ploidy levels are in­ volved (interploidal crosses) (Rowe, 1974).

11 The roles of the two parents in haploid production are very different. Accordingly it is less important to know the genetics of the seed parent effect than the genetics of the pollinator effect,i norde r to improve the efficiency of haploid production. If haploids are wanted from a certain plantthe nth eyiel d ofhaploid sca nonl yb echange db y choosing otherpol ­ linators. Pollinator improvement is possible without influencing the type ofhaploid s produced, as the haploid doesno treceiv echromosome s from the sperms. Several measurements of dihaploid frequencies have been used for com­ parisonsbetwee nparent s orbetwee nmethods .Frandse n (1967)gav e threedif ­ ferentratio s inhi s tables:dihaploid spe r 100pollinate d flowers,pe r 100 seedlings and per 100 berries. He used only the last one in his summary. Irikura (1975b) gave the same three ratios in his tables,bu t used diha­ ploids per 100 pollinated flowers in histext .Thi srati o isa measur e for the effort needed to obtain dihaploids. However, adverse weather condi­ tions, which cause pollinated flowers or young fruits to drop, influence this ratio very much. There isvariatio n in fruitse tbetwee n S. tuberosum cultivars as well. Ifn o seed marker is available the ratio per 100 seed­ lings is to be preferred (Bender, 1963) as raising and screening of seed­ lings is the most laborious part of the production of dihaploids in that situation.A s thenumbe r ofseed spe rberr y is ahighl yvariabl e character, dihaploidspe r 100 seedlings isno tappropriat e forcomparison s betweenpa ­ rents. Dihaploids per 100berrie s (d/100b)i sth emos twidel y usedmeasur e (Ga- bert, 1963;Frandsen , 1967;Hermse n& Verdenius,1973) . Iti s alsouse d in this study. It is not influenced very much by environmental conditions or the seed parent and reflects the number of crosses needed to produce one dihaploid.

2. 3. 1 Influence of the pollinator

Plants producing pollen that gives rise to haploids through pseudogamy are called pollinators, as they pollinate the seedparen twithou t fertili­ zation ofth e eggcell . The closer the parents were related, themor e research hasbee ndon e on the influence of the pollinator genotype onhaploi d frequencies.Maiz eha ­ ploids have been derived from intraspecific 2x x 2x crosses.Whe n its ha­ ploid production was studied systematically, Chase (1949,1952b )foun d the choice of the pollen parent to be important. Coe & Sarkar (1964) came to the same conclusion and they transferred genes for high haploid inducing ability from one stock into another (Sarkar & Coe, 1966). Later Sarkar (1974) reported doubling of the haploid inducing ability of a pollinator populationb y twocycle s ofselectio n fortha tcharacter .Late rh e obtained a furtherdoublin g afterth ethir d cycle (Aman& Sarkar, 1978). Thecharac -

12 terwa s highly heritable andcontrolle d by alarg enumbe r ofgene swit had ­ ditiveeffect . In the cross Populus tremula x p. alba thepolle nca nb e inactivated by toluidineblu e and still inducehaploid s (lilies,1974a,b) . Thegenotyp e of thepolle nparen t isprobabl yunimportant .Bingha m (1971)foun d pollinators to have influence in the interploidal intraspecific crosses that yielded dihaploids in Medicago sativa. Inothe r interploidal crosses, Fragaria ana- nassa x Potentilla anserina and P. fructicosa, whichproduce d Fragaria ha­ ploids, Janick& Hughe s (1974)di dno tmentio n differentgenotype s withina species. In Solanum species pseudogamic haploids are usually induced using se­ lected plants of S. phureja aspollinators . Suchpollinator s fertilize the central nucleus and thus contribute to the indispensable vital endosperm (Von Wangenheim et al., 1960). Dihaploids ofS . tuberosum originate from a 4x x 2x interploidal cross.A s S. phureja isclosel y related to S. tubero­ sum and even S. tuberosum diploids can be used as pollinator (Hermsen et al., 1974a), the interploidal character of thecros sma y bemor e important thanth e interspecific character. Influence of the genotype of the pollinator on haploid frequencies in S. tuberosum has been established by several authors (Gabert,1963 ;Houga s et al., 1964; Jakubiec, 1964; Frandsen, 1967; Hermsen & Verdenius, 1973; Irikura, 1975b). From thewor ko fHermse n& Verdeniu s (1973)i tca nb e seen thatselectio n forhig hd.i.a .i spossibl e inS . phureja. Gabert (1963)wa s the first to screen a large number of S. phureja clones and he found good and bad pollinators. He observed a discontinuity between the two groups, fewbein g goodpollinators .H e crossedpollinator s and obtained goodpolli ­ nators only from the combination good x good. His conclusion was that the characterwa sheritable ,hig h d.i.a.bein g recessive with only oneo r a few genes involved. Frandsen (1967) reported a clear difference between two of thesepollinator sbot h ind.i.a . and inhybri dproduction .Hermse n& Verde ­ nius (1973) produced a sib-F2 population of S. phureja and studied 29 plants of this population extensively for their d.i.a..Th e averaged.i.a . of these plants was higher than thato fthei rparent s andmuc hhighe r than in Gabert's (1963) experiments. Although they found a large variation in d.i.a., no clear-cut discontinuity was detected. Irikura (1975b) studied the selfprogeny of a S. phureja clone. He foundbot hpositiv e and negative transgression for d.i.a. and concluded that low d.i.a. was the dominant character and inherited quantitatively. This conclusion didno texplai n the negative transgression inhi s inbredpopulation .

The genes forhig hd.i.a .wer e recessive according tobot hGaber t (1963) and Irikura (1975b). Irikura analysed one population and Gabert (unpubl.) used rather small populations for his analyses. None of the authors men­ tioned arguments fordominanc e ofhig h d.i.a.. Intermediate inheritance was never considered, as far asth eautho r isaware .

13 Berries fromcrosse sbetwee n S. tuberosum and S. phureja usually contain hybrid seeds apart from seeds with dihaploid embryos.Hermse n & Verdenius (1973) distinguished pollinators in those with high seed set and low seed set. They suggested a relationship between seed set and dihaploid numbers perberry ,especiall ywhe nusin glo wsee d setpollinators . A good pollinator needs amarke r tob euseful .Th e firstmarke r applied in screening for dihaploids in S. tuberosum was purple hypocotyl, used by Peloquin & Hougas (1959). They suggested to use embryo spot as a seed marker foreve nearlie r screening.Thi sspo ti sa naccumulatio n ofanthocy - anin at the base of the leaves and iscause db y twogenes :B andP (or R). Dominant P provides the anthocyanin (also in the hypocotyl), dominant B causes the accumulation, provided P is present (Dodds& Long, 1955, 1956). The character is pleiotropic: the spots are visible on all parts of the plant homologous to a leaf base. The dominant alleles of the non-linked genes B and P are not present in European cultivars of S. tuberosum. At­ tempts have been made to produce S, phureja plants homozygous for embryo spot, which initially failed (Frandsen, 1967). Hermsen& Verdeniu s (1973), however,succeede d inbreedin g S. phureja plantswhic hcombine d homozygosi­ tyfo rembry o spotwit hhig hd.i.a. .Thi s increased the efficiency ofdiha ­ ploid induction considerably and also made selection for monoploids pos­ sible,wher e thousands ofseed shav et ob e screenedt o obtain onemonoploi d (VanBreukele ne t al., 1975).

2. 3. 2 Influence of the seed parent

An influence of the seed parent on the haploid frequency is expected rathertha na neffec to fth epollinato r (Hougas et al., 1964). The seedpar ­ ent influence could work through nuclear genes, through the cytoplasm or both. Lethal genes can reduce the numbero fviabl ehaploids .Th enumbe r of ovules per berry might influence the number of haploids directly. Differ­ ences innumber s ofovule s inovarie s ofsevera lpotat o cultivars have been found (Arnason, 1943;Carrol l & Low, 1975). The influence of lethal genes is difficult to determine, unless lethality occurs during the germination of the seeds or after emergence of the seedlings. Montelongo-Escobedo (1969) found lethal genes not tob e amajo r factori nth epotat o clones he used. Hermsen et al. (1978a) found four lethal genes in Gineke, which is still agoo d dihaploidproducer .Letha l genes,whic hcaus e seed abortion of dihaploids,usuall yescap edetection . Genes for pseudogamy can work at either of two levels: sporophytic or gametophytic. In other words, the mechanism to produce haploids is con­ trolled by the genotype of either theplan to r theeg gcell . Bender (1963) ascribed whath e called an •apomictic tendency' to the egg cell, but this wasmor e anam etha na nexplanation .H epropose d anexperimen tt oprov ega ­ metophytic determination. If the egg cell has ageneti c constitution which

14 promotes pseudogamy, then doubling the chromosome number of ahaploi dwil l result in plants in which genes for pseudogamy have accumulated. Such plants are expected to produce on the averagemor e haploids thanth eorig ­ inalplant .Chas e (1952a)foun d doubled monoploid lines toproduc emor emo - noploids thanunselecte d inbred lines. Differences in haploid producing ability between maize stockshav ebee n reported (Chase,1949 ,1952b ;Sarka r& Coe,1966 ;Sarkar , 1974). Coe& Sar - kar (1964)wer e able to incorporate the high haploid producing ability of their best seed parent into another stockb ybackcrosses .Homozygosit y in­ creased the haploid frequency in maize (Sarkar, 1974). This might be due either to homozygosity for genes for haploidproductio n ort oreductio n of the number of lethal genes. Bingham (1971) found differences in haploid producing ability between Medicago sativa genotypes, Stringham & Downey (1973) in Brassica napus. Differences between Nicotiana tabacum lines in haploid frequencies fromcrosse swit h N. atricana were foundb yBur k et al. (1979). Potato dihaploids can be extracted from almostan y seedparen t (Gabert, 1963). Frandsen (1967)investigate d 117clones .Althoug h 13di dno tactual ­ ly produce dihaploids, he himself was not convinced that they were unable to produce them. Irikura (1975b) extracted dihaploids from 66 out of 83 seed parents. The general occurrence of dihaploid producing ability (d.p.a.)i n S. tuberosum means thata broa d geneticbas ema yb emad eavail ­ able for the application of dihaploids. However, dihaploid frequencies vary. The influence ofth e seedparen to nth e frequency ofdihaploid s hasbee n recognized from the beginning of systematic work onpotat o dihaploids (Ga­ bert, 1963;Jakubiec , 1964;Rothacke r et al., 1966;Frandsen , 1967;Irikura , 1975b). Genetic factors that influence dihaploid production have not been studied extensively, autotetraploidy being ahindrance .Montelongo-Escobed o (1968)mad e crosses between dihaploids of a good and a bad seed parentt o determine the inheritance of the seed parent effect. The meiotically doubled tetraploid progeny was again used as seed parent. From the diha­ ploid frequencies obtained he concluded that highd.p.a .wa s dominant,re ­ latively few genes being involved. However, positive genes are overrepre- sented in dihaploids from a bad seed parent, if Bender's (1963) theory is correct. The tetraploid-diploid-tetraploid cycle concentrates good genes, butt o alarge rexten ti nba d seedparent s thani ngoo d seedparents . Apart from chromosomal influence there may be a cytoplasmic influence. Frandsen (1967) found the cytoplasmic differences between Solanum species to be important. His own data didno tsuppor t thisver y strongly.Th e uni­ formitywithi n each cytoplasm groupwa sno tver y high andrange swer eover ­ lapping: e.g. the group with S. demissum cytoplasm yielded 15-233 d/100b (mean: 65d/100b) and with S. tuberosum cytoplasm 15-433 d/100b (mean: 150 d/100b). Frandsen did not come to a conclusion about cytoplasmic differ­ enceswithi n S. tuberosum. 15 2.3.3 Interaction between pollinator and seed parent influence

Pollinators can only be evaluated using seedparents ,see dparent s only by using pollinators.Th e idea of interaction between pollinator and seed parent effect on dihaploid frequencies did not receive much attention in earlier studies. Gabert (1963), who studied both effects, gave only a slight suggestion that good pollinators were good irrespective of the seed parent. Hougas et al. (1964) stated this more clearly. This same tendency could be observed in the data ofother swh oworke dwit h severalpotat ova ­ rieties anda fe wpollinator s (Jakubiec,1964 ;Frandsen , 1967;Hermsen ,un - publ.). Onth eothe rhand ,Hermse n& Verdeniu s (1973)foun d the correlation between dihaploid production of two varieties with a range of pollinators not to be highly significant. They explained this as a differential reac­ tion of the two seed parents toth e sameserie s ofpollinators .Thi s could be interpreted as interaction. The different frequencies of hybrids found byFrandse n (1967)betwee n S. tuberosum and S. andigena with two S. phureja clones could also be seen as aninteractio nbetwee n seedparen t andpolli ­ nator. Chase (1949) noticed that certain seed parents in maize produced more monoploids than others with each of the pollinators used. The same can be concluded from data of Coe & Sarkar (1964). Inthei r data amultiplicativ e effect can be found but no interaction in a strict sense between the ef­ fectso fsee dparen tan dpollinator .

2.4 EXTERNAL INFLUENCEO NHAPLOI D FREQUENCIES

It is apparent from Section 2.3 that haploid production is genetically controlled. Inthi s sectionliteratur e onth einfluenc e ofexterna l factors on the frequency of haploids is presented. These factors include tempera­ ture, light,radiatio n andchemicals . Extreme temperatureshav einduce dhaploid s insom especies .Blakesle e et al. (1922)obtaine d their first haploids in Datura stramonium by using low temperatures. Müntzing (1937) reported a haploid plant in Secale cereale aftercol dtreatmen tan dNordenskiöl d (1939)afte rhea ttreatment . In Sola­ num tuberosum many authors founddihaploi d rates changing fromyea r to year (Gabert,1963 ;Frandsen ,1967 ;Hermsen , unpubl.). Gabert (1963) investigated the influence of temperature on dihaploid frequencies under natural light conditions in an air-conditioned green­ house. Though he stressed thebi g contribution ofth epolle nparen t todi ­ haploid frequencies, he controlled temperature influences on the female parent only. He used small compartments with temperatures ranging from 15-30 °C. He pollinated cut stem inflorescences (Peloquin & Hougas, 1959) and placed them for four days ina compartmen twit heithe r fixed oralter ­ nating temperatures. None of the treatments enhanced dihaploid frequency,

16 when compared with the frequency found at gradually changing temperatures of 21-24 °C during the day and 13-16 °C at night. Gabert (I.e.)conclude d thattemperatur e probably affected fruit setan d fruitdevelopment . Wöhrmann (1964)als ocarrie d outsystemati c experiments todetermin e the influence of temperature on the production of dihaploids in potato, es­ pecially duringpolle n tubegrowth .H e used thecu t stem technique.Flower s of fourcultivar swer eplace d ata temperatur e of20 ,2 5 or 30 °C for 10o r 20h immediately after pollination. For the remainder of the first eight days afterpollinatio n all flowerswer e kepta t2 0 °C.H e found thathighe r temperatures lowered thenumber s ofberrie spe r flower,seed spe rberr y and dihaploids per berry considerably. The number of dihaploids per seed was lowered by higher temperatures to a smaller extent. Duration of treatment seemed tohav emor e influence thantemperature . His treatments were rather short if one considers that pollen tube growth takes 36-48h unti l themomen to f fertilization.H e also didno tex ­ clude influences on the pollen parent and on the seed parent before an- thesis. The two treatments of 10 and 20h at2 0 °C, each followedb y 20° C for eight days, were not considered asreplications .Ther ewa s aconsider ­ abledifferenc ebetwee nthes e almost identical treatments for eachcultiva r withrespec tt onumber so fseed spe rberr y anddihaploid spe rberry .Appar ­ ently therewer e factors thatwer eno tcontrolle d inhi sexperiments . Frandsen (1967) was the first to note that seasonal influences before the period of fertilization might affect dihaploid rates,bu tdi dno tmen ­ tion any influence ofenvironmenta l factors onth ed.i.a .o fpollinators . Meiosis takes place in the anther beforemeiosi s inth e ovule in S. tu­ berosum (ReesLeonard , 1935;Clarke ,1940 )an d in S. phureja (Maherchandani & Pushkarnath, 1960). Thismean s thatenvironmenta l influences onbot hmei - oticprocesse s canb edifferent . An important aspect of environmental influence is formedb y the growing conditions of the parental plants as far as they influence flowering and fruit set and indirectly the dihaploid frequency (Gorea, 1968, 1970). The decapitation technique, described by Peloquin & Hougas (1959), increased fruit set five to ten times. Grafting of potato onto tomato rootstock re­ sulted inprofus e flowering over alon gperio d and improved berry set,thu s increasing the efficiency of greenhouse space and labour (Hermsen, 1979). Gorea (1968, 1970, 1973) concluded from his potato dihaploid inductions thatth etota lyiel d ofdihaploid s waspromote d by suitable culturalcondi ­ tions during fruit development. Fruit set can be reduced by high tempera­ tures during flowering (Bienz,1958 )o r lownigh ttemperature s (Bodlaender, 1960). Most Solanum species flower only at long day length, but a high light intensity might compensate for too short day lengths (Driver & Hawkes, 1943; Krug, 1957). This has consequences for the duration of the crossing season and forth echoic eo f artificial conditions ingrowt hcham ­ bers.

17 Apart from manipulations with natural factors astemperatur e and light, technicalmethod shav ebee ntrie dt orais ehaploi d frequencies (Review:La - cadena, 1974). Most methods reduced the vitality of the pollen, others aimed atinducin g restitution ofth epolle ntub emitosis . Kopecky (1960)reduce dpolle nvitalit y of Populus alba by a fermentation process. He obtained more haploids after pollinations with this pollen. Heat-treatedpopla rpolle nproduce d haploid-diploid mixaploids and maternal diploids, probably doubled haploids (Winton & Einspahr, 1968). lilies (1974a,b) obtained significantly higher frequencies of haploids after treatment ofpolle n orth epollinate d stylewit htoluidin e blue.Th e pollen tube mitosis was blocked and the subvital pollen was only a stimulus. The endosperm was apparently notimportan t forsee ddevelopmen t in Populus. To­ luidine blue did not give an increase of haploid frequency in tomato and maize (AlYasir i & Rogers, 1971). Dumas de Vaulx & Pochard (1974) treated pollinated flowers of Capsicum annuum with N20. This treatment increased the frequency of haploid-diploid twins in one cultivar. Treatment of the flowers before pollination resulted in the production of single haploids. Ecochard et al. (1974)inactivate d themal e sperm specificallywit h thermal neutrons just before fertilization and obtained two tomato haploids with thistechnique . Severalworker s appliedX-ray s to Solanum pollen.Bende r (1963)obtaine d increased dihaploid frequencies,wherea sWöhrman n (1963)reporte d similar or lower frequencies ofdihaploid spe rberry .Buka i (1973)obtaine d a fourfold increase of the dihaploid frequency. Tarasenko (1974) reported a sixfold increase in dihaploids per seed. Berry set and seed set were found to be lower after X-ray treatment by Bender (1963)an dWöhrman n (1963). Itseem s that dihaploids are less sensitive to X-rays than hybrid seeds, probably because irradiated pollen can still be functional inendosperm , when iti s not functional any more in a zygote.Montezuma-de-Carvalh o (1967)an dMon -

telongo-Escobedo & Rowe (1969)treate d S. phureja pollenwit hN 20 and col­ chicine respectively in order to increase dihaploid production by stimu­ lating restitution of the pollen tube mitosis. Increased restitution was observed in both studies and in the latter also increased dihaploid fre­ quencieswer e found (seeals oSectio n 2.5.2.3). It may be concluded that optimalization of the growing conditions of seed parents and pollinators was the onlymanipulatio n ofexterna l factors thatconsiderabl y increased dihaploid frequencies ofa see dparen tpollina ­ torcombination .

2.5 MECHANISMSO FPSEUDOGAMI CHAPLOI D PRODUCTION

When haploids are formed through pseudogamy the reproductive mechanism is switched to an abnormal development.Man y authors have studied haploid producing mechanisms applying embryological and cytological techniques. Literature about these mechanisms is given here for 2x x 2x and 4x x 2x crosses with emphasis on the latter, especially in Solanum. The origin of hybrids isals otake n intoaccount .

2. 5.1 2x x 2x crosses

The mechanism of pseudogamy is extensively studied in maize.Maiz e ha- ploids may occur in low frequencies in progenies of 2x x 2x crosses.Mos t cytological data come from experiments by Chase (1964b) and Sarkar & Coe (1966). Chase used a dominant endosperm marker,whic h regulated theinten ­ sity of yellow colour according to thenumbe r of allelespresent .Wit h the marker in one of the parents he could distinguish triploid from tetraploid endosperm. He found monoploids to have triploid endosperm. Sarkar & Coe (1966) confirmed this using another endosperm marker.Apparentl y one sperm fertilized the central nucleus and the other was lost. This is called the 'maize mechanism' (Hermsen, 1971). Chase (1969) proposed a hypothesis of early division of the egg cell or synergidbefor e fertilization to explain this. The tendency of precocious division would depend on the maternal genotype.Thi sma y leadt o thehighe rmonoploi d rate observed after delayed pollination. Smith (1946) found a higher haploid frequency after delayed pollination in Triticum monococcum.

2. 5.2 4x x 2x crosses

Pseudogamic dihaploids from S. tuberosum originate almost exclusively from 4x x 2x crosses.Thi s is an interesting type of cross asth e outcome isno ttriploi d hybrids,whic h ismathematicall y theexpectation . Triploids occur in very low frequencies per berry. The majority of the progeny con­ sists of tetraploid hybrids. Dihaploids are produced as well in certain combinations of parents, sometimes in frequencies ofu p to fivepe rberry . The total number of seeds produced is usually very low compared with that in2 xx 2x and4 xx 4 xcrosses .

2.5.2.1 Triploid hybrids

The lack of triploids in a 4x x 2x cross has been reported frequently and was described as 'triploid block' by Marks (1966a). He called it an enigma, astriploi dplant s growwell .Koopman s& Va nde rBur g (1952), Marks (1966b), Van Suchtelen (1966, 1973) and Hanneman & Peloquin (1967) men­ tioned low triploid frequencies in their crosses. Rothacker & Schäfer (1961) found 40 triploids per 100 berries. Frandsen (1967)obtaine d much lower frequencies:aroun d 1pe r 100berries .Hermse n& Verdeniu s (1973)re ­ ported relatively high triploid frequencies from crosses involving low seed set pollinators. Calculations on their data show that these pollinators

19 produced 15-30 triploids per 100 berries. Triploid frequencies from high seed setpollinator s could be of the samemagnitude .Jackso n et al. (1978) found similar frequencies in Andean potato cultivars. Hanneman & Peloquin (1968) and Ross & Jacobsen (1976)observe d lower frequencies of triploids inbigge rprogenies ,whic h isno tunexpecte d as theyexpresse d the triploid frequency as a fraction of total seed set.Triploi d frequencies can vary from year to year (Sudheer, 1977) and be influenced by the parental geno­ types (Gorea,1970 ;Va nSuchtelen , 1976). Beamish (1955) crossed hexaploid S. demissum with diploid S. phureja. She obtained less than one seed perberry ,predominantl y tetraploids,com ­ parable to the triploids in a 4x x 2x cross. In other genera low triploid frequencies in 4x x 2x crosses were found, e.g. Dacti/lis (Carroll & Bor- rill, 1965), Brassica (Nishiyama & Inomata, 1966), Primula (Skiebe, 1967) and Medicago (Bingham, 1971). A barrier to abette r understanding ofth etriploi d blockwa s theopin ­ ion, that fixed ratios between the ploidy levels ofmaterna l tissue,endo ­ sperm and embryo were necessary for the development of a seed. This ratio should be 2:3:2 or 4:6:4 (Miintzing, 1930). Cooper &Brin k (1945)foun dir - regularaties in Lucopersicon endosperm after crosses between 4x and 2x plants, followed by poor embryo development leading to shrivelled seed. They stressed the importance ofth erati obetwee nmaterna l tissue andendo ­ sperm. Study of endosperm development by VonWangenhei m (1961)showe d that the absolute ploidy level of the endosperm was important. He found 3x, 6x and 9xendosper m tobehav enormall y during seed growth, irrespective ofth e ploidy level of the embryo, sometimes even without an embryo. In 4x endo­ sperm he observed small cells without vacuoles and with large intercellu- lars. Pentaploid endosperm had its own characteristic type of degradation as it did not form cell walls.Simila r defective endosperm developmentha s been foundi n4 xx 2 xcrosse s in Brassica byNishiyam a & Inomata (1966). It isth e lacko fvitalit y of5 xendosper m thatmake s the seed setlo w in4 xx 2x crosses. The majority of ovules, containing triploid embryos, does not develop beyond initial stages. Skiebe (1973) supported the conclusions of VonWangenhei m (1961). Exceptions ofthi s theory havebee n explained byVa ­ lentine & Woodell (1963) by introducing a 'genetic value', a factor that makes endosperm chromosome numbers a multiple of3x .Th evalue s they found look unnatural. Den Nijs (1977) and Den Nijs & Peloquin (1977) introduced ahypothesi s aboutendosper mbalanc e factors (EBF)t oexplai nth e different crossing behaviour of Mexican tetraploids and S. acaule compared to other tetraploid Solanum species.Endosper m would only develop normally ifther e were a 2:1 ratio between maternal andpaterna l genomes.A tetraploid would normally have 4EBF , but S. acaule would have only 2EB F and therefore easier produce triploids in crosses with diploids than tetraploids in crosseswit h tetraploids.Johnsto n et al. (1980)introduce d thenam e 'endo­ spermbalanc e number' forth enumbe ro fEB Fo fDe nNij s (I.e.).

20 2.5.2.2 Tetraploid hybrids

Most hybrids from tetraploid S. tuberosum x diploid S. phureja crosses are tetraploid (Frandsen, 1967;Höglund , 1970;Hermse n& Verdenius , 1973). Höglund (I.e.)wante d to explain the high number of tetraploids from 4xx 2x crosses, she studied microsporogenesis and pollen mitosis in vitro of one S. phureja clone. The spindle formation inmetaphas e IIwa s irregular, resulting in about 50%2 n gametes.Afte r second pollen mitosis she found, apart from tubes with two reduced sperms, also tubes with two 'unreduced' sperms, the 2n sperms contributing to tetraploid hybrids. She suggested a cytoplasmic influence from the tetraploid pistil or diploid eggcel l lead­ ing to preferential fertilization by 2n pollen grains. She mentioned the possibility of differences in growth rate between 2n and n pollen. This should explain why the frequency of tetraploid progenywa s higher thanth e frequency of 2n pollen. She did nottak e into accountth e lethality ofth e triploids. Preferential fertilization is aver y rare phenomenon according to Haustein (1967). Differences in pollen growth rate, called 'certation' by Heribert-Nilsson (1920), are more common. The term was originally used for types of pollen differing in genotype,bu tca n alsob e used forpolle n differing inploid y level.Skieb e (1966)di dno t find differences ingrowt h rate inviv o for2 x andx polle no fdiploi d Primula malacoides, neither did Esen et al. (1978) in Citrus. In sugar beet x pollen grows fastertha n 2x pollen (Matsumura, 1958). Ramanna (1974) found aberrant cytokinesis to be the major mechanism in the occurrence of 2n pollen. Mok & Peloquin (1975) distinguished three mechanisms of 2n pollen formation in diploid potatoes on cytological grounds. Regarding the genetic consequences there are two mechanisms. All these mechanisms lead to tetraploid hybrids from tetraploid-diploid crosses. In a re-examination of the mechanisms of 2n pollen formation Ra- manna (1979)doubte d whether cytological observations couldb e interpreted genetically inth ewa yMo k& Peloqui n (I.e.)did .

2.5.2.3 Dihaploids

Dihaploids may originate from 4x x 2x crosses as well, but by which mechanism? If the maize mechanism (Section 2.5.2.1)wer e involved, there ­ sulting seeds would have dihaploid embryos in lethal pentaploid endosperm and dihaploids would occur in even lower frequencies than triploids. Von Wangenheim et al. (1960) investigated the chromosome numbers in11 2 ovules of S. tuberosum, pollinated by a S. phureja clone with high d.i.a.. They counted 35 ovules with hexaploid endosperm, of which 30 lacked an embryo and 5containe d adihaploi d embryo.Bende r (1963)di d similar counts in 614 ovules. The chromosome numbers of both endosperm and embryo could be

21 counted in11 4ovules .Fou rdihaploi d embryoswer e found inhexaploi d endo­ sperm. Three other dihaploid embryos were found in an endosperm of which chromosomes could not be counted. Again ahig hnumbe r ofovule swit hhexa ­ ploid endosperms without embryo was found. These nine dihaploids inhexa ­ ploid endosperm are the only ones in which the ploidy level of embryo and endosperm hasbee ndetermined .Neve rwa s adihaploi d mentioned inothe ren ­ dosperm. Itca nb e assumed thatdihaploid s generally occur inhexaploi d en­ dosperm. A dihaploid in pentaploid endosperm has only a small chance to survive, but with the survival rate of triploids of about 1 in 1000 (Von Wangenheim, 1956)i ti spossibl e that anoccasiona ldihaploi d inpentaploi d endosperm occurs, the frequency of which being too low for detection by chromosome counts in ovules. A hexaploid endosperm will have received 2x chromosomes from the pollen. If the egg cell is not fertilized, there are basically two possibilities for the central nucleus tob e fertilized by 2x chromosomes:eithe rb y 2npolle n orb y reducedpollen . Incas e of2 npolle n either one sperm fertilizes the central nucleus and the otherdegenerates , or - as Bender (1963) suggested -onl y one single 2n sperm is available to fertilize the central nucleus due to failing pollen tube mitosis. With n pollen twosperm s fertilize thecentra l nucleus either individually orcon ­ nectedb y achromati dbridg e or fused afterrestitution . Fertilization ofth ecentra lnucleu s only and degeneration ofth e second sperm has been observed cytologically. Hâkansson (1943) reported a prema­ ture division of the egg cell in Poa alpina, followedb y the fusiono f one spermwit h thecentra lnucleus ,whil e the other sperm degenerated.Banniko - va & Khvedynich (1974)compare d the fertilizationproces s of intraspecific crosseswit htha to fwid e crosses in Nicotiana. Inth ewid e cross N. rusti- ca x N. paniculata one sperm fused with the egg cell or the central nu­ cleus, theothe rdegenerate d often ino rnea r a synergid. The general occurrence of 2n pollen in S. phureja pollinators could be an argument in favour of the role of this pollen in dihaploid induction, but then it has to be explained why the second sperm does not function. Prematuredivisio no fth e eggcel l isno t aslikel y here as iti s inmaize . Guignard (1902) reported that the development of the embryo in Solanum started after several endosperm divisions. This was confirmed by Clarke (1940), who observed the first endosperm divisions twoday s afterpollina ­ tion and the first division of thezygot e after five days in S. tuberosum. A similar time interval was recorded by Williams (1955). Delayed pollina­ tion to exploit the tendency of premature division ofth eeg g cell did not increase dihaploid frequencies inGabert' s (1963)work . X-ray treatment (Bender, 1963;Wöhrmann , 1963;Bukai ,1973 )an dcolchi ­ cine (Montelongo-Escobedo & Rowe, 1969) increased dihaploid frequencies in S. tuberosum. Iti sno timpossibl e thaton e spermwa smad e subfunctional in theprocess .Thi swa y 2npolle nmigh tinduc edihaploids .

22 Double fertilization of the central nucleus has been reported in some species, often when apomixis was involved. Gaines& Aas e (1926)discovere d a Triticum haploid originating from a large kernel.The y assumed that the giant endosperm was caused by a fusiono fbot h spermswit h thecentra lnu ­ cleus. Rutishauser (1956)foun d double fertilization ofth ecentra l nucleus in pseudogamous Ranunculus auricomus. Fused sperms and sperms connected by abridg ehav ebee nobserve d in S. phureja. Bender (1963)foun d occasionally bridge formation between two sperm nuclei and restitution nuclei invivo . The frequency was higher after irradiation. Montezuma-de-Carvalho (1967) observed themetaphas eplat e tob e absenti npolle n tubemitosi s of S. phu­ reja invivo .H e found thechromosome s to liei nrows , anarrangemen t easi­ ly leading to abnormalities.Montelongo-Escobed o& Row e (1969)studie d pol­ len growth in vitro. They found one single restitution nucleus in 35%o f the tubes in two good pollinators, in bad pollinators only in 3%. These percentages need not be a reflection of what happens in vivo, as the authors found abnormalities in their in vitro pollen germination, e.g. a secondpolle nmitosi s inside apolle ngrain . Complete or partial restitution during pollen tube mitosis occurs in S. phureja. Several attempts have been made to increase the restitution frequency. Montezuma-de-Carvalho (1967) applied N20 to pollinated flowers atth etim e ofpolle n tubemitosis .Afte r treatmenth e found up to70 %res ­ titutionnucle i ordumb-bel l shaped nucleiwit h abridg ebetwee n sperms due to lagging chromosomes. Montelongo-Escobedo & Rowe (1969) gave different colchicine treatments to pollen before using iti npollinations .Frui t set was reduced, seed set varied and the dihaploid frequency was slightly in­ creased.Treate d pollen germinated invitr o showed 100%restitutio nnuclei . Treated pollen fromba dpollinator s didno tyiel d asman y dihaploids asun ­ treated pollen from good pollinators. Treatment doubled theperformanc e of good pollen. The origin of the hybrids obtained wasno texplaine d inthei r experiment. Some pollen must have escaped thetreatment .Bende r (1963)in ­ creased dihaploid frequencieswit hX-ra y treatmento fpollen . Inaccompany ­ ing cytological studies he found an increase in restitution nuclei and chromatid bridges from 4 in 1000 tubes studied to 10 in 800 tubes at the optimum radiationdose . These are someexperiment s thatsuppor tth etheor y that fertilization of the central nucleusb y two (fused)n sperms isth emechanis m leading todi ­ haploid formation. This was called the 'Solanum mechanism' by Hermsen (1971). Bender (1963) stated on theoretical grounds that n pollen must be im­ portant as otherwise more dihaploids would have been found from 4x x 4x crosses in potato, apart from the dihaploids claimed by Cooper & Rieman (1958)an dRiema n et al. (1959). This isno tnecessaril y true,a s dihaploids would be difficult to detect in the large progenies of 4x x 4x crosses. Apart from that, nobody ever selected for good tetraploid pollinators as

23 wasdon e fordiploi dpollinators . The role of 2n pollen could be determined by the correlation between frequencies of 2n pollen and dihaploid induction for several pollinators. This can be done directly or indirectlywhe nhybri dproductio n is takena s a measure for the frequency of 2npollen .Th e triploid frequencype r berry canb eneglected .Rothacke r etal . (1966)di dno t find correlations between dihaploids andtota l seedset .Hermse n& Verdeniu s (1973)presente d results of dihaploid induction with 29 S. phureja clones. They did not calculate between-pollinator correlations, but only within-pollinator correlations between berries. Buketova & Yashina (1973)reporte d absence of correlation between the percentage of dyads in nine pollinators and the induced diha­ ploid frequencies. Also these results point to the importance of reduced gametes.

24 3 Pollinator influence on dihaploid and hybrid frequencies

3.1 INTRODUCTION

S. tuberosum dihaploids are usually produced by pseudogamy using se­ lectedplant s of S. phureja aspollinator . Suchpollinator s dono tcontrib ­ ute toth egenotyp e ofth edihaploid ,bu t fertilize thecentra l nucleus and thus contribute to thegenotyp e ofth eendosperm . Thenatur e ofth estimu ­ lus to the egg cell causing cell division without fertilization is not known. As described in Section2.3. 1 the influence ofth e genotype ofth e pol­ linator on dihaploid frequencies was well established insevera lstudies . Frequencies of hybrid seedswer e also influenced by thepollinator .Reces ­ sive inheritance of dihaploid inducing ability (d.i.a.)i npollinator s was reported by Gabert (1963) and Irikura (1975b). The former mentioned that only one or a few loci were involved. Dominance and intermediate inheri­ tance were never reported.A s d.i.a.migh tb ebase d ona gametophyti c pro­ cess in which interaction between alleles is absent because of the ha- ploidy, intermediate inheritancemus tb e considered. Thepositiv e trangression ind/100 b foundb y Irikura (1975b)an d these ­ lection response obtained byHermse n& Verdeniu s (1973)rais e the question whether d.i.a.ca nb e further improvedb ybreedin g and towha textent . Hybrid seeds are aby-produc to fdihaploi d induction.A s dihaploids and hybrids are formed fromth e same limited number ofovule s in aberry , there must be a relationship between the number of dihaploids and hybrids pro­ duced by a pollinator. This relationship is of interest since the fre­ quencies of hybrids mightinfluenc e the frequencies ofdihaploids .Hermse n & Verdenius (1973) found a 3:1 ratio forlo wvs .hig hhybri d frequencies. They suggested amonogeni c dominant inheritance of lowhybri d frequencies, but didno texclud e thepossibilit y ofmor e genesbein g involved. The objectives ofth e experiments described inthi s chapterwere : - to confirm the influence of the pollinator on dihaploid and hybrid fre­ quencies, - to study the process of inheritance of d.i.a. and hybrid frequencies in thepollinator , - to find out whether with the existing S. phureja genotypes the maximum d.i.a.ha dbee n reached, - to study the quantitative relationship betweendihaploi d andhybri d fre­ quencies.

25 Forthes epurpose spollinator swit hknow nperformanc ewer eused , crosses betweenpollinator swer emad ean dthei rprogenie s tested.Th ed.i.a .o fth e pollinators was assessed by making pollinations on a S. tuberosum tester during four seasons.Th e number of dihaploids per 100 berries in the pro­ genywa s themeasur e forcompariso no fth ed.i.a .o fdifferen tpollinators .

3.2 MATERIALAN DMETHOD S

Most S. phureja plantswer ederive d frommateria l developedb yHermse n & Verdenius (1973), who were first to combine high d.i.a. with homozygosity forembry o spot (Section 2.3.1).Th edono r forhig h d.i.a.wa sP I 225682.22 (Gabert, 1963) and the source of embryo spot was PI 225702.2. Full-sib F2 plants from this cross were indicated as iVP-clones. All were homozygous forth ehypocoty lmarke rP ,bu tsom eprove d tob e heterozygous for the gene forembry o spotB .Th eperformance s ford.i.a . and seed setwer e known from work ofHermse n & Verdenius (1973) and confirmed in a preliminary experi­ ment. Genotypes for B and lists of pollinators with high d.i.a. and high seed set are given in Table 1. Full-sib F3 and self progenies were made from IVP-clones. They are listed in Table 2. Reciprocal crosses between IVP48 and S. tuberosum dihaploid G609 (from cv. Gineke) yielded two sib families, IVP48x G60 9 andG60 9x IVP48,whic hwer eheterozygou s forbot h B and P. Plant 57 of G609 x IVP48 was backcrossed with IVP48 as mother, whereasplan t3 wa sbackcrosse d with IVP48a s father. Inthi swa y onepopu ­ lationha d S. phureja cytoplasm andth eothe r S. tuberosum cytoplasm, coded IVP482x G60 9an dG60 9x IVP482 respectively. As seed parent in tests for d.i.a. of pollinators cultivar Gineke of autotetraploid S. tuberosum wasused ,whic hha dn ogene s from other Solanum species. In earlier studies this cultivar had shownth e ability to produce

Table 1. Genotypes forembry o spotan d listso fpollinator s withhig hd.i.a . andhig h seed set,respectively .

Genotype Highd.i.a . High seed set

BBPP BbPP

IVP6 IVP1 IVP1 IVP10 IVP10 IVP24 IVP10 IVP32 IVP26 IVP32 IVP26 IVP35 IVP35 IVP35 IVP48 IVP48 IVP66 IVP71

26 Table 2. Full-sib F3 and selfprogenie s derived from the IVP series ofpollinator s (Hermsen& Verdenius , 1973). Populations indicatedwit h Sar e segregating forembry ospot .

Population Population n

IVP1 selfed 12 S IVP(48x 10) 10 IVP(1x 10) 5 S IVP(48x 35) 24 IVP(24x 10) 18 S IVP48 selfed 5 IVP(32x 48) 47 S IVP(66x 6) 8 IVP(48x l) 5 S n. sizeo fth epopulatio n

many dihaploids. In addition, it flowered profusely for alon gperio d when grafted onto tomato. Itwa s easy to emasculate and produced many berries. The high correlation (r=0.65)betwee n d.i.a. of2 2pollinator s tested with cv. Gineke andcv .Rados a inon eyea r (Hermsen& Verdenius ,1973 )justifie d theus e ofon eteste r cultivar todetermin e therelativ e d.i.a.o fpollina ­ tors. This was corroborated by other tests by Hermsen and by ownprelimi ­ nary trials (VanBreukelen , 1972). All pollinators and seed parents were grafted onto tomato rootstock to enhance flowering.The ywer eplante d ina greenhous e withhig h relativehu ­ midity and temperature seta t2 0 °C.Temperature s didno tdro pbelo w 20 °C, but were occasionally as high as 35 °Cfo r2- 4 h onver y hot days.Flower ­ ingusuall y lasted from lateJun et omi d September. For pollination mostly fresh pollen was used. Sometimes the pollen was collected from flowers that had been dried overnight. All Gineke flowers were emasculated and labelled individually. Pollinations were made in cycles. Ten emasculated flowers of Gineke were pollinated per pollinator until all pollinators had been used. This took several days. Thecycl ewa s repeated until 15-30 berries per combination were obtained. The number of pollinations perpollinato r ranged from 15t o150 . The seeds were carefully extracted from every berry separately and screened for embryo spot with the aid of a dissecting microscope (magn. 20-40X). spotted seeds, spotless seeds and doubtful caseswer e counted and kept apart. The fraction of spotless seeds in the progeny of pollinators homozygous for embryo spot can be regarded as the fraction of dihaploids produced. However, the spotless seeds from pollinators heterozygous for gene B contained also hybrids,whic h showed apurpl e hypocotyl as a seed­ ling. All the spotless seeds from heterozygous pollinators were therefore planted in the next spring, after thedormanc ywa s over.Fro m the emerging seedlings data were taken to enter them in the group of dihaploids orhy -

27 brids.I nth e first two years all theothe r spotless seedswer eplante da s well, together with thedoubtfu l cases. This waymistake s inscreenin g could berectified . Thepercentag e ofemergenc e wascalculate d for all groupso fseeds .Th e germination rate wasver yhigh ,excep t forth edoubt ­ ful cases,wher e theabsenc eo fa nembry oha dbee n suspected. Only emerging seedlings without purple hypocotyl orlea f base were finally recordeda s dihaploids.Thi s group included lethal types thatdi dno t survive the first two weeks.Gree n tetraploid plants constituted less than1 %o fth e plants from spotless seeds.The y were detectedb ythei rwid e leaves and confirmed astetraploid sb ycountin g thenumbe r ofchloroplast s inth eguar d cellso f the stomata. This method reliably distinguishes2 xfro m4 xplant s (Frand- sen, 1968).

From the seedso f197 5 and 1976 onlya sampl e ofth e spotless seeds from Gmekex -homozygous'pollinator s was planted,t odetermin e the percentage ofemergin g dihaploids.Th etota lnumbe ro fspotles s seeds froma cros swa s then multipliedb ythi s percentage togiv ea nestimat e ofth e numbero f emergxngdihaploids .Th epercentage s didno tvar ymuc h fromyea rt o year. As a routée seeds with embryo spotwer eno tplanted . They were counted ashybrids ,togethe rwit h seedlingswit ha purpl ehypocoty lo rlea fbase . Absolute numberso fdihaploid s and hybrids froma cros s were converted tonumber spe r 100berries :d/100 b andhy/100 b Dihaploid numbers inberrie s have a distribution that isclos e to a Poisson distribution (Van Breukelen, 1972). „eans ofnumber so fdihaploid s IZnZ P6r y arS Pr°bably Cl°Ser t0 the n0rmal attribution. Notwith- lTTZYTln deViati°n fr°m n°rmality'aMlysi s° f*«*«" »<-ova ) Polle 7-Wlt h POllinat0r 3nd Year °rCrOSSin * date - -ineffects . UP t0 th e pomnatorTT." " ***"*»* ~ times per season forever y th percent 'T"" ^^ ^ ^^^ -id fuchsinan dcountin g P a re9UlarlyShaP6 dd6ePl yStainS dPOlle n lastmetho d whicI°h aives a fa-ii- ;-J;„ ^ • «»*» •Tbl .i sa poor fJanse n *. u indicationa st owhethe r polleni sgoo do r poor vJansse n & Herm^^n TQT C \ <-. -, grainsca nb ., He~' 1976)- Sam?les with more than 95%staine d pollen grainsca nb e considered a«?VOT-T ,^^„ ^; T at p p llm - -«.- «»in ^:™ r™™ " °" ° 3.3 RESULTS

Results ofcrosse svarie dwit h theyear s pr„, Per±0dS occurred in197 5 and1976 , lowered berry set anHl ' "^ " berry considerably. As for berry set 1974 was ' T °* '*** *" 1973 were good, since most r b6St year and 1971 and », «-<—«^/r::.:::-» ™ ti—i • •—- —

28 Table 3. Dihaploidspe r 100berrie s (d/100b)an dhybri d seedspe r 100ber ­ ries (hy/100b). Berries formed oncv .Gineke , afterpollinatio nwit h seven relatedpollinators .Result s are averages ofaroun d 30berrie s foreac h of threeyear s (upperpart )o r averages of5-1 0 berries foreac h of fourdif ­ ferentday s in197 5 (lower part).

Pollinator (IVPnr. )

10 26 35 48 66 71 d/100b 1968 ,37 53 46 236 111 9 0 70 1975 103 235 304 297 406 46 41 205 1976 192 156 402 237 207 27 28 178 mean 111 148 251 257 241 27 23 151 hy/100b 1968a 7 7838 42 1453 17 117 115 1370 1975 40 465 31 150 39 125 60 130 1976 72 1950 97 1323 25 196 157 546 mean 40 3418 57 975 27 146 111 682

Pollinator (IVPnr. )

6 10 26 35 48 (48 x 35) mean d/100b 18/8 68 84 348 265 290 395 242 21/8 62 115 255 225 376 541 262 25/8 119 330 312 479 568 646 409 28/8 56 367 238 256 401 974 388 mean 76 224 288 306 409 639 324 hy/100b 18/8 0 200 25 73 20 13 55 21/8 0 420 38 38 67 46 102 25/8 50 375 60 211 33 70 133 28/8 20 850 0 378 25 50 221 mean 18 461 31 175 36 45 128 a.dat a calculated fromHermse n& Verdeniu s (1973)

29 3. 3.1 Influence of the pollinator on dihaploid and hybrid frequencies

Cv. Gineke was pollinated with the sameseve npollinator s in threesea ­ sonsan dd/100 ban dhy/100 b werecalculate d (Table3 , upper part). The same data were collected from six pollinators on four different days in one month in 1975 (Table 3, lower part). Analyses ofvarianc e frombot h groups ofdat aar egive ni nTabl e4 .Th einfluenc e ofth epollinato r ond/100 bwa s significanti nbot hcases .I twa sles smarke d onhy/100b ,wher e itwa s sig­ nificant in the four date experiment, but only marginal in the three year experiment. Variations between years and days were high and even signifi­ cantfo rd/100b . Notwithstanding environmental influences,th e ranking of the pollinators was almost identical from year to year: IVP48 > IVP35 > IVP10 = IVP1 =

Table 4. Analysis ofvarianc e fromth edat a ofTabl e 3.

Source Sumo f squares d.f. Probability

Three year experiment

d/100b pollinators 187716 6 0.01 years 70959 2 0.02 error 74718 12 total 333394 20

hy/100b pollinators 28248879 6 0.12 years 5574976 2 0.31 error 25880246 12 total 59704101 20

Four date experiment

d/l00b pollinators 717522 5 0.00 dates 127022 3 0.06 error 209097 15 total 1053641 23

hy/100b pollinators 601120 5 0.00 dates 87535 3 0.16 error 219245 15 total 907900 23

d.f..degree so f freedom

30 IVP26 > IVP6 > IVP24 > IVP71 > IVP32. Thus seven levels of d/100b were found.

3.3.2 Relationship between dihaploid and hybrid frequencies

The different cytological processes leading to formation of dihaploids and hybrids will be discussed in Chapter 7.Her e only the quantitativere ­ lationship between frequencies of pseudogamic dihaploids and hybrid seeds willb e dealtwith . Among somewel l testedpollinator s the levels ofd/100 b andhy/100 b dif­ fered as follows:

low d/100b high d/100b lowhy/100 b high hy/100b IVP24,3 2 IVP10, 35,4 8 IVP24,4 8 IVP10,32 ,3 5

High d/100b did not always coincide with high hy/100b, nor with low hy/100b. The derived populations IVP(24 x 10),IVP(3 2 x 48) and IVP(48x 35) had a wide range for both d/100b and hy/100b and there was no clear positive or negative trend, as can be seen in Fig. 1. Correlation between the two characters and regression of d/100b onhy/100 bwer e calculated for everypopulatio n and forpart s ofth epopulation s (Table 5). Thepatter n in the three populations was similar. From 0-100 hy/100b, d/100b was rising

Table 5. Analysis ofth erelationshi p betweend/100 b andhy/100 b from theuntransforme d datao fFig .1 .n = (sub)population size;r = cor ­ relation coefficient;slop e= th e slopeo fth eregressio n ofd/100 b on hy/100b.

Population Rangeo fhy/100 b Slope

IVP(24x io) 0-100 0 7 0.36 0.08 1000-20000 11 -0.63 -0.01 total 18 -0.66 -0.01

IVP(32 x 48) 0- 100 9 0.44 1.58 100- 1000 13 -0.35 -0.11 1000-20000 25 -0.59 -0.01 total 47 -0.67 -0.01

IVP(48x 35) 0- 100 9 0.66 3.69 100- 1000 8 -0.19 -0.12 1000-20000 7 -0.87 -0.03 total 24 -0.39 -0.02

31 d/100b 200- IVP (24.10)

100

0 W 0 10 20 400 50 100 200 500 1000 2000 5000 10000 20000

IVP(32«48)

300 H

200'

100

0 •-/-! —i— 0 10 20 50 100 200 500- 500 1000 2000 5000 10000 20000

IVP(48«35) 400 H

300-

200

100-

20 io 100 200 500 1000 2000 —i—-, ,— 5000 10000 20000 Zl oTT P" 10° """" «« **"-«> « popu!.- io> .v::::i « ~ zr *. ' IVP<32 *« "- ™<>° -»> • °«* -

32 (not visible in IVP(24 x 10)wit h only few plants in this range). In the range of 100-1000 hy/100b correlation was very low and the level of d/100b decreased slowly. Beyond 1000 hy/100b the average number ofdihaploid sde ­ creased even slower. High values of d/100b occurred even when hy/100b was very high. The area in Fig.1 with less than 50 d/100b and less than 50 hy/100b was empty. Itwoul d have contained pollinatorswhic hproduce d less than one seed per berry. Berries with zero, one or very few seeds usually drop before maturity. Especially in IVP(48 x 35) the positive correlation was reasonably high inth erang e of0-10 0 hy/100b. The slope of the regression of d/100b on hy/100b varied from -0.01 to -0.02betwee npopulations .

3.3.3 Inheritance of dihaploid inducing ability

Apart from the pollinator populations in Fig. 1,othe r full-sib S. phu- reja populations weremad e andteste d ford.i.a .wit hcv .Ginek e (Table6) . In addition to these, also the progenies of dihaploid S. tuberosum x S. phureja crosses were tested (Table 7).Crosse s between good and bad S. phureja pollinators yielded plants with a wide range in d/100b and showed both positive and negative transgression (Table 6).Beside s giving good pollinators, selfing of IVP1 yielded also very bad ones, whereas the small self progeny of IVP48 did not give low values. In the population IVP(66 x 6),th eprogen y oftw opoo rpollinators ,on ehig hvalu e occurred. Theplant s from IVP48x G60 9wer erathe r uniform ind/100 b (Table 7). Of the two backcross populations, G609 x IVP482 and IVP482 x G609, the former had ahighe r average d/100b thanth e latter (222vs .12 4d/100b) , reflecting the difference between the two plants that were backcrossed: 157 vs. 86 d/100b in 1974.Th eparen tG60 9 induced only2 d/100 b in1973 . The cross between the best pollinators available, IVP48 and IVP35, was made to see whether pollinators with evenhighe rd.i.a .coul d beproduced . Positive transgression for d/100b was not only found from this cross, but also from other crosses, even where aba d pollinator as IVP32 was used as one ofth ep-.rent s (Fig.1 ,Tabl e 6). IVP(32x 48 )yielde dman y goodpolli ­ nators, IVP1 selfed, IVP(48 x 1) and IVP(1x 10)several .Eve n IVP(66 x 6) produced agoo dpollinator . In order to investigate whether this transgression could be reproduced over the years, test crosses were made in the following years (Table8) . Not all new plants produced tubers, so that some could not be repeated. Clones 1 and 5 from the population IVP(48 x 35) repeatedly showed trans­ gression; clones 2, 3 and 4 did not perform better than IVP48 inth e long run. The transgression of IVP48 selfed,1 was consistent; IVP(48x 10)1 and IVP(66 x 6)1 also showed repeatedlypositiv e transgression.

33 Table 6. Values ofd/lOO ban dhy/100 b inducedb yS. phureja genotypeso f F3 full-sibpopulations .Mean so f atleas t fiveberrie s from crosses made duringon e seasonwit hcv .Ginek ea s seedparent .Dat a ofparent s (PIan d P2)ar eals oincluded .

1973 1974

IVP48selfe d IVP1selfe d IVP(1x 10 ) IVP(48x 1 ) IVP(66x6 )

d/100bhy/100 bd/100 bhy/100 bd/100 bhy/100 bd/100 bhy/100 b d/100bhy/100 b

394 353 520 39 444 56 666 73 356 144 287 50 350 90 389 67 514 136 188 2413 273 80 340 128 223 8 475 244 40 940 228 21 310 433 208 133 307 57 40 447 211 1567 270 111 175 100 103 38 22 337 220 85 14 593 180 33 6 2623 160 59 0 60 140 60 120 1857 94 44 50 93

PI21 7 28 251 75 251 75 445 81 a a P2 18 290 209 8569 251 75 163 78

a.mea n of197 5an d 1976

3.3.4 inheritance of hybrid-producing ability

specTto IT"", Pr°genieS fr°m SeCti°n 3-3-3 Were als° -t«dl.d with re- theL plants ^"^ °f hy/1°°b- The »umber ofhybrid s producedb y Sh Wed a tralsores B" ^^ x^ IVP482) * ^^5an divp4 ^ 8seifed ° 5 —rkable positive ^crTiTr/ ,r^ tuberorbackcross' <*-*•• popuiation* «*v s* differe- d - dlr6Ction as the vs. 74 The Z' I u ^ *>ackcrossed Plants: 347 18000 hy/lOOb y/1°0b WaS greatSr than that °fd /100b-'value su pt o ua Wer reaChed ESP6Cially in thS hlgher f dat wewer«e verV eryVmuc :h spreadout ' . ^ ° ^ range, 0 a zzzv ?::: ^"*••*-*" * <*th e ieVeis of hy/i00b* BaMfo r imator, four plants of IVP(24 x 10), producing 32, ^ 2Q9 ^ ^ 34 Table 7. Values ofd/lOO ban dhy/100 b inducedb y S. phureja x diha- ploid S. tuberosum populations.Mean so fa tleas t fiveberrie s from crosseswit h cv.Ginek e assee dparent .Dat ao fparent s (PIan d P2) areals o included.

1974 1975

IVP48x G60 9 G609x IVP482 IVP482 xG60 9

d/100b hy/100b d/100b hy/100b d/100b hy/100b

160 67 550 914 217 67 157 347 282 86 171 59 140 235 270 53 150 75 130 27 242 79 128 56 130 24 216 3310 111 144 104 381 206 86 111 16 96 16 157 50 100 17 86 14 120 96 89 93 79 526 100 133 86 45 14 100 75 81 78 39

PI 445 81 157b 347b 322 74 P2 2a 25a 322 74 86b 14b a.measure d in197 3 b.measure d in 1974

hy/100b respectively, were used again as pollinators in 1974.I ntha tyea r they produced 92, 88, 280 and 20794 hy/100b respectively. The level was higher in 1974,bu t the order was aboutth e same.Th eplan twit h aninter ­ mediate level was still intermediate. Similarly IVP66produce d usuallybe ­ tween 100 and 200 hy/100b, more than IVP48 (<85 hy/100b) and less than IVP32 and 35 (>500 hy/100b).

3.4 DISCUSSION

3. 4.1 influence of the pollinator on dihaploid and hybrid frequencies

This study confirmed thatth epollinato r influenced the frequency ofdi - haploids formed in S. tuberosum significantly. Whereas mostauthor s worked with only a few pollinators (Jakubiec, 1964; Frandsen, 1967) or with a

35 Table 8. Positive transgression ford.i.a. .Value so fd/100 bo fsevera l clonesi ndifferen tyears ,compare dwit h thevalu eo fth eparent s recorded inth e sameyear .

Pollinator 1973 1974 1975 1976

Fl PI P2 Fl PI P2 Fl PI P2 Fl PI P2

IVP(48x 35) 1 47021 718 480 044 534 255 632 224 630 927 8 272 2 462 487 372 3 437 394 322 4 383 488 147 313 5 360 713 IVP48selfed, l 39421 721 7 41827 8 278

IVP(48x 10) 1 547 322 229 39027 8 152 IVP(66x 6)1 35 6 163 450 26 68

C ntaining nlya few rl \\Z7' ° ° 9oodpollinator s (Gabert, 1963;IriKu - ra 1975b), nowa group ofpollinator s with awid e range ofd.i.a .wa s

be"twelOVer SeVeral yearS- The °Vera11 leVel °f diha^°id Productionvarie d ea S ThlS in relitv e d ff "" ^ ' ^^ *" ^ ^^ * ^°* "* ^ V ln d i a b6tWeen 01 Ib - " - P "»*»» -d theirranking . wasno tf o , "* ^ P°llinat°-< ascribed byGaber t (1963), wasno t found m thismaterial" ^ .

fluc^LnsT,ti0n WaS alS° inflU6nCed ^ the Pollinator.Th e high yearly n0t UPSet the rSnking f the onls: ith ve ° Pollinator-, exceptfo rth l proLly dl ! W hy/10°b- The g"at Variati°n b— years and dayswa s Probablydu et oenvironmenta l influences during pollen formation (Chapter

3.4. 2 Relationship between dihaploid andhybrid frequencies frihiin:T:i:T rationship between the dihapi°id «* ^^ p«*«* P llinat r were negative , ° ° ™ »^ Correlationsbetwee n the twl (Table 5) W th V6ry hi9h in three ^^^ ^^ions asa whol e ard t0 the POSitlVe C ati hy/1 b ;mtr ™ °* inth erang eo f0-10 0 m ht h f tn! l™^« that low pollen guality could have beenth e th c n tion of iow d/ioob and ti n po :en :t ; : ^^ ^ «.«ei«*« be- low L Z Stainablllty and elther d/100b - ny/lOOb provedt ob ever y elmedtoTary rr; 7 10°"1000 hY/1°°b' ^^ «" ^^ »»*- with more thanTooTt/To^ " "" ^ ^^ C™<- inplant s !000 hy/l00b were accentuatedb ysom e very high valueso f 36 hy/lOOb. Factors for high d/lOOb and high hy/lOOb seem to have been dis­ tributed independently in the populations, but in their expression they mighthav e influenced each other.Hig hnumber s such as100-18 0hybri d seeds perberr y will limit thenumbe ro fovule s available forproductio n ofdiha - ploids. The negative slope of the regression ofd/100 b onhy/100 b canthu s be interpreted as areductio n indihaploid s of anaverag epollinator , caused by extra hybrids in aberry . This reduction was 1 to 2 dihaploidspe r 100 hybrids. Itwa s higher in IVP(48 x 35)wit h a higher mean ford/100 b than in IVP(32 x 48) and IVP(24 x 10).Thi s isno ta majo r influence on average dihaploid production. In the 10% of the pollinators producing more than 10000 hy/100b, the reduction in d/100b willhav ebee n100-200 .A high seed set pollinator as e.g. IVP32 will probably be underestimated for its ge­ netic potential for inducing dihaploids. Apart fromreducin g thenumbe r of dihaploids produced, the production of very high numbers of hybrids is an unfavourable character in a pollinator, because it involves more work in seed extraction and screening. However, it is possible for an individual pollinator to produce high numbers of both dihaploids and hybrids per berry, asca nb e seen in IVP10. Hermsen & Verdenius (1973) studied the correlation of total numbers of seeds and dihaploids per berry. They did not calculate between-pollinator correlations, but the within-pollinator correlations between berries. The positive correlations in their study were the result of using the total number of seeds instead of number of hybrids, especially for low seed set pollinators. In low seed setpollinators , like IVP48, dihaploids form the largest fraction of the seeds and a correlation between the total and its largest fractionmus tb epositiv e andhigh .

3.4.3 Inheritance of dihaploid inducing ability

Inheritance of d/100b was not controlled by the cytoplasm, as the progeny of G609 x IVP48 was notlowe rtha n itsreciproca l cross,whic h had a superiorpollinato r assourc eo fcytoplasm . The range in d/100b was almost continuous in the populations studied. This meant that many genes were involved and/or that environment had a great influence on theexpressio n ofth egenes .Amongs tth e IVPpollinator s a constant ranking was found and seven levels of d/100b could be distin­ guished (Section 3.3.1), butsom ecoul d havebee n causedb y highnumber s of hy/l00b. Several genes must have been involved in producing these differ­ ences, on the basis of the dataavailabl e itwa s attempted toestimat e the type of allelic interaction and the number ofgene s involved ind.i.a. . It was assumed that all genes for d.i.a. had the same influence and thatth e allelic interaction was of the same type. Alleles increasing d.i.a. of a pollinator are calledpositiv e alleles (+alleles).

37 3.4.3.1 Mode ofinheritanc e

In the dominant type of allelic interaction, positive alleles are not hidden in a heterozygous combination. Therefore, d.i.a. levels higher than that of the higher parent cannot be expected after selfing. As there was positive transgression in the self progenies of IVP1 and48 ,a swel l asi n that of S. phureja 253 (Irikura, 1975b), dominancewa sno t a likely wayo f inheritance. Recessivity has been suggested by Gabert (1963) and Irikura (1975b). Data fromthi s studydi dno tsuppor tthei rconclusion . Ina recessiv e model d.i.a. levels lower than that of a parent are not expected after selfing. Plants were found in selfprogenie s with amuc h lower level of d/100b than the parent (Table 6).Inbreedin g depression andba dpolle nmigh thav e been a cause for the lower d.i.a., but some of the plants with low d/100b pro­ duced many hybrids, e.g. IVPI selfed, plant 10.Dat a from Irikura (1975b) also showed negative transgression for d/100b, often combined with a high male fertility ofth epollinator . If complete dominance and recessivity are both excluded, intermediate inheritance remains.Ther e are arguments in favouro f intermediate inheri­ tance. Gineke dihaploid G609 had a very lowd.i.a . (2d/100b) ,whic h level couldb eregarde d asspontaneou s orbasic .G60 9wa s unrelated to S. phureja IVP48, a superior pollinator. It is unlikely thatbot h clones had genes in commonwhic hcontribute d topolle n irregularities (Chapter 7)an dwoul dre ­ duce generativevitality .Th ereciproca l Flpopulation s fromG60 9 and IVP48 had agoo d level ofd.i.a .an dwer e remarkably uniform (Table 7).A n excep­ tion was one plant with 14 d/100b, which had low pollen stainability and produced only 100 hy/l00b. This uniformity indicated ahig h level ofhomo ­ zygosity for .alleles in both parents. This is not unlikely in IVP48 be­ causeo finbreedin g and selection forhig hd.i.a. .Fo rG60 9 this would mean a complete lacko f+alleles . With intermediate inheritance all alleles would be expressed. If none were present in G609 and many in IVP48, the number of .alleles in the Fl thus b been half °f that ln IVP48- The leVel of

38 3.4.3.2 Number ofloc i

Assuming a square root scale a hypothesis canb e formulated onth enum ­ ber of loci for d.i.a. in IVP48an d relatedpollinators .Th euniformit y of the Fl of IVP48 and G609 pointed to high homozygosity of the parents for d.i.a..Th e small differences amongstF lplant swer e caused by heterozygous loci in IVP48 and/or environmental influences. Ifenvironmenta l influences areno tconsidered , thedifferenc ebetwee nth e twogroup s inth eF lplant s- 4 plants:79-10 4d/100 b and 5plants :130-16 0d/100 b -wil l reflectth e num­ ber of heterozygous loci in IVP48. In that case the lower group will have the same number of +alleles as IVP48 has homozygous +loci.Wit h only one heterozygous locus in IVP48, the effect ofon e+allel ewoul dbrin g thedi - haploid frequency from 90 d/100b to 145d/100b .O n asquar e rootscal e the distances 0-90 d/100b and 145-444 d/100b are each 3.8 times the distance 90-145 d/100b. With one heterozygous locus in IVP48 the level of IVP48 (445 d/100b) would have been caused by 9 +alleles on 5 loci. On a square rootscal e thevalue s for1-1 0 +alleleswoul db e anexpansio no fth e series 5.5x2: 6, 22, 50, 88, 137, 200, 270, 350, 445 and 550 d/100b (level of 1974). If IVP48 were heterozygous for 2 or 3 dihaploid inducing loci, the range in the Fl would be caused by 2 or 3 +alleles and thenumbe r of+al - leles in IVP48 would be 12 and 17 respectively on 7 and 10 loci. Another indication for the number of loci involvedwa s found inth edifference s in d.i.a. between IVP pollinators and their constant ranking. Seven levels were distinguished, but the lower pollinators were probably underrepre- sented, as better pollinators got more emphasis in studies. A number of fiveo r seven lociwoul d agreebette rwit hth e sevenlevel s thante n loci.

The transgression found in IVP48 selfed agreedwel lwit h atota l number of five loci, fourbein ghomozygou s for+allele s andon eheterozygous .

3.4.3.3 Test for intermediate inheritance

The hypothesis of intermediate inheritance can be tested in the three pollinatorpopulation s from 1973 (Fig.1) ,usin g theestimat e ofth enumbe r of alleles involved in determining the d.i.a., intermediate inheritance meaning that the progeny mean must be close to the mid parent value. The general level of d/100b was lower in197 3 than in 1974.I f IVP48 is assumed to have 9 +alleles, an adapted series for1-1 0 +alleles for 1973woul dbe : 3,13 ,29 ,52 ,81 ,117 ,160 ,208 ,26 3 and32 5d/100b .Wit h thesevalue s the number of +alleles was determined for everyplan to fth e threepopulation s and the parents. The observed numbers of d/100b were increased with 1% of hy/100b in IVP(32 x 48) and IVP(24x 10)an d2 %o fhy/100 b in IVP(48x 35) because of the negative influence of hy/100b on d/100b (Section 3.4.2). This correction was not needed for IVP48x G609, ashy/100 b was low,Popu ­ lation IVP(48 x 35)ha d a mean value of 8.8 +alleles and a range of 4-12

39 +alleles. The mid parent value was 8.5 +alleles. For IVP(24 x 10) the progeny mean was 5.9 +alleles, the midparen tvalu ewa s 5.5, assuming that IVP10 carried 10 +alleles, which is reasonable with a number of 6500 hy/100b and acorrectio n factoro f2% .Progenie s of IVP10wer e usually high in d/100b. The progeny mean of IVP(32 x 48)wa s 6.9 +alleles, whereas the mid parent value was 6.5, if 4+allele swoul db epresen t in IVP32 (in197 3 4090 hy/100b). The three population means tended tob e higher than themi d parent values, but not much. These results point to intermediate inheri­ tance.Thi s conclusiondoe sno tdepen d onth e assumption aboutth e numbero f loci involved. With seven or ten loci for d.i.a. the progeny mean would likewiseb eclos et oth emi dparen tvalue .

3.4.3.4 Transgression

With the big increase in levels of d.i.a. as a result of the selection work of Hermsen & Verdenius (1973), the limits of dihaploid induction had notye tbee nreached . Further increasesprove d tob epossibl e with the same material (Table 8).Repeatabl epositiv e transgression was found in several full-sib F3 progenies, to levels higher than the best pollinators. Within theavailabl emateria l theuppe rlimi to fd.i.a .ma y havebee n reached,bu t with othersource s ofgene s fordihaploi d induction further improvementma y bepossible .

3.4.4 Inheritance of hybrid producing ability

Frequencies of hy/100b were very variable between seasons.Thre e levels were found: 0-100 hy/100b, 100-300 hy/100b and >300 hy/100b. It is quite possible that more levels were present in the material, but with the high variation they could not be distinguished. Monogenic dominance of lownum ­ bers of hybrids, as suggested by Hermsen & Verdenius (1973), would result intw o levels only,wit h thethre e levels found,eithe r inheritance mustb e intermediate,o rmor e genesmus thav ebee n involved.

40 4 Seed parent influence on dihaploid and hybrid production

4.1 INTRODUCTION

Influence ofth e S. tuberosum seedparen to nth e frequency ofdihaploid s itproduce s after pollination with S. phureja hasbee nreporte d by several authors (Section 2.3.2). Such influence may workthroug hgene s forpseudo - gamy, cytoplasmic factors, lethal genes or thenumbe r ofovule spe rberry . The last factor canals o influencehybri d frequencies. The dihaploid producing ability (d.p.a.)o f asee dparen tca neithe r be controlled by the genotype ofth e sporophyte orb y thato fth egametophyte . Little is known about the mechanism inducing an unfertilized egg cell to divide. Therefore this question cannot be solved cytologically; itmus tb e solved genetically. To do this Bender's (1963)suggestio n was followed to check whether a haploidization-diploidization cycle would accumulate genes forpseudogamy . Cytoplasm played an important role in the level of d.p.a. according to Frandsen (1967).Montelongo-Escobed o (1968)conclude d aboutth e chromosomal inheritance that high d.p.a. was dominant and based on few genes.Va n der Knaap (unpubl.)foun d Flvalue s ford.p.a .betwee n the lowerparen t and the midparen tvalu e in astud y ofth e inheritance ofd.p.a . In this study it was assumed that the seed parent had an influence on the production of dihaploids. Data were obtained toconfir m thiswit hmor e pollinators and over several years.A tth e sametim e seedparent s fromdif ­ ferent origins were pollinated during several years to determine the ge­ netics ofd.p.a. , the influence ofcytoplasmi c inheritance and thestag e of functioning: gametophytic orsporophytic .

4.2 MATERIAL AND METHODS

Seed parents from different origins were tested for their d.p.a.. All Plants were grafted onto tomato rootstock, to enhance flowering and fruit set. Forcrosse swit h severalpollinator s the following S. tuberosum culti­ va« were used: Gineke, Radosa, Ultimus, Sirtema,Record , Libertas,Multa , Merrimack and Katahdin. In addition a S. demissum (2n=6x)genotyp ewa sin ­ cluded, a seedling from accession WAC 3095. A clone of allotetraploid S. polutrichon was testedwit h severalpollinator s aswell . To test for cytoplasmic influence, reciprocal crossesha dbee nmad ebe ­ tween Gineke and Libertas, which differed in d.p.a. and had no parents in

41 common in the last four generations. From each ofth etw o resulting Flpo ­ pulations fifteen unselected seedlings were grafted and used as seedpar ­ ents.Fo rth e samepurpos e A18, aNeotuberosu mclon e selected atWageninge n from material derived by Dr.N.w . Simmonds from pure S. andigena. was crossed with S. tuberosum cv. Sirtema. Both parents and three Fl plants were studied assee dparents . In order to distinguish between gametophytic and sporophyticdetermina ­ tiono fd.p.a. , severalGinek ederivative s wereused , asdescribe d inTabl e 9. Two different colchicine doubling methods were applied to seeds and to Plants,seed swer eplace d on filterpape r inpetridishes ,soake d with 0.25% solutxon ofcolchicin e and leftt ogerminate .Germinatin g seedswer e placed in seed boxes, vigorous plants of tetraploid appearance and having a high chlorop astnumbe rwer epotte d and latergrafted .Th e tetraploidy was later In11 1 rntin9th SnUmbe r° fChr °mOSOmeSl nth eroo ttiP - -d finally bv r CellS-Th eCOlchici -<^ledplant so fG60 9wer eobtaine d -re remolTT (R°SS " ^" "^ LM*too' 1974>" ™* *»^a faxil s Lu chTo , +cotton WOQ1 soaked with °-25% colchicine »^« in the Zl\ h-r HWSr e C°Unted °nSf W Sh0°tStha t l0°ked *»**«*• Short. fowerin gtim eTh " l ^^^ «« *»*»* onto tomato rootstock. At Is polnlat th%:hr°mOSOme nUmber »« determined inpolle nmothe rcells .

35 4 8 origiorign onfo th fth e epollinator " '' ^s an^d the^mpi-hr,ri ^ =„*^ • "' « * «" ^d. The haploidsa ™ H •„ , • methodso fcrossin g and screening fordi - napioids aredescribe d in Section3.2 .

Table 9. Derivatives ofcv .Ginek euse d assee dparent .

Name Number of Origin genotypes

Ginekepseudogami c 3 tetraploids pseudogamic tetraploids fromGineke , preselected onlea f shape;th etetra ­ ploid characterwa s determined from the chloroplast number (Gx G254)4 x 2 tetraploid progeny from Gineke and its highly fertile dihaploid G254 Gineke colchicine- 2 plants from dihaploid seeds from Gineke doubled dihaploids treatedwit h colchicine (seedtreat ­ ment) G609colchicine - 3 doubled colchicine-doubled plants from Gineke Gineke selfe d A. dihaploid G609 (plant treatment) plants from seeds obtained after self- ingo fGinek e

42 Numbers of ovules per berry were determined according to the method of Carroll & Low (1975).

4.3 RESULTS

4. 3. 1 Influence of the seed parent on dihaploid and hybrid frequencies

Six S. tuberosum cultivars and a S. demissum genotype were pollinated with four S. phureja pollinators in1974 .Value s ford/100 b and hy/100b are given in Table 10, the analyses of variance in Table 11.Th e seed parent effecto nd/100 b was highly significant,a swa s thepollinato r effect.Con ­ sidering onepollinato r ata time ,rankin g ofth e seedparent swa s similar. Six cultivars were pollinated with two good pollinators in 1975 and 1976. Data of this two year comparison aregive ni nTabl e 12an d the analyses of variance in Table 13.Agai n the seedparen teffec twa shighl y significant. The ranks of the seed parents were similar for both pollinators. In both experiments Gineke and Radosa ranked highest and Record low. Libertas was very low. More data on the seed parent effect can be found in Chapters 3 and 5 (Tables 3, 17an d 19). All,includin g the three and fouryea rcompar ­ isons ofTabl e 17,indicate d asignifican t seedparen teffect . From more than 250 flowers of S. polytrichon, pollinated with several pollinators, a fewberrie swer eobtained ;bu tn odihaploid s were found. The seedparen t effecto nhy/100 bwa sno tsignifican t inth etw oexperi -

Table 10. Numbers ofd/100 b andhy/100 bproduce db y S. tuberosum cultivars and a S. demissum (2n=6x)genotyp e afterbein g crossedwit h S. phureja pol­ linators. Crosseswer emad e in 1974.Averag enumbe r ofberrie swa s38 .

Seedparen t d/100b hy/100b poll inator (IVPnr . ) pollinator (IVP nr.)

1 10 35 48 mean 1 10 35 48 mean

Gineke 251 209 342 445 312 75 8569 2065 81 2698 Radosa 361 154 250 700 366 128 8802 2195 86 2803 Ultimus 253 291 193 283 255 87 1082 924 50 536 Sirtema 154 117 116 315 176 178 12900 2163 138 3845 Record 131 59 155 248 148 77 4334 624 33 1267 Libertas 114 81 41 113 87 100 3038 3076 75 1572 S. demissuma 125 100 109 119 113 0 1811 293 34 535 mean 198 144 172 318 208 92 5791 1620 71 1894 a- S. demissum, a hexaploid ,produce d trihaploids

43 Table 11. Analysis ofvarianc e ofth e data fromTabl e10 .

Source Sumo fsquare s d.f. Probability

d/100b seedparent s 264801 6 0.00 pollinators 121900 3 0.01 error •135993 18 total 522694 27

hy/100b seedparent s 37866392 6 0.28 pollinators 152814121 3 0.00 error 83378472 18 total 274058985 27

d.f..degree so f freedom

Table 12. Numbers ofd/100 b andhy/100 bproduce d by S. tuberosum cul- tivars afterpollinatio nwit htw o S. phureja pollinators intw oyears . Averagenumbe r ofberrie swa s 21.

Seedparen t d/100b hy/100b pollinator (IVPnr. ) pollinator (IVP nr.)

35 48 35 48 mean

Gineke 1975 297 406 352 150 39 95 1976 237 207 222 1323 25 674

Radosa 1975 301 448 375 254 25 140 1976 134 160 147 2714 27 1371

Multa 1975 279 349 314 89 33 61 1976 239 180 210 252 14 133

Record 1975 188 126 157 100 0 50 1976 89 93 91 3880 27 1954

Merrimack 1975 0 130 65 0 50 25 1976 148 248 198 455 67 261

Katahdin 1975 31 21 26 67 14 41 1976 49 91 70 4285 38 2162 mean 166 205 185 1131 30 580 Table 13. Analysis ofvarianc e ofth edat a fromTabl e12 .

Source Sumo fsquare s d.f. Probability d/100b seed parents 189331 5 0.01 pollinators 9087 1 0.10 years 20475 1 0.03 s-xp interaction 14308 5 0.38 s xy interaction 82953 5 0.02 p x y interaction 3775 1 0.24 error 10866 5 total 330794 23 hy/100b seedparent s 3770058 5 0.48 pollinators 7271004 1 0.03 years 6289408 1 0.03 s xp interaction 3930302 5 0.47 s xy interaction 3744606 5 0.49 p xy interaction 6213872 1 0.03 error 3623714 5 total 34842965 23 d.f.. degrees of freedom

ments. However, it was almost significant (p=0.08)i nth e 1974experiment , after logarithmic transformation to remove multiplicative effects (Chapter 5). variation was very high ,fo. , r ^ hy/100b \.,./-\nn .hTh Thecoeixici.c »coefficient n so fvariatio n for ,, „ /i->o/.„ j •?*<>/• from thesam e experiments the d/100b in Tables 10 and 12 were 42% and zw*, «» ""» *- values forhy/100 b were 147%an d114% . Correlation between numbers of ovules per berry of nine cultivars and d/lOObwa s low;correlatio nwit hhy/lOO bwa shig honl ywit h twopollinator s in197 4 (r=0.90).Number s ofovule spe rberr yrange s from600-950 .

4.3.2 Cytoplasmic inheritance

The possibility of cytoplasmic inheritance of d.p.a.wa s studied within 5. tuberosum and in a 5. tuberosum x S. anäigena cross.F lplant s fromre ­ ciprocal crosses between the good seed parentGinek e and the -~^dJ~d seedparen t Libertas andth eparent s themselveswer epollinate dwit hIVP35 . „•,. i^ o AT-P aiveni nTabl e14 .Notwithstandin g The resulting numbers of dihaploids are given several Flplant sdi dno t flower.Thi s the grafting onto tomato rootstock+^r>r~v, several n y PI niants from thereciproca l crossesbein g resulted in unequal numbers of Flplant ~r s nu. ^„„„on^eo fth e Flplant swa s low.Tw o used as seedparents .Th eoveral lperformanc e ofth e tx p

45 Table 14. Numberso fd/lOO bproduce db y S. tuberosum cul- tivarsGinek ean dLiberta san dthei rtw oreciproca lF l progeniesafte rpollinatio nwit h S, phureja cloneIVP3 5 in1974 .Als ogive nar eth emean spe rF lan dth emi dpar ­ entvalue .F lplant syielde d1 7berrie so nth eaverage .

d/100b Mean / Libertasx Ginek e 17 40 64 75 49 Ginekex Liberta s 6 25 36 42 54 67 64 67 83 86 108 128 Gineke 342 192 Libertas 41

outo ffiftee nplant sproduce d lessdihaploid s thanth elowe rparen t Li- r N others reached the mid with :;n r r r — — *»»£^ 1tio nw t hL Yb aSm had S SlightlY higher mean (64 d/1°0b> than ** P°P- Tt Hrwii o" T°PlaSm (49 d/10°b)' but the diff— - "o*<**- nincant (Wilcoxon rank test, p=o 14) s cytopiasm the .eo^osTcti;" 8 -a:rr; ' ' — -* — — dp a wer!ni T '' 9°°d* tüZ>e™s™ cultlvarwit hhig h level ofT , ; Wlth IVP48- ReSUltS are given in Table ". Themea n d owert h ,\ P- -tion of the progeny (186 d/100b) was slightly lowertha nth emi dparen tvalu e(20 1d/l00b) .

4.3.3 Sporophytic vs. gametophytic determination lowerfffrent ^ ^ Pr°gSnieS ^ Gineke a11 Pr°du~d dihaploids in T^bleITT " thSn GinSke itS6lf WhSn POllinatSd With ™5 or4 8 t^Llo'd I" WaS Variati°n Within and betWSen ^PB- speciallyth e threeI c!i T ^ ***** * ^ ^^ ™* ^ dihaploids The differene s 6 ° ed ^ ^^ ^^ **^ticll y identical.Thei r re POSSlbly dUe t0 after eff ItL Tow ;r d - — °*colchicine .The yforme d ^Lr, rLTdo^id::; it ; ™h T~* dihapioids (335: of rinflV^ ^ .. cens, were the highest producing group K^/o o± Gineke), but not much higher than thP =ol* • Progamic tetraploids (25°/) TI- ^^ <31%> °r ^ tet loidS originated from 2n egg cens' or ™ ^ ""**** ™* ^ S nt during pseudogamy. In the latter * ^ *" P° —ly doubled with doubled dihaploids "" ^ W°Uld b" geneti-lly comparable

46 Table 15. Numbers ofd/lOO bproduce db y a S. an- digena genotype (A18),S . tuberosum cv.Sirtem a and theirprogen y afterpollinatio nwit h S. phure- ja clone IVP48 in 1974.Als o givenar emean so f the Fl and themi dparen tvalue .F lplant syielde d 10berrie s onth eaverage .

d/100b Mean

A18x sirtema , pi. 1 190 pi. 2 267 186 pi. 3 100 A18 86 201 Sirtema 315

4.4 DISCUSSION

4. 4.1 Influence of the seed parent on dihaploid and hybrid frequencies

The effect of the seed parent on the production of dihaploids was con­ firmed. The ranking of seedparent swit h aserie so fpollinator s was rather constant over several years.Th eeffec to nhybri dproductio nwa sno tsigni ­ ficant, probably due to the high variationo fthi scharacte r andmultipli ­ cative effects (see Chapter 5).A big seed parent effect could not beex ­ pected, asmos thybrid swer etetraploids ,whos e frequencywa s determined by 2n pollen frequencies in the pollinators (Chapter 7).Femal e fertility of the seedparen t could have influenced hy/100b,bu tth ecorrelatio n ofovul e numberswit h hy/100bwa shig h ina fe wcase sonly . Trihaploids from S. demissum havebee nreporte d earlier (Bains& Howard , 1950; Irikura, 1975a,b). The overall frequency ofon etrihaploi d perberr y confirmed that the cytologically unbalanced nature ofa triploi d isappar ­ ently no obstacle to vegetative growth (Marks, 1966a). Allotetraploid 5. polutrichon has proved to be able to produce dihaploids after a cross with s. stoliniferum (Marks, 1955) or via anther culture (Irikura, 1975a,b). A reason for failure in this study might be, that S. phureja is not the right pollinator. S. polutrichon might have a number of endosperm balance factors (Section 2.5.2.1) different from S. tuberosum. The cross S. polutrichon x s. phureja can yield many triploids (Ramanna & Abdalla, 1970).

47 Table 16. Numbers ofd/100 b andhy/100 b from cv.Ginek e and tetraploid progeny ofGineke .Result s frompollination s with IVP35 and IVP48.Diha - ploidproductio n isals o expressed aspercentag e ofth eproductio n ofGi ­ nekei nth esam eyea rwit hth esam epollinator .Mea nnumbe r ofberrie s fromprogen yplant swa s14 .

Pollinator Seedparen t Dihaploid production hy/100b

d/100b as% of mean % Gineke

IVP35 1974 Gineke 342 100 2065 Gineke pt pi. 1 30 9 140 pi. 2 70 20 23 260 Pi- 3 136 40 1791 (Gx G254)4x Pi. 1 20 6 3 104 Pi. 2 0 0 109

IVP48 Gineke 445 100 81 Ginekep t Pi. 3 130 29 140

IVP48 1975 Gineke 406 100 39 Gineke selfed Pi. 1 100 25 80 pi. 2 0 0 50 31 Pi- 3 118 29 92 Pi. 4 278 68 75 (Gx G254)4 x pi. 1 32 8 16

IVP48 1976 Gineke 207 100 25 Gineke cdd pi. 1 67 32 33 133 Pi. 2 68 33 140 G609cd d Pi. 1 10 5 90 pi. 2 44 21 12 122 Pi. 3 22 11 67

G. Gineke pt. pseudogamic tetraploid cdd.colchicine-double d dihaploid

48 4.4.2 Cytoplasmic inheritance

Concerning cytoplasmic differenceswithi n S. tuberosum, theresult s from Gineke x Libertas progenies did not indicate cytoplasmic inheritance of d.p.a. (Table 14).Th ebi gdifferenc e ind.p.a .betwee nth etw oparent s was not reflected in the two reciprocal progenies. The results from the few plants from the cross between A18 andSirtem a- al lwit h S. andigena cyto­ plasm - did not support the theory of Frandsen (1967) about interspecific cytoplasm differences (Table 15).Eve n his own data did not support this cytoplasmic influence very strongly. Apparently the cytoplasm of the seed parentwa s nota majo r factor indeterminin g d.p.a..

4. 4. 3 Sporophytic vs. gametophytic determination

When studying the genetics of dihaploid production, it is important to knowwhethe r iti sth e sporophytic orth egametophyti c genotype thatinflu ­ ences pseudogamy. This question was investigated on the basis of the hy­ pothesis of Bender (1963)wh o stated thatdouble d dihaploids should beef ­ fective dihaploid producers, if the 'apomictic tendency1 was determined by thegametophyti c genotype.Th etetraploi dprogen y ofsee dparen tGinek epro ­ duced via dihaploidization should have accumulated alleles for pseudogamy and should accordingly produce more dihaploids than Gineke itself or its progeny obtained through selfing, unless Gineke was homozygous for all genes involved. The results did not support this (Table 16).Th e doubled dihaploids were similar to the self progeny of Gineke, their d.p.a. being 33% and 31%respectivel y of the d.p.a. of Gineke. The doubled G609 plants produced even fewer dihaploids. The pseudogamic tetraploids also did not seemt ob e theresul to fa successfu l selection forallele s ford.p.a. , nor did the tetraploid Fl plants of Gineke with two of its dihaploids produce dihaploids in high numbers. The bad performance of double G609 plants and doubled dihaploids of Gineke could not be explained by reduction of the fertility due to inbreeding, as the production of hybrid seeds per berry was evenhighe r thantha to fGinek eitself . The populations used were very small, but on the average they should have had a d.p.a. at least equal toGineke ,i fth egametophyti c phasewer e determinative. It can be concluded that thegeneti c determination ofpseu ­ dogamy was not onth edihaploid , gametophyticlevel . Dihaploidization works as a sieve for lethal genes. One might assume that d.p.a. was not genetically determined, but that the differences be­ tweenth e seed parents ind.p.a .wer eonl ycause db y adifferen t occurrence of lethal genes. In that case aswell ,th eprogen y ofdihaploid s ofGinek e should have performed atleas ta swel l asGinek e itself.Sinc ethi swa s not the case, apparently lethal genes were not determinative either. The pro-

49 ductiono fdihaploid s ismos tprobabl y determined inth e sporophyticphase , modified by the lethal genes,whic hwor k inth egametophyti cphase .

4.4.4 Node of inheritance of dihaploid producing ability

Assuming thatcytoplasmi c effects ond.p.a .wer emino r andtha tth e spo­ rophytic genotype determined pseudogamy, the results obtained can be ana­ lysed asbein g governedb y tetraploid chromosomal inheritance. Iti sdiffi ­ cult to reach definite conclusions as the sizes of the populations used were small and tetraploid inheritance is complicated initself .Chance s of recovering agoo d genotype inth eprogen y arerelativel y high,whe n achar ­ acter isdominant ,bu t aremuc h lower fora recessiv echaracter , especially in tetraploids. With intermediate inheritance amea n progeny value is ex­ pected closet oth emi dparen tvalue . The two reciprocal progenies of Libertas and Gineke can be regarded as oneF lpopulatio n (Table 14).Th e lowleve l ofdihaploi d production inthi s progenydi dno tpoin tt odominanc ewit h oneo r afe wloci , likeMontelongo - Escobedo (1968) concluded. Ifman y loci were involved, dominance would be possible. Fouro rmor e additive locipresen t inGinek e as simplex or duplex could also resulti na dispersa l ofth epositiv e genotype ofGinek e to such an extent thatn o plant in this small progeny would receive all theposi ­ tive alleles. The S. andigena x s. tuberosum progeny was close to the mid parent value, pointing to intermediate inheritance (Table 15).Al l differ­ ent Gineke derivatives had a lower d.p.a. than Gineke itself. Amongst the data of Tables 14, 15 and 16 no case ofpositiv e transgression occurred. Only two plants came close to the better parent: Gineke selfed, plant 4 (68%) and A18 x sirtema, plant 2 (85%). Recessivitycoul d fitwit h some of the data,bu t ifpseudogam y were recessive, the differences in d.p.a. be­ tweenth e S. tuberosum cultivarswoul d haveha d tob e causedb y asman y ad­ ditive loci. Pseudogamy is common in Solanum and most S. tuberosum clones produce dihaploids. It is not very likely that so many (tetraploid)geno ­ types are homozygous recessive for several loci for a character which has no obvious advantage toth eplant .Fro m thatpoin to fvie w intermediate in­ heritance is more likely. This was also supported by findings of Van der Knaap (unpubl.), who found progeny means close to the mid parent value in two reciprocalprogenie s ofth ecultivar s Radosa andLibertas . Iti s likely thatth einheritanc e ofdihaploi d producing abilitywa sin ­ termediate,bu tdominanc e canno tb e excluded,provide d thatman y loci were involved.

50 5 Interaction between pollinator andsee d parent influence

5.1 INTRODUCTION

Onlya fe wgenotype s wereuse d astester sfo rpollinato ran dsee dparen t influence in Chapters 3 and 4. When using only few testers one assumes that, apart from multiplicative effects, there is no interaction between theeffect s ofpollinato r andsee dparen to ndihaploi dproduction .Thi sim ­ plies that onemai n effect canb estudie d irrespectiveo fth eleve lo fth e other.On eca ndistinguis h amultiplicativ e anda non-multiplicativ e effect in interaction. The multiplicative effect will not change therankin g of thepollinator s or seed parents andthi s effect canb e eliminated from an analysis ofvarianc e (anova)b y a logarithmic transformation.Th einterac ­ tionwhic h then remains isth einteractio ni na stric tsense . The approacho fno tconsiderin g interactionwa sbase do npreliminar yex ­ periments (VanBreukelen , 1972).A ttha ttim ei twa sfoun dtha tth erankin g ofpollinator swa ssimila rwit heac ho fth etw ocultivar s theywer ecrosse d to. m thecours e of this research (late1974 )th enee dwa sfel tt overif y theassumptio no fn ointeraction ,especiall y sincea hig hvariatio nwa sob ­ served in dihaploid andhybri d frequencies ofIVP3 5wit hsevera l cultivars during a season. The 'no interaction' hypothesis could attha t momentb e tested with available data using years asreplications .However , analysis ofth eresult s ofth etw oyear swithou tpoolin gth eresult s from different crossing dates, raised the suspicion that 'crossing date' was a major source of variation in these experiments. When temperature influence was confirmed ina growt hchambe r experiment (Chapter 6), apla nwa sdevise d in which complete sets ofcrosse s were tob ecarrie dou ti na singl eda ywit h aninterva lo fthre et ofou rdays .Thi swa yth einfluenc eo fth edat ewoul d be the same for all combinations. Crossing date combines twoinfluences : the ageo fth eplant s andth eenvironmen t ontha tday ,a stemperatur e and humidity. The experiment of197 5wa splanne dwit ha limite d numbero fcul - _• •• _<•mon v<-> fth ecrosse sowin gt onig h tivars andpollinators . After failure ofman yo ftn ecr o y 4- ,. „n„ated in197 6o na smaller scale. Both temperatures the experiment was repeated in •.H« t-n determine whether among the best experiments were planned especially to determine Pollinators some would induce more dihaploids with certain cultivars than others, or whether thedihaploi d production ofa specific P°^«*"""d Parent combination could bepredicte d from theperformanc e ofth e ndivid- i.i,^-i/ inmductio nwa sinclude da shybri d «alparents .Collectio no fdat ao nhybri d production Productionmigh t influence dihaploid frequencies. Othertwo-facto r interactionswer e studieda swell . ^ 5.2 MATERIALAN DMETHOD S

Theprocedur efo rcrossin gan dgrowin gplant sha sbee ndescribe di nSec ­ tion3.2 .Fo rth einteractio nexperimen tth ecultivar sGineke ,Record ,Mul - ta, Radosaan dKatahdi nwer euse d assee dparent s and IVP6,10 ,3 5an d4 8 aspollinator s (Section 3.2). Thesee dparent swer earrange di nth egreen ­ house ina wa ytha tenvironmenta l differencesbetwee nthe mwer eminimized . Crossmg took place according to the following schedule: every threet o fourday ste npollination swer emad e forever ycombinatio no fth eselecte d pollinators and seed parents. Seedswer e screened forembry o spota sde - scribedi nSectio n3.2 . A sample wasplante d of thespotles sseed san do fth eseed swithou ta

v1S1ble embryoo feac hcultivar .Th enumbe ro femergin gdihaploid swa sde - ter^ned for each group,wit h the emergence ratioth enumbe ro femergin g

diploids per10 0 berries (d/100b)wa scalculate dfo rever ycrossin gdat e Theirnur" " ^^^^ The ^°tted ^brid) seeds were notplanted . Theirnumber swer econverte dt ohybrid spe r!0 0berrie s (hy/l00b). ofdihil o Tn7Y .eXPeriment(Va n B"ukelen, X972)ha dshow ntha tnumber s LibuUon /h /ln9le berriSS had 3 distributi- closet oa Poisso ndis - br"rthe' dt ?" a SqUare r°0t tranSf°r-^- c««*!hav e been usedt o trnsforma t *° * ^^ diSt^^- Zwever,a logarithmi c(In , from 2int"1 T Preferred' ^ " eliminat6S the multiplicativeeffec t 0 tranSf rmati hy^OOb Wlth tilf " T ^ ° - »« carriedou to nd/1 00b and y=1 ln(X+1) with ÎIrmedvalu e Th T °° < x=originalvalu e andy=trans - formedvalue .Thi sformul awa spreferre d toy= 1001nx,a sth evalu ex= 0oc -

S W e r e arried Ut riglnal threrra in e f f ect ; ^ ° "* *"»"°™* ( valueate)s regarding ing^^z;z^:-^z-: - r -— be small andconsidere d aserror ,i tc o1 T h "" " aSSUmed *° asther ewer en oreplications .V aue sfo rth e ^.^"^ f*™ the e~°r ofth evarianc erati ostatisti c ^^ ^^^^ ^ ' <"> ^^::^ ^rrr1:;;user data °;individua i »-^ - of thethree-factor^n^i^"^T^tT " " " Ut Wln9 to the variation in the number ofberrie s L!h A. ""* ° ° analysis is the sguare root transformn"^ ^ *'""**\ .^ **' f thevariance ,whic hgive sproblem si ninterpretatio n A ^^^^ ° ofanalysi sa choic ewa smad e fora lo a1 ^ ^ P°SSlble "^ three two-factor interactions.Th i £££??** "^ *"m "" interaction. The analysis resulted in .x ^v.l u. ^ •° !^ *"-*"*" thethree-facto r interactionwer ehig hcompare d tott e T ^ ^ "

Notnumber so fd/ioo ban dh y/100bwer euse d but ha ^^ance level. 'ou t dlhaploidsan dhybrid spe r

52 berry, multiplied with the mean number of berries in that experiment. The outcome of the testha s alimite dvalu e asth enumbe ro fberrie svarie dbe ­ tween combinations. It was still adhered to as the most practical way to find the order of magnitude of the three-factor interaction. High x2 values, indicating thepresenc eo fa three-facto r interaction,mea ntha ti n an anova the main effects and two-factor interactions are tested against toohig h amea n square oferror .Thi sca nobscur e smalleffects . For between-year comparisons, data from experiments described in Chap­ ters3 and 4wer e used.

5.3 RESULTS

The interaction experiments were very much affected by spells of hot weather during the peak of flowering, bothi n197 5an d1976 .Therefor e the datawer e not as complete asplanned .A nexac tanalysi s ofal l (incomplete) data would have given problems in interpretation of contrasts which are difficult to estimate. The problem of analysis with missing data was cir­ cumvented by making separate analyses of complete parts of the data. In every part the effects were orthogonal. This way some of the data served several times and others were omitted. The number ofberrie s obtained on oneda y from one seedparent-pollinato r combinationvarie d asth eberr y set wasno tconstant . X2 values for theestimate s ofth ethree-facto r interactionwer ehig h in thebetween-yea r comparisons indicating thata three-facto rinteractio nwa s present. Consequently, F-ratios in anovas were underestimated and critical levels overestimated, x2 values were low andno tsignifican t inth einter ­ action experiment in all anovas for d/100b and in one anova for hy/100b. Therefore the combined three-factor interaction was not higher than what error was expected to be.Her e the foundF-ratio s andcritica l levelswer e probably of the correct order ofmagnitude .

5. 3. 1 Dihaploid production

The first results presented are from experiments in which pollination dayswer e notnecessaril y the same foral lcombinations . Inth e firstyear s mostly good pollinators were used,s otha tn obetween-yea r comparison could be made with orthogonal effects and including bad pollinators. Data are given inTabl e 17 and anovas ofthre eorthogona l parts ofth edat a inTabl e 18, both for the original and for the transformed data.Mai n effects were significant with and without transformation. In part Ith e seed parentx Pollinator (sx p , interactionwa sno tsignificant . Inpart s IIan d HI the „ in« in the original data (p=0.18 and significance of this interaction was low in the ongi „„~ Hicannpared after transformation 0.15 resp.), but even this significance disappeared art

53 Table 17. Numberso fd/lOO b andhy/100b ,origina l values.Dat a are means fromcrosse smad e on several days in ayear .Als o given are means ofmai neffect s ofpart s ofth e datawit h orthogonal effects.

Seedparen t d/100b hy/100b pollinator (IVPnr. ) pollinator (IVP nr.)

10 35 48 10 35 48

Gmeke 1973 268 160 184 217 1974 251 209 342 445 8569 2065 81 1975 235 297 406 465 150 39 1976 156 237 207 1950 1323 25 Record 1973 110 86 69 123 1974 131 59 155 248 4334 624 33 1975 122 188 126 86 100 0 1976 151 89 93 1653 3880 27 Radosa 1974 154 250 700 8802 2195 86 1975 322 301 448 614 254 25 1976 223 134 160 4586 2714 27 Ultimus 1973 291 91 126 261 1974 253 291 193 283 Libertas 1973 100 34 41 57 1974 114 81 41 113

Means of main effects d/100fc> hy/100b part part

II III II

seedparen t Gineke 260 282 258 1630 Record 123 137 126 1193 Radosa 299 2145 Ultimus 224 Libertas 73 pollinator IVPl 190 IVPlo 126 181 147 3451 IVP35 144 221 195 1478 IVP48 218 315 233 38 year 1973 139 140 1974 201 285 243 2977 1975 272 229 192 1976 161 156 1798 meannumbe r ofberrie s 43 39 47 39 — Table 18. Analyses ofvarianc e oforthogona lpart s ofth edat a from Table 17an d of their transformed values(y=1001n(x+l)) .

Originalvalue s Transformed values

Source d.f. sumo f proba­ sumo f proba­ squares bility squares bility d/100b

seed parents 3 180892 0 .00 86541 0 .00 pollinators 3 42522 0 .03 20679 0 .02 years 1 30690 0 .01 11325 0 .01 s xp interaction 9 16985 0 .76 6564 0 .75 s xy interaction 3 5990 0 .60 88 0 .99 p x y interaction 3 13715 0 .28 3416 0 .44 error 9 27498 10363 total 31 318292 138976

II seed parents 2 142863 0 .00 32438 0,.0 0 pollinators 2 84498 0 .01 9211 0,.0 2 years 2 82973 0 .01 12254 0..0 1 s xp interaction 4 38482 0 .18 2493 0..5 0 s xy interaction 4 23130 0,.3 8 2026 0..5 9 p xy interaction 4 94976 0..0 3 15662 0..0 2 error 8 38168 5459 total 26 505089 79543

0.,0 0 III seed parents 1 104808 0.,0 0 32708 pollinators 2 29628 0.,0 1 6528 0.,0 9 10272 0. 07 years 3 48141 0.,0 1 294 0. 85 s xp interaction 2 7526 0. 15 577 0. 88 s xy interaction 3 9948 0. 18 8811 0. 27 p xy interaction 6 26745 0. 10 5222 error 6 8613 64411 total 23 235409 hy/100b

57672 0.01 H seedparent s 2 4085695 0.10 898659 0.00 pollinators 2 52840393 0.00 291334 0.00 years 2 35152158 0.00 14015 0.39 s xp interaction 4 7047428 0.11 43459 0.05 s x y interaction 4 6498731 0.13 39612 0.07 P x y interaction 4 46767826 0.00 23477 error 8 5193141 1368229 total 26 157585372 (p=0.50 andp=0.85) . Somemor e combinations ofpollinator s and seed parents in several years were analysed, apart from the onespresented , and nosig ­ nificant sx p interactionwa s found. The complete interaction experiment of 1975 consisted of a combination of five seed parents and six pollinators crossed on seven days in athre e week period. In 1976 four seed parents were used and three pollinators on

Table 19. Numbers ofd/100 b andhy/100b ,origina lvalues .Dat a are means ofcrosse smad e onon eday .Als o given aremean s ofmai n effects ofpart s ofth e datawit h orthogonaleffects .

Seedparen t d/100b hy/100b pollinator (IVPnr. ) pollinator (IVP nr.)

10 35 48 10 35 48

Gineke 75 -08 -18 68 84 265 290 200 73 20 75 -08 -21 62 115 225 376 420 38 67 75 -08'-2 5 330 479 568 75 -08 -28 56 367 256 401 76 -06 -09 43 56 123 5814 3167 33 76 -07 -23 471 463 276 1292 583 21

Radosa 75'-08 '-1 8 93 193 349 387 467 250 20 75.-08 --21 0 124 133 235 567 30 0 75--08- -25 373 440 998 75--08- -28 83 457 312 406 76--06--09 74 264 80 7447 3833 33 76--07--23 571 312 322 1800 475 20

Mult a 75- -08--18 33 197 374 246 233 50 75 75--08--21 0 60 176 183 133 0 13 75-•08-•25 337 332 543 75--08- •28 38 328 263 184

Record 75- •08-•18 107 172 44 86 133 0 75-•08-•21 0 104 88 0 33 0 76-•06-•09 107 78 43 3850 7700 33 76- 07-•23 191 39 87 371 0 20

Katahdin 75- 08- 18 13 19 12 67 75 50 75- 08-21 7 49 47 100 33 33

56 fourdays . Four complete (orthogonal)set so fdat awer eextracte d which are given in Table 19. Anovas of original and transformed data are given in Table 20. In a combination of three seed parents and three pollinators on four days all main effects were significant,th es x p interactionwa s not (part I).Par t II included five seed parents, three pollinators and two crossing days, and part III three seedparents ,fou rpollinator s and three days. Both seed parent and pollinator effect were highly significant. In both cases the s x p interaction was significant in the original data (p=0.03 and 0.00 resp.). The In transformation lowered thecontributio n of this interaction to the total variance to such an extent, that itwa s not significant any more. Where the scarce data from 1976 were combined with those from 1975 (three seedparents ,thre epollinators ,par t IV)th epolli ­ natoreffec twa s not significant,no rwa sth es x p interaction .

Table 19. Continued.

Means of main effects

d/100b hy/100b part part

II III IV II IV seedparen t Gineke 313 226 214 232 136 977 Radosa 360 223 224 247 222 1245 Multa 269 206 174 84 Record 86 88 42 1019 Katahdin 25 60 pollinator IVP6 48 IVP10 249 92 216 175 227 1860 IVP35 300 187 261 205 72 1360 IVP48 393 181 290 188 28 22

139 date 75-08-18 254 177 207 199 120 75-08-21 183 129 142 158 98 128 75-08-25 489 75-08-28 330 263 3546 76-06-09 96 509 76-07-23 304 meannumbe r °fberrie s 7 6 7 8 6 8

57 Table 20. Analyses ofvarianc e oforthogona l parts of the data from Table1 9an do fthei rtransforme d values (y=1001n(x+l)).

Original values Transformed values

Source d.f. sum of proba­ sum of proba- squares bility squares bility

d/100b

I seedparent s 2 50618 0.05 4188 0.05 pollinators 2 128644 0.00 15307 0.00 dates 3 465202 0.00 51605 0.00 s xp interaction 4 37174 0.29 3912 0.18 s xd interaction 6 63386 0.23 7169 0.10 p xd interaction 6 186184 0.01 16884 0.01 error 12 78709 6264 total 35 1009918 105328

II seedparent s 4 204364 0.00 274155 0.00 pollinators 2 56472 0.00 76680 0.04 dates 1 16898 0.01 12526 0.23 s xp interaction 8 39499 0.03 31628 0.81 s xd interaction 4 30891 0.01 37987 0.35 p xd interaction 2 14848 0.02 33885 0.16 error 8 9377 59425 total 29 372350 526285

III seedparent s 2 17183 0.00 18564 0.14 pollinators 3 315902 0.00 358154 0.00 dates 2 87020 0.00 92760 0.00 s xp interaction 6 32807 0.00 36495 0.26 s xd interaction 4 31243 0.00 37642 0.11 p xd interaction 6 90685 0.00 87534 0.03 error 12 10548 48229 total 35 585388 679378

IV seedparent s 2 184188 0.00 114761 0.01 pollinators 2 5504 0.61 19718 0.35 dates 3 205004 0.00 68480 0.10 s xp interaction 4 17795 0.52 6590 0.94 s xd interaction 6 90334 0.06 49052 0.50 P xd interaction 6 115044 0.03 74766 0.28 error 12 63012 104182 total 35 680880 437549 Table 20. Continued. hy/lOOb

III seed parents 4 126835 0.01 191563 0.06 pollinators 2 219945 0.00 259937 0.01 dates 1 3674 0.34 90640 0.03 s xp interaction 8 186581 0.01 302065 0.08 s x d interaction 4 21933 0.29 74491 0.31 p x d interaction 2 19366 0.13 16798 0.55 error 8 28785 105419 total 29 607119 1040913

IV/ seed parents 2 498735 0.78 224769 0.05 pollinators 2 21655459 0.00 895682 0.00 dates 3 73768016 0.00 704883 0.00 s xp interaction 4 6270671 0.23 45128 0.80 s x d interaction 6 1617241 0.94 138483 0.57 p x d interaction 6 37140551 0.00 146895 0.54 error 12 11593584 332539 total 35 152544257 2488377

The seedparen tx yea r (date)interactio nwa sneve r significant after In transformation. Pollinator x year (date)interaction , however,wa ssevera l times significant,eve n aftertransformation . It canb e concluded that themos t important interaction fordihaploi d production,namel yth es x p interaction ,wa sonl ypresent ,bu tno talways , where both main effects were significant inth eorigina l data. After In transformationn os x p interactio nwa sfound .

5-3. 2 Hybrid production

Variationwa shighe rfo rhybri dproductio n thanfo rdihaploi d production in thesam e experiments. Thecoefficien t ofvariatio n ford/100 b ine.g . Table 17 part u was29% ,wherea s itwa s49 %fo rhy/100 b from thesam e crosses. In a compilation of results from three years (Table 17an d18 ) s x p interaction washardl y significant (p=0.11)an deve n less after In transformation (p=0.39).Th esee dparen tmai neffec twa sno tver y strong in theorigina l data,a scoul db eexpecte d fromth eresult so fChapte r 4 Combined results from theinteractio n experiments of197 5an d197 6fo r hy/100b aregive n inTabl e 19.A sth ehybri dproductio ni sles simportant , i t „-r-i-c!o fth edat a with orthogonalef - !ess data aregiven . Two anovas ofpart s ortn e c ,,.t I includina five seed parents,th e fects aregive n inTabl e 20.I npar t II,mciuain g 59 s x p interaction was significant (p=0.01); it had even some significance after transformation (p=0.08). Part IV (three seedparents ,thre epollina ­ tors, fourdate s from two years), where the seedparen t effectwa ssignifi ­ cantafte r transformation only, showed no significant sx p interaction. Ofth eothe r two-factor interactions thepollinato r xyea r (date)inter ­ actionwa s sometimespresent ,bu t aftertransformatio n itwa s reduced;onl y in the between-year comparison (Table 17) thecritica l level was stilllo w (p=0.07). About s x p interaction it can be concluded that it was found oncefo r hy/100b. Logarithmic transformation reduced this, but did not remove it completely.

5.4 DISCUSSION

5. 4.1 Dihaploid production

The analysis of data from Table 17 showed no interaction between seed parentan dpollinato r effecto nth eproductio n ofdihaploids . Inth einter ­ action experiment (Table 19),wher e the influence of pollination date was

the same for all combinations, the s x p interaction was found only in those parts where data of a bad pollinator or low producing seed parent could be included (Table 20, II and III), itwa s probably due to themor e careful collection of data that this interaction was significant here and 2 noti nth ebetween-yea r comparison. Itwa s alsoher e thatX values forth e three-factor interaction were low. The fact thatth e Intransformatio nre ­ duced the interaction component in the anova to a not significant level, means that only the multiplicative part of the s x p interaction hadbee n present. Where only pollinators and seed parents with high d.i.a. and d.p.a.wer e involved, themultiplicativ e effect could not appear,henc eth e absence ofinteractio n there, even inth eorigina l data.Th e multiplicative effect means that the combination of apollinato r with high d.i.a. witha seed parent with high d.p.a. gives even higher dihaploid production than couldb eexpecte d from their individual performances. The absence ofinter ­ action in a strict sense implies that a good pollinator is good irrespec­ tive ofth e seedparen tused . Forpractica l dihaploid production thismean s that the best pollinator for a certain seed parent is also the best for others. For dihaploid research the implication is that testing for d.p.a. can be done with one pollinator and testing for d.i.a. with one seedpar ­ ent wxth the limited data and themultipl e use ofpar t of the data sucha conclusion should be drawn with some caution, especially as the genetic variation in the group of pollinators was limited. The parentage of the cultivarso f5 . tuberosum seems tohav ebee n sufficiently diverse to assume genera applicability ofth econclusio n atth e sideo f the seedparent . These observations did not provide an argument for introducing theter -

60 minologyo f general and specificdihaploi dproducin g orinducin g ability in S. tuberosum as Eenink (1974) proposed for Brassica oleracea. In Solanum the dihaploid production of specific parental combinations can apparently be deduced from the general performance of the parents,provide d that the multiplicative effect istake nint oaccount . Experimental errorprove d tob econsiderabl e ininteractio nexperiments , evenwhe n crossing was carried outaccordin gt oa balance dplan .Thi s error might have been the reason why the correlation for dihaploid production found by Hermsen & Verdenius (1973)betwee n two cultivars was not as high asthe y expected'.Thei r theory ofa differentia l reactiono fth etw oculti ­ varst o the same series ofpollinator s isno tnecessar y asexplanation .Th e reasonably high correlation they obtained (r=0.65;n=22 )ca nsuppor tth en o interaction conclusion. The pollinator x year (date) interaction found fittedwit h thetempera ­ ture sensitivity ofth epollinato r ford/100 b inth egrowt h chamberexperi ­ ment (Chapter 6). Apparently thedifferen tpollinator s areno tsensitiv e to theenvironmen t to the same extent,henc eth einteraction .

5.4.2 Hybrid production

The high variation in hybrid production might have been the reason why an existing interaction was not observed. Inth emor eaccurat e interaction experiment a s x p interaction was foundonce ,bu t itwa sno tver y strong. There is probably no s x p interaction for hy/100b, but the existence of this interaction could not be excluded with certainty. Even if it would exist, it would not be an important factor indeterminin g the levelo fhy ­ brid formation. As the removal of hybrids thatar eproduce d alongwit hdi - haploids only plays a minor role in the amount of work involved inmakin g dihaploids, this-possible interaction isno to fpractica limportance . Thepollinato r xyea r (date)interactio n foundwa sexpecte d from theob ­ servation that IVP10 and 35 were veryvariabl e inhy/100 b during aseason , incontras twit h IVP48.

61 6 Influence of temperature on frequencies of dihaploids and hybrids

6.1 INTRODUCTION

It follows from Chapters 3 and 4 that dihaploid and hybrid production are determined genetically, m this chapter the influence of one external factor, temperature,o nth e frequency ofdihaploid s andhybrid s is studied. Gabert (1963)di dno t findtemperatur e influence ond/100b ,bu tWöhrman n (1964) found a negative influence of high temperature on d/100b and berry set (see also Section 2.4).Bot h authors experimented with temperature treatments afterpollinatio n only. informationavailabl e onvariabilit y isno tver y detailed.Dat a ondiha - ploxd frequencies are usually pooled per year, so that the fluctuations durxng the year are lost. Several authors have found year toyea r fluctua­ te», in d/lOOb and hy/100b (Gabert, 1963; Frandsen, 1967; Hermsen, unpubl.)..A preliminary study with unpooled data from crosses on different days xn 1971 indicated that fluctuations occurred within the season ina ZZZ°Zd greenh°USe (Van kreukelen, 1972). There was more variation in andT, I" " d/10°b' eSpeCia11* in ** «ed setpollinator s like IVP35 and 32.Whe n a similar pattern was found in1973 ,a nexperimen twa s setu p todetermxn eth e xnfluence ofth e temperature on d/100b

tionsYTr °f the eXPeriment WaS to «n- the best temperature condi- ld Pr dUCti0n and als InZ o£\TT ° ° f obtainmor e insight in themech - theIc h /d Pr°dUCti0n- AS this —ot be separated completely from Product^ Ybrid Pr°dUCti0n' the ^1-nce of temperature onhybri d ch1e r "H l0°ked int°- The «*«*»«* was carried out in growth the tem'lrt ^ ^ ^ *° *"* ^ ^^* constat. To separate are cZ7^7 ^^^ °*^ ^ »^ ""° n ^ regimes :i " ^ ^^ *» Cambers with different temperature growthch l „T "^ ^ alS° P°SSible t0 transf- Plants between t0 C ntr lled de to sefat Wh h ^ ^ * ° ° <*""* ^ temperature, inor - chLgLg dihaolld T* ^ '^ ** ^^^ure becomes effective in Ybrid fre<ÏUenci-- would give additional infor- matxoTeZTlTTnabou tth emechanism °: s involved. W

6.2 MATERIALAN DMETHOD S

CultivarGinek ewa suse d assee dnaren tir , IVP35an d4 8 aspollinator s (seeSectxT s ^ T"" "* * **""*" ^^ { Sectxon 3.2). Allplant swer e grafted onto 62 tomato rootstock, potted in well drained plastic buckets of 8-10 1 and placed in growth chambers about 14 days before the first flowers opened. Gineke and S. phureja plants were distributed at random over two growth chambers.Code s indicating plants andtreatment s aregive ni nTabl e21 . Thegrowt h chambers usedwer epar to fth ephytotro na tth eDepartmen to f Field Crops and Grassland Husbandry ofWageninge nAgricultura l University. Each chamber measured 4.5 m x 3.2 m x 2.2 m. Plants wereplace d ontable s so that the flowers were about 0.5 m below the lamps. Fluorescent tubes (PhilipsTL33/4 0 W)an d some additional incandescentbulb sprovide d aradi - _2 _2 ation intensity of 175 J-m -s at plant height. C02-concentration was kept at 300 mg-kg~ .Temperature , moisture and daylength were set as fol­ lows.Light swer e on for 16h pe rday ,durin gwhic hperio d therelativ ehu ­ midity was 70%.Durin g the 8 h dark period the relative humidity was90% . Inth e chamber with low temperature (L)th e temperature was kept at 18° C throughout. Inth e other chamber (H)th etemperature s forth e first1 2h o f light were 22 °C in 1974 and 23 °C in 1975. During the remaining 4 h of light and 8 h of darkness the temperature was 18 °C (in 1975 after 6/8: 15° C in darkness). Good air circulation madeth etemperatur eequa l inal l partso fa growt h chamber. White fly (Trialeurodes vaporariorum) teemed during thecrossin g season of 1974 and was controlled with dichlorephos.Thi scause d abortiono fpol -

Table 21. Codes indicatingplant s andexample s oftreatment s intw ogrowt h chambers withdifferen ttemperatur eregime san d a greenhouse.Al l transfer dateswer ei nAugus t1975 .

G = S. tuberosum cv.Ginek e 35 = S. phureja IVP35 48 = S. phureja IVP48 L =plan t inchambe rwit h low temperature H =plan t inchambe rwit hhig h temperature Gr =plan t in greenhouse LH17= plan t transferred fromchambe r Lt ochambe rH on17/ 8 HL5 =plan t transferred fromchambe rH t ochambe rL o n 5/8

Examples oftreatments :

GLx 48 H = crossbetwee ncv .Ginek e inchambe rL wit h IVP48 inchambe r H GHx 35LH1 6 = crossbetwee n cv.Ginek ei nchambe rH wit hIVP35 , transferred from chamberL t ochambe rH on16/ 8 GHL5x 35 L = crossbetwee ncv .Gineke ,transferre d fromcham ­ berH tochambe r Lo n5/8 ,wit h IVP35 inchambe rL

63 linated flowers in crosses made from one day before until 4-5 days after spraying. The negative effect of dichlorephos on berry set was determined only by the end of the crossing period owing to misleading results ofa former sprayingexperiment . In197 5 Encarsia formosa, aparasiti c wasp,wa s successfully introduced intoth echamber s tokee p thepes t at lowlevels . Fresh pollen was applied to emasculated flowers of Gineke. Seeds were extracted fromeac hberr y separately and screened for embryo spot (seeSec ­ tion 3.2).Thre e groups were distinguished: seeds with embryo spot, seeds withoutembry o spotan d seedswithou t avisibl e embryo.Th e third groupwa s discarded. All seeds without embyro spot were classified as dihaploids. Prevaous experiments actually displayed less than 1% pseudogamic tetra- ploids and lesstha n1 %hybrids . Ifsom e spotless seedswoul d nothav ebee n dihaploids,th eresultin g errorwoul dhav ebee nver ysmall . Gineke plants in specified conditions crossed with S. phureja plantsi n specified conditions on several successive days are called atreatment .I n 1974, fourt oeigh t flowerswer epollinate d per treatmentpe r day.Thi swa s increased to ten flowers in 1975.Dat a from singleberrie s werepoole dpe r daywithi n onetreatment .Nearl y 7000 flowers in allwer e pollinated. Mostplant swer e lefti n thesam egrowt h chamber throughout the crossing

period.B yusin gpoli en from both chambers to pollinate seed parents ln one chamber, four treatments could be made perpollinator , e.g. GL x 35L,G Lx 35H GH x 35L and GH x 35H. These are called -fixed treatments'. There - ults of these treatments were compared with those of the conditioned greenhouse, described in Chapter 3. m other treatments plants were ex­ changed between the two chambers. This was done inbot hyears ,bu t onlyi n 1975dat awer ecomplet e enough tob e reported.

0Ut f thS GinekS Plants in each m, ° Camber a group of plants was chosen and pollinated every day with pollen from 35L. After eight days, on 5/8, J"tSnWere transf—d from chamber L to H (GLH5) and from H to L (GHL5). Pollinations with 35L were continued for 21 more days. Two other

Ü P llinated transfe " "d(GLH1 T"^?an d^ ' ^' °^ ° ^ ^^. **** 35^L an d after^ 12 days, on17/8 , tmued for1 6days . on16/ 8T^rJ ^ fr°m eaCh ChambSr WaS tra-ferred to the other chamber 35HL16> P llinati ns we on seed ^ ' ° ° ~ -de inth eperio d 15/8-31/8 P a re n Pl 35LH27a n d 3 5H%;:^" ^ " ^ ^ "*» ™« transferred on 27/S 7/8 31/8 S6e d analysiAnalysis ofT varianc '• "" e (anova"*" *"")wa s* rarH" ^ °"*. •, *««*• *«Cambe r H. andtemperatur e regimes of th. T ^ ** rePlicationS regimes of the seedparen t andpollinato r as factors. 6•3 RESULTS

From the experiment in 1974 it-„= „ , were suitable for FT • learned that the temperatures used -table for flowering and berry set.Berr y setwa s almostni l inth e 64 periodi nwhic h insecticides were used. Thebes tberr y setwa s obtainedi n chamber Lan dwit h pollinator IVP35.Therefor e these were useda smuc ha s possiblei n197 5i nthos e treatments,i nwhic hplant swer etransferred .Th e milder pest control in197 5le dt ohighe r berry set. Since thelas ttw o weeks showed thebes t berry set,mos t data usedar efro m that period.Fo r 1974, only data from thefirs t tenday s were used andonl yo fth efixe d treatments. Data from transfers intha t year were very incomplete. They served,however ,t oimprov eth etransfe rprocedur efo rth enex tyear .

6.3.1 Treatments with fixed temperatures

Coefficients ofvariatio n ford/100 b andhy/100 b between crossing days within treatments in197 5wer e calculated both forth ewhol e period(4 0 days)an dth elas t1 5day swit hhig hberr y set.The ywer ecompare dwit hth e variation ina conditione d greenhouse duringth ecrossin g seasonso f 1974 and197 5 (Table 22).Variatio ni nd/100 b wasabou tequa li ntreatment s with IVP35an d4 8i nth egrowt h chambers,bot hi nth ewhol ecrossin gperio dan d inth e last1 5days . Variationwa shighe ri nhy/100b ,especiall yfo r IVP35. Variation inth egreenhous e in1974 ,a goo d crossing year,wa ssimila rt o

Table 22. Range,mea nan dcoefficien to fvariatio n (c.v.=s/x)fo r diha- Ploidspe r10 0 berries (d/100b)an dhybri d seedspe r10 0berrie s (hy/100b). Data from crossesi ngrowt h chambers (1975)an dagreenhous e (1974an d 1975).Fo rcode sse eTabl e21 .

hy/100b Treatment Numbero fday s d/100b

a ™~=r, r- v ranqe mean c.v. total crosses3 range mean c.v. range GLX35L 40 20 SO-65 0 231 0.64 71-2943 572 1.29 GLx 35L b 15 8 100-40 0 199 0.49 71-29 0 1790.6 1

OOrx 35C Rc 84 12 100-54 0 318 0.55 100-6963 2059 0.88

GGrx35Gr* 73 17 33-389 171 0.60 0-300 8171.3 0

GLx48L 40 21 33-525 244 0.58 25-1 GLx 48L b 15 9 38-52 5 271 0.59 25-8 8 380.6 9

GGrx48GrC 77 n 200-1230 505 0.53 0-100 600.5 7 d GGrx48Gr 47 7 100-513 278 0.58 0- 80 271.2 1

a. numbero fdays ,o nwhic hcrosse swer emad e *>•th «.ueshorte- _...... r crossin. ..._...g perio,„dwa „M s«par narto tfo thf th ee totalcrossin gperio d m the growth chamber c- in197 4 d.i n197 5

65 thato fth eshorte rperio di nth egrowt hchambers ,wherea sth e greenhouse

vacationo f19 75wa susuall yhighe ran dclose rt oth ewhol egrowt hchambe r period.Mean so fd/100 ban d hy/lOObi nth egrowt hchamber swer eclose rt o thoseo f197 5tha nt othos eo f197 4i nth egreenhouse . se2lTnenCe °f the tertperatUre - thesee dparen tan dth e pollinator 74an d17 StUdled Wlth ^^ °fth S relativel* Stabl* "»t periodi n Uli 23 n? Perl°d ln 1975- NUmberS °f d/1°0b and h^100b *~ 9lveni n Table2 3 Dataar emean so ffou ro rfiv ecrossin gday swit hon eda yintér ­ ieures S1°nS ^ ^^ " ^^ ln£luence °fhi *h °r ^ *»" LnatorI" 6 ^ " ^ ^ °** » f°Ur treatmentswit hth esam epol - ncefo rXVPT 3 "" ""^ °Ut ^ ^ Cro"ln» <^S as replications, va sL t /: 1974 n0t en°Ugh COmPlete data Were aVailable *>ra nano - Period Result ' talled) "" applled t0 data *"» *» wholecrossin g ^e two* ^tlng S19nifiCa- 1—1- ofdifference sbetwee nth eeffect so f P llinato 2? DTfferetc 7 ^ ^^ °' ° - areals opresente di nTabl e t0 temPeratUrS Were 9^4 Hesults f r,h »o« —intent in197 5 thani n .report: ^ rr:poiiinators ——• — - ence of the ten,™ T "^ Xt Can be concluded that, whereas the influ­ ent. L tempi, : " ^ "^ '^ *»d/1 °°b WaS ^»le or ab- re a ma11 Poll nator n d7io0b F / ^ ^^ *°*^e effect on the on the seed parent whenlvpsf0015 ^ ^"^ ^ * -•*" i~ tent, LOW temperature \ 7 "" M' ^ f°' IVP48 data were in—is" 9 P tiVe inflUenCe n Pecially onT^ri m .I d ;;; ° — Pomnators,es - most marked. differences those in hy/100b for IVP35 were

— ner;:/™::;hy/ioobfro m crosses—— andhig h(H )temper^ ! "* " ^ «*"*"•» itt ^

Seedparen t Pollinator197 4 1975

IVP35 IVP48 IVP35 IVP48 d/100b L 240 218 251 266 H 121 132 199 263 L 325 156 246 224 H 145 150 181 201 hy/100b L L 850 39 236 H 31 428 25 133 19 L 1400 21 430 43 H 541 7 174 37

66 Table 24. Influence of low (L)an dhig h (H)temperature so n d/100b and hy/100bvi a seedparen tan dpollinator .Positiv e (>) andnegativ e (<)influenc e oflo wtemperature s isconclude d from datai nTabl e 23.Critica l levels (p)ar e fordifference sbe ­ tweenL an dH and calculated with ananov ao na northogona l part ofth edata , for IVP48 in197 4wit h asig ntest .

Influence via 1974 1975

IVP35 IVP48 IVP35 IVP48

d/100b seed parent LH0.6 7 L>H 0.61 pollinator L>H0.0 0 L>H0.0 2 L>H0.3 6 L>H 0.16 hy/100b seed parent LH0.1 2 LH0.0 1 L>H0.7 7 L>H 0.03 L>H 0.80

6- 3. 2 Treatments in which the seed parent was transferred

It is of interest to know whether seed parents react immediately to transfer to another temperature regime or only after a lapse of time.A s the clearest influence on the seed parent was with IVP35 in 1974, only IVP35 pollen was used on transferred seed parents in 1975.Mea n dihaploid and hybrid frequencies are given in Table 25 for periods of five toseve n consecutive days of crossing before and after transfer of two groups of seed parents. As could be expected from the fixed treatments, no big dif­ ference in d/lOOb was observed between the treatments. For hy/lOOb clear differences between the treatments were observed. Before and after the first transfer more hybrid seeds were formedo nsee dparent s originally in chamberL tha no n seedparent s fromchambe rH ,bu tofte nth edifferenc ewa s small,o fth e seedparent s ofth esecon d transferth eplant s fromchambe rL produced consistently lesshy/100b ,bot hbefor e andafte r transfe^

,.„f transfer for d/100b and hy/100b. There was no change at the moment of transier roru / The direction of the differences in hy/100b was inconsistent: itwa s dif- ferentfo r the twotransfers . 6- 3. 3 Treatments in which the pollinator was transferred

From the clear effect of temperature onth epollinator s infixe dtreat ­ ments, especially on IVP35,change scoul db eexpecte d aftertransfe r ofth e

67 Table 25. Meanso fd/lOO ban dhy/100 bfro mcrossin g periodso f5- 7days .Dat afro msee dparent strans ­ ferredfro mon egrowt hchambe rt oanother ,compare d withdat afro mnon-transferre dsee dparents .Al lpol ­ linationswit h35L .Date si nAugus t1975 .Fo rplan t codesse eTabl e21 .

Date d/100b hy/lOOb

GLH5 GHL5 GL GH GLH5 GHL5 GL GH

1-5 236 263a - - 1166 513a - - 6-12 191 189 209 210 448 188 283 370 15-20 218 221 198 202 333 302 197 232 21-26 365 332 254 275 613 541 221 562

GLH17 GHL17 GLH17 GHL17

12-17 143 160 181 390 19-24 233 238 337 530 25-30 223 269 190 500

a.onl yon eobservatio n no data

pollinators from onetemperatur e regimet oth eother .Mea n frequencieso f Table2 6 "" **"*" ** 8Ucoe"lw ^^ -"er transfer aregive ni n thanftrer «IflrS t tranSfer (35LH16 and 35HL16) d/100bwa sl°we rfo r35LH1 6 Ts2ll°\TT f0r ten daYS- ^ the neXt five da*s Vl°™ washighe rfo r H^ I a dlffSrenCe WaS Sma11- *«erth esecon dtransfe r (35LH27an d mi r toI? 3 C°Uld °nly be made f°r a Sh°rt Period- «* -suits were 0llina ducedles sd ,° " °'^ 3 Th"" edat * aWSr^ analySee^ <> ldn ^s placed inchambe rH in - trient"r - in g assource - ~^ so f variatih thepollina n - and T*T7atri : ;f::: t T ° - - - rio-edl17/8-3!/ fo r U^L^^Z'« „ lxuenc e was0.02 JZ,wherea ^ "" s^thi "^^swa s0.1 6 fo"*rth epe -^^

ob poiiin,t rs Mgh.rP;:ru„n L; " ° <*-•«*- «• «„.».er B..... auctionth „ therrtr.n.f.rr. d oounterpart.afte rbot htransfer s

68 Table 26. Means of d/100b and hy/100b from crossing periods of five days. Data from pollinators transferred from one growth chamber to another, compared with data from non- transferred pollinators. All pollinations on GL, except 35LH27 and 35HL27 which were on GH. Dates in August 1975. For plant codes see Table 21.

Date d/100b hy/100b

35LH16 35HL16 35L 35H 35LH16 35HL16 35L 35H

17-21 185 252 247 225 335 93 260 150 22-26 211 289 254 182 253 129 221 122 27-31 176 154 - - 288 143 - -

35LH27 35HL27 35LH27 35HL27

27-31 176 190 367 273

dataver y incomplete

and for the whole period. In an anova with pollen treatments and crossing days as sources of variation the critical level forpollinato r treatments wasver y low (p=0.003).

6.4 DISCUSSION

in different temperature conditions the number of dihaploids per 100 berries was a rather stable character. Variationi nd/lOO bwa s regularan d low, compared with hy/100b, both in thegreenhous e andi nth egrowt hcham ­ bers. IVP35 had highervariatio n than IVP48i nmos tconditions . W«*»» hada ninfluenc e on d/100b onlyvi ath epollinator .Especiall yi nIVP3 5 temperaturesha d apositiv e influence. differences found The higher variation in hy/lOOb was reflected in the ^nsitive u 4. ^ T,n>-*Ran d4 8wer etemperatur e sensitive between the fixed treatments. Both IVP35 and4 Uw e * ,!<= Positive influence of low tem regarding the production of hybrid seeds. Positive

Perature on the seed parent was more outspoKen n tre-en. w;th IVP35 than in those with IVP48. With the higher level ofhy/X 0b ^ wasmor e room for influences to show. Results from 1974 well ^H narentwa stransferred ,n o effect la the experiments, in which the seeP««*J treatments. was expected for d/100b, considering the results ox

69 Thiswa sconfirme db yth eresults .Fo rhy/100 ba negativ eeffec ti nchambe r Lwa sexpected .Thi sinfluenc eprove dno tt ob econsisten ti nthi sexperi ­ ment.Alread ybefor etransfe rth eplant s inchambe rL ha dhighe rhy/100b . This continued after the transfer. Before the second transfer plantsi n chamber Lha d lower hy/lOOb, as expected. These,too ,wer eunaffecte db y thetransfe rfo r1 2days . After transfer of the pollinator apositiv e influence oflo wtempera ­ tureso nth epollinato rwa sexpected ,considerin gth eresult so fth efixe d

treatments,bot hi nd/100 b and inh y/100b.Fo rd/100 b thiswa sconfirmed . After both transfers 35LH induced the lower number of dihaploids.Fo r hy/100b,however ,th erevers ewa s found:i nbot hcase s35L Hproduce dmor e hybrids. For d/100b results -adapted- toth ene wtemperatur e aftertrans ­ fer; forhy/100 bthe ydi dno tadapt ,eve nafte r1 5days . Apparently, the two temperature sensitivemechanism s inth epollinato r influencingd/iOO ban dhy/lOO bacte da tdifferen tmoment sdurin gflora lde ­ velopment.Th e mechanism for the induction of dihaploids reacted fastt o newconditions .Dat afro mth e firsttw oday safte rtransfe rwer eno tcom ­ plete butafte rtw oday sth echang eha dalread ytake nplace .Fo rdihaploi d inductionevent si nth elas tphas eo fpolle nproductio no ri nth estyl ean d LILve^ P°llinati0n °r «lately afterwards must be temperature

theToneinPratUreffeC t° nhy/10 °bVi ath ePoU-ato r•»» *nav eworke do n brid seeds ^ ^ ^^ P°ll6n **>«"••• ™» *«»up< * **' Sted mainly ftetraPl0l d hyb dS S Phureja ;o°iT ° ^ < <«*-, fromunreduce d f P lle greatestDar tf ' ^^ ° ^^ ° " -^termine*fo rth e UtS dUrlngmei0Sis llTTJTi r -^^-- i» S,phureU takesplac e antheSiS l Pa" why durZ :h " "^ '° "°*' -—> •Thi scoul dex - m numbLo f hy;rd see: —*^°« ^ - » days after t^r, the chamber ^TL^i" 'T ^^ *° *"* ^ ^ >^ **» *» ended 15dav sS t P°lllnat°r h*d been »>.fo»transfer ,whe nth eexperimen t ™T^ir::::-:: -- - —ge« ^ be

sûmes a fast teZC ^^ ^ ^ ""* *«»*»«*.. " ^ ^ reaction in hy/Zb r6aCti°n °f ^ P°lllnat°r *" *™>> «" * sl°« thatit;;^::mt:r ture infiuence onth e seed* — *>* d/i°°b ^ seedParen tfoTwoor; 5"1 7" * ^"^ ^ "^ the tófl"^ °"th e consistenti ntral ' '"" ^ *««»«*.. theresult swer ein - Parents nott o IT^LZTT' ^^ ^ *" 3 **°*™* '«" »«* thati twa sth eL *tranSfer - This »1« reactioncoul dindicat e andno to fsee dde v? " dete™ined thenumbe ro fovule s ina novar y tive. developmentafte rpollination ,whic hwa stemperatur esensi -

70 From the results of experiments with temperatures it can be concluded thatwithi n the range of 18-23 °C nopositiv e effectso nd/100 b canb eex ­ pected from aspecia l constanto r alternatingtemperatur e forth esee dpar ­ ent.Thi s confirms Gabert's (1963)findings .Th elowerin go fd/100 bb yhig h temperature found by Wöhrmann (1964)wa s notconfirme dwithi nth etempera ­ turerang e used. The fact that he and Gabert (1963)di d notpa y attention to conditions in which the seed parent was grown before pollination, will not have affected their results very much. The higher berry set on seed parents in chamber L could be areaso nt otr yt omaintai na temperatur e of 18° c during the flowering period. Conditions forth epollinato rprove dt o bemor e important. Low temperature (18 °C)ha d apositiv e influence ondi - haploid induction. Dihaploid frequencies can probably not be improved by lowertemperatures , as temperatures lowertha n1 8 °Creduc e flowering (Bod- laender,I960 )an d are difficult tomaintai n ina greenhouse . If space is limited in low temperature growing facilities, itwil l be advantageous toplac e pollinators therepreferentiall y andt ogiv eth esee d parents the second best place. Low temperature for pollinators will also resulti nhighe r hybrid frequencies,bu tthi swil lno thav enegativ econse ­ quences forpractica l dihaploidproduction .

71 7 The mechanism of dihaploid induction

7.1 INTRODUCTION

After the discussion of the genetics of the pollinator effect ondiha ­ ploid induction the question remained how the pollinator effect operated. Through what kind of mechanism did pollen influence the frequency ofdiha - ploids, towhic h itdi dno tcontribut e genetically? Considering the processes after pollination of a4 x S. tuberosum flower with pollen of a 2x S. phureja, it can be assumed that 5x endosperm is lethal and that seeds with a dihaploid embryo have 6x endosperm (VonWan ­ genheim et al., 1960;Bender , 1963) (seeSectio n 2.5.2.1). Pollenmus tpro ­ vide 2xchromosome s forthi s 6xendosperm . Isthi s 2no rn pollen ?Doe sdi ­ haploid induction follow the 'maize mechanism', in which one sperm ferti­ lizes the polar nuclei and the other sperm is lost, or does it followth e 'Solanum mechanism' (Hermsen, 1971), where two sperms both fertilize the centralnucleus ,separatel y or fused? This chapter deals with cytological and microscopical observations in order to try to elucidate these questions. Frequencies of 2n pollen were estimated in a series of pollinators in order to relate these to the fre­ quencies of dihaploids and hybrids that derive from them. Confirmationwa s sought of the lethality of seeds with 3x embryos and an estimate wasmad e of the lethality of seeds with 4x embryos.Number s of dihaploids and hy­ brids of a range of pollinators were compared inorde r to find outwhethe r dihaploids and hybrids originated from pollen with the same ploidy level. Pollen tubes were studied in vivo to screen for restitution nuclei and other abnormalities inpolle n tube mitosis. To check Höglund's (1970)hy ­ pothesis of preferential fertilization or faster growth of 2n pollen, as compared withn pollen ,th e relative growth rateo fpolle n tubes ofdiffer ­ ent ploidy levels was measured in styles. Finally, the effect of delayed pollination ondihaploi d frequencieswa s determined.

7.2 MATERIALAN DMETHOD S

Material for experiments described in this chapter consisted of fresh pollen from S. phureja clones andpollinate d styles and ovaries from S. tu­ berosum at different stages of development. The origin of the s. phureja clones is described in Section 3.2. s. tuberosum cultivars usedwer eGine - ke, Hulta and Radosa, together with Gineke dihaploids G254 and G609, de-

72 scribedi n Section 8.2, anddihaploi d derivative GB24. Frequencies of 2n pollen grains were estimated by the pollen diameter. Freshpolle n was dusted on a slide andstaine dwit hlactopheno l acidfuch - sin. The diameters of 100 to 200 pollen grains were measured in eyepiece graticule units and a frequency distribution was made. Assuming that 2n pollen grains,ha d twice the volume ofn polle ngrains ,thei r lineardimen ­ sions should differ by a factor 1.25.Grain swit ha diamete ro fabou t 1.25 timesth ediamete r atth e lowermod ewer erecorde d as2 npolle ngrains .Da ­ tawer e collected in 1975 and 1976. Frequencies ofpolle nwit h four germ pores,a n indication for2 npollen ,wer erecorde d from sampleso f2 5polle n grains. Seed collapse was checked in young berries of Gineke,pollinate d with pollen of IVP10, 24, 35, 48, Gineke or G254. Three sets of pollinations were made at monthly intervals. Three berries of each combination were picked one and two weeks after pollination.Th eberrie swer edissecte d and thenumbe r of developing ovules counted.Full ydevelope d seedswer ecounte d afterripenin g of the remainingberries . Correlation coefficients between d/100b andhy/100 bwer ecalculate d from datai nFig . 1,Table s 6an d 7an d datai nVa nBreukele n (1972). Sperms were studied in pollen tubes invivo , after pollen tube mitosis wascompleted . The method for preparing styleswa s adapted fromMontezuma - de-Carvalho (1967). Gineke styles,pollinate d with S. phureja pollen,wer e fixed after 24 h in a 3:1 mixture of ethanol and acetic^acid for one day^ andkep ti n 70%ethanol .Style swer emacerate d in1 mol- 1 HCl at 10min , rinsed, placed in Feulgen stain anddissecte d ona slid e inaceto - carmine. Only the lowest third of the style was used and dissected with needles to very fine bundles. The material was gently pressed. The number and relative size of the nuclei wasrecorde d foral lpolle ntube so fwhic h a long section near the nuclei was not broken.Afte r initial experiments the distance between the two sperms was also recorded. Pictures were made ofa numbe r of the tubes. Thenegative swer eenlarge d toth esam emagnifi ­ cation and the nuclei, measuring 1-4 cm, were cut out very close edgean dweighe d for anestimat e ofth esize . Certation was measured by placing two small amounts ofpolie n atop p •it* sites on a stigma and comparing the length of thebundle s ofpollen tubes after 24 h. Styles were prepared for UV fluorescence according to ««tin (1959). They were fixed in amixtur e of acetic acid for-.!» nd 80% ethanol in a 1:1:8 proportion for2 4h an dmacerate d m 8mo l1 HO H

** S h. staining was done in0.1 %anilin eblu e ^^V?^ acnVon -

(0.1N , K3P04. After staining the style «™^ ^^^Z* taining one bundle of tubes. Before pollination the style h

»ith a razor blade to distinguish between the two halves of the st1 ana to f •-,• ->\ BV this method thetw ohalves , witn to facmtate splitting (seeFig . 2).B y th maceration and afferent types of pollen, could be recognized even 73 Fig.

Fig. 2. Markings on the stigma for certation measurements. The stigma was split and a small part (A)remove d aton e side ofth e incision.Polle nwa s applied ateac hstigm ahal f (arrows). Fig. 3. Schematic picture of two half styles with thinning pollen tube bundles. The top half shows little thinning and the lower half much. Arrows:place s ofpolle n application.

splitting. The two halves were placed alongside on one slide in glycerine and pressed gently. Both pollen tube bundles could be seen simultaneously under lowmagnification . Foreac hreplicatio n of acompariso n oftw opolle n sources five to ten styles were used.Note swer emad e of the relative dis­ tances coveredb y thepolle nbundle s and the thinning ofth ebundle s inth e three to eight best preparations. From these notes it was decided whether onepolle n typewa s growing fastertha nth eother .Som e comparisons ofpol ­ len sources were replicated several times. Tetraploid styles,mostl y from Multa,wer e used in allcase sbu tone .Diploi d styles are usually thin.Di ­ ploid GB24wa schose nbecaus e ofit srobus tstigmas . The effect of the age of the flower atth e timeo fpollinatio n ondiha - Ploid frequencieswa s determined bypollinatin g Gineke flowers of different ageswit h pollen of IVP35 on oneda y ina growt h chamberwit h lowtempera - tures (seeSectio n6.2) .

7.3 RESULTS

7.3. 1 Determination of 2n pollen frequencit germTr dlStrlbUti°nS °£ P°"« ^i-eters and counts of thenumbe ro f Iedere T\ ""* ^ ^^ *' ^^ ^^ ***»encie-o f2 npol - ted fr m the freqUenCy xlble 27 D ° attributions and are presented in eiS f POllen fr0m IVP35 meaSUred sions as'H l ° — - *™ otherocca - as well with the following results: 23, 10, 13 and 1% 2n pollen. 74 Table 27. Percentages of 2n pollen of S. phureja polli­ nators determined by measuring pollen diameters. Also given are numbers of hy/100b produced by cv. Gineke with the same pollinators in two years.

Pollen source •% 2n pollen hy/100b

1973a 1975 1976° 1976D 1974197 6

IVP1 0C 0 0 - 75 75 IVP10 4 6 2 6 8569 1950 IVP24 1 0 0 0 85 131 IVP26 - 0 0 - - 97 IVP32 30 7 - 10 3283 7400 IVP35 0 0 1 5 2065 1323 IVP48 oc 0° 0 0C 81 25 IVP66 - 0 0 - - 196 sample size 200- 200 100 100 Perpollinato r 2000

- noobservatio n a-dat akindl y providedb yDr .M.S .Ramann a b-polle n collected ontw odifferen tdate si n 1976 c-value sbetwee n 0.0an d0. 5

1*35wa shighl y variable forthi s character.Th efrequenc ^ ^ ^ther Pollinators were variable aswell . High seedse tpollinator s (IVF , ^ 35)ha dhighe r frequencies thanlo wsee dse tpollinator s (IVP24, , cc> .•*.», ^„r aermpore s showedth esam e «). Thefrequencie s ofpolle n grains with fourger mp p.««„. „„„>.„ o£ w/100b « im »,»» .« -; ™ ™:"- °fth epollinator s were used, have been included inTabl e son.

7-3.2 ovule lethality

Numberbers o offdevelopin developing govul ovulese , counted one ^^°^^nZers de- tlchi>n™ , an... d, number. s of_ matur. e seeds are given m ovules, cease>dd sharplsharplyy witwithh th theag e eago ef th ofe berry the berry, oiuy. Only^ abou^^ crossest 5/„ _ In ^ Which .initiall y started to grow, reache_^_d maturj ™=,1-iiritiy V in ^4 X *^ « ^^ ^ Gineknekex Gineke (4xx 4x )th edecreas ewa sles s ^ ^ reductioni n leas3t40 %o fth edevelopin g ovulesdi dno tproduc roduceea s - ^ ^ ^ t40 %o fth edevelopin g ovulesdi dno tP ^ ^ crosses thani nth e ovullenumber swa si ngenera l slowe- ••ri nth eAfirs " 1"bptsfirs e tse to i 75 other two. Final seed set was high only with Gineke as pollen parent and reasonable with IVP10 in the third set. Variation within samples was not high.

7.3. 3 Correlation between dihaploids and hybrids per berry

Correlations between d/100b and hy/100b were calculated on data from full-sib S. phureja populations. Correlation coefficients, including those calculated for all pollinators used in one season, are given in Table29 . All correlationsbu ton ewer enegativ e andsom ewer ehigh .Correlation s for allpollinator s from one seasonwer eno tver yhigh .

7.3.4 Pollen tube mitosis

Information about sperms after pollen tube mitosis before fertilization was obtained from observations of pollen tubes invivo . Inth emajorit yo f the tubes two or three nuclei were seen,th evegetativ e nucleus not always beingvisible .Weight s ofphotographi c printswer e used asmeasur e ofsper m size. Sizes of sperms varied. Larger and smaller pairs of sperms wereob ­ served,possibl y corresponding with 2n andn gamete s respectively.However , thesiz eo fth esperm svarie dwit ha factortw o inGinek ex Ginek ecrosses ,

Table 28. Numbers ofdevelopin g ovuleso rseed spe r berry atdifferen t times afterpollinatio n ofcv . Ginekewit h severalpolle nparents .Dat a aremean s ofthre e ormor eberries .Pollination s weremad e on threedifferen t dates in1976 .

Date Pollenparen t 1 week 2 weeks maturity

76-06-18 IVP10 135 42 11 IVP35 203 129 6 IVP48 152 78 6

76-07-20 Gineke 360 279 218 G254 77 8 3 IVP48 177 35 9

76-08-21 IVP10 _ 64 50 IVP24 - 6 2 IVP48 - 9 3

-n o observation

76 Table 29. Correlation coefficients (r) betweend/100 b and hy/100b ofpopulation s ofpollinator s crossed with cv.Rados a (1971)o rcv .Ginek e (1973, 1974).

Population n r

1971a IVP(32 x 48) 21 -0.41 IVP(48x 35) 13 -0.13 total 34 -0.29

1973b IVP(48x 35) 24 -0.39 IVP(32 x 48) 47 -0.67 IVP(24x 10) 18 -0.66 total 89 -0.57

1974C IVP1 selfed 13 -0.15 IVP( 1x 10) 6 -0.58 IVP(48 x 1) 5 0.41 IVP(66 x 6) 8 -0.04 IVP(48x G609) 10 -0.00 total 42 -0.42 n.numbe r ofgenotype s in apopulatio n a-origina l data inVa n Breukelen (1972) b- original data inFig . 1 e original data inTable s 6an d7

.v. ' +~A from Dollen diameter observa- «here only one size of sperms was expected frompo l ^ ^ ^^ tions.Th e range in IVP10 spermswa swide rtha ni n nuclei with Thusi twa s notpossibl e by thismetho d toidentif yrestitu a

Ce t^-tubes outo fa groU po f54 2 ^^-^rll these single sperms were not big andprobabl yno t ^^ ^^ ^ ^_ sperms might have been lost during ^^ unambiguously one sperm, h wed «ations only one out of 365 intact tubes * ° J^J tubes might have m eight others only single sperms were seen ^ ^ ^ ^ ^ beenbroken , or the other sperm was leftm theo ^ ^ doubtful cases which was excluded from the preparation, mclu i* ^ having tne onl y 5.1% of the tubes contained one sperm, the m si ze ofa restitution nucleus. elative position of the sperms Inmos t of the tubes screened also the re ^^ together or w*s noted. Thirty pairs (4.8%) from 615 tubeswer e % * ï rrt- ^''•'••••--'•'•'•'•"•'••••^'••;^£?

/-i/W •* ** •'•'• *«*-* ;-••£t £*• '

# f * . >

u b e s pared in styiar tissue with tw The riLniCTe; t h e sr ° — *«*• Cl S6 (botte, respectively " "^ ^ ' ^ ^^ "* ^ ° 78 touching (Fig.4) .Percentage s ofsperm sver yclos etogethe rvarie d forth e different S. phureja pollinators. IVP48 had the highestpercentag e (8.6%), followed by IVP35 (5.5%). IVP1, 10 and 24 had about 3%. Most of these spermswer esmall .

7.3.5 Certation

Therelativ e rate ofpolle n tube growthwa sdetermine d forpair so fpol ­ lenparents . In 4x styles two batches of Ginekepolle n were compared with eachothe r to check the accuracy ofth emethod .N ogrowt hrat edifference s were found between the pollen tube bundles in three replications.Ginek e pollen was growing at the same rate aspolle n from Radosa and Multa, and fastertha n pollen from the dihaploids G254 and G609. In2 x styles Gineke proved again to be faster than thedihaploi dG25 4i ntw oreplications .Ir ­ respective of the ploidy level of the style, 2xpolle nwa s faster thanx Polleni n S. tuberosum crosses. A more complicated pattern arose when Gineke pollen was compared with thato f s. phureja. Pollenbundle s ofsevera l S. phureja pollinatorsshowe d athinning , so that some tubes were aheado fth eother s (Fig.3) .Pollina ­ tors producing much 2n (=2x) pollen, e.g. IVP10 and 32, had'man y pollen tubesgrowin g as fast as 2x pollen fromGineke .Other slik eIVP4 8ha donl y fewtube s reaching as far as Gineke,whil e themajorit y was shorter. Ina comparisono f IVP10 and 48,bot hbundle swer ethinning ,bu tno tt oth esam e «tent. Many tubes of IVP10 were as long as the few forerunners ofIVP48 . S. tuberosum dihaploid G254 was compared with IVP32,3 5an d48 .G25 4tube s "ereslowe r than the fastest tubes of S. phureja polleni nal lcases . Tubes of 2n (=2x)polle n of S. phureja did grow faster than those ofn Pollenan d as fast as those of2 xpolle n from S. tuberosum.

7 3 - -6 Delayed pollinat ion

The age of the flower at the time of pollination influenced d/lOOb, hy/lOOb and berry set (Table 30).Th e stigma was not receptivebefor e the bud started to colour. Highest dihaploid frequencies were obtained wit *»** open flowers. Closing old flowers were lower in d/lOOb,hy/100 b and be «y set. Delayed pollination did notincreas edihaploi dfrequencies .

7-4 DISCUSSION

'4-l Relationship between 2n pollen and hybrids

Percentages of 2n pollen in S. phureja clones inTabl e 27 (°"30%|' "*" Si^*r or slightly lower than those foundb yothe rauthor s nd 1963)_ observed 7%l arge sperms afterpolle n tubemitosis .Houga se tal . (19 79 Table 30. Berry set,d/lOO b andhy/100 b ofGinek e flowers of different ages, allpollinate d onth e sameda ywit hpolle n of IVP35.Pollination s weremad e in agrowt hchamber .

Age of flower Number of flowers %berr y set d/100b hy/100b greenbu d 10 0 budwit hcolou r 15 13 75 250 openingbu d 20 40 200 1180 young open flower 25 96 209 954 paleolde r flower 25 96 172 733 closing old flower 16 35 117 786

ported 5-15% 2n pollen in their bestpollinator .The y considered thisper ­ centage asexceptionall y high.Höglun d (1970)calculate d 50%2 npolle n from studies ofPM Cmeiosi si n S. phureja. Mendiburu (1971)foun d 15-25%2 npol ­ len in S. phureja x s. tuberosum diploids.Mo k & Peloquin (1975) found up to58 %i nsimila rmaterial . Frequencies of 2n pollen, as calculated from pollen size, fitted well with hy/100b (Table 27).Hybri d seeds consisted oftriploi d and tetraploid hybrids. As triploid frequencies would not be higher than 15 per 100 berries (Section 2.5.2.1), they could be neglected. Therefore the fre­ quencies of tetraploid hybrids, formed by fusion of an egg cell with a2 n sperm, could be used as an indication ofth e level of2 npolle n production ofa pollinator . The considerablevariatio n in frequencies of2 npolle nbetwee n thetime s of sampling was not unexpected, as frequencies ofhybri d seeds alsovarie d during the season (Table 3).Ramann a (1979)observe d alsovariatio n during a season in frequencies of cytological processes, which led to 2n pollen formation.Jacobse n (1976)foun d evenvariatio nbetwee n differentplant so f thesam eclon eo nth esam eday .

7.4.2 Ovule lethality

VonWangenhei m (1957,1961 )an dBende r (1963)showe d thatovule s with3 x embryos and 5x endosperm were lethal.Triploid s degenerated between 5 and 15day s after fertilization.Th ereductio n ofovule s in this study occurred in the same period. Only aver y low percentage of the ovules developedt o maturity in4 xx 2 xcrosses .Th ematur e seedswer emostl y tetraploids,wit h some dihaploids. Only tetraploids were expected in the Gineke x Gineke cross, hence the high seed set. IVP10 yielded many tetraploids as well, whxch could be expected from the highnumber s of2 ngamete s inthi sclone .

80 Where high numbers oftetraploid s were formed, relatively high numberso f ovuleswer e found intw owee kol dberries . These results implied that thelo wnumber s ofseed s insevera l4 xx 2 x crosses were notth e result of failure offertilization ,bu to fhig hle ­ thalityo f especially the ovules fertilized with npollen . They wereth e majorityo fth eovule s inmos tcrosses .

7.4.3 The role of 2n pollen in dihaploid induction

Hougase tal . (1964)suggeste d apositiv e influenceo f2 ngamete so ndi ­ haploid induction. Thebes t dihaploid inducer inthei r material produced thehighes t number of seeds (8pe rberry) . Overall seed setwa slow . The highnumbe ro fseed s without embryo gavethe mreason sno tt oexclud ea rol e forth en pollen. Frandsen (1967) worked withtw opollinators .Agai nth e highest dihaploid inducer had the highest seed set.Buketov a & Yashina (1973)di dno tfin d acorrelatio ni nnin eclones . Larger numbers ofpollinators , some with highsee dset ,wer e studiedb y Hermsen& Verdeniu s (1973). Calculations from theirdat ao n2 8pollinator s gavea correlation coefficient between d/100b andhy/lOO b of-0.21 . This lowan d negative correlation together with thenegativ e correlationsfoun d inthi s study (Table 29)thro w some lighto nth eorigi no fdihaploids . dihaploids originated from 2npollen , apositiv e correlation wouldb eex ­ pected:hig h frequencies of2 npolle n giving riset ohig hnumber s ^J^ Ploidhybrid s anda certain fraction inducing dihaploids,i ft es e sperm would not fertilize theeg gcell .Hig h numbers ofhybri d -^s d „^ ,,„ Therefore dihaploids were n°tg owit h higher dihaploid numbers generally.Tnereiui .

Probablyno tinduce db y2 npollen . diha_ Rowe (1974) also decided that 2ngamete swer eno trespons i

Ploid induction. Hisargumen t was, however, that2 npolle n «•"**^ & ducepedogam y when one sperm was lost. According tohi m e^ ^^ spermwoul d notb eunde r genetic control,wherea sth eeffe c n *toro ndihaploi d frequency was. in dihaploid in­

here areothe r arguments against arol e of2 np ^ a reason duction. If dihaploids were induced by2 npolle n there sh ^^.^ whyon esper m would notfunction . Inmaiz e thiswa sex p ^ ^ ^

Premature division ofth eeg gcel l (Chase, 1969), ^J increased the C ^1<1no tb efertilize d anymore.Thi swa ydelaye d ^^ ^ delayed haPloid frequency. Itwa sconfirme d (Table 30)tha ti n . ^ _^_ Pollinationdi dno tincreas e thedihaploi d frequency,a s ^_^ ^ re e ^Y found, several studies have showntha tprematur e ^ ^ (ig4Q) ±n ^ikely to occur in Solanum. Embryological studies o ^ ^ ^ S- tuberosum revealed that thefirs t division oftn e yy ^^ ^_ tWee * four and five days after pollination,tw oday sa f & ^ S^m division. Dnyansagar & Cooper (I960) found a 81 S. phureja. Functioningo fonl yon esper m isthu sno ta slikel yi n Solanum asi nmaize .Thi sstrengthen s thelikelihoo d that2 nsperm s dono tinduc e dihaploids. Thenegativ eslop e (-0.01—0.02)o fth eregressio no fd/100 bo nhy/100 b (Table 29)gav e an indication ofth eorde ro fmagnitud e ofth enumbe ro f ovules(50-100 )tha tneed st ob efertilize db yn polle ni norde rt oproduc e onedihaploid ,i fa naverag epollinato ri sinvolved . If n pollen induced dihaploids, it has to be explained, how bothn sperms fertilized the central nucleus to provide 6x endosperm. Studyo f pollentub emitosi scoul dgiv eth einformatio nneeded .

7.4.4 Pollen tube mitosis

Ifn polle n contributes 2x chromosomes toth e6 xendosper m ofa diha ­ ploid, these can come from two separate sperms or from two fused sperms frompolle ntub emitosi srestitution .Chance s thatdoubl efertilizatio no f thecentra lnucleu stake splac eb ytw ocompletel yseparate dsperm sar ever y low. Singlesperm sfro mpolle ntub emitosi srestitutio ni nn polle nmigh tin ­ ducedihaploids ,bu tmos tsingl esperm sfoun di npolle ntube swer epossibl y frombroke ntube san dthe ywer eofte nno tbig .Thei rfrequenc ywa sto olo w to explainth edihaploi d numbers found.Th emos treliabl e counto fsingl e spermsi npolle ntube sgav e1 %singl enucle ii n20 0tube so fIVP48 ,a goo d pollinator.A 1 %frequenc ymean seigh tpotentia ldihaploid si na berr ywit h 800 ovules. Assuming 50%genera l lethality (cf. lethality in 4x ovules. Table28) , 50%failur eo fth eunfertilize deg gcel lt odivid ean dn oletha l genes m the embryo,whic h were all generousassumptions ,onl y twodiha ­ ploids per berry would result.Thi s frequency islowe rtha nth eobserve d frequency asIVP4 8induce dmor etha nfou rdihaploid spe rberr yi n1974 . Moredihaploid scoul db einduce db ypolle ntube swit hsperm sver yclos e ZU Tc" TeCtSd bY a bridge 3S d6SCribed b* Bend- (»63)an dMonte ­ (196?) Cl S6 nt b7h, - ^ ^^ ^ ° ^^ »***actio na s onean dbot h fertilizeth ecentra lnucleus .The ywer efoun di na frequenc y

1? P tential samea s ", "* ^ ^^ ° dihaploids following the sameassumption sa sabove .A sth eassumption swer egenerous ,th efou rdiha ­ ploidspe rberr yactuall y foundcoul db eexplaine db ythi styp eo fsperms . rlonaebTvenhy;f,ClOSe '**"" ^ ^ *» *>°d *°^°* ™* wasals o reasonablyhigh ,bu tm goodIVPI Oratne r lowa scompare dwit hba dIVP24 . loo atlr "° f6 X end°SPerm f°r dih-Ploids one should notonl y reStitUtl0n WhlCh SUlts tk eint oa ! ' - - • single sperm, but also re itutLn I ,T "^ ^^^ OCCU"ln* « 'functional, restitution,wher eth etw osperm sar eno tcompletel ™°™^yseparate d Montezuma-de-carvalho (1967)describ e , ,. . d- tubemitosi si n s pnureia t deSCribed «aphase disturbances inpolle n S. phureja. He suggested thatlaggard smigh tlea dt ochro - 82 mosomal bridges. Themitosi s wasfoun d tob esensitiv e toexterna l influ­ encesa sirradatio n (Bender, 1963;Bukai , 1973),N 20 (Montezuma-de-Carvalho, 1967)an d colchicine (Montelongo-Escobedo & Rowe, 1969). Temperatureef ­ fectso ndihaploi d frequencies foundb yWöhrman n (1964)migh tals owor kvi a pollen tube mitosis. On the other hand oneo fth esperm s might have lost itsfunctionalit y by these treatments sotha t single2 nsperm s could have induceddihaploids . Irregularities inpolle n tube mitosis dono tnecessaril y mean thatth e sameirregularitie s have tooccu r inothe r typeso fmitosi si nth eparent , asth epolle n tube mitosis hasspecia l spatial conditions.I nadditio nt o mechanicalfactors , also genetic factorswil l control irregularitiesi nmi ­ tosis.Accordin gt oth econclusio n from Section3.4. 3severa l additiveloc i willb einvolved . Additivity ofgen e effectsmean stha tmor egene s for,o r more structural changes inchromosome s leading to,chromosom e bridgesbe ­ tweenth esperm s will increase thechance s ofdoubl e fertilization ofth e centralnucleus ,an dthu so fdihaploi d induction. Sucha syste mwil l follow Patternso fquantitativ e inheritance andfunctio ngametophytically .

7.4.5 Certation and dihaploid frequencies

The study ofcertatio n showed that 2xpolle ngre wfaste rtha nx pollen , bothi n4 xan di n2 x styles.Thi si sdifferen t fromth esuggestio no fHog ­ end (1970) that 2xovule s (intetraploi d plants)migh tb epreferentiall y fertilizedb y2 xpollen . Ifi twa sa matte ro fpreference ,th ex eg gcell s should also favour x pollen (in diploids), which wasno tth ecase . er othersuggestio n ofdifference s ingrowt h ratewa sconfirmed .Howeve r the tetraploid frequency she found could better be explained bytrrplo i thality.Potat o pollen behaves differently from beetpolle nwher ex polle n is faster than 2x pollen in a4 xstyl e (Matsumura, 1958). Differences in 9r°wth rate areno ta general phenomenon. Skiebe (1966)an dEse n eta ^78) found no indication fordifference s ingrowt hrat ebetwee n2 xan d PoUeni n Primula malacoides and Citrus respectively. ^_^ The faster growth of2 xpolle n found in Solanum implied a ^ iono fovule s fertilized by2 xpolle nmus tb ehighe rtha nth e 2X PoUen, if there is an overdose ofpollen . With room for300J ^ ^ 9rai an d ns ona tetraploid stigma (Janssen &Hermsen , 1976) If all

•»-ar y (Section 4.3.1} there will be enough roomfo r« ^ ^ 2n Pollen tubes would grow faster than n pollen tubes, m ^ C0*H occupy three to four times their shareo fth eovule su pt o f ° 100%. However; a comparison of the percentage of2 ngamete s

»*** ofhybrid s inTa L 27 shows that thehybri d ^^J^^. hat *uc hhighe r thanth e2 npolle n frequencies.^^^"„e. , even en tubes „ere faster than alln pollen tubes. In4 x as ^ Wlth muc h2 n pollen, many ovules were leftfo rn polle nt o seed setwa s generally lowertha ni n4 xx 4 xcrosses . Thus it can be concluded that the mechanism of certation reduced diha- ploid frequencies of pollinators with much 2n pollen, but not to a large extent.

7.4. 6 The mechanism of dihaploid induction

What happens after pollination of 4x S. tuberosum with 2x S. phureja? S. phureja pollen consists of amixtur e of 2n and n pollen. Somepollina ­ torsar egenotypicall y determined toproduc eman y2 npolle n grains,bu tth e actual frequency varies much. The 2n pollenha s ingenera l ahighe rgrowt h rate than n pollen and thus occupies more than its share of the ovules, leading to 4x hybrids.Th e n pollen occupies the remainder of the ovules. Mosto fthes eovule swil l contain 5xendosper m andb e lethal.Ver y fewtri - ploid hybrids survive. A certain fraction of the n pollen - influenced by genotype and environment - forms asingl e restitution sperm or abridg ebe ­ tweenth e two sperms after thepolle n tubemitosis .Thi s can lead todoubl e fertilization ofth ecentra lnucleu s and result in6 xendosperm . Ifth eun ­ fertilized egg cell then divides and does not carry lethal genes, adiha ­ ploid embryo in viable endosperm will result. The frequency of the diha- ploids will depend on - first the tendency of n pollen to form abnormal sperms, together with the tendency of unfertilized egg cells to divide, - secondly the number of ovules fertilized by 2n pollen. The more 2npol ­ len, the smaller the chances of npolle n to fertilize ovules.However ,th e genetic constitution of the n pollen seems to be more important than the frequency of 2n pollen, as among pollinators producing many hybrids some also induce highnumber s ofdihaploids .

84 8 Monoploids

8.1 INTRODUCTION

The first group of monoploids from S. tuberosum obtained by pseudogamy was induced by S. phureja pollinators with high d.i.a. (Van Breukelen et al-, 1975). This led to the question whether it was a general phenomenon that good dihaploid inducers also induce monoploids. A positive answer will have implications for a hypothesis on the mechanism of monoploid formation in Solanum. Solanum dihaploids, originating from a 4x x 2x cross, result from the 'Solanum mechanism', in which both reduced sperms fertilized the central nucleus. Would S. tuberosum monoploids from 2x x 2x crosses follow this mechanism or the 'maize mechanism', in which one sperm fertilizes the central nucleus and the other is lost (Hermsen, 1971)? The genotypes of both the seed parent and the pollinator influence diha- Ploid frequencies in S. tuberosum. Both parents influence monoploid rates in maize (chase, 1949; Coe & Sarkar, 1964; Sarkar * Coe, 1966). Nothing is known, however, about the relative importance of the female and male type in monoploid pseudogamy in Solanum. narf,P ^„ Kt, ^arrvina out a large An attempt to answer these questions was made by carrying ^er of s. tuberosum di(ha)ploid x 5. phureja crosses, mvo v ng sever^ Breukel ^notypes. Most of the results were presented in Van n <"77). More details and a further analysis of the data are presented this chapter. verrucoSum monoploids could Irikura & Sakaguchi (1972) claimed that S. verru 1 °nly be made via anther culture. In order to test the *™'* ™* Q£ ^

Won 0f monoploide in S. verrucosum through pseudogamy, some p secies were pollinated by S. phureja. chromosome doubling. This Most of the applications of monoploids require cnr identifi­ es done and the resulting diploid S-allele homozygotes cation are reported.

8.2 MATERIAL AND METHODS es IVP6, 10, 35< 48 and Monoploids were induced with S. phureja genotyp ^'^ (48 Se 10 * 35)4, all homozygous for embryo spot (see ^ ^og and G254 are di- di(ha)pl01 Most seed parents were S. tuberosum d*haploid from cv. Gineke a Ploids from cv. Gineke. B16 is possibly also a i tetrapioid clone (Her^en, pers. comm.), otherwise it derives from 85 B4495, developed byDr .W .Black . All three havea hig h male and female fertility. G609 hasn oletha l genes andi sself-incompatibl e (genotype S1S2). G254i s heterozygous forthre e lethal genes andi sself-compatibl e (genotype S1S3). B16contain s twosub-letha l genes andi sself-compatibl e (Hermsene tal. , 1978a).Th eself-compatibilit y ofG25 4an dB1 6i scause db y an S-allele bearing translocation, which isletha l inhomozygou s condition (Hermsen, 1978b).O fth eF lpopulation s G609x G254 , G254x G60 9an dG60 9x B16, which segregated forself-(incompatibility , mostly self-compatible Plants were usedi nthi s study. Oneself-compatibl e clone,GB37 , fromG25 4 x B16,wa s also included (Olsder& Hermsen , 1976). other dihaploidsuse d derive from fivedifferen tcultivars .Dihaploi dx dihaploi d derivativesan d onedihaploi dwer ekindl yprovide db yIr .N .va nSuchtele nan dDr .B .Mari s ofth eFoundatio nfo rAgricultura l PlantBreeding ,Wageningen . Non s tuberosum diploid seedparent swer e fromth ewil d diploid species 390«TldlSSeCtUm and 3' verrucosus TheS. multiäissectun, seedparen t(WA C 3908)wa scollecte di n197 4i nPeru .Thre epopulation s ofth eself-compat - seÎL VerrUC°SUm Were USed: A3 «"««1, «AC accessions, andCod e 41.A 3 S d fth e S6lf Pr geny fa nF 1 lant PI 19517Tr, ? ° ° ° * ** » «e 2247( 3x)x sisted , „ M°St PlantS WSre eUPl0ld (2n=24>- The WAC Populationcon - 3337 and^ « PlantS fr°m eaCh °fth e WAC —ions 3330, 3332, 3335, COde C nSiSted f S6lf Pr f 2x seed ^, 11 " ° ° ^ °^ ° *Parthenogenes e tween * 1970)' ^ 3dditi°n hybridS fr°m reciprocal crossesbe ­ tweenSS . tuberosum dihaploids and s verrucosum were used_

greenLuPseantFrere '"^ "^ ^^ r°°tSt°Ck and «rowni na controlle d rS fSee d ParSntS WGre poTeno fS r ° —^ted andpollinate dwit h See" were t ^^ h™°Z™°™ ^ embryo spot (Section 2.3.1). e d fr m thSberrie SSn dCarefUl1 tS ed w7thou t T ° - «— *« ^° -

wereno t J»Z, „! * "^ °fthSS e Seeds were s°™al l thatthe y Ppe ri npXf d t ^^ B°th "™»> <*-ed swer epu to nwe tfilte r ::::r^ïir ^nate and thosewhic h ge™inated — ™- nated were putin t d ^x TlIII ^ ^ "^ ^ ^ "* V^' Plantlets,showin ga coloure d^ J^ T^ ^ °f^ *1""t"^ *" cotyledonsw P™ r * nypocotylo ranoda l banda tth ebas e ofth e ~L :::;:::»™r;i;:r:7 v*rgo—«*—» -* • h Ch nC a 1 »ut , coloured hyPoc„tyl PL„t' ' " ' •"""'"' "° "™ "•" "^ —«« Z^LZ\l°t: ä °h Thto f• bout5- «th" •<- °" 9U Cells of the (Frandsen,1968 ) Pian+ .^ stomata was counted fch were kept The chm , * **** ^^ ° l^oPlaSts lower than12. 0 (VanBreukele n T^ ^T ^ m°n°Pl°idS *' ""»* *** ™' $> lose a single monoploid The«! / ^ mar9in "" USed ±n 0rder n0t t0 small potsi n order i-„ r •,• Plantlets were transplanted into ordert ofacilitat e thecollectio no froo t tips forchromo - 86 A.%, ^*5te *«***,*

1 Ä ^'•J fe-^ * u\.:v*?&

i

i * fig. 5. Stained chloroplasts in guard cells of strata of 5 tuftero«- Plants with different chromosome numbers: diploid (left) and monoploid (right)

some counting at a later stage.Plant swit h 12chromosome s m theroo ttip s wer«ere recorded as monoploids. A few pl.ntl.ts did not survive until this s tage. The somatic chr0mosome number was also counted inplant swit h roup than12. 0chloroplasts .N omonoploid s were found m this9 - iosis Mostmonoploid s produced flowerbuds ,whic hdroppe dprematur ey .

«asstudie dwhe npossibl e to confirm rootti pcount« . 8_hydroxyguino_ For chromosome counting root tips were pretrea ^ ^^ ^ ^ •Une for about 24 h. Fixations were made m J-i (1968).

-ained in propionic acid haematoxylin according toHenderson^L u (1^ F°rmeioti c studies young flower buds were fixed ^ ^ °f ethanol and propionic acid saturated with iron acera . Quashed in 1% acetocarmine (VanBreukele ne tal. , 1975 . ^ ^ ^ ^^ F°r chromosome doubling Dionne's method was;us e (o ^.^ ficatl angton, 1974) (see Section 4.2)wit h some "^ chloroplast number h^ beengrafte d and others lefto nthei row nroot s The ^^ ^ ^ an<*th e chromosome number were determined afterp treated plants ^ration. The method was later modified New shoo. ^ ^ ^ ^e firstchecke d for their chloroplastnumber .Sho o ^ Ch l°roPlasts were left on the plant. The plant was then prun 87 the growth of the doubled shoot. Later several cuttings were made fromth e doubled shoot for propagation. This reduced the risk of losing a doubled plant. Inaddition , flowers couldb e obtained in ashorte rtime . Frequencies of monoploids were considered as the outcome of Poisson variables and were compared using a binomial test.On e frequency was taken as test statistic and the test performed conditionally on the totalnumbe r ofmonoploids . For observation of pollen growth, styles were cut from the ovary 24h afterpollinatio n andprepare d forU V fluorescence (Section 7.2).Individu ­ alstyle swer epu to n aslid e inglycerin e and gently squashed with acove r glass. Pollen tubes could then be studied under a UV microscope with a large field.

8.3 RESULTS

Most monoploids surviving the seedling stage were well growing plants with lightgree nleave s andnarro w leaflets (Fig.6) .The y reached aheigh t

^—— •»-•* .> . . a Fig. 6. Leaves of S. tuberosum. Topro w (from left to right): Gineke(4x) , G609 (2x)an dmonoploid sM 3 andM 7 (x).Lowe r row:double d G609 and doubled monoploidsmonoploids .

88 • i-4-. i-inpice G609 and mo- Fig. 7. Tubers from S. tuberosum. From left to nght: Gxneke, noploids M40, M55, M109, M4 and M39.

, viaht): Gineke, G&09 left to riynl-'' Fi 7- 8. Flowers of S. tuberosum. Top row (fr , M39. and monoploid M39. Lower row: doubled wus

89 of about 80 cm. Most produced tuberswit h little orn o dormancy or stolons which developed into new shoots immediately (Fig. 7).Abou t 60% of the genotypes produced flower buds, but only two actually flowered (Fig.8) . Flowerswer e completely sterile. Monoploid frequencies can be expressed as monoploids per 1000 seeds (m/1000s) or as monoploids per 100 berries (m/100b). The monoploid fre­ quency of G609 was in general lower in 1974 than in 1975,whe n expressed as m/1000s (Table 31).Th e seed set, however, was higher in 1974. There seemed tob e anegativ ecorrelatio nbetwee nm/1000 s and seed set.Th emono ­ ploid frequencyo fG60 9wa smor econstan tove rth eyears ,whe nexpresse da s m/100b.Therefor em/100 bwa suse d tocompar e S. tuberosum seedparents .

8.3.1 Pollinator influence

The four S. phureja genotypes, which were used extensively for mono­ ploid induction, had also been used as pollinators in dihaploid induction in the same years.;Th e numbers and frequencies of haploids obtained from seed parents G609 and Gineke are given in Table 31. IVP35 and 48wer eth e

Table 31. Numbers and frequencies ofmonoploid s obtained from S. tuberosum dihaploid G609i n197 4an d1975 ,compare dwit h dihaploid frequencies ofcv . Ginekewit hth esam epollinator s inth e sameyears .

Pollinator Numbero f Frequency of Number of Gineke monoploids seedspe r dihaploids berries monoploids berry per 100 per 100pe r 1000 berries berries seeds

1974 IVP6 12 2 17 0.38 435 163 IVP10 23 6 26 0.76 342 209 IVP35 135 23 17 0.46 374 342 IVP48 108 18 17 0.42 393 445

Total andmean s 278 49 18 0.46 382 290

IVP6 21 2 10 0.42 224 103 IVP10 12 2 17 0.90 185 235 IVP35 34 7 21 0.80 257 297 IVP48 26 3 12 0.52 221 406

Totalan dmean s 93 14 15 0.65 230 260

90 bestdihaploi d inducers and IVP10 was also good (cf.Sectio n 3.3.1). The d.i.a.o f IVP6wa s only 41%an d24 %o fth emea no fIVP3 5an d4 8i n197 4an d 1975respectively . In monoploid induction IVP6wa sagai nth elowes tpolli ­ nator but here it induced respectively 99%an d 62%o f themea n monoploid frequency of IVP35 and 48 expressed as m/100b in the two years. On the otherhand ,th e firstplac e of IVP48 ind/100 b didno tg otogethe rwit hth e highest monoploid frequency. Another pollinator, IVP(48 * 35)4 (cf.Tabl e 8), was used on a'small scale.Thi splan twit hhig hd.i.a .yielde d 1mono ­ ploid in 15 berries, which was not anexcellen tperformance ,compare dwit h otherpollinators .

8,3.2 Seed parent influence

Three dihaploids and three Fl populations derived fromthe m-r epolli ­ natedwit h IVP35 in one season. Details ofth eberrie san dmonopl od sp r , , bestproduce ro fmonoploi d ducedar egive n inTabl e 32.G60 9wa sb y farth ebes tf r thP P +- at all Neither difl DID, plants, while G254 produced no monoploid plants ax. • , • <-v,=i-ra nb e drawn istha t BID butwit h only 7 berries the only conclusion that can +•„r to. FlDlant sproduce don eo rmor e isno tmuc h better than G609. Tenou to f3 4F Ipia n y ^ ^ recip- P pUlatl n monoploids,unevenl y distributed over thethre e ° ° ^urnber of berries rocalprogenie s of G609 and G254 produced aboutth esam en u ^^ four and seeds, but the progeny which had G609 as mother,pr o timesa sman y monoploids. seed parents, includ- Thecomplet e results ofmonoploi d induction m different Solanum ing the negative results with several seed parents rom^ ^^^ already secies, arepresente d inTabl e 33.Apar t from thepos ii

obtained fromthre edi - and frequencies ofmonoploid s Table 32. Numbers ana IICIIIICI^^-- -- - IVP35- haploidsan d their Flprogenie s afterpollinatio nw i . Frequency of Se if ed parent Number of Number c monoploi ds genotypes monoploids berries per 100 per 1000 total producing berries seeds monoploids

17.0 0.46 135 23 G609 1 1 0.0 0.00 135 0 G254 1 0 0.0 0.00 B16 7 0 1 0 4.5 0.16 247 11 G609 x G254 13 7 1.2 0.05 3 G254 x G609 14 2 250 0.9 0.03 1 G609 x B16 7 1 106 91 Table 33. Summaryo finduction so fmonoploid s insevera l seedparents , using S. phureja clonesa spollinators .

Seedparen t Numbero f

genotypes monoploidsberrie s seeds seedspe r berry G609 1(1) 67 411 G254 13680933 3 1 0 B16 333 65541 197 1 0 7 Progenieso fG609 ,G25 4 339 48 andB1 6 34(10) 15 other dihaploids 603 165631 275 21 0 dihaploidx dihaploi d 127 11008 87 6 0 111 S. multidissectum 1127410 2 1 0 S. tuberosum x s. verrucosum 244 10103 41 12 0 S. verrucosus! x s. tuberosum 135 11041 82 6 0 S. verrucosum A3,selfe d 63 5145 82 9(1) 1 S. verrucosum WAC accessions 259 7012 27 17(1) 2 S. verrucosum code4 1 146 1742 12 GB37 1 0 89 1559 18 1 0 47 4677 100 Total 111(13) 85 2575 431881 L^!!^!!l^r^"pro

mentioned (Table 31anr f i->\ <-i. of* verrucosum. Three^opl !^w^ " TT ^ *" ^^ Otherplant san dpopulation spollin g with7\ "^^ ^^^ withoutembry ospo t Thes. ^llnated Wlth *• Phureja yielded afe wseed s onopi ias 1 p..«, m^j:t ;;- r.c::.r:;:" ° - ° •"•- - The .bility to „rod,,,-. KoancM on«o r.or emonoploids . over. n ,Peo/„ J^^T^" — "« * » «•««*.«- —7 S.-Î .3 Doubled monoploids

About10 0monoploi dplant swer e treated with ,„ •• per plant). Twenty-one doubled 'd ' 7 colchicine (10-20meristem s Ual la different genotypes RM,, * *" P "ts were obtained from eight J^ °* results werp hptfû*. •J_ , plants on their ownroo t M grafted plants than with their monoploid counterparts " dOUbled plants lowered (Fig.8) ,unlik e crossed with G254 Th„ ,. ' Produced berries with 80-100 seeds when quality ofth epolle no fth edouble d monoploidswa s 92 bad, stainability being not higher than 1%. The homozygous diploids, obtained from monoploids from G609, proved to be suitable material to test S-allele genotypes. Crosses were made with different tester genotypes and pollen growth in the style was observed with a UV microscope. Normal functioning of the style was tested with pollen of G609 and G254. Results are given in Table 34. Both S-allele genotypes ex- pected from G609 were actually found. Doubled M39 gave an incompatible re­ action with pollen of tester genotype S1S1 and must have been homozygous for SI. Doubled M4 and M36, compatible with S1S1, must have been homozygous for S2, the other allele from G609.

8.4 DISCUSSION

Numbers of seeds per berry were lower in 1975 (230s/b) than in 1974 (382 s/b). The number of monoploids per 1000 seeds seemed to be negatively cor related with the number of seeds per berry. IVP10 was the höhest of four pollinators in m/1000s in both years but also the lowest in ^^ berry. This can be explained by the high number of 2n pollen - which leads to non-viable seeds with tetraploid endosperm. Apparently number of ovules is more determinative for the monoploid nunfcer number of seeds formed.

Table 34. Compatible (+) and incompatible (-) reactions in styles of S. tuberosum plants of different ploidy levels, ob­ served on the growth of pollen of diploid testers. The S-allele genotype of the monoploid M39 and the doubled monoploids (DM) are deducted from the compatibility reactions in this table.

Tested material Tester

S1S2 S1S1 S1S3

Gineke S1S2S3S5 G609 S1S2 M39 SI + + DM39 S1S1 + + DM4 S2S2 + DM36 S2S2 + +

a- inhibition only at the base of the style

93 8.4.1 Pollinator influence

Iti sapparen tfro mthi sstud ytha tmonoploi dinducer sar euseles swith ­ outa nembry oo rsee dmarker ,whic hdifferentiate spseudogami cembryo sfro m amphimictic (and paternal) ones, without the embryo spot marker allth e 45000 0seed swhic hwer eno wscreene dfo rembry ospo twoul dhav eha dt ob e Planted andthe ywoul dhav etake n90 0m *a ta 4 c mx 5 c mspacin ga sseed ­ lings.Apar t from spacelimitations ,th ewor kinvolve d inscreenin go fs o many seedlings wouldhav ebee nenormous .Consequentl y itwa sno tpossibl e totes tth epollinator s IVP24an d3 2wit hver y lowd.i.a . forthei rmono ­ ploid educing ability,a sthe ywer eheterozygou s foron eembry ospo tlo - eus. r Therol eo fth epollinato ri sver yimportan ti ndihaploi dinduction ,bu t

lesss ol n monoploid induction.Dat aobtaine dfro mfou rpollinator si ntw o seasons seemed toindicat etha tlo wdihaploi dinduce rIVP 6wa sals olo wi n m/100b (Table 31). Onth eothe rhan dth edifference sbetwee nfrequencie so f m/100b were small,wherea sdifference sbetwee n frequencies ofd/100 bwer e

as high as 300%. To testwhethe r IVP6wa s indeed loweri nm/100 b twohy ­ potheses were formulated. The firstwas : IVP6 equals IVP35 and48 .Unde r thisassumptio nth ecritica lleve l(one-sided )wa s0.6 7i n197 4an d0.3 6i n elan H thlS lt Can be C°nClUded that it is not impossible thatIVP 6 equalledIVP3 5an d48 .Th eothe rhypothesi swas :whe nIVP 6i scompare dwit h

35 and 48,l t isa smuc hinferio r inm/l00 b asi ti s ind/100b .Unde r thisassumptio nth ecritica llevel swer e0.3 7an d0.6 2 (two-sided)fo rX97 4 tarent!™ BPeCt±VBly- ™S hypothesis was also notrefute db y thedata , statement T nUmberS °f mOn0pl°idS a" neededfo rstatisticall ysoun d e SignifiCanCe f the d PoTZtrs°\T ° i«erences between theeffect so f oftheto n:I Can bS COnClUded that' alt*ough themonoploi d frequencies nm /oo br;r s iooked simnar- *±s not ^>«^ ««* ™ „L ^ Iti s : SamS eXt6nt 3S ^ d/10°b C™^« »"* IVP35an d48 . Thi couldTl thSt eVSry S- PhUreU polli-tor triedwa ssuccessful .

andtha teverTmal rymal ifLS efertil ; Tfeplan t.migh °' ^^tb ea potentia ^ ^l monoploi ^^ dinducer —. —

8.4.2 seed parent influence

no.o„0 Ploia, IVP35 'rÜ ""J ""l "''f"» • <«» »« °»* P—" Mth 32) l£ tors( Table ,3, ,, . „ " ' "««I« <Ü<1. o„it hothe rpollina -

fourletha lgene so fr« A , 9 S are known in G609- The f afacto r1 6 2 «let I ***"* ^ "^ ° ^^ ^opioidswit h expectednumbe ro fmonoploid sfro mth e33 3berrie so fG25 4 QA -i fi tha dth esam e frequency asG60 9 (49/278)- woul dhav ebee n58. 7an d thiswoul d have been reduced to3. 7monoploid* .Th eprobabilit yo ffindin g Opioids in33 3berrie s if3. 7ar eexpecte di sonl y2.9% .Thi sprobata - lityi ss olo wtha t itca nb eassume d thatth edineren -, +.>,=+rtp,-e differenc ebetwee nG60 9an a G254wa sno tbase d onth eknow n genes forlethalit r~ry onil.t-hsiity vonlv . Genes formono - ploidproductio n mightb epresen ti nG60 9an dlackin gm G254 Population (G609x G254 )produce d fourtime sa sman ymonoploid*as _(025 4 xG609 ) from the same number ofberries . Under theassumptio n_** * ^J populationsproduce d monoploidswit hth esam e frequency,th ecr i forth eactua l value of (G254x G609)wa s0.0 3 (one-sxded .* » '«^ G2 eludetha t (G254x G609 )wa sindee d lower than (G609x **> ' ^ level couldb efoun d between (G254 x G609)an d(G60 9x B16). Te for(G60 9x B16 ,compare d with (G609x G254 )wa s0.1 7^ ^ ^ tingtha tth etw opopulation s possibly differedi nmonoploi dproductio n • re.M and G254 produced monoploids in G2 5 The two reciprocal progenies of G609 and j ^ ^ average half of lower frequencies than G609 itself. Flplant s carr ^ .^

thethre e lethal genes of G254. The"""«'^J.^ . Togetherthi s Pliedtha t most ofthe m also carried aleui a ^ ^ ^ lethal genes »adetha tth eplant s from (G609x G254 )carried Io n'^^ ^ production andthos e from (G254 x G609) 2.3.Assumin g the ^ ^ ^^ ^^

perberr ya stha t from G609 with IVP35an da reductio n ^^ QQ and 8>6 theberrie s from thetw opopulation s were expece ^ ioidg actuaiiypro - m monoploids respectively, instead ofth e1 1a n not for its

âuced.Th enumber s fitted reasonably wen i^ v ^ ^G6Qg ) yielded3

reciprocal.Th eprobabilit y thatth e25 0berrie s o^ ^ 4g/278 (asG609 )

monoploids or less is3.5 %i fth eexpecte d frequenc .ficantsh ort- and2. 3letha l genes aretake n into account.Tne r a geo fmonoploid s in(G25 4x G609). ^ G254 differed and

The two reciprocal Fl populations from G6^ ^^.^ such a situa- followedth efrequenc y ofth emothe r inmonoplo i pro^^^ ^ concerned.

tionusuall y points toa cytoplasmic influenceo n e ^ .nthi s case

G609cytoplas m would be superior totha to fG2 5. cultivar. Even while

istha t G609 and G254 are dihaploids from thesa m^ ig78a)/ it ismor e G25 4ha sman y alleles incommo nwit hB1 6(Hermse ne a. , (HermSen, ll "«ly that B16 derived from Gineke than GiVk ^^ ^ ^ (G254x e P «. comm.). This leaves thedifferenc e between ved tnat wasno t G609) unexplained, unless a cell organelle were

transmittedb ycv .Ginek e toal lit sdihaploids . ploidB.Th e lownum - prodU e Only a small part ofth esee d parents ° ar"nts didno twarran tth e

berso f berries or seeds from many ofth esee dP"^ ^ a good measuret o expectation ofa single monoploid from them. M/l ^ comparing seed

<=ompare pollinators used with G609,bu ti swa sn o g^ ^ ^ faerry than

Parents, s. verrucosum produced a lower number o ^ verrucosufflpopu - S - tuberosum di(ha)ploidS. Inm/100 b thetw osucces s 95 lations were lower than the three S. tuberosum Flpopulation s (0.7vs .2. 5 m/100b), buti nm/1000 s they were higher (0.3 vs. 0.1 m/1000s). Together withth e lownumbe r ofseed s in S. verrucosum. thismean s thatth enumbe ro f ovules mightb ea nimportan t factori nth e determination ofm/100 b ina species. The number of monoploids should be expressed per ovule to compare species.A s this isinconvenien t itca nb e expressed per seed when therei s noexcessiv e lethality. The relatively high monoploid frequency inS. verrucosum mightb ere ­ lated to the self-compatibilityo fthi s species.Th e consequent inbreeding reduces thenumbe r ofletha l genes, compared with outbreeding species. About ten monoploids could beexpecte d from the genotypes without mono­ ploidso nth e basiso fth e numbero fseed s produced and assuminga mono ­ ploid frequency similar to G609. Lethality or other factors must havere ­ duced this number.A monoploi d froma S. verrucosum S S. tuberosum hybrid could survive,a sIrikur a (1975a,b)ha s shown.Th e facttha t soman ygeno ­ typesdi dno tproduc emonoploid smigh tmea n thatth e ability toproduc emo ­ noploids isrestricte d toa fe widiotypes .Th e influence of the seedparen t in monoploid production ismor e important thani ndihaploi d production, whereonl y fewclone s failed.

8. 4.3 Doubled monoploids

The male fertilityo f the doubled monoploids was unexpectedly low,con ­ sidering that most derived from G609 whichproduce d excellentpollen .Col ­ chicineha sbee nmentione d asa possibl e causeo f low fertility afterchro ­ mosome doubling (Skiebe, 1975). if that were the case,i tshoul d as well affect female fertility, which was not low. Spontaneous doubling through tissue culture (Murashige& Nakano , 1966)woul d have no adverse colchicine after-effects.Complet ehomozygosit y has alsobee nmentione d asa caus e for nloids y/S,febe' 1975)- A11 F1 PlantS fr°m CrOSses bet— doubled mono- whlchJ t T had hi9hly Stainable POllen (B°nthuis & Hermsen/unpubl.), Both7 a genetlC eXplanation °f themal e sterility. llele 9en0tyPeS (Sm and S2S2 couldbb ee™rr/"'expecte d fromG60 9derivatives . > -re found, which

8.4.4 Mechanism of monoploid formatie

linato:: (Tile^ w' d ^^ - <^^ "ies ofpol - Ploids are formedb y2^« ^ indiCati°n " *° ""*" *>'— ^ nor the absence of .^r^"*^8" " dlhaPl°idS- Neither the thepollinator s^Z/llT , * PrWBd «* ^ information from d n startl nism. ° "g Point fora hypothesi so nth e mecha- The endosperm must Dlav =»,J ust playa nimportan t rolei nmonoploi d formation, but 96 the ploidy level of endosperm with a monoploid embryo has never been deter­ mined. It probably will never be, as the frequency of ovules with monoploid embryos is extremely low. Double fertilization of the central nucleus as in dihaploid formation is a possible mechanism. This would give rise to hexa- ploid endosperm with 2n pollen and tetraploid endosperm with n pollen. Hexaploid endosperm would be vital, but it would imply that monoploid fre­ quencies of pollinators would be positively correlated with 2n pollen for­ mation. IVP48 - very low in 2n pollen - was equal in monoploid induction to IVP10 and 35, which had a high 2n pollen frequency. Tetraploid endosperm is not very vital (Von Wangenheim, 1961), but a certain percentage of the ovules with tetraploid endosperm will grow to maturity. Monoploid fre­ quencies would then be the product of the frequencies of two rare ^"^ double fertilization and tetraploid endosperm vitality. If double er i i zation would occur in 1 out of 100 ovules and the vitality of tetraploid endosperm were the same as that of pentaploid endosperm (20/100b, see j n -> /inn h This is much lower tion 2.5.2.1), the frequency would be around 0.2/100b. in -.„« -ir,

Process, or is it 'spontaneous' and due to chance. If e ^ .^^ ef_ influence of the pollinator on the monoploid frequencies^, ^ ^^ ^^ ^ fe expeC =t on the 'single fertilization' would not be ^e'lQSS of a sperm, direct genetic influence through the seed parent on ^e ^ ^^ ^^ ^ monoploid frequencies would probably be higher than 1 ou^ ^ ^ ^^ ^ f ound. A precocious division of the egg cell prevents o^ ^ division would

function in maize (Section 2.5.1), but in Solanum ^.^ ?>4#3) t0 pre- have to be advanced by more than four or five days e ^ ^^ differ­ ent fertilization of the egg cell. This would be a ra ^ _ ^^ ^ en =e. The effect of the seed parent on the monoploid q^ ^ ^.^ ,n. th at through lethal genes - probably functions difj^ ^^ ^ first en_ v°lve a stimulus on the unfertilized egg cell to divi e ^.^ ^mty ä0 101 ^Perm division. A seed parent with a high ^P ^, but increase Wou ld not increase the frequency of 'single fertile such a stiau- th* chances that 'single fertilization' leads to a mon 1Us coul d be of cytoplasmic origin. this hypothesis and Chance plays a major roie in monoploid production i ^ ^ parent the ex lain6 «use of the loss of a sperm is not P ^ absence of lethal 5en otypes Would e loit the chance process better ana 97 genes would reduce further losses of potential monoploids.A t thepollina ­ tor side only the genetic markers to facilitate screening are very impor­ tant.

98 9 Conclusion

The frequency of dihaploids from S. tuberosum is definitely influenced by the seed parent and the pollinator.Th e genetics ofth esee dparen tin ­ fluence is yet little known, but the influence probably functions in the sporophytewithou t contribution ofcytoplasmi c factors.Hig hdihaploi d pro­ duction is most likely not recessive. Contrary toth egenotyp e ofth esee d parent, the genotype ofth epollinato r canb emanipulate d toincreas ediha ­ ploid frequencies. Therefore, knowledge about genetics of the pollinator effect is more important. Several loci are involved in dihaploid inducing ability (d.i.a.), possibly five or more, and the inheritance follows the intermediate pattern. Consequently the progeny of an excellent pollinator contains a reasonable number of plants with good d.i.a.. This facilitates exchange of pollinators between institutes in the form of seeds, circum­ venting the risks oftube r shipments.Intermediat e inheritance alsoimplie s thatinbreedin g isnecessar y to attainhig hd.i.a. .Th einbreedin g will act negatively on the vigour and especially themal e fertility ofth epollina ­ tor. The environment influences the frequency ofdihaploid s aswell .Tempera ­ tures that give good growth and flowering of Solanum plants generally im­ prove the total yield of dihaploids (per plant,pe r m* greenhouse or per day work). The frequency per 100 berries doesno tappea rt ob e affectedb y a specific temperature regime as far as the seed parent is concerned. The pollinator induces more dihaploids at a constanttemperatur e of1 8° Ctha n at2 3 °c during the day and 15-18 °C at night. It is therefore goodprac ­ ticet omak epollination s fordihaploi d inductionearl y inth eseason . One ofth e objectives ofthi s studywa s todetermin ewhethe r themaximu m d.i.a . had been reached. It seems that the maximum of the IVP population has been reached in the genotype of IVP(48x 35)1 ,whic h inducesu pt o80 0 d/100b incv . Gineke.However ,dihaploi d inducingabilit y isno tlimite dt o the two S. phureja parents of the IVPpopulation .Severa lothe r S. phureja populations and also other Solanum diploids havebee nreporte d asinducin g dihaploids.Recombinatio n between thosepollinator s andth ebes t IVPpolli ­ nators would increase thedihaploi d frequencybeyon d 800d/lOOb .Thi swoul d not be easy. The search for newgenes ,incorporatio n inth e IVPpopulatio n and sib crosses forhomozygosit ywoul d takeman yyears .Notwithstandin g the fact that pollinators can be tested withon e seedparent ,selectio no fth e best pollinators will be laborious. It isdoubtfu lwhethe r sucha breedin g program is worth the effort, since the present level is sufficient for

99 practical breeding. In this study pseudogamy was used to produce haploids. How does this method compare with in vitro methods like anther and pollen culture for S. tuberosum? The two methods should be compared in costs per haploid and alsoi nth ediversit y ofth ehaploid s produced. Pseudogamy appears to have many advantages if a good pollinator is available. Investments in laboratory equipment are small and no highly trained personnel is needed. The right growing conditions have to be pro­ vided. This often means a conditioned greenhouse and good care for the plants,bu tthi s is also a- sometime s neglected -requiremen t for invitr o techniques. Also more Solanum genotypes will be suitable as seed parent than asmateria l for pollen or anther culture,becaus e female sterilityi s less frequenttha nmal e sterility.A disadvantage of invitr o techniques is thatno tal lgenotype s respond similarly toth e samemedium . In practical haploid production pseudogamy is nowadays the best method for production of dihaploids. Large numbers of dihaploids have been pro­ ducedb y this technique from adiversit y ofgenotypes ,whil e only fewdiha ­ ploidshav e so farbee nreporte d from antherculture . For production of monoploids the picture is not that clear. Pseudogamy was only successful in part of the genotypes. Anther culture has also yielded monoploids from a few genotypes only. The two techniques might be complementary as fara sdiversit y isconcerned .N o genotypeha sbee n tested bybot hmethods . As the influence of the pollinator onmonoploi d frequencies isprobabl y small, pseudogamic frequencies may not be increased by selection of good pollinators. One person can obtain 50-200 pseudogamic monoploids ina sea ­ son. Similar numbers of plants might be obtained invitro ,bu tmos tplant s will be diploids. These are either from 2n gametes or from spontaneously doubled n embryos and itwil l be difficult to separate the two. Male ste­ rility is also very frequent among diploid S. tuberosum plants. All to­ gether pseudogamy has advantages over anther culture formonoploi d produc­ tiona swell .

100 Summary

Chapter 1. Haploid plants canb eproduce d fromsevera l species.The yhav e applications inbasi c research andar ea valuabl e tool inplan tbreeding . Dihaploids of Solanum tuberosum canoccu r amongst theprogen y ofcrosse s with S. phureja. This publication describes experiments to determineth e influence of genetic and environmental factors on S. tuberosum dihaploid frequenciesan dt odetermin ewhethe rmaximu m frequencieshav ebee nreached .

Chapter 2. Ina literature review occurrence andus eo fdiploid s are de­ scribed togetherwit h literature relevantt oth eexperimenta lchapters .Ha - ploids from diploidsan dallopolyploid sca nyiel dhomozygou splant sfo rhy ­ brid breeding. Haploids from autotetraploids areuse d forbreedin g at di­ ploid level, after which an improved tetraploid canb ereconstituted ;2 n gametesca npla ya nimportan trol ei nthi sretetraploidization . Frequencieso f S. tuberosum dihaploids,obtaine d throughpseudogamy ,ar e influencedb yth egenotype so fbot hth epollinato ran dth esee dparent .Se ­ lectionan dbreedin gha sbee n donefo rsuperio rpollinators .Gene sfo rhig h dihaploid frequencieswer e describeda srecessive .Gene si nth esee dparen t forhig h dihaploid frequencies were foundt ob edominant . Dihaploid frequencies canb e influenced toa large extentb yenviron ­ mental factors like temperature,light ,radiatio nan dchemicals . Endosperm plays amajo r role indevelopmen to fseeds .Dihaploi dembryo s of5 . tuberosum have onlybee nobserve di nhexaploi dendosper m Thetwenty - four paternal chromosomes inth eendosper m areeithe r contributed byon e 2nsper mo rb ytw on sperms .Argument sfo rbot hhav ebee ngiven .

Copter 3. crosseswer emad ebetwee npollinator san dth eprogeny^teste dt o study the genetics ofth e influence ofth epollinato r ondihaploi dfte .uencies.Th enumbe ro floc ideterminin g -^^X^^L lation studied wasprobabl y five, assuming ^J^ ^ the progeny action was of the intermediate type Seve,^1 P ^ ^ ^ ^ were better than thebes t parent,an don eP»" ^ * ed high „ llM P haploids per 100 berries (d/100b).Certai n * *°" dihaploid fre. hers of hybrid seeds. This had a smaU ^^^V » ™ - quencies. The number of dihaploids wasreau u everylo oextr ahybrids . ofth esee d parento ndihaploi d Chapter 4. Theexistenc e ofa ninfluenc e 101 frequencieswa s confirmed. Results ofa ninheritanc e studyo f the seedpar ­ enteffec twer eno tconclusive .Probabl y several lociwer e involved andth e inheritance was unlikely to be recessive, possibly intermediate. A major influence ofth ecytoplas m ondihaploi d frequencies wasno t found. Thespo - rophyte rather thanth e gametophyte determined the frequency ofdihaploids .

Chapter 5. Interactionbetwee npollinato r and seedparen t influence ondi ­ haploid frequencies was studied. Apart from analysis of data over several years, a special experiment was carried out, in which 'crossing date'ef ­ fectswer e excluded.A seedparen tx pollinato r interaction was found only, but not always, when both main effects were significant. This interaction disappeared after a logarithmic transformation, indicating that a multi­ plicative effect was responsible for most of the interaction. Interaction in a strict sense between the pollinator and the seedparen t effect seemed tob e absent.A pollinato r xyea r (crossing date)interactio nwa s found. Forhybri dproductio n no seedparen tx pollinato r interaction was found, but again apollinato r xyea r (crossing date)interactio n was found.

Chapter 6. Crosses made in a greenhouse had shown considerable variation indihaploi d frequencies during aseason .A n experiment intw o growth cham­ bers with high temperature (23 °C day, 15-18 °Cnight )an d low temperature (18 °C constant)wa s set up to give information about the influence ofth e temperature.Ther ewa s noclea r effecto fth e temperaturevi a the seedpar ­ ento nd/100b . Low temperature had apositiv e effecto nd/100 b via thepol ­ linator. Hybrid frequencies were more variable thandihaploi d frequencies. High temperature had a positive influence on hybrid frequencies via the seed parent when IVP35 was the pollinator.Pollinator s exposed to lowtem ­ perature gavehighe r hybrid frequencies thanthos e exposed tohig htempera ­ ture.

Chapter 7. Cytological techniques were used to elucidate themechanis m of dihaploid formation.Hig h 2npolle n frequencies in S. phureja did notcoin ­ cidewit hhig hdihaploi d frequencies,bu t theydi dwit h hybrid frequencies. No correlation was found between dihaploid and hybrid frequencies. Ovule lethality proved to be high, especially whenman y 3x embryoswer epresent . Irregularities were found in the pollen tubemitosi s of S. phureja (invi ­ vo). In 5% of the tubes only one sperm was found, but inonl y very fewi t was certain that only one sperm was present. In a further 5% ofth e tubes the two sperms were very close or touching. This group might represent sperms with a 'functional' restitution, leading todoubl e fertilization of the central nucleus, a hexaploid endosperm and an unfertilized egg cell. Observations on the growth rate of pollen in styles showed that 2x pollen grew fastertha nx polle n in S. tuberosum. Thiswa s also true for S. phure­ ja. Delayedpollinatio n didno tincreas e dihaploid frequencies.

102 From these elements itwa s concluded,tha t2 n S. phureja pollen produced tetraploid hybrid seeds after pollination of S. tuberosum. The n pollen, however, produced lethal triploid hybrid seeds and induced dihaploids in hexaploid endosperm. The double fertilization of the central nucleus was possible through complete or incomplete,bu t functional,restitutio no fth e pollentub emitosis .

Chapter 8. Monoploids from S. tuberosum were alsoproduce d throughpseudo - gamy, using the same pollinators as for dihaploids. Differences between pollinators were not significant. Seed parents differed considerably; the highest frequency found was 17 monoploids per 100 berries,whil e certain seedparent s did notproduc e anymonoploid . Themechanis m ofmonoploi d for­ mation almost certainly differed from that of dihaploid formation. Again then polle n of S. phureja was involved,bu ton esper mdi dno t function and theothe r fused with thecentra lnucleu s to form3 xendosperm . Threemonoploid s from diploid S. verrucosum wereobtained .

Chapter 9. In a concluding chapter the implications of the findings are discussed. It will be possible to breed pollinators for higher dihaploid frequencies. However, it is most likely not worth the effort, as fre­ quencies from existingpollinator s arehig h already. Pseudogamy is more efficient than anther culture as ametho d fordiha ­ ploid production, both in costs and in genetic diversity. For monoploid production pseudogamy is more efficient as well. Both methods may be limited inth e number ofgenotype s fromwhic hmonoploid s canb e extracted.

103 Samenvatting

Geslachtscellen van planten bevatten gewoonlijk de helftva nhe t aantal chromosomen dat in de lichaamscellen voorkomt. Een plant die zich ontwik­ kelt uit één enkele geslachtscel wordt eenhaploi d genoemd.Vee l haploiden zijn onvruchtbaar, omdat ze op hun beurt geen geslachtscellen kunnen vor­ men. De cultuuraardappel, Solanum tuberosum, heeft alle chromosomen in viervoud; ditword t aangeduid als4x ,waarbi j "x"éé nvolledig e setchromo ­ somenvoorstelt .Haploide nva n de aardappel hebben 2x chromosomen enworde n dihaploiden genoemd. Ze zijn vaak vruchtbaar. Zekunne n ingrot e aantallen gevormdworde ndoo rbestuivin g vanbloeme nme tstuifmee lva n Solanum phure- ja, eenaardappelsoor t uitZuid-Amerika , diediploi d (2x)is .D ebesse n die worden gevormd na deze bestuiving bevatten meestal weinig zaden. Een paar zadenkunne ndihaploi d zijne nd eres tnormaa l kruisingsprodukt (hybride). Voor toepassingen inveredelin g enonderzoe kmoete n dihaploiden op grote schaal gemaakt kunnen worden. Hoewel de dihaploiden geen chromosomen ont­ vangenva n debestuiver ,beïnvloede n debestuiver swe l de frequentie vand e dihaploiden. Er kan op hoge frequentie geselecteerd worden. Ook is het praktisch als dihaploiden zo vroeg mogelijk vanhybride n kunnenworde non ­ derscheiden. Dit is mogelijk door gebruik te maken van zaadstip. Bepaalde bestuivers zijn homozygoot voor zaadstip en al hun hybride nakomelingen kunnenreed s alszaa dworde ngescheide nva nd edihaploiden . Wilde aardappelsoorten (2x) kunnen eigenschappen - zoals resistenties tegen ziekten - hebben, die men over zou willen brengen naar de cultuur­ aardappel. Echter, 2x-planten zijn over het algemeen slecht kruisbaar met 4x-planten. Door uit 4x-planten dihaploiden te maken, deze te kruisen met wilde 2x-planten en van hun nakomelingen het aantal chromosomen weer te verdubbelen, kunnen de gewenste eigenschappen toch benut worden voor de veredeling van decultuuraardappel . Dihaploiden zijnoo knutti gvoo r erfelijkheidsonderzoek. Bij gebruikva n 2x-planten zijnvee lminde rplante nnodi go mhe tvoorkome nva n eenbepaald e eigenschap ind enakomelinge n tevolge n danbi j4x-planten . Het doel van dit onderzoek was omui tt ezoeke n ofe rbestuiver s konden worden geselecteerd, die nog meer dihaploiden zoudenkunne n leverenda nd e beschikbare bestuivers. Daarnaast zou worden getracht om meer inzicht te krijgen in de vererving van de eigenschap om dihaploiden te leveren en in de ontstaanswijze vandihaploiden . Verschillende bestuivers zijn met elkaar gekruist. Denieuw e bestuivers die zo ontstonden, zijn getest op hun dihaploiden-leverend vermogen. Dit

104 leverdenieuw ebestuiver s diebete rware nda nd ebest eouder .D ebest e nieuwebestuive r leverde gemiddeld achtdihaploide npe rbes .Sommig ebe ­ stuivers leverdenhog e aantallenhybriden .Dez ehog eaantalle nhadde nee n negatiefeffec t op hetaanta l dihaploiden:voo r iederehonder d extrahybri ­ denwer d het aantal dihaploiden gemiddeldme téé nverlaagd . Hetteste nva n denieuw ebestuiver s leverdeoo kaanwijzinge no pove rd e verervingva n het dihaploiden-leverend vermogen.Va nieder egroe pbestui ­ verswerde n deverhoudinge n tussen goedee nslecht ebestuiver svergeleken . Hieruitwer d geconcludeerd, datd e erfelijkheidsfactoren voordez eeigen ­ schap op de chromosomen liggen ennie ti nhe tcelplasma .E rzij nmeerder e genendi edez e eigenschap bepalen;binne nd eonderzocht e groepbestuiver s zijne rwaarschijnlij k vijf.Binne n eenge ni sd eexpressi enie tdominan t ofrecessief ,maa r intermediair.D e effectenva nd eafzonderlijk e genen versterken elkaar. Naastd ebestuive r heeftoo kd emoederplan tvee l invloedo phe t aantal geproduceerde dihaploiden.Oo kbi jd emoederplante nbleke nd e factorendi e hetaanta l dihaploiden bepalen opd echromosome nt eliggen .D e eigenschap wordtbepaal d door deerfelijk e eigenschappenva nd eplan te nnie tdoo rdi e vand e eicellen.E rzij nwaarschijnlij k meerdere genenwerkzaam . Erwer d onderzocht ofe r interactiewa stusse nhe teffec tva nd emoeder - plante nda tva n debestuive r op de dihaploiden frequentie (d.f.).Al sd e d.f.va n eengoed emoederplant ,bestove ndoo ree ngoed ebestuiver ,hoge ri s dano pgron d van ieders d.f.me tee ngemiddeld e testplantverwach tmoch t worden, ise r sprakeva n eenvermenigvuldigingseffect . Alsd erangord ei n d.f.va nmoederplante n afhangtva nd ebestuive rwaarme e getestword te nd e rangordeva n debestuiver s vand emoederplant ,i se rsprak eva n interactie instrikt ezin . Erwer d gebruik gemaaktva nd ejaargemiddelde nva nverschillend ejaren . Ookwer d er een gedetailleerder experimentuitgevoerd ,waari nmilieumvloe - denzovee lmogelij k gelijkwerde n gehoudenvoo rall emoederplante n enbe ­ stuivers.Groepe n gegevens werdenme tbehul pva nee nvariantieanalys ege ­ analyseerd, m sommige combinaties werd inderdaadee ninteracti etusse n moederplant- enbestuiver-effec t gevonden.Dez e interactieverdwee nechte r nit-hBtpkent ,da tallee n naee n logarithmische bewerkingva nd egegevens .Di tbetekent, ^ ^^ eenvermenigvuldigingseffec t aanwezigwa se ngee ninteracti e ins r i «in.He t ontbrekenva n deze interactieheef t alsconsequenti e dat h^ het20eke n naar eenbestuive rme thog ed.f .voldoend e ^^^^ Plantt e toetsen enda tee nmoederplan tme thog ed.f . kanworde n pg metbehul p vanmaa r éénbestuiver . 0 Oemeest e proeven zijn inee n ka,gedaan ,waari nd etemperatuu r «^ wastenzi j debuitentemperatuu r tehoo goplie p Inde J op zoekbleke n erverschille n ind.f . tebestaa n tussen*ruisxn g g verschillende dagen.Vermoe dwerd , datme tnam ed e»«** £ J^^. „•i-ir,p rDroeve n intwe eklimaatkamer s speelde.O m deze invloed na tegaa nzij ne rproev e 105 gedaan: een met een constante lage temperatuur (18 °C)e nee nme tee nwis ­ selende, gemiddeld hogere temperatuur (23 °C overdag en 15-18 °C 'snachts) . Binnen de onderzochte temperatuursgrenzen bleek ergee n invloed van de temperatuur via de moederplanten op de d.f. te zijn. Lage tempera­ tuur verhoogde wel de d.f. van debestuivers .Oo k aantallenhybride n waren hoger als debestuiver sbi j lage temperatuur opgroeiden. Het is dus belangrijk om met name de bestuiverplanten op een koele plaats te zetten ofo m debestuivinge nvroe g inhe tseizoe n temaken ,voor ­ date rhog e temperaturen optreden. Om een beter inzicht te krijgen in de ontstaanswijze van dihaploiden zijn microscopische technieken gebruikt. Uit de literatuur wasbekend , dat dihaploiden alleen zijn gevonden in zaden met 6x-kiemwit. Als regel be­ vrucht een generatieve kern van het stuifmeel de eicel en de andere de (dubbele)central e kernva nd ekiemzak .Aangezie n demoederplan t4 xchromo ­ somen levert voor het kiemwit, moet de resterende 2xva nhe t stuifmeelko ­ men, terwijl deeice lnie tword tbevrucht . Bij sommige bestuivers bevat een deel van het stuifmeel 2x chromosomen in plaats van x. Percentages van dit2x-stuifmee l werdenbepaal d voorver ­ schillende veelgebruikte bestuivers door de doorsnede van eengroo t aantal stuifmeelkorrels te meten. Deze percentages liepen niet parallel met de aantallen dihaploiden, wel met de aantallen hybridendi e door deverschil ­ lende bestuivers werden geproduceerd. Ook bleek er geen parallel te zijn tussenhe t aantal dihaploiden enhybriden .He t2x-stuifmee l speeltdu s geen directe rol bij de dihaploidenvorming, wel bijd ehybridenvorming .He twa s al bekend, dat de meeste hybriden 4x-planten waren.Zi jworde n dus-gevormd uit een 2x-eicel van de moederplant en een 2x-generatieve kern van debe - stuiver. Een anderemogelijkhei d voorhe tontstaa nva n 6x-kiemwit is,da td e twee generatieve kernen van het x-stuifmeel beide de centrale kernva n de kiem­ zak bevruchten, na mogelijk eerst samen versmolten te zijn. Om dit te on­ derzoeken werden de kernen van stuifmeelbuizen van de bestuivers, die 24 uur in een stijl waren gegroeid, onder een microscoop bekeken. In sommige buizen werd maar één generatieve kern gevonden, maar dit kon slechts zeer zelden met zekerheid worden vastgesteld. De frequentie was te laag om de gevonden aantallen dihaploiden te verklaren. In ongeveer 5%va n de buizen werden de twee generatieve kernen vlak bij elkaar of tegen elkaar aange ­ vonden. Het is mogelijk dat deze kernenparen zich als een geheel gedragen en samend ecentral eker nva nd ekiemza kbevruchten . Hybridenme t3 xchromosomen ,he tnormal e produktva n een kruising tussen een4x -e nee n2x-plant ,hebbe n geringe levenskansen. Tellingen van deaan ­ tallen zich ontwikkelende zaadbeginsels in jonge aardappelbessen bevestig­ den, datd e3x-zade n ingrot emeerderhei d afsterven. In het kort: het 2x-stuifmeel van s. phureja levert na bestuiving van een 4x-aardappelplant levenskrachtige 4x-hybriden. Het x-stuifmeel levert

106 grote aantallen 3x-hybriden, die in een vroeg stadium afsterven, en een paar dihaploiden in 6x-kiemwit, als beide generatieve kernen op de een of anderemanie r samen de centrale kernva nd ekiemza kbevruchten . De groeisnelheid van 2x-stuifmeel werd vergeleken met die van x-stuif- meel door de in 24 uur afgelegde afstand in een stijl te vergelijken.He t 2x-stuifmeel bleek gemiddeld sneller te groeien dan het x-stuifmeel. Dit heeft tot gevolg, dat 2x-stuifmeel naar verhouding meer eicellen bevrucht en dat er meer 4x-hybriden optreden dan op grond van de frequenties van 2x-stuifmeelverwach tma gworden . Bestuiving van oudere bloemen bleek geen verhoging van de d.f. op te leveren. Het is waarschijnlijk mogelijk om bestuivers te kweken die nog grotere aantallen dihaploiden opleverenda nd eto tn uto egeselecteerd ebestuivers . Hiervoor zijn zeer grote aantallen planten nodig,di e allemaal ophu nd.f . getestmoete nworden .He t isd evraag ,o fdez e langdurige selectieprocedure demoeit ewaar d zal zijn,aangezie n dehuidig ebestuiver s alredelijk ehoe ­ veelheden dihaploiden opleveren. Het bleek mogelijk het chromosoomaantal van dedihaploide nva nd eaard ­ appel opnieuw te halveren, weer door bestuivingen met S. phureja. Dege ­ vormdeplante nme tx chromosome nworde nmonoploide n genoemd.D emonoploide n frequentie bleek niet duidelijk afhankelijk van de bestuiver. De moeder- planten hadden wel een grote invloed. Sommige produceerden in het geheel geenmonoploiden , terwijl dehoogst e gevonden frequentie 17monoploide npe r 100 bessen bedroeg. In totaal werden 82 S. tuberosum monoploiden gevonden

zowel van dihaploiden als van andere 2x-aardappelplanten.Oo k werden drie monoploiden van Solanum verrucosum (2x)verkregen . Aangezien bestuivers met meer 2x-stuifmeel niet meer monoploiden leve­ ren, moet bij het ontstaan van monoploiden weer het x-stuifmeel een rol spelen.Va n de twee generatieve kernenbevruch te réé nd ecentral e kernva n de kiemzak, terwijl de ander verloren gaat, zodat de eicel °nbe-ucht blijft. Het is niet duidelijk waarom een van de generatieve « werkzaam is.He t vermogen van eicellen om onbevrucht tochto tee nkie m tegroeie n iswaarschijnlij k erfelijkbepaald . .. Door het verdubbelen van het aantal chromosomen van enigemonoplo dn werden homozygote 2x-planten verkregen. Homozygote ^^^J2_ door inteelt niet te verkrijgen. Ze kunnen worden gebruikt in,-gerede lingssysteem dat op "hybride groeikracht" is gebaseerd.Ee n ^»^ passing ind everr e toekomst isd e teeltva nhybrid e aardappelenui tzaad .

107 References

Abdalla, M.M.F., 1970. Inbreeding, heterosis, fertility, plasmon differen­ tiation and Phytophtora resistance in Solanum verrucosum Schlechtd., and some interspecific crosses in Solanum. Ph.D. Thesis,Wageningen , 213p . Abdalla, M.M.F. & J.G.Th. Hermsen, 1972. Plasmons andmal e sterility types in Solanum verrucosum and its interspecific hybrid derivatives.Euphy - tica21 :209-220 . Al-Yasiri, S. & O.M. Rogers, 1971. Attempting chemical induction of ha- ploidyusin g toluidinblue .J .Am . Soc.Hort . Sei. 96:126-127 . Aman,M.A . & K.R. Sarkar, 1978. Selection for haploidy inducing potential inmaize . IndianJ . Genet, and Pi. Breed.38 :452:457 . Arnason, T.J., 1943. Female sterility in potatoes. Can. J. Res. 21, C: 41-56. Bains, G.S.& H.W . Howard, 1950.Haploi d plants of Solanum demissum. Nature 166: 795. Bannikova, V.P. & O.A. Khvedynich, 1974. Features of fertilization in the course of remote hybridization of plants. In: H.F. Linskens (ed.). Fertilization inhighe rplants ,Amsterdam ,p .301-307 . Beamish,K.I. ,1955 .See d failure following hybridizationbetwee n thehexa - ploid Solanum demissum and four diploid Solanum species. Am. J. Bot. 42:297-304 . Behnke, M., 1975 .Regeneratio n in Gewebekulturen einiger dihaploider Sola­ num tuberosum-Klone. Z.Pflanzenzüchtg . 75:262-265 . Behnke,M. , 1979.Selectio n ofpotat o callus forresistanc e to culture fil­ trates of Phytophtora infestans and regeneration of resistantplants . Theor.Appl . Genet.55 :69-71 . Bender, K., 1963. Über die Erzeugung und Entstehung dihaploider Pflanzen bei Solanum tuberosum. Z. Pflanzenzüchtg. 50:141-166 . Bienz, D.R., 1958. The influence of environmental factors and pollinating techniques on the success of potato pollination in the greenhouse. Am.Potat o J. 35:377-385 . Binding,H. , 1974.Mutatio n inhaploi d cell cultures. In:K.J . Kasha (ed.), Haploids inhighe rplants ,Univ . ofGuelph ,p .323-337 . Bingham, E.T., 1969. Haploids from cultivated alfalfa Medicago sativa L.. Nature 221:865-866 . Bingham, E.T.,1971 .Isolatio no fhaploid s oftetraploi d alfalfa. Crop Sei. 11: 433-435.

108 Bingham, E.T. & J.W. Saunders, 1974. Chromosome manipulation in alfalfa: scaling the cultivated tetraploid to seven ploidy levels. Crop Sei. 14:474-477 . Blakeslee, A.F., J. Belling, M.E. Farnham & A.D.Bergner , 1922.A haploid mutant intji eJimso nWeed , Datura stramonium. Science55 :646-647 . Bodlaender, K.B.A., 1960. De invloed van de temperatuur op de aardappel. Jaarboek I.B.S.: p. 69-83. Bourgin, J.P. & J.P. Nitsch, 1967.Obtentio nd e Nicotiana haploides àpar ­ tird'étamine s cultivées invitro .Ann .Physiol .Vég .9 :377-382 . Breukelen,E.W.M .van ,1972 .Overervin gva nhe tvermoge nva n Solanum phure- ja om dihaploiden te induceren in Solanum tuberosum. Ingenieursver­ slag, Landbouwhogeschool,Wageningen ,unpublished , 28p . Breukelen, E.W.M. van, M.S.Ramann a & J.G.Th. Hermsen, 1975.Monohaploid s (n=x=12) from autotetraploid Solanum tuberosum (2n=4x=48) through two successive cycles of femaleparthenogenesis .Euphytic a24 :567-574 . Breukelen, E.W.M. van, M.S. Ramanna & J.G.Th. Hermsen, 1977.Parthenoge ­ nese monohaploids (2n=x=12)fro m Solanum tuberosum L.an d S. verruco- sum Schlechtd. and the production of homozygous potato diploids.Eu ­ phytica 26:263-271 . Broksh, V.L., 1969.Method s ofproducin g haploidplant so f Solanum tubero­ sum. Shorn. Trud. Aspirant. Inst.Raste n 10:470-478 . In:P.B.A .40 : 1109.

Budin, K.Z. & V.L. Broksh, 1972. The influence of the pollinator on the yields of haploids in the potato. Trudy po Prikladnoi Botamke Gene- tike iSelektsi i 48:126-131 .In :P.B.A .43 :9770 . Bukai, J., 1973. inductiono fdihaploid s inpotat ob ymean so f X""rad^ed , A cn . ni-113 In:P.B.A .43 :7975 . pollen.Kisérletügy i Kozlemenyek,A 64.H I iu. , • „„m^tnrs forproducin g dihaploids.Kar - Buketova, M.M., 1970.Choic e o f pollinatorc s rorpi . tofel iOvosk i 6: 12.In :P.B.A .41 :1291 . Buketova, M.M. * I.M.Yashina , 1971.Th eproductio no fdihaploids_x n Sola num tuberosum and their characterization. Genetika Moskva 7. 3947 .

in:P.B.A .41 :8159 . species Buketova, M.M. & I.M. Yashma, 1973. Meiosis i P which areproducer s ofdihaploids .Tsitologiy a15 :503-507 .In .

44: 1791' 1Q79 Maternalhaploid so f Nico- Burk,L.G. , D.U. Gerstel & E.A. Wernsman, 1979.Mater n tiana tabacum L. from seed. Science206 :585 . ploidiearades von

Zuckerrüben durch Zählen der Schlieszelien

28: 309-314. L.améliorationd u caféier en Capot, J., B. Dupautex * A. Durandeau, 1968^^ " ~ Cafe Cacao Côte d'ivoire. Duplication chromosomique ethybridis a Thé 12:114-126 .

inq Carroll, C.P., 1975.Th e inheritance and expression of sterility inhybrid s ofdihaploi d and cultivated potatoes.Genetic a 45:149-162 . Carroll, C.P. & M. Borrill, 1965.Tetraploi d hybrids from crosses between diploid and tetraploid Dactylis. and their significance. Genetica 36: 65-82. Carroll, C.P. & R.J. Low, 1975. Flowering behaviour and seed fertility in dihaploid Solanum tuberosum. PotatoRes .18 :416-427 . Carroll, C.P. & R.J. Low, 1976.Aspect s ofmal e fertility ingrou p Tubero­ sum dihaploids.Potat oRes .19 :109-121 . Chase, S.S., 1949.Monoploi d frequencies in a commercial double cross hy­ brid maize, and in its component single cross hybrids and inbred lines.Genetic s 34:328-332 . Chase, S.S., 1952a.Monoploid s in maize. In: J.W. Gowen (ed.), Heterosis, Ames,p .389-399 . Chase, S.S., 1952b. Production of homozygous diploids of maize from mono­ ploids.Agronom y J.44 :263-267 . Chase, S.S., 1963a.Androgenesi s -it sus e fortransfe r ofmaiz e cytoplasm. J.Hered . 54:152-158 . Chase, S.S., 1963b. Analytic breeding in Solanum tuberosum L.. A scheme utilizing parthenotes and other diploid stocks. Can.J . Genet. Cytol. 5:359-364 . Chase, S.S., 1964a. Analytic breeding of amphipolyploid plant varieties. Crop Sei.4 :334-337 . Chase, S.S., 1964b.Parthenogenesis .Maiz e Genetics CooperationNew s Letter 38: 46. Chase, S.S., 1969.Monoploid s and monoploid-derivatives ofmaiz e (Zea maus L.). TheBotanica l Review 35:117-167 . Chase, S.S., 1973.Hig h androgenetic rates inmaize .Genetic s 74:45 . Chase, S.S., 1974.Utilizatio n of haploids inplan tbreeding :breedin gdi ­

ploid species. In:K.J . Kasha (edr)-r Haploids inhighe rplants ,Univ . ofGuelph ,p .211-230 . Chaudhari, H.K. & J.R. Barrow, 1975. Identification of cotton haploids by stomatal chloroplast-counttechnique .Cro p Sei. 15:760-763 . Clarke,A.E. , 1940.Fertilizatio n and earlyembry o development inth epota ­ to. Am. Potato J. 17:20-25 . Coe, E.H. & K.R. Sarkar, 1964. The detection of haploids in maize. J. Hered. 55:231-233 . Collins, G.B. & P.D. Legg, 1974. Theus eo fhaploid s inbreedin g allopoly­ ploid species. In: K.J. Kasha (ed.), Haploids inhighe rplants ,Univ . ofGuelph ,p .231-247 . Cooper,D.C .& R. ABrink , 1945.See d collapse followingmating sbetwee ndi ­ ploid and tetraploid races of Lucopersicon pimpinellifolium. Genetics 30:376-401 .

110 Cooper, D.C. & G.H. Rieman, 1958.Diploi d plants in a seedling population ofth e cultivated potato.Am .Potat o J. 35:642 . Daker, M.G., 1966. 'Kleine Liebling' a haploid cultivar of Pelargonium. Nature 211:549-550 . De laPuente , F.F., 1973. Haploids of group Andigena. Diss. Univ. Wise Diss.Abstr . Intern.B 197434 :4762B . De laPuente-Ciudad , F. & S.J. Peloquin, 1974.Haploid s ofgrou pAndigena . Am. Potato J. 51:278 . Devreux, M. & D. de Nettancourt, 1974. Screening mutations in haploid plants. In: K.J. Kasha (ed.), Haploids in higher plants, Univ. of Guelph,p .309-322 . Dnyansagar, V.R. & D.C. Cooper, 1960. Development of the seed of Solanum phureja. Am. J. Bot.47 :176-186 . Dodds, K.S.,1950 .Polyhaploid s of Solanum demissum. Nature 166:795 . Dodds, K.S. & D.H. Long, 1955.Th e inheritance of colour in diploidpota ­ toes I. Types of anthocyanidins and theirgeneti c loci.J .Genet .53 : 136-149. Dodds, K.S. & D.H. Long, 1956.Th e inheritance of colour in diploidpota ­ toes II.A three-factor linkage group.J . Genet.54 :27-41 . Driver, CM. & J.G. Hawkes,1943 .Photoperiodis m inth epotato . Imp.Burea u ofPlan t Breeding andGenetics ,Cambridge ,3 6p . Dublin, P., 1974. Les haploides de Theobroma cacao L.. Diploidisation et obtention d'individus homozygotes.Caf é CacaoTh é 18:83-96 . Dumas de Vaulx, R. & E. Pochard, 1974.Essa i d'induction de laparténoge - nèsehaploid epa r action duprotoxyd e d'azote surle s fleursd epimen t (Capsicum annuum L.).Ann .Amélior .Plante s 24:283-306 . Dunbier, M.W. & E.T. Bingham, 1975.Maximu m heterozygosity inalfalfa :re ­ sults using haploid-derived autotetraploids.Cro p Sei.15 :527-531 . Dunwell, J.M. & N. Sunderland^ ••, 1973/• TQ-7-.Antne 5 Anther culturr culture e of Solanum tuberosum. Euphytica 22:317-323 . Ecochard,R. , G. Merks &M .Matteoli ,1974 .Parthenogenesi s inducedb yspe ­ cific radioinactivation of the male gamete. In:K.J .Kash a (ed.),Ha ­

ploids inhighe rplants, Univ .o fGuelph ,p .136 . Differences in

Eenink. A.H., 1974.Matromorph y in Brassica oleracea L. II. i ±_ parthenogenetic ability andparthenogenesi s inducing ability.Euphyt i

C 3 4 3 5 4 4 5 ESen, :.: R. K . ; o o s;a G. Geraci, 1978.See d set,siz ean ddevelopmen taf ­ ter4 x x 2x and4 xx 4 x crosses in Citrus^ Euphyic a2 -S . Foroughi-Wehr, B., H.M. Wilson G. Mi>,* H G -^ ^ _ ^^ from anthers of adihaploi d genotype of Solanum

ca 26:361-367 . K j Kasha Fossard, R.A. de, 1974. Terminology in 'Haploid work. n (ed.), Haploids inhighe rplants, Univ .o fGuelph ,p .40 3

111 Frandsen, N.O., 1967. Haploidproduktion aus einem Kartoffelzuchtmaterial mitintensive rWildarteinkreuzung .De r Züchter37 :120-134 . Frandsen, N.O., 1968. Die Piastidenzahl als Merkmal bei der Kartoffel. Theor.Appl . Genet.38 :153-167 . Gabert, A.C., 1963. Factors influencing the frequency of haploids in the common potato {Solanum tuberosum L.). Ph.D. Thesis,Univ . of Wiscon­ sin, 53p . Gaines,E.F . &H.C .Aase ,1926 .A haploid wheatplant .Am .J . Bot. 13:373 - 385. Gorea, T.V., 1968. Untersuchungen zur Erzeugung und züchterischen Nutzung dihaploider Kartoffeln (Solanum tuberosum L.). Diss. Landw. Fak. Rostock.Berlin , 96p . Gorea, T.V., 1970.Fertilitä t und Kreuzbarkeit derDihaploide n von Solanum tuberosum L.un d derenF lBastarden .Z .Pflanzenzüchtg . 64:201-220 . Gorea, T.V., 1973. Creare formelor dihaploide de cartof si perspectivile folosirii lor in ameliorare. Probleme de Genetica Teoretica si Apli- cata 5: 74-99. Guha, S. & S.C. Maheshwari, 1964. Invitr o production of embryos from an­ thers of Datura. Nature 204:497 . Guha, S. & S.C. Maheshwari, 1966. Cell division and differentiation ofem ­ bryos inth epolle ngrain s of Datura invitro .Natur e 212:97-98 . Guignard, M.L., 1902.L a double fécondation chez les Solanées. Journal de Botanique 16:145-167 . HSkansson, A., 1943. Die Entwicklung des Embryosacks und die Befruchtung bei Poa alpina. Hereditas 29:25-61 . Hanneman, R.E. & S.J. Peloquin, 1967.Crossabilit y of24-chromosom e potato hybridswit h48-chromosom e cultivars.Eur .Potat o J. 10:62-73 . Hanneman, R.E. & S.J. Peloquin, 1968. Ploidy levels of progeny from di- ploid-tetraploid crosses inth epotato .Am . Potato J. 45:255-261 . Harland, S.C, 1936.Haploid s in polyembryonic seeds of Sea IslandCotton . J.Hered . 27:229-231 . Harten,A.M .van ,1978 .Mutatio nbreedin g techniques andbehaviou r ofirra ­ diated shootapice s ofpotato .Ph.D .Thesis, Wageningen , 132p . Haustein, E., 1967.Di e selektive Befruchtung. In:W . Ruhland (ed.), Hand­ buch desPflanzenphysiologie . Bd 18,Berlin ,p .479-505 . Henderson, S.A. & B.C. Lu, 1968.Th eus eo fhaematoxyli n forsquas hprepa ­ rations ofchromosomes .Stai nTechnolog y 43:233-236 . Heribert-Nilsson, N., 1920. Zuwachsgeschwindigkeit der Pollensläuche und gestörteMendelzahle n bei Oenothera lamarckiana. Hereditas 1: 41-67. Hermsen, J.G.Th., 1969. Induction ofhaploid s and aneuploids incolchicine - induced tetraploid Solanum chacoense Bitt..Euphytic a 18:183-189 . Hermsen, J.G.Th., 1971. Tetrahaploids from colchicine-induced octoploid Solanum acaule subsp. aemulans. Euphytica 20:490-492 .

112 Hermsen, J.G.Th., 1974. Summation: Haploids in plant breeding. In. K.J. Kasha (ed.), Haploids inhighe rplants ,Univ .o fGuelph ,p .281-285 . Hermsen, J.G.Th., 1977.Toward s the cultivation indevelopin g countries of hybrid populations ofpotat o frombotanica l seeds.In :Repor tplannin g conference on utilization of genetic resources in the potato II. C.I.P.,p . 101-109. .,..,* Hermsen, J.G.Th., 1978a. Genetics of self-compatibility in diploids; of Solanum tuberosum L. 3. Lethality of S-bearing translocate homozy-

gotes. Euphytica 27:13-17 . . .. .ftf Hermsen, J.G.Th., 1978b. Genetics of self-incompatibility in ihapl dso f Solanum tuberosum L.4 .Linkag ebetwee n anS-bearm g translocation

a locus forvirescens .Euphytic a 27:381-384 . ..,.,+.„an d Hermsen, J.G.Th., 1979. Factors controlling interspecific crossai nd their bearing on the strategy for breaking *"»«•*» ££T ^

of tuber-bearing Solanum species. In: Proc. Conf Broadenlng Genet. Base crops,Wageningen , 1978."--79 , Pjl — ^ Sele Hermsen, J.G.Th. & J. Verdemus, 1973. ^° id induction group Phureja of genotypes combining high-frequency hp with homozygosity forembryo-spot .Euphytic a22 :244- 2 Hermsen, J.G.Th., M.Wagenvoor t& M.S .Ramanna ,197 0 ^uplo ral and colchicine-induced autotetraploids of Solanum. Can.J . Cytol. 12:601-613 . ^ ^.^ of a reCes- Hermsen, J.G.Th., M.S.Ramann a &J - Vogel, Solanum tuberosum by

means of trisomie analysis.Can .J -t*e n . i Haploids from t 97 4 p „.„«„,,...».. ». v,„ -«— .* o^»;:; m; h I;loW , M«-t autotetraploid Solanum tuberosum cv.Ginek et

cultivar aspollinators .H.I-S . 9:6. Hovinq 1974b. Acceptance Hermsen, J.G.Th., J. Olsder, P.C. ^^m ^rrlcosul in diploids from Soi of self-compatible pollen from ™ ^liMtlon in nigherplants , S. tuberosum. In:H.F . Linskens (ed.). Fertile Amsterdam, p. 37-40. Lipski, 1978a.In - BreU n Hermsen,J.G.Th. , L.M. Taylor,E.W.M ,va n ^ ploids and tneir re- heritance of genetic markers from two P^^1

spectiveparen t cultivars.Euphytic a 27. ^ ^ Ballegooijen & E.

Hermsen, J.G.Th., J. Sonneveld D. Vreug« ^ ^^ in dihaploids of

Staas-Ebrecht, 1978b. Genetics of /^trapioid potato cultivar r •; The S-genotype of tetrapioiu y Solanum tuberosum L. 5. ine ay

Gineke.Euphytic a 27:385-388 . ^rsica Batch.: Description Hesse, CO., 1971. Monoploid peaches, Prunus pers^ 326-330. andmeioti c ^[^ '^JZnJo l ofchromosom e eliminationdur - Ho, K.M. & K.J. Kasha, 1975.Geneti c ^ 263-275.

inghaploi d formation inbarley .Gen e ics " 183-188. Höglund M., 1970.Meiosi s in Solanum pnureja. Hereditas 11 t Hougas, R.W. & S.J. Peloquin, 1957.A haploid plant of the potato variety Katahdin.Natur e 180:1209-1210 . Hougas, R.W. & S.J. Peloquin, 1958. The potential of potato haploids in breeding and genetic research.Am .Potat o J. 35:701-707 . Hougas, R.W. & S.J. Peloquin, 1960. Crossability of Solanum tuberosum ha­ ploids withdiploi d Solanum species.Eur .Potat o J. 3:325-330 . Hougas,R.W. , S.J. Peloquin& R.W . Ross, 1958.Haploid s ofth e commonpota ­ to. J. Hered.49 :103-107 . Hougas, R.W., S.J. Peloquin & A.C. Gabert,1964 .Effec t ofsee dparen t and pollinator on the frequency of haploids in Solanum tuberosum. Crop Sei.4 :593-595 . lilies,Z.M. ,1974a . Induction ofhaploi d parthenogenesis in Populus tremu- la by male gametes inactivated with toluidine blue. In: K.J. Kasha (ed.), Haploids inhighe rplants ,Univ . ofGuelph ,p .136 . lilies, Z.M., 1974b. Experimentally induced haploid parthenogenesis inth e Populus section leuce after inactivation of the male gamete with to- luidine-blue-o. In: H.F. Linskens (ed.), Fertilization in higher plants,Amsterdam ,p .335-340 . Irikura,Y. , 1975a. Inductiono fhaploi dplant sb y anther culture intuber - -bearing species and interspecific hybrids of Solanum. Potato Res.18 : 133-140. Irikura, Y., 1975b. Cytogenetic studies on haploidplant s of tuber-bearing Solanum species I. Induction of haploid plants oftuber-bearin g Sola- nums. Res.Bull .Hokkaid o Natl. Agric. Exp. Stn.,No .112 ,p . 1-74. Irikura,Y & S.Sakaguchi ,1972 .Inductio n of12-chromosom eplant s froman ­ ther culture ina tuberou s Solanum. Potato Res.15 :170-173 . Ivanovskaja, E.V., 1939. A haploid plant of Solanum tuberosum L.. CR. (Doklady)Akad . Sei.U.S.S.R . 1939,24 :517-520 . In:P.B.A .10 :478 . Jackson, M.T., P.R. Rowe & J.G. Hawkes, 1978. Crossability relationships between Andean potato varieties of three ploidy levels.Euphytic a27 : 541-551. Jacobsen, E., 1976.Cytologica l studies on diplandroid production ina di - haploid potato clone and its correlation with seed set in 4x-2x crosses. 2.Pflanzenziichtg . 77:10-15 . Jacobsen, E., 1978.Haploid , diploid and tetraploid parthenogenesis inin ­ terspecific crosses between Solanum tuberosum interdihaploids and S. phureja. Potato Res.21 :15-17 . Jacobsen, E. & S.K. Sopory, 1978.Th e influence andpossibl e recombination of genotypes on the production of microspore embryoids in anther cul­ tures of Solanum tuberosum and dihaploid hybrids.Theor .Appl . Genet. 52: 119-123. Jakubiec, J., 1964. Wytwazanie form haploidalnych Solanum tuberosum. Hod. Ros. Aklim.Nas .8 :489-530 .

114 Janick, J. & H. Hughes, 1974. Production of polyhaploids in octoploid strawberries. In: Proceedings of the XIX International Horticultural Congress (IA Section VII, Fruits, p. 361),Warsaw , 1974. In:P.B.A . 46, 8294. Janssen, A.W.B. & J.G.Th. Hermsen, 1976. Estimating pollen fertility in Solanum species and haploids.Euphytic a 25:577-586 . Johnston, S.A., A.P.M. den Nijs, S.J. Peloquin & R.E.Hanneman , 1980.Th e significance of genie balance to endosperm development ininterspeci ­ ficcrosses .Theor .Appl . Genet.56 :5-9 .

J0rgensen, CA., 1928.Th e experimental formation ofheteroploi dplant si n the genus Solanum. J. Genet.19 :133-211 . Kasha,K.J . (ed.), 1974.Haploid s inhighe rplants .Univ .o fGuelph ,42 1p . Kasha, K.J. & K.N. Kao, 1970.Hig h frequency haploid production xnbarle y (Hordeum vulgare L.).Natur e 225:874-876 . Kermicle, J.L., 1969. Androgenesis conditioned by a mutation in maize. Science 166:1422-1424 . _ -„^^a Kermicle, J.L., 1974.Origi n of androgenetic haploids anddiploid s induced by the indeterminate gametophyte (ig) mutation in "-aize. In: .• Kasha (ed.), Haploids inhighe rplants ,Univ .o fGuelph ,p .13 7

Kessel, R. & P.R. Rowe, 1974. Inheritance of two qualitative **"•***«* proposed genetic map for their linkage group indiploi dpotatoes . tato Res.17 :283-295 . , Kidane-Mariam, H-M. * S.J. Peloquin, 1974.Th e effect of «*™c*™ o£ £ bridization (4xx 2xvs .2 x* 4x )o nyiel d ofcultivate d potatoes.Am .

Potato J. 51:330-336 . m~t-hnd Kihara, H. & K. Tsunewaki, 1962.Us e ofa nalie ncytoplas m asa ne wmetho d ofproducin g haploids.Jap .J . Genet.37 :310-313 . Kimber, G. &R . Riley, 1963.Haploi d angiosperms.Bot .Re v 19: _^ Klopfer, K., 1965. über den Nachweis von drei selbständigen Schichten Sprossscheitel derKartoffel .Z .Pflanzenzüchtg . 53:67-8 7 ,r Bl]ro 1952. chromosome number and chromosome Koopmans, A. & A.H. van der Burg, J-»^- ^ ,.„•=* Planum •,+..„*• thP cross Solanum phureja x Solanum behaviour of Fl and F2 plants of the crossa tuberosum. Genetica 26:102-116 . Weisspappeln •miio Fr?euaunq von haploiden weisspew Kopecky, F., 1960. Experimentelle Erzeugung (Populus alba L.).Silv .Genet . 9: 102-105. rpnetica 13: Kostoff, D., 1942. The problem of haploidy. Bibliographia Genetica

1-148 • ^ n4_ Aor Kartoffel (Solanum tube- Krug,H. , 1957.Da sphotoperiodisch e Verhalten derKartoffe l rosum). EineÜbersicht .Angew .Botani k 31:29-4 4 year_ Kruse, A., 1961.Haploid s inpolyembryo s of beet, Beta vulgaris book,Roya lVeterinar y andAgricultura l College,Copen hgen ^p

Lacadena, J.R., 1974.Spontaneou s and inducedP««^ ^ ^ of esis. in: K.J. Kasha (ed.), Haploids in higher plants, Guelph,p . 13-32. 115 Lamm,R. , 1938.Note s on ahaploi dpotat o hybrid.Heredita s 24:391-396 . Langton, F.A., 1974.A re-evaluation of the Dionne method of vegetatively doubling thechromosom e number inpotato .Potat oRes .17 :296-306 . Magoon,M.K . & K.R. Khanna,1963 .Haploids .Caryologi a 16:191-235 . Maherchandani, N. & Pushkarnath, 1960. Embryological studies in Solanum rubinii. IndianPotat oJ . 2:66-69 . Marks, G.E., 1955.A polyhaploi d plant of Solanum polytrichon Rydb..Natur e 175:469 . Marks, G.E., 1966a.Th eenigm a of triploid potatoes.Euphytic a 15:285-290 . Marks, G.E., 1966b. The origin and significance of interspecific poly­ ploidy: experimental evidence from Solanum chacoense. Evolution 20: 552-557. Martin, F.W., 1959. Staining and observing pollen tubes in the style by means of fluorescence.Stai nTechnolog y 34:125-128 . Matsumura, S., 1958.Breedin g of sugar beetsb ymean s oftriploidy . Seiken Zihô,Rpt .Kihar a Institute forBiol .Res .9 : 29-38. Melchers,G. , 1960.Haploid eBlütenpflanze nal sMateria l derMutationszüch ­ tung.De r Züchter 30:129-134 . Mendiburu, A.O., 1971. The significance of 2n gametes in potato breeding and genetics.Ph.D .Thesis ,Univ . ofWisconsin ,8 9p . Mendiburu, A.O. & S.J. Peloquin, 1969.Hal ftetra d analysis inpotatoe s by meanso f4x-2 xcrosses .Am .Potat o J. 46:438 . Mendiburu,A.O .& S.J. Peloquin, 1971.Hig hyieldin g tetraploids from 4x-2x and2x-2 xmatings .Am .Potat o J. 48:300-301 . Mendiburu, A.O. & S.J. Peloquin, 1979. Gene-centromere mapping by 4x-2x matings inpotatoes .Theor .Appl . Genet.54 :177-180 . Mendiburu,A.O. , S.J. Peloquin& D.W.S .Mok , 1974.Potat obreedin gwit hha ­ ploids and 2n gametes. In: K.J. Kasha (ed.), Haploids in higher plants,Univ . ofGuelph ,p .249-258 . Mendoza, H.A. & F.L. Haynes, 1974.Geneti cbasi s ofheterosi s foryiel d in theautotetraploi d potato.Theor .Appl .Genet .45 :21-25 . Mok, D.W.S. & S.J. Peloquin, 1975.Thre emechanism s of2 npolle n formation indiploi d potatoes.Can .J . Genet.Cytol . 17:217-225 . Mok, D.W.S., S.J. Peloquin & A.O. Mendiburu, 1976. Genetic evidence for mode of 2n pollen formation and S-locus mapping in potatoes. Potato Res. 19:157-164 . Montelongo-Escobedo, H., 1968.Factor s influencing haploid frequency inte - traploid Solanum tuberosum. Diss. Univ. wise. Diss. Abstr. Intern. B 1969 29:4533b . Montelongo-Escobedo,H .& P.R . Rowe,1969 .Haploi d induction inpotato :Cy - tologicalbasi s forth epollinato r effect.Euphytic a 18:116-123 .

Montezuma-de-Carvalho, J., 1967.Th e effect ofN 20 on pollen tube mitosis in styles and its potential significance for inducinghaploid y inpo ­ tato.Euphytic a 16:190-198 .

116 Morgan, D.T. & R.D. Rappleye, 1954. A cytogenetic study on the origin of multiple seedlings of Capsicum fi-utescens. Am.J .Bot .41 :576-585 . Müntzing, A., 1930.Übe r Chromosomenvermehrung in Galeopsis Kreuzungenun d ihrephylogenetisch e Bedeutung.Heredita s 14:153-172 . Müntzing,A. , 1937.Not e on ahaploi d ryeplant .Heredita s23 :401-404 . Murashige, T. & R. Nakano, 1966.Tissu ecultur e asa potentia l tool inob ­ taining polyploid plants.J .Hered . 57:115-118 . Najcevska, CM. & G.J. Speckmann, 1968.Number so fchloroplast s andpolle n grainpore s indiploi d andtetraploi dvarietie so fsom e Trifolium spe­ cies. Euphytica 17:357-362 . Nakamura, A., T. Yamada, N. Kadotani & R. Itagaki, 1974a. Improvement of flue-cured tobaccovariet yM C 1610b ymean so fhaploi dbreedin gmetho d and some problems of this method. In: K.J. Kasha (ed.), Haploids in higherplants ,Univ . ofGuelph ,p .277-278 . Nakamura, A., T. Yamada, N. Kadotani, R. Itagaki &M . Oka, 1974b. Studies on the haploid method of breeding in tobacco. Sabrao Journal 6: 107-132. Nanda, D.K. & S.S. Chase, 1966.A n embryo marker for detecting monoploids ofmaiz e (Zea mays L.).Cro p Sei.6 :213-215 . Nishiyama, I.& N . Inomata, 1966.Embryologica lstudie so ncross-incompati ­ bility between 2x and4 x in Brassica. Jap.J .Genet .41 :27-42 . Nitsch, C, 1974.Polle n culture - a new technique"fo rmas s production of haploid and homozygous plants. In: K.J. Kasha (ed.), Haploids in higherplants ,Univ . ofGuelph ,p .123-135 . Nitsch, J.P. & C. Nitsch, 1969.Haploi dplant s frompolle ngrains .Scienc e 163: 85-87. Nordenskiöld, H., 1939. Studies of a haploid rye plant. Hereditas 25: 204-210. .. Hijs, A.P.M, den, 1977.2 n gametes inpotat o species;occurrenc e genetic basis and function in sexual polyploidization. Ph.D.Thesis ,Univ . Wisconsin, 135 p. V,„V,T-Ï • iQT7 Th?rol eo fth eendosper m in nybn- Nijs,A.P.M . den& S.J. Peloqum, 1977.Th erol eo i dization.Am .Potat o J. 54:488-489 .

Obajimi, A.O. & E.T. Bingham, 1973. Inbreeding cultivated alfalfa in - tetraploid-haploid-tetraploid cycle: effect on morphology, fertility

and cytology. Crop Sei. 13:36-39 . „-u-to-v indiha -

Olsder, J. & J.G Th.Hermsen , 1976.Genetic s of "l^^™ ^^ Ploids of Solanum tuberosum L. I.Breedin gbehaviou r oftw o

patible dihaploids.Euphytic a 25:597-607 . markers asre -

Peloguin, S.J. & B.W.Hougas, 1959 .Decapitatio n andgeneti c «^»^ lated to haploidy in Solanum tuberosum. Eur.Potat oJ - • rap <~abert,1966 .Haploid y asa ne w

118 Rowe, P.R., 1974.Method s of producing haploids:parthenogenesi s following interspecific hybridization. In:K.J . Kasha (ed.), Haploids inhighe r plants, Univ. ofGuelph ,p .43-52 . Rutishauser, A., 1956.Di e Entwicklungserregung des Endospermsbe ipseudo - gamen Ranunculusarten. Mitteilungen der Naturforschenden Gesellschaft Schaffhausen 25:1-45 . Sadasivaiah, R.S., 1974.Haploid s in genetic andcytologica l research.In : K.J. Kasha (ed.), Haploids in higher plants,Univ .o fGuelph ,p .355 - 386. Sarkar,K.R. , 1974.Geneti c selection techniques forproductio no fhaploid s in higher plants. In: K.J. Kasha (ed.), Haploids in higher plants, Univ. ofGuelph ,p . 33-41. Sarkar, K.R. & E.H. Coe, 1966.A genetic analysis of the origin ofmater ­ nal haploids inmaize .Genetic s 54:453-464 . Skiebe, K., 1966. Polyploidie und Fertilität. Z. Pflanzenzüchtg. 56:301 - 342. Skiebe,K. , 1967.Übe r die genetischenUrsache n derSamenbildung .De rZüch ­ ter 37:75-82 . Skiebe, K., 1973. Genetische Voraussetzungen für dieSamenbildun g nachin - tergenomatisch verschiedenen Kreuzungen. Arch. Züchtungsforsch. 3: 183-202. Skiebe, K., 1975. Fertilität und vegetative Leistungbe imeiotisc hun dmi ­ totisch entstandenen polyploiden Formen. Arch. Züchtungsforsch. 5: 3-9. Smith, L., 1946.Haploid y inEinkorn .J .Agric .Res .73 :291-301 . Song, L.S.P., S.J. Park, E. Reinbergs, T.M. Choo & K.J. Kasha, 1978. Doubled haploids vs.th ebul kplo tmetho d forproductio no fhomozygou s lines inbarley . Z.Pflanzenzüchtg . 80:271-280 . ,„, ti=r,ini d freauenciesi n Stringham, G.R. &R.K .Downey , 1973.Haploi d irequen Brassica napus. Can. J. Plant Sei. 53:229-231 . . v., 1071 selective chromosome elimination Subrahmanyam, N.C. & K.J. Kasha„, 1973. selective . . , ..• „ i v following interspecific hybndiza- during haploid formatio nf in harbarle e y lonowo.y tion. Chromosoma42 :111-125 . .. 4-,'„„ r,r haDloidpotatoe s inth eNetner - Suchtelen,N.J . van,1966 . Investigationo fnapioi a P lands. Eur.Potat o J. 9: 64-68. „ . •- , -^ ~f Planum tuberosum. Eucarpia Confernn for- Suchtelen, N. van, 1973.Tnploid s of Solanum

ence onPotat o haploids,Wageningen-Köln . +„horn^m Suchtelen, N. van, 1976. Triploids of thecommo npotato , Solanum tuberosum L.. Potato Res.19 :377-380 . cmithAmerica n

Madrid, Spain,p .313-320 .

119 Sunderland, N., 1974.Anthe r culture as a means of haploid induction. In: K.J. Kasha (ed.), Haploids in higher plants,Univ . of Guelph, p. 91- 122. Tarasenko, N.D., 1974.Productio n of diploids in Solanum tuberosum L.. In: K.J.Kash a (ed.), Haploids inhighe rplants ,Univ .o fGuelph ,p .138 . Thévenin, L., 1974.Haploid s in asparagus breeding. In: K.J. Kasha (ed.), Haploids inhighe rplants ,Univ . ofGuelph ,p .279 . Thompson, K.F., 1972.Oi l seed rape. Report Plant Breeding Institute Cam­ bridge,p .94-96 . Tsunewaki,K. , T.Ry uEnd o &Y .Mukai , 1974.Furthe r discovery of aliency ­ toplasms inducing haploids and twins in common wheat. Theor. Appl. Genet.45 :104-109 . Turcotte,E.L .& C.V. Feaster,1967 .Semigam y inPim a cotton.J . Hered.58 : 54-57. Turcotte, E.L. & C.V. Feaster, 1969. Semigameticproductio n ofhaploid s in Pimacotton .Cro p Sei. 9:653-655 . Turcotte, E.L. & C.V. Feaster, 1974.Method s of producing haploids: semi­ gametic production of cotton haploids. In:K.J . Kasha (ed.), Haploids inhighe rplants, Univ . ofGuelph ,p . 53-64. Valentine,D.H . & S.R.J.Woodell , 1963.Studie s inBritis hprimula s X. Seed incompatibility in intraspecific and interspecific crosses atdiploi d and tetraploid levels.Th eNe wPhytologis t 62:125-143 . Vogt, G.E. & P.R. Rowe, 1968.Aneuploid s from triploid-diploid crosses in the series Tuberosa of the genus Solanum. Can J. Genet. Cytol. 10: 479-486. Wagenvoort, M. & W. Lange, 1975. Theproductio n of aneudihaploids in Sola­ num tuberosum L. group Tuberosum (the common potato). Euphytica 24: 731-741. Walsh, E.J., 1974.Efficienc y of the haploidmetho d ofbreedin g autogamous diploid species: a computer simulation study. In: K.J. Kasha (ed.), Haploids inhighe rplants ,Univ . ofGuelph ,p .195-210 . Wang,H .& S.J. Peloguin, 1971.Cytolog y of2 npolle n formation intw oPhu - reja-haploidTuberosu m hybrids.Am . Potato J. 48:301 . Wangenheim, K.H. von, 1956.Ursache n von Sterilität und schlechter Kreuz­ barkeitbe i Kartoffeln. Z. Pflanzenzüchtg.36 :445-456 . Wangenheim, K.H. von, 1957. Untersuchungen über den Zusammenhang zwischen Chromosomenzahl und Kreuzbarkeit bei So1anum- Arten. 2. Indukt. Ab- stamm.-u .Vererb.-Lehr e 88:21-37 . Wangenheim, K.H. von, 1961. Zur Ursache der Abortion von Samenanlagen in Diploid-Polyploid-Kreuzungen I.Di e Chromosomenzahlen von mütterlichen Gewebe,Endosper m undEmbryo .Z . Pflanzenzüchtg. 46:13-19 . Wangenheim, K.H. von, s.J.Peloqui n & R.W. Hougas,1960 .Embryologica l in­ vestigations on the formation ofhaploid s inth epotat o (Solanum tube­ rosum). Z.Vererb.-Lehr e 91:391-399 .

120 Wenzel, G., 1979.Neu eWeg e inde rKartoffelzüchtung .Kartoffelba u 30:126 - 129. Wenzel,G. ,1980 .Recen tprogres s inmicrospor e culture ofcro pplants .In : D.R. Davies & D.A. Hopwood (eds.). Fourth John Innes Symposium and Second International Haploid Conference.Th ePlan tGenome ,p .185-196 . Wenzel, G., O. Schieder, T. Przewozny, S.K. Sopory & G. Melchers, 1979. Comparison of single cell culture derived Solanum tuberosum L.plant s and a model for their application inbreedin gprograms .Theor .Appl . Genet. 55:49-55 . Williams,E.L. , 1955.See d failure inth eChippew avariet y of Solanum tube­ rosum. Botanical Gazette 117:10-15 . Winton, L.L. & D.W. Einspahr, 1968.Th eus eo fheat-treate dpolle n foras ­ penhaploi d production. Forest Sei. 14:406-407 . Winton, L.L. & R.F. Stettler, 1974.Utilizatio no fhaploid y intre ebreed ­ ing, in: K.J. Kasha (ed.), Haploids inhighe rplants ,Univ .o fGuelph , p. 259-273. Wöhrmann, K., 1963.übe rde nEinflus svo nröntgenbestrahlte m Pollen aufde n Anteil dihaploider Sämlinge in den Kreuzungsnachkommenschaften von Solanum tuberosum L.x Solanum phureja Juz.e tBuk. .Z .Pflanzenzüchtg . 50: 132-139. Wöhrmann,K. , 1964.übe rde nEinflus s derTemperatu r aufdi eDihaplddenra - tebe i Solanum tuberosum L..Z .Pflanzenzüchtg . 52:1-7 . Zenk, M.H., 1974. Haploids inphysiologica l and biochemical research. Ins K.J. Kasha (ed.), Haploids in higher plants,Univ .o fGuelph ,p .339 - 354.

121