Introduction to on-farm Organic Plant Breeding

Embed Size (px)

DESCRIPTION

http://ofrf.org/about Introduction to On-farm Organic Plant Breeding provides farmers with an overview of basic genetics, farm-based experimental design, and breeding techniques appropriate for organic farms. - See more at: http://ofrf.org/blogs/new-tools-organic-farmers-teach-diy-plant-breeding#sthash.TXCEgy0P.dpuf

Citation preview

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org

    Introduction to On-farmOrganic Plant Breeding

    Organic Seed AllianceAdvancing the ethical development and stewardship of the genetic resources of agricultural seedPO Box 772, Port Townsend, WA 98368

    This publication was made possible through a grant from Organic Farming Research Foundation and Seed Matters

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org2

    Section I: Introduction...................................................................................................................................................3 Whyplantbreedingisimportant......................................................................................................................................3 Aphilosophyoforganicplantbreeding.........................................................................................................................4 Ourfarmingancestorsneverstoppedbreeding...........................................................................................4 Returningfarmerstotheirroleasseedstewards........................................................................................4

    Section II: Plant breeding basics...............................................................................................................................5 Selectionintheoryandpractice........................................................................................................................................5 Howtoselect..............................................................................................................................................................................8 Acropsmatingsystemandhowitaffectsplantbreeding.....................................................................................8 Self-pollinatedcrops................................................................................................................................................9 Cross-pollinatedcrops............................................................................................................................................9 Breedingself-pollinatedcropsvs.breedingcross-pollinatedcrops..................................................................9

    Section III: Developing a plant breeding plan.....................................................................................................12 Thinkingaboutyourtargetenvironment...................................................................................................................12 Determiningtraits.................................................................................................................................................................13 Prioritizingtraits...................................................................................................................................................................13 Howcanthetraitsbemeasured?.....................................................................................................................13 Howeasilycanthetraitsbeinherited?.........................................................................................................14 Choosingparents...................................................................................................................................................................13 Creatingabreedingtimeline............................................................................................................................................15

    Section IV: Theories of field-based organic plant breeding...........................................................................21 Howgenestravelfromparentstooffspring..............................................................................................................21 Howgenesdeterminetheappearanceandperformanceofplants..................................................21 Howgenestraveltogetherduringreproduction......................................................................................22 Howgenesoperateinpopulations................................................................................................................................24 Howtoseethegeneticdifferencesbetweenplants...............................................................................................28 Understandtheeffectsoftheenvironment.................................................................................................28 Ensurethatplantsreceiveconsistenttreatment.......................................................................................29 Usesufficientpopulationandplotsizes.......................................................................................................29

    Section V: Examples of farmers breeding for organic systems......................................................................30 AbundantBloomsdaleorganicspinachbreedingproject...................................................................................30 Whatwerethegoalsofthisproject?..............................................................................................................30 Breedingprocedure...............................................................................................................................................30 WinterSproutingBroccoli...............................................................................................................................................31 Whatwerethegoalsofthisproject?..............................................................................................................31 BreedingProcedure...............................................................................................................................................31

    Glossary and index.......................................................................................................................................................33

    References and resources..........................................................................................................................................36

    Table of Contents

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org3

    Why plant breeding is importantThehistoryofthedomesticationofourcropplantsisalsothehistoryofplantbreeding.Throughmanygenerationsoflaborioushumanselection,wildplantsbecameusableasagriculturalcrops.

    Ourhunter-gathererancestorshadtohaveanintimaterelationshipwithplantstonoticethegeneticvariationthatexistedandcontinuallyaroseinordertomakemeaningfulprogressinthedo-mesticationprocess.Thisisoneofthefundamentalpreceptsofplantbreedingandgivestheseearlyinnovatorsfullstatusasplantbreeders.

    Coupledwithearlyhumanselectionwasastrongelementofnatural selection.Alloftheancestorstoourmoderncropspeciesweregrownwithoutsup-plementalnutrition,fertility,orpestprotectants,andwereconstantlyexposedtothechallengesoftheenvironment.Thebestofthefarmer-breed-ersspedthisprocessalongbyselectingfromthemostdesirableplantsfortheirfoodandfibervalue,usingtheseastheparentalstocksforsubsequentgenerations.Thisbalancebetweenhumanandnaturalselectionpersisteduntilthebeginningoftheindustrialagriculturaleraofthe20thcentury.

    Allagriculturalcropshavecontinuouslychangedwiththeselectionpressuresappliedfrombothfarmersandtheenvironmentsinwhichtheywerecultivated.Whenfarmersandtheenvironmentchanged,thecropchanged.Manyofourcropscon-tinuetochangeandadapttonewchallenges.

    Thenumberofenvironmentsthatcropsarebredin,andthenumberofbreeders,havebothshrunkoverthepast100years.Thisisbecausefarmersinmostagriculturalregionsoftheglobedolittleornoplantbreedinganymore,andthenumberofre-gionalizedseedcompaniesdoingbreedinghasalsodrasticallyshrunkinthepast40to60years.

    Withtheadventoftheindustrialagriculturalmodeltherehasbeenanarrowingofthesetoftraitsthatareconsideredcriticaltotheproductionofmostmoderncrops.Theindustrialmodelhasalsoled

    toagradualspecializationshiftinwhereandhowmanyofthemosteconomicallyimportantcropsaregrown.Inthepre-industrialagriculturalera,almostallcropswerespreadacrossmanyenvironmentalandgeographicnichestosatisfythediversityofagriculturalsettlements.Thismeantthattheywereconstantlyadaptingtoadiversityofclimaticcondi-tionsandtheever-changingneedsofthehumanag-riculturalsocietiesthatwereusingthesecrops.Thefarmersineachregionwerethebreeders.Somefarmerswereundoubtedlybetterbreedersthanothers.Therehasalwaysbeensomespecializationinmanyaspectsofagriculturalsocieties,butinthispre-industrialera,everyagriculturalcommunitycertainlyhadtheseedneededbythatcommunity.Seedofeachofthesecommunityvarietiesweregeneticallyuniquefromasimilarvarietyofthesamecropinthenextcommunity.Thispatchworkofuniquevarietiesisthenmultipliedthousandsoftimesacrosstheuniqueenvironmentsofagricul-turalcommunitiesaroundtheglobe.

    Cropgeneticdiversityinthepre-industrialerawasalsomuchgreaterthanwhatcurrentlyexistsinmodernagriculturebecausethediversitywithineachcommunityvarietywasfarricher.Farmersdidnotdemandthekindoftraituniformitythatisnowthenorminourmoderncrops.Thisallelic*di-versitywaspartofeverycroppopulationandwasreflectedbymanyuniquegenotypes(andpheno-types)ineachfield.

    TherewasagreatblossomingofseedcompaniesinNorthAmericaandEuropeinthesecondhalfofthe19thcentury.Withtheadventofseedasacommercialcommodity,andtheemergenceofseedcompanies,therewasanincentivetoproduceseedofcropvarietiesthatweremuchmoreuniformandpredictable.Theseearlyseedcompanieswereregionalinnature.Theyreliedonacentralstoreinametropolitancenterandsupplementedwithmailordersformuchoftheirbusiness.

    Theindustrializationofagricultureinthe20thcenturyledtoamuchhigherdegreeofspecializa-tion.Specificregionswithmoreidealclimatesforacertaincroptypeandprimeagriculturallandbe-camecentersoflarge-scaleproductionforcertaincrops.Inaddition,plantbreedingprogramsstarted

    I. Introduction

    *Definitionsfortechnicalwordsthatareitalicizedandboldedcanbefoundintheglossary

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org

    cateringtoindustrializedsystems,frommechanicalplanting,cultivation,andharvesting,totheneedsofthepackers,shippers,andretailershandlingthecroppost-harvest.

    The philosophy of organic plant breedingAsorganicplantbreeders,wehaveanopportunitytoshapeorganicagricultureinfundamentalways.Bymakingnewvarietiesavailabletoorganicfarm-ers,weareabletogivethesefarmersnewoptions.

    Buildingahealthy,sustainableagriculturefuturerequiresfarmer-centricseedsystemsatthere-gionallevel,wherefarmersandthecommunitiestheyserveconsciouslychoosewhichcropgeneticstheyuse,howtheyaremaintained,andhowthesegeneticsarecontrolled.

    Donewell,organicplantbreedingcanhelpensurethatfarmerscontroltheseedtheyuse.Breedingworkcangivefarmersfreeaccesstogeneticre-sourcesandthefreedomtogrowwhattheyplease,whilesupplyingthemwiththeknowledgeandskillstobothgrowseedandimproveacropschar-acteristicstobestmeettheirneeds.

    Overthepastseveraldecades,plantbreedinghasbecomeincreasinglyformalizedandcentral-ized.Breedingworkthatwasoncedonemainlyinfarmersfieldsisnowdonealmostexclusivelybylargeseedcompaniesandonstate-runagricul-turalexperimentstations.Mostbreedersworkingforlargeseedcompanies,aswellasmostpublicbreeders,focustheirattentionprimarilyonthelargestmarketsforseed,whicharetypicallylarge,conventionalfarmingsystemslocatedinprimeagriculturalregions.Thisfocusontheneedsoflarge-scale,conventionalagricultureleavesorganicfarmerswithoutvarietiesthatareadaptedtotheneedsoftheirsystems.Researchhasfoundthatthelackoforganicallyadaptedvarietiesleavesorganicfarmersatadisadvantage,resultingin,amongotherthings,loweryieldsthaniftheyhadvarietiesadaptedtotheirsystems.Organic,on-farmbreed-ingworkisdevotedtoservingtheneedsoftheseorganicfarmers.

    Our farming ancestors never stopped breedingAllgoodfarmerswhosurvivedandflourishedwere,bynecessity,plantbreeders.Theyusedob-

    servationalskillstodetermineandselectthebestadaptedplantseveryyear.Thismeantselectingthehighestyielding,besttasting,andmostdisease-re-sistantplants.Alloftheheirloomcropvarietiesweknowandenjoytodayweredevelopedoveralong,richperiodofourhistory,tellingthetaleofourplantbreedingancestorswhoweremakingselec-tionsandsavingseedfromthehealthiestandmostvigorousplantsyearafteryear.Thesefarmerswereconstantlystrivingtoimprovevarietiestobettersuittheirneeds,engaginginaconstantdanceofimprovementandco-evolutionwiththeirfood.

    Theseancestorswereneverfullysatisfiedwiththecropstheyhad.Theywantedtoimprovetheirfarmer-bredvarieties(sometimescalledlandra-ces),sinceitcouldmeanthedifferencebetweenlifeanddeathfortheirfamiliesandcommuni-ties.Modernorganicfarmerswhoarechoosingtoproduceandselectcropvarietiestoflourishintheirregionalagro-ecosystemsarerubbingshoul-derswiththebestfarmer-breedersofthepastwhodomesticatedandcontinuallyimprovedourcropgeneticresources.

    Therewillalwaysbeaneedtoadaptnewvariet-iestocurrentchallenges,suchasclimatechange.Classicalplantbreedingallowsustoselect,adapt,andcontinuallyco-evolvewithourfoodcrops.Thisformofplantbreedingisnottobeviewedasmessingwithnature,anditiscertainlynotgenet-icengineering.Theprinciplesofevolutionshowusthatthereisalwaysvariationinbiologicalpopula-tions,environmentalconditionsalwayschange,andnobiologicalentityeverstaysthesameinresponsetoitsenvironment.Inotherwords,noorganismiseverstaticinnature.Selectionpressureandchangeovertimeareinevitable.Weadvocateforanevolu-tionarybreedingmodelthatallowsthevarietiesweusetobepartofthesustainableagriculturalsys-temswearepioneering.

    Returning farmers to their role as seed stewardsOurworkatOrganicSeedAlliance(OSA)isintend-edtoempowerfarmerstoagainbepartoftheseedsystemthatmostfarmershavebeendivorcedfromforalmost100years.Overthelastcentury,wehavenotonlylostvaluablegeneticdiversitythroughmodernagriculturalpractices,wehavealsolostmuchofthefarmerknowledgeneededtosteward

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org

    thegeneticresourcesthataremostimportanttothemandtheircommunities.Webelievethereisjustasmuchimportantworktobedoneintrainingfarmerstobreed,maintain,andgrowhigh-qualityseedcropsadaptedtotheneedsoflow-input,agro-ecosystemsasthereisforpreservingthegeneticresourcesofthepast.Thebestwayforvaluablegeneticresourcestobecomepartofmodernag-ricultureistogetthemintothehandsoffarmers.Asskilledstewards,thesefarmerswillcontinuetoadoptandfurtheradaptthegeneticstotheirneeds.Itisthereforecriticalthatfarmersfirsthaveaccesstothesegeneticresources.

    Ourbreedingworkisdoneonagro-ecologicalfarmsbyapplyingsoundclassicalplantbreedingmethodology.Weemphasizebuildinggeneticresil-ienceanddiversityintoeverycroppopulation.Ourplantbreedingworkstrivestoestablishadiversi-fiedandecologically-basedagriculture,tofosterthemanagementofcropsgeneticsattheregionallevel,andtostrengthenaregionseconomy.OSApromotesthestewardshipofseedinamannerthatencouragestheadaptationofplantgeneticstotheecologicalandmarketneedsoffarmersandthecommunitiestheyfeed.Thisincludeshelpingfarmersperformon-farmplantbreedinginthesamemannerasourmostinnovativeancestors.Tonotencouragesuchimprovementswouldbeadisservicetofarmers,eaters,theplanet,andfuturegenerations.

    Thissectioncoversthebasicsofhowtocreateandimprovevarietiesforyourorganicfarm.Firstwedescribethebasicstepsofplantbreeding.Thenwedescribesomebasicgeneticprinciplesandhowtheyapplytoplantbreeding.

    Initsessence,plantbreedingcanbeboileddowntothefollowing:

    Select plants with superior genes from a genetically variable population of plants.

    Howyoudothiswilldependonyourgoals,thecropsyouareworkingon,andthetimeandre-

    sourcesyouwanttoinvest.

    Selection in theory and practiceSelectioninplantbreedingreferstotheprocessofchoosingwhichplantsproducethenextgenerationofapopulation.

    Therearetwomajortypesofselectionappropriateforon-farmbreeding:massselectionandfamilyselection.Therearebenefitsanddrawbackstoeachtypeofselection.Whetheryouuseoneortheother,oracombinationofboth,willdependonthecropyouareimprovingaswellasyourtimeframe,resources,andultimategoals.

    Withmass selection,youmakeselectionsbasedonhowindividualplantswithinapopulationperform.Massselectionisthemoststraightforwardbreed-ingmethod.Assuch,itrequiresminimumtime,labor,andrecordkeeping.Becauseofthesimplicity,itcanmakeworkingwithlargepopulationsmorepracticalduringthebreedingprocess.However,itdoeshaveafewdrawbackswhencomparedtofamilyselection.First,becauseyourselectionsarebasedsolelyontheplantsphenotype(itsoutwardappearance),youaremorelikelytochooseaplantbasedonenvironmentalinfluenceratherthanonsuperiorgenetics.Forexample,thesoilwherethatplantisgrowingmaybemorefertileincomparisontoaneighboringspotinthefield.Theseconddraw-backwithusingmassselectionisthepotentialthatundesirablerecessive genesmaybecarriedbytheplantsyouselectinahiddenheterozygousstate.

    Griddedselectionisonetechniquethatcanbeusedwhendoingmassselectiontohelpaccountfortheinfluenceoffieldvariationonplantpheno-types.Whenusinggriddedselection,imaginethefieldthatisplantedwithyourbreedingpopulationdividedintoagridofsmallersections.Forexample,thefieldcouldbedividedintoquarters.Whense-lecting,attempttoselectanequalnumberofplantsfromeachsection.Evenif,forexample,theplantsinthenorthwestcornerofthefieldlookinferioringeneraltotheotherquarters,selectthebestplantsfromthatsection.

    Incontrasttomassselection,withfamilyselec-tionyouarelookingattheoverallperformanceofagroupofrelatedplants(afamily)andusingthe

    II. Plant Breeding Basics

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org6

    Figure 2. This diagram exhibits the steps to execute family selection in three successive steps.

    familysoverallperformancetomakeselections.Infamilyselection,youplantaseriesofrowsofplants,whereeachrowcomesfromseedsavedfromasingleselectedplantfromthepreviousgeneration.Theadvantagesoffamilyselectionaretwo-fold.First,becauseyouaremakingselectionsbasedonawholerowofplantsthatarecloselyre-lated(siblingsorsibs),ormorethanonerowforeachfamilyifyouareusingreplication(seemore

    Key: Each star represents a selected plant

    Gridded Selection

    Figure 1. Gridded selection is one technique that can be used when doing mass selection to help account for the influ-ence of field variation on plant phenotypes.

    aboutthisconceptbelowinthetheorysection),itbecomeslesslikelythatyoumightselectplantsbasedontheinfluencesofverylocalizedenviron-mentaleffects.Second,becauseyouareabletoseemanycloselyrelatedplantsnexttoeachother,youaremorelikelytospotsomeplantsexhibitingdeleteriousrecessivetraitsthatmaybecarriedinagivenfamily.

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org7

    Figure2exhibitsthestepstoexecutefamilyselec-tioninthreesuccessivesteps.Instep1,eachoftherowsisafamilyinacornbreedingprogram,withthesamerowrepresentedlefttorightsequentiallyacrossthethreeselectionsteps.Step1involvesevaluatingeachfamilyasawholebasedontheper-formanceoftheentirerow.Step2istoeliminateanyfamiliesthatdonothaveahighproportionofsuperiorplants.Althoughtherearenohardandfastrulesinplantbreeding,whendecidingwhichfamiliestoselect,atOSAweeliminateafamilyunlessithasatleast60to70%acceptableplants.Onceyouhavedeterminedwhichfamiliestokeep,step3istoeliminatethepoorest30to40%oftheplantswithinthesefamilies.Step4showsthefinalfieldarrangement,withonlytheselectedplantsintheselectedfamiliesremaining.

    Thefamiliesthatareusedinfamilyselectioncancomeinvariousforms.Threeofthemostcommonformsarehalf-sibfamilies,full-sibfamilies,andS1families.Half-sib(HS)familiesarefamilieswherealloftheplantsshareacommonmotherbutmayhavedifferentfathers.Youwouldproduceahalf-sibfamilybysavingseedfromaplantthathasbeenallowedtoopenlypollinatewithmanyotherplants.Alloftheseedyousavefromthatplantsharesamother,buteachseedwaspotentiallycreatedbyfertilizationwithpollenfromdifferentfathers.Plantsinafull-sib(FS)familysharebothamotherandfather.Youwouldproduceafull-sibfamilybyconductingacontrolledpollinationbetweentwoplants.PlantsinanS1familyallcomefromaself-pollination.Inotherwords,theyallsharethesamesingleparent,whichactedasbothmotherandfather.Allthreeofthesefamilieshavetheiruses.Ingeneral,asyougofromfamilieswithlessrelat-ednesstofamilieswithmorerelatedness(HStoFStoS1),thefamiliesbecomemoreuniform.Thisuniformityallowsyoutobetterpredicthowtheir offspringwillperform.Inaddition,withS1familiesinparticular,youarebetterabletoseethetraitscontrolledbyrecessivegenesinthatfamily.How-ever,FSandS1familiesoftentakemoretimeandeffecttoproduce,andtheycreateagreaterriskforinbreedingincross-pollinatingspecies.

    Selectioncan,andshould,bedoneatmanypointsthroughouttheseason.Selectionfortraitssuchasgermination,vigor,plantstature,leafshape,color,

    andweedcompetitivenesscanoccurearlyintheseason.Byplantingatahighinitialdensity,someofthesetraitscanbeselectedforduringthethinningprocess.Manyvegetablecrops,suchascarrots,let-tuce,cabbage,onions,andkaleproduceaharvest-ablecropbeforepollination.Ontheotherhand,theharvestablecropofsomecroptypes,suchascorn,wintersquash,tomatoes,andmelons,canonlybeevaluatedafterpollinationoccurs.Itisdesirabletodoasmuchselectionaspossiblebeforepollina-tionoccurstoincreasetherateofprogressofyourbreedingwork.Anyselectiondoneafterpollinationislesseffectivethanselectiondonebeforepollina-tion.Thisisbecauseinthelatterscenariothepol-lenfrominferiorplantsisallowedtofertilizethesuperiorplants,thuspassingtheirgenesontothenextgeneration.

    Thefinalconcepttounderstandinselectionistheselection intensity.Inessence,theselectioninten-sityisbasedonthefractionofplants,orfamilies,youselectfromthestartingpopulation.Forexam-ple,youwouldbepracticingagreaterintensityofselectionifyouselectedonly50plantsoutof500tosaveseedfromthanifyouselected300plantsoutof500.Asyouincreasetheselectionintensity,youwillmakemoreprogressinyourbreeding.However,especiallyincross-pollinatedcrops,youneedtoavoidendingwithtoosmallofapopula-tion,whichcanincreasetheriskofinbreedingdepressionorlosingvaluablegeneticdiversitythatyoumayneedinthefuture.Ageneralguideontheminimumnumberofplantstokeepinyourpopu-lationcanbefoundinOSAsSeed Saving Guide for Gardeners and Farmers.Becauseitisimportanttoavoidgoingbelowtheseminimumpopulationsizerecommendations,thebestwaytoincreaseyourbreedingefficiencythroughincreasingtheselec-tionintensityistostartwithmoreplants.

    Therearetwobroadtermsthatconveytheinten-sityofselection.Thetermpositive selectionisusedwhenasubstantialfractionoftheplantsinapopulationarebeingremoved.Negative selection,alsoknownasroguing,referstotheremovalofonlyasmallfractionofthepopulation,usually10to20%orless,duringaseason.Manytimes,suchaswhenmaintainingavariety,negativeselectionmaybeallthatisneeded.Evenifyouareunsureofwhichplantsarebestinyourpopulation,youwill

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org8

    probablybeabletospotandremovethepoorestperformingplantsandtheplantsthatareobviousoff-types.Thisconceptisbestexpressedintheoldplantbreedingaxiom:IdontalwaysknowwhatIlike,butIalwaysknowwhatIdontlike!

    How to selectHowdoyouactuallypracticeselectioninthefield?Thereareanumberofmethodsforchoosingwhichplantswillproducethenextgenerationofthebreedingpopulation.Afewofthesemethodsaredescribedbelow.

    Eliminate inferior plants or entire families as you identify them.Thismethodcanbeeasilyusedwhenselectingfortraitsduringthinning,asmentionedabove,orwhenpracticingnegativeselection.Thismethodhastheadvantageofbeingverysimpleanddirect:cutdown,hoeout,orpullupanyplantsthatdonotmeetyourselectioncriteria.Themethodhasafewlimitations,however.First,ifyourbreed-ingpopulationhasrelativelyfewgoodplantsandrequiresahighselectionintensity,itcanbemoretime-consumingtoidentifytheinferiorplantsinsteadofidentifyingthefewsuperiorplants.Sec-ond,ifyouarelookingformultipletraitsatonce,andyoudontexpecttofindplantsthatperfectlyexhibitthesetraits,youmayneedamorerobustwayofselectingplantsbasedontheiroverallper-formanceacrossmultipletraits.

    Select superior plants or families as you see them. Thismethodistheconverseoftheonejustde-scribed.Typicallyyouwalkyourbreedingfieldonanumberofoccasionsthroughouttheseason,evalu-atingforoneormoretraitsatappropriatetimes.Toselectaplantorfamily,placeaflagorstakenexttotheplantorfamily.Youcanusedifferentcoloredflagsduringevaluationsfordifferenttraits,oryoucanmakenotesonastakeorinanotebooktodescribewhyyouareselectingcertainplantsorfamilies.Thismethodisusefulwhenfindingasmallpercentageofsuperiorplantsorfamiliesoutofalargerpopulation.Thismethodcanalsocutdownonextensiverecordkeepingwhenthechoicesofwhichplantsorfamiliestoselectareobvious.

    Select superior plants of families based on data. Thismethodinvolvesrecordingmeasurementsorscoresforeachplantorfamilyforthetraitsof

    interest.Thedataisthenusedtomakeselections.Thismethodisthemosttimeandlaborintensiveofthethreeandonlybecomespracticalwhendealingwithsmallernumbersofplantsorfamilies(lessthanafewhundred).However,thereareanumberofreasonstousethismethod.First,thismethodisusefulforidentifyingasmallnumberoffamiliesthatperformedbestbasedonarangeoftraitsthataremeasuredovertheseason.Forexample,sayyouwanttopickoutthebestkalefamiliesbasedonearlyvigor,leafcolor,resistancetoearlyaphidpressure,andlengthofharvestseason.Inthiscase,youmayneedtomakeyourselectionattheendoftheseason,butyouwillneeddetailedinformationabouteachfamilysperformancethroughoutthewholeseason.Second,thismethodhelpsincaseswherethetraitsarenotimmediatelyvisiblebutyouneedsomemeasurementstoevaluate.Thesemeasurementsmightincludeweighingyieldsormeasuringplantheights.Third,thismethodworkswellwhenyouhaveplantedyourfamiliesinmulti-plereplicationsandwanttofindtheaverageper-formanceofafamilyacrossallreplications.

    Whenrecordingdatatoguideselections,atraitcaneitherbemeasuredorscored.Tomeasureatrait,toolssuchasscalesoryardsticksareusedtogetavalue.Toscoreatrait,youassignavaluetoaplantorfamilybasedonaratingsystem.Forexample,youmightscoreleafsizeonascalefrom1to9,where1representsatinyleafand9representsalargeone.Whenscoringtraits,beginbywalkingtheentirefieldtoassesstherangeofexpressionwithinthepopulationforthetraitofinterest.Theplantsorfamilieswiththeworstorleastdesirableexpressionofthetraitwillbethe1onyourscaleandtheplantsorfamilieswiththemostdesir-ableexpressionwillbeyour9.Thenproceedtoscoreallplantsorfamiliesbasedonthatrange,attemptingtousethewholerangeofscoresfrom1to9.MoreinformationonhowtoconducttheseevaluationscanbefoundinOSAsOn-farm Variety Trials: A Guide for Organic Vegetable, Herb, and Flower Producers.Boththisguideanddatasheetsforevaluatingandselectingplantsareavailableforfreedownloadatwww.seedalliance.org.

    A crops mating system and how it affects plant breedingThewayaplantmatescanbedescribedasfalling

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org9

    alongaspectrumbetweenstrongly self-pollinating andstrongly cross-pollinating.Whereacropfallsonthisspectrumhasseveralimplicationsforwhichbreedingmethodwillbemosteffective.Thisideawillbediscussedindetailbelow.

    Self-pollinated cropsInplants,self-pollinationoccurswhenthespermofanindividualplantfertilizesanovuleofthesameplant.Essentially,theplantmateswithitself.Theplantsinbredoffspringreceivealloftheirgenesfromthisoneandonlyparent.Thus,theoffspringareverygeneticallysimilartothisparent.Allpre-dominatelyself-pollinatingplantshaveperfectorbisexualflowers,whereboththemaleandfemalesexualstructuresareborneineachflower.Self-pol-linatingplantsalsohaveflowersthatexcludecross-pollinationasdescribedbelow.

    Therearetwoimportantadvantagesofself-pol-lination.First,inplantsthathaveevolvedtobe-comeextremelywelladaptedtotheirenvironment,self-pollinationoffersawaytoensurethataplantsoffspringwillbejustaswelladaptedasitsparents.Inotherwords,self-pollinationisawayofreplicat-ingsuccess.Second,self-pollinationhelpstoensurereproductivesuccessunderavarietyofenviron-mentalcircumstances,asitdoesnotrequirethepresenceofwind,insects,oranimalstotransferpollenfromoneplanttoanother.

    Theplantsthatwecallstrongly self-pollinatingrelyalmostcompletelyonself-fertilizationtoreproduce.Theperfectflowersoftheseplantshaveevolvedwaysofexcludingpollenfromotherflowerstopreventcross-pollination.Forexam-ple,theflowersofmostmoderntomatovarietieshaveanthersthatformaconearoundthepistil,effectivelysealingofftheflowersstigmafromanypollenotherthanitsown.Inpeas,thepet-alsofthefloweraretypicallyclosedwhenthestigmafirstbecomesreceptivetopollen,givingpollenfromitsownanthersexclusiveaccesstothestigma.Whenthepeafloweropens,itisnearlyalwaysalreadyfertilized.Otherexamplesofstronglyself-pollinatingplantsincludecom-monbeans,wheat,andoats.

    However,evenastronglyself-pollinatingplantwilloccasionallycrosswithanotherplantofthe

    samespecies.Forexample,beescaneatthroughortearopenflowersanddepositpollenfromadifferentplantofthatspecies.Anumberofplantbreedersandseedgrowershavealsonotedthatwiththeincreaseofinsectactivitythatusuallyoccursinorganicallymanagedsystems,thereisoftenanincreaseinthepercentageofcrossesthatoccurinstronglyself-pollinatingplantspecies.Also,undercertainclimaticconditions,theflow-ersofsomeself-pollinatingplantswillopenearli-erthannormalandthestigmawillbereceptivetoreceivingpollenfromanotherplantbeforeithasachancetoself-fertilize.Andofcoursehumanscanperformcontrolledcrossesbetweenself-pol-linatingplantsbyintentionallytransferringpollenfromoneplanttoanother.

    Cross-pollinated cropsOntheotherendofthebiologicalspectrumarecross-pollinatedplants.Cross-pollinationoccurswhenthespermofoneplantfertilizesaneggofadifferentplantofthesamespecies.Inotherwords,aplantmateswithanotherplantofthesamespe-cies.Whentwoplantscross-pollinate,theyproduceoffspringthataregeneticallydifferentfromeitherparent.Ingeneral,allplantshavethepotentialtocross-pollinate,becauseeveryplantcanbefertil-izedbyanotherplantofthesamespecies.

    Thecentraladvantageofcross-pollinationisthatitfacilitatesaplantspeciesabilitytoadapttochang-ingenvironments.Plantspeciesthatconstantlymixtheirgenesintonewcombinationsincreasethelikelihoodthatatleastafewindividualswithinthespecieswillhavetherightcombinationofgenestowithstandnewenvironmentalchallenges.

    Breeding self-pollinated crops vs. breed-ing cross-pollinated cropsThematingsystemofthecropaffectsbreedingstrategiesinsomefundamentalways.Self-polli-natedcropsneedlessisolationdistancefromeachotherandothercompatiblespeciesthancross-pol-linatedcropstoavoidunintendedcross-pollination.Theneedforlessisolationmakesiteasiertoman-ageseveralsimultaneousbreedingprojectsofthesamespeciesinalimitedspace,andwillallowyoutogrowyourbreedingprojectnearacommercialcropwithoutthreatofcross-contamination.Itisgenerallyhardertomakecrossesbetweenself-pol-

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org0

    linatedplantsthancross-pollinatedones.Inordertoovercomeaself-pollinatedplantstendencytoonlyself-pollinate,somehandlaborisoftenre-quired,suchasremovingthemalepartsofoneplant(knownasemasculation)andcollectingandbringinginpollenfromanotherplant.Foron-farmbreedingprojects,cross-pollinatedcropscanbehandledwithsimpletechniqueslikestraincross-ing,explainedbelow.

    Thechallengeofcross-pollinationinself-pollinatedcropshasanimportantadditionaleffect.Popula-tionsofcross-pollinatedcropscaneasilycross-pollinateandrecombinetheirgenes.Ontheotherhand,afteraninitialcrossismadeinaself-polli-natedcrop,eachsubsequentgenerationofself-pol-linationrapidlyfixesgroupsofgenes.Seebelowforanillustrationofthisphenomenon.

    Doingbreedingworkwithself-pollinatedcropsallowsyoutoeasilycreateuniformvarietiesfromcrosses.Oncethesevarietiesaredeveloped,theyaremucheasiertomaintaincomparedtocross-pollinatedvarieties.

    Finally,self-pollinatedcropssufferlessfromin-breedingdepression.Thisallowsyoutoworkwithfewerplantsinaself-pollinatedbreedingprogramcomparedtoacross-pollinatedbreedingprogramandstillproduceavigorousvariety.

    Asfigure3shows,ifwefollowthegenerationsofplantsafteracrossismadeineitheraself-pol-

    Comparison of Attributes of Self- and Cross-pollinated Crops

    Self-pollinated crops Cross-pollinated crops

    Isolation Less More

    Crossing Harder Easier

    Self-pollinating Easier Harder

    Inbreeding depression Less More

    linatedspeciesorinacross-pollinatedspecies,fundamentaldifferencesariseinthestructureofthepopulations.Inacross-pollinatedspecies,theplantsineachgenerationareabletointermate.Intermatingcreatesaunifiedgene poolandallowsnewcombinationsbetweenplantstobemadeeachyear.Incontrast,unlesshumansintervenetomakeadditionalcrosses,self-pollinatedplantswillself-pollinateineachgenerationaftertheinitialcross.Self-pollinatingineachgenerationleadstoeachplantderivingseparatefamiliesorlines.Ineachgeneration,theplantsineachfamilywillbecomemoreuniformwithinthefamily,whilestrongerdifferenceswillarisebetweenfamilies.Thegeno-typesofeachfamilybecomefixed,andnewgeneticcombinationsbetweenfamiliesonlyoccurthroughhand-pollinationorthroughrarenaturalcross-pollination.Theconsequenceforbreedersisthat,whendealingwithself-pollinatedcrops,asufficientnumberoflinesmustbecreatedandmaintainedtofindalinethathasallofthetraitsofinterestfrombothparents.Incontrast,thegenepoolofcross-pollinatedcropscanbegraduallyimprovedbyeliminatingtheworstplantsorfamilies,andallow-ingthebesttocross-pollinatetocreatenewcom-binationsthatmaycontainthedesiredtraitsfrombothparents.

    Inbreeding depressionisalossofvigorduetoself-pollinationorthecrossingofgeneticallysimi-larplants.Essentially,inbreedingdepressionisthereverseofheterosis(hybridvigor).Inbreedingdepressionresultsfromtheaccumulationofho-

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org

    mozygouspairsofrareanddetrimental recessive alleles.Tosimplifytheidea,figure4servesasanexamplewherethereisararemutantrecessiveallele,a,inapopulation.Ifaplanthastwocop-iesofthisallele(aa),theplantcannoteffectivelyproducechlorophyll.Theplantappearsverylightgreenandgrowspoorly.Aslongasaplanthasonecopyofthefunctioningallele,A,itwillbeabletoproducechlorophyll.Inotherwords,plantswiththegenotypeAAorAaarehealthy.Asmallfractionofthepopulationcarriesthisalleleasaheterozy-gote(Aa).WhentheseAaplantscross-pollinate(ontheright-sideoffigure4),almostallofthepollentheyreceivewillcarrythedominantAallele,andtheiroffspringwillbehealthy,evenifhalfoftheoffspringarecarryingthedeleteriousrecessiveaallele.However,iftheseheterozygousAaplantsself-pollinate,oriftheypollinatewithcloserela-

    tives,thesituationchanges.Asyoucanseeontheleftinfigure4,whenoneoftheseplantsself-pol-linates,25%oftheiroffspringwillbehomozygousaaandwillshowthedetrimentaleffectsfromhav-ingtwocopiesofthisallele.Mostcross-pollinatingplantscarrymanydeleteriousrecessivealleles.Theseplantsrelyoncross-pollinationamonglargepopulationstoavoidinbreeding.Withself-pollinat-ingplants,ontheotherhand,selectionhasactedovermanygenerationstoweedouttheplantswithdeleteriousalleles.

    Figure 3. This diagram show how populations of self-pollinated and cross-pollinated crops differ in their genetic structures.

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org2

    Thinking about your target environmentInaperfectworld,plantbreederscoulddevelopva-rietiesthatthrivedinanysituationandanywhereontheplanet.Infact,mostlargeseedcompaniesstrivetoproducevarietiesthatcanbegrownacrosswidegeographicregionsinmanyagriculturalsitua-tions.Thisconceptisknownasbroad adaptation.However,varietiesthatarebroadlyadaptedmaynotbethebestvarietyforanygivenarea,andtheymayrequirehighlevelsofinputstorecreatethegrowingenvironmentunderwhichtheywerebred.Creatingbroadlyadaptedvarietiesmayalsore-quireextensivetestingthatmostsmallerbreedingprogramsdonothavetheresourcestoundertake.Incontrasttobreedingbroadlyadaptedvarieties,specifically adaptedvarietiesarebredtoexcelinspecificenvironments.Thesevarietiesarenotex-pectedtodowelleverywhere,justinthetargeten-vironment.Theconceptoftherelativeperformance

    ofavarietychangingbasedontheenvironmentitisgrowinginiscalledthegenotype by environ-ment interaction(GxE).

    Oneofthefirststepsinplanningabreedingproj-ectistoidentifythetargetenvironment.Theenvi-ronmentcanbeasbroadornarrowasyouchoose,keepinginmindthatadditionalresourceswillberequiredasthebreadthoftheprojectincreases.Developingabroadlyadaptedvarietywillrequiremoretimeandresourcesthandevelopingavari-etyspecificallyadaptedtoyourlocalconditionsandchallenges.

    Whendetailingthetargetenvironmentwhereyouwantyourvarietytoexcel,thereareseveralimpor-tantfactorstoconsider.Someoftheseare:

    Geographicregion(daylengthandfrost-freeperiod)Temperatureregime(highs,lows,fluctuations,andcumulativeheatunits)Rainfall(amountandtiming)

    Figure 4. Inbreeding depression results from the accumulation of homozygous pairs of rare and detrimental recessive alleles. This diagram serves as an example where there is a rare mutant recessive allele, a, in a population.

    III. Developing a Plant Breeding Plan

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org3

    SoiltypeDiseasepressuresWeedpressures

    Youshouldalsoconsiderculturalfactorsinyourtargetenvironment,suchas:

    Croppingsystem(directseedingvs.transplanting)Irrigationsystem(overhead,drip,orflood/furrow)Fertilization(typeandfrequency)Cultivationtechniques(frequencyofplowing,mechanicalvs.handtechniques)

    Finally,considerthemarketandsocialfactorsthatareintegraltoyourtargetenvironment,includingexpectationsfor:

    ColorStoragepropertiesFlavorCulinaryusesQualityofproductTransportationability

    Determining traitsKnowingwhatyouwantiscriticaltogettingwhatyouwant.Themostsuccessfulbreedingprojectshaveclearlydefinedtraitsthatarebeingselectedfororagainst.Priortostartingthebreedingeffort,itisimportanttotakethebroadlydefinedgoalsforthevarietyanddigdowntodefineexactlywhatyouwillbelookingforinthefield.Itmaybeveryeasytocomeupwithalistofthetraitsthroughbrain-storming.Anotherwaytounderstandwhichtraitsareimportantisbydiscussingthebestcurrentlyavailablevarieties.Whatarethetraitsthatmakecertainvarietiessogreat?Whataretheylacking?Ausefulconceptisthatofan ideotype.Anideotypeisessentiallyanidealizedcropvarietythatisper-fectlysuitedtoaparticularsetofproductionandmarketcriteria.Whatwouldthisidealvarietylooklike?Howwoulditperform?Whattraitswoulditneedtohavetobesuccessful?Whatarethetraitsthatareimportanttothefarmerandmarketplace?

    Someexamplesoftraitsthatyoumaywanttose-lectforinclude:

    PlantheightPlantstature

    DaystomaturityHarvestableyieldColorFlavorTextureStoragelifeSeedlingvigorPestResistanceDiseaseResistance

    Prioritizing traitsOnceyouhavecompiledalistoftraits,thenextstepistoprioritizethem.Ingeneral,abreedingprojectshouldnotattempttoactivelyselectformorethanfiveorsixtraitsatanygiventime.Theremaybemorethanfivetraitsthatareimportant,butitsnotfeasibletoworkoneverythingallatonce.Alsoconsiderthatsometraitsmaybeeasilyfixed,orlockedin,bychoosingparentsthatsharethesetraits.Inthiscase,effectiveselectionmaybeachievedsimplybythrowingoutacoupleofoff-typeplantshereandthere.

    Toprioritizethetraitsandlimitthelisttofive,twoquestionstoconsiderare:

    Isthetraiteasilymeasurable?Howheritableisthetrait?

    How can the traits be measured?Plantbreedingisanexerciseinbalancingwhatwewantwiththeresourceswehavetodothework.Mostplantbreedingprojectshavelimitedresources,soitmakessensetoavoidbreedingfortraitsthatrequireexcessivelabor.Itmightbewonderfultode-velopalettucevarietythatsupportsadiversefloraofmicroorganismsarounditsroots,butunlessyourprogramhastheresourcestodeterminetheidentityofawidevarietyofspeciesofsoilmicrobesinhabit-inghundredsoflettucelines,youwillneedtoworkwithtraitsthatareeasiertomeasureandselect.Intheprocessofprioritizingselectioncriteria,consid-erwhenandhowthetrait(s)willbemeasured.Thisplandoesntneedtobefleshedout,butitshouldbesufficientlydetailedsothatyouhaveanideaofhowmuchworkwillberequired.Willyoubeabletovisuallyratethetraitona1to9scale?Willyoubeabletoquicklymeasurethetraitwithayardstickorascale?Howmanytimesthroughouttheseasonwillthetraitneedtobemeasured?

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org

    How easily can the traits be inherited? Theseconditemtoconsiderwhendeterminingwhichtraitstoprioritizeinyourbreedingproj-ectistheheritability ofthetraits.Simplystated,heritabilityexpressesthelikelihoodthatatraityouseeinaplantthisyearwillshowupintheplantsoffspring.Putanotherway,heritabilityexpresseswhatfractionoftheappearanceofatraitcomesfromtheplantsgeneticsasopposedtotheenvironment.Althougheachcropandsituationisdifferent,traitsthatarehighlyheritableoftenincludesuchqualitiesascolor,shape,andspinesversusspinelessness.Themorehighlyheritableatraitis,theeasieritistoselectandseerapidprog-ressfromonegenerationtothenext.Traitswithlowheritabilitytypicallyincludeyieldandvariousformsofstresstolerance(i.e.,drought,salinity,andheattolerance).Thesetraitsrequiremorebreedingexpertise,aswellasamorefinelytunedexperi-mentalfieldmodeltoteaseoutsubtledifferencesinthesetraitstomakeimprovementsacrosscyclesofselection.Werecommendthatyouchooseonlyacoupleoftraitswithlowheritabilitytoselectforinyourbreedingproject,especiallyatthebeginning.Manytraits,suchasflavor,plantheight,anddiseaseresistance,maybeeitherhighlyheritableorlessheritable,dependingonthepopulation.Gettingadvicefromaformallytrainedbreederwhohasexperiencewithyourcropofinterestcanhelpyoudefinewhichtraitsaremoreorlessheritable.

    Choosing parentsGermplasmisageneraltermthatreferstothecollectionofgeneticresourcesforaplantspecies.Varieties,landraces,collections,breedinglines,andunimprovedmaterialarealltypesofgerm-plasm.Essentiallygermplasmisanyformoflivingtissue(e.g.,seeds,cuttings,roots,andtubers)fromwhichplantscanbegrown.Germplasmistheparentalmaterialusedtobeginyourbreed-ingwork.Sourcingqualitygermplasmdetermineshoweasilyandrapidlyyouwillbeabletodevelopyourdesiredvariety.

    Therearetwomainfactorstoconsiderwhenchoosinggermplasm:

    Quality:thegermplasmshouldincludeelementsoftheideotypeyouarestrivingtoproduce

    Variability:thegermplasmshouldhavevaria-tionforthetrait(s)youwanttoimprove

    Considervarietiesthat:

    PerformwellinyourtargetenvironmentAreconsideredtobecommercialstandardsContainuniquetraitsthatyouwanttoincorpo-rateintoyourproject

    Manyvaluablevarietiesforbreedingworkcanbefoundinthecatalogsofdomesticandinternationalseedcompanies.However,beawareoflegalpro-tectionsonsomevarietiesthatrestricttheiruseforbreedingorlimitwhatyoucandowiththeminotherways.Theseprotectionsincludeplantpat-ents,utilitypatents,plantvarietyprotectioncertifi-cates,andlicenses.YoucanfindmoreinformationabouttheseprotectionsoneOrganicsintellectualpropertyprotectionpageathttp://www.extension.org/pages/18449.Alwaysinvestigateyourgerm-plasmsourcestoensuretherearenolegalrestric-tionsonbreeding.

    Besidesseedcompanies,youcanalsosourcegermplasmfrom:

    FarmersSeedexchangesGermplasmResourceInformationNetwork:http://www.ars-grin.gov/

    Unlessyouarealreadyfamiliarwiththegermplasmyouwanttouseinyourproject,itiswisetocon-ductvarietytrialstoevaluatetheavailablesourcesofgermplasm.MoreinformationonhowtoconductvarietytrialscanbefoundinOSAsOn-farm Vari-ety Trials: A Guide for Organic Vegetable, Herb, and Flower Producers.

    Herearefourcategoriesofgermplasmthatareusefultoevaluateandconsiderforuseinorganicplantbreeding:

    1) Heirloom varieties:Thiscategoryincludesfamilyorcommunity-basedheirloomsaswellasnon-hybridcommercialcropvarietiesthatwereinexistencebeforeWorldWarII.Heirloomsareopen-pollinated(OP)varietiesthatwereselected,adapted,andmaintainedovergenerationswithin

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org

    familiesoragriculturalcommunities.Theywereoftenuniquelyadaptedtothechallengesofthesoil,climate,andendemicdiseasesofparticularregions.Inmanycases,thecommercialvarietiesreleasedduringthisperiodbyregionalseedcompanieswereimprovedorrefinedheirloomvarietiesal-readyimportanttothatparticularregion.ThesecommercialOPvarietieswereeitherdevelopedbyoneofthemanyregionalseedcompaniesthatflourishedduringtheearly20thcenturyorbyplantbreedersatoneofthelandgrantuniversities.Thesevarietieswerethenmadewidelyavailabletofarmersandgardenersacrosstheregion.

    2) Early modern varieties:Thisisatermforcropvarietiesdevelopedbybothregionalseedcompa-niesandlandgrantuniversitiesfromtheendofWorldWarIIuntilabout1980.Mostofthesevari-etiesarenotF1hybridsbutopen-pollinatedvariet-iesincross-pollinatedcropsandpurelinevariet-iesinself-pollinatedcrops.Earlymodernvarietieswerebredfortheconditionsofspecificregions.Thebreedingofvegetablesandotherminorcropswasntusuallydoneunderoptimumfertilityorheavysprays.Manyofthesevarietiesthatarestillusedtodayhaveproventheirworthasworkhorsevarieties,andanumberoffarmer-breedershavechosenvarietiesoutofthiscategorytoserveasgermplasmfortheirbreedingprojects.

    3) F1 hybrid varieties:Hybridvarietieswerefirstextensivelydevelopedincornduringthe1920sandbegantorepresentalargepartofthevarietyofferingsofmanycross-pollinatedcropsafterWorldWarII.Hybridsbecamepopularfor

    anumberofreasons.Theyareaneffectivewaytocreategeneticallyuniform,vigorousvariet-ies.Inaddition,theyrepresentaformofbiologi-calintellectualpropertyprotectionfortheseedcompaniesthatdevelopthembecausetheoff-springofhybridsdonotgrowtrue-to-type(duetotheirhighlyheterozygousnature).BecausemanybreedingeffortssinceWorldWarIIhavefocusedondevelopinghybridvarieties,alotofthemostelitecropvarietiescurrentlyarehybrids.Hybridsareoftengeneticallydiverseandcanrepresentgoodgermplasmformanybreedingprojects.

    4) Modern open-pollinated varieties (non-hybrids):Variousseedcompaniesarebreedinggoodopen-pollinatedvarieties(generallysmall-tomedium-sizedregionalcompanies)asaresomepublicbreedersatlandgrantuniversities.Thebestmodernopen-pollinatedvarietiesprovideexcellentgermplasmforbreedingprograms.

    Beforebeginningabreedingproject,itisagoodideatoscreencropvarietiesfromawideswathofgerm-plasmsources,includingsomefromeachofthefourcategoriesabove.Thisprocesswillhelpyouidentifythehighestqualitygermplasmcurrentlyavailable.Thepotentialforincreasingyoursuccessandde-creasingyourfrustrationisgreatestwhenyoubeginaprojectwiththebestpossiblegermplasm.

    Creating a breeding timelineFigure5onthefollowingpageisasummaryofageneralbreedingtimeline.Afterthesummary,wewillgothroughthestepsinmoredetail.

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org6

    Step 1.Conductvarietytrialstoidentifythebestpotentialparentalgermplasm.

    Step 2. Ifnecessary,makecrossesbetweenparents.Ifadequatevariationexistsinafavorableexistingvariety,thennocrossesarenecessary.

    Step 3.Ifyouareworkingwithacross-pollinatingcropandyoumadecrosses,allowtheoffspringtorandomlymateforatleasttwoorthreegenera-tionswhileremovingobviousundesirableplants.Inthecaseofself-pollinatedcrops,allowtheplantstonaturallyself-pollinateforatleasttwoorthreegenerationsafterthecross,removingonlytrulydysfunctionalplants.

    At this point, you can conduct a breeding program purely based on mass selection by proceeding to step 8. To conduct family selection, first go through steps 4 through 7.

    Step 4.Growalargepopulationandsaveseedseparatelyfromindividualplants.Theseedfromindividualplantsbecomefamilies.

    Step 5. Thefollowinggeneration,growfamiliesoutinrows.Planteachrowfromseedharvestedfromanindividualplant.Evaluateandselectthebestperformingfamiliesandplants.

    Step 6. Harvestseedfromtheselectedplantswith-intheselectedfamilies.Dependingonwhatyouseeandyourgoals,eitherkeepseedfromindividualplantsseparateorcombinewithinfamilies.

    Step 7. Repeattheprocessofgrowingandselectingfamilyrowsuntilyouaresatisfiedwiththeperfor-manceofallfamilies.

    Step 8.Oncetheselectedfamiliesareallaccept-able,combinetheseedofthesefamiliesintoabulkpopulation.

    Step 9.Growthebulkedpopulationagainandre-movetheleastdesirableplants.Repeatasnecessary.

    Step 10.Oncethepopulationisuniformlyaccept-able,increaseseedforrelease.

    Figure 5. This diagram shows a summary of a general breeding timeline.

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org7

    Herearefurtherdetailsoneachstep:

    Step 1. Conduct variety trials to identify the best potential parental germplasm.

    Beforecommittingtoalong-termbreedingproj-ect,carefullyevaluatepotentialparentalgerm-plasm.Thisevaluationtypicallytakestheformofavarietytrial.Avarietytrialisasystematicwaytoevaluateasetofvarietiesusinggoodtrialdesign.Thepurposeofatrialdesignistohelpminimizetheenvironmentalvariationtoensurethatthedif-ferencesobservedrepresentgeneticdifferencesbetweenplants.Moreinformationonhowtocon-ductvarietytrialscanbefoundinOSAsOn-farm Variety Trials: A Guide for Organic Vegetable, Herb, and Flower Producers.

    Step 2. If necessary, make crosses between parents. If adequate variation is present in a favorable existing variety, then no crosses are necessary.

    Toimproveyourpopulation,somedegreeofvaria-tionforimportanttraitsisnecessary.SeeSectionIV. Theories of field-based organic breedingbelowformoreinformationabouthowvariationworks.Tohaveavariablepopulation,youneedtoeitherstartwithgermplasmthatisvariableforthetraitsyouintendtoimprove,oryouneedtocreatevariationbycrossingtwoormorevarietiestogeth-er.Thesecrossescanbemadeinanumberofways,dependingonyourgoalsandresources.

    Types of Crosses

    Controlled pollinations:Plantbreedingmightevokeanimageofbreedersmakingcontrolledcrosses,perhapswithtweezers,magnifyingglasses,andsmallpaintbrushes.Thegeneralstrategywilldependonwhetherthespeciesyouworkwithhasperfectflowersornot.Per-fectflowershavebothmalestamenandfemalepistilsonthesameflower.Iftheplanthasper-fectflowers,youwillgenerallyneedtoemas-culatethefemaleparentbyremovingthemalestamenandthentransferpollenfromthemaleparent.Iftheplanthasonlyunisexualflow-ersflowersthatcontaineitherallmaleorall

    femalepartsthereisnoneedtoemasculate.Youjusttransferpollenfromamaleflowertoafemaleflower.Afterpollinating,somesortofcoverisoftenusedtoexcludeotherpollenfromcontaminatingthecross.

    Open pollinations:Thesecrossesarealsoknowasstrain crossesorblindcrosses,andaredonebyallowingoneplantorasetofplantstocrosswithanotherplantorsetofplantswithoutmakinganattempttoensurethatpollenisonlytransferredbetweencertainplants.Thisisaccomplishedbybringingalloftheplantstogetherandallowingthemtofreelycross-pollinate.Outsidepollenisexcludedeitherbyisolationdistanceorbycagingtheplantsinsideastructurethatpreventsinsectsandpollenfromentering.

    Hybrids:Thesevarietiescanbeusedaspre-madecrosses,savingastepinthebreedingpro-cess.Keepinmind,however,thatmosthybridsaremadebycrossingtwonarrowlyselectedin-bredparentsandthereforeapopulationformedfromasinglehybridwillhavecomparativelylittlegeneticdiversityrelativetoonecreatedfromastrainorblindcrossbetweenopen-pol-linatedvarieties.

    Step 3. If you are working with a cross-pol-linating crop and you made crosses, allow the offspring to randomly mate for at least two or three generations while removing obvious un-desirable plants. In the case of self-pollinated crops, allow the plants to naturally self-polli-nate for at least two or three generations after the cross while removing undesirable plants.

    Priortoselection,growyournewbreedingpopula-tionforafewgenerations,allowingtheplantstoflowerandfreelycross-pollinate.Duringthistime,youmaypracticenegativemass selectiontorogueouttheworstperformingplants.

    Itistemptingtobeginintensiveselectioninthegenerationafteryourcrossesaremade,asplantbreedingcanbesuchalongprocess.However,werecommendthatyoudonotskipthisstep.Incross-pollinatedcrops,allowingafewgenerationsofran-

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org8

    dommatingbreaksupthelinkagebetweengenes.Simplystated,linkagereferstohowgenesinaplantwilltendtotraveltogether.Forexample,ifyoucrossatallplantwithdarkleaveswithashortplantwithlightleaves,youremorelikelytogeteitheratalldarkplantorashortlightplantthanyouaretogetatall,lightplantorshort,darkplant.Randommatingforacouplegenerationsallowsthegenestoshuffleandmakesitmorelikelythatyouwillseenewcom-binationsofgenesinthepopulation.

    Forself-pollinatedcrops,itisalsovaluabletowaitforatleasttwoorthreegenerationsbeforemak-inganysignificantselections.Thereasonforthisisdifferentthanforcross-pollinatedcrops.Asde-scribedinthepreviousdiscussionofdifferencesingeneticstructuresbetweencross-andself-pol-linatedcrops,randommatingandthebreakingoflinkagesdoesnotoccurinself-pollinatedcrops.Instead,eachplantderivedfromtheoffspringofacrosshasthepotentialtobecomeaseparate,independentfamilylinewithuniquecombina-tionsoffixedpairsofgenes.Aseachoftheselinescontinuestoself-pollinateoverthecourseofafewgenerations,eachlinewillbecomemoreuniformandthedifferencesbetweenlineswillbecomemoreobvious.Itwillbeeasiertomakedistinctionsbetweenfamiliesifsufficienttimehaspassedtoallowthefamiliestoessentiallybecomefixedanddistinctfromoneanother.

    Additionally,themoregenerationsyouspendimprovingthepopulationbythisgradualprocess(massselection),thebetterthegeneticsofthepopulationwillbe,andthemorelikelyyouwillbesuccessfulinthelaterphasesofbreeding.

    Thesegenerationswillalsogiveyouachancetogettoknowthepopulation,toseewhatkindsofvaria-tionexistinit,andtoevaluatewhatitsstrengthsandflawsare.Usetheseyearstorefineyourgoals,traitsofinterest,andevaluationtechniques.Re-member,whenpracticingselection,makesurethefieldlocationrepresentsthelocationyouwanttoselectforandisrelativelyuniformthroughout.

    Step 4. Grow a large population and save seed separately from individual plants. The seed from individual plants become families.

    Duringthisgenerationyouwillbeselectingyourbestplantstoserveasparentsforfamilies.Evalu-ateyourplantsmultipletimesthroughouttheseason,consideringthegoalsandtraitsyouhaveprioritized.Sincetheremaybesomevariationinthesoilquality,weedpressures,andotherfactorsinyourfield,selectthebestplantsfromallpartsofthefieldequally.Ifalloftheplantsinonecor-nerofyourfieldlookworseoffthantherest,stilltrytopicksomeofthebestlookingplantsinthatpatchthoseareplantsthatcansurviveundersomeenvironmentalstressorpotentiallysub-op-timalconditions.

    Ifyourcropiscross-pollinatingandisacropwhereyouhaveanopportunitytoevaluateandselectthebestplantsbeforepollinationoccurs,youwillmakefasterprogressbyremovingallinferiorplantsfromthefieldbeforepollinationoccurs.Inthisway,anycross-pollinationthatoccurswillonlybebetweentheselectedplants.

    Whenyouharvestseedfromselectedplants,putseedfromeachplantintoaseparatebag.Theseedineachbagrepresentsafamily.Labeleachfamilywithauniquenumber.Alongwiththenumberoneachlabel,youmaywanttomakenotesaboutwhyyouselectedthatplant.Theseplantswillrepresentfamiliesforthenextgenerationsroundofselec-tion.Inthecaseofcross-pollinatedcrops,trytosaveseedfromatleast20to50individualplants.Inthecaseofself-pollinatingcrops,saveseedfromatleast30to100plants.

    Step 5. The following generation, grow family plots by planting each separate plot with seed harvested from an individual plant. Evaluate and select the best performing families.

    Thisgenerationyouwillbelookingattheprog-enyofyourselectedplantsasfamilies.Inbreed-ingterms,thisistheprocessoffamily selection.Thismeansyouwillplantseedfromeachplantthatyouselectedinseparateplotsorrows,justasifyouwereconductingatrial,followingthebestpracticesoftrialdesignasoutlinedinOSAsOn-farm Variety Trials: A Guide for Organic Vegetable, Herb, and Flower Producers.

    Examineeachplotasawhole.Howdoesthisfamily

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org9

    ofrelatedplantslook?Onaverage,whichfamilieslookbestgivenyourgoalsandtraits?Selectthebestfamiliesbasedontheiroverallquality,andeliminatetheotherfamilies,preferablybeforetheyhaveachancetocross-pollinatewiththeselectedfamilies.Lookforfamilieswhereatleast60to80%oftheplantsareacceptable.

    Next,selectthebestplantswithinthebestfami-lies.Aimtoeliminate30to40%oftheplantswithinafamily,keepinginmindtheminimumpopulationsizeyouwanttomaintaintokeepthepopulationhealthy(seebox,Whymaintainlargepopulationsizes?).

    Step 6. Harvest seed from the best plants within the best families.

    Atthispoint,savetheseedfromeachfamily.Inotherwords,combinetheseedfromallofthese-lectedplantswithineachfamilysothatyouendupwithaseparatebagforeachfamily.

    Why maintain large population sizes?

    Whatistheadvantageofmaintaininglargenumbersofplantsinyourbreedingpopula-tions?Afterall,ifyouhaveafewexcellentlook-ingplantsinthefield,whysaveseedfromanyplantsthatdontlookasnice?Therearetwomainreasonstomaintainlargepopulations.

    Inthecaseofcross-pollinatingcrops,largepopulationshelpavoidinbreedingdepression.

    Inbreedingdepressionisalossofvigorduetothecrossingofgeneticallysimilarplants.Plantssufferingfrominbreedingdepressiongerminatepoorly,yieldpoorly,andsuccumbmorequicklytoenvironmentalstress.Ingen-eral,plantsthatarestronglycross-pollinatingaremostsusceptibletoinbreedingdepression,whilethosethatarestronglyself-pollinatingareleastsusceptible.

    Thesecondreasontomaintainlargepopula-tionsizesistoavoidlosingimportantgenesthroughgenetic drift.Geneticdriftistherandomchangeinthefrequencyofgenesinthepopulationasitreproduces.Evenintheabsenceofselection,thepopulationthatgrowsfromonegenerationtothenextmayhaveaslightlydifferentpercentageofgenesjustbychance.If,bychance,noneoftheplantsinagivengenerationcarryacertaingene,itwillbelosttothepopulation.Thisismuchmorelikelytohappeninsmallpopulationsthaninlargeones.Thethreechartsbelowshowasimula-tionofthechangeingenefrequencyafter25generationsforsomehypotheticalgenesthatarefoundinbetween20to80%oftheplantsinapopulationtostartwith.Whenthepopula-tionsizeiskeptat200plants,allofthegenesarestillpresentinthepopulationattheendof25generations.Withonly30plants,one-fifthofthegenesarefixedorlostwithin25genera-tions.Whenseedissavedfromonly10plantseverygeneration,themajorityofgenesarefixedorlostwithin25generations.

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org20

    Step 7. Repeat the process of growing and se-lecting family plots until you are satisfied with the performance of all selected families.

    Thefollowinggeneration,continuetoplanttheselectedfamiliesinseparateplotsandevaluatethemagainthroughouttheseason.If,duringthesecondroundofevaluations,someofthefamiliesnolongerseemacceptable,discardthosefami-lies.Alsoeliminateallinferiorplantswithintheselectedfamiliesthatyoudecidetocarryontothenextgeneration.

    Step 8. Once the selected families are all accept-able, combine the seed of these families togeth-er into a bulk population.

    Forthesepastfewgenerations,youhavebeenmaintainingthepopulationasasetofseparatefamilies.Oncealloftheremainingfamiliesmeet

    yourgoals,youcancombinethemtogethertore-formthepopulation.

    Step 9. Grow the population again and remove the least desirable plants. Repeat as necessary.

    Oncethefamilieshavebeenreformedintoasinglepopulation,youcanmaintainthispopulationinsubsequentgenerationsthroughmassselection(i.e.,selectingthebestindividualplantstorepro-duceeachgeneration).Ifatanypointyoubelievethevarietyneedstobemoreintensivelyselected,youcancompleteanothercycleoffamilyselectionbysavingseedseparatelyfromindividualplantsandplantingtheminindividualplotsthefollowinggeneration,aspreviouslydescribed.

    Step 10. Once the population is uniformly ac-ceptable, increase seed for release.

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org2

    Asyouincreasetheseed,continuetorogueoutanyundesirableplantsastheyarise.

    Lookingatbroccoliseedandkaleseed,youmaynotbeabletotellthemapart.Theyarebothsmall,brownish-greyseeds,essentiallyindistinguishablefromoneanother.Butplanttheseedintheground,andastheygerminateandgrowintomatureplants,thedifferencesbecomegreater.Thekaleseedpro-duceskaleplantswithbroad,curlygreenleaves,whilethebroccoliseedproducesbroccoliplantswithlarge,tightflowerheads.Whatexactlyiscontainedinthatseedthatcausesbroccoliseedtogrowintobroccoliandkaleseedtogrowintokale?Genes,madeupofDNA,programthewayplants(andalllivingthings)grow,thewaytheylook,andthewaytheytaste,amongmanyothercharacteris-tics.Genesarehowaplantsprogrammingpassesdownfromonegenerationtothenext.Asplantbreeders,ourjobistopicktheplantswiththebestgenesandmakesurethosegenesgetpasseddown.

    Thissectionwillintroduceyoutothebasicsofgeneticsbyexplaininghowgenesarepassedfromonegenerationtothenextandhowthosegenesde-terminetheappearanceandperformanceofplants.

    How genes travel from parents to offspringJustasitisinhumans,DNAisbundledtogetherintopackagescalledchromosomesinplants.Plantsalsohavemultiplechromosomes,eachofwhichcontainsmanygenes.Mostcropspeciesarediploid,meaningtheyhavetwosetsofchromosomes:onesetfromtheirmotherandonefromtheirfather.Forexample,tomatoes,ingeneral,have12pairsofchromosomes,withtwo#1chromosomes,two#2chromosomes,etc.,foratotalof24chromosomes.However,someplantsarepolyploidy,meaningtheyhavemorethantwosetsofchromosomes.Forexample,commonwheatisahexaploid,meaningithassixsetsofchro-mosomes.Althoughpolyploidyisfoundinanumberofimportantcropplants,forthesakeofsimplicity,wewillfocusonlyondiploidspeciesforthefollow-ingexplanationsandexamples.

    Toreproduce,plantsproducespermandeggs.Spermandeggshaveonlyonesetofchromosomesratherthantwo.Again,lookingattomatoes,whiletherestoftheplanthas24chromosomes,thespermandeggsonlyhave12.Wheneachspermoreggiscreated,itreceives,atrandom,eitherachromosomefromtheplantsmotherorachromosomefromtheplantsfather.Thismakeseachspermandeggunique.Later,whenthespermiscarriedonthepollenandfertilizesanegg,auniquenewplantwillbecreated.Becauseofthis,plantscreatedfromthesamecrossbetweentwoplantscanalllookdifferent.Theexceptiontothisisiftheparentplantsarehighlygeneticallyuniform,whichcanoccureitherthroughinbreedingacross-pollinatedspeciesortheuseofpurelineself-pollinat-ingvarieties.Inthiscasetheoffspringfromthecrosswillalllooknearlyidentical.

    How genes determine the appearance and per-formance of plantsAswelearnmoreaboutthecomplexityoflife,weseemoredeviationsandexceptionstohowweoriginallyunderstoodgenestofunction.Neverthe-less,althoughthefollowingdescriptionsaresimpli-fiedversionsofhighlycomplexsystems,theyarestillrelevantandusefulforplantbreedingwork.

    Genesareinstructionsforourcells,writteninthelanguageofthenucleicacidsthatmakeupDNA.Cellsreadtheseinstructionsandmakeproteins.Typi-cally,onegenemakesoneprotein.Eachchromosomecontainshundredsorthousandsofgenesmostoftheinstructionsnecessarytocreateaplant.

    Asmentioned,halfofthosegeneswillcomefromthemother(maternal)andhalffromthefather(paternal).Foranygiventraitthatiscontrolledbyonegene,suchasgrowthhabitintomatoes,aplantwillhaveamaternalgeneandapaternalgene.Whatifthematernalandpaternalgenesdif-fer?Forexample,whatifthegenereceivedfromthemotherproducesproteinsthatmakeashorter,determinantplant,whilethegenefromthefathermakesalarger,indeterminateplant?Formanytraitstherearegenesthatresultinproperlyfunc-tioningproteinsandgenesthatresultindefective,non-functioningproteins.Also,formanytraits,oneproperlyfunctioninggeneisenoughtocreatesuf-ficientproteintomeettheplantsneeds.Inthesecases,ifaplanthasjustonecopyofagenethat

    IV. Theories of Field-based Organic Plant Breeding

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org22

    Tomato 1Sp Sp

    Tomato 2Sp sp

    Tomato 3sp sp

    encodesforfunctioningproteins,itwilllookandfunctionthesameasifithadtwoofthosegenes.Onlywhentherearenocopiesofthatgenepresenttoencodeforthefunctioningproteindoweseeadifferenceinappearance.Toillustrate,letuslookatdeterminantversusindeterminategrowthhabitintomatoes.Figure6illustratestomatoesthathave:(1)twofunctionalcopiesoftheself-pruninggene(Sp),(2)onecopyofSpandonenon-functionalself-pruninggene(sp),and(3)twonon-function-alself-pruninggenes.

    Noticethattomatoes1and2lookthesame.ThatisbecauseonlyonecopyofSpwasrequiredtopro-ducetheproteinsthatallowedthetomatotogrowcontinuously.Onlywhenbothcopieswerenon-functional,asintomato3,doesthetomatoendupshortanddeterminate.Inthiscase,wewouldcallsparecessivegeneandSpadominantgene.Inthecaseofdominantandrecessivegenes,thefunda-mentalconcepttounderstandisthatrecessivegenesareonlyvisibleinthephenotypeofaplant(whataplantlookslike)whentwocopiesofthegenearepresent,whereasasinglecopyofadomi-nantgenewillresultinthedominantphenotype.

    How genes travel together during reproductionUltimately,whenweengageinplantbreedingweareparticipatinginthesexualreproductionofplants.Inthissexualprocess,someoftheparentalcellsundergomeiosis,splittingtheirchromosomepairsintogametes:eitherspermoreggs.Thesegametesthencombinewithacomplementarygam-eteandformazygote,whichisthebeginningofanewplant.Oneofthewaysinwhichthisprocessprofoundlyaffectsplantbreedingisthroughthephenomenonoflinkage.

    Linkagereferstothefactthatsomegenesresideonthesamechromosomeandarethusphysicallylinkedtogether.Inbreeding,weoftenrepresentlinkedgeneswithtwolines,eachlineconnectingoneofthelinkedpairsofgenes:

    AbaB

    InthisexampleAandbarelinkedtogetherononehalfofachromosomepair,whileaandBarelinkedontheotherhalf.Linkagecausessomeoutcomesthatwewouldnotanticipateifthegeneswere

    Figure 6. This figure illustrates determinant versus indeterminate growth habit in tomatoes.

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org23

    notlinked.Forexample,ifweself-pollinatedtheaboveplant,wewouldgetthefollowingoffspring:AAbb,aaBB,andAaBb.WewouldnotseeanyAABBoraaBB,becauseAbandaBarelinkedpairsthattraveltogetherthroughtheprocessofmeiosis.Thisiswhathappensduringreproduction:

    AbGametes:AbaBaB

    Zygotes:AbAbaB AbaBaB

    ButwhatifthemostdesirablegenotypeisAABB?Unfortunately,linkagepreventsthisgenotypefromoccurring.Whenlinkagecausesanabundanceorlackofgenotypesthatisdisproportionatetowhatwewouldexpectbasedonthegenesthatareinthepopulation,wecallitlinkage disequilibrium.Fortunately,naturehasawaytoreducetheeffectsoflinkagebycrossing overchromosomes.

    Duringmeiosis,pairsofchromosomesconnecttogether.Whentheycomeapartagain,partsofonechromosomecanendupswappedwithpartsoftheotherchromosomethatitpairedwith.Thisswappingofchromosomepiecesbreaksuplinkedgenes.Dependingonhowclosetwogenesaretoeachother,thelinkagesbetweenthemmaybreakupmoreorlessquicklyandeasily.Tightlylinkedgenes,orthosethatareveryclosetogether,aremoredifficulttobreakapartthanthosethatarelooselylinkedorfartherapartfromeachother.Itispossibleforgenestobesotightlylinkedthatbreakingthemapartisvirtuallyimpossible.Hereisanillustrationofhowcrossingoveractstobreakapartlinkedgenes:

    Ab Ab ABaB aB ab

    Whyarelinkageandlinkagedisequilibriumimpor-tantforusasbreeders?Therearemanyreasons,buttwoofthemostimportantaretheneedforcautionwhenselectinginearlygenerationsandlinkagesrelationtoheterosis.

    Aftermakingacrossorcrossestogenerateanewpopulation,youneedtobecarefulaboutselectingjustafewofthebestplantsforthefirstfewgenera-

    tions.Toillustrate,letscontinuewiththetomatoexamplewherethelinkedgenesthatcomefromoneparentareA b,andthegenesfromtheotherparentarea B.IfwewanttofindplantsthathavetheA Bgenotypethenweneedtowaituntilthepopulationhasgonethroughenoughgenerationstoallowtimeforthoseparticularchromosomestohavecrossedoverbetweenthoseparticulargenes.OnlythencanweendupwithAandBonthesamechromosome.Ifselectionsaremadetooearlyontheycanreducethepotentialforcertaingenecombinationstooccur,ultimatelylimitingwhatyouhavetoworkwith.

    Heterosisisthetermusedwhentheoffspringofacrossaresuperiortotheirparents.Heterosisisalsocalledhybridvigor.Therehavebeenmanyargu-mentsovertheyearsaboutwhatcausesheterosis.Recently,mostscientistsagreethattheprimarycauseofheterosisinvolvestwoprincipleswehavealreadydiscussed:dominanceandlinkage.Tounderstandheterosis,letslookatthecrossoftwocornplants.

    Parent1

    Parent2

    Offspring

    Figure 7. This diagram illustrating heterosis shows the cross of two corn plants that creates an offspring superior to its parents.

    Whyistheoffspringsomuchtaller?Inthesecornplants,therearetwoloci(locationofagene)thatdetermineplantheight:AandB.PlantswithAaretallerthanplantswitha,andplantswithBaretallerthanplantswithb.BothAandBarecom-pletelydominantgenes,soaslongastheplanthasonecopyofthedominantgene,itisastallasifithadtwocopies.Inthisexampleparent1isAAbbandparent2isaaBB.

    Sowhathappenswhenwecrossthem?Theoff-springgetAbfromparent1andaBfromparent2,

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org2

    sotheirgenotypeisAaBb.BecausebothAandBaredominant,theyaffectplantheightinthesamewayregardlessofwhetherthegenotypeisAABBorAaBb.Thechartbelowshowsthatwhenparents1and2arecrossed,theiroffspringgetsaboosttoheightfromparent1sdominantAgene(shownasa+)andaboosttoheightfromparent2sdomi-nantBgene.Thisresultsinthehybridoffspringbeingtallerthaneitherparent.

    Parent 1

    Parent 2

    Offspring

    Genes AAbb aaBB AaBb

    Effect of A + +

    Effect of B + +

    Overall height

    + + ++

    Inthiscase,theparentseachhadcomplementarygenes.Whentheycrossed,eachparentcontributedallelesthat,throughtheirdominanteffects,maskedtheinferiorrecessiveallelesoftheotherparent.

    Heterosisisquitecommoninmanycropplants.Why,then,betweennaturalselectionandtheef-fortsofplantbreeders,havewenotbeenabletocombineallofthebestgenesintooneplant/vari-ety?Whydowehavetomakehybridstohidethepoorgenesoftheparents?Thisiswherelinkagecomesintoplay.LetsassumethelociforAandBarerightnexttoeachotheronthechromosomeandaretightlylinked.Inparent1scase,A bwilltraveltogetheruntilthechromosomecrossesoverintotheregionbetweenthegenes.Likewise,inparent2scase,aBwilltraveltogether.Untilacrossoveroccurstobreakthislinkageapart,thebestwecandoiskeepremakingthehybrid.Mosttraits,suchasplantheight,aredeterminedbymul-tiplegenes,notjustone.Manyofthesegenes,bothdesirableanddeleterious/undesirable,arelinkedtogether.Sinceitisunlikelywewillevercreateorcomeacrossaplantwiththeperfectcombinationofgenesforeverytraitwefinddesirable,wemustcontinuetoexploittheeffectofdominantgenestoproducesuperioroffspringbyfindingparents

    whosegenecombinationscomplementeachother.

    How genes operate in populations Upuntilnowwehavebeenfocusingonhowgenespassfromonegenerationtothenextandhowtheyaffecttheappearanceoftheplantswegrow.Tokeepourdiscussionsimple,wehavepresentedexampleswherejustoneortwogenescontrolatraitlikeplantheight.Now,letustakeastepclosertoreality.Manyofthetraitsthatwecareaboutplantheight,vigor,flavor,stresstolerancearecontrolledbymultiplegenes.Ratherthanplantsappearingeitherbigorsmall,theirheightmightex-istalongaspectrum.Inthisspectrum,afewoftheplantswillbeverysmall,afewwillbeverylarge,andmostwillbesomewhereinthemiddle.Ifyoumeasuredalltheplantheightsandgraphedhowcommonthevariousheightswere,youmightseethemappearasabellcurve,asshowninfigure8.

    When,asbreeders,weworkwiththesepoly-genic(multiple-gene)traits,ourgoalistoselecttheplantswiththetraitswewantandslowlyshiftthewholepopulationinthedesireddirec-tion.Figure9showstheslightimprovementyoumightexpectfromonegenerationtothenextthankstoyourselection.

    Tobesuccessfulasplantbreeders,wewanttoimproveplantpopulationsforthetraitsweareinterestedin.Fundamentally,thewaytoimproveapopulationistofindplantsthatwillproducesupe-rioroffspring.Howdowefindtheseplants?Letslookatfigure10toseeanexampleofafieldplant-edwithapopulationofcornthatweareinterestedinimprovingforheight.

    Thereisobviousvariationinthephenotypesoftheseplants.Ifthetaskistopickoutthetallestplants,itwouldbesimple:plants2,4,5,6,and12arethetallest.However,whatwereallywanttoknowiswhichoftheseplantswillproducethetallestoffspring.Thefirststepinfindingtheseplantsistounderstandthatthephenotype(P)isinfluencedbytwoprimaryfactors:thegenetics,orgenotype(G),oftheplantandtheenvironment(E).Inbreedingterms:P=G+E.Thegenotypeisthesumofallthegeneticinformationwithintheplant,althoughwegenerallyusethetermtorefertothegenesintheDNA.Theenvironmentencompasses

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org2

    Frequency of plants

    Plant height

    Figure 8. Rather than plants appearing either big or small, their height might exist along a spectrum. In this spectrum, a few of the plants will be very small, a few will be very large, and most will be somewhere in the middle as shown in this bell curve.

    Frequency of plants

    Plant height

    Generation 1 Generation 2

    Figure 9. This graph illustrating how polygenic traits operate in populations shows the slight improvement you might expect from one generation to the next thanks to your selection.

    Figure 10. This figure represents a field planted with a population of corn that we are interested in improving for height.

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org26

    allofthenon-geneticinfluencesonthephenotype.Examplesofenvironmentalfactorsaffectingplantgrowthincludefertility,temperatures,water,weedcompetition,plantingdensity,andpestanddis-easepressure.Sayweareabletogrowthesesametwelveplantswithverylittlewaterandtheyallgrow10inchesshorterthanbeforeasseeninthebelowfigure:

    Clearly,theenvironmentaffectedtheplantpheno-types,butdiditaffecttheminawaythatchangesourabilitytopickthebestparents?Inthiscase,no.Wewouldstillpick2,4,5,6,and12asourfavor-ites.Ifalleffectsfromenvironmentalstresseswereuniformlikethisexample,itwouldbesimpleforustopickthebestparentsbasedontheirpheno-types.Unfortunately,lifeisnotsosimple.Thereareenvironmentaleffectsthatmakeitmorechalleng-ing.Theseeffectsarebasicallyclassifiedintotwotypes:errorandgenotypebyenvironmentinterac-tion(GxE).Errorisacatch-alltermforeffectsthatwecannot,ordidnot,measureoraccountfor.Forexample,imaginethatadditionalfertilizerwasac-cidentallyappliedtothesoilunderplant4,provid-ingafertilityboostthatnoneoftheotherplants

    experienced.Thisfertilityboostwouldbeaneffectthatwasnotreplicable,andwouldinfluenceourselectioninanunpredictableandprobablyunhelp-fulway.Thisiserror.Someerrorwecanreducethroughgoodexperimentaldesign,especiallytheuseofreplicationandrandomization.Toillustrate,letstakeanotherlookatthecornplants.Whatifplants7through12aresmalleroverallbecausethe

    fieldisdrierontheright?Totestthiswecouldsplitthefieldintotwosectionsandreplicatetheexperi-mentbyplantingeachcornrandomlyoneachsideofthefieldasshowninfigure11.

    Byreplicatingourtrial,weseetheplantsontherightareconsistentlysmallerthanthoseontheleft.Replicationallowsustomeasurethiseffectandcorrectforit.Unlikeanyoftheothereffectsonplantphenotype,weareabletoaccountforandremoveeffectsthatarecapturedbyreplication.

    ThesecondeffectthatconfoundsourabilitytopickthebestparentsistheGxEinteraction.GxEisastatisticaltermthataccountsforthefactthatdif-ferentgenotypessometimesresponddifferentlyto

    Figure 11. Here is the above field split into two sections and the experiment replicated by planting each corn randomly on each side of the field.

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org27

    Original environment

    Dry environment

    thesameenvironment.Tocontinuewithourcornexample,saythatwegrewourcornbothintheenvironmentandinamuchdrierenvironmentasshownbelow.

    Inthisexample,theheightofmostplantswasre-ducedinthedryenvironmentrelativetotheorigi-nalenvironment.However,theplantheightswerenotuniformlyreduced.Number4ismuchsmaller

    thanintheoriginalenvironment,andnumber12isthesameheightinbothenvironments.Thisunequalresponsefromdifferentgenotypestodif-ferentenvironmentsiscapturedbytheconceptofGxE.Asbreeders,wehavetwochoicesofwhattodowiththisinformation.Wecanfocusourbreed-ingeffortstowardacertainenvironmentinthiscasetoselectforadryenvironment.Thismethodiscalledselectingforspecificadaptation.Alternately,wecanselectforthegenotypesthat,onaverage,performbestinallenvironments.Thismethodiscalledselectingforgeneral adaptation.Inreal-ity,breedingoftenmixesthesetwoapproachesbybreedingforadaptationwithinacertainspecifiedtarget population of environments(TPE).

    Asplantbreeders,weareinterestedintheplantsthataremostlikelytoproducethebestoffspring,notsimplyperformthebestthemselves.Hereisanotherkeytoplantbreeding:Theplantswiththebestgenotypesmaynotproducethebestoffspring.

    Thisisbecause,unlesstheyareclonallyreproduced,plantspassongenes,notgenotypes.Thenextsec-tionwillexplainwhy,includingtheimplications.

    Dominanceeffectsconfoundoureffortstofindthebestparentsbecausetheyaregeneticeffectsthatarenottransferablefromparenttooffspring.Tounderstanddominanceeffects,wewillreturntoourtomatoes(seenextpage).

    Again,givenidenticalenvironments,tomatoes1and2havethesameappearance,becausetheybothhaveatleastonecopyofthedominantSpallele.However,whiletomato1willonlyproduceindeter-minateoffspring,tomato2hasachanceofproduc-ingdeterminantoffspring.

    Sincetomato1(Sp Sp)andtomato2(Sp sp)areindistinguishablefromeachother,youareequallyaslikelytochoosethemasparents.Thisisprob-lematicbecauseyoumaypreferoneortheotherforyourbreedingwork.ThedominanceeffectthatallowsSp spplantstoappearthesameastheSp Spplantsisagenotypiceffectthatisnotreliablytransferabletoitsoffspring.Thebottomlineisthatgenotypiceffectsduetodominancearenottransferabletooffspring.

    Insummary,geneticeffectsandenvironmentaleffectsinfluenceplantphenotypes.Ifproperex-perimentaldesignsareusedwecanaccountfor

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org28

    environmentaleffects.Theseexperimentaldesignsallowustoachieveoneofourprimarygoalsasplantbreeders,whichistominimizeandaccountforthevariabilitycausedbynon-geneticeffects.

    Onemoreimportantplantbreedingtermisherita-bility.Heritabilitydescribeshowpredictablyatraitispassedonfromparenttooffspring.Plantbreed-ersoftentalkabouttheheritabilityofatraitandhowtoincreaseit.Sometraitsarenaturallyhighinheritability.Oftenthesetraitsarerelativelyunaf-fectedbyenvironment,likeseedcolororflowershape.Wecanincreasetheheritabilityinalltraitsbyusingmatingdesignsthatincreaseormaintainadditivegeneticvariance,andbyusingexperimen-taldesignsthatreducethevarianceofothereffects.

    How to see the genetic differences between plantsEarlierwetalkedabouttheimportanceofreduc-ingenvironmentalerror.Webrieflymentionedreplicationasonetooltohelpreducethiserror.Nextwellgointomoredepthabouthowtousereplicationandothertoolsofexperimental de-signtohelpmakemoreeffectiveselectionsinourbreedingwork.

    Experimentaldesignreferstohowtheexperimentissetuporarrangedinthefield.Oneofthemore

    commondesignsforbreedingiscalledrandom-izedcompleteblockdesign,orRCBD.ARCBDisreplicatedusingmultipleblocks,witheachblockcontainingthecompletesetofbreedinglineswithinit.TocreateaRCBD,assigneachbreedinglinetoaplotineachblockandarrangetheplotsinarandomizedorderwithineachblock.Thebreed-ingtrialshouldideallybelocatedwherethereisaslittlevariabilityaspossible.Wherevariabilityexists,trytoorienttheblockssothevariabilityisbetweenblocksratherthanwithinthem,becausebreedinglinesarecomparedtoeachotherwithineachblock.Replication,randomization,andblock-ingaretheessentialelementsoftheexperimentaldesignthathelpstoseparateoutthegenotypicef-fectsfromtheenvironmentaleffects.

    Understand the effects of the environmentEnvironmentaleffectsrefertotheinfluencethatvariablefieldconditionshaveontheperformanceofplants.Forexample,ifthefieldisonaslopeandthesoilisricheronthedownhillside,thenplantsonthedownhillsidemaygrowlargerbecauseofbettersoilconditions,notbecausetheyhavesupe-riorgenetics.Likewise,plantsgrownonthesouth-westsideofafieldmaygrowlargerbecausetheyreceivemoreafternoonsun,orsmalleriftheyaresun-sensitiveandtheirrootsareeasilydriedoutbyafternoonsun.Naturalvariabilityoftenexistsinthefieldthataffectscropgrowth,includingvariations

    Tomato 1Sp Sp

    Tomato 2Sp sp

    Tomato 3sp sp

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org29

    insoilquality,soiltexture,soilpH,soilfertility,drainage,pestpressures(insect,weed,disease,andanimals),sunexposure,irrigation,andtemperature(bothwarmandcoldpockets).Somesourcesofenvironmentalvariationarerelativelystable,likesoiltypeandproductionpractices,whileothersvaryfromyear-to-year,likeweatherandclimate.Pests(diseases,insects,andweeds)maybestableorvariabledependingontheorganism.Forex-ample,thesoilinsectsknownassymphylansareusuallyfoundinthesamespotyearafteryearandwouldberegardedasstable,whereasaphidsblownintotheplotsfromanotherregionwouldbemorevariable.Stablesourcesofvariabilitycanbecom-pensatedforbytheuseofrandomizationwhereasseasonallyvariablesourcesofvariabilityrequiretestingovermultiplelocationsand/orseasons.Seasons,asusedhere,maybedifferenttimesofyearordifferentyears.Theobjectiveistoobtaindatainanumberofdifferentgrowingenviron-ments.Thegoaloftheexperimentaldesignistominimizetheeffectsofenvironmentalvariabilityontrialresultsbyblockingreplicationssofieldvariationisminimizedwithinblocks.Inessence,replicationinthebreedingtrialaccountsforstableeffectsandrepeatingtheexperimentovertimecompensatesforvariableeffects.

    Ensure that plants receive consistent treatmentManagementpracticesshouldbeconsistentacrossallbreedingplots,includingirrigation,soilfertil-ization,pestmanagement,staking,cultivationandweeding,oranyotherhorticulturalaspectofcropproduction.Ifpracticesarenotconsistent,differ-encesinperformancemaybetheresultofunequaltreatmentsasopposedtogeneticdifferencesinvarieties.Forexample,ifhalfthefieldwasculti-vatedandtheotherhalfleftuncultivated,theplotsintheuncultivatedsectionwouldhaveanunfair

    disadvantage,astheywouldlikelybesubjectedtogreaterweedcompetition.

    Randomize the plotsRandomizingplotswithineachblockhelpstoincreasetheconfidencethattheevaluationresultsareduetogeneticdifferencebetweenvarietiesratherthanvariationintheenvironment.Themorereplicationsusedthegreatertheassurancethatmeasurementswillreflectdifferencesinthevariet-iesratherthanvariationinthefield.Randomizingisalsoimportanttominimizetheinfluencethatvarietieshaveononeanother.Forexample,ifeachbreedinglinewasplantedinthesameorderineachblockratherthanifonelineisplantednexttoanotherverytall,vigorousline,itmightbeunfairlyaffectedbythecompetitionforlightandnutrients.

    Theorderofplotsmayberandomizedusinganyba-sicrandomizationtool,asbasicasdrawingnumbersfromahat.Aneasymethodistomarkalloftheplantstakesforeachblockandthenmixthemup,walkingtheblockandinsertingwhicheverstakecomestohandfirsttomarkthefirstplot.Continuedowntherowmarkingplotsinthismanner.

    Use sufficient population and plot sizesThesizeofplotsinthetrialshouldbelargeenoughthattheresultingnumberofplantsprovidesagoodrepresentationofthewholefamilyorpopulation.Thegreatertheplotsize,themorerepresentativeitwillbeofthepopulation.However,thelargertheplotthemoredifficultitmaybetomanage,especiallyifyouaredealingwithalargenumberofvarietiesandthreeormorereplications.Ingeneral,whenevaluatingopen-pollinatedvarietiesofself-pollinatedcrops,fewerplantsarenecessarytogetarepresentativesamplingthanincross-pollinatedcrops.Thisisduetothegenerallygreaterinherent

    Example of RCBD, where all varieties would be in each block or rep

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org30

    geneticvariabilityincross-pollinatedpopulations.

    Ideallyplotsshouldbelargeenoughthatyoucantreatthemthesameastherestofyourfarmandmaintainthemwithyournormalfarmingpractices.Forexample,ifyouusuallyplantyourcabbageon36-inchcentersandfieldcultivatewithatractor,youwilllikelygetdifferentresultsthensomeonewithasmallgardentrialon24-inchspacingus-inghandhoeing.Similarly,ifyouusuallyapplydripirrigation,thenoverheadirrigationmayskewresults.Forbestresults,locateyourbreedingtrialwithinyourproductionfieldofthesamecropsothatallproductionpracticesarethesameasforcommercialproduction.Plotsizemaybedictatedbyyournormalplantingandmanagementprac-tices.Forexample,aminimumamountofseedmayberequiredtoplantwithaseeddrillorvacuumplanter,andtheplantingequipmentmayneedtorunacertaindistancetofunctionproperly.Me-chanicalharvestingequipmentmayalsorequireacertainminimumscaletooperateappropriately.

    Anothervariabletoconsiderwhendeterminingplotsizeisthenumberofrowstoincludeperplot.Singlerowplotsallowyoutoplacemorelinesorfamiliesinyourtrial,butmaygivemorevariableresultsbecauseoffewerplantsperplotandpoten-tialcompetitionbetweenvarietiesinadjacentplots.Withthreeorfourrowplots,thecenteroneortworowscanbeevaluatedwithoutconcernofcompeti-tioneffectswithneighboringplots.Whilethisap-proachtakesmorespace,itdoesminimizeinter-plotcompetitionandproducesmoreaccuratedata.Aswithplotlength,aminimumnumberofrowsmaybedeterminedbyequipment.Forexample,youmaybelimitedbythesetupofyourplantingequipment.Borderrowsaroundyourexperimentalplotswillminimizewhatisknownastheedgeeffect,where-byplotsontheedgeexperienceslightlydifferentgrowingconditionsrelativetotherestofthetrial.Withoutborders,plotsontheedgemaybemoreproductivebecausetheylackcompetingplantsaroundthemorlessproductivebecausetheyarenexttoamorecompetitivecropordamagedbyfieldoperationsoccurringinadjacentcrops.Assuch,datafromun-borderedplotsaroundtheout-sideofthetrialwillbebiasedrelativetothecenterofthefield.Borderrowsshouldbeplantedatthefrontandbackofatrialaswellasalongthesides.

    Abundant Bloomsdale organic spinach breeding projectWhat were the goals of this project?Thereisastrongdemandformoreopen-pollinatedsavoyspinachvarieties,whichareimportanttodiversifiedorganicfarms.ThegoaloftheAbundantBloomsdaleprojectwastodevelopadarkgreen,open-pollinatedsavoyspinachvarietywithhighnutritionalcontent,goodflavor,andsuperiorboltresistance.Thevarietyneededtohaveanuprightplanthabitandbeusefulatallstagesfrombabyleaftofullymaturebunchingtypes.

    ThePacificNorthwestisaprimespinachseedproducingregion,yetthereislittleorganicspin-achseedproducedthere,asmostseedcompanieshaveignoredtheorganicspinachseedmarket.Anendresultofthisbreedingprojectwillbetohelpbothorganicfarmersandsmallorganicseedcompaniesbysupplyingthemwithanalternativetothespinachhybridsthatarelargelybredforconventionalsystems.

    Breeding procedureTheprojectbeganwithacrossofWinterBlooms-daleandEvergreen.WinterBloomsdaleisacoldhardy,open-pollinatedspinachvarietywithdarkgreenleaves,deepcurl,andgoodflavor.ThevarietywasoriginallybredforwinterandearlyspringharvestincoastalVirginia.ItisthelastoftheoldfullysavoyspinachvarietiesthatisstillwidelygrowninNorthAmerica.Evergreenisasemi-savoyspinachvarietydevelopedbyDr.TeddyMorelockoftheUniversityofArkansasforwinterproductioninArkansasandTexas.Thevarietyhasstronghorizontalresistancetomanyofthemajordiseasesofspinach.

    Inthefirstyearofbreeding,OSAbreederJohnNa-vaziomadeastraincross,whereapproximately20plantsofbothparentalvarietieswereplantedside-by-sideandallowedtocrossthroughopen-pollina-tion.Bymakingthecrossinthismanner,ratherthanpollinatingbyhand,morecrossesbetweenthevarietiescouldoccurandmoreofthegeneticvari-abilityofthetwoparentsisretained.

    V. Examples of Farmers Breeding for Organic Systems

  • Introduction to On-farm Organic Plant Breeding www.seedalliance.org3

    Afterthatinitialcross,theresultingpopulationwasgrownforfivegenerationsonvariousfarmsinWesternWashington.Thisintermatingprocessbrokeupsomeofthegeneticlinkages(seelinkagedescriptioninprevioussection)andallowedthegenesofthetwoparentstothoroughlymix.Duringthisprocess,alowintensityofmassselectionwasappliedforappropriateleafshapeandvigor.

    Inthesixthseason,afterfivegenerationsofin-termating,NavazioandlocalfarmerMarkoColbyplanteda2,000plantspaceplantnursery,whereeachplantwasspacedfarenoughfromallotherssotheycouldbeevaluatedindividually.Fromthese2,000plantstheyselectedintenselyforthetraitsalreadymentioned,andendedupwith130spin-achplants.Theseselectedplantswereallowedtoopenlypollinate.

    Asspinachisdioecious,withplantscontainingeithermaleflowersorfemaleflowers,onlyabouthalfoftheselectedplantswereseedbearing.Ulti-mately,seedfrom67femaleplantswereharvestedintoseparatebagsthatwerenumberedsequential-lyfrom1to67.Sevenfamiliesweresubsequentlydroppedbecauseoftheirpoorseedqualityoryield.Inthefollowingyear(yearseven),seedoftheremaining60familieswereplantedinfamilyrowsandevaluatedatseveralstagesofgrowthbetweenthebabyleafandmaturebunchingstages.Ofthese60families,onlyfiveprovedtobesuperiorthroughallgrowthstages.Theother55undesirablefami-lieswereeliminatedandtheremainingfivese-lectedfamilieswereallowedtointermatethroughopen-pollination.Theresultingseedofeachofthefivefamilieswereharvestedseparatelyintofivefamilybags.

    Inyeareight,NavazioandColbyplantedseedfromthesefivefamiliesasfamilyrows,withthegoalofidentifyingthebestoftheseselectedfamiliestocomprisethefinalpopulationthatwouldbere-leasedasAbundantBloomsdale.Thefivefamilieswereagainevaluatedatallstagesofgrowthbe-tweenthebabyleafandmatureleafstage.Inthisfinalevaluation,onlyoneofthefivefamilieswasuniforminallofthetraitsthatsatisfiedthebreed-inggoals.Thisfamilyhaddarkgreen,fullysavoyedleavesonaveryupright,vigorousplant,andarich,

    sweetflavorandsuperiorcoldhardiness.Theotherfourfamilieswereeliminatedandthechosenfam-ilywasroguedforofftypes.TheremainingplantswereallowedtointermateandproducethestockseedofAbundantBloomsdale.

    Winter sprouting broccoliWhat were the goals of this project?Thisbreedingprojectisongoing,andthegoalsaretodevelopapurplesproutingbroccolivarietythatreliablyoverwintersinthePacificNorthwestandproducesgoodyieldsoftender,flavorful,6to8inchshootswithpurpleflorets.Thepurplewintersproutingbroccolivarietiesthatar