ESTIMATIONOFGENETICPARAMETERS,OPTIMALTESTDESIGNSAND PREDICTIONOFTHEGENETICMERITOFCLONALANDSEEDLING MATERIALOFEUCALYPTUSGRANDIS By LUISFERNANDOOSORIO ADISSERTATIONPRESENTEDTOTHEGRADUATESCHOOL OFTHEUNIVERSITYOFFLORIDAINPARTIALFULFILLMENT OFTHEREQUIREMENTSFORTHEDEGREEOF DOCTOROFPHILOSOPHY UNIVERSITYOFFLORIDA 1999 Dedicatedto: Mywife,Astrid Mychildren,Felipe,Daniel,AndresandLisa Mymother,Cecilia,auntIsabelandmother-in-lawMariela ACKNOWLEDGMENTS I would liketothankthemembersofmy supervisorycommittee, Doctors D. Huber,R.C.Littel,P.Lyrene,D.RockwoodandT.L.White,fortheirtime,effortand helpduringalltheseyears.Inparticular,myspecialgratitudetoDr.TimWhiteforhis continuingguidance,appropriateadvise,friendshipandsupportthroughalltheprogram. MyappreciationalsogoestoDr.DudleyHuberforhisexcellentsuggestionsandhelpto understanddifficultissues. ThisprojectcouldnothavebeendonewithoutthefinancialsupportofSmurfit Carton de Colombia, Fundacion Carton de Colombia and Colciencias. I want to especiallythankDr. JeffWright, Mr. EdgarLondono, Mr. VictorGiraldo andJaime Munozforalltheirsupport.ManythanksareextendedtoByronUrrego,NorhaIsaza, LilianaPerafan,HugoEspana,HaroldCampo,DiegoRamirezandMauricioArcefor theircollaborationwiththeestablishment,maintenanceandmeasurementoftheteststhat form the basis ofmydissertation. Also, I would liketo express my appreciationto AnabellaMontanofortakingcareofmypersonalissuesinColombiathroughthewhole program. IespeciallywanttothankGregPowellforhiscontinuingavailabilityandhelpin thegeneticslaboratory.Itwasagreatexperiencetosharethistimewithgraduatestudents UilsonLopes,PengxinLu,RebecaSanhuezaandRyanAtwood.Myspecialgratitudeto JavierLopez-UptonandVictorSierra-Lucerofortheirunconditionalhelp,friendshipand iii supportindifficulttimes.Mywife,Astrid,whowentthroughitallwithme,deserves muchofthecreditforthecompletionofthisprojectandIalsothankherandmychildren who are the source of my happiness for their love, support, and understanding. iv TABLEOFCONTENTS page ACKNOWLEDGMENTS iii LISTOFTABLES viii LISTOFFIGURES x ABSTRACT xii CHAPTERS 1 INTRODUCTION 1 2 AGETRENDSOFHERITABILITIESANDGENOTYPEBY ENVIRONMENTINTERACTIONFROMCLONALTESTSOF EUCALYPTUSGRANDIS 5 Introduction 5 PreviousWork 6 Goals 8 MaterialsandMethods 8 FieldEstablishmentandTestDesign 8 DataCollection 9 DataEditingandStandardization 11 PreliminaryAnalyses 11 AnalysesofSurvival,GrowthTraitsandWoodDensity 13 Genotype-by-EnvironmentInteraction 18 EstimationofGeneticGains 19 ResultsandDiscussion 21 StemBreakage 21 AverageTrendsinSurvivalandGrowth 22 TrendsofBroadSenseHeritabilities 25 Survival 25 GrowthTraits 27 Genotype-by-EnvironmentInteraction 30 WoodDensity 33 GeneticGains 35 v Conclusions 36 3 AGE-AGEANDTRAIT-TRAITCORRELATIONFOREUCALYPTUS GRANDISANDTHEIRIMPLICATIONSFOROPTIMALSELECTION AGEANDDESIGNOFCLONALTESTS 40 Introduction 40 MaterialsandMethods 43 FieldTestDesignandTraitsMeasured 43 EstimationofGeneticCorrelations 44 OptimalTestDesign 48 DeterminingOptimalSelectionAge 51 ResultsandDiscussion 54 GeneticParameterEstimates 54 Age-agecorrelations 54 Trait-traitcorrelations 56 Standardparameterestimates 59 OptimalTestDesigns 61 OptimalSelectionAge 64 Conclusions 68 4 ESTIMATIONOFGENETICPARAMETERSFORSEEDLING POPULATIONSOFEUCALYPTUSGRANDISANDCOMPARISONS WITHCLONALPOPULATIONS 70 Introduction 70 MaterialsandMethods 72 DataCollection 75 DataEditingandStandardization 75 GeneticGroupComparisons 76 EstimatesofGeneticParametersfromSeedlingTests 78 ComparingofSeedlingandClonalPerformance 82 ResultsandDiscussion 84 GeneticGroupDifferences 84 GeneticParameteEstimatesofSeedlingsandClones 88 Conclusions 96 APPENDICES 1 TBRRAIOTASDANSDENWSOEOHDERDIETNASBIITLYIITNYEU(HC2A)LEYSPTTIUMSAGTREASNDFIOSR GROWTH 103 2 PERCENT OF STEM BREAKAGE AND RESULTS FROM THE ANALYSESOFVARIANCE 104 vi 3 SINGLE SITE AND ACROSS-SITE STATISTICAL MODELS FOR ESTIMATIONOFGENETICPARAMETERSFORSTEMBREAKAGE 106 4 VARIANCE COMPONENTS, PAIRED-SITE BROAD SENSE HERITABILITIES AND TYPE B GENETIC CORRELATIONS OF GROWTHTRAITS 108 5AVERAGESURVIVAL(%),ANDBROADSENSEHERITABILITYBASED ON 0/1 DATA AND ON THE UNDERLYING SCALE FOR CLONAL TESTSINTHREEDIFFERENTTYPESOFENVIRONMENTS Ill 6STATISTICALMODELSFORGENOTYPE-BY-ENVIRONMENT INTERACTIONANDSTABILITYINDEX 112 7 VARIANCE COMPONENTS FROM MULTIVARIATE ANALYSES OF GROWTHTRAITS 114 REFERENCES 117 BIOGRAPHICALSKETCH 128 vii LISTOFTABLES Table page 2-1.Locationandclimateforclonaltests,series2-36,ofEucalyptusgrandisin SmurfitCartondeColombia 9 2-2.SitepreparationandlayoutofclonaltestsofEucalyptusgrandisatSmurfit CartondeColombia.Fertilizationwasattimeofplantingandratesshownare amountsoffertilizerappliedpertree 10 2-3.Broad-senseheritabilitiesandgeneticcorrelationsamonggroupsof environmentstreatedasdifferent"traits"(1=targetenvironment, 2=Guachiconaand3=Maravillas),forgrowthvariablesandwooddensityof growthanalyzedatsixyearsofage" 31 3-1.Climaticconditionsandfieldlayoutofclonaltests,series2-36,ofEucalyptus grandis planted in three different environments at Smurfit Carton de Colombia 44 3-2.Standardgeneticparameterestimatesformeanannualincrement(MAI)tobe usedinthesimulation.Thevariancecomponents cj2CSjo^earevariances duetoclonaleffects,clonebyenvironmentinteractioneffectsandresidualeffects respectively. Variancecomponents areexpressed asafraction ofaphenotypic varianceequaltounityduetodatastandardization.Thevalue isthebroad-sense individual heritability (Equation 3-3), rBis the type B genetic correlation M(EAqIuatatiosnix2y-e6a)rasn(dEqrxuyaitsitohne3g-e2n)eticcorrelationbetweenMAIateachageand 58 3-3.Trait-traitgeneticcorrelationsbetweenheight(HT),meanannualdensity (MAI) and wood density (WD) of clonal trials Eucalyptus grandis in differentenvironmentsatthreeandsixyears.Broadsenseheritabilitiesare shownonthediagonals 60 4-1. Locationandclimatic conditions ofseedlingtests ofEucalyptusgrandis correspondingtoseries2-71and2-73inSmurfitCartondeColombia 73 4-2.GeneticgroupsofEucalyptusgrandisandtheircorrespondingfamiliestested inseries2-73acrossfoursites 74 viii 4-3. Analysis ofvariance for height (HT) andmean annual increment (MAI) pooledacrossthefourseriesatthreeyearsofage 85 4-4.Narrow-senseheritabilitiesandtypeBgeneticcorrelationsforheight(HT), individual tree volume (VOL), mean annual increment (MAI) and wood density(WD)fromtwoseriesofseedlinggenetictrialsandoneseriesof clonal trials of Eucalyptus grandis. For the seedling trials parameter estimateswerederivedfrombivariateanalyses. Thosefortheclonaltrials werederivedfrommultivariateanalysesincludingtwoextremeenvironments (seeChapter2) 88 4-5Narrow-senseheritabilitiesandtypeBgeneticcorrelationsforheight(HT), individual tree volume (VOL), mean annual increment (MAI) and wood density(WD)fromtwoseriesofseedlinggenetictrialsandoneseriesof clonaltrialswith44commonortetsofEucalyptusgrandis 92 ix LISTOFFIGURES Figure page 2-1.Trendsofaveragesurvival,height,individualtreevolumeandmeanannual incrememt(MAI)ofEucalyptusgrandisacrosstime 24 2-2.Trendsofbroadsenseheritabilities(H2)forsurvival,height,individualtree volume and mean annual increment (MAI) ofEucalyptus grandis across time 26 2-3.TrendsoftypeBgeneticcorrelationsforgrowthtraitsofEucalyptusgrandis atfivelocationsrepresentingthetargetenvironmentacrosstime 32 2-4.Trendsofaveragebasicwooddensityandbroadsenseheritabilitiesofseven clonaltrialsofEucalyptusgrandisacrosstime 34 2-5. Clonalgeneticgainsinmeanannualincrement(MAI)andindividualtree volume,inthetargetenvironment,estimatedby usingpredictedbreeding valuesandthestandardformula.Eachgainestimateiscomputedbyselecting the bestcloneoutofatotalof 67untilselectingallofthem.Gainbvstands forgain computed frompredictedclonalvaluesandGainforreferstogain computedfromthestandardformula(Falconer1993) 37 3-1.dAigffee-raengteaggeesnewtiitchcMoArrIelaattisoinxsye(arrgs)ifnocrlomnealantriaalnsnoufalEucianlcyrpetmuesntgra(nMdAiIs) at 55 3-2.Age-agegeneticcorrelations(rg)forheightatdifferentageswithMAIatage 6yearsinclonaltrialsofEucalyptusgrandis 55 3-3.Selectionefficiencyformeanannualincrement(MAI)ofdirectselectionat three and four years with respect to maximum gain at the same age achievablefromaninfinitenumberofsitesandblocks.Geneticparameters usedtoconductthesesimulationswerefromTable3-2 62 3-4.Effectofdiscountedselectionefficiencyatperpetuity(DSE fromequation3- 16)onselectionageusingadiscountrateof5%and10% for arandomized completeblockdesignwithsingletree plots andafixednumberofsix blocksfromagetwotosixinclonaltrialsofEucalyptusgrandis.Thetimeto x