摘要
Identificationofhigh-yieldingstablepromisingricelinesanddeterminationofsuitableareasforricelineswouldbedonebyadditivemaineffectsandmultiplicativeinteraction(AMMI)model.SevenpromisingricegenotypesplustwocheckvarietiesShiroudiand843wereanalyzedusingarandomizedcompleteblockdesignwiththreereplicationsinthreeconsecutiveyears(2012,2013and2014).Homogenouserrorvariancewasindicatedinthenineenvironmentsforgrainyield.Thecombinedanalysisofvarianceindicatedsignificanteffectsofenvironment,genotypeandgenotype×environment(GE)interactionsongrainyield.ThesignificanteffectofGEinteractionreflectedonthedifferentialresponseofgenotypesinvariousenvironmentsanddemonstratedthatGEinteractionhadremarkableeffectongenotypicperformanceindifferentenvironments.TheapplicationofAMMImodelforpartitioningtheGEinteractioneffectsshowedthatonlythefirsttwotermsofAMMIweresignificantbasedonGollob'sF-test.ThelowestAMMI-1wasobservedforG7,G2andG6.G7andG6hadhighergrainyield.Accordingtothefirsteigenvalue,whichbenefitsonlythefirstinteractionprincipalcomponentscores,G1,G6,G2andG9werethemoststablegenotypes.Thevaluesofthesumoffirsttwointeractionprincipalcomponentscorescouldbeusefulinidentifyinggenotypestability,andG6,G5andG2werethemostdynamicstablegenotypes.AMMIstabilityvalueintroducedG6asthemoststableone.AccordingtoAMMIbiplotview,G6washighyieldingandhighlystablegenotype.Inconclusion,thisstudyrevealedthatGEinteractionswereanimportantsourceofriceyieldvariation,anditsAMMIbiplotswereforcefulforvisualizingtheresponseofgenotypestoenvironments.
出版日期
2017年03月13日（中国期刊网平台首次上网日期，不代表论文的发表时间）
