U.S. flag

An official website of the United States government


Main content area

A comparison of dry matter yield stability assessment methods for small numbers of genotypes of bermudagrass

Rose, Louis W. IV, Das, Modan K., Taliaferro, Charles M.
Euphytica 2008 v.164 no.1 pp. 19
Cynodon dactylon, forage grasses, plant growth, dry matter accumulation, genotype, statistical analysis, mathematics and statistics, genotype-environment interaction, Oklahoma
No information is available on the efficacy of various nonparametric stability parameters when compared with GGE biplot methodology in assessing the stability of dry matter yield in bermudagrass (Cynodon dactylon L. Pers.) when a small number of genotypes is assayed. This study was conducted to compare the results of four nonparametric stability parameters developed by Huehn and Nassar ( [graphic removed] , [graphic removed] , [graphic removed] , [graphic removed] ), Kang's rank-sum method and the GGE biplot method for five genotypes over 11 location-year environments at Oklahoma State University experiment stations. Results from analysis of variance procedures indicated highly significant levels of genotype-by-environment interaction (P < 0.01), which further indicated the need for stability analysis measures to be conducted. Results of the stability analysis indicated agreement among [graphic removed] , [graphic removed] , Kang's rank-sum method, and the biplot method for the stability rankings of the genotypes and between these methods and the overall yield rankings of the genotypes. The [graphic removed] and [graphic removed] statistics were not in agreement with each other or any of the previously mentioned methods concerning the stability rankings of the genotypes. From examination of the formulae for the nonparametric statistics it was concluded that, when a small number of genotypes is assayed, the [graphic removed] , [graphic removed] , [graphic removed] and [graphic removed] statistics have the potential to be extremely sensitive and to produce misleading results. It was further concluded that for assessment of small numbers of genotypes the GGE biplot stability analysis method, augmented with Kang's rank-sum method, would produce the most reliable estimates of genotype stability.