Association among stability measurements in rubber tree traits

Abstract The simultaneous use of different methods of stability analysis is becoming more widely used in the selection of genotypes in breeding programs. Knowing how these different methods are associated can contribute to a more efficient selection. The objective of the present study was to assess association among different methods of stability analysis in the rubber tree [ Hevea brasiliensis (Willd. ex Adr. de Juss.) Muell.-Arg.], using different traits and different groups of genotypes for the same trait. Two open-pollinated progeny populations (POP1 and POP2) and a group of clones (Group 1) were analyzed. POP1 contained 22 progeny, assessed in the municipalities of Pindorama, Votuporanga and Jau, Sao Paulo state, Brazil. POP2 included 30 progeny assessed in Selviria, Mato Grosso do Sul state, Votuporanga and Colina, Sao Paulo state, Brazil. Group 1 consisted of 25 clones assessed in Votuporanga. The following traits were assessed in POP1: girth, rubber yield, bark thickness and number of latex vessel rings. Rubber yield was assessed in POP2 and Group 1 clones. The following methods of stability analysis were used: Wricke; Eberhart and Russell; Lin and Binns; AMMI (Principal Additive Effect and Multiplicative Interaction) and HMRPGV (Harmonic Mean of the Relative Performance of the Genetic Values) predicted by Blup (Best Linear Unbiased Prediction). The Spearman correlation was used to verify the association between the stability parameters. Three scenarios were observed; in the first, some parameters did not show significant association in any of the analyses: Eberhart and Russell with Lin and Binns; Eberhart and Russell with HMRPGV; Wricke with HMRPGV; AMMI with Lin and Binns and AMMI with HMRPGV. In the second, some parameters showed significant associations in all the analyses such as: Wricke with AMMI, Lin and Binns with HMRPGV. Finally, in the third, the significance of association varied according to the analysis: Eberhart and Russell with Wricke, Eberhart and Russell with AMMI and Wricke with Lin and Binns. Therefore, the association among some stability parameters varies according to the trait and/or the group of genotypes analyzed. The association observed in one analysis cannot be applied to another analysis. In this data analysis is possible to observe that the AMMI method can be used with HMRPGV or Lin and Binns. In addition, Eberhart and Russell can be used with HMRPGV or Lin and Binns, since they offer differentiation and provide additional approaches to the study of stability.