Vol: 33 Page: 221

Authors:
Wiesław Mądry
Anna Rajfura

Title:
 Statistical analysis of stability measures using genotype x environment data. Part III. Comparison of estimators and tests in respective models from two groups and a numerical example

Language: Polish

Keywords:
classification of genotypes with respect to stability and adaptation
comparison of estimators and tests for stability measures based on the respective models in two groups
genotype by environment data
grain yield
multi-environment (location) trials (MET)
oat
stability analysis

Summary:
In this part of the paper similarities of optimal or the best among known estimators and tests for the same parameters in S-C and Sh models as well as C-K and E-R-S models are presented. Statistical properties of these tools and their practical usefulness in stability analysis are given and discussed. Illustration of applications of these statistical tools concerning the respective models from two groups was done using grain yield data of 29 oat varieties tested in multi-location trials (MET) consisted of 14 locations chosen from all polish agricultural area in 1996. S-C and Sh models as well as C-K and E-R-S joint regression models may be considered as pairs of alternative, real and statistically proper tools that enable us to conduct similarly effective analysis of genotypic effects and stability analysis of these genotypes on the base of data coming from multi-environment (location) trials put in two-way genotype by environment classification. Numerical calculations made by program SERGEN in the analysis of data according to S-C and C-K models cover most calculation demands in Sh and E-R-S models; program SERGEN does not deliver estimates of stability variances σi2 and residual variances σδ(i)2 as well as respective F statistics in the last models.