Mohammadi, R., Roostaei, M., Ansari, Aghaee, M. and Amri, M. 2010. Relationships of phenotypic stability measures for genotypes of three cereal crops. Can. J. Plant Sci. 90: 819Á830. Multi-environment trial (MET) data are required to obtain stability performance parameters as selection tools for effective genotype evaluation. The main objective of this study was to investigate the interrelationships among nine phenotypic stability methods using grain yield from three sets of cereal experiments [15 durum wheat (Triticum turgidum var. durum) genotypes )12 environments; 20 bread wheat (T. aestivum L.) genotypes)18 environments; and 13 barley (Hordeum vulgare L.) genotypes )18 environments]. The experiments were conducted in representative rain-fed areas of Iran in collaboration with the International Center for Agricultural Research in the Dry Areas (ICARDA). The combined ANOVA for environments (E), genotypes (G) and G )E interaction was highly significant (PB0.01) for each set of data, suggesting differential genotypic responses and the need for stability analysis. The inter-relationships among the parameters and their association with mean yield based on Spearman's rank correlation were determined in each of the three cereal experiments. Highly significant correlations were found between several of the stability measures indicating that several of the statistics probably measure similar aspects of phenotypic stability for these crop species. The AMMI stability value (ASV), variance of regression deviation (S 2 di ) and Wricke's ecovalence (W 2 i ) were consistently and highly correlated with each other over these crops and, therefore, could be used if selection is to be based primarily on stability. The superiority index (Pi) and geometric adaptability index (GAI), which are related to the dynamic concept of stability showed significant correlation with mean yield over these crops, suggesting P i and GAI would be the best methods for ranking genotypes across environments. The coefficient of variation (CV), regression coefficient (b i ), yield reliability index (I i ), and environmental variance (S 2 x ) showed inconsistent relationships with either the static or dynamic concepts of stability over these crops. The correlation analysis provided a good description of static and dynamic concepts of stability for interpreting the G)E interaction and verified that the groups of stability methods (dynamic vs. static) discriminated genotypes in different fashions in these crops.Key words: Rank correlation, phenotypic stability measures, dynamic and static stability Mohammadi, R., Roostaei, M., Ansari, Aghaee, M. et Amri, M. 2010. Liens entre la mesure de la stabilite´du phe´notype et le ge´notype de trois cereals. Can. J. Plant Sci. 90: 819Á830. On a besoin de donne´es multi-environnementales pour e´tablir les parame`tres de la stabilite´qui serviront de moyen de se´lection en vue d'une e´valuation efficace du ge´notype. Cette e´tude avait pour principal objectif d'examiner les relations entre neuf me´thodes d'analyse de...

Th e main objective of this work was the analysis of barley (Hordeum vulgare L.) multienvironment trials (MET) fi rst to identify superior genotypes for the barley crop area in Iran, and second to investigate if diff erent megaenvironments exist. Th e analyses were performed via GGE (genotype + genotype × environment) biplot methodology on the grain yield of 13 barley genotypes, selected from the joint project of Iran/ICARDA, grown for three consecutive seasons (2003-2005) at six locations, representative of major barley growing areas in Iran. Th e by-year and combined analyses showed that the variation between locations was always the most important source of yield variability. Collective analyses of yearly and combined GGE biplots were able to identify high-yielding genotypes and their areas of adaptation, and suggested the existence of four barley megaenvironments. Th e fi rst megaenvironment consisted of the location of Ardabil (ARDA), where genotype G11 was the highest yielder; another megaenvironment is represented by the location of Kermanshah (KERM), with G2 as the best genotype. Th e megaenvironment was comprised of the locations Maragheh (MARA) and Zanjan (ZANJ), with G13 as the highest yielder, whereas the Ghamlo (GHAM) and Shirvan (SHIR) locations made up the fourth megaenvironment, with G4 and G10 as the recommended genotypes. Th e SHIR, ZANJ, and GHAM were more representative and are considered desirable for selecting genotypes adapted to the whole target region. In contrast, ARDA was best at discriminating high-and low-yielding genotypes. Th e highest-yielding genotypes (G4, G3, and G13) were more unstable over locations and years, while high-yielding genotypes (G5 and G8) were more stable.