Stability through years in a non-balanced trial of globe artichoke varietal types.

Abstract Yield and precocity performance through the years of 10 of globe artichoke ( Cynara cardunculus var. scolymus ) varietal types were evaluated in a non-balanced trial. The varietal types stability was analyzed by four different statistical methods (Fisher's least significant difference, relative yield, consistency of performance and AMMI1) to explore the advantages and disadvantages of each one in selecting stable genotypes with high yield and precocity. All methods identified “Gauchito” as the most stable varietal type for yield, for this reason it is highly recommended in Argentine's cultivation area. With respect to the precocity, “Violeta de Provence” and “Frances” were the most precocious and stable, followed by “53B3” and “Gauchito”. All methods agreed in classified some varietal types as stable or non-stable, nevertheless, LSD rankings were the least consonant with the others. The AMMI1 model could be considered as the most reliable analysis because it summarizes genotype × environment interaction patterns through analysis of variance and principal components, diminishing the possibility of misleading classification of a varietal type.