Genetic Progress of Plant Architecture Using Recurrent Selection in Common Beans

Common bean (Phaseolus vulgaris) farmers in Brazil have been looking over time for plants with better upright architecture,which was found to be a complex trait with quantitative nature. Thus, recurrent selection (RS) represents the best breeding technique to successfully attain genetic gains. Plant architecture is evaluated using a grade scale, which usually has a high heritability, so RS to this trait has been done upon visual selection on S₀ generation.

This study aimed to verify if we had genetic progress and if the genetic gain estimative changes when different number of progenies are analyzed by generation or when different generations are considered. The effect of RS to grain yield was also assed. Populations S_0:3 and S_0:4, from the fifth, eight and twelfth RS cycle were used. Combined analysis of variance based on adjusted means of 47 best progenies from each cycle and generation, using two common controls as reference was carried out. The joint analyses of the two generations in each cycle were also performed.

Genetic gains were estimated using linear regression between the adjusted mean of each cycle.  RS was efficient and the genetic progress based on average of progenies was 4.46%. When fewer progenies were evaluated, the genetic gain was also of smaller magnitude. Our results support that it is possible to successfully select for increasing grain yield after three or more cycles of selection upon plant architecture, making possible to obtain upright high yielding plants.