Search

Routine evaluations of cooking time trait in common bean (Phaseolus vulgaris L.) can be performed in different ways resulting in different variables. At the same time, the univariate statistical analysis does not consider the interdependencies between the variables, and may omit important information regarding the genotypes. With this, the objective of this work was to present an alternative proposal for analysis of the cooking time in common bean, allowing the discrimination between genotypes. The experiment used for this approach was conducted under field conditions in the 2017/18 agricultural season in Lages, Santa Catarina, Brazil. The treatments consisted of twelve genotypes, (four parents, structured in two crossings BAF50 x BAF07 and BAF09 x IPR 88 Uirapuru, with their generations F2, F3, F8 and F9). The design used was randomized blocks, with two blocks and two observations in each experimental unit. After the harvest, the response variable cooking time of the grains was measured with a Mattson cooker, considering the cooking time of the 13 initial stems. In the multivariate analysis, the variables cooking time of the second (TCH2), twelfth (TCH12) and thirteenth stem (TCH13) were used based on their significance by the stepwise variable selection method. Multivariate analysis of variance showed differences between genotypes (P<0.05). From the dissimilarity matrix with the Mahalanobis distances and the clustering dendrogram, it was possible to verify the distances of the genotypes derived from crosses BAF50 x BAF07 and BAF09 x IPR 88 Uirapuru. With that, the multivariate analysis enabled the genotypes, additionally the crossing BAF50 x BAF07 showed higher estimates of dissimilarity in the progenies. [ABSTRACT FROM AUTHOR]