PROBABILITY OF WHEAT QUALITY TRAITS FALLING WITHIN ACCEPTABLE LIMITS

Improving consistency of wheat (Triticum aestivum L.) end-use quality requires simultaneous consideration of a large number of quality traits evaluated in multiple growing environments. Stability analyses have inherent limitations that make analyses of large numbers of intercorrelated variables, or non-normally distributed values, difficult. Univariate and multivariate approaches were used to measure genotypic consistency of wheat quality traits based on the probability of traits falling within acceptable limits. Eighteen wheat genotypes were evaluated across 14 environments for flour protein concentration, mixograph mixing time and tolerance, sodium dodecylsulfate sedimentation volume, and kernel hardness. Acceptability of traits was defined by (i) values falling within chosen upper and lower limits of acceptability across environments, and (ii) values exceeding those for a check genotype at each location. Multivariate probabilities of all five traits falling within upper and lower limits of acceptability ranged from 0.0 to 0.45 among genotypes. Univariate probabilities of acceptance were calculated for each trait. They ranged from 0.03 to 1.00 for genotype-trait combinations and indicated the relative contributions of individual quality traits to the multivariate probability value. Several genotypes had low multivariate probabilities as the result of only one or two traits with low univariate probabilities. The probability based approach provides a simple and flexible decision-making tool to identify genotypes with high probability of providing acceptable quality when grown across multiple environments.