Genotype-by-Environment Interaction in Sweet and Bitter Sensory Attributes of Peanut

Abstract Although there have been recent studies of the roasted peanut sensory attribute of roasted peanut (Arachis hypogaea L.), there is little information on the factors influencing the sweet and bitter attributes. A study was conducted to test the significance of the effects of genotype, environment, and their interaction on the sweet and bitter attributes, to estimate genotypic means of these sensory attributes, and to determine an efficient scheme of resource allocation for multiple-site testing of breeding lines for sweet and bitter attributes. Samples of sound mature kernels from 17 genotypes grown at 42 different location-year combinations were stored under 5 C and 60% RH, roasted to nearly common color, ground into paste, and tasted by a trained sensory evaluation panel. Results for the roasted peanut attribute were similar to those found in previous studies. Significant variation among years was observed for bitter but not for sweet. There were no consistent differences among the three main peanut production regions for either attribute, but there was significant year-by-region interaction for sweet. Variation among locations within years and regions was significant for both attributes as was variation among genotypes. The mean flavor profile for the runner market class was significantly better than the mean for the virginia class in sweet, bitter and roasted peanut, but the distributions of the two classes did overlap. Significant correlations of genotypic means for the three attributes indicated that chemical assays for currently unidentified sweet or bitter principles could be used for indirect selection to improve roasted peanut attribute without the need for expensive and time-consuming sensory evaluation. Significant genotype-by-year interaction for bitter makes it necessary to acquire sensory data from 2 yr with three replications at each of four locations or two replications at each of five locations to differentiate statistically between means differing by half a flavor intensity unit. The precision of comparisons for sweet and roasted peanut will be more precise than that for bitter at a given arrangement of testing resources.