Improvement of flavor quality is a breeding objective that merits increased attention. To obtain further information on broad-sense heritability of selected sensory attributes, 30 virginia-type genotypes sources were grown in replicated experiments at two locations during 1988 for evaluation. Roasted peanut paste samples were evaluated for 14 sensory attributes. Fruity attribute was confirmed as having a significant suppressive effect on roasted peanut attribute. Session-to-session variation was significant and use of an incomplete block design provided for control of panel variation in the experimental error. Broad-sense heritability estimates for roasted peanut, sweet, and nutty sensory attributes were higher than previously reported, 0.31 vs. 0.24; 0.68 vs. 0.14; 0.37 vs. 0.05, respectively. Nine genotypes were found to have statistically significant higher roasted peanut intensity than the industry accepted standard Florigiant. Further calculations showed that experiments with two replications at each of four locations should have an 80% chance of detecting statistical significance for roasted peanut attribute differences of 0.5 units among genotypes when testing at the 5% level. A similar experiment with two replications at two locations should have a 40% chance of detecting statistical significance.