Diallel analysis over two environments for blossom-end scar size in tomato

To investigate genetic regulation of blossom-end scar size in tomato (Lycopersicon esculentum Mill.), a half diallel cross including 10 parents was grown under warm fall conditions in Bradenton, Florida, and cool winter conditions in Hazeva, Israel. The parents were a random sample representing all available fresh market tomato breeding lines commercially grown under subtropical field conditions. A randomized, complete block design with three replications was used and the blossom-end scar index (BSI), a measure for scar size relative to truit size, was measured on 40 and 25 fruits per plot in Bradenton and Hazeva, respectively. Analysis of variance for BSI indicated highly significant (P=0.001) variation among parents and among F1's in both environments. In Bradenton and Hazeva both, the difference in average BSI between parents and F1's was not significant suggesting insignificant overall dominance effects. Further partitioning of variation within F1's indicated that general combining ability (GCA) effects were highly significant in both environments, whereas specific combining ability (SCA) effects were significant at P=0.001 in Bradenton, but only at P=0.05 in Hazeva. Estimated variance components for GCA and SCA effects indicated that BSI inherited mainly additively in both environments. Analysis combined over environments indicated that variation in sensitivity to environments was 5-fold higher among parents than among F1's. The genotype x environment variance component was not large enough to justify testing over more than one environment for population improvement purposes and early testing of hybrids. Evaluation of inbred lines, however, may have to be done in more than one environment, especially for 'Suncoast'-derived material. Under temperatures in Hazeva, genotype differences were more pronounced and heritabilities higher than under high temperatures in Bradenton. The estimated overall single plot heritability was 0.63. Breeding lines with a pointed blossom-end morphology (e.g. NC 8276 and NC 140 in this study) generally had small blossom-end scars, and intercrossing of these lines or crossing with non-pointed, moderately smooth breeding lines generally resulted in smooth hybrids.