Heritability and number of genes controlling partial resistance to Exserohilum turcicum in maize inbred H99.

摘要

High levels of partial resistance are essential to manage northern corn leaf blight (NCLB) of maize, incited by the casual agent Exserohilum turcicum, in disease-prone environments. Maize inbred H99 displays high levels of partial resistance to E. turcicum and has been used as a donor line for the improvement of host resistance in susceptible temperate-adapted germplasm. This study was undertaken to determine the potential of H99 as a prospective donor of partial resistance to E. turcicum in a disease-prone, mid-altitude tropical environment of East Africa. Estimates of the heritability of resistance in, and the number of genes for, resistance to E. turcicum in H99 were determined using 120 F_2:3 families and their F_2:4 progenies derived from a cross between H99 and susceptible inbred A619. The lines were evaluated in replicated field experiments at Namulonge and Kamenyamiggo, Uganda, during 1993- Disease severity was assessed by counting lesion number, visually estimating percent leaf area affected (PLAA), and calculating the area under disease progress curves (AUDPC). Heritability estimates were computed from variance component analyses and parent-offspring regressions. Broad-sense heritability estimates of F_2:3 and F_2:4 lines were determined using plot-mean and family-mean bases across locations. Estimates of the number of genes controlling quantitative resistance to E. turcicum based on the Castle-Wright formula were made using the square of differences of pheno-typic means of inbreds A619 and H99, and the square of the phenotypic range between the most susceptible and most resistant families in each generation. Heritability estimates for PLAA, and AUDPC based on PLAA, were higher than those for lesion number and AUDPC based on lesion number, and ranged from 0.73 to 0.78. Narrow-sense heritability estimates calculated using parent-offspring regression also tended to be higher than those based on lesion number data. Gene number estimates were two to three for lesion number and four to five for PLAA based on the parental phenotypic means. Gene number estimates based on the phenotypic range of families were approximately eight for lesion number and five to seven for PLAA. Use of a modified Castle-Wright formula raised all gene number estimates.