Genetic analysis of resistance to bacterial blight in some crosses of cotton (Gossypium hirsutum L.)

Bacterial blight of cotton affects all plant parts during all growth stages. The disease can be extremely damaging under Sudan conditions particularly, to long staple cotton (gossypium barbadense). All commercial cotton cultivars grown in Sudan are susceptible to the new race of the bacterium. Breeding for blight resistance has become a major objective in cotton breeding program. Therefore, if improvement of resistance is to be made, an understanding of its genetical basis is essential. Quantitative analysis to bacterial blight indicated additive, dominance, and epistatic gene action (Wallace and El-Zik, 1990). A substantial number of diallel studies indicated the importance of both additive and dominance effects on bacterial blight resistance (Innes and Brown, 1969; Wallace and El-Zik 1990). Estimates of gene effects involved in the inheritance of bacterial blight resistance from four cotton cultivars crosses are reported in this study. Four cotton cultivars (Acala (93) H, S295, LEBO-1-78 and Tamcot HQ 95) were crossed in a half diallel mating system. These parents were selected on the basis of their moderate resistance to bacterial blight and their different genetic background. Acala (93) H is a newly released cotton cultivar in the Sudan, S295 is a cotton cultivar from Chad, LEBO-1-78, and Tamcot HQ 95 are cotton cultivars from the USA. Parents, F1, F2, and F3 generations were grown at Gezira Research Station field, Wad Medani, Sudan, in July 2000, using a randomized complete block design with three replications. Field inoculation and scoring of disease severity described by Knight (1946) were used. Disease severity was recorded 21 days after inoculation using a scale of 0-10, where zero represents immunity and 10 represents full susceptibility. Adequacy of the additive-dominance model was tested using Hayman and Mather scales (1955). Gene effects for resistance to bacterial blight were estimated following Hayman (1958) using the five-parameter model. The significance of gene effects was tested by calculation variances, standard errors, and 't' value separately for each effect. Analysis of variance indicated significant differences for leaf disease grade among different experimental populations (data not shown). The results obtained by scaling tests and model-fitting procedures to the generation means are shown in Table 1 and 2. A satisfactory fit of expected and observed generation means was obtained with the three-parameter model for all crosses (Table 2). Both simple scaling test C and D were not significant indicating the absence of non-allelic interaction. Therefore, the additive dominance model is adequate to explain genetic variability for blight resistance in these crosses