Bean fly (Ophiomyia spp.) is a major pest of common bean (Phaseolus vulgaris L.) throughout eastern and southern Africa. In the semi-arid areas, apart from drought, the insect pest is reported to cause high crop losses up to 100%, particularly when drought occurs and under low soil fertility. Host-plant resistance is part of the integrated pest management strategies that have been widely employed against major insect pests of tropical legumes. However, information regarding its use in control of bean fly in common bean is limited. Therefore, the objectives of this study were to: (1) validate farmers’ perceptions of major constraints responsible for yield losses, particularly the major insect pests of beans; (2) asses the level of adoption of improved bean varieties and determine factors that influence farmers’ preferences of the varieties and criteria for selection; (3) identify sources of resistance to bean fly available in landraces; (4) determine the nature of gene action controlling bean fly resistance and seed yield in common bean; (5) describe a procedure for generating optimal bean fly populations for artificial cage screening for study of the mechanisms of resistance available in common bean against bean fly. Farmers considered drought and insect pest problems as main causes for low yields. The adoption rate for improved varieties was high but self-sufficiency in beans stood at 23% in the dry transitional (DT) agro-ecology and at 18% in the dry mid-altitude (DM) agroecology, respectively. Drought, earliness, yield stability, and insect pest resistance were the factors determining the choice of varieties by farmers. Bean fly (Ophiomyia spp.), African bollworm (Helicoverpa armigera) and bean aphid (Aphis fabae) were identified as key crop pests of beans limiting yield. The study to identify new sources of resistance included 64 genotypes consisting of landraces, bean fly resistant lines and local checks. The experiment was done under drought stressed (DS) and non-stressed (NS) environments and two bean fly treatments (insecticide sprayed and natural infestation) for three cropping seasons between 2008 and 2009. Genotypes differed in their reaction to natural bean fly attack under drought stressed (DS) and non-stressed environments (NS) over different cropping seasons. However, the effect of bean fly appeared to vary between the long rains (LR) and short rains (SR). It was observed that an increase in the number of pupae per stem resulted in a higher plant mortality. The range of seed yield was from 345 to1704 kg ha-1 under natural infestation and from 591 to 2659 kg ha-1 under insecticide protection. Seed yield loss ranged from 3 to 69 %. The resistance of most of the bean fly resistant lines seemed to be ineffective in presence of DS. To determine the nature of gene action controlling the inheritance of resistance to bean fly, four parents with known reaction to bean fly were crossed with four locally adapted genotypes in an 8 x 8 half-diallel mating design. Similarly, two resistant and two susceptible parents were selected and crossed to produce populations for generations means and variance components analysis. Results revealed that both general combining ability (GCA) and specific combining ability (SCA) mean squares were significant (p A 0.05) for all four traits studied, except SCA for stem damage during one cropping season. Among the parents, GBK 047858 was the best general combiner for all the traits studied across seasons except for stem damage during LR 2009. Genotypes GBK 047821 and Kat x 69 (a locally adapted variety) were generally good general combiners for resistance traits as well as seed yield. General predictability ratio values ranging from 0.63 to 0.90 were obtained for plant mortality, stem damage, pupae in stem and seed yield across cropping seasons. These results established the predominance of additive gene effects (fixable variation) over the non-additive effects in controlling the traits. Low to moderate narrow sense heritability values ranging from 0.22 to 0.45 were obtained for pupae in stem. Such heritability estimates indicate that although additive gene components were critical in the inheritance of resistance for the trait, non-additive gene action was also important in addition to the environmental effects. A major disadvantage in screening for resistance to bean fly in common bean by controlled means in net cages has been the lack of a method to use for raising adequate fly populations for screening. Due to this problem, a simple procedure for raising sufficient numbers of adult bean flies required for screening was described. Through this method, up to 62 % emergence of the adult flies was achieved. Moreover, the flies retained their ability to infest bean plants. To determine the presence of antibiosis and antixenosis mechanisms of resistance in common bean, five genotypes [CC 888 (G15430), GBK 047821, GBK 047858, Ikinimba and Macho (G22501)] and two local check varieties (Kat B1 and Kat B9) were screened under free-choice in outdoor net cages and no-choice conditions in net cages placed in a shadehouse. All the five resistant genotypes tested had relatively long internodes. It was established that long internode was a morphological trait associated with reduced pupation rate in bean stems, hence an antixenosis component of resistance. Both ovipositional non-preference and antibiosis mechanisms were found to exist in three genotypes namely CC 888 (G45430), GBK 047858 and Macho (G22501). These genotypes were resistant when they were subjected to bean fly under both free-choice and no-choice conditions. They had fewer feeding/oviposition punctures, low number of pupae in the stem, reduced damage to the stems and low percent plant mortality. The remaining genotypes, Ikinimba and GBK 047821 only expressed antixenosis. To maximize the effectiveness of host-plant resistance against bean fly, multiple insect resistances should be incorporated into a single bean genotype in order to ensure durability. However, this should be within the background of integrated pest management strategy. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2010.