Finger millet is the second most important cereal in Uganda after maize. The yields however, have remained low due to several constraints, such as finger millet blast disease and limited technology options. Therefore breeding investigations were conducted to determine farmer preferred traits, genetic variation, combining ability and genetic effects for head blast disease and head shapes, and other quantitative traits in finger millet. Among other traits, farmers preferred high grain yield potential, brown seed colour, compact head shape, tolerance to blast disease, high tillering ability, medium plant height, early maturity, tolerance to shattering and ease of threshing in new finger millet varieties. Path coefficient analysis indicated that the most important traits were grain mass head-1, tillering ability and reaction to head blast disease. Overall, the high heritabilities and genetic advance (GA) as a percentage of mean revealed the existence of variability which can be utilised through selection and/or hybridisation. The genotype x environment interaction (GEI) and stability analysis showed significant differences due to genotypes (58%), environments (10%) and GEI (32%). Twelve genotypes that combined high yield potential and stability were identified for advancement in the program. Both general (GCA) and specific combining ability (SCA) were significant for most traits, but GCA effects were more important for all the traits except for number of fingers head-1, finger width and panicle width. The Hayman genetic analysis confirmed importance of additive gene action for most of the traits and that additive-dominance model was adequate for explaining genetic variation in finger millet. The results also indicated that yield was controlled by recessive genes whereas blast resistance was controlled by dominant genes. At least two genes, probably three gene pairs and their interactions seemed to control head shape in finger millet. The interactions observed suggest recessive and dominant epistasis, and probably an inhibitor were involved. Seemingly, the gene for curving of fingers, when present in a dominant form prohibits opening of the heads; whereas the recessive form leads to open head shape irrespective of the gene conditions in the other loci. This study forms the baseline for future investigations and the basis for devising breeding strategy on finger millet head shapes. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2013.