Seven lines or varieties of faba beans were top crossed to each of the three different genetic base testers, Alfred,crystal and Sham. The 21 top crosses and ten parental Unes were grown In a random ized complete block design with three replicatfon.s . Four crosses cxceedecj signijىcargly the mid parrent and two of them were sign1fىcan higher than the better parent for number of pods per maui stem. For total shedding percentage, the three crosses exhibited sign ffcant negative heterotic effects relative to mid-parent. WhiLe . the most desirabLe , hezero& effest relative to better parent was obtained by the cross Shani x 61/1311166. The mean squares due to males . females arid males by femaLes were signflcanfor all the studied traits . The results showed that non-additive type of gene action was the more imporcaru part of the total genetic variability for all the studied traits. The male tester Alfred gave the desirable ‘gi effect for no. of set pods, no. of mature pods and shedding percentage of pod setting. Females Giza 2and 61/131.j/ expressed sign ijicwit negative gi effec for shedding % of flowers and total shedding percentage, respectively. The desirable inter- and infra - alLelic interact ions were represent by two, six, three ,flue, two top crossesfor number of flowers, no. of mature pods / main stem. shedding of flowers . shedding of pod setting and shedding percentage, respectively.