Utilization of exotic germplasm resources for population improvement in peanut (Arachis hypogaea L.) has increased as the need to increase genetic diversity among peanut cultivars was recognized. Progeny of crosses of two unadapted germplasm lines (GP-NC 343 and FESR-11-P11-32) with an adapted cultivar ('NCV11') of peanut were evaluated for the genetic factors influencing the inheritance of yield and fruit characters in crosses among diverse lines. Objectives were to (1) estimate the relative importance of additive and nonadditive genetic effects in the inheritance of yield and fruit characters in two diverse peanut crosses; (2) determine the proportion of exotic germplasm that gave the optimum combination of mean productivity and genetic variability for each of the crosses; (3) relate the results to theories regarding the transfer of desirable alleles from exotic germplasm into adapted breeding populations. Crosses and backcrosses were made to generate germplasm lines (ten generations) ranging from 0 to 100% exotic germplasm for each cross. The populations were evaluated in replicated field trials. Yield and six fruit characters were measured, and a weighted analysis of variance was conducted to determine if significant differences existed among generations. Generation means analyses were performed for each trait measured in each of the crosses using both three- and six-parameter models, which were tested for goodness-of-fit with a joint-scaling test. Significant differences were detected among generations for most traits measured in both crosses. Estimates of additive genetic effects were significant for pod weight and seed weight in cross 1 ('NC-V11' x GP-NC 343) and for all traits in cross 2 ('NC-V11' x FESR-11-P11-32) except seed∶pod ratio. Significant estimates of dominance effects were found for pod length, pod width, and pod weight in cross 1 and for pod length in cross 2. No significant estimates of digenic effects were observed in cross 1, whereas in cross 2 estimates of additive x dominance epistatic effects were significant for yield and pod length, while estimates of additive x additive effects were significant for seed number. Regression of trait means on generations showed a curvilinear response for all traits in cross 1 except seed weight, which gave a linear response. For all traits in cross 2, the relationship between productivity and proportion of unadapted germplasm was effectively linear. Based on generation means and variances, progeny from the first or second backcross generation to the recurrent parent should be expected to give an optimum combination of mean productivity and relative variability in the population.