AbstractIn order to evaluate the effects of heterogeneity and heterozygosity on yield stability in faba beans, genotypes were generated with contrasting population structures, differing only in their levels of heterogeneity and heterozygosity. All entries were based on either eight or 36 inbred lines, respectively. The population structures tested consisted of pure stands of the inbred lines (1) and of their F1,‐hybrids (2), blends of the inbred lines (3) and of their F1hybrids (4), four‐component synthetics in different Syngenerations (5), the open‐pollinated source varieties (6) and 36 polycross progenies (7). Five different entry sets (= data sets), each covering several of the seven population structures listed, were grown during 1986 to 1991 in at least four different environments in Western Germany and tested for yield.Stability parameters, i.e., the regression coefficient and deviation from regression (EBERHART and RUSSELL 1966), and the ecovalence (WRICKE 1962), revealed that yield stability of the different population structures was improved by increasing heterozygosity as well as heterogeneity. But the impact of both factors of diversity on yield stability varied between the different data sets. Positive effects of both factors on yield stability seemed to be additively combined in the hybrid blends (data set I). The same proved to be partly true for the four different Syn‐generations of the nine synthetics tested. Yield advantage of the different heterozygous variety structures over the homozygous ones in creased with increasing yield level at test sites, as indicated by regression coefficients larger