The paper reviews theoretical considerations and empirical facts on the productivity of mixed species forest stands versus pure stands. The theoretical part first draws attention on the combined effect of facilitation and competition and the net result for both individual and mean tree growth in pure and mixed forest stands (Figs. 1-3). By species mixing, competition as well as facilitation may be changed. Mixing effects can modify the productivity of individual trees and the productivity of the mean tree in a stand, and they can also modify the carrying capacity, i.e., the stand density of mixed versus pure stands (Figs. 4). The combination of mean tree productivity and density results in the overall stand productivity. Stand productivity of mixed versus pure stands can be quantified by ratios of relative productivity (LER, RYT, RPR) and illustrated by cross-diagrams (Fig. 5). How mixed stands come off in terms of over-or underyielding on different sites is conceptualized by the stress-gradient-hypothesis. Latter predicts that positive effects dominate on poor sites due to increased facilitation, while neutral or negative mixing effects in terms of productivity are predicted for rich sites where competition dominates (Fig. 6). In the second part of the paper these theoretical concepts are scrutinized on the basis of long-term experimental plot data from mixed stands experiments in Norway spruce/European beech, Sessile and Common oak/European beech, Scots pine/European beech, and Norway spruce/Silver fir. The included mixed stand trials consist of triplets of plots with pure stands of both species as well as one or more mixed species stands differing in mixing portion, mixing structure or stand density. The oldest of these plots are under survey since about 100 years. The experiments are located along an ecological gradient from Denmark through Poland and Germany to Switzerland; with a concentration in Bavaria and Lower Saxony in Germany. Evaluation of above experiments shows that mixing can increase mean tree productivity (dry matter) up to 60% and stand density up to 50% (Fig. 7). While the conifers react on mixing with productivity rise, deciduous tree react with an increase of the packing density. On average productivity increases by 25-50% (Fig. 8). The data reflects a strong variation depending on tree species combination, mixing portion, mixing structure and site conditions. In most cases both species contribute to the positive mixing effect. Positive mixing effects on species 1 are not on the expense of the productivity of species 2, but mostly coupled with also positive reactions of the admixed species. Latter reaction pattern indicates predominance of mutualistic interactions. Analysis of the relation between mixing effects and site conditions (site index) revealed a rather general pattern for all mixtures: Mixing effects in terms of over-yielding are the highest on poor sites and diminish the better the sites are (Fig. 9 and Table 1). As causes of the mixing reactions spatial and temporal niche complementary and the involved competition reduction are discussed. Furthermore, effects of facilitation by mutual support of nutrients exploitation, water capture, or humus improvement are discussed. The relevance of the revealed over- and under-yielding for a resource-efficient landscape management and for adaptation of forest ecosystems to climate change and other disturbances are discussed. Finally, we sketch the impact of our empirical findings on the modelling of mixed species stands and the theory of ecology.