Winter cover crops are grown in temperate regions to capture nitrogen, build organic matter in soil, and suppress weeds during autumn and winter. This thesis aimed to identify whether growing species mixtures instead of pure stands strengthens the ecosystem services provided by cover crops. Emphasis was given to productivity, resource capture and yield stability. From 2015 to 2019, I conducted fourteen field experiments across three locations in the Netherlands and one location in Germany.Twenty-five cover crop species were characterized as to their productivity to produce a shortlist of promising species for further studies. I found large differences in productivity between species. Species with early ground cover, such as crucifers and black oats, had the highest productivity and they showed the greatest suppression of weeds. An analysis was made of complementarity and dominance effects in mixtures. Complementarity occurred more often when the mixed species were equally productive, e.g. when combining oilseed radish and black oats, while mixing species with different productivity resulted more often in dominance of the highly productive species in the mixture. Mixing cover crop species within the row or in alternate rows had a small and variable effect on the competitive relationship between mixed species. Across all experiments, mixtures produced 25 to 42% more biomass and captured 11 to 34% more nitrogen than pure stands. This large difference was reduced when the comparison was restricted to the most productive pure stands and species mixtures. In this case, mixtures and pure stands produced similar biomass in most experiments, though, on average, mixtures captured 9% more nitrogen from soil than the best pure stands. In an analysis of data collected in multiple treatments at multiple sites and years, I analysed how the variability of yield differed between mixtures and pure stands. The contributions to yield variability were estimated as variance components using Bayesian mixed models. Yield variability was related to the variation in growing conditions as related to variation in soils and weather between sites and years. At this level there was no difference between mixtures and pure stands. Therefore, I conclude that mixtures do not provide insurance against variability in growing conditions. Within each combination of site and year (site-year), mixing species, however, significantly reduced the yield variability between treatments. This difference was mainly due to the low yield of some species when grown in pure stands, illustrating the importance of species choice. Moreover, mixtures had smaller within-field variation than pure stands. I investigated how biomass formation of cover crops can be separated in the processes of radiation interception and radiation use efficiency. Of the high yielding species, some intercepted large amounts of radiation (crucifers) and some had high radiation use efficiency (oats).Combining species with high radiation interception and high radiation use efficiency did, howev...