The coexistence of multiple seed size strategies within plant communities have been considered puzzling, based on a theoretical expectation of the existence of an optimal seed size under each set of specific environmental conditions. A model aimed at explaining the coexistence of different seed sizes has been suggested, where a seed size – seed number tradeoff is connected to a tradeoff between competition and colonization, leading to a competitive advantage in larger-seeded species and a colonization advantage in smaller-seeded species. Recently an alternative model has been suggested, based on a tradeoff between stress tolerance and fecundity, associated with the variation from large to small seeds. Here, we examine the role of seed size for recruitment in two-species contests subjected to various treatments. In a garden experiment seeds of 14 plant species were combined pair-wise into seven pairs, each with one larger-seeded species and one smaller-seeded species. Each species-pair was sown with sparse and dense seed densities and subjected to different treatments of shading and litter. Recruitment was recorded during two years. Our results showed a general advantage of larger-seeded species over smaller-seeded species. This seed size advantage increased in treatments with litter, whereas there were minor effects of shade, and no effect of seed density was found. We thus found little support for a density dependent seed size game as assumed in models of a competition-colonization tradeoff, whereas our results fit well with a model based on a tradeoff between stress tolerance and fecundity. Our experiment provides novel empirical data to theoretical models on co-existence between multiple seed size strategies.