Plant community functional structure may drive ecosystem functions in relation with (i) the trait values characterizing dominant species, according to the "biomass ratio hypothesis" proposed by Grime, and (ii) thanks to trait dissimilarity among species, according to the "diversity hypothesis" proposed by Tilman. Both propositions have already yielded support but their relative importance and how they impact biomass production in field situations is still not well known. This study therefore tested (i) whether or not there was a close relationship between biomass production and the community-weighted mean trait values (CWM), as expected from the "biomass ratio hypothesis", and (ii) the impact of the functional diversity (FDQ) on biomass production, which is expected to be positive according to the "diversity-hypothesis". The study considered a range of plant assemblages occurring in a wet grassland and five above-ground and four below-ground plant traits were measured to characterize their functional structure. The effects of species diversity, soil water content, soil nitrogen availability and grazing intensity on biomass production were also determined. We showed that biomass production was not related either to species richness and diversity or to any of the resource and disturbance parameters considered. Conversely, the functional structure was found to explain up to 55% of the variability of the biomass production. The results obtained clearly supported the "biomass-ratio hypothesis". Functional diversity was mainly found to negatively impact biomass production with only poor support to the "diversity hypothesis". We suggest that such a dilution effect of increasing FDQ on community primary production may be typical of fertile habitats. In order to significantly improve our understanding of the relationship between functional diversity and ecosystem processes, further studies should consider plant assemblages that have been shaped over the long term and habitats across a wide range of productivity.