Plant diversity experiments generally find that increased diversity causes increased productivity; however, primary productivity is typically measured in the presence of a diverse food web, including pathogens, mutualists and herbivores. If food web impacts on productivity vary with plant diversity, as predicted by both theoretical and empirical studies, estimates of the effect of plant diversity on productivity may be biased. We experimentally removed arthropods, foliar fungi and soil fungi from the longest-running plant diversity experiment. We found that fungi and arthropods removed a constant, large proportion of biomass leading to a greater reduction of total biomass in high diversity plots. As a result, the effect of diversity on measured plant productivity was much higher in the absence of fungi and arthropods. Thus, diversity increases productivity more than reported in previous studies that did not control for the effects of heterotrophic consumption.