We present a theoretical model to quantify the influence of diversity on productivity and nutrient acquisition in plant communities during exponential growth. The model fractionates diversity into three components, namely species richness (S), species evenness (E ) and the degree of difference between species (D). The influence of each of these components is assessed individually: S is varied by changing the number of species, E by changing their population size, and D by changing the range of species traits critical to productivity (specific nutrient uptake rate, Σr, or nutrient use efficiency, NUE). D was quantified as the coefficient of variation of Σr or NUE. All three components of diversity enhance the biomass and nutrient stocks in the community, but the response patterns are different. Species richness has a saturating influence, whereas effects of E and D are linear and exponential, respectively. In all cases the non-linear dependence of productivity and nutrient acquisition on Σr and NUE during exponential growth was the single mechanism underlying these effects. This causes the presence of plants with extreme traits to promote productivity, and S, E and D all affect the abundance and/or intensity of these extremes. The model offers a framework to explain differences between experimental observations, and suggests a concept of diversity where S and E are structural components and D a qualitative or functional component, which modulates the influence of the two others. We propose to explicitly recognise D as an integral constituent of plant diversity in future studies.