Correlative studies have shown a ‘hump-backed’ relation between the vegetation N:P ratio and plant species diversity with the highest diversity at balanced N:P ratios (between 10 and 14). We tested the hypothesis that adding growth-limiting nutrients to mesotrophic grasslands that were in shortage of either N (N:P ratio<10) or P (N:P ratio>14) would lead to an increase of plant diversity. Thereto, we studied the effects of long-term (11 yr) experimentally increased N and/or P supply on soil nutrient pools, vegetation nutrient dynamics and biodiversity in a riverine grassland in the Netherlands with a low soil N:P ratio (N shortage) and a peat grassland with a high soil N:P ratio (P shortage), respectively. Eleven years of nutrient addition hardly had any effects on the total stocks of C, N and P in the soils of both sites, due to the large size of the soil nutrient pools already present and to the management at both sites (annual hay-making and -removal). However, in the riverine grassland the treatments increased the cycling of the small pool of labile N and P compounds resulting in large increases in annual fluxes of especially N. In the unfertilised controls, species establishments balanced more or less species losses during an 11 year period, thus leading to a dynamic equilibrium of the species pool. However, contrary to our hypothesis, addition of the growth-limiting nutrient led at both sites to a reduction of species diversity even when total biomass remained below critical levels. Species diversity and species evenness were strongly determined by N mineralisation and to a lesser extent by total soil N and extractable P, respectively. Total aboveground biomass of the vegetation was determined by total soil N.

Our study shows that patterns found in correlative studies of the relation between plant diversity and soil and vegetation N:P ratio can not be translated into successful experimental manipulations to enhance biodiversity. The most likely explanation is that colonization limitation occurred in the fertilized plots and that not sufficient diaspores of potentially new species could reach and/or colonize the plots to compensate for the species extinctions as a result of increased nutrient supply.