Plant species effects on microbial communities are attributed to changes in microbial community composition and biomass, and may depend on plant species specific differences in the quality of resources (carbon) inputs. We examined the idea that plant-soil feedbacks can be explained by a chance effect, which is the probability of a highly productive or keystone plant species is present in the community and will influence the functions more than the number of species per se. A C-13 pulse labelling technique was applied to three plant species and a species mixture in a greenhouse experiment to examine the carbon now from plants to soil microbial communities. The C-13 label was given as CO2 to shoots of a legume (Lotus comiculatus), a forb (Plantago lanceolata), a grass (Holcus lanatus) and a mixture of the three species. Microbial phospholipid fatty acids (PLFA)was analysed in order to determine the biomass and composition of the soil microbial community. The incorporation of the stable isotope into soil microorganisms was determined through GC-IRMS analyses of the microbial PLFAs. Plant species identity did not influence the microbial biomass when determined as total carbon of microbial phospholipid fatty acids. However, the labelled carbon showed that the grass monoculture (H. lanatus) and the plant mixture allocated more C-13 into bacteria and actinomycete biomass than the other plant species. H. lanatus monocultures had also the highest amounts of C-13 allocated to AM-fungi and saprophytic fungi. The carbon allocation from plants to soil microorganisms in a plant species mixture can thus be explained by the presence of a highly productive species that influence soil functions. (C) 2009 Elsevier Ltd. All rights reserved.