Functional redundancy predicts that some species may play equivalent roles in ecosystem functioning therefore conferring a kind of ‘insurance’ to perturbation when species richness is reduced, by the compensation of species of the same functional group on ecosystem processes. We evaluate functional redundancy on grassland plant communities by a removal experiment in which the evaluated treatments were: GG – clipping two graminoid species, FF – clipping two forb species, GF – clipping one graminoid and one forb species and Control – no removal. We tested the hypothesis that the above-ground biomass removal of one species of each functional group would cause less change in the community composition (community persistence) and less decrease in biomass production than the above-ground biomass removal of two species of the same functional group. Functional redundancy was corroborated for community persistence since treatments FG and C caused less change in community composition than treatments GG and FF, although no differences were found between treatments for above-ground biomass. We verified that clipped species tend to be compensated by an increase in the percent cover of the remaining species of the same functional group. This work provides experimental evidence of early responses after plant clipping in small spatial scale of functional redundancy in naturally established grassland plant communities. We highlight redundancy as an intrinsic feature of communities insuring their reliability, as a consequence of species compensation within functional groups.