Grazing rates and grazing selectivity of Antarctic copepods (Rhincalanus gigas, Calanus simillimus and copepods 0.8 to 2 mm in size) were determined using incubation experiments comparing plankton community development with and without added grazers. Estimates of grazing rates were corrected for trophic cascade effects using a simple ecosystem model composed of 5 compartments: (1) nanophytoplankton, (2) nanozooplankton, (3) microphytoplankton, (4) microzooplankton (mainly protozoans) and (5) copepods. Copepod grazing rates on the different planktonic compartments were derived by solving for model parameters as constrained by microscopy enumeration of nano- and microphytoplankton and nano- and microprotozooplankton in treatments with and without added copepods. Simulations based on different parameter values and model configurations support the notion that grazing estimates were robust and between 18 and 30 % higher than uncorrected estimates. Furthermore, model analyses suggest the presence of an additional trophic cascade effect due to grazing of R. gigas on small copepods and copepod developmental stages present in incubation bottles. © Inter-Research 2008.