Reliable estimates of microbial growth yield efficiency (Y = microbial production/substrate utilization) are needed to quantify and predict soil carbon (C) dynamics. We examined patterns of C utilization in two soils, a Paleustoll (USA) and Rhodoxeralf (Australia), under two levels of protozoan grazing (low vs high) when substrate was not limiting. Soil, either amended with unlabeled or C-14-labeled glucose was incubated at 25 degreesC and glucose-C concentration, CO2-C evolution, and microbial biomass-C were determined over a 12-20 h period. Three approaches were used for estimating Y:Y-s = (dS(C) - Sigma CO2-C)/dS(C), Y-b = dB(C)/(dB(C) + Sigma CO2-C), and Y-c = dB(C)/dS(c) where dS(C) is the change in substrate concentration (substrate utilization), Sigma CO2-C the cumulative amount Of CO2-C evolved, and dB(C) the change in microbial biomass (biomass production). Calculation of Y, assumes that all substrate-C utilized, minus that respired, is used for biomass and metabolite production. Calculation of Y-b assumes that substrate use equals biomass-C plus respired-C and does not account for biomass production consumed by grazers. Under low grazing, the three estimates of Y were similar with an average value of 0.58 and 0.55 for the Paleustoll and Rhodoxeralf, respectively. Under high grazing, the value of Y varied depending on the calculation used, with values of Y-b (0.44) and Y-e (0.26) being significantly lower than Y-s (0.67). The total amount of glucose utilized did not vary with protozoan grazing intensity, but a high level of grazing increased the rate of glucose use and significantly reduced the amount of measurable biomass C. Substrate-based yield (Y-s) provided the most reliable C assimilation efficiency estimate under both grazing treatments. (C) 2001 Elsevier Science Ltd. All rights reserved.