The maximum turbulence energy dissipation rate per unit mass, epsilon(max), is an important variable in dispersion systems, particularly for drop breakup and coalescence, and for gas dispersion. The effect of tank geometry (number of baffles, impeller diameter, and off-bottom clearance) on epsilon(max) for four impellers (the Rushton turbine, RT; the pitched blade turbine, PET; the fluidfoil turbine, A310; and the high-efficiency turbine, HE3) is examined. Mean and fluctuating velocity profiles close to the impellers were measured in a cylindrical baffled tank using laser doppler velocimetry. Local and maximum turbulence energy dissipation rates in the impeller region were estimated using epsilon = Av(3)/L with A = 1 and L = D/10 for all four impellers. Factorial designs were used to test for the effects of single geometric variables under widely varying conditions and interactions between variables. Several factorial designs were used to ensure that real effects were separated from effects that appeared as an artifact of the experimental design. Results show that the rank geometry has a significant effect on epsilon(max), primarily with respect to variations in impeller diameter and interactions between the off-bottom clearance and impeller diameter : For the same power input and tank geometry the RT consistently produces the largest epsilon(max) and/or epsilon(max) scaled with (ND2)-D-3.