The Row and turbulence fields in a fully baffled Vessel stirred by a Rushton turbine have been simulated by means of computational fluid dynamics (CFD) techniques. The simulation techniques adopted (the "Sliding-grid" and "Inner-outer" methods) are fully predictive and require no experimental data as boundary or initial conditions. The effect of clearance (C) on the flow patterns in stirred vessels was simulated as the flow field transition observed can be considered a particularly stringent benchmark for the testing of CFD techniques. The results are compared with previously obtained LDA data and show that the double- to single-loop transition experimentally observed when impeller clearance from the vessel bottom is suitably reduced, can be well reproduced by the CFD simulations. The single-loop flow structure present for C/T = 0.15 is compared in detail with the experimental data and good overall agreement is shown between the experiment and simulation. The periodic component of the kinetic energy is well predicted, but the random component is underestimated, a finding also observed in earlier investigations. The mean Row in most of the vessel is also well predicted but the angle to the horizontal of the impeller discharge flow is overestimated. Predictions are also reported with grid refinements and different turbulence models, in an effort to identify means of improving agreement with the measurements. The findings indicate that further improvements in turbulence modelling might be necessary.