The effect of four-wall shear rates (34.9, 74.8, 142.5, and 194.5 s(-1)) on bacterial deposition on glass slides in drinking water flow chambers was studied. Biofilm image acquisition was performed over a 50-day period. Bacterial accumulation and surface coverage curves were obtained. Microscopic observations allowed us to obtain information about the dynamics and spatial distribution of the biofilm. During the first stage of biofilm formation (210-518 h), bacterial accumulation was a function of the wall shear rate: the higher the wall shear rate, the faster the bacterial deposition (1.1 and 1.9 x 10(4) bacterial cells . cm(-2) for wall shear rates of 34.9 and 142.5 s(-1), respectively). A new similarity relationship characteristic of a non-dimensional time and function of the wall shear rate was proposed to describe initial bacterial deposition. After 50 days of exposure to drinking water, surface coverage was more or less identical under the entire wall shear rates (7.44 +/- 0.9%), suggesting that biofilm bacterial density cannot be controlled using hydrodynamics. However, the spatial distribution of the biofilm was clearly different. Under low wall shear rate, aggregates were composed of bacterial cells able to "vibrate" independently on the surface, whereas, under a high wall shear rate, aggregates were more cohesive. Therefore, susceptibility to the hydraulic discontinuities occurring in drinking water system may not be similar. In all the flow chambers, significant decreases in bacterial biomass (up to 77%) were associated with the presence of amoebae. This grazing preferentially targeted small, isolated cells.