Hydraulic properties (half-time of water exchange, T1/2, and hydraulic conductivity, Lp; T1/2 approximately 1/Lp) of individual cells in the cortex of young corn roots were measured using a cell pressure probe for up to 6 h to avoid variations between cells. When pulses of turgor pressure of different size were imposed, T1/2 (Lp) responded differently depending on the size. Pulses of smaller than 0.1 MPa, which induced a small proportional water flow, caused no changes in T1/2 (Lp). Medium-sized pulses of between 0.1 and 0.2 MPa caused an increase in T1/2 (decrease in Lp) by a factor of 4 to 23. The effects caused by medium-sized pulses were reversible within 5-20 min. When larger pulses of more than 0.2 MPa were employed, changes were not reversible within 1-3 h, but could be reversed within 30 min in the presence of 500 nM of the stress hormone ABA. Cells with a short T1/2 responded to the aquaporin blocker mercuric chloride (HgCl2). The treatment had no effect on cells which exhibited long T1/2 following a mechanical inhibition by the large-pulse treatment. Step decreases in pressure resulted in the same inhibition as step increases. Hence, the treatment did not cause a stretch-inhibition of water channels and was independent of the directions of both pressure changes and water flows induced by them. It is concluded that inhibition is caused by the absolute value of intensities of water flow within the channels, which increased in proportion to the size of step changes in pressure. Probable mechanisms by which the mechanical stimuli are perceived are (i) the input of kinetic energy to the channel constriction (NPA motif of aquaporin) which may cause a conformational change of the channel protein (energy-input model) or (ii) the creation of tensions at the constriction analogous to Bernoulli's principle for macroscopic pores (cohesion-tension model). Estimated rates of water flow within the pores were a few hundred micro m s-1, which is too small to create sufficient tension. They were much smaller than those proposed for AQP1. Based on literature data of single-channel permeability of AQP1, a per channel energy input of 200 kBxT (kB=Boltzmann constant) was estimated for the energy-input model. This should be sufficient to initiate changes of protein conformation and an inactivation of channels. The data indicate different closed states which differ in the amount of distortion and the rates at which they relax back to the open state.