Xylem probe measurements in the roots of intact plants of wheat and barley revealed that the xylem pressure decreased rapidly when the roots were subjected to osmotic stress (NaCl or sucrose). The magnitude of the xylem pressure response and, in turn, that of the radial reflection coefficients (σr) depended on the transpiration rate. Under very low transpiration conditions (darkness and high relative humidity), σr assumed values of the order of about 0·2–0·4. The σr values of excised roots were also found to be rather low, in agreement with data obtained using the root pressure probe of Steudle. For transpiring plants (light intensities at least 10 μmol m−2 s−1; relative humidity 20–40%) the response was nearly 1:1, corresponding to radial reflection coefficients of σr= 1. Further increase of the light intensity to about 400 μmol m−2 s−1 resulted in a slight but significant decrease of the σr values to about 0·8. Similar measurements on maize roots confirmed our previous results (Zhu et al. 1995, Plant, Cell and Environment 18, 906–912) that, in intact transpiring plants at low light intensities of about 10 μmol m−2 s−1 and at relative humidities of 20–40% as well as in excised roots, the xylem pressure response was much less than expected from the external osmotic pressure (σr values 0·3–0·5). In contrast to wheat and barley, very high light intensities (about 700 μmol m−2 s−1) were needed to shift the radial reflection coefficients of maize roots to values of about 0·9. Osmotically induced xylem pressure changes were apparently linked to changes in turgor pressure in the root cortical parenchyma cells, as shown by simultaneous measurements of xylem and cell turgor pressure. In analogy to the σr values of the respective glycophytes, the σc values of the root cortical cells of wheat and barley were close to unity, whereas σc for maize was significantly smaller (about 0·7) under laboratory conditions. When the light intensity was increased up to about 700 μmol m−2 s−1 the cellular reflection coefficient of maize roots increased to about 0·95. In contrast to the σr values, the σc values of the three species investigated remained almost unchanged when the leaves were exposed to darkness and humidified air or when the roots were cut. The transpiration-dependent (species-specific) pattern of the cellular and radial reflection coefficients of the root compartment of the three glycophytes apparently resulted from (flow-dependent) concentration-polarization and sweep-away effects in the roots of intact plants. The data could be explained straightforwardly terms of theoretical considerations outlined previously by Dainty (1985, Acta Horticulturae 171, 21–31). The far-reaching consequences of this finding for root pressure probe measurements on excised roots, for the occurrence of pressure gradients under transpiring conditions, and for the non-linear flow-force relationships in roots found by other investigators are discussed.