A model of maize stomatal behaviour has been developed, in which stomatal conductance is linked to the concentration of abscisic acid ([ABA]) in the xylem sap, with a sensitivity dependent upon the leaf water potential (PSI(l)). It was tested against two alternative hypotheses, namely that stomatal sensitivity to xylem [ABA] would be linked to the leaf-to-air vapour pressure difference (VPD), or to the flux of ABA into the leaf. Stomatal conductance (g(s)) was studied: (1) in field-grown plants whose xylem [ABA] and PSI(l) depended on soil water status and evaporative demand; (2) in field-grown plants fed with ABA solutions such that xylem [ABA] was artificially raised, thereby decreasing g(s) and increasing PSI(l) and leaf-to-air VPD; and (3) in ABA-fed detached leaves exposed to varying evaporative demands, but with a constant and high PSI(l). The same relationships between g(s), xylem [ABA] and PSI(l), showing lower stomatal sensitivity to [ABA] at high PSI(l), applied whether variations in xylem [ABA] were due to natural increase or to feeding, and whether variations in PSI(l) were due to changes in evaporative demand or to the increased PSI(l) observed in ABA-fed plants. Conversely, neither the leaf-to-air VPD nor the ABA flux into the leaf accounted for the observed changes in stomatal sensitivity to xylem [ABA]. The model, using parameters calculated from previous field data and the detached-leaf data, was tested against the observations of both ABA-fed and droughted plants in the field. It accounted with reasonable accuracy for changes in g(s) (r2 ranging from 0.77 to 0.81). These results support the view that modelling of stomatal behaviour requires consideration of both chemical and hydraulic aspects of root-to-shoot communication.