ABSTRACTDaily patterns of stomatal conductance (gs), xylem pressure potential (P) and canopy microclimatic variables were recorded on 11 sample days as part of a one‐year study of the water use ofEucalyptus grandisHill ex Maiden in the eastern Transvaal, South Africa. Measured gswas found to be largely controlled by quantum flux density (Q) and ambient vapour pressure deficit (D). Canopy conductance (gc) was determined for hourly intervals using gsmeasurements and leaf areas in four different canopy levels. A simple model was constructed to allow the prediction of gcand transpiration from Q, D and season of year. The model was used to estimate transpiration rates from 10 trees in a later study of similarly‐agedE. grandistrees, in which sap flow in each tree was measured using the heat pulse velocity (HPV) technique. Five of the trees were monitored on a summer day and five on a winter day. Correspondence between HPV sap flow and modelled transpiration was good for the summertime comparisons, but measured winter‐time sap flow rates were underestimated by the model, especially under conditions of high sap flow. The discrepancy is believed to result from having insufficient data from the conductance study to describe the response of gsto relatively high D in winter. Marked variation in transpiration per unit leaf area indicates that a relatively large number of trees must be sampled for the HPV technique to be used to obtain a mean rate for an entire stand in w