Seasonal water use data are presented for 4-year-old Pyrus serotina Rehder cv. Hosui growing in drainage lysimeters and trained onto a Tatura trellis. Weekly water use (WU) was calculated using the mass balance approach. For 8 consecutive weeks during late summer, instantaneous WU was also measured by the compensation heat-pulse technique for measuring sap flow. Although good agreement was found between the two methods for 4 weeks after probe installation, discrepancies increased after this time. Water use was highest in early to mid-January in New Zealand, averaging almost-equal-to 8 liters/tree per day, or 2 liters.m-2 canopy surface area/day. Total water use over the growing season was 1070 liters/tree, or 245 liters-M-2 canopy surface area. The correlation coefficient between weekly WU and evaporation from a nearby Class A pan was 0.81 for the season. Weekly crop coefficients thus calculated for the well-watered trees ranged from 0.15 to 0.55 and 0.20 to 0.83 when calculated using canopy surface area and projected ground area, respectively. Low values were due to low values of canopy leaf area early in the season. Withholding irrigation during three periods resulted in a gradual decline in water use. Water-stressed trees had a lower predawn water potential than fully irrigated trees. This pattern was followed by a more-rapid decline during the morning, and a slower recovery during late afternoon and early evening. Midday leaf water potential never fell below -2.5 MPa.
