In search of new innovations for saving irrigation water, fruit yield response and N-fertiliser recovery of greenhouse grown spring-planted tomato (Lycopersicon esculentum Mill., cv. F1 Fantastic) were assessed as influenced by deficit irrigation, imposed using either conventional deficit irrigation (DI) or partial root drying (PRD). Three irrigation treatments were tested: (1) FULL, control treatment where the full amount of irrigation water, which was measured using Class-A pan evaporation data, was applied uniformly on the two halves of plant-root zone; (2) PRD, 50% deficit irrigation in which wetted and partially dry halves of the root-zone were interchanged every irrigation; (3) DI, conventional deficit irrigation maintained at 50% deficit, compared to FULL irrigation, with water applied on the both halves of the root-zone. During a growth period of 153 days, the highest fruit yield of 145.4 t ha(-1) was measured under FULL irrigation treatment, which was followed by PRD and DI treatments with statistically lower (P < 0.01) yields of 114.6 and 103.4 t ha(-1), respectively. Irrigation water use efficiencies (IWUE) of both deficit treatments were significantly (P < 0.01) higher (52.7% for PRD and 38.3% for DI) compared to FULL irrigation. Nitrogen-fertiliser recovery was over 70%, with no significant difference among the irrigation treatments. Both deficit treatments (DI and PRD) showed lower values of leaf water potential, photosynthetic rate and stomatal conductance compared to FULL irrigation. Before irrigation, xylem-sap abscisic acid (ABA) concentrations were 28% and 38% higher under water-stressed deficit treatments DIO and PRD, respectively, compared to FULL irrigation, and the high ABA concentrations was maintained only under PRD effect, following irrigation. The results of this work suggest that PRD practices can be viable and advantageous compared to conventional techniques to minimise crop-yield reductions during deficit irrigation. (c) 2006 Elsevier B.V. All rights reserved.