Within the framework of a long-term experiment started in 1986, a 3-year field study was carried out under rainfed Mediterranean conditions to determine the effect of tillage, crop rotation and N fertilization on the efficiency of nitrogen in wheat. Tillage treatments included no tillage (NT) and conventional tillage (CT). Crop rotations were wheat-sunflower (Helianthus annuus L.) (WS), wheat-chickpea (Cicer arietinum L.) (WC), wheat-faba bean (Vicia faba L.) (WB), wheat-fallow (WF) and continuous wheat (WW). Nitrogen fertilizer rates were 0, 50, 100 and 150 kg N ha(-1). A split-split plot design with four replications was used. The high rainfall recorded over the three study years (60% above average) had a negative influence on wheat growth and yield. Soil NO3--N and mineralized N content did not vary with tillage treatment. Residual NO3--N was greater in WW and in the WE and WF rotations, whereas lower levels were recorded with the WS rotation. Mineralized N levels fell sharply for WW and were higher for WS and WE Wheat yield was greater with CT than with NT. Clear differences in yield were recorded for all 2-year rotations with respect to WW. Wheat yield showed no additional response to N fertilizer rates above 100 kg ha(-1). N use efficiency (NUE) and N uptake efficiency (NUpE) were greater with CT than with NT These indices, together with N utilization efficiency (NU,E), N harvest index (NHI), N physiological efficiency (NPE), N agronomic efficiency (NAE) and N apparent recovery fraction (NRF) generally pointed to greater NUE for the WS rotation, and poor efficiency for wheat monoculture. Moreover, all indices except NPE fell as N fertilizer rates increased, indicating poor N utilization. (C) 2001 Elsevier Science B.V. All rights reserved.