In rice (Oryza sativa L.) culture the effect of puddling (wet tillage) on puddle quality, weed growth and yield of crop depends upon initial soil manipulations by pre-puddling tillage. However, the role of pre-puddling tillage on these aspects has not been studied adequately. These effects were studied for three years (1994±1996) in a ®eld experiment with a rice±wheat (Triticum aestivum L.) cropping system at Punjab Agricultural University, Ludhiana. Treatments included pre-puddling tillage treatments no tillage (PT0), one discing one harrowing (PT2) and one discing and three harrowings (PT4)) in rice in combination with four tillage systems, varying in depth and intensity of soil disruption in wheat on puddle quality, weed growth and yield of rice and wheat on a sandy loam soil (Dystric Cambisol). Pre-puddling tillage improved puddle quality in terms of increased puddle depth and tended to decrease percolation rate. Weed infestation in rice decreased with increase in intensity of pre-puddling tillage. Mean dry weed biomass 35±40 days after transplanting was 1.6 Mg haÿ1 in PT0, 0.6 Mg ha ÿ1 in PT2 and 0.5 Mg ha ÿ1 in PT4. Leaving some area untilled between rows (strip tillage) in wheat resulted in a larger weed biomass in rice, than with inversion of soil. Pre-puddling tillage did not affect rice yield during the ®rst two years but signi®cantly increased it during the third year when rice yield was 4.1 Mg haÿ1 in PT0 compared with 5.8 Mg ha ÿ1 in PT2 and 6.1 Mg haÿ1 in PT4. Manual weeding at 40 days after transplanting masked the effect of pre-puddling tillage on rice yield. In general, rice yield decreased exponentially with increase in weed biomass recorded at harvest of the rice crop. Pre-puddling tillage did not affect wheat yield signi®cantly. The results suggest that for effective weed control, high rice yield and water use ef®ciency, the ®eld must receive pre-puddling tillage at least once. # 1999 Elsevier Science B.V. All rights reserved.