The experiment was conducted at the Central Research Station of Bangladesh Agricultural Research Institute, Joydebpur (24°N, 92.3°E, 8 m above sea-level) Gazipur, from 1996-97 to 1998-99. Joydebpur is above river flood level within the Modhupur tract of Agro-ecological zone 28 (BARC, 1997). The soil was silty clay of the Chhiata series of Grey Terrace soil having moderately slow permeability and moderately neutral pH of 6.0 (Bramner, 1996). The soil site was low in organic matter and N (0.1 %). The soil had available P (9.2 ppm), DTPA-extractable Zn (3.4 ppm), available S (9.0 ppm) and available K (0.1 meql 100 g). The experiment was laid out in randomized complete-block design with 4 replications. Plot size was 20 m x 10 m. The treatments were: T1, wheat- mungbean (stubble rernovedj-transplanted, aman rice, fertilizers for all crops based on soil test; T 2, wheat-rnungbean (stubble removed) transplanted aman rice; farmers' fertilizer dose based on survey of farmers' practice; T ,' wheat-mungbean (stubble retainedj-transplanted aman rice, fertilizers based on soil test; T6, wheat-maize-trans- planted aman rice, farmers' fertilizer dose based on survey. The fertilizers for soil- test-based treatments were for wheat 120, 22,66,20 and 2 kg N, P,K,S, and B/ha; mungbean 20,20 and 20 kg N, P and K/ha; maize 120,26,33 and 20 kg N, P, K and S/ha and for rice transplanted aman 90, 26, 33 and 20 kg N, P, K, and S/ha, whereas farmers' practice fertilizer included levels fertilizer for wheal 60, 11 and 25 kg N, P and K/ha; mungbean 0,0 and 0 kg N, P and Kha; maize 60, 13 and 17 kg N, P and K, and for rice 60, 9,17 and 10 kg N, P, K and S kg/ha. Wheat (,Kanchan'), mungbean ('Kanti'), maize ('Bamali') and rice ('BR 32') were used with a spacing of 20 cm in solid line for wheat, 25 cm x 10 cm for mungbean, 75 cm x 25 cm for maize and 25 cm x15 cm for rice. Seeds of wheat were sown on 26 November, 25 November and 8 December in 1996, 1997 and 1998; of mungbean I maize on 10, 13 and 6 April in 1997, 1998 and 1999; and rice was transplanted on 5,9 and I August in 1997, 1998 and 1999 respectively. For wheat, two-thirds of N and all other fertilizers were broadcast at sowing, while the remaining N was top- dressed at crown-root initiation (CRI) stage. Three irrigations were given to wheat, 1 each at crown-root initiation (CRI), maxi- mum tillering (MT) and grain-filling (GF) stages. Irrigation was not given to mungbean and maize crop but 1 supplementary irrigation was given for transplanted man rice. For mungbean, all fertilizers were applied at sowing, while for maize one- third N and the remaining two-thirds N was top-dressed in equal amounts 21 and 42 days after sowing (DAS). For rice, fertilizer other than N was applied at the time of final land preparation, while N was applied in 3 equal splits 15,30 and 45 days after trans- planting (DA T). Mungbean pods were hand-picked twice 60 and 70 days after sowing, and plant residues were incorporated into the soil (stubble retained) and removed (stubble removed). Maize straw was removed from the field but roots were incorporated. Intercultural operations such as weeding, mulching and earthling up (only in maize) were performed as and when necessary. Harvest dates in 1997, 1998 and 1999 were 25, 18 and 23 March for wheat; 25, 28 and 15 June for mungbean; 15, 10 and 10 July for maize; and 16,25 and 29 November for rice respectively. Data were analysed by mean comparison test using LSD. Rice- equivalent yield was calculated using commodity price of individual crop. Overall, the soil may be rated poor in fertility. Soil analysis be- fore sowing of wheat crop (initial) and after 2 years of experimentation, showed that organic matter content (1. 78%) increased in mungbean stubbles incorporated soil.