Development of a ploughpan has been reported in Bangladesh for almost all ploughed soils which are puddled for transplanted rice cultivation. Field information on the water requirement of dryland crops such as wheat and the effects of loosening the dense layer on crop yield and water use efficiency are very limited. Field experiments were, therefore, conducted in the grey flood-plain soil of Sonatala series (Aeric Haplaquept) to study the irrigation and tillage effects on the yield and water relations of wheat (Triticum aestivum L. cv. Sonalika). The split plot design experiment comprised four irrigation treatments in the mainplots viz. W0 = no irrigation, W1 = irrigation of 5 cm at 4 weeks after planting, W2 - W1 + irrigation(s) of 5 cm each at irrigation water to cummulative pan evaporation (IW/CPE) ratio of 0.75 and W3 - W1 + irrigation(s) of 5 cm each at IW/CPE ratio of 0.50. The sub-plot tillage depth treatments were: A - 7.5 cm (traditional), B - 15 cm, C - 22.5 cm, D - 22.5 cm practised in alternate wheat seasons. Measurements were made of grain and straw yield, soil water depletion and water expense efficiency. Irrigation had no effect on grain or straw yield. Tillage to 15 cm increased wheat yield by about 15% over traditional depth to ploughing. In general, deep tillage coupled with one irrigation at four weeks after planting produced the largest wheat yield. Soil water depletion (SWD) in the 0 - 90 cm profile was greatest in the treatment receiving two irrigations, one at 4 weeks and again at IW/CPE ratio of 0.50. The average SWD in this treatment was 113 in 1982-83 and 82 mm in 1983-84. Plots receiving traditional tillage (7.5 cm) had the greatest SWD. Total water expense were the greatest in treatments receiving three irrigations. The maximum water expense efficiency (WEE) of wheat was observed in the non-irrigated plots in 1982-83 and 1983-84, respectively. Deep tillage treatments, in general, had significantly greater WEE than those under traditional ploughing.