The growth and yield of winter wheat was studied after waterlogging at different stages of growth under outdoor conditions, using lysimeters (80 cm diameter, 135 cm deep) containing undisturbed monoliths of sandy loam and clay soils, and cylinders (20 cm diameter, 100 cm deep) filled with sandy loam. Soil oxygen concentration fell from 20% to 3% or less after waterlogging to the soil surface for five days at 12°C, but after up to 21 days at less than 5°C. Winter wheat was most sensitive to waterlogging before emergence, when sixteen days of waterlogging killed all seedlings. Six days of waterlogging depressed plant populations by 88% (clay) and 61% (sandy loam) but compensatory growth limited yield losses to around 18%. After emergence, wheat survived surface waterlogging for up to 120 days. In several experiments on the sandy soil, grain yields were depressed by around 8% after waterlogging for seven weeks during seedling growth, six weeks during tillering, and three weeks during stem elongation. Losses of grain from one experiment on the clay were 16% and 12% respectively at the latter two stages of growth. On both soils, yield losses were largest when yields were high (10 t ha-1), and small when yields were low (4 t ha- and influenced by disease, late frosts and high summer temperatures. Waterlogging increased leaf chlorosis, and depressed tillering, although ear numbers were less affected. Tiller survival appeared related to soil nitrogen availability; waterlogging increased nitrous oxide concentrations, and decreased the nitrate loss in drainage, indicating average denitrification losses of around 12 kg N ha-1. Additional nitrogen fertilizer after winter waterlogging increased ear numbers and grain yield, and a higher proportion of the nitrogen in a winter waterlogged crop was obtained from fertilizer nitrogen than in the freely drained crop.