At Rothamsted during 1997/98, 1998/99 and 1999/2000 winter oilseed rape growing seasons, numbers of air-borne ascospores of Leptosphaeria maculans were often > 4 m(-3) from autumn (September/October) to spring (April/May), while few or no ascospores were detected during the summer. Mature pseudothecia were generally not observed on debris of the previous crop until September. One-year-old debris (harvested in July 1998) had 95% discharged and 5% mature pseudothecia in August 1999, but by 15 September new pseudothecia (of which 30% were mature) were observed and the first increase in air-borne ascospores (> 4 m(-3) ) occurred. Phoma leaf spotting appeared in untreated field plots 14-25 days after the first increase in air-borne ascospores in autumn. The fungicide mixture difenoconazole plus carbendazim decreased the incidence of new leaf lesions for 1 month after application in autumn and for 2 months in mid-winter. When L. maculans was isolated from infected leaves, the growth rate of isolates from leaves to which fungicide was applied was less than that of those from untreated leaves. Foliar applications of fungicide to field plots in the autumn and winter not only decreased the incidence of crown cankers but also reduced the rate of canker development on stem bases in the spring and early summer (when severity of crown cankers increased linearly with time). In untreated crops, when phoma leaf spots appeared early in the autumn, crown cankers developed early in the spring but only became severe enough before harvest to reduce yield greatly in 1997/98. Yield loss was associated with crown cankers that girdled more than half of the stem by harvest (mean severity > 3 on a 0-5 scale). Infections of new leaves produced after stem extension, from January onwards, led to phoma stem lesion development above the crown. In the three seasons, phoma stem lesions became moderately severe (> 2) by harvest only in untreated plots in 1997/98.