The temporal and spatial dynamics of Cercospora leaf spot on susceptible and resistant lines of faba bean grown in or at defined distances from soil with residues infested by Cercospora zonata were examined in South Australia in 2005 and 2006. The disease was first observed on susceptible seedlings 49 days after sowing (DAS) in soil that had been sown with faba bean every 3 years since 1997 (positive soil zone for C. zonata) but was delayed by 1 week in adjacent soil (0 to 16 m away) with no history of cultivation of faba bean (negative soil zone). The incidence of diseased seedlings from 49 to 63 DAS showed a gradient from 4 to 16 m from the infested soil and was significantly greater for susceptible plants grown in the positive versus negative soil zones in field trials conducted in 2005 and 2006 (92 versus 30% in 2005, χ21 = 32.2, P < 0.001; 98 versus 55% in 2006, χ21 = 12.1, P < 0.001). The incidence of Cercospora leaf spot on the resistant line 1322/2 was significantly less (χ26 = 171.7; P < 0.001) than on the susceptible line 'Farah' at that time in both years, with fewer than 5% of the seedlings showing the disease. However, a gradient was shown at 70 to 84 DAS, where disease incidence was significantly greater on line 1322/2 in the positive soil zone than on plants in the negative soil zone in both years (62 and 18%, respectively, with χ21 = 27.9, P < 0.001 in the 2005 trial; and 47 and 6%, respectively, with χ21 = 33.3, P < 0.001 in the 2006 trial). At peak disease severity on Farah, Cercospora leaf spot mean leaf area diseased (%LAD) was severe (85 ± 4.3%) on leaves of the three nodes closest to the soil surface, and much less severe (1 ± 0.6%) in the upper canopy. Defoliation combined with %LAD was used to describe the loss of photosynthetic leaf area (%LPLA) in both cultivars, on both soil zones, in each year. Nonlinear regression analyses using a logistic model described disease development over time on susceptible plants grown in infested soil (e.g., for +12-m blocks within infested soil, y = 2.66 + 46.08/[1 + exp(-0.23 × [X - 40.92])] in 2005 and y = 0.49 + 5.02/[1 + exp(-0.14 × [X - 28.30])] in 2006, where X = DAS and y = %LPLA, with both regressions significant at P < 0.001), whereas an exponential model (e.g., for -12-m blocks from infested soil, y = 0.23 + 0.77 × 1.04X in 2005 and y = 0.44 + 0.56 × 1.04X in 2006, both at P < 0.001) best described disease gradients with increasing distance from the inoculum source. Paired t tests of %LPLA at 77 and 98 DAS showed significant differences in disease severity in the positive versus negative soil zones and a steep gradient in %LPLA from 0 to 4 m from the inoculum source. The role of infested faba bean residue in survival of C. zonata over time was also examined using a pot-bioassay and in situ field assay. When residues were removed from the soil surface or depleted rapidly by animal grazing, the amount of C. zonata inoculum in the soil was significantly less (P < 0.001) than for soil with residue remaining on the soil surface. C. zonata survived in soil and remained infective for at least 30 months after harvest of an infected faba bean crop.