ANALYSIS AND QUANTIFICATION OF SOYBEAN RUST EPIDEMICS FROM 73 SEQUENTIAL PLANTING EXPERIMENTS

Epidemics of soybean rust (caused by Phakopsora pachyrhizi) were studied in Taiwan by sequentially planting soybean cultivars G 8587 and TK 5 at weekly intervals in 1980 and 1981. For most planting dates, lesions were first observed at soybean growth stage V2 or V3 (unrolled leaf occurred at second or third nodes on main stem), indicating the presence of inoculum near the time of planting. The development of disease was affected by the growth of soybean, of which the growth periods varied from 83 days to 207 days for different planting dates. On the autumn-seeded crop, disease reached the maximum level 40-60 days after planting. On the summer-seeded crop, disease onset was 20-40 days later than that of spring- or autumn-seeded crop, and disease reached the maximum level 80-100 days after planting. Physiological days were used as a measure of biological time for both the soybean plant and the rust pathogen. Total leaf area per plant was closely described using plant physiological day as an independent variable with R2 greater than 0.88 (n = 215-235). A disease model using physiological day of plant and physiological day of pathogen as independent variables was used to calculate the disease development with data pooled from all the plantings, and R2 = 0.69 (n = 696) for G 8587 and 0.75 (n = 589) for TK 5. Defoliation was significantly explained with regression on disease severity (R2 = 0.625-0.809, P < 0.0001); however, error increased when disease severity was low. Estimation of yield loss using area under disease progress curve (AUDPC) as an independent variable resulted in a greater R2-value than using area under defoliation curve for each cultivar. The results were significantly improved when AUDPC adjusted for planting date effect was used as an independent variable for the regression.