Abstract
The utilization of production systems with reduced chemical input renewed the interest in Ustilago nuda and Pyrenophora graminea. The investigations of seed fungicide treatments are more related to their efficacy than to their contribution to yield gain. The data were collected from research and development trials on fungicide efficacy against U. nuda and P. graminea conducted from 2014 to 2020 in Serbia. Partial least squares, multiple stepwise regression and best subset regression were used for statistical modeling. The total number of plants infected with U. nuda and P. graminea per plot differed significantly in the seven-year period. Shifts in the predominance of one pathogen over the other were also shown. Temperature, total rainfall and relative humidity in flowering time (p < 0.001) influenced the occurrence of both pathogens. The strongest impact on yield loss was observed for temperature in the phenological phases of leaf development (p = 0.014), temperature in flowering time (p < 0.001) and total number of plants infected with U. nuda and P. graminea per plot (p < 0.001). Our results indicated that regression models consisting of both biotic and abiotic factors were more precise in estimating regression coefficients. Neither fungicidal treatment had a stable contribution to yield gain in the seven-year period.