The relationship between incidence of infection by the take-all fungus ( Gaeumannomyces graminis var. tritici ), rainfall and yield of wheat in South Australia

. This paper describes results obtained from an 8-year field trial on a calcareous sandy loam in South Australia. Different crop rotations resulted in varying percentages of plants with take-all [caused by Gaeumannomyces graminis var. tritici (Ggt)]. The results demonstrated that in a wheat-grass/medic pasture rotation, take-all caused an average annual yield loss in wheat of 29%. These yield losses were strongly correlated with disease incidence and rainfall in September (r2 = 0.91, P = 0.07) but only moderately correlated to disease incidence alone (r2 = 0.44, P = 0.09). The level of early infection (at 10 weeks) by Ggt was influenced by spring rainfall in the previous season. A regression model was developed to predict the incidence of take-all in a wheat crop from the incidence of take-all and the August-September rainfall the previous season (r2 = 0.96, P = 0.007) for a wheat-grass/medic pasture rotation (successive host plants).