Relationship between the incidences of ear and spikelet infection of Fusarium ear blight in wheat

There is a urgent need to develop a rational strategy for managing Fusarium ear blight in order to reduce current reliance on routine fungicide applications, based on an objective assessment of disease risks. One of important components for such a management strategy is a fast, easy, accurate and reliable method for disease assessment. The relationship between incidence of Fusarium ear blight ear infection and number of spikelets infected on an ear ( or incidence of spikelet infection) were investigated during three seasons and in four countries in order to derive a simple relationship for predicting disease at the spikelet level using ear incidence. More than half of the data sets of the number of infected spikelets on an ear could not be fitted satisfactorily by a Poisson distribution. Three two-parameter discrete distributions (negative binominal, Neyman type A and Polya-Aeppli) provided a significantly better fit than the Poisson distribution, indicating a degree of aggregation of number of infected spikelets on an ear. Taylor's power-law satisfactorily described the observed variance-mean relationship for the number of infected spikelets on an ear; this relationship was generally consistent over years and countries. A robust relationship between incidence of ear infection and average number of infected spikelets per ear was obtained assuming a fixed variance mean relationship and a negative binomial distribution for the number of infected spikelets. A relationship between incidences of spikelet and ear infection was also obtained based on the complementary log-log or logit transformation of ear and spikelet infection incidence. These models appeared to be consistent over years and countries and thus may be used in making practical disease management decisions involving fungicide applications.