Environmental factors that affect the fumonisin content in maize grain

Abstract

Regression logistic techniques were used to study the environmental variables associated with fumonisin content in maize grain. The dependent variable consisted of 29 values of fumonisins B1 and B2 (FB1 and FB2) from samples of susceptible hybrid sowed in several sites of the Pampas region from 2006/07 to 2008/09. Meteorological variables were recorded in relation to the silking stage. The Np2 variable (number of periods of two days with simultaneous recording of rainfall and relative humidity >= 81 % (day 1) and relative humidity > 70 % at the second day) was the most strongly correlated with FB1 and FB2 (Kendall coefficient: τ: 0.63 and 0.52 respectively) and was included in simple and bivariate regression logistic models with Txm (mean maximum temperature). A trivariate model integrated for two interactive variables and Txm achieved the highest accuracy of prediction, classifying correctly 93 % of cases. The environmental variables identified in this study stimulate F. verticillioides infection and can explain the variability in fumonisin levels at harvest. These results could be useful for the rational management of both disease and mycotoxin.