FIELD-EVALUATION OF PREDICTIVE MODEL TO CONTROL ANTHRACNOSE DISEASE OF MANGO IN THE PHILIPPINES

Infection estimates determined by a predictive model developed by Dodd et al. (1991b) were used to time fungicide sprays to control anthracn ose disease of mango in the Philippines. For an amount of disease on f ruits after harvest which was acceptable to growers, this approach res ulted in the application of five fewer sprays compared with a standard spray programme used by the growers in a field trial conducted in 199 1-1992. The model predicted only two high anthracnose-risk periods (>4 0% of conidia forming appressoria) throughout the duration of the grow ing period. Rainfall intensity and its time of occurrence during fruit development was found to greatly influence the amount of anthracnose and stem-end rot disease on fruits after harvest. Three relatively str ong precipitations (>20 mm) within a month before harvest resulted in relatively high anthracnose infection of fruits after harvest. At a se cond field trial, rainfall periods during fruit development did not ex ceed 4 mm and resulted in virtually disease-free fruits after harvest, including those not treated with fungicide. Again the use of the pred ictive model resulted in the elimination of five fungicide treatments compared with the standard programme. The amount of rainfall and the t ime of its occurrence should be considered when planning a disease man agement scheme for the control of anthracnose on mango fruit.