Determining the probability of bitter pit in 'Golden Delicious' apples through the post-harvest mineral content of individual fruit

Bitter pit fruit in commercial consignments of apples still poses an economic threat to exporters from South Africa. Mineral analysis of fruit has been used with variable success to predict bitter pit prior to harvest. The possibility of increasing the accuracy of existing predictive models by using analysis of individual fruit rather than pooled samples was investigated. By improving the normality of the distributions of the different minerals and decreasing the overlap between pitted and non-pitted fruit classes, we attempted to improve the reliability of predictions based on variable threshold values. Even though our model produced a correct classification of 85% for non-pitted fruit which can be useful, this was still below the required tolerance expected on the market which, at present, is less than 2% bitter pit in an overseas consignment. The classification for pitted fruit, 63%, was not satisfactory.