This article shows that acoustic measurement can be a useful tool to discriminate different apple batches with a low error rate. Starting from the spectrum of the signal recorded by a microphone after the impact of a small hammer on the fruit, 18 key features were identified and used for the classification of apples belonging to 10 different varieties. A capability study of the gage has been done and the results show that the measuring instrument is repeatable and reproducible for the resonance frequency and amplitude of the signal. The prospects of this study are interesting in the context of automatic sorting of fruits.