Crack in the shell of eggs was detected by rolling eggs on a plate that had seven steps. This produced seven separate acoustic signals to produce characteristic vectors. The vectors from intact and cracked eggs were analyzed using the Mahalanobis distance so that intact and cracked eggs could be separated. The receiver operating characteristic curve and area under curve were employed to determine the discrimination threshold and optimize the step number - in this case being six. The analysis indicated that the method proposed in this paper could detect 90% of the cracks with a 10% false rejection. (C) 2014 Published by Elsevier Ltd.