Comment on 'Two-way protocols for quantum cryptography with a nonmaximally entangled qubit pair'

Three protocols of quantum cryptography with a nonmaximally entangled qubit pair [Phys. Rev. A 80, 022323 (2009)] were recently proposed by Shimizu, Tamaki, and Fukasaka. The security of these protocols is based on the quantum-mechanical constraint for a state transformation between nonmaximally entangled states. However, we find that the second protocol is vulnerable under the correlation-elicitation attack. An eavesdropper can obtain the encoded bit M although she has no knowledge about the random bit R.