We compare different notions of stability in three firm merger games. We discuss some of their shortcomings and introduce an alternative notion of stability which overcomes them. The paper concludes with an illustrative example.