Comment on “First-order phase transitions: Equivalence between bimodalities and the Yang–Lee theorem”

I discuss the validity of a result put forward recently by Chomaz and Gulminelli [Physica A 330 (2003) 451] concerning the equivalence of two definitions of first-order phase transitions. I show that distributions of zeros of the partition function fulfilling the conditions of the Yang–Lee Theorem are not necessarily associated with nonconcave microcanonical entropy functions or, equivalently, with canonical distributions of the mean energy having a bimodal shape, as claimed by Chomaz and Gulminelli. In fact, such distributions of zeros can also be associated with concave entropy functions and unimodal canonical distributions having affine parts. A simple example is worked out in detail to illustrate this subtlety.