Epidemic quorum systems enable highly available agreement even when a quorum is not simultaneously connected, and are therefore very interesting for mobile networks. Although recent work has proposed epidemic quorum algorithms, their properties and trade-offs are not well studied. This paper sheds some light on less known aspects of epidemic quorum systems. With simple counter-examples and combinatorial exercises, we contradict common misbeliefs that are often associated with epidemic quorum systems. Our claims advocate the need for a deeper study of these promising systems.