Interesting question, and it is true that if females were polymorphic like described by Härdling and Bergsten (2005 Am Nat) this would have to be taken into account. However none of the 13 species of Acilius in the present study have polymorphic females. Two forms of females (one structured and one male-like non-structured) are only known for some diving beetle species and genera including e.g. Graphoderus and Dytiscus as referred to by Härdling and Bergsten (2005) and also in species of some smaller diving beetles e.g. Hydroporus, Hygrotus. Taking polymorphism into account in a similar phylogenetic context would probably require that two characters were scored. One for polymorphism itself and infer the origins and losses by scoring a character of polymorphism as present or absent for terminal taxa. A second character for the presence or absence of the structured female morph, scored as present for both monomorphic (if structured) and polymorphic conditions. Inferred origins in this second character should be unproblematic, but inferred reversals would then have to be interpreted in light of the optimized first character. The first character would inform whether the reversal was a loss from a polymorphic ancestor or a true reversal from one state to another. In principal inferring the evolutionary history of traits that are polymorphic in recent taxa are not different from other traits since all new traits start as polymorphic between the time of mutation to fixation in a species. The question is whether this time will span any speciation events, something usually unknown for “normal” traits, but could be tested if traits are polymorphic in some recent taxa.