In figs (Ficus, Moraceae) there are two breeding systems: monoecy is the ancestral condition but approximately half the 750 odd species are functionally dioecious. Three hypotheses have been proposed for the evolution of dioecy in figs, invoking seasonality (Kjellberg et al. 1987), the reduction of non-pollinating wasp species (Kerdelhue and Rasplus 1996), and the persistence of pollinator populations within small groups of trees (Kameyama et al. 1999). However, there are two major problems with these ideas. Firstly, dioecy has probably evolved only twice (Weiblen 2000), which severely limits our ability to test between alternative hypotheses. Secondly, it is very simple to suggest ways in which dioecy can evolve from monoecy (Charnov 1982). To illustrate this problem, and enlarge on some recent progress in our understanding of functionally dioecious figs, we are proposing a few more hypotheses.