Introduction

The origin of the angiosperm flower and its subsequent evolution have been major topics of discussion and controversy for over a century. Because so many of the distinctive synapomorphies of angiosperms involve the flower, its origin and the homologies of its parts are closely tied to the vexed problem of the origin of angiosperms as a group. From a phylogenetic point of view, the origin of angiosperms involves two related problems: identification of the closest outgroups of angiosperms, which may clarify homologies of their distinctive features with structures seen in other plants, and rooting of the angiosperm phylogenetic tree and identification of its earliest branches, which may allow reconstruction of the flower in the most recent common ancestor of living angiosperms. It is this second topic that we address in this chapter (for the first study, see Frohlich and Chase, 2007; Doyle, 2008). This task has become much easier in the past ten years, thanks to molecular phylogenetics.

Ideas on the ancestral flower have varied greatly since early in the last century. Two extremes were euanthial theories, which postulated that the flower was a simple strobilus that was originally bisexual and had many free parts (Arber and Parkin, 1907), and pseudanthial theories, which assumed that the first angiosperms had unisexual flowers with few parts, as in ‘Amentiferae’ (now mostly Fagales), which were later grouped to form bisexual flowers (Wettstein, 1907; review in Friis and Endress, 1990). Later variations on the pseudanthial theory proposed that the angiosperms were polyphyletic (Meeuse, 1965, 1975), while recognition of chloranthoid pollen, leaves and flowers in the Early Cretaceous fossil record (Muller, 1981; Upchurch, 1984; Walker and Walker, 1984; Friis et al., 1986; Pedersen et al., 1991; Eklund et al., 2004) contributed to suggestions that Chloranthaceae, which combine putatively primitive wood and monosulcate pollen with extremely simple flowers, often consisting of just one stamen or one carpel, might provide another model for the ancestral flower (Endress, 1986b; Taylor and Hickey, 1992).