Leaf morphology in angiosperms varies both within and across species, suggesting that leaf characteristics are highly labile in development and evolution. This apparent lability in leaf morphology has important ecological and evolutionary implications. For instance, the compound leaf form may have an impact on processes such as photosynthesis and temperature control via transpiration. The extent of variation and plasticity of leaf form is known to differ in different taxonomic groups, and this might point to developmental or phylogenetic constraints in evolution. In order to understand factors underlying the evolution of compound leaves, we examine systematic patterns in the distribution of this feature among angiosperms. Using phylogenetic hypotheses based on analyses of rbcL sequences (Chase et al 1993), and mindful of uncertainties in the branching order of major lineages (Rice, Donoghue & Olmstead 1997), we map the distribution of compound leaves using data drawn from the literature and personal observations. We show that compound leaves evolved several times in different major lineages of angiosperms. We hypothesize that this evolutionary pattern of multiple origins should be reflected in underlying developmental patterns. One prediction of this hypothesis is that we should observe systematic differences in the expression patterns of genes important in shoot development (e.g., KNOX, a class of homeodomain containing proteins). We report results of expression studies that allow preliminary tests of this hypothesis.

Key words: Angiosperms, compound leaf, development, evolution