Evolution has explored and enormous range of body plans, yet species numbers between different taxa vary enormously. For example, the vast majority of land plants are eudicots or monocots whereas the remaining taxa (e.g. early diverging angiosperms, extant gymnosperms, mosses, ferns) make up only a minority of land plant species. Ecological opportunities and evolutionary innovations have been proposed to be among the main drivers for species richness. However, we suggest that developmental robustness is an additional and currently underestimated factor that contributes to species diversity. For example, perianth organs (petals, sepals or tepals) are considered a key innovation of angiosperms that may have contributed to their radiation as they facilitated interaction with insects and other animal pollinators. However, perianth organs evolved at the base of the angiosperms and are already present in the relatively species-poor early diverging angiosperms (water lilies, star anise etc.). Intriguingly, one of the main differences between early diverging angiosperms and monocots or eudicots is that the former have a quite variable number of perianth organs whereas the perianth is mostly pentamerous and trimerous in eudicos and monocots, respectively. We speculate that is was eventually the fixation (i.e. increase of robustness) of perianth organ numbers – that spurred species diversification. In general, the robust implementation of an evolutionary innovation into a pre-existing developmental program might be as important for fostering species diversity as the innovation itself.
We are currently analyzing the robustness of flower development using Arabidopsis thaliana as a model system. For that purpose we employ mutation accumulation lines as well as the analysis of plants carrying mutations in robustness candidate genes (SEPALLATA genes and HSP90 genes are currently our prime candidates). One aim of those analyses is to understand whether robustness of floral traits was selected for during evolution and how the robustness of different traits (i.e. petal number vs. stamen number, petal number vs. petal length) compares to each other.