The evolution of wind pollination from animal pollination has been one of the most enduring problems in the study of flowering plants. Unlike other common transitions, there has been very little research on the evolution of wind pollination. This is surprising because ~10% of angiosperm species rely on wind to transport pollen and this pollination system has originated at least 65 times in diverse lineages. The evolution of wind pollination is puzzling because it is viewed as a wasteful process at the mercy of unpredictable conditions, especially compared to animal pollination. Why should such an apparently wasteful process evolve?
To understand the selective mechanisms underlying the evolution of wind pollination, we used a comparative phylogenetic analysis to investigate the correlated evolution of reproductive traits and the order in which they arise in lineages. This allowed us to provide mechanistic explanations about why traits become associated. Next, we addressed the assumption that wind pollination is more inefficient than animal pollination by measuring pollen capture in 19 wind-pollinated species. We demonstrated that pollen-transfer efficiencies were not substantially lower than in animal-pollinated taxa challenging the prevailing view. This result challenges the common assumption that wind-pollinated species typically have uniovulate flowers because they capture few pollen grains. To explore the adaptive value of uniovulate flowers we developed a theoretical model that demonstrated that when flowers are inexpensive, as in wind-pollinated species, it is beneficial to distribute ovules among numerous flowers because pollen capture per flower saturates and plants that make many flowers sample more of the airstream maximizing fitness.
I also use experimental systems to examine the adaptive value of traits in wind-pollinated species. In a study examining floral strategies thought to promote outcrossing we demonstrated that on the contrary neither unisexual flowers nor delayed male flowering served this role in 7 species of sympatric sedges (Carex). Based on this result we revised the prevailing view of the maladaptive nature of between flower self-pollination and proposed instead that this form of selfing can be adaptive. This idea provides a new insight on the functional differences between animal and wind pollination. Next we addressed the problem of why sex ratio biases occur in some wind-pollinated plants. By investigating populations of the herb Rumex nivalis using sex–specific genetic markers we showed that the strongly female-biased sex ratios that characterize populations result from demographic aspects of their mating environment. This was the first demonstration of environmentally determined sex ratios in plants. Lastly, we used the colonizing plant, ragweed (Ambrosia artemisiifolia), to test theoretical models of sex allocation. We found that sex allocation is plastic and occurs in the direction predicted by theory based on changes in plant size and environment. These case studies provided novel insights and I believe they will help to change prevailing paradigms about reproduction in wind-pollinated plants.