Why does the flower stalk of Pulsatilla cernua (Ranunculaceae) bend during anthesis?

Flower stalks of Pulsatilla cernua, an early spring herb in north temperate Asia, changed position from erect to pendulous and back to erect during 6-10 d anthesis. We tested three possible explanations for this movement. Our results showed that (1) this movement is unlikely to be a mechanism to attract pollinators or enhance pollen output, because no pollinator preference was observed between erect and pendulous flowers and we found no buzz-pollination in this species; (2) hand self-pollination yielded higher seed set than open pollination in the field, but spontaneous selfing rarely occurred. Among open-pollinated flowers, seed set was depressed by emasculation, indicating that in the presence of insects, self-pollen provided reproductive assurance in this protogynous and self-compatible species. However, the change in flower orientation cannot be explained as reproductive assurance in that even self-pollination largely depended on pollinator visits rather than gravity. (3) A pollen germination experiment indicated that pollen damage by water is serious in this species. We deduced that the bending of the flower stalk during anthesis was to avoid rain damage to pollen grains in this species. During the 3-6 d period of pollen presentation, the petals elongated and were covered with unwettable hairs. Together with flower stalk movement, this was enough to protect the organs inside the flower from rain. This movement of the flower stalk seems to be important to maintain pollen viability in a rainy habitat with a scarcity of pollinators.