Skip to main content
Dryad logo

Evidence for hawkmoth pollination in the chiropterophilous African baobab (Adansonia digitata)

Citation

Karimi, Nisa (2021), Evidence for hawkmoth pollination in the chiropterophilous African baobab (Adansonia digitata), Dryad, Dataset, https://doi.org/10.5061/dryad.xwdbrv1d3

Abstract

Chiropterophily, or bat pollination, is typically considered a highly specialized pollination system that has evolved independently numerous times across the angiosperm phylogeny, with distinct lineages often converging on a similar suite of floral traits. The African baobab, Adansonia digitata, occurs widespread across continental Africa and introduced throughout much of the tropics, possesses floral traits classically associated with bat pollination, namely nocturnal anthesis, pendulous white flowers, and a “musky” fragrance. Despite this, our observations and pollination exclusion experiments in South African baobab populations suggested little if any role for bats, but instead showed that hawkmoths are the main pollinators. Hand pollination indicated strong self-incompatibility and crossing experiments suggest minimal diurnal receptivity. Furthermore, our analyses of floral volatiles revealed not only sulfur-containing compounds, commonly associated with bat pollination, but also a high concentration of the sesquiterpene β-caryophyllene, a compound generally more typical of hawkmoth pollination. Comparing previous pollination studies and published floral scent profiles from West Africa suggests that the classic bat pollination system in baobabs may be more labile than previously thought. This study provides an empirical example of a species that most likely evolved due to bat pollination yet has some degree of generality and possible geographic variation in floral traits and pollinator visitation patterns across its range.

Methods

This dataset includes the Gas chromatography–mass spectrometry (GC-MS) data files, standard curves for emsission calcuations, and pollinator observation and treatment data.