More than fruity scents: floral biology, scent and spectral reflectance of Annonaceae species
Liu, M.-F.; Chen, J.; Pang, C.-C.; Scharaschkin, T.; Saunders, R. M. K.
Show abstract
PremiseThe family Annonaceae possesses a broad array of floral phenotypes and pollination specialisations, and are important in the plant-pollinator interactions of tropical rainforests. Although there has been considerable effort to assess their interactions with pollinators, attempts to characterise their visual and olfactory communication channels are scarce. MethodsHere, we investigated the pollination biology of 12 Annonaceae species from five genera, viz. Meiogyne, Monoon, Polyalthia, Pseuduvaria, and Uvaria. Furthermore, their floral colour was characterised by reflectance spectroscopy and floral odour chemistry was assessed using gas chromatography-mass spectrometry. Floral scent was further compared across the whole family using non-metric dimensional scaling plots to identify specialisation in floral odour. ResultsThe Meiogyne species are likely pollinated by small beetles; the Polyalthia and Pseuduvaria species are likely pollinated by beetles and flies; and the Uvaria species is likely pollinated by beetles and bees. Flowers of most species are UV non-reflective, and have various spectral reflectance profile across the remaining visible spectra. Multiple species produce floral odour resembling ripe fruits. The flowers of Meiogyne species and Polyalthia xanthocarpa emitted mostly branched-chain esters, while flowers of Uvaria released mainly straight-chain esters. The Pseuduvaria species instead emitted scent reminiscent of rotten fruits, largely consisting of 2,3-butanediol. The inner petal corrugation in Meiogyne functions as a food reward, and the inner petal growth serves as a nectary gland for Pseuduvaria. ConclusionsOur study identifies the visual and olfactory cues of multiple Annonaceae species and provides insights into how Annonaceae flowers attract different guilds of pollinators.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.