Annonaceae is a major tropical plant family particularly diverse in tropical rain forests of the world. Although the classification and systematics of the family has significantly improved over the past decade, the most recent classification was based on a reduced set of plastid markers and incomplete genus‐level taxon sampling. This classification recognised 4 subfamilies and 20 tribes. Yet, several important problems persisted, especially the phylogenetic placement of the African genus Meiocarpidium, resolution of intertribal relationships within subfamily Malmeoideae, resolution and classification within the diverse tribe Miliuseae with 23 genera, and the contrasting placement of the liana genus Artabotrys when using nuclear versus plastid data. Here, using a previously published Annonaceae‐specific nuclear bait kit, we generated for the first time a complete genus‐level (108 taxa) phylogenomic tree of the family based on 373 loci. We show that Meiocarpidium is sister to Ambavioideae and should be considered as a tribe and not a separate subfamily. Artabotrys is recovered as belonging to tribe Duguetieae, and not Xylopieae as previously inferred based on plastid data, and is sister to two other African liana genera, Letestudoxa and Pseudartabotrys. Finally, we were able to resolve intertribal relationships within subfamily Malmeoideae and most of the relationships within tribe Miliuseae. Nevertheless, we recovered strong gene conflict mainly at the backbone of the tribe, probably linked to a rapid diversification at its origin, leading to substantial incomplete lineage sorting. We suggest that this conflict will be hard to resolve. Using this novel phylogenomic framework we recognize 25 subtribes, 21 as new, to improve the infrafamilial classification of Annonaceae.
Two new species of Artabotrys are described from Thailand. Artabotrys tanaosriensis J. Chen, Chalermglin & R.M.K. Saunders, sp. nov., is similar to A. oblanceolatus Craib but differs in its symmetrical, cuneate or decurrent leaf base, externally distinct outer petal blades and claws, deltoid and undulate outer petal blades, rhomboid and undulate inner petal blades and shorter, subsessile and slightly beaked monocarps. Artabotrys spathulatus J. Chen, Chalermglin & R.M.K. Saunders, sp. nov., is most similar to A. tanaosriensis but differs in having flat outer petal blades, broadly rhomboid outer petal claws, broadly spathulate and strongly concave inner petal blades and strongly beaked monocarps. Two new records for the Flora of Thailand are furthermore reported here: A. punctulatus C.Y.Wu ex S.H.Yuan and A. byrsophyllus I.M.Turner & Utteridge, which were previously confused with A. aeneus Ast and A. grandifolius King, respectively. A key to Artabotrys species indigenous to Thailand is provided here.
Floral mimics deceive their pollinators by developing visual and olfactory resemblance to various models. We report a flower that exhibits phenotypes like aerial litter and deceives an aerial litter specialist beetle to achieve pollination. We assessed the floral phenology and the effective pollinators of an Australian understorey treelet, Meiogyne heteropetala (Annonaceae). The similarities of morphology, colour and odour between the flowers and co‐occurring aerial litter were investigated. The terpene synthase involved in floral scent emission was identified by expression patterns and product profile. The behavioural responses of the pollinator to various odours were assessed using bioassays. The erotylid beetle Loberus sharpi is the most likely effective pollinator because it was the only pollen‐laden visitor during the pistillate phase. Loberus sharpi was exclusively found in aerial litter and M. heteropetala flowers. The flowers offer an honest shelter reward. The beetle also oviposits there, but most larvae eventually perished as the petals dropped onto the forest floor. The morphology and spectral reflectance of the flowers overlap with aerial litter. The floral scent was dominated by monoterpenes, especially 1,8‐cineole. The cineole synthase MhCINS was the only highly expressed floral terpene synthase and possessed a highly similar product profile to the floral scent composition. The volatile composition of M. heteropetala flowers is distinct from other congeners and highly similar to aerial litter, indicating advergence to aerial litter. Visual and odour resemblance, coupled with low larval survivorship, provides evidence that the beetles were deceived into pollinating the flowers. Behavioural experiments showed that the pollinator was attracted to both aerial litter and M. heteropetala flowers. The beetles were also attracted to 1,8‐cineole and synthetic mixes of floral odour and MhCINS products. The beetles were unable to distinguish floral scent from MhCINS products nor from 1,8‐cineole, suggesting MhCINS alone sufficed to attract the pollinator olfactorily. The beetles, however, preferred aerial litter over flowers. The beetles likely categorised the flower as a general, but not the most preferred, brood substrate. Synthesis. This study reports the first case of floral mimicry of aerial litter and characterises the biochemical process responsible for olfactory mimicry.