Large‐flowered stapeliad Edithcolea grandis (Apocynaceae: Asclepiadoideae) is pollinated by dung‐ and blood‐feeding African face flies (Muscidae: Musca xanthomelas

Abstract

The more than 360 species of stem‐succulent stapeliads in the tribe Ceropegieae represent a perfect example for species radiation and diversification in response to pollination by flies. In the majority of species, the myiophilous flowers have visual and olfactory characteristics that suggest carrion or dung mimicry. However, pollination studies in stapeliads are scarce which hampers our understanding of floral mimicry strategies and pollinator specialization. This study investigated pollination and floral scent in the large‐flowered stapeliad Edithcolea grandis in its natural savanna habitat in Tanzania (East Africa). The foul‐scented flowers were visited and pollinated exclusively by females of the blood‐feeding fly species Musca xanthomelas . This is the first time that this species is recorded as a pollinator. Pollen transfer efficiency was high with 61.5% of removed pollinia being deposited in the guide‐rails of the gynostegium. The chemical analyses of the strong foetid floral scent revealed a rather unusual combination of volatiles, mainly p ‐cresyl isobutyrate, p ‐cresyl butyrate, δ ‐cadinene, α ‐copaene and pentadecane, together with various minor compounds. The main floral scent compound, p ‐cresyl isobutyrate, is documented for the first time as floral volatile. Overall, the floral scent profile of E. grandis is different from other stapeliads so far analysed, and the volatile compounds do not fit a typical dung mimicry strategy. The salient flower colouration suggests that visual signals, besides floral scent, play an important role in the attraction of pollinating flies. In particular, the central depression of the corolla tube not only seems to serve as a morphological filter limiting access to the gynostegium, it also appears to be visually similar to myiatic lesions caused by parasitic flies with subcutaneous larval development. This implies that E. grandis flowers might mimic wounds to specifically attract blood‐feeding female M. xanthomelas which, after landing on the corolla, start probing the “fake wound” in the flower centre with their proboscis and thereby make contact with the gynostegium for successful pollination. It is therefore assumed that E. grandis flowers mimic a food source for adult M . xanthomelas flies rather than an oviposition site, although there may be overlap between both of these strategies as a single fly egg was found deposited at a gynostegium.