Pollination syndrome accurately predicts pollination by tangle-veined flies (Nemestrinidae: Prosoeca s.s.) across multiple plant families

Abstract

The idea that a syndrome of floral traits predicts pollination by a particular functional group of pollinators remains simultaneously controversial and widely used because it allows plants to be rapidly assigned to pollinators. To test the idea requires demonstrating that there is an association between floral traits and pollinator type. I conducted such a test in the Cape Floristic Region of South Africa, by studying the pollination of eight plant species from six families that flower in spring and have scentless, actinomorphic, upwards-facing flowers, with orbicular petals all held in the same plane. The petals are brilliant-white with red-purple nectar guides. The tubes are short and hold small volumes of concentrated nectar, except in the rewardless Disa fasciata (Orchidaceae). Pollinators were photographed and captured, pollen loads were analysed and pollination networks were constructed. Consistent with the pollination syndrome hypothesis, the species with the defined syndrome shared a small group of pollinators. The most frequent pollinators belonged to a clade of four tangle-veined fly species with relatively short proboscises (Nemestrinidae: Prosoeca s.s.), while functionally similar Bombyliidae and Tabanidae played minor roles. Among the four Prosoeca species, only Prosoeca westermanni has been described, a result that highlights our ignorance about pollinators. The demonstration of an association between the syndrome of traits and pollination by this group of flies explains the repeated evolution of the syndrome across multiple plant families, and allows prediction of pollinators in additional species. More generally, the result validates the idea that the traits of organisms determine their ecology.

Keywords: Adenandra; Fynbos vegetation; Stuckenbergina africana; coevolution; convergent evolution; floral deception; long-proboscid flies.

Fig. 1

Spectral reflectance for confirmed members of the Prosoeca westermanni pollination guild. [Colour figure can be viewed at wileyonlinelibrary.com]

Fig. 2

Bipartite networks of interaction between the plants that conform to the Prosoeca westermanni pollination syndrome and their pollinators. (a) Network of 354 observed visits, with line thickness varying from 1 to 110 visits. (b) Pollen transport network, where line thickness is the average number of pollen grains carried per sampled insect. The thinnest lines represent 0.125 pollen grains, and the thickest is 666 grains. The number of individuals sampled is in brackets. The link from Disa fasciata is shown at 1% of its actual width because its pollen is aggregated into pollinaria that contain thousands of pollen grains. [Colour figure can be viewed at wileyonlinelibrary.com]

Fig. 3

Interactions between members of the Prosoeca s.s. clade (Nemestrinidae) and their nectar plants: (a) Prosoeca westermanni pollinates Geissorhiza ovata (Iridaceae) in burnt vegetation at Houwhoek. The orange pollen of G. ovata is on the frons while ivory pollen of Gladiolus debilis is on the thorax. (b) A female Pr. westermanni pollinates Pelargonium longicaule at Eerste Waterval, Jonkershoek; the stigma contacts the ventral abdomen. (c) A female Prosoeca “dimorpha” pollinates Adenandra villosa at Steenberg Peak, Cape Peninsula. (d) A male Prosoeca “alticola” pollinates Erica fastigiata at Pic‐sans‐Nom, Jonkershoek. The sizes of the flies are in Table 2. [Colour figure can be viewed at wileyonlinelibrary.com]

Fig. 4

Pollination of the rewardless Disa fasciata (Orchidaceae). (a) A female horsefly Stuckenbergina africana (Tabanidae), with a pollinarium and several viscidia of D. fasciata attached to its proboscis. Orange pollen grains of Geissorhiza ovata adhere to the frons and eyes. Scale bar = 2 mm. (b) D. fasciata is unusual in Disa in having a lip that resembles the sepals, and in this way actinomorphy is achieved. Scale bar = 5 mm. (c) Pollen massulae on the stigma flanked by the lateral petals (Voorberg, Kogelberg). Scale bar = 1 mm. [Colour figure can be viewed at wileyonlinelibrary.com]

Fig. 5

The network of interactions between plants that conform to the Prosoeca s.s. pollination syndrome and their most important pollinators. Connecting lines represent the number of sites at which an interaction was observed. The thinnest lines represent interaction observed at one site; the thickest, five sites. Clockwise, the flies are: Stuckenbergina africana, Prosoeca “dimorpha”, Prosoeca “alticola”, Prosoeca westermanni, Prosoeca “brevirostris”. The flowers: Disa fasciata (Orchidaceae), Aristea spiralis (Iridaceae), Adenandra villosa (Rutaceae), Geissorhiza ovata (Iridaceae), Pelargonium longicaule (Geraniaceae), Erica fastigiata (Ericaceae), Gladiolus debilis (Iridaceae). Pollinators for which 20 or fewer visits were recorded are excluded from the figure. When the image measures 29.7 cm across, the flowers are life size and the flies are 1.5 × life size. [Colour figure can be viewed at wileyonlinelibrary.com]

Fig. 6

Pollen distribution on the bodies of sampled flies. Standard deviations are not displayed because they are often larger than the means, as can be seen from the raw data (Appendix S5). The scale is logarithmic and values <1 are not shown. [Colour figure can be viewed at wileyonlinelibrary.com]