The evolution of flower colour variation in the endangered grassland daisy Gerbera aurantiaca (Asteraceae)

Abstract

Flower colour variation is widespread within and between populations and has long been a focus for research into natural selection. I investigated the potential factors influencing the capitulum colour variation exhibited by Gerbera aurantiaca (Asteraceae), an endangered grassland daisy endemic to the sub escarpment mistbelt grasslands of eastern South African. Widely scattered and fragmented flower colour polymorphic and monomorphic populations comprise long-lived clones with inflorescences ranging from yellow through orange to deep red. An earlier study investigated some aspects of pollination biology and population structure in the southern monomorphic, red-flowered G. aurantiaca populations, but little was previously known about the yellow and flower colour polymorphic communities in the central and northern parts of the distribution. Geographical variation in flower colour of a plant species may reflect the outcome of selection by pollinators or may reflect abiotic factors such as soil chemistry or neutral processes such as genetic drift. I found no strong associations between morph colour and abiotic factors, including soil chemistry, and plants in a common garden retained the capitulum colour of the source population, even when grown from seed, suggesting that flower colour variation is not a result of phenotypic plasticity. Distribution of the various colour morphs in a flower colour polymorphic population further supported this lack of association. Hopliine beetles were the most abundant flower visitors in all populations and in a colour polymorphic population, all colour forms were equally attractive to these beetles and did not differ significantly in terms of fruit set. Native honeybees visited only yellow capitula in a monomorphic population. I found that G. aurantiaca is largely self-incompatible, completely dependent on pollinators for pollen transfer, and pollen quantity limited in some of the red coloured populations. This species is therefore highly susceptible to pollinator colour sensitivities for visits and fruit set. Hopliine beetle abundance appeared to be the most important factor in fecundity in all populations, but visited all colour forms equally, suggesting that it is maintaining the current colour pattern rather than driving change. The absence of this pollinator in two red-flowered populations has resulted in critically low fecundity. Pre-dispersal tephritid fly seed predators damaged similar numbers of capitula in all study populations making it unlikely that these biotic agents are driving colour variation. Since community flower has been shown to influence the suite of pollinators and subsequently predominant flower colour, I compared the petal reflectance spectra of all coflowering species across five representative G. aurantiaca populations but found no differences. Insect vision models showed that the yellow morphs were visible to honeybees, but the red forms were not different to the background, while all colour forms, in particular red, were conspicuous to the beetles. These results confirmed field pollinator observations. Pigment distribution and petal structure may be important in pollinator attraction. In G. aurantiaca pigments are concentrated in the upper epidermal cells of the ray florets which have an epidermal cuticular surface micro structuring suggesting that pollinator attraction is important in this species. Flower colour inheritance may play a vital role in self-incompatible species in flower colour polymorphic populations. Here I crossed true breeding red and yellow colour forms from different populations which produced primarily yellow and orange F1 offspring, suggesting a polygenic inheritance of anthocyanins in G. aurantiaca. The results presented here show that capitulum colour in G. aurantiaca is geographically structured, but the ultimate evolutionary basis of this colour variation remains elusive, and further research using genetic techniques should be pursued to answer this question. Some recommendations for the long-term conservation of this flagship endangered species are detailed.