Do pollinators drive the evolution of flowers?

Angiosperms, the flowering plants, are the astonishingly diverse. But what drives the selection pressures to create this diversity? One explanation is the Grant-Stebbins model. This model looks at pollination as a selection process. Many flowers need pollinators. If there are no pollinators, the plants don’t get pollinated and they have no offspring. It means that plants have to live where the ranges of their pollinators are. Plants that evolve to pull in more pollinators will have more success. Flowers are therefore constantly evolving to attract visitors.

The diversity of flower colours has been occupying some botanists thoughts. Image: Spiderwort/Flickr

It’s an interesting idea, but how do you test it?

Newman, Anderson and Johnson investigated the South African plant Disa ferruginea for their paper ‘Flower colour adaptation in a mimetic orchid‘. This is a clever piece of work based on a mimetic orchid. Disa ferruginea doesn’t waste time producing a reward for visiting insects, but this means there’s no obvious reason for a pollinator to want to visit. Instead D. ferruginea looks like the flowers of other nectar producing plants. What makes D. ferruginea odd is that it doesn’t always look like the same plant. In the west it has red flowers. In the east it’s orange.

Newman et al. thought this might be that pollinators were selecting the plants that looked most like the nectar bearing species, so they conducted an experiment. They swapped some orchids, so some orange orchids were found in the west, and some red orchids were moved to the east. Then they watched to see what happened.

The only insect that pollinates D. ferruginea is a butterfly Aeropetes tulbaghia. Sure enough they found that in the west they tended to ignore the orange plants and stick with the red orchids. In the east they skipped over the red orchids but stuck with the orange flowers. In both sites it was the orchids that looked most like the reward-bearing plants that attracted pollinators, so here was proof that pollinator selection was happening.

I suppose the follow-up experiment would be to relocate butterflies from one range to another, but I’ve no idea how you’d manage to fit them with radio transmitters to track them.

This isn’t the only test of Grant-Stebbins.

Ellis and Johnson (the same Johnson mentioned above) have published a paper ‘The Evolution of Floral Variation Without Pollinator Shifts in Gorteria Diffusa (Asteraceae)‘. Gorteria Diffusa is a South African daisy that’s usually pollinated by the bee fly Megapalpus capensis. Ellis and Johnson found fourteen different varieties of G. Diffusa. Each of the forms seems to be inherited, not a plastic adaptation to the environment, but they couldn’t find fourteen different pollination scenarios. If pollination isn’t the selective force, what is going on? Ellis and Johnson point to other work that argues that there’s more to plant survival than pollination. For example once you have the seeds you have to ensure that at least some of them get past predators to germinate. Also, while you want a flower to look tasty to a visiting pollinator, do you really want to attract larger herbivores that will simply eat the flower?

It’s clearly a problem that needs more research, and there’s a whole special issue on pollinator-driven speciation on the way from Annals of Botany. The earliest papers are available as advanced access to subscribers including ‘Do pollinator distributions underlie the evolution of pollination ecotypes in the Cape shrub Erica plukenetii?‘ by Van der Niet, Pirie, Shuttleworh, Johnson and Midgley. Yes, the same Johnson.

Malachite sunbird visiting subsp. plukenetii at Rooiberg. Image: Van der Niet et al.
Malachite sunbird visiting subsp. plukenetii at Rooiberg. Image: Van der Niet et al.

Erica plukenetii is a shrub that grows waist-high, if your waist is typically 90cm above the ground. You can find it on the slopes of South African mountains, but you’ll find it with more than one pollinator. The typical E. plukenetii has a corolla (set of petals) of medium length. These are pollinated by the Orange-breasted sunbird (Anthobaphes violacea). A sunbird is a big like a hummingbird in that it’s a small nectar-feeding bird. A difference is that sunbirds tend to perch to feed instead of hovering.

The Orange-breasted sunbird is not the only pollinator of E. plukenetii. In the north of their range the plants are pollinated by Malachite sunbirds (Nectarinia famosa). Malachite sunbirds have a longer bill and the E. plukenetii in this region have longer corollas. In the centre of their range there’s also a shorter corolla variety of E. plukenetii and this is not bird pollinated. Instead it’s pollinated by a moth. These means there are three ways to pollinate E. plukenetii and they seem to have developed from the middle form. How does this fit with the Grant-Stebbins model?

Van der Niet et al. argue that the longer corolla plant fits the Grant-Stebbins model very well. As you go north Orange-breasted sunbirds become rarer and Malachite sunbirds more common, so the flowers better suited to the Malachite sunbirds will produce more offspring. Here pollinator selection makes sense. What about the short-corolla plants?

These are in the middle of the range, but here there are already plenty of Orange-breasted sunbirds, so they didn’t need to change to attract them. In fact, they’ve evolved to move away from the pollinators. This contradicts the Grant-Stebbins model. Van der Niet et al. suggest that other pressures must have an influence, and compare the bird pollinated E. plukenetii, found on mountainsides with the moth-pollinated E. plukenetii, found on flatter land. The moth-pollinated plants have slender branches that don’t support birds well. If this is necessary to colonise the flats, then it’s the habitat that changes the pollinator for the plant, instead of the pollinator defining the plant’s habitat.

It seems likely that it’s not just pollinators that select plants, but that plants can also select pollinators. One example is the paper ‘Domestication of cardamom (Elettaria cardamomum) in Western Ghats, India: divergence in productive traits and a shift in major pollinators‘ by Kuriakose, Sinu and Shivanna. Domesticating crops brings about a lot of changes. In the case of cardamom one change is there are flowers around for a lot longer than in the wild. For wild cardamom the pollinators tend to be solitary bees. For domesticated cardamom the flowers attracted social bees, the Purple sunbird and the Little Spiderhunter, a bird which – despite its name – is fond of nectar. The change in floral display seems to attract an entirely different kind of pollinator. Wild and cultivated cardamom don’t seem to share pollinators, despite being compatible with each other.

What seems to be happening is that where selection can occur, it will occur. While there are many scenarios where that can happen, like herbivory or habitat, competition for pollinators is in some cases a major factor in driving the evolution of plants.

The Annals of Botany special issue on pollinator-driven speciation is due out early 2014, with subscribers getting early access to some papers now. The issue will become free-access in 2015.


Ellis A.G. & Johnson S.D. (2009). The evolution of floral variation without pollinator shifts in Gorteria diffusa (Asteraceae), American Journal of Botany, 96 (4) 793-801. DOI: (free access)

Kuriakose G., Sinu P.A. & Shivanna K.R. (2008). Domestication of cardamom (Elettaria cardamomum) in Western Ghats, India: divergence in productive traits and a shift in major pollinators, Annals of Botany, 103 (5) 727-733. DOI:

Newman E., Anderson B. & Johnson S.D. (2012). Flower colour adaptation in a mimetic orchid, Proceedings of the Royal Society B: Biological Sciences, 279 (1737) 2309-2313. DOI: (free access)

Van der Niet T., Pirie M.D., Shuttleworth A., Johnson S.D. & Midgley J.J. Do pollinator distributions underlie the evolution of pollination ecotypes in the Cape shrub Erica plukenetii?, Annals of Botany, DOI: (subscription access till 2015)


Pansies. Image by Spiderwort/Flickr. This image licensed under a Creative Commons by-nd licence.