Can the Grant-Stebbins model explain where all the flowers come from?

Rainbow rose
Rainbow rose by Pamela Carls/Flickr

Valentine’s Day is coming, and many people are looking for the right flower to express their feelings, though it’s hard to beat a multi-coloured rose. But where do all the different sorts of flowers come from?

The Grant-Stebbins model suggests that pollinators drive speciation. Shifts in pollinators cause angiosperms (flowering plants) to adapt and form ‘ecotypes’ which then become new species. To someone like me, who is not a botanist, this is an appealing explanation. In reality it’s more complicated than that.

Pollinator-Driven Speciation is the subject of a recent special issue for Annals of Botany. I’ll be blogging about papers from the issue during the week. In their introductory paper for the issue, Van der Niet, Peakall and Johnson point out there’s plenty of evidence of floral adaptation to pollinators, and evidence at the large-scale of angiosperms diversifying with pollinators, but there’s a gap in the middle. How do you get from floral adaptation to speciation?

Van der Niet et al. identify four key factors in their paper. First you need to show that pollinators are selecting which plants get fertilised and which don’t. Next you need to show that this selection has consequences for floral traits. After this you’d need to look at the geographical context. What is it that causes you to find a certain plant here but not there? Finally, you’ll want to show that the pollinators are helping isolate populations so the differences between plant populations don’t spread back through the parent population.

Breaking down the problem really helps, because it means we can move from a general idea to some testable hypotheses. This is what the viewpoint paper does and each step is peppered with citations showing how each one of them can be tested with evidence. In the case of pollinator selection it’s possible to do direct experiments.

One of the most important factors in experimenting is that you also have to accept you may be wrong about something. In the conclusions Van der Niet et al. say the Grant-Stebbins model does a good job of explaining speciation, but there are some non-pollinator factors involved too. Strangely I think this is actually an excellent result for the model.

A model that explains everything risks being a just-so story. The fact it breaks down in places shows that scientists aren’t simply recording what they expect to see. However, the model breaking down doesn’t make it useless. The fact that it usually works means that when it does fail, it is pointing out that something really interesting and unexpected is going on. This is one of the most useful things a scientific model can do, take a lot of different observations and help you sort out what results are excitingly weird.

You can pick up the viewpoint paper that introduces the pollinator-driven speciation issue from the Annals of Botany.


Rainbow rose by Pamela carls/Flickr. This image licensed under a Creative Commons by licence.