Super indelible flower colours

How do you fire a pollinator?

That was the question in last week’s Ecology, Evolution and Behaviour departmental seminar. The speaker was James Thomson, an evolutionary ecologist from the University of Toronto who specializes in the interactions between plants and their animal pollinators. His research shows that nectar-addled hummingbirds are like corporate ladder climbers. Bees, on the other hand, are always getting canned.

Pollination syndromes have been a major focus of Thomson’s work1. These are not garden ailments. “Syndrome” here refers to a suite of traits that tend to be found together, in this case because they help a plant attract a certain kind of pollinator.

Bird-pollinated flowers tend to be red and tube-shaped, producing lots of nectar but relatively little scent. Birds have sharp vision, and do not depend much on their sense of smell. Honeysuckle is an example of this type of flower – or anything that looks like a hummingbird feeder. Bee-pollinated flowers come in shades of yellow, blue, and purple, because bees cannot see the colour red. Familiar examples are sunflowers, snapdragons and wild pansies. These often have petals modified into special bee landing platforms. Flowers that specialize on birds and bees are common, but there are many other pollination syndromes. If a flower is orange-brown and smells like rot, it probably depends on carrion flies. Mammal-pollinated flowers often smell fruity, like synthetic grape flavouring.

In his talk, James Thomson reviewed a decade’s worth of work on beardtongue flowers from the genus Penstemon2. In 2007, Thomson and his collaborators used genetic analyses to build the evolutionary tree for close to 200 of the species in this group3. When flower traits were mapped on to the Penstemon family tree, interesting patterns were revealed.

First, the bird and bee pollinated species were distributed broadly throughout, implying frequent transitions between these two syndromes in the history of the Penstemon group. Like an evolutionary magnet, pollination by one type of animal or another exerts a strong pull on multiple flower traits in concert. Evolving species are drawn rapidly towards a new form, so you almost never find intermediates.

This lability or changeable nature of floral traits was not much of a surprise, but the Penstemon tree also suggested something incredible. Floral evolution was directional.

The transition from bird to bee pollination happened frequently in the Penstemon history, but the flowers apparently never evolve in the opposite direction3. Thomsom suggests that this might not be limited to Penstemon: the bee-bird directionality might also apply to floral evolution in general2. In other words, once a plant gets close to the threshold, it gets sucked in to the bird pollination syndrome and never looks back. Borrowing from Michael Gilpin’s analogy for extinction, Thomson compared this process to an evolutionary vortex. He thinks that a complex set of ecological and genetic factors, interacting with evolutionary history, drive plant species rapidly towards one type of pollination over another.

But the question remains: How? What might ease the transition between these different floral lifestyles? And why can’t bees hold down a job?

The answer might depend on where you look. Morning glories provide a fascinating example of how the genetics of flower colour contribute to this directionality. In morning glories, the red flower colour that hummingbirds prefer is permanent, evolutionarily speaking.

Most of the 1000 or so species in the morning glory family Convolvulaceae are typical bee-lovers, with blue or purple flowers. In the genus Ipomoea, however, a number of red, hummingbird-pollinated flowers evolved from a single blue ancestor4.

The red and blue pigments in these plants are derived from slightly different anthocyanidin precursors: the blue precursor has an extra hydroxyl group added on to it. Genetic mutations that disrupt any of the steps involved in adding this hydroxyl group cause all of the raw materials to funnel into the red pigment pathway, instead. This is how you make a red morning glory.

Back in 2004, Rebecca Zufall and Mark Rausher used the red cypressvine morning glory (Ipomoea quamoclit) to look at this evolutionary transition in greater detail5. In an elegant series of experiments, they used techniques from molecular biology to show that morning glory red is like a super indelible marker5. First, they took gene sequences from the red cypressvine morning glories and their blue ancestors, and compared how well they functioned in lab strains of genetically engineered Arabidopsis plants that lacked the pigment genes. They found that the red morning glories have multiple loss-of-function mutations along the blue pigment pathway. Second, Zufall and Rausher compared normal gene expression levels in the red and blue species. Their results demonstrated that the red cypressvine morning glories also have mutations in the regulatory genes for the blue pigment pathway.

Since the cypressvine transition to red, the genetic information required to produce blue pigment has decayed. Zufall and Rausher’s work shows how extensive these mutational losses have been: the chance of a red morning glory regaining all of the lost steps required to produce blue flowers is near zero. Zufall and Rausher also showed that the evolutionary tree of the Ipomoea morning glories is consistent with their permanent colour hypothesis5. Blue colour was lost once, and and has never been regained.

At the time, this was touted as concrete evidence of Dollo’s Law: evolution does not go backwards. More on this to come in a future post.

Permanent colour change cannot explain directional evolution towards bird pollination in all plant lineages. In monkey-flowers, the transition between bee and bird syndromes seems to involve different carotenoid pigments6. However, anthocyanin shifts similar to those seen in the morning glories might also be at play in James Thomson’s Penstemon flowers7.

Thomson thinks there is much more to the Penstemon story. In his view, ecology has a major role in directional floral evolution. He outlined his “pollen transfer-efficiency” hypothesis: hummingbirds are better at getting the job done than bees, even for plants that have not yet evolved the bird-loving traits8.

Bees get fired because they pocket the merchandise. When they are hard at work traveling from flower to flower, they store it in special “pollen basket” pouches on their legs. They even groom pollen that sticks to their bodies into these pouches, where it gets mixed with nectar; these stores are brought back to the hive as bee food. This is a loss to the plant.

Birds, on the other hand, are nectar addicts. Unlike bees, hummingbirds do not eat pollen and they do not groom in flight. Any pollen that sticks to their feathers is readily transferred to the next flower they visit. Indeed, experimental studies by researchers in Thomson’s lab demonstrate that hummingbirds are more efficient at pollinating Penstemon species of both the bird and bee-loving type8.

Thomson’s idea is that in the evolutionary history of this group, as soon as hummingbirds appear on the scene in high enough numbers, bees go from being helpers to functional parasites2. Bees will always consume some pollen, so when birds are around the net impact of bees on Penstemon reproduction is negative.

Thomson concluded his talk by saying that his role in this story has ended, since he is no longer working on Penstemon. An audience member asked whether anyone was doing the ultimate experiment, to see if removing one of the pollinators in a species that normally experiences both would cause the plants to evolve.

Thomson’s answer was no: for one thing, it is too hard to work with birds in a long-term project like this. Years ago, he had proposed a similar pollinator evolution experiment, but nobody seemed to go for the 35-year PhD program it would have entailed. Thomson added that he had heard rumours that professors are allowed to do research, too, and even department heads – but the latter turned out not to be true in his experience.

In an ideal world, graduate degrees would be measured in the generation time of your study organism. For now, though, there are plenty of questions about the factors involved in pollination syndromes, and whether they are common, or unique, across the plant kingdom.


  1. Fenster, C. B. et al. 2004. Annual Review of Ecology, Evolution, and Systematics 35: 372-403.
  2. Thomson, J. D. and Wilson, P. 2008. International Journal of Plant Sciences 169: 23-28.
  3. Wilson, P. et al. 2007. New Phytologist 176: 883-890.
  4. Miller, R. E. et al. 2004. American Journal of Botany 91: 1208-1218.
  5. Zufall, R. and Rausher, M. D. 2004. Nature 428: 847-850.
  6. Bradshaw, H. D. and Schemske, D. W. 2003. Nature 426: 176-178.
  7. Rausher, M. D. 2008. International Journal of Plant Sciences 169: 7-21.
  8. Castellanos, M. C. et al. 2003. Evolution 57: 2742-2752.