Nothing in Biology Makes Sense: Making sense of pollination syndromes

2010.07.15 - Eastern Tiger Swallowtail Pollinator at work. Photo by jby.

Over at Nothing in Biology Makes Sense! I’m discussing pollination syndromes—suites of traits held in common by plants that use similar pollinators.

  • Bee-pollinated flowers are usually blue or yellow, often with contrasting “guides” that point towards nectar rewards, and they usually have some sort of scent.
  • Bird-pollinated flowers tend to be red and tubular, and often open downwards. They produce lots of relatively weak nectar, and generally don’t have very strong scents …
  • Moth-pollinated flowers are usually white, opening in the evenings, and strongly scented.

To find out how evolution makes sense of these handy rules of natural historical thumb, go read the whole thing, and check out the new meta-analysis of pollination syndromes that I discuss.◼


Nothing in Biology Makes Sense: A flower is an evolutionary compromise

Pollination IMG_4730D Welcome, pollinators! But, um, everyone else can just stay out, okay? Photo by Yeoh Thean Kheng.

Over at Nothing in Biology Makes Sense!, I’ve written about a neat study of the tropical vine Dalechampia scandens, which has to solve an evolutionary puzzle that confronts most flowering plants:

But there’s a downside to making a big, showy display to attract pollinators—you might also attract visitors who have less helpful intentions than gathering up some pollen and moving on to the next flower. Showy flowers might attract animals that steal the rewards offered to pollinators—or they might attract animals that eat the flowers themselves, or the developing seeds created by pollination.

To see how a team of biologists directly measured this evolutionary compromise (spoiler: it involves counting pollen grains with a hand lens) go read the whole thing.◼


Nothing in Biology Makes Sense: Timing is everything

A euglossine bee gathers scent compounds inside an orchid. Photo by Alex Popovkin, Russian in Brazil.

This week at Nothing in Biology Makes Sense, the big science post comes from … me. It’s about a big new study of orchids and the perfume-collecting euglossine bees that pollinate them.

The study by a team out of Harvard—lead-authored by Santiago R. Ramírez—tests three predictions arising from the proposition that bees and orchids are equally dependent on the scent-collection mutualism. First, as I noted above, a mutually-dependent relationship should mean that bee and orchid species often form in tandem, and that the euglossine bees and the orchids have spent most of their histories together. Second, the euglossines should rely mainly on scents from orchids, not from other sources. Finally, euglossines and orchids should show similar degrees of dependency. An orchid that relies on only one bee species should use a bee species that only collects scent from that one orchid; bees that collect scent from multiple orchids should use orchids that are, themselves, involved with multiple bee species.

To find out whether or not these predictions are borne out, go read the whole post. ◼


Flowers stay open for pollinators, not daylight

A honeybee explores the depths of a dandelion, one of the species used in Fründ et al.‘s experiments. Photo by je-sa.

ResearchBlogging.orgIf you’ve ever stopped to admire morning glory flowers opening first thing in the morning, then noticed they’ve closed by evening, you’re at least dimly aware of one of the longest-established ideas in plant biology: that flowers open and close on a reliable daily schedule. Different species are open at different times of day, of course, but each flowering plant has its preferred open period, and it sticks to that schedule during its flowering season.

This idea led Carolus Linneaus, the father of modern biological taxonomy, to propose an Horologium florae, or “floral clock” using plantings of species with known flowering times to mark the hours. You can find his table of proposed species in the online version of Linneaus’ 1783 treatise Philosophia Botanica, if you’re not averse to Latin. Studies of flowers’ daily schedules go back to well before English was the language of international science, and continue to the present day [$a].

Yet no one seems to have spent much time considering how flowers’ schedules might respond to the activity of their very reason for being: pollinators. Flowers don’t open just to be open in a particular kind of sunlight—they’re open to attract animals that can carry pollen to another plant, and maybe leave some, too. If a flower receives enough pollen to make seeds by noon, why would it stay open until two o’clock?

According to some new experimental results, the answer to that question is that they don’t [$a].

Jochen Fründ, Carsten F. Dormann, and Teja Tscharntke set out to see whether a selection of European wildflowers adjusted their opening schedules in response to pollination, with two major experiments and a broader-scale observation project. The experiments address whether pollinator activity could change flowers’ schedules; the observations help determine how important those changes might be in studies of plant-pollinator interaction.

A floral clock in Geneva—not quite what Linneaus had in mind. Photo by aranmanoth.

In the first experiment, the team planted wildflowers—Crespis capillaris, a close relative of common dandelions—in experimental plots spaced across a field. Plots were either caged or left open to insect visitors, and Fründ et al introduced bees into some of the caged plots. So some plots had a controlled set of pollinators, some had none at all, and some had whatever pollinators were already active in the field.

The team then watched the flowers’ daily opening and closing in the experimental plots. (They had a lot of help—a long list of names in the paper’s Acknowledgements section ends with “and many others.”) Over the same period of time, flowers in the un-caged plots received more insect visitors than flowers in either other treatment, and had mostly closed by midafternoon; flowers in the caged plots with bees introduced received fewer visitors and closed hours later; and flowers in the plots with no pollinators at all stayed open till evening.

So flowers experiencing the same daylight pattern closed earlier if they received more pollinator visits. The team followed up this result by hand-pollinating flowers of C. capillaris and a handful of closely related species growing in the same field, including dandelions—and flowers of three out of four species closed more rapidly when hand pollinated. Dandelions didn’t respond to hand pollination, a result the authors explain by noting that dandelions often self-pollinate, and so don’t need to wait for animal pollinators.

Finally, the team compiled observations of plant-pollinator interactions from sites similar to their study field located across Germany, and divided them into observations taken before solar noon, when the focal flower species from the experiments above tend to be open, and after solar noon. Which pollinator species visited which flowering plants depended significantly on when the observations were made—to the extent that the apparent importance of C. capillaris and its relatives is entirely different before and after noon.

Of course, these results apply directly to only a handful of species representing a particular group of flowering plants—but it’s a group with a lot of widespread and abundant members, and the result is straightforward and striking. Animal-pollinated plants may not behave much like clocks at all. Instead, they’re more like the patrons of a singles bar: they show up at about the same time and hang around until they find someone to buy them a drink. That’s a dynamic worth keeping in mind for studies of plant-pollinator interaction, since it suggests that the partners a pollinator chooses will depend, at least in part, on whether or not it’s out after closing time. ◼


Ewusie, J., & Quaye, E. (1977). Diurnal periodicity in some common flowers. New Phytologist, 78 (2), 479-485 DOI: 10.1111/j.1469-8137.1977.tb04854.x

Fründ, J., Dormann, C., & Tscharntke, T. (2011). Linné’s floral clock is slow without pollinators – flower closure and plant-pollinator interaction webs. Ecology Letters DOI: 10.1111/j.1461-0248.2011.01654.x

von Hase, A., Cowling, R., & Ellis, A. (2005). Petal movement in cape wildflowers protects pollen from exposure to moisture Plant Ecology, 184 (1), 75-87 DOI: 10.1007/s11258-005-9053-8


Deprived of pollinators, flowers evolve to do without

Who needs pollinators? Not monkeyflowers—at least not after a few generations of evolution. Photo by Brewbooks.

ResearchBlogging.orgThe loss of animal pollinators poses a potentially big problem for plants. However, many plant species that rely on animals to move pollen from anther to stigma have the capacity to make due if that service goes undone—and, as a new study released online early by the journal Evolution demonstrates, such plants can rapidly evolve to do without pollinators [$a] if they must.

The paper’s authors, Sarah Bodbyl Roels and John Kelly, demonstrate this using a simple greenhouse experiment with the monkeyflower Mimulus guttatus, a wildflower native to western North America, and a member of a genus rapidly developing into a major model system for studying the evolution of ecological isolation and floral evolution.

Mimulus species vary in their reliance on animal pollinators—some grow minimalistic flowers, with the anther so close to the stigma that pollen transfers without any assistance. In natural populations, M. guttatus is usually pollinated by bees, but individual plants vary in the distance between anther and stigma, and this variation has a genetic basis. So a population of M. guttatus deprived of pollinators would have the raw material to evolve a solution—natural selection would favor plants that are better able to self-pollinate. As the population evolved to be more self-fertilizing, it might also evolve to look more like self-pollinating Mimulus species, losing the bright petals that attract pollinators.

To see whether this could actually happen, Bobdyl Roels and Kelly challenged an experimental population of Mimulus guttatus to do without pollinators, and tracked its response.

The authors raised seeds derived from a natural wild population of Mimulus guttatus in greenhouses under two trial conditions: control populations were provided with hives of bumblebees to pollinate them when their flowers were ready for servicing; and experimental populations were left to produce what seed they could without pollinators. The authors collected the seeds produced by each population, and planted them to form the next generation.

A bumblebee digs for nectar in flowers of Mimulus moschatus. Photo by Mollivan Jon.

Early on in the experiment, the experimental populations deprived of pollinators fared badly. Without pollinators, the average plant produced two seeds or fewer by the end of the generation, compared to eight or ten seeds per plant in the population provided with bees. By the fifth generation, however, this was starting to improve—plants in both populations without pollinators were producing more seeds, and one of the two experimental populations produced nearly as many seeds as the control plants.

Examining the traits of plants produced by this final generation (actually, the grand-offspring of the fifth generation, to control for effects of inbreeding), the authors found that the average distance between the pollen-producing anther and the pollen-receiving stigma had shrunk significantly in plants from the experimental population. Across all the treatments, plants with a shorter distance between stigma and anther produced more self-pollinated seeds. There was no evolved change in other floral measurements, however—plants in the no-pollinators treatment had petals as big and showy as plants evolved with bumble bees.

In a natural population of Mimulus guttatus, the drop-off in seed production created by loss of pollinators should have much the same effect as in this experiment, creating a strong selective advantage for individual plants that can make more seeds on their own. The fact that the experimental plants did not evolve reduced petals could mean that in the cushy conditions of a greenhouse, there wasn’t much need to stop spending resources making showy flowers. Or maybe, when the major source of natural selection is the need to make any seeds at all, selection to save resources on flower production is relatively weak and correspondingly slow-acting.

As the authors point out, one of many changes humans are making to natural communities around the world is to disrupt pollination relationships. In a sense, experiments like theirs are being carried out worldwide, on hundreds of plant species—and each species will adapt, or fail to adapt, in its own way.


Bodbyl Roels, S., & Kelly, J. (2011). Rapid evolution caused by pollinator loss in Mimulus guttatus. Evolution DOI: 10.1111/j.1558-5646.2011.01326.x


Pollinating birds leave plants in the lurch

ResearchBlogging.orgPlants’ ancient relationship with animal pollinators is pretty crazy, when you think about it. On the one hand, it gives plants access to mates they can’t go find on their own, and it’s more efficient than making scads of pollen and hoping the wind blows some onto another member of your species. On the other hand, it can leave a plant totally dependent upon another species for its reproduction.

This catch is probably why lots of plants still use wind pollination strategies, or reserve the option to pollinate themselves if animals don’t do the job for them. Avoiding complete dependence on animal pollinators is likely to become more important in the modern era, as human disruption of the environment amplifies the inherent risk of entrusting your reproduction to another species [$a], a study in the latest issue of Science shows.

A flower of Rhabdothamnus solandri, waiting for pollinators who may never show up. Photo by Tonyfoster.

Sandra Anderson and her coauthors examined the health of populations of Rhabdothamnus solandri, a forest shrub native to the North Island of New Zealand. The flowers of R. solandri are classic examples of the pollination syndrome associated with birds—bright red-orange, with long nectar tubes. Rhabdothamnus solandri is incapable of self-pollinating, because its The flowers attract three native bird species, the tui, the bellbird, and the stitchbird. Thanks to human activity, all three of these birds “functionally extinct” in most of the range where R. solandri grows.

The bellbird and the stitchbird were eliminated from much of the North Island in the Nineteenth Century as European colonists cleared forests for farmland and introduced cats, rats, and dogs that preyed on the native fauna. Tuis have persisted, but tend to stay in the upper forest canopy—possibly to avoid rat predation—and don’t visit lower-growing shrubs. However, all three birds are still living as they did before Europeans arrived on two island nature preserves just a few kilometers off the North Island’s shores. This creates an inadvertent experiment in pollinator loss, allowing Anderson et al. to compare R. solandri populations on the mainland with those on the preserve islands to see how the plant gets on without its pollinators.

The short answer is: not well.

The three principle pollinators of R. solandri, the tui, the bellbird, and the stitchbird. Only the Tui is still common in most R. solandri habitat. Photos by kookr, angrysunbird, and digitaltrails.

To test whether R. solandri‘s reproduction is limited by pollen supply (as opposed to water or nutrients), the authors compared flowers that were either enclosed to prevent pollinator access, left open to natural pollination, or pollinated artificially. On the islands, plants left open set about as much fruit as plants pollinated by hand—but on the mainland, plants pollinated by hand set much more fruit than those left open. Mainland plants also produced smaller fruits, with fewer seeds per fruit, than island plants. The enclosed flowers set very little fruit, so it seems clear that pollen is the limiting factor for island and mainland R. solandri populations, and mainland populations aren’t getting enough.

The age structure of island and mainland R. solandri populations bears this out. Anderson et al. surveyed the island and mainland sites and counted the number of “adult” shrubs in a given area relative to recently sprouted seedlings. Island and mainland sites had similar densities of adult shrubs, but mainland sites had much lower densities of seedlings. It looks very likely that R. solandri populations on the North Island mainland are in decline as a direct result of losing pollinator services.

As Cagan Sekercioglu points out in an invited commentary [$a], this study demonstrates that species’ ecological roles can be strongly compromised even if they don’t go extinct. Tuis and bellbirds are not considered particularly endangered, and the stitchbird is classified as “vulnerable,” the lowest level of “threatened” under the system used by the International Union for the Conservation of Nature. Yet these birds’ local losses and adaptation to human activity have left R. solandri without adequate pollination services. Conserving biodiversity requires more than preventing extinction—but it can be quite a bit harder to preserve important relationships between species such as this one.


Anderson, S., Kelly, D., Ladley, J., Molloy, S., & Terry, J. (2011). Cascading effects of bird functional extinction reduce pollination and plant density. Science, 331 (6020), 1068-1071 DOI: 10.1126/science.1199092

Sekercioglu, C. (2011). Functional extinctions of bird pollinators cause plant declines. Science, 331 (6020), 1019-20 DOI: 10.1126/science.1202389


Finding the middle road: Flowers evolve to work with multiple pollinators

“I’ve had a lot of jobs in my life: boxer, mascot, astronaut, baby proofer, imitation Krusty, truck driver, hippie, plow driver, food critic, conceptual artist, grease salesman, carny, mayor, grifter, body guard for the mayor, country western manager, garbage commissioner, mountain climber, farmer, inventor, Smithers, Poochie, celebrity assistant, power plant worker, fortune cookie writer, beer baron, Kwik-E-Mart clerk, homophobe, and missionary, but protecting people, that gives me the best feeling of all.”
—Homer Simpson

In twenty-two seasons of The Simpsons, the eponymous family’s bumbling father Homer has tried his hand at dozens of different jobs, and failed hilariously at most of them. Homer is a one-man illustration of “Jack of all trades, master of none,” the idea that it’s hard to do many different things well. This principle applies more broadly than the curriculum vitae; in biology, it means that living things face trade-offs between different ways of making a living.

A wild radish (Raphanus raphaistrum) flower. Photo by Valter Jacinto.

For instance, a plant whose pollen is carried from flower to flower by just one pollinating animal only needs to match that one pollinator very well. But most plants’ flowers are visited by many different potential pollinators, and matching all of them probably means finding a middle ground among the best ways to match each individual pollinator. A study of one such “generalist” flower, the wild radish, has found exactly this: working with multiple partners takes evolutionary compromise [$a].

Wild radishes are visited by a wide variety of different insects, including honeybees, bumblebees, syrphid flies, and cabbage butterflies, among others. Each of these pollinators comes to a radish flower with a slightly different agenda. Butterflies are there for nectar, but bees like to eat pollen as well—and bumblebees will sometimes bite into the base of a flower and “steal” nectar without ever coming into contact with pollen. Figuring out how natural selection from each of these different pollinators adds up required some clever experimental design.

The study’s authors arrayed potted radish flowers inside a big mesh flight cage, and then introduced either bumblebees, honeybees, cabbage butterflies, or all three pollinators to visit the plants and circulate pollen from flower to flower. They measured the plants’ flowers before putting them in the flight cage, then let the pollinators do their thing. Afterward, the authors collected seeds resulting from the pollinators’ activity, grew them up, and measured the offspring to see whether their traits differed. The procedure was essentially one generation of experimental evolution.

A cabbage white butterfly (Pieris rapae), one of many pollinator species exerting natural selection on wild radishes. Photo by ComputerHotline.

By taking DNA fingerprints of both the parents and the offspring, the authors could also estimate the relationship between each parental plant’s floral measurements and the number of offspring it produced, either from its own seeds or by pollinating another plant.

The results are complex. Depending on the floral measurement under consideration, different pollinators selected in different directions, or the same direction, or not at all. One particularly interesting result, though, was in the effects each pollinator had on the “dimorphism” of the radish flowers’ stamens—the difference between the length of the shortest, and longest, of the male parts of the flower. Flowers only visited by honeybees evolved less dimorphic stamens, while flowers visited by either bumblebees or cabbage butterflies evolved more dimorphic stamens. Flowers in the treatment visited by all three pollinators, however, evolved to find a happy medium, an evolutionary compromise to work with the different partners.

The way these interactions played out in a flight cage probably don’t reflect exactly how they operate in the wild, but this is a pretty cool result all the same. I’ve written in the past about how incorporating multiple interactions can alter the way coevolution works. Gerbils under attack by fleas are less careful about watching for predators; but for the protists living inside pitcher plants, competitors can help distract predators. Here we have an example of multiple similar interactions pulling a generalized plant in different evolutionary directions.


Sahli, H., & Conner, J. (2011). Testing for conflicting and non-additive selection: Floral adaptation to multiple pollinators through male and female fitness. Evolution DOI: 10.1111/j.1558-5646.2011.01229.x


Pollination before flowers

ResearchBlogging.orgWhich came first, the pollinator or the pollinated? An article in this week’s Science suggests that a diverse group of insects may have been drinking nectar and pollinating plants millions of years before the appearance of modern flowering plants [$-a].

Panorpis communis, a modern scorpionfly species, and a sketch of ancient, pollinating scorpionflies. Photo by JR Guillaumin; sketch from Ollerton and Coulthard (2009).

Prior to the origins of modern flowering plants, or angiosperms, in the early-middle Cretaceous period, most of the diversity of land plants were gymnosperms. These plants are characterized by “naked seeds” — reproductive organs exposed to the air, where the wind can catch pollen and carry it from one plant to fertilize the ovules of another. In a world dominated by gymnosperms, the thinking used to be, animal pollinators were mostly unnecessary.

The new paper by Ren et al. challenges this idea with the description of a set of fossilized scorpionflies, all of which have strikingly long probosces that are clearly suited to sucking up liquid. The earliest of these fossils are from the Jurassic, tens of millions of years before the flowering plants began to diversify. In modern insects, sucking mouthparts like the ones described are associated with two kinds of feeding: drinking pollen, and drinking blood. To determine which was most likely in this case, Ren et al. performed energy-dispersive spectroscopy on the best-preserved fossil, and found no sign of the elevated levels of iron in the proboscis that would result from the residue of blood meals. This suggests that the scorpionflies were drinking nectar, or something like it.

Nectar has one major function in plants: to attract insects. Ant-protected plants reward their ants with nectar, and flowering plants use nectar to lure animal pollinators close enough to pick up or drop off pollen. If these ancient scorpionflies were, in fact, living on nectar, Ren et al. reason they were probably pollinating contemporary plants, which were all gymnosperms. The authors identify a diverse list of candidate host plants, including seed ferns and a relative of the modern ginkgo, whose reproductive structures were (1) too well-sheltered for efficient wind pollination or (2) included tubular structures similar to those that modern plants use to guide nectar-feeding pollinators. Finally, the authors point out, many modern gymnosperms produce “ovular secretions” that are very similar to the nectar produced by angiosperms.

As a neontologist, I’m often amazed how much can be told from million-years-old fossils — who knew there was a way to test for residual blood in a fossilized proboscis? At the same time, Ren et al. connect some mighty scattered dots to build their hypothesis. The real clincher is that it seems mighty unlikely that animal pollination would be rare in a world that already has both flying insects and pollen-producing plants. Animal pollination is much more efficient than wind pollination, and if there’s one constant in evolutionary history, it’s that living things rarely miss an opportunity like that.


Ollerton, J., & Coulthard, E. (2009). Evolution of animal pollination. Science, 326 (5954), 808-9 DOI: 10.1126/science.1181154

Ren, D., Labandeira, C., Santiago-Blay, J., Rasnitsyn, A., Shih, C., Bashkuev, A., Logan, M., Hotton, C., & Dilcher, D. (2009). A probable pollination mode before angiosperms: Eurasian, long-proboscid scorpionflies. Science, 326 (5954), 840-7 DOI: 10.1126/science.1178338


Video of yucca pollination

ResearchBlogging.orgWith permission from my doctoral advisor, Olle Pellmyr, I’ve just uploaded a unique video to Vimeo: a yucca moth laying eggs in, then pollinating, a yucca flower. I don’t know why I didn’t think of this earlier — it’s great footage, and deserves to be seen more widely.

A female yucca moth mates, then collects pollen from a yucca flower in specialized mouthparts. She carries it to another flower where, as shown in the video, she drills into the floral pistil with her ovipositor and lays eggs inside, then climbs to the tip of the pistil and applies pollen to fertilize the flower. When the flower develops into a fruit, the eggs hatch and the caterpillars eat some of the seeds inside.

Yuccas and yucca moths are completely dependent on each other [PDF] — nothing else pollinates yuccas, and the moths have no other source of food (they don’t eat as adults). Recently, the Pellmyr lab has shown that this interaction may be leading to speciation in one yucca species, the Joshua tree — Joshua trees pollinated by two different species of yucca moths have differently-shaped flowers [PDF], but these two tree types may not be totally genetically isolated [PDF]. I’ve written about this work before — for more information about the interaction, check out Olle’s publication page.


Godsoe, W., Yoder, J., Smith, C., & Pellmyr, O. (2008). Coevolution and Divergence in the Joshua Tree/Yucca Moth Mutualism. The American Naturalist, 171 (6), 816-23 DOI: 10.1086/587757

Pellmyr, O. (2003). Yuccas, yucca moths, and coevolution: A review. Annals of the Missouri Botanical Garden, 90 (1) DOI: 10.2307/3298524

Smith, C., Godsoe, W., Tank, S., Yoder, J., & Pellmyr, O. (2008). Distinguishing coevolution from covicariance in an obligate pollination mutualism: Asynchronous divergence in Joshua tree and its pollinators. Evolution, 62 (10), 2676-87 DOI: 10.1111/j.1558-5646.2008.00500.x


Pollinator isolation and divergence in floral shape

ResearchBlogging.orgThis post was written for The Giant’s Shoulders, a monthly blog carnival focusing on classic research.

Since Darwin, evolutionary biologists have thought that interactions between species cause diversification. However, it wasn’t until the second half of the Twentieth Century that scientists began to draw a connection between species interactions and speciation. One of the earliest of these studies was Verne Grant’s 1949 discovery of cleverly indirect evidence that pollinator isolation shapes the evolution of flowers [$-a].

A bee at work. (Flickr: jby)

Pollinator isolation is reproductive isolation created when animal pollinators don’t transfer pollen between plants of two different species. This could be because of pollinator behavior – say, because pollinator species tend to prefer a single plant. Or it could be because of the mechanics of pollen presentation by a flower, with each plant species applying pollen to a different part of a pollinator’s body so that foreign pollen is less likely to come into contact with the female floral parts. In either case, flowers are the key to the isolation – either to guide pollinators to their preferred target, or to make sure that the wrong pollen isn’t delivered.

Grant reasoned that pollinator isolation should have a real effect on how plant species are classified. Pollinator-isolated species probably have very different flowers; taxonomists, who look for characteristics that easily differentiate between related organisms, might therefore be more likely to use floral characteristics to tell pollinator-isolated plant species apart. To test this, Grant collected published classifications of plants pollinated by specialized animals (birds, bees, and long-tongued flies) and plants pollinated either by non-specialized animals, by water, or by wind.

Figure 1 from Grant (1959), showing the effect of pollinator isolation.

The result is presented in the tidy graph seen here. In plants pollinated by birds (A) and bees or long-tongued flies (CD), a much larger of the characteristics used by taxonomists to identify species were floral traits, compared to plants with non-specialized pollinators (E) or wind- and water-pollinated plants (FG). To follow up this result, Grant took systematic observations of pollinators’ movements through an experimental garden planted with three subspecies of Gilia capitata, each of which had differently colored flowers. The bees seemed to forage mainly among plants with similar flowers, and when Grant raised seeds from the experimental plants in the greenhouse, he found that there were fewer hybrids between subspecies than would be expected from random pollinator movement.

Generally, today, we wouldn’t assume that species classifications are an unbiased proxy for biological diversity – to some degree, they’re human constructs. But the basic idea that Grant develops, that speciation is an accidental consequence of plants’ interactions with pollinators, is still very important to how we understand the history of life. Together, flowering plants and insects make up the majority of the diversity of life on Earth, and it seems reasonable to think that this may be because the two groups interact so intimately.

More than fifty years after Grant’s study, pollinator isolation is a well-established mechanism for speciation. And the principle that Grant proposed, that increased divergence in floral traits is a sign of pollinator isolation, is still very useful. My lab, for instance, recently found that two forms of Joshua trees pollinated by different moth species are more different in certain floral dimensions than in non-floral traits [PDF]. That’s only the first step in what promises to be a long program of research (including my dissertation), seeking answers to some of the same questions that motivated Grant’s study.


W. Godsoe, J.B. Yoder, C.I. Smith, & O. Pellmyr (2008). Coevolution and divergence in the Joshua tree/yucca moth mutualism. The American Naturalist, 171 (6), 816-23 DOI: 10.1086/587757

V. Grant (1949). Pollination systems as isolating mechanisms in angiosperms. Evolution, 3, 82-97