Weed species (part 1)

This is the first in a series of essays titled, "Practical Evolution." Here are links to the whole series and the series introduction. I've decided to break the articles up into two parts, so that a full essay will appear in two successive weeks. So if you enjoy the current installment, by all means come back on Friday, when I will follow the threads of dispersal by way of an obnoxious animal pest right back to hominids.
Dandelion seeds

Probably the most well-known weed species in America is the one growing so prolifically in my yard: the dandelion. Any kid can tell you why dandelions spread so quickly: Their seeds are carried by the wind.

But for all their blowing of the puff-heads, as Goodwin calls them, my kids haven’t noticed what seems so obvious to me. Dandelions are almost indestructible. Pull them up and their leaves will dry, but the flowers on the dead-looking plants continue to develop seeds and puff out. Leave just a little of the taproot in the ground, and new leaves will sprout hydra-like from all directions. Gretchen drenched a goodly portion of our sidewalk dandelions last year with boiling water, wilting the leaves like spinach. Some of them really died. The rest sprouted right back.

Even when we’re talking about dainty little parasol-like seeds, much depends on the size of the whole package. The part that looks like a “seed,” a botanist calls an achene, which is technically a kind of fruit. The sunflower seeds that teenagers spit at softball games are fruits, too, but don’t try telling that to their mothers. The dandelion’s parachute, or pappus, is connected to the achene by one long shaft. A large, massive achene will tend to fall faster and won’t blow very far except in a very strong wind. Sunflowers and coneflowers go for the big achene strategy, which the winter-resident birds really appreciate come January. Dandelions go the opposite way: small achene, big pappus. That’s why weedy coneflowers are only a problem within five feet of my perennial beds, while weedy dandelions are in every disturbed field in the country.

Barkley Sound, Broken Island group

Barkley Sound, Broken Islands group

A remarkable botanical field study has been carried out over the last twenty-seven years by Martin Cody, in Barkley Sound, British Columbia. In 1981, Cody began keeping track of the plants on a few of the small islands that dot the Sound. After a few years, the study had changed into something grander than a simple plant census: Cody’s 2006 book on the subject begins, “This is a book about a field study that just grew and grew.” For more than twenty years, Cody and his students have counted plants on hundreds of tiny islands, most under a few hundred square meters in size. Many of these islands are so small that a species may disappear from them entirely, only to be re-established later by new colonists. With his careful censuses, Cody could determine when populations were newly founded, and when they became locally extinct.

By 1996, Cody had amassed information about weedy plants that had colonized dozens of the tiny islands of Barkley Sound. These plants all originally had come from Europe, including dandelions, their close relative cat’s-ear (Hypochaeris radicata), wall lettuce (Mycelis (Lactuca) muralis), and woodland groundsel (Senecio sylvaticus). That means that all the plants are recent colonists, and they have all been carried to the islands by the wind, each using similar seed packages that include an achene and pappus.

What makes a successful colonist? The first seeds to reach one of these tiny islands have done an exceptional thing: Wind has carried them across a large body of water to reach a tiny speck of virgin soil. Only the tiniest achenes tend to make this unlikely journey, and only if their parachutes are big enough to carry them.

Cody and his student, Jacob Overton, took some of these seeds from both island and mainland populations and embarked on a series of experiments. First, they set a stopwatch and dropped the seeds to see how fast they fell. The results of this seed drop were pretty much the same as when kids tie little parachutes onto green army men. Generally, the bigger the parachute, the slower the descent. But if the achene is small enough, even a very small pappus may be enough to keep it aloft. The key is the ratio: the volume of the pappus compared to that of the achene.

Wall lettuce seedhead

Seed head of wall lettuce (Mycelis muralis) with five achenes remaining

There’s nothing very surprising about this: The wind carries small seeds farther, particularly if they have big parachutes. Plants that bear this kind of seed have only a tiny chance of colonizing an island. But plants with big achenes and little parachutes may have no chance of colonizing new islands at all. So by sorting seeds, the wind and sea also sort the genes of the colonizing plants.

Indeed, Cody and Overton found that new populations of these plants on islands produce the kind of seeds that occasionally skip across to new islands. For example, the youngest island populations of wall lettuce — less than four years old — have the tiniest achenes. Their parachutes are small, too, but they generate a lot of loft. If the seeds of mainland wall lettuce are like a green army man tied to a kleenex, the first island colonists are like a BB stuck to a postage stamp.

Wall lettuce populations that manage to stick around on an island make a change. After ten years, they are producing achenes just as big as the mainland populations. But instead of big, fluffy parachutes, they grow smaller, stunted ones. Like green army men tied to smaller and smaller postage stamps, these seeds aren’t made for dispersal. They drop fast, right next door to their parents.

More dispersal is not always better, even if you’re a weed. A windy day might carry your seeds far away, but what if you live on a tiny island? Your seeds will be carried right off into the ocean. Worse, even if you live on a large island in Barkley Sound, there may be only a small strip of good habitat. Too far toward the water and you’re drowned by storm surges; too far inland and you have to compete with larger, more established plants. For these island weeds, the plants that disperse the least have more new seedlings.



Adaptable as they are, dandelions have done a poor job colonizing these tiny islands. The best of these weed species have managed to invade one island in five. Dandelions are found on less than a fourth that many. Despite their fame for being windblown, dandelion seeds actually drop faster than the wispy seeds of the other species on these islands. The rare colonists that do reach an island tend to last less than two years.

Their relative, cat’s-ear, has done better. Since it carries its flowers higher, on branched stems, cat’s-ear makes more headway into grassy areas away from the beach, giving it a bit more area to work with. As a result, its populations become extinct only a third as much turnover as dandelions. LIke the wall lettuce, the island cat’s-ear populations are different from the mainland, sporting big achenes and small parachutes. Dropping their seeds nearby seems to make a big difference as to whether their populations will persist.

Natural selection seems to have changed the island populations of both these weeds, favoring lower dispersal. But how do we know it is really selection on dispersal, and not just something about the islands affecting seed size?

Woodland groundsel, another of the weedy asters on the islands, provides one natural experiment. Cody and Overton dropped dozens of groundsel seeds from twenty populations, and found that the sizes of achenes and pappus made no difference at all to the speed that the seeds drop to the ground. Small groundsel achenes drop just as fast as large ones, no less than if Galileo had dropped them from the Leaning Tower. Even if selection had favored faster drop times, woodland groundsel achenes and pappus volumes should not have changed. And in fact, Cody and Overton found that they didn’t change: island populations of woodland groundsel have proportions almost exactly the same as the mainland.

Another experiment has been carried out in the most unlikely location. Hawksbeards (genus Crepis), also relatives of the dandelion, are common weeds spread across the Old World and into North America. Like dandelions, the hawksbeard Crepis sancta can colonize urban habitat, including the tiny patches of bare ground that cities leave in sidewalks for trees to grow. These are the urban equivalent of the tiny islands in Barkley Sound: sometimes far apart, they are separated by stretches of inhospitable concrete. In Montpelier, France, during the last few years Pierre-Olivier Cheptou and his colleagues have performed the urban version of Martin Cody’s work: counting plants and collecting seeds from the small hawksbeard colonies around the city. What it lacks in rugged charm, Montpelier makes up in record-keeping — knowing when the sidewalks were built, the researchers could work out the ages of different hawksbeard populations.

Unlike dandelions, C. sancta produces two different kinds of seeds: one with the achene and pappus carried on the wind, and another with no pappus at all. These seeds without parachutes do not go far; they scatter near the mother plant. That makes them bad for colonizing new patches of ground, but really good for establishing a sustained presence in one patch.

Compared to the rural C. sancta population, urban plants have more of the no-parachute seeds. These plants cast fewer seeds to the wind, and drop more on the ground to sprout nearby. Like the islands of Barkley Sound, these urban islands select for low dispersal.

Cheptou and his coworkers were able to go farther than Cody and Overton: They planted the seeds from urban patches in a greenhouse alongside seeds from the rural areas outside of town. The urban/rural differences were still present in the greenhouse plants, showing that growing in a small patch is not enough to change the seeds. Instead, the differences are caused by differences in genes, the result of selection on the urban plants.

What is more, they grew their own small patches of C. sancta, replicating the conditions of the urban plants. The seed heads of the real urban populations were the equal of those after 12 to 13 generations of selection on their fake urban patches. This rapid pace of selection on the urban plants matches that on the islands of Barkley Sound, where wall lettuce and cat’s-ear repeatedly showed strongest changes after only a few generations.

Cody and Overton 1996, Figure 1

The cycle of founding new island populations and their subsequent extinction.

Colonization is a filter: Only certain individuals may have the right combination of traits to cross a significant barrier and inhabit a new place. But the traits that enable individuals to make such crossings are not always very good for staying in the new place.

For certain species, selection yields a complicated pattern. A plant that is a local failure may be a global success, as it casts its seeds to far-off places, a few of which may grow into significant new populations. Such high-dispersing plants may do very poorly as a new population grows, but may be the only ones with a chance to escape when local extinction looms. High-dispersing individuals are also the ones most likely to carry their genes from one population to another, providing the potential for adaptations across the species as a whole.

Without dispersal, a species may become moribund, unable to change in the face of new challenges. At the extreme, this has been called “evolutionary suicide.” Traits that are selected locally, because they enhance the reproduction of individuals in the species, may nevertheless doom the species as a whole by reducing adaptability to changing conditions. The cost of maintaining the ability to change is a constant unstable balance between dispersal and staying power.

For the weeds of Barkley Sound, this unstable balance ultimately favors the colonizers. New island populations do not last long. The local selection in favor of lower dispersal helps them to persist for a time, but on the scale of centuries all of them will fail, only to be replaced by new colonists from elsewhere.

Friday: We turn from plant parachutes to animal dispersal mechanisms. And by “mechanisms” I mean, “legs.”


I am indebted to George Cox, whose excellent Alien Species and Evolution: The Evolutionary Ecology of Exotic Plants, Animals, Microbes, and Interacting Native Species (2004) gives a good account of Cody and Overton’s study of island weeds. I can’t say enough good things about this book.

Martin Cody’s book, Plants on Islands: Diversity And Dynamics on a Continental Archipelago, gives an account of the history of his field project, as well as reviews of the many areas of ecology that it has helped to illuminate.

Many readers may notice that there is another factor besides low dispersal potential by which selection might favor large achenes in these plant species. Larger seeds may be advantageous because, with more stored energy, they may tend to germinate more readily or grow more quickly, thus making them more competitive. In this circumstance, selection would pertain to energy allocation or reproductive effort, rather than dispersal. I’ll be discussing that form of selection in a later essay. In the present cases, dispersal is indicated as the target of selection, but seed size and energy allocation may also be involved.

If seed morphology makes so little difference to drop time in Senecio sylvaticus, you may wonder what maintains achene size and pappus volume in this species. At least, that’s what I wondered. Without more information it is hard to say, but I would hypothesize that achene and pappus volume may be genetically correlated in a way that is hard to select for one and not the other. In the other species, the selection on achene size and pappus volume is apparently antagonistic – larger achenes are favored at the same time as smaller pappus. If these two traits are strongly genetically correlated, it becomes harder to select for lower dispersal. It would also be worth investigating whether other traits besides these volumes may have evolved in the island populations.

The rapid evolution of urban hawksbeard (Crepis sancta) was examined by Cheptou et al. (2008). Pierre-Olivier Cheptou has a webpage at CEFE, including a description of Crepis sancta as a model system.

Photo credits:

Dandelion seeds: Photo by Piccolo Namek, on Wikimedia. GPL license.

Barkley Sound: Photo by Sam Wilson (BaylorBear78) on Flickr. Creative commons noncommercial share-alike license.

Cat’s-ear: Photo by Ian Boyd, on Flickr. Creative Commons noncommercial attribution license.

Wall lettuce: Photo by Mollivan John, on Flickr. Creative Commons noncommercial attribution license.

Colonization figure: Figure 1 from Cody and Overton, 1996.


Cheptou, P-O, Carrue O, Souifed S, Cantarel A. 2008. Rapid evolution of seed dispersal in an urban environment in the weed Crepis sancta. Proc Nat Acad Sci USA 105:3796-3799. doi:10.1073/pnas.0708446105

Cody ML. 2006. Plants on Islands: Diversity and Dynamics on a Continental Archipelago. University of California Press, Berkeley, CA.

Cody ML, Overton JM, 1996. Short-term evolution of reduced dispersal in island plant populations. J Ecol 84:53-61.

Cox G. 2004. Alien Species and Evolution. Island Press, Washington.