Why Evolution Is True (31 page)

This prediction too is fulfilled: we see many sister species divided by a geographic barrier. Each side of the Isthmus of Panama, for example, harbors seven species of snapping shrimp in shallow waters. The closest relative of each species is another species on the
other
side. What must have happened is that seven ancestral species of shrimp were divided when the isthmus arose from beneath the sea three million years ago. Each ancestor formed an Atlantic and a Pacific species. (Snapping shrimp, by the way, are a biological marvel. Their name comes from the way they kill. The shrimp doesn’t touch its prey but, by snapping together its single oversized claw, creates a high-pressure sonic blast that stuns its victim. Large groups of these shrimp can be so noisy that they confuse the sonar of submarines.)

It’s the same with plants. You can find pairs of sister species of flowering plants in eastern Asia and eastern North America. All botanists know that these areas have similar flora, including skunk cabbage, tulip trees, and magnolias. One survey of plants uncovered nine pairs of sister species, including trumpetvines, dogwoods, and mayapples, with each pair having one species in Asia and its closest relative in North America. Botanists theorized that each of the nine pairs used to be a single species continuously distributed across both continents, but these became geographically isolated (and began to evolve separately) when the climate became cooler and dryer about five million years ago, wiping out the intervening forest. Sure enough, DNA-BASED dating of these nine pairs puts their divergence times at around five million years.

Archipelagoes are a good place to find out whether speciation requires physical isolation. If a group has produced species within a cluster of islands, then we should find that the closest relatives live on different islands rather than the same one. (Single islands are often too small to allow the geographic separation of populations that is the first step in speciation. Different islands, on the other hand, are isolated by water, and should allow new species to arise easily.) This prediction also turns out to be generally true. In Hawaii, for instance, sister species of
Drosophila
flies usually occupy different islands; this is also true of the lesser-known but still dramatic radiations of flightless crickets and lobelia plants. What’s more, the dates of the speciation events in
Drosophila
have been determined using the flies’ DNA, and we find, exactly as predicted, that the oldest species are found on the oldest islands.

Still another prediction of the geographic-speciation model rests on the reasonable assumption that geographic speciation is still occurring in nature. If that’s so, we should be able to find isolated populations of a single species that are beginning to speciate, and show small amounts of reproductive isolation from other populations. And sure enough, there are many examples. One is the orchid
Satyrium hallackii,
which lives in South Africa. In the northern and eastern parts of the country it is pollinated by hawkmoths and long-tongued flies. To attract these pollinators, the orchid has evolved long nectar tubes in its flowers; pollination can occur only when the long-tongued moths and flies get close enough to the flower to stick their tongues into the tubes. But in coastal regions, the only pollinators are short-tongued bees, and here the orchid has evolved much shorter nectar tubes. If the populations were to live in an area containing all three types of pollinators, the long- and short-tubed flowers would undoubtedly show some genetic isolation, for long-tongued species can’t easily pollinate short-tubed flowers, and vice versa. And there are many examples of animal species in which individuals from different populations mate less readily than do individuals from the same population.

There’s a final prediction we can make to test geographic speciation: we should find that reproductive isolation between a pair of physically isolated populations increases slowly with time. My colleague Allen Orr and I tested this by looking at many pairs of Drosophila species, each pair having diverged from its own common ancestor at various times in the past. (With the molecular-clock method described in chapter 4, we could estimate the time when a pair of species began diverging by counting the number of differences in their DNA sequences.) We measured three types of reproductive barriers in the laboratory: mating discrimination between the pairs, and the sterility and inviability of their hybrids. Just as predicted, we found that the reproductive isolation between species increased steadily with time. Genetic barriers between groups became strong enough to completely prevent interbreeding after about 2.7 million years of divergence. That’s a long time. It’s clear that, at least in fruit flies, the origin of new species is a slow process.

The way we discovered how species arise resembles the way astronomers discovered how stars “evolve” over time. Both processes occur too slowly for us to see them happening over our lifetime. But we can still understand how they work by finding snapshots of the process at different evolutionary stages and putting these snapshots together into a conceptual movie. For stars, astronomers saw dispersed clouds of matter (“star nurseries”) in galaxies. Elsewhere they saw those clouds condensing into protostars. And in other places they saw protostars becoming full stars, condensing further and then generating light as their core temperature became high enough to fuse hydrogen atoms into helium. Other stars were large “red giants” like Betelgeuse; some showed signs of throwing off their outer layers into space; and others still were small, dense white dwarfs. By assembling all these stages into a logical sequence, based on what we know of their physical and chemical structure and behavior, we’ve been able to piece together how stars form, persist, and die. From this picture of stellar evolution, we can make predictions. We know, for example, that stars about the size of our sun shine steadily for about ten billion years before bulging out to form red giants. Since the sun is about 4.6 billion years old, we know that we’re roughly halfway through our tenure as a planet before we’ll finally be swallowed up by the sun’s expansion.

And so it is with speciation. We see geographically isolated populations running the gamut from those showing no reproductive isolation, through those having increasing degrees of reproductive isolation (as the populations become isolated for longer periods), and, finally, to complete speciation. We see young species, descended from a common ancestor, on either side of geographic barriers like rivers or the Isthmus of Panama, and on different islands of an archipelago. Putting all this together, we conclude that isolated populations diverge, and that when that divergence has gone on for a sufficiently long time, reproductive barriers develop as a by-product of evolution.

Creationists often claim that if we can’t see a new species evolve during our lifetime, then speciation doesn’t occur. But this argument is fatuous: it’s like saying that because we haven’t seen a single star go through its complete life cycle, stars don’t evolve, or because we haven’t seen a new language arise, languages don’t evolve. Historical reconstruction of a process is a perfectly valid way to study that process, and can produce testable predictions.
³⁹
We can predict that the sun will begin to burn out in about five billion years, just as we can predict that laboratory populations artificially selected in different directions will become genetically isolated.

Most evolutionists accept that geographic isolation of populations is the most common way that speciation takes place. This means that when closely related species live in the same area-a common situation-they actually diverged from each other during an earlier time when their ancestors were geographically isolated. But some biologists think that new species can arise without the need for any geographic separation. In
The
Origin, for example, Darwin repeatedly suggested that new species, especially plants, could arise within a very small, circumscribed area. And since Darwin’s time, biologists have argued fiercely about the likelihood that speciation could occur without geographic barriers (this is called
sympatric
speciation, from the Greek for “same place”). The problem with this, as I mentioned before, is that it’s hard to split one gene pool in two while its members remain in the same area, because interbreeding between the diverging forms will constantly be pulling them back into a single species. Mathematical theories show that sympatric speciation is possible, but only under restrictive conditions that may be uncommon in nature.

It’s relatively easy to find evidence for geographic speciation, but it’s much harder for sympatric speciation. If you see two related species living in one area, that doesn’t necessarily mean that they arose in that area. Species constantly shift their ranges as their habitats expand and contract during long-term changes in climate, episodes of glaciation, and so on. Related species living in the same place may have arisen elsewhere and come into contact with each other only later. How can we be sure, then, that two related species living in one place actually
arose
in that place?

Here’s one way to do it. We can look at habitat islands: small patches of isolated terrain (like oceanic islands) or water (like tiny lakes) that are generally too small to contain any geographic barriers. If we see closely related species in these habitats, we could conclude that they formed sympatrically, since the possibility of geographic isolation is remote.

There are only a few examples. The best involves cichlid fish in two tiny lakes in Cameroon. These isolated African lakes, filling the craters of volcanoes, are too small to permit populations within them to become spatially separated (their areas are 0.2 and 1.6 square miles, respectively). Nevertheless, each lake contains a different miniradiation of species, each recently descended from a common ancestor: one lake has eleven species, the other nine. This is perhaps the best evidence we have for sympatric speciation, although we don’t know how and why it happened.

Another case involves palm trees on Lord Howe, an oceanic island lying in the Tasman Sea about 350 miles off the east coast of Australia. Although the island is small-about five square miles-it contains two native species of palms, the kentia and curly palms, which happen to be each other’s closest relatives. (The kentia palm may be familiar-it’s a popular houseplant throughout the world.) These appear to have arisen from an ancestral palm that lived on the island about five million years ago. The chance that this speciation involved geographic isolation appears quite small, especially because the palms are pollinated by wind, which can spread pollen over a large area.

There are a few more examples of sympatric speciation, though they’re not quite as convincing as these. What is most surprising, however, is the number of times that sympatric speciation has not occurred given the opportunity. There are many habitat islands that contain a fair number of species, but none of these are each other’s closest relatives. Obviously, sympatric speciation has not occurred on those islands. My colleague Trevor Price and I surveyed bird species on isolated oceanic islands, looking for the presence of close relatives that might indicate speciation. Of forty-six islands we examined, not a single one contained endemic bird species that were each other’s closest relatives. A similar result was seen for Anolis lizards, the small green animals often sold in pet shops. Closely related
Anolis
species simply aren’t found on islands smaller than Jamaica, which is large, mountainous, and varied enough to allow geographic speciation. The absence of sister species on these islands shows that sympatric speciation can’t be common in these groups. It also counts as evidence against creationism. After all, there’s no obvious reason why a creator would produce similar species of birds or lizards on continents but not on isolated islands. (By “similar,” I mean so similar that evolutionists would regard them as close relatives. Most creationists do not accept species as “relatives,” since that presupposes evolution.) The rarity of sympatric speciation is precisely what evolutionary theory predicts, and is further support for that theory.

There are, however, two special forms of sympatric speciation that are not only common in plants, but also give us our only cases of “speciation in action”: species actually forming during a human lifetime. One of them is called
allopolyploid speciation.
The curious thing about this form of speciation is that instead of beginning with isolated populations of the same species, it starts with the hybridization of two
different
species that live in the same area. And it usually requires that those two different species also have different numbers or types of chromosomes. Because of this difference, a hybrid between the species won’t undergo proper pairing of chromosomes when it tries to make pollen or ovules, and it will be sterile. However, if there was a way to double every chromosome in that hybrid, each chromosome would now have a pairing partner, and the doubled-chromosome hybrid would be fertile. And it would also be a new species, because while interfertile with other similar hybrids, it would be unable to interbreed with either of the original two parent species, for such a mating would yield sterile offspring with odd numbers of chromosomes. In fact, such “doubled-chromosome” allopolyploids occur with regularity, giving rise to new species.
⁴⁰

Polyploid speciation doesn’t always require hybridization. A polyploid can arise simply by doubling all of the chromosomes of a single species—a process called
autopolyploidy.
This too results in a new species, for each autopolyploid is able to produce fertile hybrids when mating with other autopolyploids, but produces only sterile hybrids when mating with the original parental species.
⁴¹

To get either type of polyploid speciation, you need a rare event to occur in two successive generations: the formation and union of sperm and eggs with abnormally high numbers of chromosomes. Because of this, you might have thought that such speciation would be very rare indeed. But it isn’t. Given that a single plant can produce millions of eggs and pollen grains, an improbable event eventually becomes probable. Estimates vary, but in well-studied areas of the world it’s been estimated that as many as a quarter of all species of flowering plants were formed via polyploidy. The fraction of existing species that had a polyploidy event occurring
somewhere
in their ancestry, on the other hand, could be as high as 70 percent. This is obviously a common way that new plant species arise. What’s more, we find polyploid species in nearly all groups of plants (a notable exception is trees). And many plants used for food or decoration are polyploids or sterile hybrids that had a polyploid parent, including wheat, cotton, cabbage, chrysanthemums, and bananas. This is because humans recognized the hybrids in nature as having useful traits from both parental species, or they deliberately produced the polyploids to create desirable gene combinations. Two everyday examples from your kitchen show this. Many forms of wheat have six sets of chromosomes, arising from a complicated series of crosses, involving three different species, that were made by our ancestors. Commercial bananas are sterile hybrids between two wild species, having two sets of chromosomes from one species and one set from the other. Those black specks in the middle of your banana are, in fact, aborted plant ovules that don’t become seeds because their chromosomes can’t pair properly. Since banana plants are sterile, they must be propagated from cuttings.