Testing Natural Selection: Part 2

 
There are several interesting articles about evolution in the Januray 2009 issue of Scientific American. One of the most interesting is an article by H. Allen Orr of the University of Rochester (NY, USA). The magazine title is "Testing Natural Selection"¹ and, as the title implies, the focus is on evolution by natural selection.

Orr's article gives us an opportunity to compare and contrast the views of an adaptationist (Orr) and a pluralistic approach to evolution.

In Testing Natural Selection: Part 1 we discussed two of Orr's opinions: (1) random genetic drift is not as common as most people think, and (2) most (if not all) visible phenotypic change is driven by natural selection.

Here, we discuss Orr's ideas about speciation.

When we say "speciation" we're talking about the biological species concept. Speciation occurs when two formerly compatible populations evolve to the point at which they can no longer interbreed. The key question is what causes this reproductive isolation and how does it evolve?

To contemporary biologists, then, the question of whether natural selection drives the origin of species reduces to the question of whether natural selection drives the origin of reproductive isolation.

For much of the 20th century, many evolutionists thought the answer was no. Instead they believed that genetic drift was the critical factor in speciation. One of the most intriguing findings from recent research on the origin of species is that the genetic drift hypothesis about the origin of species is probably wrong. Rather natural selection plays a major role in speciation.

Orr is correct to point out that random genetic drift is important in speciation. It's the mechanism described in many evolutionary biology textbooks, though it's not the mechanism that most people think about when they think about speciation.

Many biologists have always believed that natural selection plays a much more important role in speciation than random genetic drift. They aren't happy with the textbook description. Orr is one of these biologists. He now claims that the drift explanation is "probably wrong."

Let's think about what has to happen when two species become reproductively isolated. We'll use allopatric speciation as an example.²

We begin with a situation where two populations (races, subspecies) are geographically separated. There is very little gene flow between them so they evolve independently of each other. Over time, they may come to be different because each is adapting to different environments or they may just drift apart by accident. With respect to the actual speciation event, these differences don't matter.

From time to time, individuals from the two subspecies will interbreed to produce fertile offspring. This is responsible for limited gene flow between the subspecies and it proves that speciation has not occurred. If the barrier between the two populations breaks down they will merge back into a single population.

But if the two species have been separated for a long period of time, mutations that prevent interbreeding will accumulate and hybrids will become less and less viable until eventually no fertile hybrids are produced and speciation is complete. There are many ways that this can happen but a common hypothesis involves the build-up of post-zygotic genetic incompatibilities called Dobzhansky-Muller (D-M) incompatibilities.

How are D-M incompatibilites fixed in the population? If they interfere with the matings of individuals from the two populations then how come they don't contribute to infertility when individuals from the same population mate with each other? When the mutation first arises it seems to have a very strange property. It doesn't affect matings between an individual carrying the D-M allele and an individual not carrying that allele from the same population but it does affect matings between the individual carrying the new D-M allele and and an individual from the other population.

In order for this to happen there must already be some genetic differences between the two populations in terms of mating and reproduction. Those differences have accumulated in each of the populations but they must not have an effect on hybrid crosses. Presumably, the new D-M allele is not harmful in one genetic background but it is in the other.

Are these pre-existing potentiators neutral within a population, in which case they become fixed by random genetic drift? Or, are they beneficial in one of the populations, and not in the other, in which case they are fixed by natural selection? The general consensus has been that they are neutral within a population and they accumulate by accident. When enough of them become fixed the cumulative effect is to prevent hybridization. The last allele to arise, the D-M incompatibility allele, is the straw that breaks the camel's back.

Orr believes that the alleles are beneficial in one of the populations. Thus, according to him, reproductive isolation is driven by natural selection. He gives two examples.

The first one is the incomplete speciation in monkeyflower subspecies. I described this in an earlier posting [Speciation in Monkeyflowers], where I pointed out that the role of natural selection was not clear. The differences in flower color and pollinators could have arisen by selection if one postulates changes in the bee population but they could also be due to chance.

For an adaptationist like Allen Orr there's no doubt about what happened.

A good example is the evolutionary history of the two monkeyflower species mentioned earlier. Because their pollinators seldom visit the “wrong” species of monkeyflower, the two species are almost completely isolated reproductively. Even though both species sometimes occur in the same locations in North America, a bumblebee that visits M. lewisii almost never visits M. cardinalis, and a hummingbird that visits M. cardinalis almost never visits M. lewisii. Thus, pollen is rarely transferred between the two species. In fact, Schemske and his colleagues showed that pollinator differences alone account for 98 percent of the total blockage in gene flow between the two species. In this case, then, there can be no doubt that natural selection shaped the plants’ adaptations to distinct pollinators and gave rise to strong reproductive isolation.

This is not a good example of speciation by natural selection. We simply don't know if the flower color mutation spread in one of the populations because it conferred a selective advantage on individuals within that population.

Besides, these two "species" will still form viable hybrids so they're not really species in the first place.

The other example of presumed speciation by natural selection comes from studies on Drosophila There are several example of D-M incompatibility alleles that have been identified. In some of them, there is evidence at the sequence level for rapid fixation. If correct, this is a good indication that the alleles have become fixed by natural selection. The resulting reproductive isolation is an epiphenomenon.³

One example is OdsH in Drosophila mauritiana. It appears to result in an increase in sperm production so it may have been selected in the early population of this species, before it became a species. Presumably, the allele was beneficial in the genetic background that had evolved up to that point and presumably it was detrimental in the genetic background of whatever subspecies it was related to.

The genetic background is obviously part of the speciation event. I suppose that if even one of the D-M alleles is selected then it's fair to say that speciation by natural selection took place.

The question is whether this is common or not. Shucker et al. (2005) looked at post-zygotic reproduction isolation in two populations of grasshopper and provided evidence that all the D-M incompatibilities could be adequately explained by random genetic drift. We'll need to have many more examples in order to decide whether natural selection explains most speciation events.

Personally, I find it easier to understand how reproductive isolation could arise by accidental accumulation of many neutral alleles that eventually lead to reproductive isolation. It's harder to envisage alleles that confer a selective advantage within one population but are extremely detrimental in the other.

Orr doesn't agree.

The studies of the monkeyflower and of hybrid sterility in fruit flies only begin to scratch the surface of a large and growing literature that reveals the hand of natural selection in speciation. Indeed, most biologists now agree that natural selection is the key evolutionary force that drives not only evolutionary change within species but also the origin of new species. Although some laypeople continue to question the cogency or adequacy of natural selection, its status among evolutionary biologists in the past few decades has, perhaps ironically, only grown more secure.

I'm not an expert on speciation and I don't hang out with people who work in the field. However, my general impression from reading the scientific literature is that Orr's statements may be somewhat exaggerated. From what I can see, there are a great many evolutionary biologists who question the hegemony of natural selection. Their numbers seem to be growing, not shrinking.

I don't know where Orr is coming from when he implies that laypeople question the adequacy of natural selection. In my experience laypeople only think about natural selection. They have no idea that there are any other mechanisms of evolution.

---

1. The website title is "Testing Natural Selection with Genetics."

2. In allopatric speciation the two diverging populations are geographically separated. That's what makes them distinct populations. In sympatric speciation the two populations may exist in the same geographical and restricted gene flow between them is due to other factors. It's easier to visualize what's happening during allopatric speciation but the logic can apply to sympatric speciation as well.

3. I don't think Orr is actually proposing that there would be selection for reproductive isolation. How would that work?

Shuker, D.M., Underwood, K., King, T.M., and Butlin, R.K. (2005) Patterns of male sterility in a grasshopper hybrid zone imply accumulation of hybrid incompatibilities without selection. Proc. Roy. Soc. B 272:2491-2497. [DOI: 10.1098/rspb.2005.3242]