« Less-vicious viruses evolve in viscous cannibal populations | Main | Evolution of color vision: transgenic mice see red »

Splitting species: sneak attacks from strategic hamlets

This week's paper is "Colour pattern as a single trait driving speciation in Hypoplectrus coral reef fishes?" by Oscar Puebla (Smithsonian Tropical Research Institute, Panama) and colleagues in Canada and the UK, published on-line (no volume or page number yet) in Proceedings of the Royal Society. (I was planning to review a paper on the evolutionary history of genetic differences between chimps and humans, suggested by a reader, but decided I didn't understand it well enough myself to explain it clearly. Is there a volunteer guest blogger out there?)

Actually, there's a bit of a connection between the two papers. At some point, the ancestors of humans must have stopped having babies with the ancestors of chimps. Otherwise, we'd still be one species. We might have evolved a lot from our common ancestor, but we'd be evolving together, not separately. Interbreeding is a problem for the production of new species in general; the resulting "gene flow" can prevent differentiation into separate species.

One easy solution is geographic separation. Finches on different islands in the Galapagos group rarely interbreed with each other, and never with their ancestral species on the mainland. So natural selection, working in different directions on the different islands, isn't swamped by interbreeding. This eventually produced enough change that the finches would at least hesitate to mate if brought back together.

But can species separate without being physically separated? There are already a few known examples of this, but the authors of this week's paper may have caught "sympatric" speciation in reef fish known as hamlets red-handed. Uh, finned.

Hamlets have a variety of color patterns (morphs) on the same coral reef. From the picture in the article, I would have guessed that they were several different species, just as I would have guessed that brocolli and cabbage are different species, but in both cases the different morphs sometimes produce offspring together, as if they were all one species.

Puebla and colleagues collected several individuals of each of three different morphs for genetic analysis. There were significant genetic differences among morphs. In particular, one gene (or one region of a chromosome, anyway) differed consistently among fish with different color patterns. Field observations showed that the fish usually mate with another of the same color pattern (247 out of 251 observations). Within-species mating of like with like is is known as "assortative mating." This behavior and the genetic differences were enough to classify the morphs as "incipient species."

Assortative mating based on color morph would favor eventual speciation, although you might think this would require one gene for being blue, and another gene for mating with blue. But there's more to the story.

Hamlets are predatory fish, so their prey should avoid them. But apparently some mutants arose that somewhat resembled various nonpredatory fish. The hamlets exploit this resemblance. When divers followed 12 individuals of one of the morphs, the hamlets spent a lot of time following nonpredatory fish that they resembled -- if you have access to this journal, e.g., via a university library, check out the video! -- and were most successful in catching prey when in this nonthreatening company.

But we seem to be swimming into a sea of troubles here. Do the blue morphs now need a gene for being blue, a gene for mating with blue, and a gene for associating with blue nonpredatory fish, while yellow morphs need a gene for being yellow, a gene for mating with yellow, and a gene for associating with yellow nonpredatory fish?

Maybe not. It seems to me that there are at least three hypotheses consistent with the data in this week's paper:

1) Maybe color, mating preference, and mimicry behavior are indeed controlled by separate genes at present. How could this system of genes have evolved? Here's one scenario: the original genetic variation among the ancestors of today's morphs was variation in a gene that influenced which type of nonpredatory fish they associate with. That would lead to some assortative mating, through propinquity, even without preference for a mate that shows a particular color pattern. Differences among hamlets in the nonpredatory fish they associate with would then drive divergent natural selection, favoring mutations making each mimic look even more like its model. There would also be selection for mating preferences that would strengthen assortative mating, since hybrids between two incipient morphs wouldn't resemble either of the nonthreatening model species and would therefore be less successful predators and less likely to survive to reproduce themselves.

2) Or maybe morphs initially varied only in appearance; each individual hamlet learned which nonpredatory model fish to hang out with by trial and error (based on hunting success), and then mated with others associating with the same model. These behaviors could subsequently have been strengthened by natural selection, as in hypothesis 1.

3) If all of the hamlets share a gene for "following nonpredatory fish that look like me" and another gene for "mating with fish that look like me" (or genes for associating with fish, of whatever species, that look like themselves and for mating with nearby members of their own species), then as few as one gene, controlling color pattern, would need to vary among morphs. This assumes, among other things, that each individual knows what it looks like. They don't have access to mirrors, but maybe their fisheye-lens eyes have a sufficiently wide-angle view that they can see themselves. My impression is that this is the hypothesis preferred by the authors.

I hope we will be seeing more research on these interesting fish. For example, mate-choice experiments under laboratory conditions, where the fish wouldn't have the opportunity to associate with any nonpredatory fish, could potentially disprove the "mate with those who associate with the same nonpredators" hypothesis. (Alexandra Basolo uses clever three-part aquaria to study the mating preferences of female swordtail fish. Crossed polarizing glass in the two dividers lets the female see and approach males on each side, while they can see her but not each other.) It would also be interesting to interbreed morphs in the laboratory and see how the hybrids behave, both with respect to mating preferences (including with nonhybrids) and with respect to associating with different nonpredatory fish.

The same issue of Proceedings of the Royal Society has a review article on the evolution of "family living" in birds, experiments on the evolution of mutation rate in bacteria by Angus Buckling (an evolutionary biologist mentioned last week), a new fossil of a dinosaur that dug underground dens, research on mate choice in birds, and a computer modeling paper that tries to explain some aspects of human mating behavior, including societies that consider children to have more than one biological father. This seems unlikely, but maybe we should teach the controversy?


So, just to clear this up...Homo sapiens and Pan troglodytes (as the species we know today) had a single common ancestor and a direct split? Or are we and they descended from two different ancestors, which split a few hundred mya? Does that question make sense?

There's also an interesting review article on hybrid speciation by James Mallet in the March 15 2007 issue of Nature, where he says the phenomenon may be more widespread than originally believed.

Based on molecular evidence (J. Heredity 92:469, for example), the last common ancestor of humans and chimps lived about 5 million years ago. Of course, we had earlier common ancestors that we share with gorillas, even earlier ones that we share with monkeys, etc.

hypothesis 3 feels (highly unscientific, i know) bogus to me. it seems unlikely the fish "know" (apart from the fact that fish have no perception of self) what they look like. it seems more logical they would only base their behaviour on the reaction of other fish to them, pointing to hypothesis 2. this would also better fit a gradual diversification of the colour patterns.

You may be right; I like hypothesis 2 also, though I would like to see experimental tests of learning ability in these fish. I may be wrong that hypothesis 3 was preferred by the authors; I couldn't tell for sure.

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Type the characters you see in the picture above.