A number of years ago, I came across a comparison of the Tasmanian wolf (from Tasmania, of course, the southernmost province of Australia) and a North American gray wolf. Their skulls, as the image below shows, are remarkably similar.
You may think that this is not that unusual because they are, after all, both wolves. Well, you would be both wrong and right. They are both “wolves” but they are only distantly related. They are both mammals, but the Tasmanian wolf, also known as a thylacine, was a marsupial mammal and the gray wolf is a placental mammal, with about 180 million years of evolutionary separation between them.
For reference, humans and chimpanzees split off from the same evolutionary line about 6 million years ago — 1/30th of the time separating the placental and marsupial wolves from each other.
So how did these two very different species evolve such astounding similarities with such a vast span of time separating them? Well, no one really knows.
But there are some good ideas attempting to explain this phenomenon of “convergent evolution,” and the prevailing view among evolutionary biologists is that similar conditions of life, with the same physical laws (as is always the case for any species on this little planet of ours), will often lead to similar evolutionary outcomes.
This explanation only gets us so far, however, as we’ll see in the interview below.
Jonathan Losos is director of the Living Earth Collaborative and the William H. Danforth Distinguished University Professor at Washington University. He is formerly a professor of evolutionary ecology at Harvard University and the curator of herpetology at the Harvard Museum of Comparative Zoology.
He is the author of the excellent book Improbable Destinies: Fate, Chance, and the Future of Evolution. His book is all about convergent evolution, whether it is the norm or the exception to the rule, and what we can conclude about evolution more generally from the many cases of convergent evolution that he discusses — including his own long-term research on anole lizards in the Caribbean.
I really enjoyed the book, and it’s a great example of what a popular science book can be: accessible and interesting but also presenting original science in a detailed manner and in a way that leads to real understanding.
We conducted this interview via email in 2018.
• • •
Tam Hunt: What led you to a career in evolutionary biology?
Jonathan Losos: I was a typical 5-year-old dinosaur enthusiastic, but I never lost my interest in reptiles. Raising caimans (a relative of alligators) helped deepen my interest in herpetology, and college and graduate school took it from there. Reading Stephen Jay Gould’s monthly columns in Natural History magazine was very mind-expanding as well.
TH: A large focus of your book is the idea of “convergent evolution,” which refers to cases of different species apparently evolving similar solutions to the problems that nature presents. What are the most striking examples of convergent evolution?
JL: Well, there are so many! My book is full of examples, and there have been several massive volumes published in the last few years documenting case after case of sometimes extraordinary convergence. You can define “striking” in different ways. The fact that two species of seasnakes were considered to be a single species until DNA data showed that they were related to different species is one example. On the other hand, no one would mistake a porpoise for a shark, and yet they are extremely similar. Cactus and their old world, euphorb, counterparts … there are just so many!
TH: Charles Darwin’s theory of natural selection relies on two components: 1) random variation (whatever the source of such variation, genetic or otherwise); 2) selection of organisms based on differential survival and reproduction. How do we explain convergence if variation is truly random?
Isn’t it very unlikely in the vast majority of cases that truly random variation would lead to the generation of similar traits, regardless of whether those traits are selected? Are there factors limiting (canalizing) what variations do occur?
JL: Of course, variation isn’t random at all, it is constrained in many ways. No species has ever produced a mutation leading to a wheel, for example. Or, more realistically, why are there no six-legged vertebrates? Because for some reason, vertebrates seem unable to produce three sets of independent limbs.
What biologists mean when they say variation is random is that it is random with respect to fitness in the environment in which a species occurs: mutations don’t occur when they are particularly useful. So, this is really a nonquestion: Variation occurs randomly with respect to environmental conditions, but species living in similar environments experience similar selective pressures, and thus often evolve similar adaptations.
TH: You have studied the evolution of anole lizards on Caribbean islands for decades now and have found surprisingly similar lizard evolutionary responses to similar conditions on different islands, which is now a classic case of convergent evolution.
Your book is largely about the debate between convergence as the broad theme of evolutionary change and Gould’s “contingency” school of thought that sees evolution as fundamentally unpredictable and not convergent. Are we at the point now where we can soon quantify the case for either contingency or convergence to provide a good answer to this longstanding debate?
JL: We can certainly enumerate long lists of convergent evolution, and also long lists of nonconvergence. But quantifying the two is much harder. As Rich Lenski recently pointed out in a paper in the American Naturalist, we don’t know the denominator. That is, we can count the number of times that species experiencing a similar environment have evolved a particular feature, but we don’t know how many other species have experienced that environment, but failed to evolve that adaptation.
So, in that sense, we can’t really quantify convergence and non-convergence.
TH: If that’s the case how can we form a solid conclusion about the prevalence of convergence vs. divergence? It seems to me, as a layperson with a strong biology background, that convergence is the more special case, remarkable because it is more rare, and in some cases (such as, for example, the Tasmanian marsupial wolf and the placental wolves) truly remarkable for the degree to which certain traits have evolved highly convergently.
JL: Well, that’s the traditional view, that convergence can lead to remarkably similar species, but that it is quite rare. The latter conclusion has been challenged by Simon Conway Morris and others, who have shown that convergence is a lot more common than we used to realize.
TH: Why does the debate over “contingency” (Gould’s school of thought on evolution) and convergence (Conway Morris’ school of thought) matter in biology? To laypeople?
JL: Simply put, has the course of evolution been haphazard and contingent, or was the outcome predestined by Earth’s environment. Were we — the human species, or something very similar to us — destined to have evolved, or are we a cosmic fluke, the happy accidental result of happenstance?
TH: Even with your clarification above of what “random” evolution of traits means, isn’t it still the case that genetic variations are thought to be random in terms of where they occur in the genome, since they are either simply transcription errors during replication or mutations from cosmic rays, etc.?
And if that’s the case, isn’t it still somewhat (highly?) mysterious how we have witnessed so many cases of convergent evolution? Do you find the “facilitated variation” hypothesis of Kirschner and Gerhart, described in their book The Plausibility of Life, as, well, plausible in terms of explaining at least some cases of convergence?
JL: No, it’s very clear that mutations are more likely to occur in some parts of the genome — sometimes called “mutational hotspots” than in others. More generally, it’s really not that surprising that species independently produce similar mutational phenotypes, sometimes by experiencing the same mutations, sometimes by producing very different mutations that have the same phenotypic effect.
Given enough time and enough individuals, it’s not surprising that such similar variation is produced. And the prevalence of convergence shows that in many cases, developmental constraints do not prevent the species from evolving the same adaptive change.
TH: Given that much of your discussion of the debate between convergence and divergence in evolution relies on long-term controlled lab experiments on bacteria in highly artificial environments, and much less experimental evidence in the wild — your own work on anoles being a large exception to this rule — how confident are you that the lab evidence in this context can provide good insights into the dynamics of natural selection in the wild?
JL: I’m not confident, but that’s all we’ve got. We need more studies, both in the lab and particularly in the field.
TH: Similarly, are there a lot of “natural selection in the wild” experiments ongoing now or upcoming, in order to flesh out this body of evidence?
JL: Yes, there’s a lot more of this type of work now underway.
TH: At the risk of belaboring this point, can we drill down a little more on the origins of phenotypic variation? A new paper looking at cichlid evolution states: “We know relatively little about their [phenotypic novelties] genetic origins.”
And your colleagues at Harvard and UC Berkeley state in the 2006 book The Plausibility of Life: “The question unanswered by the two well-established pillars of evolutionary theory (selection and heredity) is whether, given the rate and nature of changes in the DNA, enough of the right kind of phenotypic variation will occur to allow selection to do its work, powering complex evolutionary change. If the organism were a machine, like Paley’s watch, we would expect that random alterations either would have little effect or would lead to catastrophic failure. We would not expect random change to cause the clock to run more accurately or to develop new features, such as a snooze alarm!”
Are you saying, instead, that this should be viewed as a non-issue, that the notion of mutational hotspots makes this a nonmystery?
JL: I think this is a nonissue, though I don’t think that mutational hotspots are necessarily the reason. I’m not exactly sure what the question is, but the power of artificial selection to successfully select for almost any trait and get an evolutionary response indicates the power of selection and that variation is abundant.
In addition, the now many documented cases of rapid evolutionary response to changing conditions again shows that lack of appropriate variation is not a problem. Most mutations probably are neutral or detrimental, as the quote suggests, but that doesn’t mean that the occasional beneficial mutation doesn’t occur. And, of course, although mutation may in some sense be random, natural selection is very much not a random process.
TH: Does the idea of phenotypic plasticity answer this question? That is, do you agree that the vast course of evolutionary time has led to certain traits evolving, lost phenotypically but stored nevertheless in the, and then brought back into the phenotype as environmental changes demand?
JL: That’s certainly possible, but my guess is that relatively few examples of convergent evolution are the result of this sort of phenomenon.
TH: You discuss this critique of your work in your book a little, in response to persistent claims that the adaptive radiations you’ve identified could be the result of phenotypic plasticity rather than novel evolution of traits. You actually conducted some experiments on phenotypic plasticity with anoles and concluded that there was indeed some plasticity responsible for the observed evolutionary change, but that it couldn’t explain all of the effect, thus genetic evolution had still occurred.
You state at the end of this discussion, however, that there are still a number of question marks here and that more study is required, made more difficulty by the numerous Caribbean hurricanes mucking up your experiments over the years! So where are you on this question now: do you still feel that novel genetic differences explain the lion’s share of observed convergent adaptations or do you think plasticity may still play a large(r) role in this explanation?
JL: As I said in the book, the extent of plasticity documented in empirical studies is not of sufficient scale to explain the great differences observed between species specialized to use different habitats. I find it extremely unlikely that we will discover that the morphological differences among habitat specialists are the result of phenotypic plasticity (i.e., the different phenotypes that can be produced by the same genotype when exposed to different conditions) rather than genetic differences.
We are just on the cusp of having in-depth genome-scale data for about a dozen anole species. I think we soon will know a great deal more about the genetics of anole adaptive radiation.
— Tam Hunt is a lawyer and writer, and creator of the new Forever Young? blog on all things related to anti-aging.