I’m a bit late on this one, but the Albatross has kept me occupied. This post will be of interest mainly to science buffs, particularly those who already know a bit about evolution. But I’ll weigh in anyway, for, like an egg-eating snake expelling the shell, I had to get this out.
On October 9, the journal Nature published a longish comment by two groups of investigators called “Does evolutionary theory need a rethink?“. (Reference below; I believe the pdf is free. If not, you can get one by judicious inquiry.) It’s a “Point/Counterpoint” in which one group of evolutionists (whose part is called “Yes, urgently,”), suggests that modern evolutionary theory needs a rethink, and will be replaced by something quite different, while the other group (“No, all is well”) maintains that the “revolutionary” discoveries fit comfortably within the existing evolutionary paradigm, so no drastic overhaul is needed.
I read the “Yes, urgently” part first and decided to respond here without having read the other part, as I wanted to critique their views without being influenced by the “No, all is well” side.
Let me say first that I’m a bit puzzled by the continual appearance of these “Does evolution need a revolution?” pieces. If our field really was undergoing a revolution, we wouldn’t have to debate it. I doubt, for instance, that when there was a genuine paradigm shift in physics—from classical to quantum mechanics—we saw many physicists writing “Does physics need a rethink?” articles. The answer was obviously “yes.”
But even among those who see a paradigm shift in evolution, there’s nobody who sees anything like a complete overturning of our worldview, as happened when quantum mechanics appeared as a deeper and weirder supplement to classical mechanics. Although I haven’t read the “all is well” side, I agree with their conclusion as expressed in the title, and would probably agree with their arguments, which I’ll read after I write this. For the “revolutionary” phenomena touted by the “yes, urgently” side are either not new, fit comfortably within the modern view of evolution, are limited in scope, or, even if fairly frequent, wouldn’t cause a paradigm shift. By paradigm shift, I mean the view of evolution as gradual, based on variation in DNA sequences that change by either selection or genetic drift (or a few rarer processes), often propelled by natural selection, and producing branches—new species—that yield reproductively isolated populations that cannot interbreed (that’s what species are).
It’s also telling that nearly all the authors calling for an “urgent rethink” of evolutionary theory are those who have published or proposed the “revolutionary” ideas that motivate their views. And they severely overrate the nature of the scientific discussion going on:
This is no storm in an academic tearoom, it is a struggle for the very soul of the discipline.
“Struggle for the soul of our discipline”? That seems a bit dramatic and self-aggrandizing, and is simply untrue. There is no such “struggle” going on, except, perhaps, in the minds of those who feel that their work is not sufficiently appreciated or advertised.
That aside, let me discuss briefly the new phenomena that, the “yes” authors say, call for a new paradigm, an overthrow of what they call “standard evolutionary theory” (SET). There are four, which I’ll take in order:
1. The evolution of development (“evo devo”). The “yessers” claim that developmental biology was never properly incorporated into SET, and would change it drastically were this to happen. They further argue that evo devo has shown that some developmental pathways are more likely to evolve than others:
In our view, this concept — developmental bias — helps to explain how organisms adapt to their environments and diversify into many different species. For example, cichlid fishes in Lake Malawi are more closely related to other cichlids in Lake Malawi than to those in Lake Tanganyika, but species in both lakes have strikingly similar body shapes. In each case, some fish have large fleshy lips, others protrud- ing foreheads, and still others short, robust lower jaws.
SET explains such parallels as convergent evolution: similar environmental conditions select for random genetic variation with equivalent results. This account requires extraordinary coincidence to explain the multiple parallel forms that evolved independently in each lake. A more succinct hypothesis is that developmental bias and natural selection work together. Rather than selection being free to traverse across any physical possibility, it is guided along specific routes opened up by the processes of development.
I’m not sure why the authors feel that “developmental bias” is a more “succinct” hypothesis than convergent evolution. In fact, I see it the other way around. While some developmental pathways surely are easier to evolve than others, the remarkable plasticity of animals and plants under selection suggests that parallel selection may be a more important cause of convergent evolution. After all, icthyosaurs (reptiles), porpoises (mammals) and fish (fish) all evolved similar streamlined shapes independently. It’s more “succinct” to see this as groups of unrelated organisms responding to an environmental challenge (a watery milieu) by using different genes to produce similar shapes than by invoking similar developmental channels. After all, fish, reptiles, and mammals are only distantly related, and it seems unlikely that these shape changes reflect the constraints of development. (In fact, it would be hard to see them as anything other than selection having acted on different developmental pathways.)
And the same may be true for the cichlids in Lake Malawi and Lake Tanganyika. Why would one think that the repeated paths of evolution, which we also see in Australian marsupials versus distantly related placental mammals (both have “flying squirrels,” “moles,” and other similar forms) reflect similar developmental biases? There’s no evidence for developmental channelling here, but there’s evidence from many fronts (e.g., the remarkable plasticity of species under artificial selection) that animals have genetic variation to change into almost anything you want.
Evo devo is a fascinating field, and has come up with some stunning results: one is the discovery that developmental switches can be similar for traits even in distantly related organisms, like the Pax-6 gene controlling eye formation in flies and mammals. But that doesn’t argue for developmental channelling of entire phenotypes, for the eyes involve many different genes in mice and flies. There is still no good empirical evidence for “convergence” due to similarity of developmental pathways that are constrained.
2. Developmental plasticity. This is the notion that a single organism can change its appearance (phenotype) or physiology in different environments. Mammals may grow longer fur, or change their color from brown to white, when the weather gets colder and snowier; plants can grow toward the sun, or change their leaf shapes depending on how much sun they get; the two different claws of the lobster (crushing versus pinching) develop differently depending on which claw grabs an object first.
This is nothing new, for that plasticity is often adaptive and has evolved by natural selection. Those mammals who had genes for changing coat color in winter left more offspring (they were hidden from predators or prey), and those plants that could change their leaf shape or direction to catch the sun would photosynthesize more.
But the “revolution” advocates also propose another form of plasticity: an adaptive change in an organism is caused by phenotypes alone, with the genetic change lagging well behind:
If selection preserves genetic variants that respond effectively when conditions change, then adaptation largely occurs by accumulation of genetic variations that stabilize a trait after its first appearance. In other words, often it is the trait that comes first; genes that cement it follow, sometimes several generations later.
But all this really is is what we call “genetic assimilation”: that those traits that prove adaptive in a new environment, though perhaps a result of a malleable appearance, still have genes underlying them, and it is the accumulation of those genes that cause evolution.
Let me give an example. Suppose there are some Tiktaalik-like fishapods swimming near the shore. A few individuals venture out onto the land, as they show “behavioral plasticity” and are adventurous. It turns out that getting on the shore gives you all kinds of new food, particularly insects. They leave more offspring. Over time, the fishapod becomes a proto-amphibian. This is, in fact, the way that terrestrial tetrapods may have evolved.
But this is nothing new: in fact, it was suggested by Ernst Mayr in his famous 1963 book Animal Species and Evolution. Mayr said that many drastic changes in lifestyle may have begun with simple behavioral plasticity.
But the important thing to recognize is that behavioral change and its sequelae (like all the other adaptations for living on land) cannot evolve unless those changes have a genetic basis. What we see here is simply phenotypic (trait) variation that has some underlying genetic basis, and proves to be adaptive. The genetic changes accumulate, and eventually we get a big change in form, lifestyle, and so on. That’s simply conventional natural selection, not a revolution in SET. Further, we don’t know how often major changes in lifestyle happen this way. But whatever the case, this is not a paradigm shift.
There are other theories that the changes in phenotype are adaptive but have nothing to do with genetic variation, and the genes simply come along later to somehow “stabilize” the phenotype. This isn’t likely because it’s not obvious how lifestyle or form changes could evolve in such a way, since the initial changes would be lack any genetic underpinning. There’s one possible case involving loss of eyes in cave fish, but until that phenomenon is shown to be frequent, we can’t use it to tout a “revolution.”
3. “Nongenetic” forms of evolution. If evolution really weren’t based on heritable and permanent changes in DNA sequence, that would be surprising, and at least a major change in perspective. The “revolution” proponents argue that this does happen in two ways.
First, there is cultural evolution: stuff is passed on not by genes, but by learning. This, of course, is nothing new: Dawkins wrote about memes—units of cultural inheritance—way back in 1976, drawing a parallel between genetic and cultural evolution. But that was a parallel, and one that I don’t find terribly enlightening. But cultural inheritance is of course important in some species, including all animals that teach their young. The authors give some examples:
In addition, extra-genetic inheritance includes socially transmitted behaviour in animals, such as nut cracking in chimpanzees or the migratory patterns of reef fishes.
So what’s new? Yes, we can model how this works, but learning it itself an evolved ability, and modeling social evolution will involve things beyond the purview of evolutionary theory. Cultural evolution is not genetic evolution, and hence not part of the SET, which rests on changes in genes. Cultural evolution is important, but it’s no more part of SET than is the “evolution” of changes in automobile style over the years.
The “revolution” authors also include epigenetics as an important component of nongenetic inheritance, one that will revolutionize evolution. By “epigenetics,” they mean environmentally induced changes in the DNA (e.g., methylation of DNA bases) that somehow become coded in the DNA, so that the environment by itself can change the genome and eventually produce adaptive evolution.
While adaptive methylation has been known for a while (male versus female DNA in zygotes is often differentially methylated, and in ways that favor one parent’s genes), all of this adaptive methylation depends on changes in the DNA code itself—changes that tell the DNA to let itself be methylated. That’s different from the new proposal, which claims that such changes aren’t initially coded in the genes, but directly produced by the environment. (This is “Lamarckian” evolution for those of you who know what that means.)
The problem with this is that such cases of environmental changes in DNA are always temporary, for they’re not coded in the DNA and thus cannot persist forever. And if they’re temporary, they cannot cause long-term adaptive evolution. In fact, there is not a single known case of any new organismal trait based on environmentally-induced change of DNA that has persisted for more than a few generations. And we know of no adaptive change based on such a process. In contrast, there are lots of cases of evolutionary changes and adaptations based on heritable, non-environmentally-induced changes in DNA—that is, “conventional” changes caused by mutations. In view of this paucity of evidence for environmentally-caused epigenetic change as an evolutionary factor, why are the authors calling for us to overturn SET?
4. “Niche construction.” This is a recent buzzword in evolutionary biology that is an interesting notion, and one that certainly holds true in many cases of evolution. It is the idea that the organism’s own activities modify its environment in a way that changes the direction of natural selection acting on that organism. The classic example is the beaver. These rodents have evolved to build dams and live in those dams, and thereby have modified their environment in a way that affects their subsequent evolution. Due to its own evolution, the beaver now lives in a lake it made itself, and lives inside a house of sticks that it also built. That must surely influence its future evolution, and which mutations could be adaptive (ones promoting better swimming and tree-felling, for instance). Ditto for social insects, who now live in complex burrows (built by them, of course) that must surely affect their later evolution.
While this idea is getting new attention, and deservedly so, it doesn’t call for a revolution in SET. First of all, it’s not particularly new. The idea of “gene-culture” coevolution has been around a long time. One example is pastoralism, in which humans changed their environment by keeping domestic animals that give milk. And that has changed our evolution, for cultures that are pastoral have undergone evolution involving the use of lactose. Genes that break lactose down into digestible components are usually inactivated after weaning in humans, who, over most of our history, didn’t have a source of milk after they stopped suckling. That’s why many of us are “lactose intolerant.” When we suddenly got a rich source of nutrition from our sheep and cows, pastoral cultures evolved so that the genes metabolizing lactose weren’t inactivated,but were turned on for life. (Individuals with genes allowing them to digest milk had up to 10% more offspring on average than intolerant individuals!) Thus, our own culture affected our subsequent evolution. This did not cause us to dismantle SET; rather, it was an interesting sidelight on how culture itself caused genetic change.
Second, we don’t know how pervasive this process is. That is, while many organisms do affect their environments, we don’t know how often that environmental change feeds back to the organism to cause additional evolution. In some cases it probably doesn’t: fish adapt to an unchanging fluid medium, the coat color of polar bears cannot affect their environment of ice or snow, and the hooves of the chamois don’t affect the granitic structure of the Swiss Alps. So how often “niche construction” is important is an open question, albeit an interesting one. But I don’t see it overthrowing SET, for it’s simply a novel way that the environment can change and affect organismal evolution.
That is not to disparage this phenomenon—or any of these phenomena. Niche construction seems more likely to be important than “genes follow phenotype” plasticity, or than adaptive epigenetic evolution, of which we have not a single example. All these ideas deserve empirical study. But none call for a new paradigm.
Now I can read what the “conservatives” (Wray, Hoekstra, Futuyma, Mackay, Lenski, Strassmann, and Schluter) have to say.
Laland, K., T. Uller, M. Feldman, K. Sterelny, G. B. Müller, A. Moczek, E. Jablonka, J. Odling-Smee, G. A. Wray, H. E. Hoekstra, D. J. Futuyma, R. E. Lenski, T. F. C. Mackay, D. Schluter, and J. E. Strassmann. 2014. Does evolutionary theory need a rethink? Nature 514:161-164.