This time I’m not going to discuss Jim Shapiro’s misguided dismissal of natural selection (I’m done with him), but, like the Lernaean hydra, when you cut off one antiselectionist head, another pops up elsewhere. This has just happened at PuffHo, where Stuart Newman, a professor of cell biology and anatomy at New York Medical College, has taken arms in support of Shapiro’s views. (He mentions Shapiro explicitly.)
In his new column in the Science section, “Where do complex organisms come from?“, Newman’s answer is this: “Not from natural selection, but from the self-organizing properties of molecules and tissues.” This is a popular answer among contrarians, creationists, and those who know little about evolution, but it’s wrong. It’s wrong because “self-organization” cannot explain adaptations: those features of organisms which have obviously appeared to aid their survival and reproduction.
I’m sure you’ve noticed that I’m losing patience with those biologists who claim that something is deeply wrong with modern evolutionary theory, which needs either a drastic overhaul or a complete junking. Now, it’s clear that there are new discoveries every day that affect our view of how evolution works (horizontal gene transfer is one, the conservatism of gene regulation across long-diverged taxa another), but I haven’t seen anything that makes me think that natural selection is an outmoded way to think about the evolution of adaptations. And yet that’s what is claimed not only by religious people like Alvin Plantinga and philosophers like Tom Nagel, but also by biologists like James Shapiro and Stuart Newman. I’ll start believing them when they show real problems with the idea of natural selection and, especially, propose a credible process that can explain the evolution of adaptations like mimicry, hearts, and flippers.
I’ve divided Newman’s argument into three parts:
Newman’s first contention: natural selection can explain minor evolutionary changes (“microevolution”), but not big ones (“macroevolution”). Sound familiar? That’s a recurring theme of creationists like Michael Behe and Jon Wells. Newman says this:
While it may be an adequate scenario for the refinement of some already-existing characters — the beaks of finches, color intensity of moths — the “microevolutionary” process envisioned by Darwin and his successors does not account in any plausible way for “macroevolutionary” patterns such as the differences between oysters and grasshoppers, fish and birds. In fact, adaptationist gradualism, though still popular in some scientific circles, is increasingly questioned and found wanting by evolutionary biologists working in an expanded set of disciplines. [JAC: Who are those evolutionists? He doesn’t name any.]
But why not? Newman gives no explanation why selection can’t explain macrovolution. And I see no reason why adaptive differences that characterize major taxa, like phyla, must have evolved by a process different from than that changing the beak sizes of finches. Extrapolated over millions of years, natural selection can wreak huge changes in organisms. The evolution of whales from a land-dwelling artiodactyl (even-toed mammal) took about eight million years; this is documented in the fossil record. That’s a pretty short time for substantial change. Our own evolutionary divergence modern chimps and bonobos wasn’t much shorter—it took about six million years—and that involved far less remodeling of the body plan. As the paper of Nilsson and Pelger on eye evolution demonstrates (reference below), one can evolve complex characters in a remarkably short period of time: about 86,000 for a camera eye in their case.
Newman’s second contention: thanks to the labors of developmental biologists and geneticists, we now know that macroevolutionary changes involve saltations and “niche construction” rather than natural selection.
By incorporating embryonic development and its underlying physico-genetic processes into evolutionary theory, investigators are learning that abrupt alterations in body plan and other aspects of organismal form can occur in response to environmental change or gene mutation in ways that affect multiple members of a population and exhibit consistent patterns of inheritance. Furthermore, there is increasing emphasis on the resourcefulness of organisms and their ability to construct their own niches. Having a “phenotype” (the outward manifestation of biological identity), very different from that of one’s progenitors is no longer considered disqualifying for survival.
“Niche construction” is the idea that organisms, by their own behavior, can change the selection pressures on themselves, and in that way forge their own ecological niche. This is a reasonable idea for some (but not all) cases of adaptation. The classic example is the beaver: by evolving behaviors to cut down trees, dam streams, and build lodges, they have constructed a new environment for themselves (the lake and lodge), which undoubtedly exerts new selection pressures on beaver morphology and behavior.
This idea has occupied biologists recently, and is an intruiguing one. But it is not in any way a replacement for natural selection, any more than is the view that the first “fishapod” that crawled on land suddenly, by that behavioral quirk, exposed itself to a bunch of new selection pressures involving living on land. In fact, the idea that many new life forms begin with a nongenetic change in behavior was suggested by the evolutionist Ernst Mayr in his classic 1963 book, Animal Species and Evolution.
Note that natural selection is still a critical part of “niche construction” theory. It’s not a replacement for natural selection, but the view that, by virtue of their own behavior, animals can expose themselves to forms of natural selection that they wouldn’t experience otherwise.
So how does Newman think that “adaptations” or “macroevolutionary differences” evolve? The paragraph above doesn’t give us much help, for the critical first sentence makes no sense unless you’re a Lamarckian.
Newman’s third contention: macroevolutionary differences result from the self-organizing properties of molecules and tissues. Natural selection is not involved.
By the end of Darwin’s life new physical theories were being put forward to explain abrupt and large-scale changes in such materials, and by extension, the character and transformations of organisms and their organs.
Note how the next paragraph, which follows right after the sentence above, does not explain “by extension” how evolution proceeds. It explains, perhaps, how some mechanisms of development work:
Here is a partial list of late nineteenth and early twentieth-century physical concepts that have proved relevant to developmental processes (with the phenomenon they explain, at least partly, in parentheses): dynamical systems (ability of cells having the same genome to switch between different “types”), phase separation of liquids (capacity of embryonic tissues to form several non-mixing layers), chemical oscillations (propensity of embryonic tissues to organize into tandem segments), “Turing-type” reaction-diffusion systems (the formation in tissues of regularly spaced structures like feather and hair buds, pigment stripes, or the bones of the limb skeleton). All or most of these processes (termed “mesoscale,” being most relevant to objects the size and texture of cell clusters), along with several others, are harnessed and mobilized by the secreted products of specific genes during embryogenesis in every one of the animal phyla (e.g., arthropods, mollusks, nematodes, chordates and so forth).
Yes, maybe a “Turing-type” diffusion system can explain how stripes are formed in zebras or reef fishes (we’re not sure about this yet), but that’s a proximate explanation: how the genome produces a phenotype. It doesn’t provide an ultimate explanation: why does the genome produce that phenotype in the first place? And if the phenotype is an adaptation, like the stripes of zebras may be (again, there are ideas about this, but we’re not sure), then you still have to invoke natural selection as a cause of those “Turing-type” patterns, i.e., the particular evolutionary process that has altered the genes to produce such patterns.
Finally, Newman gives a list of observations about evolution that, he says, can’t be explained by natural selection or conventional evolutionary theory. Here it is:
What can the existence and action of such protean generative processes tell us about the origin of organismal complexity? First, let’s look at some of the expectations of the natural selection-based modern synthesis: (i) the largest differences within given categories of multicellular organisms, the animals or plants, for example, should have appeared gradually, only after exceptionally long periods of evolution; (ii) the extensive genetic changes required to generate such large differences over such vast times would have virtually erased any similarity between the sets of genes coordinating development in the different types of organism; and (iii) evolution of body types and organs should continue indefinitely. Since genetic mutation never ceases, novel organismal forms should constantly be appearing.
All these predictions of the standard model have proved to be incorrect. The actual state of affairs however, are expected outcomes of the “physico-genetic” picture outlined above.
ORLY? Let’s look at Newman’s three contentions that supposedly violate the “standard model” of evolution. On examination, none of them hold up.
1. Gradualism. Modern evolutionary theory does predict that complex adaptations (like the evolution of whales from land-dwelling artiodactyls) can’t occur instantly; they take thousands to millions of years. And that’s what we see in the fossil record. Remember, the Cambrian explosion, which produced many phyla still extant today, was not “instant” but probably took between 10 and 50 million years. That is a long time—certainly a time during which selection, if it were strong (as many paleobiologists think), could produce diverse phyla. Just think of all the changes that humans have wrought via strong artificial selection in plants and animals in the last 10,000 years, and that’s only about .03% of the duration of the Cambrian explosion.
Now there is a strain of thought in modern evolutionary biology that evolutionary change can proceed more rapidly and jerkily than people like Darwin thought—that it need not involve a gradual and insensible change in form over millions of years. It can be faster than that, with change sometimes not occurring at all. Indeed, Allen Orr and I were one of the first people who suggested this possibility (Orr and Coyne 1992; reference below), and of course rapid and sporadic evolutionary change was an important part of Eldredge and Gould’s theory of punctuated equilibrium.
Perfect gradualism—the continual and insensible change of phenotype due to changes in many genetic factors—is no longer a tenet of evolutionary biology, though a modified form of gradualism still is (“complex adaptations take time; they don’t occur instantly or within a few generations”). But we’ve seen no adaptive changes in the fossil record or otherwise that force us to reject natural selection as the causal process. A million years is a long time! And we have no evidence that any complex feature evolved instantly (by “instantly” I mean, say, a thousand years).
2. Conservatism of genes. Newman thinks that the long periods of evolutionary time separating major groups predict that there would be no similarity between the genes controlling development among these groups. Yet genes like the Hox genes are generally conserved among phyla separated by hundreds of millions of years. The gene Pax6, for example, controls eye development in both mice and fruit flies, groups separated by nearly 800 million years of evolution. But does this refute natural selection or any tenet of evolution? Certainly not! If a gene is useful in recruiting other genes to produce a feature, natural selection in the form of “stabilizing selection” (selection retaining genes that are useful for something), can maintain them. And Newman doesn’t mention that Hox genes are the exception: the genetic differences between flies and mice involve many more drastically changed or new genes than conserved ones.
3. Evolutionary theory predicts that organisms should evolve indefinitely: “novel evolutionary forms should constantly be appearing.” This is Newman’s most ridiculous “refutation” of evolutionary theory. Whether natural selection causes adaptive change depends on two things: whether a change would facilitate the organism’s reproduction, and whether there is genetic variation for that change. While genetic variation for most traits is ubiquitous, it’s not always true that it’s in the organism’s best “interest” (I’m speaking in shorthand here) to change. Organisms like deep-sea fish, for example, may be well adapted to their environment, and that environment might not change much over time! So why on earth would they constantly be spawning new forms?
Second, it’s certain that a lot of evolutionary change creating “novel forms” is happening, but it’s simply too slow for us to see. The kea of New Zealand may be evolving into the world’s only carnivorous parrot. The hippo may be becoming fully aquatic, and in a few million years will no longer be tied to the land, but evolve into a sort of freshwater whale, like the manatee. If Newman had seen Indohyus 48 million years ago, he’d have said “Why aren’t new forms developing?” But it took the small deerlike Indohyus 8 million years to evolve into modern whales. Humans have been categorizing life for only a few thousand years; evolution takes millions. It’s simply dumb to say that “novel evolutionary forms aren’t appearing.” And does Newman know what is happening in organisms that live in the deep sea, or underground? Where does he get the notion that the production of novel species and taxa has simply stopped?
In short, none of Newman’s three “observations” mandate that we toss modern evolutionary biology on the scrap heap along with the idea of natural selection.
Further, there are predictions that natural selection makes which Newman’s “physiochemical” explanation doesn’t, and these predictions are met. Here are three:
1. If there is no genetic variation, there will be no evolution, for both natural selection and genetic drift require variation for evolutionary change. And this is met: inbred or highly genetically uniform species are resistant to artificial and natural selection.
2. “True” genetically-based altruism, in which genes mandating that behavior sacrifice their ability to replicate (i.e., organisms do something without their genes getting anything in return) should be nonexistent. (“Reciprocal altruism”, like blood regurgitation in vampire bats, doesn’t count, as the organism that sacrifices also reaps benefits.) Indeed, I know of no cases of true altruism in animals outside of humans, where it’s almost certainly a cultural rather than genetically hardwired behavior. Chalk another one up for natural selection.
3. We should not see “adaptations” in one species that are useful only for a second species. One example I use is the presence of nipples on one animal (say, a wild pig) that can be used only to suckle young of another species, say a wild deer. Natural selection predicts that that can’t happen, and we never see such things. In cases where species have “adaptations” that help another species, natural selection predicts that the first species should benefit too. And that is what we see in the many cases of mutualisms like cleaner fish and their “cleanees” or symbiotic flagellates in the gut of termites.
Newman’s final paragraph, which sums up his thoughts, seems impenetrable and garbled to me:
With a 19th century notion of incremental material transformations no longer relevant to comprehending the range of organismal variation that has appeared throughout the history of life on Earth, the other pillar of the standard model can be discarded along with it. Specifically, if, as affirmed by niche construction theory, phenotypically novel animals or plants can invent new modes of existence in novel settings, rather than succumbing to a struggle for survival in the niches of their origin, there is no need for cycles of selection for marginal adaptive advantage to be the default explanation for macroevolutionary change.
Yes, niche construction can create new forms of selection, but by itself does not and cannot create new genetically-based adaptive change. And niche construction can’t always work, because some animal behaviors simply cannot change important aspects of their environment. Does the color of a polar bear’s coat affect the reflective properties of ice and snow? Does the shape of a chamois’s hoof affect the granitic structure of the Swiss Alps? Does the shape of a fish affect the hydrodynamic properties of water? In many cases animals simply must adapt to static and unchanging features of their environment.
At any rate, Newman’s argument for discarding modern evolutionary theory is, like Shapiro’s argument, totally unconvincing. And it irks me no end that PuffHo continues to give voice to such people, misleading the public about the solidity of evolutionary theory.
Nilsson, D.-E., and S. Pelger. 1994. A pessimistic estimate of the time required for an eye to evolve. Proc. Roy. Soc. Lond. B 256:53-58.
Orr, H. A., and J. A. Coyne. 1992. The genetics of adaptation: a reassessment. Amer. Natur. 140:725-742.