UPDATE: Dr. Shapiro sent me a response to my critique below and asked me to post it; I have done so in the comments (here), and have written a brief response directly below it.
Jim Shapiro is a professor in Biochemistry and Molecular Biology here at the University of Chicago, so I suppose he could be considered a colleague, though I’ve had almost no interaction with him. I have, however, followed his activities in the literature—and with some dismay. Shapiro, it seems, has devoted much of his writing to pointing out that the modern theory of evolution (“neo-Darwinism”) is deeply flawed and needs a new paradigm.
That, I think, is the thesis of his book, Evolution: A View from the 21st Century, which was published in June of last year. I haven’t read it, but Shapiro presents a summary of its thesis in a piece at the HuffPo science page, “What is the key to a realistic theory of evolution?” His thesis is that, contrary to Darwin’s views, and those of modern evolutionary biologists, evolution is not as “gradual” as we think:
Darwin put it this way in Chapter 6: “If it could be demonstrated that any complex organ existed, which could not possibly have been formed by numerous, successive, slight modifications, my theory would absolutely break down. But I can find no such case.”
There has always been controversy about whether random variation and natural selection for improved fitness can truly explain biological evolution over time. Today we can apply genome sequence data to test Darwin’s theory. It answers clearly about gradualism.
Many genome changes at key stages of evolution have been neither small nor gradual.
Now what one means by “gradual” is a matter of taste. Nearly all of us agree that large changes in phenotype—the complex of traits that distinguish, say, genera or families, don’t occur instantaneously but take thousands to millions of years to accomplish. Sometimes these changes are faster than others (this is the basis of the “pattern” part of punctuated equilibrium theory), but they never happen overnight. And we know from the fossil record that gradualism is often the case: the transformation of land-dwelling artiodactyls into whales, for example, took about ten million years, and the transformation of reptiles into modern mammals took three or four times that span. We know this because we see the gradual transformation of fossils. Theory also tells us that such profound evolution can’t happen instantly: it’s unlikely to see a single mutation that would, for example, transform a theropod dinosaur into a bird (many traits must change before that happens), and the coordinated change in many traits that differentiate these groups takes time, as we must await the buildup of many mutations. There’s also a limit to how fast selection can transform populations.
I’m comfortable with saying, then, that big changes in form in evolution are—and must be—gradual, with “gradual” meaning tens of thousands to millions of years.
Shapiro, however, points out that there are instances of more rapid evolution:
- Polyploidy and other forms of speciation involving hybridization. Sometimes different species in nature can hybridize, and, though the hybrids may be sterile, they can undergo doubling of the hybrid set of chromosomes to produce a fertile “allotetraploid” species that is fertile because each species’ chromosomes can pair during gamete formation. This has happened many times in nature (we estimate that 2-7% of all flowering plants arose this way), and it’s very quick: about 2 generations. We have actually seen this occur in a human lifetime in nature: this is described in WEIT. New species can also arise in this way by doubling of the chromosomes of a single species, forming an “autotetraploid”. Or, sometimes, interspecies hybrids can form new species without doubling their chromosomes; their genes sort out in a way that makes them reproductively isolated (and morphologically different) from the two parental species. This process, called “diploid hybrid speciation,” occurs relatively quickly as well: I believe that some diploid hybrid sunflowers arose in about 50 generations. Darwin, of course, didn’t know about this because the genetics wasn’t worked out till much later. Polyploidy is a rapid form of evolution and speciation, one that is fairly common in plants, but very rare in animals. (The reason for its rarity in animals isn’t understood, but we discuss the theories in the book I wrote with Allen Orr, Speciation.
Many domesticated species arose by either deliberate hybridization by humans or selection on naturally-occurring hybrids. Wheat, for example, cited by Shapiro as an example of near-instantaneous evolution, did arise by humans picking out a naturally-occurring hybrid; this probably happened twice, so that modern wheat contains the genomes from three wild species of grass. But much of the selection that makes for modern wheat was also artificial: humans picked out the variants with bigger heads that remained on the plant. So the differences between wild and cultivated wheat don’t completely exemplify near-instantaneous evolution in nature.
- Symbiosis. We all know, thanks to Lynn Margulis, that the evolution of the eukaryotic cell involved two rapid evolutionary events: the acquisition of mitochondria via the ingestion of one bacterium by another cell, and, in plants, the origin of chloroplasts via a similar route. Centrioles (a group of microtubules involved in cell division) may have originated via symbiosis. And some species, like lichens, are actually a mixture of two distinct species—in the case of lichens, an alga (or cyanobacterium) and a fungus. That fusion probably happened quickly as well.
Margulis theorized that symbiosis was not only important in evolution, but ubiquitous, involved in nearly all cases of speciation and macroevolution. She was wrong. We know now that the rapid origin of new taxa by symbiosis, while critical for some evolutionary transformations, is rare. It can hardly be used to discount the notion that “Darwinian” evolution is usually gradual.
- Acquisition of DNA from other species. This uptake of DNA between species, also called “horizontal gene transfer” (HGT) happens sometimes, and although rare (especially in animals), may be more common than we think. I regard it as a form of mutation, since the acquired gene is, after acquisition, subject to evolution via natural selection. Given its rarity (if it were common we wouldn’t be able to make good DNA-based trees of various species) and the subsequent modification of the acquired DNA by natural selection, it’s doubtful whether HGT qualifies as “instantaneous evolution.”
- Genomic changes other than slight changes in DNA sequence. Shapiro cites several of these which have been involved in the evolution of new genes, including gene duplication (a whole gene gets copied in its entirety, which can lead to the origin of “gene families”) and the formation of new genes by pasting together bits of old ones (this is the topic of research by my Chicago colleague Manyuan Long). But while these pheomena are “rapid” on the DNA level, they’re not necessarily a cause of rapid evolution on the organismal, biochemical, or even genomic level. After all, those new genes, particularly the duplicated ones, must undergo their own gradual divergence via natural selection: that’s how the different opsins in our eyes, the divergent genes in our immune system, and the different types of human hemoglobin arose. That differentiation took a long time, because the new genes have to await mutations.
So yes, evolution can be almost instantaneous in the case of symbiosis and polyploidy, and there are mechanisms of mutation not dreamt of in the genetics of Watson and Crick. But does this invalidate the paradigm of neo-Darwinism? Shapiro thinks it does:
Was Darwin simply mistaken about the gradual nature of hereditary variation? Such ignorance would be unavoidable before we knew about Mendelian genetics and DNA. Or was there a deeper flaw in the theory that he (and Alfred Russell Wallace) propounded? The answer may well be that it was a basic mistake to think that optimizing fitness is the source of biological diversity.
Here he’s completely wrong. Yes, evolution can sometimes be rapid, though most of the time it’s gradual, but in all cases—all cases—adaptive evolution occurs by natural selection, and that means “optimizing fitness.” Shapiro is getting Darwinism wrong here, for he’s conflating the materials of evolutionary diversity (mutations and new genes that arise randomly) with their disposition by natural selection, a nonrandom process that does involve “optimizing fitness.” Given that adaptive differentiation between species involves natural selection, you simply cannot say that “optimizing fitness’ is not a part of biological diversity. That attack on neo-Darwinism is misguided, and it’s wrong to suggest it to lay readers.
Shapiro’s piece then rapidly goes downhill as he starts repeating creationist arguments. Here’s one:
The first problem with selection as the source of diversity is that selection by humans, the subject of Darwin’s opening chapter, modifies existing traits but does not produce new traits or new species. Dogs may vary widely as a result of selective breeding, but they always remain dogs.
You’ll recognize this as the old creationist canard. Yes, of course we can’t turn a dog into a cat by artificial selection, because that would take millions of years, and we’ve only been selecting on dogs for a couple of thousand years. But the true refutation of this idea is in the fossil record: we can see land-living artiodactyls (resembling small deer) turning into whales, we can see fish turning into amphibians, we can see early reptiles turning into mammals, we can see theropod dinosaurs turning into birds, and we can see our apelike ancestors turning into more modern humans. In other words, we find in fossils precisely those transformations that Shapiro says are impossible. I deplore that a colleague of mine makes this misguided argument, and in the Science section of HuffPo, which I’m increasingly beginning to deplore as well.
Shapiro also goes after the neo-Darwinian idea that mutations are random:
The second problem is that Darwin understood only “numerous, successive, slight modifications” as the sources of inherited change. His neo-Darwinian followers have modified this position to assert that all mutations occur randomly. They insist there is no biological input into the change process. For them, the genome determines organism characteristics. They think of it as a read-only memory (ROM), which only changes by accident.
Well, he’s wrong again here. What we mean when we say that mutations are “random” is not that there’s no biological input into the mutation process (that would be a truly stupid assertion), but that the biological inputs into diversity occur irrespective of whether they would be useful to the organism. Mutational change occurs by accident, or as a byproduct of something else (like a gene being accidentally duplicated, or the ingestion of DNA from another species), but those changes occur whether or not they’d be “good” (i.e., increase the reproductive output of) individuals in the species that has mutated. Gene shuffling, gene duplication, horizontal gene transfer: all of these are biological accidents that are just the precursors of adaptive differentiation—the fuel of evolutionary change. They’re the gas that fills the tank of the car, which moves by natural selection. Adaptive differentiation still requires the process of natural selection: the nonrandom differential reproduction of genes (and individuals), usually involving greater success in the environment.
While Shapiro rightfully points out that there are more sources of genetic variation than were dreamt of in your genetics, Watson and Crick, he is misguided in thinking that these new sources of variation completely destroy the Darwinian idea of gradualism and natural selection. We see many examples of gradual evolutionary change in the fossil record, and natural selection is still the only game in town when in comes to explaining adaptations. If Shapiro knows any other way that adaptation can occur, let him tell us.
Neo-Darwinism has certainly expanded since it began to form in the 1930s, and it should: evolution is a vibrant science with new discoveries occurring weekly. But it’s simply wrong to think that those new discoveries have invalidated (or even caused a serious reassessment of) the major tenets of neo-Darwinism. That theory is regularly pronounced dead—as it is here by Shapiro—but it refuses to lie down.